aboutsummaryrefslogtreecommitdiffstats
path: root/net/ipv6/route.c
diff options
context:
space:
mode:
Diffstat (limited to 'net/ipv6/route.c')
-rw-r--r--net/ipv6/route.c640
1 files changed, 421 insertions, 219 deletions
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index e8b241cb60bc..210a079cfc6f 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -36,10 +36,12 @@
36#include <linux/route.h> 36#include <linux/route.h>
37#include <linux/netdevice.h> 37#include <linux/netdevice.h>
38#include <linux/in6.h> 38#include <linux/in6.h>
39#include <linux/mroute6.h>
39#include <linux/init.h> 40#include <linux/init.h>
40#include <linux/if_arp.h> 41#include <linux/if_arp.h>
41#include <linux/proc_fs.h> 42#include <linux/proc_fs.h>
42#include <linux/seq_file.h> 43#include <linux/seq_file.h>
44#include <linux/nsproxy.h>
43#include <net/net_namespace.h> 45#include <net/net_namespace.h>
44#include <net/snmp.h> 46#include <net/snmp.h>
45#include <net/ipv6.h> 47#include <net/ipv6.h>
@@ -87,14 +89,16 @@ static void ip6_link_failure(struct sk_buff *skb);
87static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu); 89static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu);
88 90
89#ifdef CONFIG_IPV6_ROUTE_INFO 91#ifdef CONFIG_IPV6_ROUTE_INFO
90static struct rt6_info *rt6_add_route_info(struct in6_addr *prefix, int prefixlen, 92static struct rt6_info *rt6_add_route_info(struct net *net,
93 struct in6_addr *prefix, int prefixlen,
91 struct in6_addr *gwaddr, int ifindex, 94 struct in6_addr *gwaddr, int ifindex,
92 unsigned pref); 95 unsigned pref);
93static struct rt6_info *rt6_get_route_info(struct in6_addr *prefix, int prefixlen, 96static struct rt6_info *rt6_get_route_info(struct net *net,
97 struct in6_addr *prefix, int prefixlen,
94 struct in6_addr *gwaddr, int ifindex); 98 struct in6_addr *gwaddr, int ifindex);
95#endif 99#endif
96 100
97static struct dst_ops ip6_dst_ops = { 101static struct dst_ops ip6_dst_ops_template = {
98 .family = AF_INET6, 102 .family = AF_INET6,
99 .protocol = __constant_htons(ETH_P_IPV6), 103 .protocol = __constant_htons(ETH_P_IPV6),
100 .gc = ip6_dst_gc, 104 .gc = ip6_dst_gc,
@@ -124,7 +128,7 @@ static struct dst_ops ip6_dst_blackhole_ops = {
124 .entries = ATOMIC_INIT(0), 128 .entries = ATOMIC_INIT(0),
125}; 129};
126 130
127struct rt6_info ip6_null_entry = { 131static struct rt6_info ip6_null_entry_template = {
128 .u = { 132 .u = {
129 .dst = { 133 .dst = {
130 .__refcnt = ATOMIC_INIT(1), 134 .__refcnt = ATOMIC_INIT(1),
@@ -134,8 +138,6 @@ struct rt6_info ip6_null_entry = {
134 .metrics = { [RTAX_HOPLIMIT - 1] = 255, }, 138 .metrics = { [RTAX_HOPLIMIT - 1] = 255, },
135 .input = ip6_pkt_discard, 139 .input = ip6_pkt_discard,
136 .output = ip6_pkt_discard_out, 140 .output = ip6_pkt_discard_out,
137 .ops = &ip6_dst_ops,
138 .path = (struct dst_entry*)&ip6_null_entry,
139 } 141 }
140 }, 142 },
141 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP), 143 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP),
@@ -148,7 +150,7 @@ struct rt6_info ip6_null_entry = {
148static int ip6_pkt_prohibit(struct sk_buff *skb); 150static int ip6_pkt_prohibit(struct sk_buff *skb);
149static int ip6_pkt_prohibit_out(struct sk_buff *skb); 151static int ip6_pkt_prohibit_out(struct sk_buff *skb);
150 152
151struct rt6_info ip6_prohibit_entry = { 153struct rt6_info ip6_prohibit_entry_template = {
152 .u = { 154 .u = {
153 .dst = { 155 .dst = {
154 .__refcnt = ATOMIC_INIT(1), 156 .__refcnt = ATOMIC_INIT(1),
@@ -158,8 +160,6 @@ struct rt6_info ip6_prohibit_entry = {
158 .metrics = { [RTAX_HOPLIMIT - 1] = 255, }, 160 .metrics = { [RTAX_HOPLIMIT - 1] = 255, },
159 .input = ip6_pkt_prohibit, 161 .input = ip6_pkt_prohibit,
160 .output = ip6_pkt_prohibit_out, 162 .output = ip6_pkt_prohibit_out,
161 .ops = &ip6_dst_ops,
162 .path = (struct dst_entry*)&ip6_prohibit_entry,
163 } 163 }
164 }, 164 },
165 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP), 165 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP),
@@ -167,7 +167,7 @@ struct rt6_info ip6_prohibit_entry = {
167 .rt6i_ref = ATOMIC_INIT(1), 167 .rt6i_ref = ATOMIC_INIT(1),
168}; 168};
169 169
170struct rt6_info ip6_blk_hole_entry = { 170static struct rt6_info ip6_blk_hole_entry_template = {
171 .u = { 171 .u = {
172 .dst = { 172 .dst = {
173 .__refcnt = ATOMIC_INIT(1), 173 .__refcnt = ATOMIC_INIT(1),
@@ -177,8 +177,6 @@ struct rt6_info ip6_blk_hole_entry = {
177 .metrics = { [RTAX_HOPLIMIT - 1] = 255, }, 177 .metrics = { [RTAX_HOPLIMIT - 1] = 255, },
178 .input = dst_discard, 178 .input = dst_discard,
179 .output = dst_discard, 179 .output = dst_discard,
180 .ops = &ip6_dst_ops,
181 .path = (struct dst_entry*)&ip6_blk_hole_entry,
182 } 180 }
183 }, 181 },
184 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP), 182 .rt6i_flags = (RTF_REJECT | RTF_NONEXTHOP),
@@ -189,9 +187,9 @@ struct rt6_info ip6_blk_hole_entry = {
189#endif 187#endif
190 188
191/* allocate dst with ip6_dst_ops */ 189/* allocate dst with ip6_dst_ops */
192static __inline__ struct rt6_info *ip6_dst_alloc(void) 190static inline struct rt6_info *ip6_dst_alloc(struct dst_ops *ops)
193{ 191{
194 return (struct rt6_info *)dst_alloc(&ip6_dst_ops); 192 return (struct rt6_info *)dst_alloc(ops);
195} 193}
196 194
197static void ip6_dst_destroy(struct dst_entry *dst) 195static void ip6_dst_destroy(struct dst_entry *dst)
@@ -211,7 +209,7 @@ static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
211 struct rt6_info *rt = (struct rt6_info *)dst; 209 struct rt6_info *rt = (struct rt6_info *)dst;
212 struct inet6_dev *idev = rt->rt6i_idev; 210 struct inet6_dev *idev = rt->rt6i_idev;
213 struct net_device *loopback_dev = 211 struct net_device *loopback_dev =
214 dev->nd_net->loopback_dev; 212 dev_net(dev)->loopback_dev;
215 213
216 if (dev != loopback_dev && idev != NULL && idev->dev == dev) { 214 if (dev != loopback_dev && idev != NULL && idev->dev == dev) {
217 struct inet6_dev *loopback_idev = 215 struct inet6_dev *loopback_idev =
@@ -239,7 +237,8 @@ static inline int rt6_need_strict(struct in6_addr *daddr)
239 * Route lookup. Any table->tb6_lock is implied. 237 * Route lookup. Any table->tb6_lock is implied.
