aboutsummaryrefslogtreecommitdiffstats
path: root/net/ipv4
diff options
context:
space:
mode:
Diffstat (limited to 'net/ipv4')
-rw-r--r--net/ipv4/Kconfig6
-rw-r--r--net/ipv4/arp.c55
-rw-r--r--net/ipv4/devinet.c2
-rw-r--r--net/ipv4/fib_frontend.c6
-rw-r--r--net/ipv4/fib_semantics.c80
-rw-r--r--net/ipv4/icmp.c2
-rw-r--r--net/ipv4/igmp.c87
-rw-r--r--net/ipv4/inet_connection_sock.c2
-rw-r--r--net/ipv4/inet_diag.c2
-rw-r--r--net/ipv4/ip_fragment.c39
-rw-r--r--net/ipv4/ip_gre.c4
-rw-r--r--net/ipv4/ip_output.c2
-rw-r--r--net/ipv4/ip_sockglue.c14
-rw-r--r--net/ipv4/ipcomp.c9
-rw-r--r--net/ipv4/ipconfig.c2
-rw-r--r--net/ipv4/ipip.c7
-rw-r--r--net/ipv4/netfilter/arp_tables.c4
-rw-r--r--net/ipv4/netfilter/ip_tables.c4
-rw-r--r--net/ipv4/netfilter/ipt_ECN.c2
-rw-r--r--net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c2
-rw-r--r--net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c4
-rw-r--r--net/ipv4/netfilter/nf_defrag_ipv4.c21
-rw-r--r--net/ipv4/netfilter/nf_nat_core.c22
-rw-r--r--net/ipv4/proc.c4
-rw-r--r--net/ipv4/route.c9
-rw-r--r--net/ipv4/syncookies.c30
-rw-r--r--net/ipv4/tcp.c30
-rw-r--r--net/ipv4/tcp_input.c24
-rw-r--r--net/ipv4/tcp_ipv4.c46
-rw-r--r--net/ipv4/tcp_minisocks.c10
-rw-r--r--net/ipv4/tcp_output.c40
-rw-r--r--net/ipv4/tcp_probe.c19
-rw-r--r--net/ipv4/tcp_timer.c6
-rw-r--r--net/ipv4/udp.c26
-rw-r--r--net/ipv4/udplite.c4
-rw-r--r--net/ipv4/xfrm4_policy.c14
36 files changed, 392 insertions, 248 deletions
diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig
index 70491d9035eb..0c94a1ac2946 100644
--- a/net/ipv4/Kconfig
+++ b/net/ipv4/Kconfig
@@ -166,7 +166,7 @@ config IP_PNP_DHCP
166 166
167 If unsure, say Y. Note that if you want to use DHCP, a DHCP server 167 If unsure, say Y. Note that if you want to use DHCP, a DHCP server
168 must be operating on your network. Read 168 must be operating on your network. Read
169 <file:Documentation/filesystems/nfsroot.txt> for details. 169 <file:Documentation/filesystems/nfs/nfsroot.txt> for details.
170 170
171config IP_PNP_BOOTP 171config IP_PNP_BOOTP
172 bool "IP: BOOTP support" 172 bool "IP: BOOTP support"
@@ -181,7 +181,7 @@ config IP_PNP_BOOTP
181 does BOOTP itself, providing all necessary information on the kernel 181 does BOOTP itself, providing all necessary information on the kernel
182 command line, you can say N here. If unsure, say Y. Note that if you 182 command line, you can say N here. If unsure, say Y. Note that if you
183 want to use BOOTP, a BOOTP server must be operating on your network. 183 want to use BOOTP, a BOOTP server must be operating on your network.
184 Read <file:Documentation/filesystems/nfsroot.txt> for details. 184 Read <file:Documentation/filesystems/nfs/nfsroot.txt> for details.
