2013-03-25 22:49:35 +08:00
|
|
|
/*
|
|
|
|
|
* Copyright (c) 2013 Nicira, Inc.
|
|
|
|
|
*
|
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
|
* modify it under the terms of version 2 of the GNU General Public
|
|
|
|
|
* License as published by the Free Software Foundation.
|
|
|
|
|
*
|
|
|
|
|
* This program is distributed in the hope that it will be useful, but
|
|
|
|
|
* WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
|
* General Public License for more details.
|
|
|
|
|
*
|
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
|
|
|
|
|
* 02110-1301, USA
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
|
|
|
|
|
|
|
|
|
|
#include <linux/capability.h>
|
|
|
|
|
#include <linux/module.h>
|
|
|
|
|
#include <linux/types.h>
|
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
|
#include <linux/slab.h>
|
|
|
|
|
#include <linux/uaccess.h>
|
|
|
|
|
#include <linux/skbuff.h>
|
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
|
#include <linux/in.h>
|
|
|
|
|
#include <linux/tcp.h>
|
|
|
|
|
#include <linux/udp.h>
|
|
|
|
|
#include <linux/if_arp.h>
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
|
#include <linux/in6.h>
|
|
|
|
|
#include <linux/inetdevice.h>
|
|
|
|
|
#include <linux/igmp.h>
|
|
|
|
|
#include <linux/netfilter_ipv4.h>
|
|
|
|
|
#include <linux/etherdevice.h>
|
|
|
|
|
#include <linux/if_ether.h>
|
|
|
|
|
#include <linux/if_vlan.h>
|
|
|
|
|
#include <linux/rculist.h>
|
2014-01-27 14:43:57 +08:00
|
|
|
#include <linux/err.h>
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
#include <net/sock.h>
|
|
|
|
|
#include <net/ip.h>
|
|
|
|
|
#include <net/icmp.h>
|
|
|
|
|
#include <net/protocol.h>
|
|
|
|
|
#include <net/ip_tunnels.h>
|
|
|
|
|
#include <net/arp.h>
|
|
|
|
|
#include <net/checksum.h>
|
|
|
|
|
#include <net/dsfield.h>
|
|
|
|
|
#include <net/inet_ecn.h>
|
|
|
|
|
#include <net/xfrm.h>
|
|
|
|
|
#include <net/net_namespace.h>
|
|
|
|
|
#include <net/netns/generic.h>
|
|
|
|
|
#include <net/rtnetlink.h>
|
2014-09-18 03:25:58 +08:00
|
|
|
#include <net/udp.h>
|
2016-09-16 04:00:29 +08:00
|
|
|
#include <net/dst_metadata.h>
|
2014-11-05 01:06:51 +08:00
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
|
#include <net/ipv6.h>
|
|
|
|
|
#include <net/ip6_fib.h>
|
|
|
|
|
#include <net/ip6_route.h>
|
|
|
|
|
#endif
|
|
|
|
|
|
2014-01-19 16:43:42 +08:00
|
|
|
static unsigned int ip_tunnel_hash(__be32 key, __be32 remote)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
|
|
|
|
return hash_32((__force u32)key ^ (__force u32)remote,
|
|
|
|
|
IP_TNL_HASH_BITS);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static bool ip_tunnel_key_match(const struct ip_tunnel_parm *p,
|
|
|
|
|
__be16 flags, __be32 key)
|
|
|
|
|
{
|
|
|
|
|
if (p->i_flags & TUNNEL_KEY) {
|
|
|
|
|
if (flags & TUNNEL_KEY)
|
|
|
|
|
return key == p->i_key;
|
|
|
|
|
else
|
|
|
|
|
/* key expected, none present */
|
|
|
|
|
return false;
|
|
|
|
|
} else
|
|
|
|
|
return !(flags & TUNNEL_KEY);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Fallback tunnel: no source, no destination, no key, no options
|
|
|
|
|
|
|
|
|
|
Tunnel hash table:
|
|
|
|
|
We require exact key match i.e. if a key is present in packet
|
|
|
|
|
it will match only tunnel with the same key; if it is not present,
|
|
|
|
|
it will match only keyless tunnel.
|
|
|
|
|
|
|
|
|
|
All keysless packets, if not matched configured keyless tunnels
|
|
|
|
|
will match fallback tunnel.
|
|
|
|
|
Given src, dst and key, find appropriate for input tunnel.
|
|
|
|
|
*/
|
|
|
|
|
struct ip_tunnel *ip_tunnel_lookup(struct ip_tunnel_net *itn,
|
|
|
|
|
int link, __be16 flags,
|
|
|
|
|
__be32 remote, __be32 local,
|
|
|
|
|
__be32 key)
|
|
|
|
|
{
|
|
|
|
|
unsigned int hash;
|
|
|
|
|
struct ip_tunnel *t, *cand = NULL;
|
|
|
|
|
struct hlist_head *head;
|
|
|
|
|
|
2014-01-19 16:43:42 +08:00
|
|
|
hash = ip_tunnel_hash(key, remote);
|
2013-03-25 22:49:35 +08:00
|
|
|
head = &itn->tunnels[hash];
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(t, head, hash_node) {
|
|
|
|
|
if (local != t->parms.iph.saddr ||
|
|
|
|
|
remote != t->parms.iph.daddr ||
|
|
|
|
|
!(t->dev->flags & IFF_UP))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!ip_tunnel_key_match(&t->parms, flags, key))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (t->parms.link == link)
|
|
|
|
|
return t;
|
|
|
|
|
else
|
|
|
|
|
cand = t;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(t, head, hash_node) {
|
|
|
|
|
if (remote != t->parms.iph.daddr ||
|
2014-07-05 06:26:37 +08:00
|
|
|
t->parms.iph.saddr != 0 ||
|
2013-03-25 22:49:35 +08:00
|
|
|
!(t->dev->flags & IFF_UP))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!ip_tunnel_key_match(&t->parms, flags, key))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (t->parms.link == link)
|
|
|
|
|
return t;
|
|
|
|
|
else if (!cand)
|
|
|
|
|
cand = t;
|
|
|
|
|
}
|
|
|
|
|
|
2014-01-19 16:43:42 +08:00
|
|
|
hash = ip_tunnel_hash(key, 0);
|
2013-03-25 22:49:35 +08:00
|
|
|
head = &itn->tunnels[hash];
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(t, head, hash_node) {
|
2014-07-05 06:26:37 +08:00
|
|
|
if ((local != t->parms.iph.saddr || t->parms.iph.daddr != 0) &&
|
|
|
|
|
(local != t->parms.iph.daddr || !ipv4_is_multicast(local)))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!(t->dev->flags & IFF_UP))
|
2013-03-25 22:49:35 +08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (!ip_tunnel_key_match(&t->parms, flags, key))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (t->parms.link == link)
|
|
|
|
|
return t;
|
|
|
|
|
else if (!cand)
|
|
|
|
|
cand = t;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (flags & TUNNEL_NO_KEY)
|
|
|
|
|
goto skip_key_lookup;
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(t, head, hash_node) {
|
|
|
|
|
if (t->parms.i_key != key ||
|
2014-07-05 06:26:37 +08:00
|
|
|
t->parms.iph.saddr != 0 ||
|
|
|
|
|
t->parms.iph.daddr != 0 ||
|
2013-03-25 22:49:35 +08:00
|
|
|
!(t->dev->flags & IFF_UP))
|
|
|
|
|
continue;
|
|
|
|
|
|
|
|
|
|
if (t->parms.link == link)
|
|
|
|
|
return t;
|
|
|
|
|
else if (!cand)
|
|
|
|
|
cand = t;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
skip_key_lookup:
|
|
|
|
|
if (cand)
|
|
|
|
|
return cand;
|
|
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
t = rcu_dereference(itn->collect_md_tun);
|
2017-09-12 17:47:56 +08:00
|
|
|
if (t && t->dev->flags & IFF_UP)
|
2015-08-08 14:51:42 +08:00
|
|
|
return t;
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
if (itn->fb_tunnel_dev && itn->fb_tunnel_dev->flags & IFF_UP)
|
|
|
|
|
return netdev_priv(itn->fb_tunnel_dev);
|
|
|
|
|
|
|
|
|
|
return NULL;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_lookup);
|
|
|
|
|
|
|
|
|
|
static struct hlist_head *ip_bucket(struct ip_tunnel_net *itn,
|
|
|
|
|
struct ip_tunnel_parm *parms)
|
|
|
|
|
{
|
|
|
|
|
unsigned int h;
|
|
|
|
|
__be32 remote;
|
2014-02-21 15:41:09 +08:00
|
|
|
__be32 i_key = parms->i_key;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
if (parms->iph.daddr && !ipv4_is_multicast(parms->iph.daddr))
|
|
|
|
|
remote = parms->iph.daddr;
|
|
|
|
|
else
|
|
|
|
|
remote = 0;
|
|
|
|
|
|
2014-02-21 15:41:09 +08:00
|
|
|
if (!(parms->i_flags & TUNNEL_KEY) && (parms->i_flags & VTI_ISVTI))
|
|
|
|
|
i_key = 0;
|
|
|
|
|
|