240 */ 238 */
241 239
242static __inline__ struct rt6_info *rt6_device_match(struct rt6_info *rt, 240static inline struct rt6_info *rt6_device_match(struct net *net,
241 struct rt6_info *rt,
243 int oif, 242 int oif,
244 int strict) 243 int strict)
245{ 244{
@@ -268,7 +267,7 @@ static __inline__ struct rt6_info *rt6_device_match(struct rt6_info *rt,
268 return local; 267 return local;
269 268
270 if (strict) 269 if (strict)
271 return &ip6_null_entry; 270 return net->ipv6.ip6_null_entry;
272 } 271 }
273 return rt; 272 return rt;
274} 273}
@@ -409,9 +408,10 @@ static struct rt6_info *find_rr_leaf(struct fib6_node *fn,
409static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict) 408static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)
410{ 409{
411 struct rt6_info *match, *rt0; 410 struct rt6_info *match, *rt0;
411 struct net *net;
412 412
413 RT6_TRACE("%s(fn->leaf=%p, oif=%d)\n", 413 RT6_TRACE("%s(fn->leaf=%p, oif=%d)\n",
414 __FUNCTION__, fn->leaf, oif); 414 __func__, fn->leaf, oif);
415 415
416 rt0 = fn->rr_ptr; 416 rt0 = fn->rr_ptr;
417 if (!rt0) 417 if (!rt0)
@@ -432,15 +432,17 @@ static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)
432 } 432 }
433 433
434 RT6_TRACE("%s() => %p\n", 434 RT6_TRACE("%s() => %p\n",
435 __FUNCTION__, match); 435 __func__, match);
436 436
437 return (match ? match : &ip6_null_entry); 437 net = dev_net(rt0->rt6i_dev);
438 return (match ? match : net->ipv6.ip6_null_entry);
438} 439}
439 440
440#ifdef CONFIG_IPV6_ROUTE_INFO 441#ifdef CONFIG_IPV6_ROUTE_INFO
441int rt6_route_rcv(struct net_device *dev, u8 *opt, int len, 442int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
442 struct in6_addr *gwaddr) 443 struct in6_addr *gwaddr)
443{ 444{
445 struct net *net = dev_net(dev);
444 struct route_info *rinfo = (struct route_info *) opt; 446 struct route_info *rinfo = (struct route_info *) opt;
445 struct in6_addr prefix_buf, *prefix; 447 struct in6_addr prefix_buf, *prefix;
446 unsigned int pref; 448 unsigned int pref;
@@ -488,7 +490,8 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
488 prefix = &prefix_buf; 490 prefix = &prefix_buf;
489 } 491 }
490 492
491 rt = rt6_get_route_info(prefix, rinfo->prefix_len, gwaddr, dev->ifindex); 493 rt = rt6_get_route_info(net, prefix, rinfo->prefix_len, gwaddr,
494 dev->ifindex);
492 495
493 if (rt && !lifetime) { 496 if (rt && !lifetime) {
494 ip6_del_rt(rt); 497 ip6_del_rt(rt);
@@ -496,7 +499,7 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
496 } 499 }
497 500
498 if (!rt && lifetime) 501 if (!rt && lifetime)
499 rt = rt6_add_route_info(prefix, rinfo->prefix_len, gwaddr, dev->ifindex, 502 rt = rt6_add_route_info(net, prefix, rinfo->prefix_len, gwaddr, dev->ifindex,
500 pref); 503 pref);
501 else if (rt) 504 else if (rt)
502 rt->rt6i_flags = RTF_ROUTEINFO | 505 rt->rt6i_flags = RTF_ROUTEINFO |
@@ -515,9 +518,9 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
515} 518}
516#endif 519#endif
517 520
518#define BACKTRACK(saddr) \ 521#define BACKTRACK(__net, saddr) \
519do { \ 522do { \
520 if (rt == &ip6_null_entry) { \ 523 if (rt == __net->ipv6.ip6_null_entry) { \
521 struct fib6_node *pn; \ 524 struct fib6_node *pn; \
522 while (1) { \ 525 while (1) { \
523 if (fn->fn_flags & RTN_TL_ROOT) \ 526 if (fn->fn_flags & RTN_TL_ROOT) \
@@ -533,7 +536,8 @@ do { \
533 } \ 536 } \
534} while(0) 537} while(0)
535 538
536static struct rt6_info *ip6_pol_route_lookup(struct fib6_table *table, 539static struct rt6_info *ip6_pol_route_lookup(struct net *net,
540 struct fib6_table *table,
537 struct flowi *fl, int flags) 541 struct flowi *fl, int flags)
538{ 542{
539 struct fib6_node *fn; 543 struct fib6_node *fn;
@@ -543,8 +547,8 @@ static struct rt6_info *ip6_pol_route_lookup(struct fib6_table *table,
543 fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src); 547 fn = fib6_lookup(&table->tb6_root, &fl->fl6_dst, &fl->fl6_src);
544restart: 548restart:
545 rt = fn->leaf; 549 rt = fn->leaf;
546 rt = rt6_device_match(rt, fl->oif, flags); 550 rt = rt6_device_match(net, rt, fl->oif, flags);
547 BACKTRACK(&fl->fl6_src); 551 BACKTRACK(net, &fl->fl6_src);
548out: 552out:
549 dst_use(&rt->u.dst, jiffies); 553 dst_use(&rt->u.dst, jiffies);
550 read_unlock_bh(&table->tb6_lock); 554 read_unlock_bh(&table->tb6_lock);
@@ -552,8 +556,8 @@ out:
552 556
553} 557}
554 558
555struct rt6_info *rt6_lookup(struct in6_addr *daddr, struct in6_addr *saddr, 559struct rt6_info *rt6_lookup(struct net *net, const struct in6_addr *daddr,
556 int oif, int strict) 560 const struct in6_addr *saddr, int oif, int strict)
557{ 561{
558 struct flowi fl = { 562 struct flowi fl = {
559 .oif = oif, 563 .oif = oif,
@@ -571,7 +575,7 @@ struct rt6_info *rt6_lookup(struct in6_addr *daddr, struct in6_addr *saddr,
571 flags |= RT6_LOOKUP_F_HAS_SADDR; 575 flags |= RT6_LOOKUP_F_HAS_SADDR;
572 } 576 }
573 577
574 dst = fib6_rule_lookup(&fl, flags, ip6_pol_route_lookup); 578 dst = fib6_rule_lookup(net, &fl, flags, ip6_pol_route_lookup);
575 if (dst->error == 0) 579 if (dst->error == 0)
576 return (struct rt6_info *) dst; 580 return (struct rt6_info *) dst;
577 581
@@ -604,7 +608,7 @@ static int __ip6_ins_rt(struct rt6_info *rt, struct nl_info *info)
604int ip6_ins_rt(struct rt6_info *rt) 608int ip6_ins_rt(struct rt6_info *rt)
605{ 609{
606 struct nl_info info = { 610 struct nl_info info = {
607 .nl_net = &init_net, 611 .nl_net = dev_net(rt->rt6i_dev),
608 }; 612 };
609 return __ip6_ins_rt(rt, &info); 613 return __ip6_ins_rt(rt, &info);
610} 614}
@@ -660,8 +664,8 @@ static struct rt6_info *rt6_alloc_clone(struct rt6_info *ort, struct in6_addr *d
660 return rt; 664 return rt;
661} 665}
662 666
663static struct rt6_info *ip6_pol_route(struct fib6_table *table, int oif, 667static struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table, int oif,
664 struct flowi *fl, int flags) 668 struct flowi *fl, int flags)
665{ 669{
666 struct fib6_node *fn; 670 struct fib6_node *fn;
667 struct rt6_info *rt, *nrt; 671 struct rt6_info *rt, *nrt;
@@ -680,8 +684,9 @@ restart_2:
680 684
681restart: 685restart:
682 rt = rt6_select(fn, oif, strict | reachable); 686 rt = rt6_select(fn, oif, strict | reachable);
683 BACKTRACK(&fl->fl6_src); 687
684 if (rt == &ip6_null_entry || 688 BACKTRACK(net, &fl->fl6_src);
689 if (rt == net->ipv6.ip6_null_entry ||
685 rt->rt6i_flags & RTF_CACHE) 690 rt->rt6i_flags & RTF_CACHE)
686 goto out; 691 goto out;
687 692
@@ -699,7 +704,7 @@ restart:
699 } 704 }
700 705
701 dst_release(&rt->u.dst); 706 dst_release(&rt->u.dst);
702 rt = nrt ? : &ip6_null_entry; 707 rt = nrt ? : net->ipv6.ip6_null_entry;
703 708
704 dst_hold(&rt->u.dst); 709 dst_hold(&rt->u.dst);
705 if (nrt) { 710 if (nrt) {
@@ -732,15 +737,16 @@ out2:
732 return rt; 737 return rt;
733} 738}
734 739
735static struct rt6_info *ip6_pol_route_input(struct fib6_table *table, 740static struct rt6_info *ip6_pol_route_input(struct net *net, struct fib6_table *table,
736 struct flowi *fl, int flags) 741 struct flowi *fl, int flags)
737{ 742{
738 return ip6_pol_route(table, fl->iif, fl, flags); 743 return ip6_pol_route(net, table, fl->iif, fl, flags);
739} 744}
740 745
741void ip6_route_input(struct sk_buff *skb) 746void ip6_route_input(struct sk_buff *skb)
742{ 747{
743 struct ipv6hdr *iph = ipv6_hdr(skb); 748 struct ipv6hdr *iph = ipv6_hdr(skb);
749 struct net *net = dev_net(skb->dev);
744 int flags = RT6_LOOKUP_F_HAS_SADDR; 750 int flags = RT6_LOOKUP_F_HAS_SADDR;
745 struct flowi fl = { 751 struct flowi fl = {
746 .iif = skb->dev->ifindex, 752 .iif = skb->dev->ifindex,
@@ -758,16 +764,17 @@ void ip6_route_input(struct sk_buff *skb)
758 if (rt6_need_strict(&iph->daddr)) 764 if (rt6_need_strict(&iph->daddr))
759 flags |= RT6_LOOKUP_F_IFACE; 765 flags |= RT6_LOOKUP_F_IFACE;
760 766
761 skb->dst = fib6_rule_lookup(&fl, flags, ip6_pol_route_input); 767 skb->dst = fib6_rule_lookup(net, &fl, flags, ip6_pol_route_input);
762} 768}
763 769
764static struct rt6_info *ip6_pol_route_output(struct fib6_table *table, 770static struct rt6_info *ip6_pol_route_output(struct net *net, struct fib6_table *table,
765 struct flowi *fl, int flags) 771 struct flowi *fl, int flags)
766{ 772{
767 return ip6_pol_route(table, fl->oif, fl, flags); 773 return ip6_pol_route(net, table, fl->oif, fl, flags);
768} 774}
769 775
770struct dst_entry * ip6_route_output(struct sock *sk, struct flowi *fl) 776struct dst_entry * ip6_route_output(struct net *net, struct sock *sk,
777 struct flowi *fl)
771{ 778{
772 int flags = 0; 779 int flags = 0;