185 185
186config IP_PNP_RARP 186config IP_PNP_RARP
187 bool "IP: RARP support" 187 bool "IP: RARP support"
@@ -194,7 +194,7 @@ config IP_PNP_RARP
194 older protocol which is being obsoleted by BOOTP and DHCP), say Y 194 older protocol which is being obsoleted by BOOTP and DHCP), say Y
195 here. Note that if you want to use RARP, a RARP server must be 195 here. Note that if you want to use RARP, a RARP server must be
196 operating on your network. Read 196 operating on your network. Read
197 <file:Documentation/filesystems/nfsroot.txt> for details. 197 <file:Documentation/filesystems/nfs/nfsroot.txt> for details.
198 198
199# not yet ready.. 199# not yet ready..
200# bool ' IP: ARP support' CONFIG_IP_PNP_ARP 200# bool ' IP: ARP support' CONFIG_IP_PNP_ARP
diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c
index c95cd93acf29..1940b4df7699 100644
--- a/net/ipv4/arp.c
+++ b/net/ipv4/arp.c
@@ -70,6 +70,7 @@
70 * bonding can change the skb before 70 * bonding can change the skb before
71 * sending (e.g. insert 8021q tag). 71 * sending (e.g. insert 8021q tag).
72 * Harald Welte : convert to make use of jenkins hash 72 * Harald Welte : convert to make use of jenkins hash
73 * Jesper D. Brouer: Proxy ARP PVLAN RFC 3069 support.
73 */ 74 */
74 75
75#include <linux/module.h> 76#include <linux/module.h>
@@ -524,12 +525,15 @@ int arp_bind_neighbour(struct dst_entry *dst)
524/* 525/*
525 * Check if we can use proxy ARP for this path 526 * Check if we can use proxy ARP for this path
526 */ 527 */
527 528static inline int arp_fwd_proxy(struct in_device *in_dev,
528static inline int arp_fwd_proxy(struct in_device *in_dev, struct rtable *rt) 529 struct net_device *dev, struct rtable *rt)
529{ 530{
530 struct in_device *out_dev; 531 struct in_device *out_dev;
531 int imi, omi = -1; 532 int imi, omi = -1;
532 533
534 if (rt->u.dst.dev == dev)
535 return 0;
536
533 if (!IN_DEV_PROXY_ARP(in_dev)) 537 if (!IN_DEV_PROXY_ARP(in_dev))
534 return 0; 538 return 0;
535 539
@@ -548,6 +552,43 @@ static inline int arp_fwd_proxy(struct in_device *in_dev, struct rtable *rt)
548} 552}
549 553
550/* 554/*
555 * Check for RFC3069 proxy arp private VLAN (allow to send back to same dev)
556 *
557 * RFC3069 supports proxy arp replies back to the same interface. This
558 * is done to support (ethernet) switch features, like RFC 3069, where
559 * the individual ports are not allowed to communicate with each
560 * other, BUT they are allowed to talk to the upstream router. As
561 * described in RFC 3069, it is possible to allow these hosts to
562 * communicate through the upstream router, by proxy_arp'ing.
563 *
564 * RFC 3069: "VLAN Aggregation for Efficient IP Address Allocation"
565 *
566 * This technology is known by different names:
567 * In RFC 3069 it is called VLAN Aggregation.
568 * Cisco and Allied Telesyn call it Private VLAN.
569 * Hewlett-Packard call it Source-Port filtering or port-isolation.
570 * Ericsson call it MAC-Forced Forwarding (RFC Draft).
571 *
572 */
573static inline int arp_fwd_pvlan(struct in_device *in_dev,
574 struct net_device *dev, struct rtable *rt,
575 __be32 sip, __be32 tip)
576{
577 /* Private VLAN is only concerned about the same ethernet segment */
578 if (rt->u.dst.dev != dev)
579 return 0;
580
581 /* Don't reply on self probes (often done by windowz boxes)*/
582 if (sip == tip)
583 return 0;
584
585 if (IN_DEV_PROXY_ARP_PVLAN(in_dev))
586 return 1;
587 else
588 return 0;
589}
590
591/*
551 * Interface to link layer: send routine and receive handler. 592 * Interface to link layer: send routine and receive handler.