|
|
|
|
h = ip_tunnel_hash(i_key, remote);
|
2013-03-25 22:49:35 +08:00
|
|
|
return &itn->tunnels[h];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void ip_tunnel_add(struct ip_tunnel_net *itn, struct ip_tunnel *t)
|
|
|
|
|
{
|
|
|
|
|
struct hlist_head *head = ip_bucket(itn, &t->parms);
|
|
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
if (t->collect_md)
|
|
|
|
|
rcu_assign_pointer(itn->collect_md_tun, t);
|
2013-03-25 22:49:35 +08:00
|
|
|
hlist_add_head_rcu(&t->hash_node, head);
|
|
|
|
|
}
|
|
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
static void ip_tunnel_del(struct ip_tunnel_net *itn, struct ip_tunnel *t)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2015-08-08 14:51:42 +08:00
|
|
|
if (t->collect_md)
|
|
|
|
|
rcu_assign_pointer(itn->collect_md_tun, NULL);
|
2013-03-25 22:49:35 +08:00
|
|
|
hlist_del_init_rcu(&t->hash_node);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct ip_tunnel *ip_tunnel_find(struct ip_tunnel_net *itn,
|
|
|
|
|
struct ip_tunnel_parm *parms,
|
|
|
|
|
int type)
|
|
|
|
|
{
|
|
|
|
|
__be32 remote = parms->iph.daddr;
|
|
|
|
|
__be32 local = parms->iph.saddr;
|
|
|
|
|
__be32 key = parms->i_key;
|
2014-06-08 07:03:08 +08:00
|
|
|
__be16 flags = parms->i_flags;
|
2013-03-25 22:49:35 +08:00
|
|
|
int link = parms->link;
|
|
|
|
|
struct ip_tunnel *t = NULL;
|
|
|
|
|
struct hlist_head *head = ip_bucket(itn, parms);
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_rcu(t, head, hash_node) {
|
|
|
|
|
if (local == t->parms.iph.saddr &&
|
|
|
|
|
remote == t->parms.iph.daddr &&
|
|
|
|
|
link == t->parms.link &&
|
2014-06-08 07:03:08 +08:00
|
|
|
type == t->dev->type &&
|
|
|
|
|
ip_tunnel_key_match(&t->parms, flags, key))
|
2013-03-25 22:49:35 +08:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
return t;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct net_device *__ip_tunnel_create(struct net *net,
|
|
|
|
|
const struct rtnl_link_ops *ops,
|
|
|
|
|
struct ip_tunnel_parm *parms)
|
|
|
|
|
{
|
|
|
|
|
int err;
|
|
|
|
|
struct ip_tunnel *tunnel;
|
|
|
|
|
struct net_device *dev;
|
|
|
|
|
char name[IFNAMSIZ];
|
|
|
|
|
|
2018-04-05 21:39:27 +08:00
|
|
|
err = -E2BIG;
|
|
|
|
|
if (parms->name[0]) {
|
|
|
|
|
if (!dev_valid_name(parms->name))
|
|
|
|
|
goto failed;
|
2013-03-25 22:49:35 +08:00
|
|
|
strlcpy(name, parms->name, IFNAMSIZ);
|
2018-04-05 21:39:27 +08:00
|
|
|
} else {
|
|
|
|
|
if (strlen(ops->kind) > (IFNAMSIZ - 3))
|
2013-03-25 22:49:35 +08:00
|
|
|
goto failed;
|
2018-06-07 06:56:54 +08:00
|
|
|
strcpy(name, ops->kind);
|
|
|
|
|
strcat(name, "%d");
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
ASSERT_RTNL();
|
net: set name_assign_type in alloc_netdev()
Extend alloc_netdev{,_mq{,s}}() to take name_assign_type as argument, and convert
all users to pass NET_NAME_UNKNOWN.
Coccinelle patch:
@@
expression sizeof_priv, name, setup, txqs, rxqs, count;
@@
(
-alloc_netdev_mqs(sizeof_priv, name, setup, txqs, rxqs)
+alloc_netdev_mqs(sizeof_priv, name, NET_NAME_UNKNOWN, setup, txqs, rxqs)
|
-alloc_netdev_mq(sizeof_priv, name, setup, count)
+alloc_netdev_mq(sizeof_priv, name, NET_NAME_UNKNOWN, setup, count)
|
-alloc_netdev(sizeof_priv, name, setup)
+alloc_netdev(sizeof_priv, name, NET_NAME_UNKNOWN, setup)
)
v9: move comments here from the wrong commit
Signed-off-by: Tom Gundersen <teg@jklm.no>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-07-14 22:37:24 +08:00
|
|
|
dev = alloc_netdev(ops->priv_size, name, NET_NAME_UNKNOWN, ops->setup);
|
2013-03-25 22:49:35 +08:00
|
|
|
if (!dev) {
|
|
|
|
|
err = -ENOMEM;
|
|
|
|
|
goto failed;
|
|
|
|
|
}
|
|
|
|
|
dev_net_set(dev, net);
|
|
|
|
|
|
|
|
|
|
dev->rtnl_link_ops = ops;
|
|
|
|
|
|
|
|
|
|
tunnel = netdev_priv(dev);
|
|
|
|
|
tunnel->parms = *parms;
|
2013-06-26 22:11:28 +08:00
|
|
|
tunnel->net = net;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
err = register_netdevice(dev);
|
|
|
|
|
if (err)
|
|
|
|
|
goto failed_free;
|
|
|
|
|
|
|
|
|
|
return dev;
|
|
|
|
|
|
|
|
|
|
failed_free:
|
|
|
|
|
free_netdev(dev);
|
|
|
|
|
failed:
|
|
|
|
|
return ERR_PTR(err);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static int ip_tunnel_bind_dev(struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct net_device *tdev = NULL;
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
const struct iphdr *iph;
|
|
|
|
|
int hlen = LL_MAX_HEADER;
|
|
|
|
|
int mtu = ETH_DATA_LEN;
|
|
|
|
|
int t_hlen = tunnel->hlen + sizeof(struct iphdr);
|
|
|
|
|
|
|
|
|
|
iph = &tunnel->parms.iph;
|
|
|
|
|
|
|
|
|
|
/* Guess output device to choose reasonable mtu and needed_headroom */
|
|
|
|
|
if (iph->daddr) {
|
|
|
|
|
struct flowi4 fl4;
|
|
|
|
|
struct rtable *rt;
|
|
|
|
|
|
2018-02-27 21:53:38 +08:00
|
|
|
ip_tunnel_init_flow(&fl4, iph->protocol, iph->daddr,
|
|
|
|
|
iph->saddr, tunnel->parms.o_key,
|
|
|
|
|
RT_TOS(iph->tos), tunnel->parms.link,
|
|
|
|
|
tunnel->fwmark);
|
2014-01-03 03:48:26 +08:00
|
|
|
rt = ip_route_output_key(tunnel->net, &fl4);
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
if (!IS_ERR(rt)) {
|
|
|
|
|
tdev = rt->dst.dev;
|
|
|
|
|
ip_rt_put(rt);
|
|
|
|
|
}
|
|
|
|
|
if (dev->type != ARPHRD_ETHER)
|
|
|
|
|
dev->flags |= IFF_POINTOPOINT;
|
2016-04-28 17:04:51 +08:00
|
|
|
|
|
|
|
|
dst_cache_reset(&tunnel->dst_cache);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!tdev && tunnel->parms.link)
|
2013-08-13 23:51:11 +08:00
|
|
|
tdev = __dev_get_by_index(tunnel->net, tunnel->parms.link);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
if (tdev) {
|
|
|
|
|
hlen = tdev->hard_header_len + tdev->needed_headroom;
|
2018-05-31 16:59:32 +08:00
|
|
|
mtu = min(tdev->mtu, IP_MAX_MTU);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
dev->needed_headroom = t_hlen + hlen;
|
|
|
|
|
mtu -= (dev->hard_header_len + t_hlen);
|
|
|
|
|
|
2017-12-11 23:17:39 +08:00
|
|
|
if (mtu < IPV4_MIN_MTU)
|
|
|
|
|
mtu = IPV4_MIN_MTU;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
return mtu;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static struct ip_tunnel *ip_tunnel_create(struct net *net,
|
|
|
|
|
struct ip_tunnel_net *itn,
|
|
|
|
|
struct ip_tunnel_parm *parms)
|
|
|
|
|
{
|
2014-05-15 11:43:20 +08:00
|
|
|
struct ip_tunnel *nt;
|
2013-03-25 22:49:35 +08:00
|
|
|
struct net_device *dev;
|
2016-10-21 01:55:24 +08:00
|
|
|
int t_hlen;
|
2018-03-23 01:53:33 +08:00
|
|
|
int mtu;
|
|
|
|
|
int err;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
dev = __ip_tunnel_create(net, itn->rtnl_link_ops, parms);
|
2013-03-25 22:49:35 +08:00
|
|
|
if (IS_ERR(dev))
|
2014-02-14 20:14:39 +08:00
|
|
|
return ERR_CAST(dev);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2018-03-23 01:53:33 +08:00
|
|
|
mtu = ip_tunnel_bind_dev(dev);
|
|
|
|
|
err = dev_set_mtu(dev, mtu);
|
|
|
|
|
if (err)
|
|
|
|
|
goto err_dev_set_mtu;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
nt = netdev_priv(dev);
|
2016-10-21 01:55:24 +08:00
|
|
|
t_hlen = nt->hlen + sizeof(struct iphdr);
|
|
|
|
|
dev->min_mtu = ETH_MIN_MTU;
|
2018-05-31 16:59:32 +08:00
|
|
|
dev->max_mtu = IP_MAX_MTU - dev->hard_header_len - t_hlen;
|
2013-03-25 22:49:35 +08:00
|
|
|
ip_tunnel_add(itn, nt);
|
|
|
|
|
return nt;
|
2018-03-23 01:53:33 +08:00
|
|
|
|
|
|
|
|
err_dev_set_mtu:
|
|
|
|
|
unregister_netdevice(dev);
|
|
|
|
|