773 780
@@ -776,8 +783,17 @@ struct dst_entry * ip6_route_output(struct sock *sk, struct flowi *fl)
776 783
777 if (!ipv6_addr_any(&fl->fl6_src)) 784 if (!ipv6_addr_any(&fl->fl6_src))
778 flags |= RT6_LOOKUP_F_HAS_SADDR; 785 flags |= RT6_LOOKUP_F_HAS_SADDR;
786 else if (sk) {
787 unsigned int prefs = inet6_sk(sk)->srcprefs;
788 if (prefs & IPV6_PREFER_SRC_TMP)
789 flags |= RT6_LOOKUP_F_SRCPREF_TMP;
790 if (prefs & IPV6_PREFER_SRC_PUBLIC)
791 flags |= RT6_LOOKUP_F_SRCPREF_PUBLIC;
792 if (prefs & IPV6_PREFER_SRC_COA)
793 flags |= RT6_LOOKUP_F_SRCPREF_COA;
794 }
779 795
780 return fib6_rule_lookup(fl, flags, ip6_pol_route_output); 796 return fib6_rule_lookup(net, fl, flags, ip6_pol_route_output);
781} 797}
782 798
783EXPORT_SYMBOL(ip6_route_output); 799EXPORT_SYMBOL(ip6_route_output);
@@ -886,12 +902,12 @@ static void ip6_rt_update_pmtu(struct dst_entry *dst, u32 mtu)
886 902
887static int ipv6_get_mtu(struct net_device *dev); 903static int ipv6_get_mtu(struct net_device *dev);
888 904
889static inline unsigned int ipv6_advmss(unsigned int mtu) 905static inline unsigned int ipv6_advmss(struct net *net, unsigned int mtu)
890{ 906{
891 mtu -= sizeof(struct ipv6hdr) + sizeof(struct tcphdr); 907 mtu -= sizeof(struct ipv6hdr) + sizeof(struct tcphdr);
892 908
893 if (mtu < init_net.ipv6.sysctl.ip6_rt_min_advmss) 909 if (mtu < net->ipv6.sysctl.ip6_rt_min_advmss)
894 mtu = init_net.ipv6.sysctl.ip6_rt_min_advmss; 910 mtu = net->ipv6.sysctl.ip6_rt_min_advmss;
895 911
896 /* 912 /*
897 * Maximal non-jumbo IPv6 payload is IPV6_MAXPLEN and 913 * Maximal non-jumbo IPv6 payload is IPV6_MAXPLEN and
@@ -904,21 +920,21 @@ static inline unsigned int ipv6_advmss(unsigned int mtu)
904 return mtu; 920 return mtu;
905} 921}
906 922
907static struct dst_entry *ndisc_dst_gc_list; 923static struct dst_entry *icmp6_dst_gc_list;
908static DEFINE_SPINLOCK(ndisc_lock); 924static DEFINE_SPINLOCK(icmp6_dst_lock);
909 925
910struct dst_entry *ndisc_dst_alloc(struct net_device *dev, 926struct dst_entry *icmp6_dst_alloc(struct net_device *dev,
911 struct neighbour *neigh, 927 struct neighbour *neigh,
912 struct in6_addr *addr, 928 const struct in6_addr *addr)
913 int (*output)(struct sk_buff *))
914{ 929{
915 struct rt6_info *rt; 930 struct rt6_info *rt;
916 struct inet6_dev *idev = in6_dev_get(dev); 931 struct inet6_dev *idev = in6_dev_get(dev);
932 struct net *net = dev_net(dev);
917 933
918 if (unlikely(idev == NULL)) 934 if (unlikely(idev == NULL))
919 return NULL; 935 return NULL;
920 936
921 rt = ip6_dst_alloc(); 937 rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
922 if (unlikely(rt == NULL)) { 938 if (unlikely(rt == NULL)) {
923 in6_dev_put(idev); 939 in6_dev_put(idev);
924 goto out; 940 goto out;
@@ -936,8 +952,8 @@ struct dst_entry *ndisc_dst_alloc(struct net_device *dev,
936 atomic_set(&rt->u.dst.__refcnt, 1); 952 atomic_set(&rt->u.dst.__refcnt, 1);
937 rt->u.dst.metrics[RTAX_HOPLIMIT-1] = 255; 953 rt->u.dst.metrics[RTAX_HOPLIMIT-1] = 255;
938 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev); 954 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev);
939 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_mtu(&rt->u.dst)); 955 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
940 rt->u.dst.output = output; 956 rt->u.dst.output = ip6_output;
941 957
942#if 0 /* there's no chance to use these for ndisc */ 958#if 0 /* there's no chance to use these for ndisc */
943 rt->u.dst.flags = ipv6_addr_type(addr) & IPV6_ADDR_UNICAST 959 rt->u.dst.flags = ipv6_addr_type(addr) & IPV6_ADDR_UNICAST
@@ -947,18 +963,18 @@ struct dst_entry *ndisc_dst_alloc(struct net_device *dev,
947 rt->rt6i_dst.plen = 128; 963 rt->rt6i_dst.plen = 128;
948#endif 964#endif
949 965
950 spin_lock_bh(&ndisc_lock); 966 spin_lock_bh(&icmp6_dst_lock);
951 rt->u.dst.next = ndisc_dst_gc_list; 967 rt->u.dst.next = icmp6_dst_gc_list;
952 ndisc_dst_gc_list = &rt->u.dst; 968 icmp6_dst_gc_list = &rt->u.dst;
953 spin_unlock_bh(&ndisc_lock); 969 spin_unlock_bh(&icmp6_dst_lock);
954 970
955 fib6_force_start_gc(); 971 fib6_force_start_gc(net);
956 972
957out: 973out:
958 return &rt->u.dst; 974 return &rt->u.dst;
959} 975}
960 976
961int ndisc_dst_gc(int *more) 977int icmp6_dst_gc(int *more)
962{ 978{
963 struct dst_entry *dst, *next, **pprev; 979 struct dst_entry *dst, *next, **pprev;
964 int freed; 980 int freed;
@@ -966,8 +982,8 @@ int ndisc_dst_gc(int *more)
966 next = NULL; 982 next = NULL;
967 freed = 0; 983 freed = 0;
968 984
969 spin_lock_bh(&ndisc_lock); 985 spin_lock_bh(&icmp6_dst_lock);
970 pprev = &ndisc_dst_gc_list; 986 pprev = &icmp6_dst_gc_list;
971 987
972 while ((dst = *pprev) != NULL) { 988 while ((dst = *pprev) != NULL) {
973 if (!atomic_read(&dst->__refcnt)) { 989 if (!atomic_read(&dst->__refcnt)) {
@@ -980,30 +996,33 @@ int ndisc_dst_gc(int *more)
980 } 996 }
981 } 997 }
982 998
983 spin_unlock_bh(&ndisc_lock); 999 spin_unlock_bh(&icmp6_dst_lock);
984 1000
985 return freed; 1001 return freed;
986} 1002}
987 1003
988static int ip6_dst_gc(struct dst_ops *ops) 1004static int ip6_dst_gc(struct dst_ops *ops)
989{ 1005{
990 static unsigned expire = 30*HZ;
991 static unsigned long last_gc;
992 unsigned long now = jiffies; 1006 unsigned long now = jiffies;
993 1007 struct net *net = ops->dst_net;
994 if (time_after(last_gc + init_net.ipv6.sysctl.ip6_rt_gc_min_interval, now) && 1008 int rt_min_interval = net->ipv6.sysctl.ip6_rt_gc_min_interval;
995 atomic_read(&ip6_dst_ops.entries) <= init_net.ipv6.sysctl.ip6_rt_max_size) 1009 int rt_max_size = net->ipv6.sysctl.ip6_rt_max_size;
1010 int rt_elasticity = net->ipv6.sysctl.ip6_rt_gc_elasticity;
1011 int rt_gc_timeout = net->ipv6.sysctl.ip6_rt_gc_timeout;
1012 unsigned long rt_last_gc = net->ipv6.ip6_rt_last_gc;
1013
1014 if (time_after(rt_last_gc + rt_min_interval, now) &&
1015 atomic_read(&ops->entries) <= rt_max_size)
996 goto out; 1016 goto out;
997 1017
998 expire++; 1018 net->ipv6.ip6_rt_gc_expire++;
999 fib6_run_gc(expire); 1019 fib6_run_gc(net->ipv6.ip6_rt_gc_expire, net);
1000 last_gc = now; 1020 net->ipv6.ip6_rt_last_gc = now;
1001 if (atomic_read(&ip6_dst_ops.entries) < ip6_dst_ops.gc_thresh) 1021 if (atomic_read(&ops->entries) < ops->gc_thresh)
1002 expire = init_net.ipv6.sysctl.ip6_rt_gc_timeout>>1; 1022 net->ipv6.ip6_rt_gc_expire = rt_gc_timeout>>1;
1003
1004out: 1023out:
1005 expire -= expire>>init_net.ipv6.sysctl.ip6_rt_gc_elasticity; 1024 net->ipv6.ip6_rt_gc_expire -= net->ipv6.ip6_rt_gc_expire>>rt_elasticity;
1006 return (atomic_read(&ip6_dst_ops.entries) > init_net.ipv6.sysctl.ip6_rt_max_size); 1025 return (atomic_read(&ops->entries) > rt_max_size);
1007} 1026}
1008 1027
1009/* Clean host part of a prefix. Not necessary in radix tree, 1028/* Clean host part of a prefix. Not necessary in radix tree,
@@ -1025,15 +1044,17 @@ static int ipv6_get_mtu(struct net_device *dev)
1025 return mtu; 1044 return mtu;
1026} 1045}
1027 1046
1028int ipv6_get_hoplimit(struct net_device *dev) 1047int ip6_dst_hoplimit(struct dst_entry *dst)
1029{ 1048{
1030 int hoplimit = ipv6_devconf.hop_limit; 1049 int hoplimit = dst_metric(dst, RTAX_HOPLIMIT);
1031 struct inet6_dev *idev; 1050 if (hoplimit < 0) {
1032 1051 struct net_device *dev = dst->dev;
1033 idev = in6_dev_get(dev); 1052 struct inet6_dev *idev = in6_dev_get(dev);
1034 if (idev) { 1053 if (idev) {
1035 hoplimit = idev->cnf.hop_limit; 1054 hoplimit = idev->cnf.hop_limit;
1036 in6_dev_put(idev); 1055 in6_dev_put(idev);
1056 } else
1057 hoplimit = ipv6_devconf.hop_limit;
1037 } 1058 }
1038 return hoplimit; 1059 return hoplimit;
1039} 1060}
@@ -1045,6 +1066,7 @@ int ipv6_get_hoplimit(struct net_device *dev)
1045int ip6_route_add(struct fib6_config *cfg) 1066int ip6_route_add(struct fib6_config *cfg)
1046{ 1067{
1047 int err; 1068 int err;
1069 struct net *net = cfg->fc_nlinfo.nl_net;
1048 struct rt6_info *rt = NULL; 1070 struct rt6_info *rt = NULL;
1049 struct net_device *dev = NULL; 1071 struct net_device *dev = NULL;
1050 struct inet6_dev *idev = NULL; 1072 struct inet6_dev *idev = NULL;
@@ -1059,7 +1081,7 @@ int ip6_route_add(struct fib6_config *cfg)
1059#endif 1081#endif
1060 if (cfg->fc_ifindex) { 1082 if (cfg->fc_ifindex) {
1061 err = -ENODEV; 1083 err = -ENODEV;
1062 dev = dev_get_by_index(&init_net, cfg->fc_ifindex); 1084 dev = dev_get_by_index(net, cfg->fc_ifindex);
1063 if (!dev) 1085 if (!dev)
1064 goto out; 1086 goto out;
1065 idev = in6_dev_get(dev); 1087 idev = in6_dev_get(dev);
@@ -1070,13 +1092,13 @@ int ip6_route_add(struct fib6_config *cfg)