552 */ 593 */
553 594
@@ -833,8 +874,11 @@ static int arp_process(struct sk_buff *skb)
833 } 874 }
834 goto out; 875 goto out;
835 } else if (IN_DEV_FORWARD(in_dev)) { 876 } else if (IN_DEV_FORWARD(in_dev)) {
836 if (addr_type == RTN_UNICAST && rt->u.dst.dev != dev && 877 if (addr_type == RTN_UNICAST &&
837 (arp_fwd_proxy(in_dev, rt) || pneigh_lookup(&arp_tbl, net, &tip, dev, 0))) { 878 (arp_fwd_proxy(in_dev, dev, rt) ||
879 arp_fwd_pvlan(in_dev, dev, rt, sip, tip) ||
880 pneigh_lookup(&arp_tbl, net, &tip, dev, 0)))
881 {
838 n = neigh_event_ns(&arp_tbl, sha, &sip, dev); 882 n = neigh_event_ns(&arp_tbl, sha, &sip, dev);
839 if (n) 883 if (n)
840 neigh_release(n); 884 neigh_release(n);
@@ -863,7 +907,8 @@ static int arp_process(struct sk_buff *skb)
863 devices (strip is candidate) 907 devices (strip is candidate)
864 */ 908 */
865 if (n == NULL && 909 if (n == NULL &&
866 arp->ar_op == htons(ARPOP_REPLY) && 910 (arp->ar_op == htons(ARPOP_REPLY) ||
911 (arp->ar_op == htons(ARPOP_REQUEST) && tip == sip)) &&
867 inet_addr_type(net, sip) == RTN_UNICAST) 912 inet_addr_type(net, sip) == RTN_UNICAST)
868 n = __neigh_lookup(&arp_tbl, &sip, dev, 1); 913 n = __neigh_lookup(&arp_tbl, &sip, dev, 1);
869 } 914 }
diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index 5cdbc102a418..cd71a3908391 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -1397,6 +1397,7 @@ static struct devinet_sysctl_table {
1397 DEVINET_SYSCTL_RW_ENTRY(ACCEPT_SOURCE_ROUTE, 1397 DEVINET_SYSCTL_RW_ENTRY(ACCEPT_SOURCE_ROUTE,
1398 "accept_source_route"), 1398 "accept_source_route"),
1399 DEVINET_SYSCTL_RW_ENTRY(ACCEPT_LOCAL, "accept_local"), 1399 DEVINET_SYSCTL_RW_ENTRY(ACCEPT_LOCAL, "accept_local"),
1400 DEVINET_SYSCTL_RW_ENTRY(SRC_VMARK, "src_valid_mark"),
1400 DEVINET_SYSCTL_RW_ENTRY(PROXY_ARP, "proxy_arp"), 1401 DEVINET_SYSCTL_RW_ENTRY(PROXY_ARP, "proxy_arp"),
1401 DEVINET_SYSCTL_RW_ENTRY(MEDIUM_ID, "medium_id"), 1402 DEVINET_SYSCTL_RW_ENTRY(MEDIUM_ID, "medium_id"),
1402 DEVINET_SYSCTL_RW_ENTRY(BOOTP_RELAY, "bootp_relay"), 1403 DEVINET_SYSCTL_RW_ENTRY(BOOTP_RELAY, "bootp_relay"),
@@ -1407,6 +1408,7 @@ static struct devinet_sysctl_table {
1407 DEVINET_SYSCTL_RW_ENTRY(ARP_IGNORE, "arp_ignore"), 1408 DEVINET_SYSCTL_RW_ENTRY(ARP_IGNORE, "arp_ignore"),
1408 DEVINET_SYSCTL_RW_ENTRY(ARP_ACCEPT, "arp_accept"), 1409 DEVINET_SYSCTL_RW_ENTRY(ARP_ACCEPT, "arp_accept"),
1409 DEVINET_SYSCTL_RW_ENTRY(ARP_NOTIFY, "arp_notify"), 1410 DEVINET_SYSCTL_RW_ENTRY(ARP_NOTIFY, "arp_notify"),
1411 DEVINET_SYSCTL_RW_ENTRY(PROXY_ARP_PVLAN, "proxy_arp_pvlan"),