return ERR_PTR(err);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_rcv(struct ip_tunnel *tunnel, struct sk_buff *skb,
|
2015-08-08 14:51:42 +08:00
|
|
|
const struct tnl_ptk_info *tpi, struct metadata_dst *tun_dst,
|
|
|
|
|
bool log_ecn_error)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2014-01-04 13:57:59 +08:00
|
|
|
struct pcpu_sw_netstats *tstats;
|
2013-03-25 22:49:35 +08:00
|
|
|
const struct iphdr *iph = ip_hdr(skb);
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
#ifdef CONFIG_NET_IPGRE_BROADCAST
|
|
|
|
|
if (ipv4_is_multicast(iph->daddr)) {
|
|
|
|
|
tunnel->dev->stats.multicast++;
|
|
|
|
|
skb->pkt_type = PACKET_BROADCAST;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
|
|
|
|
|
if ((!(tpi->flags&TUNNEL_CSUM) && (tunnel->parms.i_flags&TUNNEL_CSUM)) ||
|
|
|
|
|
((tpi->flags&TUNNEL_CSUM) && !(tunnel->parms.i_flags&TUNNEL_CSUM))) {
|
|
|
|
|
tunnel->dev->stats.rx_crc_errors++;
|
|
|
|
|
tunnel->dev->stats.rx_errors++;
|
|
|
|
|
goto drop;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (tunnel->parms.i_flags&TUNNEL_SEQ) {
|
|
|
|
|
if (!(tpi->flags&TUNNEL_SEQ) ||
|
|
|
|
|
(tunnel->i_seqno && (s32)(ntohl(tpi->seq) - tunnel->i_seqno) < 0)) {
|
|
|
|
|
tunnel->dev->stats.rx_fifo_errors++;
|
|
|
|
|
tunnel->dev->stats.rx_errors++;
|
|
|
|
|
goto drop;
|
|
|
|
|
}
|
|
|
|
|
tunnel->i_seqno = ntohl(tpi->seq) + 1;
|
|
|
|
|
}
|
|
|
|
|
|
2014-05-05 06:20:04 +08:00
|
|
|
skb_reset_network_header(skb);
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
err = IP_ECN_decapsulate(iph, skb);
|
|
|
|
|
if (unlikely(err)) {
|
|
|
|
|
if (log_ecn_error)
|
|
|
|
|
net_info_ratelimited("non-ECT from %pI4 with TOS=%#x\n",
|
|
|
|
|
&iph->saddr, iph->tos);
|
|
|
|
|
if (err > 1) {
|
|
|
|
|
++tunnel->dev->stats.rx_frame_errors;
|
|
|
|
|
++tunnel->dev->stats.rx_errors;
|
|
|
|
|
goto drop;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
tstats = this_cpu_ptr(tunnel->dev->tstats);
|
|
|
|
|
u64_stats_update_begin(&tstats->syncp);
|
|
|
|
|
tstats->rx_packets++;
|
|
|
|
|
tstats->rx_bytes += skb->len;
|
|
|
|
|
u64_stats_update_end(&tstats->syncp);
|
|
|
|
|
|
2013-11-13 06:39:13 +08:00
|
|
|
skb_scrub_packet(skb, !net_eq(tunnel->net, dev_net(tunnel->dev)));
|
|
|
|
|
|
2013-06-18 08:50:02 +08:00
|
|
|
if (tunnel->dev->type == ARPHRD_ETHER) {
|
|
|
|
|
skb->protocol = eth_type_trans(skb, tunnel->dev);
|
|
|
|
|
skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
|
|
|
|
|
} else {
|
|
|
|
|
skb->dev = tunnel->dev;
|
|
|
|
|
}
|
2013-08-13 23:51:09 +08:00
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
if (tun_dst)
|
|
|
|
|
skb_dst_set(skb, (struct dst_entry *)tun_dst);
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
gro_cells_receive(&tunnel->gro_cells, skb);
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
drop:
|
2017-06-15 10:29:29 +08:00
|
|
|
if (tun_dst)
|
|
|
|
|
dst_release((struct dst_entry *)tun_dst);
|
2013-03-25 22:49:35 +08:00
|
|
|
kfree_skb(skb);
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_rcv);
|
|
|
|
|
|
2014-11-13 03:54:09 +08:00
|
|
|
int ip_tunnel_encap_add_ops(const struct ip_tunnel_encap_ops *ops,
|
|
|
|
|
unsigned int num)
|
|
|
|
|
{
|
2014-12-17 04:05:20 +08:00
|
|
|
if (num >= MAX_IPTUN_ENCAP_OPS)
|
|
|
|
|
return -ERANGE;
|
|
|
|
|
|
2014-11-13 03:54:09 +08:00
|
|
|
return !cmpxchg((const struct ip_tunnel_encap_ops **)
|
|
|
|
|
&iptun_encaps[num],
|
|
|
|
|
NULL, ops) ? 0 : -1;
|
2014-09-18 03:25:58 +08:00
|
|
|
}
|
2014-11-13 03:54:09 +08:00
|
|
|
EXPORT_SYMBOL(ip_tunnel_encap_add_ops);
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_encap_del_ops(const struct ip_tunnel_encap_ops *ops,
|
|
|
|
|
unsigned int num)
|
|
|
|
|
{
|
|
|
|
|
int ret;
|
|
|
|
|
|
2014-12-17 04:05:20 +08:00
|
|
|
if (num >= MAX_IPTUN_ENCAP_OPS)
|
|
|
|
|
return -ERANGE;
|
|
|
|
|
|
2014-11-13 03:54:09 +08:00
|
|
|
ret = (cmpxchg((const struct ip_tunnel_encap_ops **)
|
|
|
|
|
&iptun_encaps[num],
|
|
|
|
|
ops, NULL) == ops) ? 0 : -1;
|
|
|
|
|
|
|
|
|
|
synchronize_net();
|
|
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL(ip_tunnel_encap_del_ops);
|
2014-09-18 03:25:58 +08:00
|
|
|
|
|
|
|
|
int ip_tunnel_encap_setup(struct ip_tunnel *t,
|
|
|
|
|
struct ip_tunnel_encap *ipencap)
|
|
|
|
|
{
|
|
|
|
|
int hlen;
|
|
|
|
|
|
|
|
|
|
memset(&t->encap, 0, sizeof(t->encap));
|
|
|
|
|
|
|
|
|
|
hlen = ip_encap_hlen(ipencap);
|
|
|
|
|
if (hlen < 0)
|
|
|
|
|
return hlen;
|
|
|
|
|
|
|
|
|
|
t->encap.type = ipencap->type;
|
|
|
|
|
t->encap.sport = ipencap->sport;
|
|
|
|
|
t->encap.dport = ipencap->dport;
|
|
|
|
|
t->encap.flags = ipencap->flags;
|
|
|
|
|
|
|
|
|
|
t->encap_hlen = hlen;
|
|
|
|
|
t->hlen = t->encap_hlen + t->tun_hlen;
|
|
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_encap_setup);
|
|
|
|
|
|
2013-07-03 01:57:33 +08:00
|
|
|
static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,
|
2015-07-07 13:34:13 +08:00
|
|
|
struct rtable *rt, __be16 df,
|
|
|
|
|
const struct iphdr *inner_iph)
|
2013-07-03 01:57:33 +08:00
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
2013-07-12 04:12:22 +08:00
|
|
|
int pkt_size = skb->len - tunnel->hlen - dev->hard_header_len;
|
2013-07-03 01:57:33 +08:00
|
|
|
int mtu;
|
|
|
|
|
|
|
|
|
|
if (df)
|
|
|
|
|
mtu = dst_mtu(&rt->dst) - dev->hard_header_len
|
|
|
|
|
- sizeof(struct iphdr) - tunnel->hlen;
|
|
|
|
|
else
|
|
|
|
|
mtu = skb_dst(skb) ? dst_mtu(skb_dst(skb)) : dev->mtu;
|
|
|
|
|
|
2018-01-26 02:03:03 +08:00
|
|
|
skb_dst_update_pmtu(skb, mtu);
|
2013-07-03 01:57:33 +08:00
|
|
|
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP)) {
|
|
|
|
|
if (!skb_is_gso(skb) &&
|
2015-07-07 13:34:13 +08:00
|
|
|
(inner_iph->frag_off & htons(IP_DF)) &&
|
|
|
|
|
mtu < pkt_size) {
|
2013-07-03 01:57:33 +08:00
|
|
|
memset(IPCB(skb), 0, sizeof(*IPCB(skb)));
|
|
|
|
|
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu));
|
|
|
|
|
return -E2BIG;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6)) {
|
|
|
|
|
struct rt6_info *rt6 = (struct rt6_info *)skb_dst(skb);
|
|
|
|
|
|
|
|
|
|
if (rt6 && mtu < dst_mtu(skb_dst(skb)) &&
|
|
|
|
|
mtu >= IPV6_MIN_MTU) {
|
|
|
|
|
if ((tunnel->parms.iph.daddr &&
|
|
|
|
|
!ipv4_is_multicast(tunnel->parms.iph.daddr)) ||
|
|
|
|
|
rt6->rt6i_dst.plen == 128) {
|
|
|
|
|
rt6->rt6i_flags |= RTF_MODIFIED;
|
|
|
|
|
dst_metric_set(skb_dst(skb), RTAX_MTU, mtu);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (!skb_is_gso(skb) && mtu >= IPV6_MIN_MTU &&
|
|
|
|
|
mtu < pkt_size) {
|
|
|
|
|
icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
|
|
|
|
|
return -E2BIG;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
|
2016-09-16 04:00:29 +08:00
|
|
|
void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, u8 proto)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
u32 headroom = sizeof(struct iphdr);
|
|
|
|
|
struct ip_tunnel_info *tun_info;
|
|
|
|
|
const struct ip_tunnel_key *key;
|
|
|
|
|
const struct iphdr *inner_iph;
|
|
|
|
|
struct rtable *rt;
|
|
|
|
|
struct flowi4 fl4;
|
|
|
|
|
__be16 df = 0;
|
|
|
|
|
u8 tos, ttl;
|
|
|
|
|
|
|
|
|
|
tun_info = skb_tunnel_info(skb);
|
|
|
|
|
if (unlikely(!tun_info || !(tun_info->mode & IP_TUNNEL_INFO_TX) ||
|
|
|
|
|
ip_tunnel_info_af(tun_info) != AF_INET))
|
|
|
|
|
goto tx_error;
|
|
|
|
|
key = &tun_info->key;
|
|
|
|
|
memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