1070 if (cfg->fc_metric == 0) 1092 if (cfg->fc_metric == 0)
1071 cfg->fc_metric = IP6_RT_PRIO_USER; 1093 cfg->fc_metric = IP6_RT_PRIO_USER;
1072 1094
1073 table = fib6_new_table(cfg->fc_table); 1095 table = fib6_new_table(net, cfg->fc_table);
1074 if (table == NULL) { 1096 if (table == NULL) {
1075 err = -ENOBUFS; 1097 err = -ENOBUFS;
1076 goto out; 1098 goto out;
1077 } 1099 }
1078 1100
1079 rt = ip6_dst_alloc(); 1101 rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
1080 1102
1081 if (rt == NULL) { 1103 if (rt == NULL) {
1082 err = -ENOMEM; 1104 err = -ENOMEM;
@@ -1117,12 +1139,12 @@ int ip6_route_add(struct fib6_config *cfg)
1117 if ((cfg->fc_flags & RTF_REJECT) || 1139 if ((cfg->fc_flags & RTF_REJECT) ||
1118 (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK))) { 1140 (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK))) {
1119 /* hold loopback dev/idev if we haven't done so. */ 1141 /* hold loopback dev/idev if we haven't done so. */
1120 if (dev != init_net.loopback_dev) { 1142 if (dev != net->loopback_dev) {
1121 if (dev) { 1143 if (dev) {
1122 dev_put(dev); 1144 dev_put(dev);
1123 in6_dev_put(idev); 1145 in6_dev_put(idev);
1124 } 1146 }
1125 dev = init_net.loopback_dev; 1147 dev = net->loopback_dev;
1126 dev_hold(dev); 1148 dev_hold(dev);
1127 idev = in6_dev_get(dev); 1149 idev = in6_dev_get(dev);
1128 if (!idev) { 1150 if (!idev) {
@@ -1159,7 +1181,7 @@ int ip6_route_add(struct fib6_config *cfg)
1159 if (!(gwa_type&IPV6_ADDR_UNICAST)) 1181 if (!(gwa_type&IPV6_ADDR_UNICAST))
1160 goto out; 1182 goto out;
1161 1183
1162 grt = rt6_lookup(gw_addr, NULL, cfg->fc_ifindex, 1); 1184 grt = rt6_lookup(net, gw_addr, NULL, cfg->fc_ifindex, 1);
1163 1185
1164 err = -EHOSTUNREACH; 1186 err = -EHOSTUNREACH;
1165 if (grt == NULL) 1187 if (grt == NULL)
@@ -1226,10 +1248,13 @@ install_route:
1226 if (!rt->u.dst.metrics[RTAX_MTU-1]) 1248 if (!rt->u.dst.metrics[RTAX_MTU-1])
1227 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(dev); 1249 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(dev);
1228 if (!rt->u.dst.metrics[RTAX_ADVMSS-1]) 1250 if (!rt->u.dst.metrics[RTAX_ADVMSS-1])
1229 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_mtu(&rt->u.dst)); 1251 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
1230 rt->u.dst.dev = dev; 1252 rt->u.dst.dev = dev;
1231 rt->rt6i_idev = idev; 1253 rt->rt6i_idev = idev;
1232 rt->rt6i_table = table; 1254 rt->rt6i_table = table;
1255
1256 cfg->fc_nlinfo.nl_net = dev_net(dev);
1257
1233 return __ip6_ins_rt(rt, &cfg->fc_nlinfo); 1258 return __ip6_ins_rt(rt, &cfg->fc_nlinfo);
1234 1259
1235out: 1260out:
@@ -1246,8 +1271,9 @@ static int __ip6_del_rt(struct rt6_info *rt, struct nl_info *info)
1246{ 1271{
1247 int err; 1272 int err;
1248 struct fib6_table *table; 1273 struct fib6_table *table;
1274 struct net *net = dev_net(rt->rt6i_dev);
1249 1275
1250 if (rt == &ip6_null_entry) 1276 if (rt == net->ipv6.ip6_null_entry)
1251 return -ENOENT; 1277 return -ENOENT;
1252 1278
1253 table = rt->rt6i_table; 1279 table = rt->rt6i_table;
@@ -1264,7 +1290,7 @@ static int __ip6_del_rt(struct rt6_info *rt, struct nl_info *info)
1264int ip6_del_rt(struct rt6_info *rt) 1290int ip6_del_rt(struct rt6_info *rt)
1265{ 1291{
1266 struct nl_info info = { 1292 struct nl_info info = {
1267 .nl_net = &init_net, 1293 .nl_net = dev_net(rt->rt6i_dev),
1268 }; 1294 };
1269 return __ip6_del_rt(rt, &info); 1295 return __ip6_del_rt(rt, &info);
1270} 1296}
@@ -1276,7 +1302,7 @@ static int ip6_route_del(struct fib6_config *cfg)
1276 struct rt6_info *rt; 1302 struct rt6_info *rt;
1277 int err = -ESRCH; 1303 int err = -ESRCH;
1278 1304
1279 table = fib6_get_table(cfg->fc_table); 1305 table = fib6_get_table(cfg->fc_nlinfo.nl_net, cfg->fc_table);
1280 if (table == NULL) 1306 if (table == NULL)
1281 return err; 1307 return err;
1282 1308
@@ -1316,7 +1342,8 @@ struct ip6rd_flowi {
1316 struct in6_addr gateway; 1342 struct in6_addr gateway;
1317}; 1343};
1318 1344
1319static struct rt6_info *__ip6_route_redirect(struct fib6_table *table, 1345static struct rt6_info *__ip6_route_redirect(struct net *net,
1346 struct fib6_table *table,
1320 struct flowi *fl, 1347 struct flowi *fl,
1321 int flags) 1348 int flags)
1322{ 1349{
@@ -1359,8 +1386,8 @@ restart:
1359 } 1386 }
1360 1387
1361 if (!rt) 1388 if (!rt)
1362 rt = &ip6_null_entry; 1389 rt = net->ipv6.ip6_null_entry;
1363 BACKTRACK(&fl->fl6_src); 1390 BACKTRACK(net, &fl->fl6_src);
1364out: 1391out:
1365 dst_hold(&rt->u.dst); 1392 dst_hold(&rt->u.dst);
1366 1393
@@ -1375,6 +1402,7 @@ static struct rt6_info *ip6_route_redirect(struct in6_addr *dest,
1375 struct net_device *dev) 1402 struct net_device *dev)
1376{ 1403{
1377 int flags = RT6_LOOKUP_F_HAS_SADDR; 1404 int flags = RT6_LOOKUP_F_HAS_SADDR;
1405 struct net *net = dev_net(dev);
1378 struct ip6rd_flowi rdfl = { 1406 struct ip6rd_flowi rdfl = {
1379 .fl = { 1407 .fl = {
1380 .oif = dev->ifindex, 1408 .oif = dev->ifindex,
@@ -1391,7 +1419,8 @@ static struct rt6_info *ip6_route_redirect(struct in6_addr *dest,
1391 if (rt6_need_strict(dest)) 1419 if (rt6_need_strict(dest))
1392 flags |= RT6_LOOKUP_F_IFACE; 1420 flags |= RT6_LOOKUP_F_IFACE;
1393 1421
1394 return (struct rt6_info *)fib6_rule_lookup((struct flowi *)&rdfl, flags, __ip6_route_redirect); 1422 return (struct rt6_info *)fib6_rule_lookup(net, (struct flowi *)&rdfl,
1423 flags, __ip6_route_redirect);
1395} 1424}
1396 1425
1397void rt6_redirect(struct in6_addr *dest, struct in6_addr *src, 1426void rt6_redirect(struct in6_addr *dest, struct in6_addr *src,
@@ -1400,10 +1429,11 @@ void rt6_redirect(struct in6_addr *dest, struct in6_addr *src,
1400{ 1429{
1401 struct rt6_info *rt, *nrt = NULL; 1430 struct rt6_info *rt, *nrt = NULL;
1402 struct netevent_redirect netevent; 1431 struct netevent_redirect netevent;
1432 struct net *net = dev_net(neigh->dev);
1403 1433
1404 rt = ip6_route_redirect(dest, src, saddr, neigh->dev); 1434 rt = ip6_route_redirect(dest, src, saddr, neigh->dev);
1405 1435
1406 if (rt == &ip6_null_entry) { 1436 if (rt == net->ipv6.ip6_null_entry) {
1407 if (net_ratelimit()) 1437 if (net_ratelimit())
1408 printk(KERN_DEBUG "rt6_redirect: source isn't a valid nexthop " 1438 printk(KERN_DEBUG "rt6_redirect: source isn't a valid nexthop "
1409 "for redirect target\n"); 1439 "for redirect target\n");
@@ -1448,7 +1478,8 @@ void rt6_redirect(struct in6_addr *dest, struct in6_addr *src,
1448 nrt->rt6i_nexthop = neigh_clone(neigh); 1478 nrt->rt6i_nexthop = neigh_clone(neigh);
1449 /* Reset pmtu, it may be better */ 1479 /* Reset pmtu, it may be better */
1450 nrt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(neigh->dev); 1480 nrt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(neigh->dev);
1451 nrt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_mtu(&nrt->u.dst)); 1481 nrt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dev_net(neigh->dev),
1482 dst_mtu(&nrt->u.dst));
1452 1483
1453 if (ip6_ins_rt(nrt)) 1484 if (ip6_ins_rt(nrt))
1454 goto out; 1485 goto out;
@@ -1476,9 +1507,10 @@ void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
1476 struct net_device *dev, u32 pmtu) 1507 struct net_device *dev, u32 pmtu)
1477{ 1508{
1478 struct rt6_info *rt, *nrt; 1509 struct rt6_info *rt, *nrt;
1510 struct net *net = dev_net(dev);
1479 int allfrag = 0; 1511 int allfrag = 0;
1480 1512
1481 rt = rt6_lookup(daddr, saddr, dev->ifindex, 0); 1513 rt = rt6_lookup(net, daddr, saddr, dev->ifindex, 0);
1482 if (rt == NULL) 1514 if (rt == NULL)
1483 return; 1515 return;
1484 1516
@@ -1511,7 +1543,7 @@ void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
1511 rt->u.dst.metrics[RTAX_MTU-1] = pmtu; 1543 rt->u.dst.metrics[RTAX_MTU-1] = pmtu;
1512 if (allfrag) 1544 if (allfrag)
1513 rt->u.dst.metrics[RTAX_FEATURES-1] |= RTAX_FEATURE_ALLFRAG; 1545 rt->u.dst.metrics[RTAX_FEATURES-1] |= RTAX_FEATURE_ALLFRAG;
1514 dst_set_expires(&rt->u.dst, init_net.ipv6.sysctl.ip6_rt_mtu_expires); 1546 dst_set_expires(&rt->u.dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
1515 rt->rt6i_flags |= RTF_MODIFIED|RTF_EXPIRES; 1547 rt->rt6i_flags |= RTF_MODIFIED|RTF_EXPIRES;
1516 goto out; 1548 goto out;
1517 } 1549 }
@@ -1537,7 +1569,7 @@ void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
1537 * which is 10 mins. After 10 mins the decreased pmtu is expired 1569 * which is 10 mins. After 10 mins the decreased pmtu is expired
1538 * and detecting PMTU increase will be automatically happened. 1570 * and detecting PMTU increase will be automatically happened.