1410 1412
1411 DEVINET_SYSCTL_FLUSHING_ENTRY(NOXFRM, "disable_xfrm"), 1413 DEVINET_SYSCTL_FLUSHING_ENTRY(NOXFRM, "disable_xfrm"),
1412 DEVINET_SYSCTL_FLUSHING_ENTRY(NOPOLICY, "disable_policy"), 1414 DEVINET_SYSCTL_FLUSHING_ENTRY(NOPOLICY, "disable_policy"),
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index 3323168ee52d..9b3e28ed5240 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -252,6 +252,8 @@ int fib_validate_source(__be32 src, __be32 dst, u8 tos, int oif,
252 no_addr = in_dev->ifa_list == NULL; 252 no_addr = in_dev->ifa_list == NULL;
253 rpf = IN_DEV_RPFILTER(in_dev); 253 rpf = IN_DEV_RPFILTER(in_dev);
254 accept_local = IN_DEV_ACCEPT_LOCAL(in_dev); 254 accept_local = IN_DEV_ACCEPT_LOCAL(in_dev);
255 if (mark && !IN_DEV_SRC_VMARK(in_dev))
256 fl.mark = 0;
255 } 257 }
256 rcu_read_unlock(); 258 rcu_read_unlock();
257 259
@@ -881,7 +883,7 @@ static void nl_fib_input(struct sk_buff *skb)
881 netlink_unicast(net->ipv4.fibnl, skb, pid, MSG_DONTWAIT); 883 netlink_unicast(net->ipv4.fibnl, skb, pid, MSG_DONTWAIT);
882} 884}
883 885
884static int nl_fib_lookup_init(struct net *net) 886static int __net_init nl_fib_lookup_init(struct net *net)
885{ 887{
886 struct sock *sk; 888 struct sock *sk;
887 sk = netlink_kernel_create(net, NETLINK_FIB_LOOKUP, 0, 889 sk = netlink_kernel_create(net, NETLINK_FIB_LOOKUP, 0,
@@ -1002,7 +1004,7 @@ fail:
1002 return err; 1004 return err;
1003} 1005}
1004 1006
1005static void __net_exit ip_fib_net_exit(struct net *net) 1007static void ip_fib_net_exit(struct net *net)
1006{ 1008{
1007 unsigned int i; 1009 unsigned int i;
1008 1010
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index ed19aa6919c2..1af0ea0fb6a2 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -62,8 +62,8 @@ static DEFINE_SPINLOCK(fib_multipath_lock);
62#define for_nexthops(fi) { int nhsel; const struct fib_nh * nh; \ 62#define for_nexthops(fi) { int nhsel; const struct fib_nh * nh; \
63for (nhsel=0, nh = (fi)->fib_nh; nhsel < (fi)->fib_nhs; nh++, nhsel++) 63for (nhsel=0, nh = (fi)->fib_nh; nhsel < (fi)->fib_nhs; nh++, nhsel++)
64 64
65#define change_nexthops(fi) { int nhsel; struct fib_nh * nh; \ 65#define change_nexthops(fi) { int nhsel; struct fib_nh *nexthop_nh; \
66for (nhsel=0, nh = (struct fib_nh *)((fi)->fib_nh); nhsel < (fi)->fib_nhs; nh++, nhsel++) 66for (nhsel=0, nexthop_nh = (struct fib_nh *)((fi)->fib_nh); nhsel < (fi)->fib_nhs; nexthop_nh++, nhsel++)
67 67
68#else /* CONFIG_IP_ROUTE_MULTIPATH */ 68#else /* CONFIG_IP_ROUTE_MULTIPATH */
69 69