|
|
|
|
|
inner_iph = (const struct iphdr *)skb_inner_network_header(skb);
|
|
|
|
|
tos = key->tos;
|
|
|
|
|
if (tos == 1) {
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP))
|
|
|
|
|
tos = inner_iph->tos;
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6))
|
|
|
|
|
tos = ipv6_get_dsfield((const struct ipv6hdr *)inner_iph);
|
|
|
|
|
}
|
2018-02-27 21:53:38 +08:00
|
|
|
ip_tunnel_init_flow(&fl4, proto, key->u.ipv4.dst, key->u.ipv4.src, 0,
|
|
|
|
|
RT_TOS(tos), tunnel->parms.link, tunnel->fwmark);
|
2016-09-16 04:00:29 +08:00
|
|
|
if (tunnel->encap.type != TUNNEL_ENCAP_NONE)
|
|
|
|
|
goto tx_error;
|
|
|
|
|
rt = ip_route_output_key(tunnel->net, &fl4);
|
|
|
|
|
if (IS_ERR(rt)) {
|
|
|
|
|
dev->stats.tx_carrier_errors++;
|
|
|
|
|
goto tx_error;
|
|
|
|
|
}
|
|
|
|
|
if (rt->dst.dev == dev) {
|
|
|
|
|
ip_rt_put(rt);
|
|
|
|
|
dev->stats.collisions++;
|
|
|
|
|
goto tx_error;
|
|
|
|
|
}
|
|
|
|
|
tos = ip_tunnel_ecn_encap(tos, inner_iph, skb);
|
|
|
|
|
ttl = key->ttl;
|
|
|
|
|
if (ttl == 0) {
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP))
|
|
|
|
|
ttl = inner_iph->ttl;
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6))
|
|
|
|
|
ttl = ((const struct ipv6hdr *)inner_iph)->hop_limit;
|
|
|
|
|
else
|
|
|
|
|
ttl = ip4_dst_hoplimit(&rt->dst);
|
|
|
|
|
}
|
|
|
|
|
if (key->tun_flags & TUNNEL_DONT_FRAGMENT)
|
|
|
|
|
df = htons(IP_DF);
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IP))
|
|
|
|
|
df = inner_iph->frag_off & htons(IP_DF);
|
|
|
|
|
headroom += LL_RESERVED_SPACE(rt->dst.dev) + rt->dst.header_len;
|
|
|
|
|
if (headroom > dev->needed_headroom)
|
|
|
|
|
dev->needed_headroom = headroom;
|
|
|
|
|
|
|
|
|
|
if (skb_cow_head(skb, dev->needed_headroom)) {
|
|
|
|
|
ip_rt_put(rt);
|
|
|
|
|
goto tx_dropped;
|
|
|
|
|
}
|
2017-09-07 14:08:34 +08:00
|
|
|
iptunnel_xmit(NULL, rt, skb, fl4.saddr, fl4.daddr, proto, tos, ttl,
|
|
|
|
|
df, !net_eq(tunnel->net, dev_net(dev)));
|
2016-09-16 04:00:29 +08:00
|
|
|
return;
|
|
|
|
|
tx_error:
|
|
|
|
|
dev->stats.tx_errors++;
|
|
|
|
|
goto kfree;
|
|
|
|
|
tx_dropped:
|
|
|
|
|
dev->stats.tx_dropped++;
|
|
|
|
|
kfree:
|
|
|
|
|
kfree_skb(skb);
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_md_tunnel_xmit);
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev,
|
2014-09-18 03:25:58 +08:00
|
|
|
const struct iphdr *tnl_params, u8 protocol)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
2018-09-24 21:48:19 +08:00
|
|
|
unsigned int inner_nhdr_len = 0;
|
2013-03-25 22:49:35 +08:00
|
|
|
const struct iphdr *inner_iph;
|
|
|
|
|
struct flowi4 fl4;
|
|
|
|
|
u8 tos, ttl;
|
|
|
|
|
__be16 df;
|
2014-02-04 04:52:14 +08:00
|
|
|
struct rtable *rt; /* Route to the other host */
|
2013-03-25 22:49:35 +08:00
|
|
|
unsigned int max_headroom; /* The extra header space needed */
|
|
|
|
|
__be32 dst;
|
2014-05-16 13:34:39 +08:00
|
|
|
bool connected;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2018-09-24 21:48:19 +08:00
|
|
|
/* ensure we can access the inner net header, for several users below */
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP))
|
|
|
|
|
inner_nhdr_len = sizeof(struct iphdr);
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6))
|
|
|
|
|
inner_nhdr_len = sizeof(struct ipv6hdr);
|
|
|
|
|
if (unlikely(!pskb_may_pull(skb, inner_nhdr_len)))
|
|
|
|
|
goto tx_error;
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
inner_iph = (const struct iphdr *)skb_inner_network_header(skb);
|
2014-05-16 13:34:39 +08:00
|
|
|
connected = (tunnel->parms.iph.daddr != 0);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2016-02-22 07:58:05 +08:00
|
|
|
memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
dst = tnl_params->daddr;
|
|
|
|
|
if (dst == 0) {
|
|
|
|
|
/* NBMA tunnel */
|
|
|
|
|
|
2015-04-03 16:17:26 +08:00
|
|
|
if (!skb_dst(skb)) {
|
2013-03-25 22:49:35 +08:00
|
|
|
dev->stats.tx_fifo_errors++;
|
|
|
|
|
goto tx_error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP)) {
|
|
|
|
|
rt = skb_rtable(skb);
|
|
|
|
|
dst = rt_nexthop(rt, inner_iph->daddr);
|
|
|
|
|
}
|
|
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6)) {
|
|
|
|
|
const struct in6_addr *addr6;
|
|
|
|
|
struct neighbour *neigh;
|
|
|
|
|
bool do_tx_error_icmp;
|
|
|
|
|
int addr_type;
|
|
|
|
|
|
|
|
|
|
neigh = dst_neigh_lookup(skb_dst(skb),
|
|
|
|
|
&ipv6_hdr(skb)->daddr);
|
2015-04-03 16:17:26 +08:00
|
|
|
if (!neigh)
|
2013-03-25 22:49:35 +08:00
|
|
|
goto tx_error;
|
|
|
|
|
|
|
|
|
|
addr6 = (const struct in6_addr *)&neigh->primary_key;
|
|
|
|
|
addr_type = ipv6_addr_type(addr6);
|
|
|
|
|
|
|
|
|
|
if (addr_type == IPV6_ADDR_ANY) {
|
|
|
|
|
addr6 = &ipv6_hdr(skb)->daddr;
|
|
|
|
|
addr_type = ipv6_addr_type(addr6);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if ((addr_type & IPV6_ADDR_COMPATv4) == 0)
|
|
|
|
|
do_tx_error_icmp = true;
|
|
|
|
|
else {
|
|
|
|
|
do_tx_error_icmp = false;
|
|
|
|
|
dst = addr6->s6_addr32[3];
|
|
|
|
|
}
|
|
|
|
|
neigh_release(neigh);
|
|
|
|
|
if (do_tx_error_icmp)
|
|
|
|
|
goto tx_error_icmp;
|
|
|
|
|
}
|
|
|
|
|
#endif
|
|
|
|
|
else
|
|
|
|
|
goto tx_error;
|
2014-01-03 03:48:26 +08:00
|
|
|
|
|
|
|
|
connected = false;
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
tos = tnl_params->tos;
|
|
|
|
|
if (tos & 0x1) {
|
|
|
|
|
tos &= ~0x1;
|
2014-01-03 03:48:26 +08:00
|
|
|
if (skb->protocol == htons(ETH_P_IP)) {
|
2013-03-25 22:49:35 +08:00
|
|
|
tos = inner_iph->tos;
|
2014-01-03 03:48:26 +08:00
|
|
|
connected = false;
|
|
|
|
|
} else if (skb->protocol == htons(ETH_P_IPV6)) {
|
2013-03-25 22:49:35 +08:00
|
|
|
tos = ipv6_get_dsfield((const struct ipv6hdr *)inner_iph);
|
2014-01-03 03:48:26 +08:00
|
|
|
connected = false;
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
2018-03-06 13:53:44 +08:00
|
|
|
ip_tunnel_init_flow(&fl4, protocol, dst, tnl_params->saddr,
|
|
|
|
|
tunnel->parms.o_key, RT_TOS(tos), tunnel->parms.link,
|
|
|
|
|
tunnel->fwmark);
|
2014-01-03 03:48:26 +08:00
|
|
|
|
2014-09-18 03:25:58 +08:00
|
|
|
if (ip_tunnel_encap(skb, tunnel, &protocol, &fl4) < 0)
|
|
|
|
|
goto tx_error;
|
|
|
|
|
|
2016-02-12 22:43:55 +08:00
|
|
|
rt = connected ? dst_cache_get_ip4(&tunnel->dst_cache, &fl4.saddr) :
|
|
|
|
|
NULL;
|
2014-01-03 03:48:26 +08:00
|
|
|
|
|
|
|
|
if (!rt) {
|
|
|
|
|
rt = ip_route_output_key(tunnel->net, &fl4);
|
|
|
|
|
|
|
|
|
|
if (IS_ERR(rt)) {
|
|
|
|
|
dev->stats.tx_carrier_errors++;
|
|
|
|
|
goto tx_error;
|
|
|
|
|
}
|
|
|
|
|
if (connected)
|
2016-02-12 22:43:55 +08:00
|
|
|
dst_cache_set_ip4(&tunnel->dst_cache, &rt->dst,
|
|
|
|
|
fl4.saddr);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
2014-01-03 03:48:26 +08:00
|
|
|
|
2013-06-18 08:49:56 +08:00
|
|
|
if (rt->dst.dev == dev) {
|
2013-03-25 22:49:35 +08:00
|
|
|
ip_rt_put(rt);
|
|
|
|
|
dev->stats.collisions++;
|
|
|
|
|
goto tx_error;
|
|
|
|
|
}
|
|
|
|
|
|
2015-07-07 13:34:13 +08:00
|
|
|
if (tnl_update_pmtu(dev, skb, rt, tnl_params->frag_off, inner_iph)) {
|
2013-07-03 01:57:33 +08:00
|
|
|
ip_rt_put(rt);
|
|
|
|
|
goto tx_error;
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (tunnel->err_count > 0) {
|
|
|
|
|
if (time_before(jiffies,
|
|
|
|
|
tunnel->err_time + IPTUNNEL_ERR_TIMEO)) {
|
|
|
|
|
tunnel->err_count--;
|
|
|
|
|
|
|
|
|
|
dst_link_failure(skb);
|
|
|
|
|
} else