1539 */ 1571 */
1540 dst_set_expires(&nrt->u.dst, init_net.ipv6.sysctl.ip6_rt_mtu_expires); 1572 dst_set_expires(&nrt->u.dst, net->ipv6.sysctl.ip6_rt_mtu_expires);
1541 nrt->rt6i_flags |= RTF_DYNAMIC|RTF_EXPIRES; 1573 nrt->rt6i_flags |= RTF_DYNAMIC|RTF_EXPIRES;
1542 1574
1543 ip6_ins_rt(nrt); 1575 ip6_ins_rt(nrt);
@@ -1552,7 +1584,8 @@ out:
1552 1584
1553static struct rt6_info * ip6_rt_copy(struct rt6_info *ort) 1585static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
1554{ 1586{
1555 struct rt6_info *rt = ip6_dst_alloc(); 1587 struct net *net = dev_net(ort->rt6i_dev);
1588 struct rt6_info *rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
1556 1589
1557 if (rt) { 1590 if (rt) {
1558 rt->u.dst.input = ort->u.dst.input; 1591 rt->u.dst.input = ort->u.dst.input;
@@ -1583,14 +1616,15 @@ static struct rt6_info * ip6_rt_copy(struct rt6_info *ort)
1583} 1616}
1584 1617
1585#ifdef CONFIG_IPV6_ROUTE_INFO 1618#ifdef CONFIG_IPV6_ROUTE_INFO
1586static struct rt6_info *rt6_get_route_info(struct in6_addr *prefix, int prefixlen, 1619static struct rt6_info *rt6_get_route_info(struct net *net,
1620 struct in6_addr *prefix, int prefixlen,
1587 struct in6_addr *gwaddr, int ifindex) 1621 struct in6_addr *gwaddr, int ifindex)
1588{ 1622{
1589 struct fib6_node *fn; 1623 struct fib6_node *fn;
1590 struct rt6_info *rt = NULL; 1624 struct rt6_info *rt = NULL;
1591 struct fib6_table *table; 1625 struct fib6_table *table;
1592 1626
1593 table = fib6_get_table(RT6_TABLE_INFO); 1627 table = fib6_get_table(net, RT6_TABLE_INFO);
1594 if (table == NULL) 1628 if (table == NULL)
1595 return NULL; 1629 return NULL;
1596 1630
@@ -1614,7 +1648,8 @@ out:
1614 return rt; 1648 return rt;
1615} 1649}
1616 1650
1617static struct rt6_info *rt6_add_route_info(struct in6_addr *prefix, int prefixlen, 1651static struct rt6_info *rt6_add_route_info(struct net *net,
1652 struct in6_addr *prefix, int prefixlen,
1618 struct in6_addr *gwaddr, int ifindex, 1653 struct in6_addr *gwaddr, int ifindex,
1619 unsigned pref) 1654 unsigned pref)
1620{ 1655{
@@ -1625,6 +1660,9 @@ static struct rt6_info *rt6_add_route_info(struct in6_addr *prefix, int prefixle
1625 .fc_dst_len = prefixlen, 1660 .fc_dst_len = prefixlen,
1626 .fc_flags = RTF_GATEWAY | RTF_ADDRCONF | RTF_ROUTEINFO | 1661 .fc_flags = RTF_GATEWAY | RTF_ADDRCONF | RTF_ROUTEINFO |
1627 RTF_UP | RTF_PREF(pref), 1662 RTF_UP | RTF_PREF(pref),
1663 .fc_nlinfo.pid = 0,
1664 .fc_nlinfo.nlh = NULL,
1665 .fc_nlinfo.nl_net = net,
1628 }; 1666 };
1629 1667
1630 ipv6_addr_copy(&cfg.fc_dst, prefix); 1668 ipv6_addr_copy(&cfg.fc_dst, prefix);
@@ -1636,7 +1674,7 @@ static struct rt6_info *rt6_add_route_info(struct in6_addr *prefix, int prefixle
1636 1674
1637 ip6_route_add(&cfg); 1675 ip6_route_add(&cfg);
1638 1676
1639 return rt6_get_route_info(prefix, prefixlen, gwaddr, ifindex); 1677 return rt6_get_route_info(net, prefix, prefixlen, gwaddr, ifindex);
1640} 1678}
1641#endif 1679#endif
1642 1680
@@ -1645,7 +1683,7 @@ struct rt6_info *rt6_get_dflt_router(struct in6_addr *addr, struct net_device *d
1645 struct rt6_info *rt; 1683 struct rt6_info *rt;
1646 struct fib6_table *table; 1684 struct fib6_table *table;
1647 1685
1648 table = fib6_get_table(RT6_TABLE_DFLT); 1686 table = fib6_get_table(dev_net(dev), RT6_TABLE_DFLT);
1649 if (table == NULL) 1687 if (table == NULL)
1650 return NULL; 1688 return NULL;
1651 1689
@@ -1662,8 +1700,6 @@ struct rt6_info *rt6_get_dflt_router(struct in6_addr *addr, struct net_device *d
1662 return rt; 1700 return rt;
1663} 1701}
1664 1702
1665EXPORT_SYMBOL(rt6_get_dflt_router);
1666
1667struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr, 1703struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr,
1668 struct net_device *dev, 1704 struct net_device *dev,
1669 unsigned int pref) 1705 unsigned int pref)
@@ -1674,6 +1710,9 @@ struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr,
1674 .fc_ifindex = dev->ifindex, 1710 .fc_ifindex = dev->ifindex,
1675 .fc_flags = RTF_GATEWAY | RTF_ADDRCONF | RTF_DEFAULT | 1711 .fc_flags = RTF_GATEWAY | RTF_ADDRCONF | RTF_DEFAULT |
1676 RTF_UP | RTF_EXPIRES | RTF_PREF(pref), 1712 RTF_UP | RTF_EXPIRES | RTF_PREF(pref),
1713 .fc_nlinfo.pid = 0,
1714 .fc_nlinfo.nlh = NULL,
1715 .fc_nlinfo.nl_net = dev_net(dev),
1677 }; 1716 };
1678 1717
1679 ipv6_addr_copy(&cfg.fc_gateway, gwaddr); 1718 ipv6_addr_copy(&cfg.fc_gateway, gwaddr);
@@ -1683,13 +1722,13 @@ struct rt6_info *rt6_add_dflt_router(struct in6_addr *gwaddr,
1683 return rt6_get_dflt_router(gwaddr, dev); 1722 return rt6_get_dflt_router(gwaddr, dev);
1684} 1723}
1685 1724
1686void rt6_purge_dflt_routers(void) 1725void rt6_purge_dflt_routers(struct net *net)
1687{ 1726{
1688 struct rt6_info *rt; 1727 struct rt6_info *rt;
1689 struct fib6_table *table; 1728 struct fib6_table *table;
1690 1729
1691 /* NOTE: Keep consistent with rt6_get_dflt_router */ 1730 /* NOTE: Keep consistent with rt6_get_dflt_router */
1692 table = fib6_get_table(RT6_TABLE_DFLT); 1731 table = fib6_get_table(net, RT6_TABLE_DFLT);
1693 if (table == NULL) 1732 if (table == NULL)
1694 return; 1733 return;
1695 1734
@@ -1706,7 +1745,8 @@ restart:
1706 read_unlock_bh(&table->tb6_lock); 1745 read_unlock_bh(&table->tb6_lock);
1707} 1746}
1708 1747
1709static void rtmsg_to_fib6_config(struct in6_rtmsg *rtmsg, 1748static void rtmsg_to_fib6_config(struct net *net,
1749 struct in6_rtmsg *rtmsg,
1710 struct fib6_config *cfg) 1750 struct fib6_config *cfg)
1711{ 1751{
1712 memset(cfg, 0, sizeof(*cfg)); 1752 memset(cfg, 0, sizeof(*cfg));
@@ -1719,14 +1759,14 @@ static void rtmsg_to_fib6_config(struct in6_rtmsg *rtmsg,
1719 cfg->fc_src_len = rtmsg->rtmsg_src_len; 1759 cfg->fc_src_len = rtmsg->rtmsg_src_len;
1720 cfg->fc_flags = rtmsg->rtmsg_flags; 1760 cfg->fc_flags = rtmsg->rtmsg_flags;
1721 1761
1722 cfg->fc_nlinfo.nl_net = &init_net; 1762 cfg->fc_nlinfo.nl_net = net;
1723 1763
1724 ipv6_addr_copy(&cfg->fc_dst, &rtmsg->rtmsg_dst); 1764 ipv6_addr_copy(&cfg->fc_dst, &rtmsg->rtmsg_dst);
1725 ipv6_addr_copy(&cfg->fc_src, &rtmsg->rtmsg_src); 1765 ipv6_addr_copy(&cfg->fc_src, &rtmsg->rtmsg_src);
1726 ipv6_addr_copy(&cfg->fc_gateway, &rtmsg->rtmsg_gateway); 1766 ipv6_addr_copy(&cfg->fc_gateway, &rtmsg->rtmsg_gateway);
1727} 1767}
1728 1768
1729int ipv6_route_ioctl(unsigned int cmd, void __user *arg) 1769int ipv6_route_ioctl(struct net *net, unsigned int cmd, void __user *arg)
1730{ 1770{
1731 struct fib6_config cfg; 1771 struct fib6_config cfg;
1732 struct in6_rtmsg rtmsg; 1772 struct in6_rtmsg rtmsg;
@@ -1742,7 +1782,7 @@ int ipv6_route_ioctl(unsigned int cmd, void __user *arg)
1742 if (err) 1782 if (err)
1743 return -EFAULT; 1783 return -EFAULT;
1744 1784
1745 rtmsg_to_fib6_config(&rtmsg, &cfg); 1785 rtmsg_to_fib6_config(net, &rtmsg, &cfg);
1746 1786
1747 rtnl_lock(); 1787 rtnl_lock();
1748 switch (cmd) { 1788 switch (cmd) {
@@ -1821,21 +1861,22 @@ struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
1821 const struct in6_addr *addr, 1861 const struct in6_addr *addr,
1822 int anycast) 1862 int anycast)
1823{ 1863{
1824 struct rt6_info *rt = ip6_dst_alloc(); 1864 struct net *net = dev_net(idev->dev);
1865 struct rt6_info *rt = ip6_dst_alloc(net->ipv6.ip6_dst_ops);
1825 1866
1826 if (rt == NULL) 1867 if (rt == NULL)
1827 return ERR_PTR(-ENOMEM); 1868 return ERR_PTR(-ENOMEM);
1828 1869
1829 dev_hold(init_net.loopback_dev); 1870 dev_hold(net->loopback_dev);
1830 in6_dev_hold(idev); 1871 in6_dev_hold(idev);
1831 1872
1832 rt->u.dst.flags = DST_HOST; 1873 rt->u.dst.flags = DST_HOST;
1833 rt->u.dst.input = ip6_input; 1874 rt->u.dst.input = ip6_input;
1834 rt->u.dst.output = ip6_output; 1875 rt->u.dst.output = ip6_output;
1835 rt->rt6i_dev = init_net.loopback_dev; 1876 rt->rt6i_dev = net->loopback_dev;
1836 rt->rt6i_idev = idev; 1877 rt->rt6i_idev = idev;
1837 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev); 1878 rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev);
1838 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_mtu(&rt->u.dst)); 1879 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, dst_mtu(&rt->u.dst));
1839 rt->u.dst.metrics[RTAX_HOPLIMIT-1] = -1; 1880 rt->u.dst.metrics[RTAX_HOPLIMIT-1] = -1;
1840 rt->u.dst.obsolete = -1; 1881 rt->u.dst.obsolete = -1;
1841 1882
@@ -1852,26 +1893,39 @@ struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
1852 1893
1853 ipv6_addr_copy(&rt->rt6i_dst.addr, addr); 1894 ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
1854 rt->rt6i_dst.plen = 128; 1895 rt->rt6i_dst.plen = 128;
1855 rt->rt6i_table = fib6_get_table(RT6_TABLE_LOCAL); 1896 rt->rt6i_table = fib6_get_table(net, RT6_TABLE_LOCAL);
1856 1897
1857 atomic_set(&rt->u.dst.__refcnt, 1); 1898 atomic_set(&rt->u.dst.__refcnt, 1);
1858 1899
1859 return rt; 1900 return rt;
1860} 1901}
1861 1902
1903struct arg_dev_net {
1904 struct net_device *dev;
1905 struct net *net;
1906};
1907
1862static int fib6_ifdown(struct rt6_info *rt, void *arg) 1908static int fib6_ifdown(struct rt6_info *rt, void *arg)
1863{ 1909{
1864 if (((void*)rt->rt6i_dev == arg || arg == NULL) && 1910 struct net_device *dev = ((struct arg_dev_net *)arg)->dev;
1865 rt != &ip6_null_entry) { 1911 struct net *net = ((struct arg_dev_net *)arg)->net;
1912
1913 if (((void *)rt->rt6i_dev == dev || dev == NULL) &&
1914 rt != net->ipv6.ip6_null_entry) {
1866 RT6_TRACE("deleted by ifdown %p\n", rt); 1915 RT6_TRACE("deleted by ifdown %p\n", rt);
1867 return -1; 1916 return -1;
1868 } 1917 }
1869 return 0; 1918 return 0;
1870} 1919}
1871 1920
1872void rt6_ifdown(struct net_device *dev) 1921void rt6_ifdown(struct net *net, struct net_device *dev)
1873{ 1922{
1874 fib6_clean_all(fib6_ifdown, 0, dev); 1923 struct arg_dev_net adn = {
1924 .dev = dev,
1925 .net = net,
1926 };
1927
1928 fib6_clean_all(net, fib6_ifdown, 0, &adn);
1875} 1929}
1876 1930
1877struct rt6_mtu_change_arg 1931struct rt6_mtu_change_arg
@@ -1884,6 +1938,7 @@ static int rt6_mtu_change_route(struct rt6_info *rt, void *p_arg)
1884{ 1938{
1885 struct rt6_mtu_change_arg *arg = (struct rt6_mtu_change_arg *) p_arg; 1939 struct rt6_mtu_change_arg *arg = (struct rt6_mtu_change_arg *) p_arg;
1886 struct inet6_dev *idev; 1940 struct inet6_dev *idev;
1941 struct net *net = dev_net(arg->dev);
1887 1942
1888 /* In IPv6 pmtu discovery is not optional, 1943 /* In IPv6 pmtu discovery is not optional,
1889 so that RTAX_MTU lock cannot disable it. 1944 so that RTAX_MTU lock cannot disable it.