@@ -72,7 +72,7 @@ for (nhsel=0, nh = (struct fib_nh *)((fi)->fib_nh); nhsel < (fi)->fib_nhs; nh++,
72#define for_nexthops(fi) { int nhsel = 0; const struct fib_nh * nh = (fi)->fib_nh; \ 72#define for_nexthops(fi) { int nhsel = 0; const struct fib_nh * nh = (fi)->fib_nh; \
73for (nhsel=0; nhsel < 1; nhsel++) 73for (nhsel=0; nhsel < 1; nhsel++)
74 74
75#define change_nexthops(fi) { int nhsel = 0; struct fib_nh * nh = (struct fib_nh *)((fi)->fib_nh); \ 75#define change_nexthops(fi) { int nhsel = 0; struct fib_nh *nexthop_nh = (struct fib_nh *)((fi)->fib_nh); \
76for (nhsel=0; nhsel < 1; nhsel++) 76for (nhsel=0; nhsel < 1; nhsel++)
77 77
78#endif /* CONFIG_IP_ROUTE_MULTIPATH */ 78#endif /* CONFIG_IP_ROUTE_MULTIPATH */
@@ -145,9 +145,9 @@ void free_fib_info(struct fib_info *fi)
145 return; 145 return;
146 } 146 }
147 change_nexthops(fi) { 147 change_nexthops(fi) {
148 if (nh->nh_dev) 148 if (nexthop_nh->nh_dev)
149 dev_put(nh->nh_dev); 149 dev_put(nexthop_nh->nh_dev);
150 nh->nh_dev = NULL; 150 nexthop_nh->nh_dev = NULL;
151 } endfor_nexthops(fi); 151 } endfor_nexthops(fi);
152 fib_info_cnt--; 152 fib_info_cnt--;
153 release_net(fi->fib_net); 153 release_net(fi->fib_net);
@@ -162,9 +162,9 @@ void fib_release_info(struct fib_info *fi)
162 if (fi->fib_prefsrc) 162 if (fi->fib_prefsrc)
163 hlist_del(&fi->fib_lhash); 163 hlist_del(&fi->fib_lhash);
164 change_nexthops(fi) { 164 change_nexthops(fi) {
165 if (!nh->nh_dev) 165 if (!nexthop_nh->nh_dev)
166 continue; 166 continue;
167 hlist_del(&nh->nh_hash); 167 hlist_del(&nexthop_nh->nh_hash);
168 } endfor_nexthops(fi) 168 } endfor_nexthops(fi)
169 fi->fib_dead = 1; 169 fi->fib_dead = 1;
170 fib_info_put(fi); 170 fib_info_put(fi);
@@ -395,19 +395,20 @@ static int fib_get_nhs(struct fib_info *fi, struct rtnexthop *rtnh,
395 if (!rtnh_ok(rtnh, remaining)) 395 if (!rtnh_ok(rtnh, remaining))
396 return -EINVAL; 396 return -EINVAL;
397 397
398 nh->nh_flags = (cfg->fc_flags & ~0xFF) | rtnh->rtnh_flags; 398 nexthop_nh->nh_flags =
399 nh->nh_oif = rtnh->rtnh_ifindex; 399 (cfg->fc_flags & ~0xFF) | rtnh->rtnh_flags;
400 nh->nh_weight = rtnh->rtnh_hops + 1; 400 nexthop_nh->nh_oif = rtnh->rtnh_ifindex;
401 nexthop_nh->nh_weight = rtnh->rtnh_hops + 1;
401 402
402 attrlen = rtnh_attrlen(rtnh); 403 attrlen = rtnh_attrlen(rtnh);
403 if (attrlen > 0) { 404 if (attrlen > 0) {
404 struct nlattr *nla, *attrs = rtnh_attrs(rtnh); 405 struct nlattr *nla, *attrs = rtnh_attrs(rtnh);
405 406
406 nla = nla_find(attrs, attrlen, RTA_GATEWAY); 407 nla = nla_find(attrs, attrlen, RTA_GATEWAY);