|
|
|
|
|
tunnel->err_count = 0;
|
|
|
|
|
}
|
|
|
|
|
|
2013-09-26 00:57:47 +08:00
|
|
|
tos = ip_tunnel_ecn_encap(tos, inner_iph, skb);
|
2013-03-25 22:49:35 +08:00
|
|
|
ttl = tnl_params->ttl;
|
|
|
|
|
if (ttl == 0) {
|
|
|
|
|
if (skb->protocol == htons(ETH_P_IP))
|
|
|
|
|
ttl = inner_iph->ttl;
|
|
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
|
else if (skb->protocol == htons(ETH_P_IPV6))
|
|
|
|
|
ttl = ((const struct ipv6hdr *)inner_iph)->hop_limit;
|
|
|
|
|
#endif
|
|
|
|
|
else
|
|
|
|
|
ttl = ip4_dst_hoplimit(&rt->dst);
|
|
|
|
|
}
|
|
|
|
|
|
2013-07-03 01:57:33 +08:00
|
|
|
df = tnl_params->frag_off;
|
2016-06-15 05:53:02 +08:00
|
|
|
if (skb->protocol == htons(ETH_P_IP) && !tunnel->ignore_df)
|
2013-07-03 01:57:33 +08:00
|
|
|
df |= (inner_iph->frag_off&htons(IP_DF));
|
|
|
|
|
|
2013-06-18 08:49:56 +08:00
|
|
|
max_headroom = LL_RESERVED_SPACE(rt->dst.dev) + sizeof(struct iphdr)
|
2014-10-04 06:48:07 +08:00
|
|
|
+ rt->dst.header_len + ip_encap_hlen(&tunnel->encap);
|
2013-10-01 17:33:59 +08:00
|
|
|
if (max_headroom > dev->needed_headroom)
|
2013-03-25 22:49:35 +08:00
|
|
|
dev->needed_headroom = max_headroom;
|
2013-10-01 17:33:59 +08:00
|
|
|
|
|
|
|
|
if (skb_cow_head(skb, dev->needed_headroom)) {
|
2014-06-06 08:34:37 +08:00
|
|
|
ip_rt_put(rt);
|
2013-10-01 17:33:59 +08:00
|
|
|
dev->stats.tx_dropped++;
|
2014-01-19 10:27:49 +08:00
|
|
|
kfree_skb(skb);
|
2013-10-01 17:33:59 +08:00
|
|
|
return;
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
|
2015-12-25 06:34:54 +08:00
|
|
|
iptunnel_xmit(NULL, rt, skb, fl4.saddr, fl4.daddr, protocol, tos, ttl,
|
|
|
|
|
df, !net_eq(tunnel->net, dev_net(dev)));
|
2013-03-25 22:49:35 +08:00
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
#if IS_ENABLED(CONFIG_IPV6)
|
|
|
|
|
tx_error_icmp:
|
|
|
|
|
dst_link_failure(skb);
|
|
|
|
|
#endif
|
|
|
|
|
tx_error:
|
|
|
|
|
dev->stats.tx_errors++;
|
2014-01-19 10:27:49 +08:00
|
|
|
kfree_skb(skb);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_xmit);
|
|
|
|
|
|
|
|
|
|
static void ip_tunnel_update(struct ip_tunnel_net *itn,
|
|
|
|
|
struct ip_tunnel *t,
|
|
|
|
|
struct net_device *dev,
|
|
|
|
|
struct ip_tunnel_parm *p,
|
2017-04-20 00:30:54 +08:00
|
|
|
bool set_mtu,
|
|
|
|
|
__u32 fwmark)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2015-08-08 14:51:42 +08:00
|
|
|
ip_tunnel_del(itn, t);
|
2013-03-25 22:49:35 +08:00
|
|
|
t->parms.iph.saddr = p->iph.saddr;
|
|
|
|
|
t->parms.iph.daddr = p->iph.daddr;
|
|
|
|
|
t->parms.i_key = p->i_key;
|
|
|
|
|
t->parms.o_key = p->o_key;
|
|
|
|
|
if (dev->type != ARPHRD_ETHER) {
|
|
|
|
|
memcpy(dev->dev_addr, &p->iph.saddr, 4);
|
|
|
|
|
memcpy(dev->broadcast, &p->iph.daddr, 4);
|
|
|
|
|
}
|
|
|
|
|
ip_tunnel_add(itn, t);
|
|
|
|
|
|
|
|
|
|
t->parms.iph.ttl = p->iph.ttl;
|
|
|
|
|
t->parms.iph.tos = p->iph.tos;
|
|
|
|
|
t->parms.iph.frag_off = p->iph.frag_off;
|
|
|
|
|
|
2017-04-20 00:30:54 +08:00
|
|
|
if (t->parms.link != p->link || t->fwmark != fwmark) {
|
2013-03-25 22:49:35 +08:00
|
|
|
int mtu;
|
|
|
|
|
|
|
|
|
|
t->parms.link = p->link;
|
2017-04-20 00:30:54 +08:00
|
|
|
t->fwmark = fwmark;
|
2013-03-25 22:49:35 +08:00
|
|
|
mtu = ip_tunnel_bind_dev(dev);
|
|
|
|
|
if (set_mtu)
|
|
|
|
|
dev->mtu = mtu;
|
|
|
|
|
}
|
2016-02-12 22:43:55 +08:00
|
|
|
dst_cache_reset(&t->dst_cache);
|
2013-03-25 22:49:35 +08:00
|
|
|
netdev_state_change(dev);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_ioctl(struct net_device *dev, struct ip_tunnel_parm *p, int cmd)
|
|
|
|
|
{
|
|
|
|
|
int err = 0;
|
2014-04-16 17:19:32 +08:00
|
|
|
struct ip_tunnel *t = netdev_priv(dev);
|
|
|
|
|
struct net *net = t->net;
|
|
|
|
|
struct ip_tunnel_net *itn = net_generic(net, t->ip_tnl_net_id);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
switch (cmd) {
|
|
|
|
|
case SIOCGETTUNNEL:
|
2014-04-16 17:19:32 +08:00
|
|
|
if (dev == itn->fb_tunnel_dev) {
|
2013-03-25 22:49:35 +08:00
|
|
|
t = ip_tunnel_find(itn, p, itn->fb_tunnel_dev->type);
|
2015-04-03 16:17:26 +08:00
|
|
|
if (!t)
|
2014-04-16 17:19:32 +08:00
|
|
|
t = netdev_priv(dev);
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
memcpy(p, &t->parms, sizeof(*p));
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case SIOCADDTUNNEL:
|
|
|
|
|
case SIOCCHGTUNNEL:
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
|
|
|
|
|
goto done;
|
|
|
|
|
if (p->iph.ttl)
|
|
|
|
|
p->iph.frag_off |= htons(IP_DF);
|
2014-06-08 06:06:25 +08:00
|
|
|
if (!(p->i_flags & VTI_ISVTI)) {
|
|
|
|
|
if (!(p->i_flags & TUNNEL_KEY))
|
|
|
|
|
p->i_key = 0;
|
|
|
|
|
if (!(p->o_flags & TUNNEL_KEY))
|
|
|
|
|
p->o_key = 0;
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
t = ip_tunnel_find(itn, p, itn->type);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2014-09-22 15:11:08 +08:00
|
|
|
if (cmd == SIOCADDTUNNEL) {
|
|
|
|
|
if (!t) {
|
|
|
|
|
t = ip_tunnel_create(net, itn, p);
|
|
|
|
|
err = PTR_ERR_OR_ZERO(t);
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
err = -EEXIST;
|
2014-05-15 13:07:02 +08:00
|
|
|
break;
|
2014-02-14 20:14:39 +08:00
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
if (dev != itn->fb_tunnel_dev && cmd == SIOCCHGTUNNEL) {
|
2015-04-03 16:17:27 +08:00
|
|
|
if (t) {
|
2013-03-25 22:49:35 +08:00
|
|
|
if (t->dev != dev) {
|
|
|
|
|
err = -EEXIST;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
} else {
|
|
|
|
|
unsigned int nflags = 0;
|
|
|
|
|
|
|
|
|
|
if (ipv4_is_multicast(p->iph.daddr))
|
|
|
|
|
nflags = IFF_BROADCAST;
|
|
|
|
|
else if (p->iph.daddr)
|
|
|
|
|
nflags = IFF_POINTOPOINT;
|
|
|
|
|
|
|
|
|
|
if ((dev->flags^nflags)&(IFF_POINTOPOINT|IFF_BROADCAST)) {
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
t = netdev_priv(dev);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (t) {
|
|
|
|
|
err = 0;
|
2017-04-20 00:30:54 +08:00
|
|
|
ip_tunnel_update(itn, t, dev, p, true, 0);
|
2014-02-14 20:14:39 +08:00
|
|
|
} else {
|
|
|
|
|
err = -ENOENT;
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
case SIOCDELTUNNEL:
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
|
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
|
|
if (dev == itn->fb_tunnel_dev) {
|
|
|
|
|
err = -ENOENT;
|
|
|
|
|
t = ip_tunnel_find(itn, p, itn->fb_tunnel_dev->type);
|
2015-04-03 16:17:26 +08:00
|
|
|
if (!t)
|
2013-03-25 22:49:35 +08:00
|
|
|
goto done;
|
|
|
|
|
err = -EPERM;
|
|
|
|
|
if (t == netdev_priv(itn->fb_tunnel_dev))
|
|
|
|
|
goto done;
|
|
|
|
|
dev = t->dev;
|
|
|
|
|
}
|
|
|
|
|
unregister_netdevice(dev);
|
|
|
|
|
err = 0;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
err = -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
done:
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_ioctl);
|
|
|
|
|
|
vxlan, gre, geneve: Set a large MTU on ovs-created tunnel devices
Prior to 4.3, openvswitch tunnel vports (vxlan, gre and geneve) could
transmit vxlan packets of any size, constrained only by the ability to
send out the resulting packets. 4.3 introduced netdevs corresponding
to tunnel vports. These netdevs have an MTU, which limits the size of
a packet that can be successfully encapsulated. The default MTU
values are low (1500 or less), which is awkwardly small in the context
of physical networks supporting jumbo frames, and leads to a
conspicuous change in behaviour for userspace.