@@ -1915,7 +1970,7 @@ static int rt6_mtu_change_route(struct rt6_info *rt, void *p_arg)
1915 (dst_mtu(&rt->u.dst) < arg->mtu && 1970 (dst_mtu(&rt->u.dst) < arg->mtu &&
1916 dst_mtu(&rt->u.dst) == idev->cnf.mtu6))) { 1971 dst_mtu(&rt->u.dst) == idev->cnf.mtu6))) {
1917 rt->u.dst.metrics[RTAX_MTU-1] = arg->mtu; 1972 rt->u.dst.metrics[RTAX_MTU-1] = arg->mtu;
1918 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(arg->mtu); 1973 rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(net, arg->mtu);
1919 } 1974 }
1920 return 0; 1975 return 0;
1921} 1976}
@@ -1927,7 +1982,7 @@ void rt6_mtu_change(struct net_device *dev, unsigned mtu)
1927 .mtu = mtu, 1982 .mtu = mtu,
1928 }; 1983 };
1929 1984
1930 fib6_clean_all(rt6_mtu_change_route, 0, &arg); 1985 fib6_clean_all(dev_net(dev), rt6_mtu_change_route, 0, &arg);
1931} 1986}
1932 1987
1933static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = { 1988static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
@@ -1964,7 +2019,7 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
1964 2019
1965 cfg->fc_nlinfo.pid = NETLINK_CB(skb).pid; 2020 cfg->fc_nlinfo.pid = NETLINK_CB(skb).pid;
1966 cfg->fc_nlinfo.nlh = nlh; 2021 cfg->fc_nlinfo.nlh = nlh;
1967 cfg->fc_nlinfo.nl_net = skb->sk->sk_net; 2022 cfg->fc_nlinfo.nl_net = sock_net(skb->sk);
1968 2023
1969 if (tb[RTA_GATEWAY]) { 2024 if (tb[RTA_GATEWAY]) {
1970 nla_memcpy(&cfg->fc_gateway, tb[RTA_GATEWAY], 16); 2025 nla_memcpy(&cfg->fc_gateway, tb[RTA_GATEWAY], 16);
@@ -2010,13 +2065,9 @@ errout:
2010 2065
2011static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg) 2066static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
2012{ 2067{
2013 struct net *net = skb->sk->sk_net;
2014 struct fib6_config cfg; 2068 struct fib6_config cfg;
2015 int err; 2069 int err;
2016 2070
2017 if (net != &init_net)
2018 return -EINVAL;
2019
2020 err = rtm_to_fib6_config(skb, nlh, &cfg); 2071 err = rtm_to_fib6_config(skb, nlh, &cfg);
2021 if (err < 0) 2072 if (err < 0)
2022 return err; 2073 return err;
@@ -2026,13 +2077,9 @@ static int inet6_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *a
2026 2077
2027static int inet6_rtm_newroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg) 2078static int inet6_rtm_newroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
2028{ 2079{
2029 struct net *net = skb->sk->sk_net;
2030 struct fib6_config cfg; 2080 struct fib6_config cfg;
2031 int err; 2081 int err;
2032 2082
2033 if (net != &init_net)
2034 return -EINVAL;
2035
2036 err = rtm_to_fib6_config(skb, nlh, &cfg); 2083 err = rtm_to_fib6_config(skb, nlh, &cfg);
2037 if (err < 0) 2084 if (err < 0)
2038 return err; 2085 return err;
@@ -2058,7 +2105,7 @@ static inline size_t rt6_nlmsg_size(void)
2058static int rt6_fill_node(struct sk_buff *skb, struct rt6_info *rt, 2105static int rt6_fill_node(struct sk_buff *skb, struct rt6_info *rt,
2059 struct in6_addr *dst, struct in6_addr *src, 2106 struct in6_addr *dst, struct in6_addr *src,
2060 int iif, int type, u32 pid, u32 seq, 2107 int iif, int type, u32 pid, u32 seq,
2061 int prefix, unsigned int flags) 2108 int prefix, int nowait, unsigned int flags)
2062{ 2109{
2063 struct rtmsg *rtm; 2110 struct rtmsg *rtm;
2064 struct nlmsghdr *nlh; 2111 struct nlmsghdr *nlh;
@@ -2118,11 +2165,27 @@ static int rt6_fill_node(struct sk_buff *skb, struct rt6_info *rt,
2118 } else if (rtm->rtm_src_len) 2165 } else if (rtm->rtm_src_len)
2119 NLA_PUT(skb, RTA_SRC, 16, &rt->rt6i_src.addr); 2166 NLA_PUT(skb, RTA_SRC, 16, &rt->rt6i_src.addr);
2120#endif 2167#endif
2121 if (iif) 2168 if (iif) {
2122 NLA_PUT_U32(skb, RTA_IIF, iif); 2169#ifdef CONFIG_IPV6_MROUTE
2123 else if (dst) { 2170 if (ipv6_addr_is_multicast(&rt->rt6i_dst.addr)) {
2171 int err = ip6mr_get_route(skb, rtm, nowait);
2172 if (err <= 0) {
2173 if (!nowait) {
2174 if (err == 0)
2175 return 0;
2176 goto nla_put_failure;
2177 } else {
2178 if (err == -EMSGSIZE)
2179 goto nla_put_failure;
2180 }
2181 }
2182 } else
2183#endif
2184 NLA_PUT_U32(skb, RTA_IIF, iif);
2185 } else if (dst) {
2124 struct in6_addr saddr_buf; 2186 struct in6_addr saddr_buf;
2125 if (ipv6_get_saddr(&rt->u.dst, dst, &saddr_buf) == 0) 2187 if (ipv6_dev_get_saddr(ip6_dst_idev(&rt->u.dst)->dev,
2188 dst, 0, &saddr_buf) == 0)
2126 NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf); 2189 NLA_PUT(skb, RTA_PREFSRC, 16, &saddr_buf);
2127 } 2190 }
2128 2191
@@ -2162,12 +2225,12 @@ int rt6_dump_route(struct rt6_info *rt, void *p_arg)
2162 2225
2163 return rt6_fill_node(arg->skb, rt, NULL, NULL, 0, RTM_NEWROUTE, 2226 return rt6_fill_node(arg->skb, rt, NULL, NULL, 0, RTM_NEWROUTE,
2164 NETLINK_CB(arg->cb->skb).pid, arg->cb->nlh->nlmsg_seq, 2227 NETLINK_CB(arg->cb->skb).pid, arg->cb->nlh->nlmsg_seq,
2165 prefix, NLM_F_MULTI); 2228 prefix, 0, NLM_F_MULTI);
2166} 2229}
2167 2230
2168static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void *arg) 2231static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void *arg)
2169{ 2232{
2170 struct net *net = in_skb->sk->sk_net; 2233 struct net *net = sock_net(in_skb->sk);
2171 struct nlattr *tb[RTA_MAX+1]; 2234 struct nlattr *tb[RTA_MAX+1];
2172 struct rt6_info *rt; 2235 struct rt6_info *rt;
2173 struct sk_buff *skb; 2236 struct sk_buff *skb;
@@ -2175,9 +2238,6 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void
2175 struct flowi fl; 2238 struct flowi fl;
2176 int err, iif = 0; 2239 int err, iif = 0;
2177 2240
2178 if (net != &init_net)
2179 return -EINVAL;
2180
2181 err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy); 2241 err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv6_policy);
2182 if (err < 0) 2242 if (err < 0)
2183 goto errout; 2243 goto errout;
@@ -2207,7 +2267,7 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void
2207 2267
2208 if (iif) { 2268 if (iif) {
2209 struct net_device *dev; 2269 struct net_device *dev;
2210 dev = __dev_get_by_index(&init_net, iif); 2270 dev = __dev_get_by_index(net, iif);
2211 if (!dev) { 2271 if (!dev) {
2212 err = -ENODEV; 2272 err = -ENODEV;
2213 goto errout; 2273 goto errout;
@@ -2226,18 +2286,18 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void
2226 skb_reset_mac_header(skb); 2286 skb_reset_mac_header(skb);
2227 skb_reserve(skb, MAX_HEADER + sizeof(struct ipv6hdr)); 2287 skb_reserve(skb, MAX_HEADER + sizeof(struct ipv6hdr));
2228 2288
2229 rt = (struct rt6_info*) ip6_route_output(NULL, &fl); 2289 rt = (struct rt6_info*) ip6_route_output(net, NULL, &fl);
2230 skb->dst = &rt->u.dst; 2290 skb->dst = &rt->u.dst;
2231 2291
2232 err = rt6_fill_node(skb, rt, &fl.fl6_dst, &fl.fl6_src, iif, 2292 err = rt6_fill_node(skb, rt, &fl.fl6_dst, &fl.fl6_src, iif,
2233 RTM_NEWROUTE, NETLINK_CB(in_skb).pid, 2293 RTM_NEWROUTE, NETLINK_CB(in_skb).pid,
2234 nlh->nlmsg_seq, 0, 0); 2294 nlh->nlmsg_seq, 0, 0, 0);
2235 if (err < 0) { 2295 if (err < 0) {
2236 kfree_skb(skb); 2296 kfree_skb(skb);
2237 goto errout; 2297 goto errout;
2238 } 2298 }
2239 2299
2240 err = rtnl_unicast(skb, &init_net, NETLINK_CB(in_skb).pid); 2300 err = rtnl_unicast(skb, net, NETLINK_CB(in_skb).pid);
2241errout: 2301errout:
2242 return err; 2302 return err;
2243} 2303}
@@ -2245,6 +2305,7 @@ errout:
2245void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info) 2305void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info)
2246{ 2306{
2247 struct sk_buff *skb; 2307 struct sk_buff *skb;
2308 struct net *net = info->nl_net;
2248 u32 seq; 2309 u32 seq;