407 nh->nh_gw = nla ? nla_get_be32(nla) : 0; 408 nexthop_nh->nh_gw = nla ? nla_get_be32(nla) : 0;
408#ifdef CONFIG_NET_CLS_ROUTE 409#ifdef CONFIG_NET_CLS_ROUTE
409 nla = nla_find(attrs, attrlen, RTA_FLOW); 410 nla = nla_find(attrs, attrlen, RTA_FLOW);
410 nh->nh_tclassid = nla ? nla_get_u32(nla) : 0; 411 nexthop_nh->nh_tclassid = nla ? nla_get_u32(nla) : 0;
411#endif 412#endif
412 } 413 }
413 414
@@ -527,10 +528,6 @@ static int fib_check_nh(struct fib_config *cfg, struct fib_info *fi,
527 if (nh->nh_gw) { 528 if (nh->nh_gw) {
528 struct fib_result res; 529 struct fib_result res;
529 530
530#ifdef CONFIG_IP_ROUTE_PERVASIVE
531 if (nh->nh_flags&RTNH_F_PERVASIVE)
532 return 0;
533#endif
534 if (nh->nh_flags&RTNH_F_ONLINK) { 531 if (nh->nh_flags&RTNH_F_ONLINK) {
535 struct net_device *dev; 532 struct net_device *dev;
536 533
@@ -738,7 +735,7 @@ struct fib_info *fib_create_info(struct fib_config *cfg)
738 735
739 fi->fib_nhs = nhs; 736 fi->fib_nhs = nhs;
740 change_nexthops(fi) { 737 change_nexthops(fi) {
741 nh->nh_parent = fi; 738 nexthop_nh->nh_parent = fi;
742 } endfor_nexthops(fi) 739 } endfor_nexthops(fi)
743 740
744 if (cfg->fc_mx) { 741 if (cfg->fc_mx) {
@@ -808,7 +805,7 @@ struct fib_info *fib_create_info(struct fib_config *cfg)
808 goto failure; 805 goto failure;
809 } else { 806 } else {
810 change_nexthops(fi) { 807 change_nexthops(fi) {
811 if ((err = fib_check_nh(cfg, fi, nh)) != 0) 808 if ((err = fib_check_nh(cfg, fi, nexthop_nh)) != 0)
812 goto failure; 809 goto failure;
813 } endfor_nexthops(fi) 810 } endfor_nexthops(fi)
814 } 811 }
@@ -843,11 +840,11 @@ link_it:
843 struct hlist_head *head; 840 struct hlist_head *head;
844 unsigned int hash; 841 unsigned int hash;
845 842
846 if (!nh->nh_dev) 843 if (!nexthop_nh->nh_dev)
847 continue; 844 continue;
848 hash = fib_devindex_hashfn(nh->nh_dev->ifindex); 845 hash = fib_devindex_hashfn(nexthop_nh->nh_dev->ifindex);
849 head = &fib_info_devhash[hash]; 846 head = &fib_info_devhash[hash];
850 hlist_add_head(&nh->nh_hash, head); 847 hlist_add_head(&nexthop_nh->nh_hash, head);
851 } endfor_nexthops(fi) 848 } endfor_nexthops(fi)
852 spin_unlock_bh(&fib_info_lock); 849 spin_unlock_bh(&fib_info_lock);
853 return fi; 850 return fi;
@@ -1080,21 +1077,21 @@ int fib_sync_down_dev(struct net_device *dev, int force)
1080 prev_fi = fi; 1077 prev_fi = fi;
1081 dead = 0; 1078 dead = 0;
1082 change_nexthops(fi) { 1079 change_nexthops(fi) {
1083 if (nh->nh_flags&RTNH_F_DEAD) 1080 if (nexthop_nh->nh_flags&RTNH_F_DEAD)
1084 dead++; 1081 dead++;