Instead, set the MTU on openvswitch-created netdevs to be the relevant
maximum (i.e. the maximum IP packet size minus any relevant overhead),
effectively restoring the behaviour prior to 4.3.
Signed-off-by: David Wragg <david@weave.works>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-10 08:05:58 +08:00
|
|
|
int __ip_tunnel_change_mtu(struct net_device *dev, int new_mtu, bool strict)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
int t_hlen = tunnel->hlen + sizeof(struct iphdr);
|
2018-05-31 16:59:32 +08:00
|
|
|
int max_mtu = IP_MAX_MTU - dev->hard_header_len - t_hlen;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2016-10-21 01:55:24 +08:00
|
|
|
if (new_mtu < ETH_MIN_MTU)
|
2013-03-25 22:49:35 +08:00
|
|
|
return -EINVAL;
|
vxlan, gre, geneve: Set a large MTU on ovs-created tunnel devices
Prior to 4.3, openvswitch tunnel vports (vxlan, gre and geneve) could
transmit vxlan packets of any size, constrained only by the ability to
send out the resulting packets. 4.3 introduced netdevs corresponding
to tunnel vports. These netdevs have an MTU, which limits the size of
a packet that can be successfully encapsulated. The default MTU
values are low (1500 or less), which is awkwardly small in the context
of physical networks supporting jumbo frames, and leads to a
conspicuous change in behaviour for userspace.
Instead, set the MTU on openvswitch-created netdevs to be the relevant
maximum (i.e. the maximum IP packet size minus any relevant overhead),
effectively restoring the behaviour prior to 4.3.
Signed-off-by: David Wragg <david@weave.works>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-10 08:05:58 +08:00
|
|
|
|
|
|
|
|
if (new_mtu > max_mtu) {
|
|
|
|
|
if (strict)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
new_mtu = max_mtu;
|
|
|
|
|
}
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
dev->mtu = new_mtu;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
vxlan, gre, geneve: Set a large MTU on ovs-created tunnel devices
Prior to 4.3, openvswitch tunnel vports (vxlan, gre and geneve) could
transmit vxlan packets of any size, constrained only by the ability to
send out the resulting packets. 4.3 introduced netdevs corresponding
to tunnel vports. These netdevs have an MTU, which limits the size of
a packet that can be successfully encapsulated. The default MTU
values are low (1500 or less), which is awkwardly small in the context
of physical networks supporting jumbo frames, and leads to a
conspicuous change in behaviour for userspace.
Instead, set the MTU on openvswitch-created netdevs to be the relevant
maximum (i.e. the maximum IP packet size minus any relevant overhead),
effectively restoring the behaviour prior to 4.3.
Signed-off-by: David Wragg <david@weave.works>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-02-10 08:05:58 +08:00
|
|
|
EXPORT_SYMBOL_GPL(__ip_tunnel_change_mtu);
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_change_mtu(struct net_device *dev, int new_mtu)
|
|
|
|
|
{
|
|
|
|
|
return __ip_tunnel_change_mtu(dev, new_mtu, true);
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_change_mtu);
|
|
|
|
|
|
|
|
|
|
static void ip_tunnel_dev_free(struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
|
|
|
|
|
gro_cells_destroy(&tunnel->gro_cells);
|
2016-02-12 22:43:55 +08:00
|
|
|
dst_cache_destroy(&tunnel->dst_cache);
|
2013-03-25 22:49:35 +08:00
|
|
|
free_percpu(dev->tstats);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void ip_tunnel_dellink(struct net_device *dev, struct list_head *head)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
struct ip_tunnel_net *itn;
|
|
|
|
|
|
2013-08-13 23:51:11 +08:00
|
|
|
itn = net_generic(tunnel->net, tunnel->ip_tnl_net_id);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
if (itn->fb_tunnel_dev != dev) {
|
2015-08-08 14:51:42 +08:00
|
|
|
ip_tunnel_del(itn, netdev_priv(dev));
|
2013-03-25 22:49:35 +08:00
|
|
|
unregister_netdevice_queue(dev, head);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_dellink);
|
|
|
|
|
|
2015-01-15 22:11:17 +08:00
|
|
|
struct net *ip_tunnel_get_link_net(const struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
|
|
|
|
|
return tunnel->net;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL(ip_tunnel_get_link_net);
|
|
|
|
|
|
2015-04-02 23:07:02 +08:00
|
|
|
int ip_tunnel_get_iflink(const struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
|
|
|
|
|
return tunnel->parms.link;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL(ip_tunnel_get_iflink);
|
|
|
|
|
|
netns: make struct pernet_operations::id unsigned int
Make struct pernet_operations::id unsigned.
There are 2 reasons to do so:
1)
This field is really an index into an zero based array and
thus is unsigned entity. Using negative value is out-of-bound
access by definition.
2)
On x86_64 unsigned 32-bit data which are mixed with pointers
via array indexing or offsets added or subtracted to pointers
are preffered to signed 32-bit data.
"int" being used as an array index needs to be sign-extended
to 64-bit before being used.
void f(long *p, int i)
{
g(p[i]);
}
roughly translates to
movsx rsi, esi
mov rdi, [rsi+...]
call g
MOVSX is 3 byte instruction which isn't necessary if the variable is
unsigned because x86_64 is zero extending by default.
Now, there is net_generic() function which, you guessed it right, uses
"int" as an array index:
static inline void *net_generic(const struct net *net, int id)
{
...
ptr = ng->ptr[id - 1];
...
}
And this function is used a lot, so those sign extensions add up.
Patch snipes ~1730 bytes on allyesconfig kernel (without all junk
messing with code generation):
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
Unfortunately some functions actually grow bigger.