2249 int err; 2310 int err;
2250 2311
@@ -2256,18 +2317,38 @@ void inet6_rt_notify(int event, struct rt6_info *rt, struct nl_info *info)
2256 goto errout; 2317 goto errout;
2257 2318
2258 err = rt6_fill_node(skb, rt, NULL, NULL, 0, 2319 err = rt6_fill_node(skb, rt, NULL, NULL, 0,
2259 event, info->pid, seq, 0, 0); 2320 event, info->pid, seq, 0, 0, 0);
2260 if (err < 0) { 2321 if (err < 0) {
2261 /* -EMSGSIZE implies BUG in rt6_nlmsg_size() */ 2322 /* -EMSGSIZE implies BUG in rt6_nlmsg_size() */
2262 WARN_ON(err == -EMSGSIZE); 2323 WARN_ON(err == -EMSGSIZE);
2263 kfree_skb(skb); 2324 kfree_skb(skb);
2264 goto errout; 2325 goto errout;
2265 } 2326 }
2266 err = rtnl_notify(skb, &init_net, info->pid, 2327 err = rtnl_notify(skb, net, info->pid, RTNLGRP_IPV6_ROUTE,
2267 RTNLGRP_IPV6_ROUTE, info->nlh, gfp_any()); 2328 info->nlh, gfp_any());
2268errout: 2329errout:
2269 if (err < 0) 2330 if (err < 0)
2270 rtnl_set_sk_err(&init_net, RTNLGRP_IPV6_ROUTE, err); 2331 rtnl_set_sk_err(net, RTNLGRP_IPV6_ROUTE, err);
2332}
2333
2334static int ip6_route_dev_notify(struct notifier_block *this,
2335 unsigned long event, void *data)
2336{
2337 struct net_device *dev = (struct net_device *)data;
2338 struct net *net = dev_net(dev);
2339
2340 if (event == NETDEV_REGISTER && (dev->flags & IFF_LOOPBACK)) {
2341 net->ipv6.ip6_null_entry->u.dst.dev = dev;
2342 net->ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(dev);
2343#ifdef CONFIG_IPV6_MULTIPLE_TABLES
2344 net->ipv6.ip6_prohibit_entry->u.dst.dev = dev;
2345 net->ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(dev);
2346 net->ipv6.ip6_blk_hole_entry->u.dst.dev = dev;
2347 net->ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(dev);
2348#endif
2349 }
2350
2351 return NOTIFY_OK;
2271} 2352}
2272 2353
2273/* 2354/*
@@ -2316,13 +2397,33 @@ static int rt6_info_route(struct rt6_info *rt, void *p_arg)
2316 2397
2317static int ipv6_route_show(struct seq_file *m, void *v) 2398static int ipv6_route_show(struct seq_file *m, void *v)
2318{ 2399{
2319 fib6_clean_all(rt6_info_route, 0, m); 2400 struct net *net = (struct net *)m->private;
2401 fib6_clean_all(net, rt6_info_route, 0, m);
2320 return 0; 2402 return 0;
2321} 2403}
2322 2404
2323static int ipv6_route_open(struct inode *inode, struct file *file) 2405static int ipv6_route_open(struct inode *inode, struct file *file)
2324{ 2406{
2325 return single_open(file, ipv6_route_show, NULL); 2407 int err;
2408 struct net *net = get_proc_net(inode);
2409 if (!net)
2410 return -ENXIO;
2411
2412 err = single_open(file, ipv6_route_show, net);
2413 if (err < 0) {
2414 put_net(net);
2415 return err;
2416 }
2417
2418 return 0;
2419}
2420
2421static int ipv6_route_release(struct inode *inode, struct file *file)
2422{
2423 struct seq_file *seq = file->private_data;
2424 struct net *net = seq->private;
2425 put_net(net);
2426 return single_release(inode, file);
2326} 2427}
2327 2428
2328static const struct file_operations ipv6_route_proc_fops = { 2429static const struct file_operations ipv6_route_proc_fops = {
@@ -2330,24 +2431,46 @@ static const struct file_operations ipv6_route_proc_fops = {
2330 .open = ipv6_route_open, 2431 .open = ipv6_route_open,
2331 .read = seq_read, 2432 .read = seq_read,
2332 .llseek = seq_lseek, 2433 .llseek = seq_lseek,
2333 .release = single_release, 2434 .release = ipv6_route_release,
2334}; 2435};
2335 2436
2336static int rt6_stats_seq_show(struct seq_file *seq, void *v) 2437static int rt6_stats_seq_show(struct seq_file *seq, void *v)
2337{ 2438{
2439 struct net *net = (struct net *)seq->private;
2338 seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n", 2440 seq_printf(seq, "%04x %04x %04x %04x %04x %04x %04x\n",
2339 rt6_stats.fib_nodes, rt6_stats.fib_route_nodes, 2441 net->ipv6.rt6_stats->fib_nodes,
2340 rt6_stats.fib_rt_alloc, rt6_stats.fib_rt_entries, 2442 net->ipv6.rt6_stats->fib_route_nodes,
2341 rt6_stats.fib_rt_cache, 2443 net->ipv6.rt6_stats->fib_rt_alloc,
2342 atomic_read(&ip6_dst_ops.entries), 2444 net->ipv6.rt6_stats->fib_rt_entries,
2343 rt6_stats.fib_discarded_routes); 2445 net->ipv6.rt6_stats->fib_rt_cache,
2446 atomic_read(&net->ipv6.ip6_dst_ops->entries),
2447 net->ipv6.rt6_stats->fib_discarded_routes);
2344 2448
2345 return 0; 2449 return 0;
2346} 2450}
2347 2451
2348static int rt6_stats_seq_open(struct inode *inode, struct file *file) 2452static int rt6_stats_seq_open(struct inode *inode, struct file *file)
2349{ 2453{
2350 return single_open(file, rt6_stats_seq_show, NULL); 2454 int err;
2455 struct net *net = get_proc_net(inode);
2456 if (!net)
2457 return -ENXIO;
2458
2459 err = single_open(file, rt6_stats_seq_show, net);
2460 if (err < 0) {
2461 put_net(net);
2462 return err;
2463 }
2464
2465 return 0;
2466}
2467
2468static int rt6_stats_seq_release(struct inode *inode, struct file *file)
2469{
2470 struct seq_file *seq = file->private_data;
2471 struct net *net = (struct net *)seq->private;
2472 put_net(net);
2473 return single_release(inode, file);
2351} 2474}
2352 2475
2353static const struct file_operations rt6_stats_seq_fops = { 2476static const struct file_operations rt6_stats_seq_fops = {
@@ -2355,42 +2478,8 @@ static const struct file_operations rt6_stats_seq_fops = {
2355 .open = rt6_stats_seq_open, 2478 .open = rt6_stats_seq_open,
2356 .read = seq_read, 2479 .read = seq_read,
2357 .llseek = seq_lseek, 2480 .llseek = seq_lseek,
2358 .release = single_release, 2481 .release = rt6_stats_seq_release,
2359}; 2482};
2360
2361static int ipv6_route_proc_init(struct net *net)
2362{
2363 int ret = -ENOMEM;
2364 if (!proc_net_fops_create(net, "ipv6_route",
2365 0, &ipv6_route_proc_fops))
2366 goto out;
2367
2368 if (!proc_net_fops_create(net, "rt6_stats",
2369 S_IRUGO, &rt6_stats_seq_fops))
2370 goto out_ipv6_route;
2371
2372 ret = 0;
2373out:
2374 return ret;
2375out_ipv6_route:
2376 proc_net_remove(net, "ipv6_route");
2377 goto out;
2378}
2379
2380static void ipv6_route_proc_fini(struct net *net)
2381{
2382 proc_net_remove(net, "ipv6_route");
2383 proc_net_remove(net, "rt6_stats");
2384}
2385#else
2386static inline int ipv6_route_proc_init(struct net *net)
2387{
2388 return 0;
2389}
2390static inline void ipv6_route_proc_fini(struct net *net)
2391{
2392 return ;
2393}
2394#endif /* CONFIG_PROC_FS */ 2483#endif /* CONFIG_PROC_FS */
2395 2484
2396#ifdef CONFIG_SYSCTL 2485#ifdef CONFIG_SYSCTL
@@ -2399,10 +2488,11 @@ static
2399int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file * filp, 2488int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file * filp,
2400 void __user *buffer, size_t *lenp, loff_t *ppos) 2489 void __user *buffer, size_t *lenp, loff_t *ppos)
2401{ 2490{
2402 int delay = init_net.ipv6.sysctl.flush_delay; 2491 struct net *net = current->nsproxy->net_ns;
2492 int delay = net->ipv6.sysctl.flush_delay;
2403 if (write) { 2493 if (write) {
2404 proc_dointvec(ctl, write, filp, buffer, lenp, ppos); 2494 proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
2405 fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay); 2495 fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay, net);
2406 return 0; 2496 return 0;
2407 } else 2497 } else
2408 return -EINVAL; 2498 return -EINVAL;
@@ -2419,7 +2509,7 @@ ctl_table ipv6_route_table_template[] = {
2419 { 2509 {
2420 .ctl_name = NET_IPV6_ROUTE_GC_THRESH, 2510 .ctl_name = NET_IPV6_ROUTE_GC_THRESH,
2421 .procname = "gc_thresh", 2511 .procname = "gc_thresh",