1085 else if (nh->nh_dev == dev && 1082 else if (nexthop_nh->nh_dev == dev &&
1086 nh->nh_scope != scope) { 1083 nexthop_nh->nh_scope != scope) {
1087 nh->nh_flags |= RTNH_F_DEAD; 1084 nexthop_nh->nh_flags |= RTNH_F_DEAD;
1088#ifdef CONFIG_IP_ROUTE_MULTIPATH 1085#ifdef CONFIG_IP_ROUTE_MULTIPATH
1089 spin_lock_bh(&fib_multipath_lock); 1086 spin_lock_bh(&fib_multipath_lock);
1090 fi->fib_power -= nh->nh_power; 1087 fi->fib_power -= nexthop_nh->nh_power;
1091 nh->nh_power = 0; 1088 nexthop_nh->nh_power = 0;
1092 spin_unlock_bh(&fib_multipath_lock); 1089 spin_unlock_bh(&fib_multipath_lock);
1093#endif 1090#endif
1094 dead++; 1091 dead++;
1095 } 1092 }
1096#ifdef CONFIG_IP_ROUTE_MULTIPATH 1093#ifdef CONFIG_IP_ROUTE_MULTIPATH
1097 if (force > 1 && nh->nh_dev == dev) { 1094 if (force > 1 && nexthop_nh->nh_dev == dev) {
1098 dead = fi->fib_nhs; 1095 dead = fi->fib_nhs;
1099 break; 1096 break;
1100 } 1097 }
@@ -1144,18 +1141,20 @@ int fib_sync_up(struct net_device *dev)
1144 prev_fi = fi; 1141 prev_fi = fi;
1145 alive = 0; 1142 alive = 0;
1146 change_nexthops(fi) { 1143 change_nexthops(fi) {
1147 if (!(nh->nh_flags&RTNH_F_DEAD)) { 1144 if (!(nexthop_nh->nh_flags&RTNH_F_DEAD)) {
1148 alive++; 1145 alive++;
1149 continue; 1146 continue;
1150 } 1147 }
1151 if (nh->nh_dev == NULL || !(nh->nh_dev->flags&IFF_UP)) 1148 if (nexthop_nh->nh_dev == NULL ||
1149 !(nexthop_nh->nh_dev->flags&IFF_UP))
1152 continue; 1150 continue;
1153 if (nh->nh_dev != dev || !__in_dev_get_rtnl(dev)) 1151 if (nexthop_nh->nh_dev != dev ||
1152 !__in_dev_get_rtnl(dev))
1154 continue; 1153 continue;
1155 alive++; 1154 alive++;
1156 spin_lock_bh(&fib_multipath_lock); 1155 spin_lock_bh(&fib_multipath_lock);
1157 nh->nh_power = 0; 1156 nexthop_nh->nh_power = 0;
1158 nh->nh_flags &= ~RTNH_F_DEAD; 1157 nexthop_nh->nh_flags &= ~RTNH_F_DEAD;
1159 spin_unlock_bh(&fib_multipath_lock); 1158 spin_unlock_bh(&fib_multipath_lock);
1160 } endfor_nexthops(fi) 1159 } endfor_nexthops(fi)
1161 1160
@@ -1182,9 +1181,9 @@ void fib_select_multipath(const struct flowi *flp, struct fib_result *res)
1182 if (fi->fib_power <= 0) { 1181 if (fi->fib_power <= 0) {
1183 int power = 0; 1182 int power = 0;
1184 change_nexthops(fi) { 1183 change_nexthops(fi) {
1185 if (!(nh->nh_flags&RTNH_F_DEAD)) { 1184 if (!(nexthop_nh->nh_flags&RTNH_F_DEAD)) {
1186 power += nh->nh_weight; 1185 power += nexthop_nh->nh_weight;
1187 nh->nh_power = nh->nh_weight; 1186 nexthop_nh->nh_power = nexthop_nh->nh_weight;
1188 } 1187 }
1189 } endfor_nexthops(fi); 1188 } endfor_nexthops(fi);