This is a semmingly random artefact of code generation with register
allocator being used differently. gcc decides that some variable
needs to live in new r8+ registers and every access now requires REX
prefix. Or it is shifted into r12, so [r12+0] addressing mode has to be
used which is longer than [r8]
However, overall balance is in negative direction:
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
function old new delta
nfsd4_lock 3886 3959 +73
tipc_link_build_proto_msg 1096 1140 +44
mac80211_hwsim_new_radio 2776 2808 +32
tipc_mon_rcv 1032 1058 +26
svcauth_gss_legacy_init 1413 1429 +16
tipc_bcbase_select_primary 379 392 +13
nfsd4_exchange_id 1247 1260 +13
nfsd4_setclientid_confirm 782 793 +11
...
put_client_renew_locked 494 480 -14
ip_set_sockfn_get 730 716 -14
geneve_sock_add 829 813 -16
nfsd4_sequence_done 721 703 -18
nlmclnt_lookup_host 708 686 -22
nfsd4_lockt 1085 1063 -22
nfs_get_client 1077 1050 -27
tcf_bpf_init 1106 1076 -30
nfsd4_encode_fattr 5997 5930 -67
Total: Before=154856051, After=154854321, chg -0.00%
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-17 09:58:21 +08:00
|
|
|
int ip_tunnel_init_net(struct net *net, unsigned int ip_tnl_net_id,
|
2013-03-25 22:49:35 +08:00
|
|
|
struct rtnl_link_ops *ops, char *devname)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel_net *itn = net_generic(net, ip_tnl_net_id);
|
|
|
|
|
struct ip_tunnel_parm parms;
|
2013-08-06 13:51:37 +08:00
|
|
|
unsigned int i;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
itn->rtnl_link_ops = ops;
|
2013-08-06 13:51:37 +08:00
|
|
|
for (i = 0; i < IP_TNL_HASH_SIZE; i++)
|
|
|
|
|
INIT_HLIST_HEAD(&itn->tunnels[i]);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
if (!ops || !net_has_fallback_tunnels(net)) {
|
|
|
|
|
struct ip_tunnel_net *it_init_net;
|
|
|
|
|
|
|
|
|
|
it_init_net = net_generic(&init_net, ip_tnl_net_id);
|
|
|
|
|
itn->type = it_init_net->type;
|
2013-03-25 22:49:35 +08:00
|
|
|
itn->fb_tunnel_dev = NULL;
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2013-08-06 13:51:37 +08:00
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
memset(&parms, 0, sizeof(parms));
|
|
|
|
|
if (devname)
|
|
|
|
|
strlcpy(parms.name, devname, IFNAMSIZ);
|
|
|
|
|
|
|
|
|
|
rtnl_lock();
|
|
|
|
|
itn->fb_tunnel_dev = __ip_tunnel_create(net, ops, &parms);
|
2013-08-19 15:05:10 +08:00
|
|
|
/* FB netdevice is special: we have one, and only one per netns.
|
|
|
|
|
* Allowing to move it to another netns is clearly unsafe.
|
|
|
|
|
*/
|
2013-10-01 17:34:48 +08:00
|
|
|
if (!IS_ERR(itn->fb_tunnel_dev)) {
|
2013-08-23 16:15:37 +08:00
|
|
|
itn->fb_tunnel_dev->features |= NETIF_F_NETNS_LOCAL;
|
2014-05-19 17:36:56 +08:00
|
|
|
itn->fb_tunnel_dev->mtu = ip_tunnel_bind_dev(itn->fb_tunnel_dev);
|
2013-10-01 17:34:48 +08:00
|
|
|
ip_tunnel_add(itn, netdev_priv(itn->fb_tunnel_dev));
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
itn->type = itn->fb_tunnel_dev->type;
|
2013-10-01 17:34:48 +08:00
|
|
|
}
|
2013-08-23 16:15:37 +08:00
|
|
|
rtnl_unlock();
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2014-01-27 14:43:57 +08:00
|
|
|
return PTR_ERR_OR_ZERO(itn->fb_tunnel_dev);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_init_net);
|
|
|
|
|
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
static void ip_tunnel_destroy(struct net *net, struct ip_tunnel_net *itn,
|
|
|
|
|
struct list_head *head,
|
2013-08-13 23:51:11 +08:00
|
|
|
struct rtnl_link_ops *ops)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2013-08-13 23:51:11 +08:00
|
|
|
struct net_device *dev, *aux;
|
2013-03-25 22:49:35 +08:00
|
|
|
int h;
|
|
|
|
|
|
2013-08-13 23:51:11 +08:00
|
|
|
for_each_netdev_safe(net, dev, aux)
|
|
|
|
|
if (dev->rtnl_link_ops == ops)
|
|
|
|
|
unregister_netdevice_queue(dev, head);
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
for (h = 0; h < IP_TNL_HASH_SIZE; h++) {
|
|
|
|
|
struct ip_tunnel *t;
|
|
|
|
|
struct hlist_node *n;
|
|
|
|
|
struct hlist_head *thead = &itn->tunnels[h];
|
|
|
|
|
|
|
|
|
|
hlist_for_each_entry_safe(t, n, thead, hash_node)
|
2013-08-13 23:51:11 +08:00
|
|
|
/* If dev is in the same netns, it has already
|
|
|
|
|
* been added to the list by the previous loop.
|
|
|
|
|
*/
|
|
|
|
|
if (!net_eq(dev_net(t->dev), net))
|
|
|
|
|
unregister_netdevice_queue(t->dev, head);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-09-20 07:27:09 +08:00
|
|
|
void ip_tunnel_delete_nets(struct list_head *net_list, unsigned int id,
|
|
|
|
|
struct rtnl_link_ops *ops)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2017-09-20 07:27:09 +08:00
|
|
|
struct ip_tunnel_net *itn;
|
|
|
|
|
struct net *net;
|
2013-03-25 22:49:35 +08:00
|
|
|
LIST_HEAD(list);
|
|
|
|
|
|
|
|
|
|
rtnl_lock();
|
2017-09-20 07:27:09 +08:00
|
|
|
list_for_each_entry(net, net_list, exit_list) {
|
|
|
|
|
itn = net_generic(net, id);
|
net: do not create fallback tunnels for non-default namespaces
fallback tunnels (like tunl0, gre0, gretap0, erspan0, sit0,
ip6tnl0, ip6gre0) are automatically created when the corresponding
module is loaded.
These tunnels are also automatically created when a new network
namespace is created, at a great cost.
In many cases, netns are used for isolation purposes, and these
extra network devices are a waste of resources. We are using
thousands of netns per host, and hit the netns creation/delete
bottleneck a lot. (Many thanks to Kirill for recent work on this)
Add a new sysctl so that we can opt-out from this automatic creation.
Note that these tunnels are still created for the initial namespace,
to be the least intrusive for typical setups.
Tested:
lpk43:~# cat add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait
lpk43:~# echo 0 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m37.521s
user 0m0.886s
sys 7m7.084s
lpk43:~# echo 1 >/proc/sys/net/core/fb_tunnels_only_for_init_net
lpk43:~# time ./add_del_unshare.sh
real 0m4.761s
user 0m0.851s
sys 1m8.343s
lpk43:~#
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2018-03-09 04:51:41 +08:00
|
|
|
ip_tunnel_destroy(net, itn, &list, ops);
|
2017-09-20 07:27:09 +08:00
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
unregister_netdevice_many(&list);
|
|
|
|
|
rtnl_unlock();
|
|
|
|
|
}
|
2017-09-20 07:27:09 +08:00
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_delete_nets);
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
int ip_tunnel_newlink(struct net_device *dev, struct nlattr *tb[],
|
2017-04-20 00:30:54 +08:00
|
|
|
struct ip_tunnel_parm *p, __u32 fwmark)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
|
|
|
|
struct ip_tunnel *nt;
|
|
|
|
|
struct net *net = dev_net(dev);
|
|
|
|
|
struct ip_tunnel_net *itn;
|
|
|
|
|
int mtu;
|
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
nt = netdev_priv(dev);
|
|
|
|
|
itn = net_generic(net, nt->ip_tnl_net_id);
|
|
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
if (nt->collect_md) {
|
|
|
|
|
if (rtnl_dereference(itn->collect_md_tun))
|
|
|
|
|
return -EEXIST;
|
|
|
|
|
} else {
|
|
|
|
|
if (ip_tunnel_find(itn, p, dev->type))
|
|
|
|
|
return -EEXIST;
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2013-06-26 22:11:28 +08:00
|
|
|
nt->net = net;
|
2013-03-25 22:49:35 +08:00
|
|
|
nt->parms = *p;
|
2017-04-20 00:30:54 +08:00
|
|
|
nt->fwmark = fwmark;
|
2013-03-25 22:49:35 +08:00
|
|
|
err = register_netdevice(dev);
|
|
|
|
|
if (err)
|
2018-03-23 01:53:33 +08:00
|
|
|
goto err_register_netdevice;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
if (dev->type == ARPHRD_ETHER && !tb[IFLA_ADDRESS])
|
|
|
|
|
eth_hw_addr_random(dev);
|
|
|
|
|
|
|
|
|
|
mtu = ip_tunnel_bind_dev(dev);
|
2018-03-16 00:16:28 +08:00
|
|
|
if (tb[IFLA_MTU]) {
|
2018-05-31 16:59:32 +08:00
|
|
|
unsigned int max = IP_MAX_MTU - dev->hard_header_len - nt->hlen;
|
2018-03-16 00:16:28 +08:00
|
|
|
|
2018-03-29 23:42:14 +08:00
|
|
|
mtu = clamp(dev->mtu, (unsigned int)ETH_MIN_MTU,
|
|
|
|
|
(unsigned int)(max - sizeof(struct iphdr)));
|
2018-03-23 01:53:33 +08:00
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
|
2018-03-29 23:42:14 +08:00
|
|
|
err = dev_set_mtu(dev, mtu);
|
|
|
|
|
if (err)
|
|
|
|
|
goto err_dev_set_mtu;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
ip_tunnel_add(itn, nt);
|
2018-03-23 01:53:33 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
|
|
err_dev_set_mtu:
|
|
|
|
|
unregister_netdevice(dev);
|
|
|
|
|
err_register_netdevice:
|
2013-03-25 22:49:35 +08:00
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_newlink);
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_changelink(struct net_device *dev, struct nlattr *tb[],
|
2017-04-20 00:30:54 +08:00
|
|
|
struct ip_tunnel_parm *p, __u32 fwmark)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
2013-08-13 23:51:11 +08:00
|
|
|
struct ip_tunnel *t;
|
2013-03-25 22:49:35 +08:00
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
2013-08-13 23:51:11 +08:00
|
|
|
struct net *net = tunnel->net;
|
2013-03-25 22:49:35 +08:00
|
|
|
struct ip_tunnel_net *itn = net_generic(net, tunnel->ip_tnl_net_id);
|
|
|
|
|
|
|
|
|
|
if (dev == itn->fb_tunnel_dev)
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
|
|
t = ip_tunnel_find(itn, p, dev->type);
|
|
|
|
|
|
|
|
|
|
if (t) {
|
|
|
|
|
if (t->dev != dev)
|
|
|
|
|
return -EEXIST;
|
|
|
|
|
} else {
|
2013-08-13 23:51:11 +08:00
|
|
|
t = tunnel;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
|
|
|
|
if (dev->type != ARPHRD_ETHER) {
|
|
|
|
|
unsigned int nflags = 0;
|
|
|
|
|
|
|
|
|
|
if (ipv4_is_multicast(p->iph.daddr))
|
|
|
|
|
nflags = IFF_BROADCAST;
|
|
|
|
|
else if (p->iph.daddr)
|
|
|
|
|
nflags = IFF_POINTOPOINT;
|
|
|
|
|
|
|
|
|
|
if ((dev->flags ^ nflags) &
|
|
|
|
|
(IFF_POINTOPOINT | IFF_BROADCAST))
|
|
|
|
|
return -EINVAL;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2017-04-20 00:30:54 +08:00
|
|
|
ip_tunnel_update(itn, t, dev, p, !tb[IFLA_MTU], fwmark);
|
2013-03-25 22:49:35 +08:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_changelink);
|
|
|
|
|
|
|
|
|
|
int ip_tunnel_init(struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
struct iphdr *iph = &tunnel->parms.iph;
|
2014-02-14 03:46:28 +08:00
|
|
|
int err;
|
2013-03-25 22:49:35 +08:00
|
|
|
|
net: Fix inconsistent teardown and release of private netdev state.