2422 .data = &ip6_dst_ops.gc_thresh, 2512 .data = &ip6_dst_ops_template.gc_thresh,
2423 .maxlen = sizeof(int), 2513 .maxlen = sizeof(int),
2424 .mode = 0644, 2514 .mode = 0644,
2425 .proc_handler = &proc_dointvec, 2515 .proc_handler = &proc_dointvec,
@@ -2505,33 +2595,143 @@ struct ctl_table *ipv6_route_sysctl_init(struct net *net)
2505 table = kmemdup(ipv6_route_table_template, 2595 table = kmemdup(ipv6_route_table_template,
2506 sizeof(ipv6_route_table_template), 2596 sizeof(ipv6_route_table_template),
2507 GFP_KERNEL); 2597 GFP_KERNEL);
2598
2599 if (table) {
2600 table[0].data = &net->ipv6.sysctl.flush_delay;
2601 table[1].data = &net->ipv6.ip6_dst_ops->gc_thresh;
2602 table[2].data = &net->ipv6.sysctl.ip6_rt_max_size;
2603 table[3].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
2604 table[4].data = &net->ipv6.sysctl.ip6_rt_gc_timeout;
2605 table[5].data = &net->ipv6.sysctl.ip6_rt_gc_interval;
2606 table[6].data = &net->ipv6.sysctl.ip6_rt_gc_elasticity;
2607 table[7].data = &net->ipv6.sysctl.ip6_rt_mtu_expires;
2608 table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
2609 }
2610
2508 return table; 2611 return table;
2509} 2612}
2510#endif 2613#endif
2511 2614
2615static int ip6_route_net_init(struct net *net)
2616{
2617 int ret = 0;
2618
2619 ret = -ENOMEM;
2620 net->ipv6.ip6_dst_ops = kmemdup(&ip6_dst_ops_template,
2621 sizeof(*net->ipv6.ip6_dst_ops),
2622 GFP_KERNEL);
2623 if (!net->ipv6.ip6_dst_ops)
2624 goto out;
2625 net->ipv6.ip6_dst_ops->dst_net = hold_net(net);
2626
2627 net->ipv6.ip6_null_entry = kmemdup(&ip6_null_entry_template,
2628 sizeof(*net->ipv6.ip6_null_entry),
2629 GFP_KERNEL);
2630 if (!net->ipv6.ip6_null_entry)
2631 goto out_ip6_dst_ops;
2632 net->ipv6.ip6_null_entry->u.dst.path =
2633 (struct dst_entry *)net->ipv6.ip6_null_entry;
2634 net->ipv6.ip6_null_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2635
2636#ifdef CONFIG_IPV6_MULTIPLE_TABLES
2637 net->ipv6.ip6_prohibit_entry = kmemdup(&ip6_prohibit_entry_template,
2638 sizeof(*net->ipv6.ip6_prohibit_entry),
2639 GFP_KERNEL);
2640 if (!net->ipv6.ip6_prohibit_entry) {
2641 kfree(net->ipv6.ip6_null_entry);
2642 goto out;
2643 }
2644 net->ipv6.ip6_prohibit_entry->u.dst.path =
2645 (struct dst_entry *)net->ipv6.ip6_prohibit_entry;
2646 net->ipv6.ip6_prohibit_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2647
2648 net->ipv6.ip6_blk_hole_entry = kmemdup(&ip6_blk_hole_entry_template,
2649 sizeof(*net->ipv6.ip6_blk_hole_entry),
2650 GFP_KERNEL);
2651 if (!net->ipv6.ip6_blk_hole_entry) {
2652 kfree(net->ipv6.ip6_null_entry);
2653 kfree(net->ipv6.ip6_prohibit_entry);
2654 goto out;
2655 }
2656 net->ipv6.ip6_blk_hole_entry->u.dst.path =
2657 (struct dst_entry *)net->ipv6.ip6_blk_hole_entry;
2658 net->ipv6.ip6_blk_hole_entry->u.dst.ops = net->ipv6.ip6_dst_ops;
2659#endif
2660
2661#ifdef CONFIG_PROC_FS
2662 proc_net_fops_create(net, "ipv6_route", 0, &ipv6_route_proc_fops);
2663 proc_net_fops_create(net, "rt6_stats", S_IRUGO, &rt6_stats_seq_fops);
2664#endif
2665 net->ipv6.ip6_rt_gc_expire = 30*HZ;
2666
2667 ret = 0;
2668out:
2669 return ret;
2670
2671out_ip6_dst_ops:
2672 release_net(net->ipv6.ip6_dst_ops->dst_net);
2673 kfree(net->ipv6.ip6_dst_ops);
2674 goto out;
2675}
2676
2677static void ip6_route_net_exit(struct net *net)
2678{
2679#ifdef CONFIG_PROC_FS
2680 proc_net_remove(net, "ipv6_route");
2681 proc_net_remove(net, "rt6_stats");
2682#endif
2683 kfree(net->ipv6.ip6_null_entry);
2684#ifdef CONFIG_IPV6_MULTIPLE_TABLES
2685 kfree(net->ipv6.ip6_prohibit_entry);
2686 kfree(net->ipv6.ip6_blk_hole_entry);
2687#endif
2688 release_net(net->ipv6.ip6_dst_ops->dst_net);
2689 kfree(net->ipv6.ip6_dst_ops);
2690}
2691
2692static struct pernet_operations ip6_route_net_ops = {
2693 .init = ip6_route_net_init,
2694 .exit = ip6_route_net_exit,
2695};
2696
2697static struct notifier_block ip6_route_dev_notifier = {
2698 .notifier_call = ip6_route_dev_notify,
2699 .priority = 0,
2700};
2701
2512int __init ip6_route_init(void) 2702int __init ip6_route_init(void)
2513{ 2703{
2514 int ret; 2704 int ret;
2515 2705
2516 ip6_dst_ops.kmem_cachep = 2706 ret = -ENOMEM;
2707 ip6_dst_ops_template.kmem_cachep =
2517 kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0, 2708 kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0,
2518 SLAB_HWCACHE_ALIGN, NULL); 2709 SLAB_HWCACHE_ALIGN, NULL);
2519 if (!ip6_dst_ops.kmem_cachep) 2710 if (!ip6_dst_ops_template.kmem_cachep)
2520 return -ENOMEM; 2711 goto out;;
2521
2522 ip6_dst_blackhole_ops.kmem_cachep = ip6_dst_ops.kmem_cachep;
2523 2712
2524 ret = fib6_init(); 2713 ret = register_pernet_subsys(&ip6_route_net_ops);
2525 if (ret) 2714 if (ret)
2526 goto out_kmem_cache; 2715 goto out_kmem_cache;
2527 2716
2528 ret = ipv6_route_proc_init(&init_net); 2717 /* Registering of the loopback is done before this portion of code,
2718 * the loopback reference in rt6_info will not be taken, do it
2719 * manually for init_net */
2720 init_net.ipv6.ip6_null_entry->u.dst.dev = init_net.loopback_dev;
2721 init_net.ipv6.ip6_null_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
2722 #ifdef CONFIG_IPV6_MULTIPLE_TABLES
2723 init_net.ipv6.ip6_prohibit_entry->u.dst.dev = init_net.loopback_dev;
2724 init_net.ipv6.ip6_prohibit_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
2725 init_net.ipv6.ip6_blk_hole_entry->u.dst.dev = init_net.loopback_dev;
2726 init_net.ipv6.ip6_blk_hole_entry->rt6i_idev = in6_dev_get(init_net.loopback_dev);
2727 #endif
2728 ret = fib6_init();
2529 if (ret) 2729 if (ret)
2530 goto out_fib6_init; 2730 goto out_register_subsys;
2531 2731
2532 ret = xfrm6_init(); 2732 ret = xfrm6_init();
2533 if (ret) 2733 if (ret)
2534 goto out_proc_init; 2734 goto out_fib6_init;
2535 2735
2536 ret = fib6_rules_init(); 2736 ret = fib6_rules_init();
2537 if (ret) 2737 if (ret)
@@ -2543,7 +2743,10 @@ int __init ip6_route_init(void)
2543 __rtnl_register(PF_INET6, RTM_GETROUTE, inet6_rtm_getroute, NULL)) 2743 __rtnl_register(PF_INET6, RTM_GETROUTE, inet6_rtm_getroute, NULL))
2544 goto fib6_rules_init; 2744 goto fib6_rules_init;
2545 2745
2546 ret = 0; 2746 ret = register_netdevice_notifier(&ip6_route_dev_notifier);
2747 if (ret)
2748 goto fib6_rules_init;
2749
2547out: 2750out:
2548 return ret; 2751 return ret;
2549 2752
@@ -2551,22 +2754,21 @@ fib6_rules_init:
2551 fib6_rules_cleanup(); 2754 fib6_rules_cleanup();
2552xfrm6_init: 2755xfrm6_init:
2553 xfrm6_fini(); 2756 xfrm6_fini();
2554out_proc_init:
2555 ipv6_route_proc_fini(&init_net);
2556out_fib6_init: 2757out_fib6_init:
2557 rt6_ifdown(NULL);
2558 fib6_gc_cleanup(); 2758 fib6_gc_cleanup();
2759out_register_subsys:
2760 unregister_pernet_subsys(&ip6_route_net_ops);
2559out_kmem_cache: 2761out_kmem_cache:
2560 kmem_cache_destroy(ip6_dst_ops.kmem_cachep); 2762 kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
2561 goto out; 2763 goto out;
2562} 2764}
2563 2765
2564void ip6_route_cleanup(void) 2766void ip6_route_cleanup(void)
2565{ 2767{
2768 unregister_netdevice_notifier(&ip6_route_dev_notifier);
2566 fib6_rules_cleanup(); 2769 fib6_rules_cleanup();
2567 ipv6_route_proc_fini(&init_net);
2568 xfrm6_fini(); 2770 xfrm6_fini();
2569 rt6_ifdown(NULL);
2570 fib6_gc_cleanup(); 2771 fib6_gc_cleanup();
2571 kmem_cache_destroy(ip6_dst_ops.kmem_cachep); 2772 unregister_pernet_subsys(&ip6_route_net_ops);
2773 kmem_cache_destroy(ip6_dst_ops_template.kmem_cachep);
2572} 2774}
generated by cgit v1.2.2 (git 2.25.0) at 2025-02-09 06:57:23 -0500