Network devices can allocate reasources and private memory using
netdev_ops->ndo_init(). However, the release of these resources
can occur in one of two different places.
Either netdev_ops->ndo_uninit() or netdev->destructor().
The decision of which operation frees the resources depends upon
whether it is necessary for all netdev refs to be released before it
is safe to perform the freeing.
netdev_ops->ndo_uninit() presumably can occur right after the
NETDEV_UNREGISTER notifier completes and the unicast and multicast
address lists are flushed.
netdev->destructor(), on the other hand, does not run until the
netdev references all go away.
Further complicating the situation is that netdev->destructor()
almost universally does also a free_netdev().
This creates a problem for the logic in register_netdevice().
Because all callers of register_netdevice() manage the freeing
of the netdev, and invoke free_netdev(dev) if register_netdevice()
fails.
If netdev_ops->ndo_init() succeeds, but something else fails inside
of register_netdevice(), it does call ndo_ops->ndo_uninit(). But
it is not able to invoke netdev->destructor().
This is because netdev->destructor() will do a free_netdev() and
then the caller of register_netdevice() will do the same.
However, this means that the resources that would normally be released
by netdev->destructor() will not be.
Over the years drivers have added local hacks to deal with this, by
invoking their destructor parts by hand when register_netdevice()
fails.
Many drivers do not try to deal with this, and instead we have leaks.
Let's close this hole by formalizing the distinction between what
private things need to be freed up by netdev->destructor() and whether
the driver needs unregister_netdevice() to perform the free_netdev().
netdev->priv_destructor() performs all actions to free up the private
resources that used to be freed by netdev->destructor(), except for
free_netdev().
netdev->needs_free_netdev is a boolean that indicates whether
free_netdev() should be done at the end of unregister_netdevice().
Now, register_netdevice() can sanely release all resources after
ndo_ops->ndo_init() succeeds, by invoking both ndo_ops->ndo_uninit()
and netdev->priv_destructor().
And at the end of unregister_netdevice(), we invoke
netdev->priv_destructor() and optionally call free_netdev().
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-05-09 00:52:56 +08:00
|
|
|
dev->needs_free_netdev = true;
|
|
|
|
|
dev->priv_destructor = ip_tunnel_dev_free;
|
2014-02-14 03:46:28 +08:00
|
|
|
dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
|
2013-03-25 22:49:35 +08:00
|
|
|
if (!dev->tstats)
|
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
2016-02-12 22:43:55 +08:00
|
|
|
err = dst_cache_init(&tunnel->dst_cache, GFP_KERNEL);
|
|
|
|
|
if (err) {
|
2014-01-03 03:48:33 +08:00
|
|
|
free_percpu(dev->tstats);
|
2016-02-12 22:43:55 +08:00
|
|
|
return err;
|
2014-01-03 03:48:33 +08:00
|
|
|
}
|
|
|
|
|
|
2013-03-25 22:49:35 +08:00
|
|
|
err = gro_cells_init(&tunnel->gro_cells, dev);
|
|
|
|
|
if (err) {
|
2016-02-12 22:43:55 +08:00
|
|
|
dst_cache_destroy(&tunnel->dst_cache);
|
2013-03-25 22:49:35 +08:00
|
|
|
free_percpu(dev->tstats);
|
|
|
|
|
return err;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
tunnel->dev = dev;
|
2013-08-13 23:51:11 +08:00
|
|
|
tunnel->net = dev_net(dev);
|
2013-03-25 22:49:35 +08:00
|
|
|
strcpy(tunnel->parms.name, dev->name);
|
|
|
|
|
iph->version = 4;
|
|
|
|
|
iph->ihl = 5;
|
|
|
|
|
|
2015-08-08 14:51:42 +08:00
|
|
|
if (tunnel->collect_md) {
|
|
|
|
|
dev->features |= NETIF_F_NETNS_LOCAL;
|
|
|
|
|
netif_keep_dst(dev);
|
|
|
|
|
}
|
2013-03-25 22:49:35 +08:00
|
|
|
return 0;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_init);
|
|
|
|
|
|
|
|
|
|
void ip_tunnel_uninit(struct net_device *dev)
|
|
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
2013-08-13 23:51:11 +08:00
|
|
|
struct net *net = tunnel->net;
|
2013-03-25 22:49:35 +08:00
|
|
|
struct ip_tunnel_net *itn;
|
|
|
|
|
|
|
|
|
|
itn = net_generic(net, tunnel->ip_tnl_net_id);
|
|
|
|
|
/* fb_tunnel_dev will be unregisted in net-exit call. */
|
|
|
|
|
if (itn->fb_tunnel_dev != dev)
|
2015-08-08 14:51:42 +08:00
|
|
|
ip_tunnel_del(itn, netdev_priv(dev));
|
2014-01-03 03:48:26 +08:00
|
|
|
|
2016-02-12 22:43:55 +08:00
|
|
|
dst_cache_reset(&tunnel->dst_cache);
|
2013-03-25 22:49:35 +08:00
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_uninit);
|
|
|
|
|
|
|
|
|
|
/* Do least required initialization, rest of init is done in tunnel_init call */
|
netns: make struct pernet_operations::id unsigned int
Make struct pernet_operations::id unsigned.
There are 2 reasons to do so:
1)
This field is really an index into an zero based array and
thus is unsigned entity. Using negative value is out-of-bound
access by definition.
2)
On x86_64 unsigned 32-bit data which are mixed with pointers
via array indexing or offsets added or subtracted to pointers
are preffered to signed 32-bit data.
"int" being used as an array index needs to be sign-extended
to 64-bit before being used.
void f(long *p, int i)
{
g(p[i]);
}
roughly translates to
movsx rsi, esi
mov rdi, [rsi+...]
call g
MOVSX is 3 byte instruction which isn't necessary if the variable is
unsigned because x86_64 is zero extending by default.
Now, there is net_generic() function which, you guessed it right, uses
"int" as an array index:
static inline void *net_generic(const struct net *net, int id)
{
...
ptr = ng->ptr[id - 1];
...
}
And this function is used a lot, so those sign extensions add up.
Patch snipes ~1730 bytes on allyesconfig kernel (without all junk
messing with code generation):
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
Unfortunately some functions actually grow bigger.
This is a semmingly random artefact of code generation with register
allocator being used differently. gcc decides that some variable
needs to live in new r8+ registers and every access now requires REX
prefix. Or it is shifted into r12, so [r12+0] addressing mode has to be
used which is longer than [r8]
However, overall balance is in negative direction:
add/remove: 0/0 grow/shrink: 70/598 up/down: 396/-2126 (-1730)
function old new delta
nfsd4_lock 3886 3959 +73
tipc_link_build_proto_msg 1096 1140 +44
mac80211_hwsim_new_radio 2776 2808 +32
tipc_mon_rcv 1032 1058 +26
svcauth_gss_legacy_init 1413 1429 +16
tipc_bcbase_select_primary 379 392 +13
nfsd4_exchange_id 1247 1260 +13
nfsd4_setclientid_confirm 782 793 +11
...
put_client_renew_locked 494 480 -14
ip_set_sockfn_get 730 716 -14
geneve_sock_add 829 813 -16
nfsd4_sequence_done 721 703 -18
nlmclnt_lookup_host 708 686 -22
nfsd4_lockt 1085 1063 -22
nfs_get_client 1077 1050 -27
tcf_bpf_init 1106 1076 -30
nfsd4_encode_fattr 5997 5930 -67
Total: Before=154856051, After=154854321, chg -0.00%
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2016-11-17 09:58:21 +08:00
|
|
|
void ip_tunnel_setup(struct net_device *dev, unsigned int net_id)
|
2013-03-25 22:49:35 +08:00
|
|
|
{
|
|
|
|
|
struct ip_tunnel *tunnel = netdev_priv(dev);
|
|
|
|
|
tunnel->ip_tnl_net_id = net_id;
|
|
|
|
|
}
|
|
|
|
|
EXPORT_SYMBOL_GPL(ip_tunnel_setup);
|
|
|
|
|
|
|
|
|
|
MODULE_LICENSE("GPL");
|