1190 fi->fib_power = power; 1189 fi->fib_power = power;
@@ -1204,9 +1203,10 @@ void fib_select_multipath(const struct flowi *flp, struct fib_result *res)
1204 w = jiffies % fi->fib_power; 1203 w = jiffies % fi->fib_power;
1205 1204
1206 change_nexthops(fi) { 1205 change_nexthops(fi) {
1207 if (!(nh->nh_flags&RTNH_F_DEAD) && nh->nh_power) { 1206 if (!(nexthop_nh->nh_flags&RTNH_F_DEAD) &&
1208 if ((w -= nh->nh_power) <= 0) { 1207 nexthop_nh->nh_power) {
1209 nh->nh_power--; 1208 if ((w -= nexthop_nh->nh_power) <= 0) {
1209 nexthop_nh->nh_power--;
1210 fi->fib_power--; 1210 fi->fib_power--;
1211 res->nh_sel = nhsel; 1211 res->nh_sel = nhsel;
1212 spin_unlock_bh(&fib_multipath_lock); 1212 spin_unlock_bh(&fib_multipath_lock);
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index fe11f60ce41b..4b4c2bcd15db 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -114,7 +114,7 @@ struct icmp_bxm {
114/* An array of errno for error messages from dest unreach. */ 114/* An array of errno for error messages from dest unreach. */
115/* RFC 1122: 3.2.2.1 States that NET_UNREACH, HOST_UNREACH and SR_FAILED MUST be considered 'transient errs'. */ 115/* RFC 1122: 3.2.2.1 States that NET_UNREACH, HOST_UNREACH and SR_FAILED MUST be considered 'transient errs'. */
116 116
117struct icmp_err icmp_err_convert[] = { 117const struct icmp_err icmp_err_convert[] = {
118 { 118 {
119 .errno = ENETUNREACH, /* ICMP_NET_UNREACH */ 119 .errno = ENETUNREACH, /* ICMP_NET_UNREACH */
120 .fatal = 0, 120 .fatal = 0,
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 76c08402c933..d28363998743 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1799,7 +1799,7 @@ int ip_mc_join_group(struct sock *sk , struct ip_mreqn *imr)
1799 iml->next = inet->mc_list; 1799 iml->next = inet->mc_list;
1800 iml->sflist = NULL; 1800 iml->sflist = NULL;
1801 iml->sfmode = MCAST_EXCLUDE; 1801 iml->sfmode = MCAST_EXCLUDE;
1802 inet->mc_list = iml; 1802 rcu_assign_pointer(inet->mc_list, iml);
1803 ip_mc_inc_group(in_dev, addr); 1803 ip_mc_inc_group(in_dev, addr);
1804 err = 0; 1804 err = 0;
1805done: 1805done:
@@ -1807,24 +1807,46 @@ done:
1807 return err; 1807 return err;
1808} 1808}
1809 1809
1810static void ip_sf_socklist_reclaim(struct rcu_head *rp)
1811{
1812 struct ip_sf_socklist *psf;
1813
1814 psf = container_of(rp, struct ip_sf_socklist, rcu);
1815 /* sk_omem_alloc should have been decreased by the caller*/
1816 kfree(psf);
1817}
1818
1810static int ip_mc_leave_src(struct sock *sk, struct ip_mc_socklist *iml, 1819static int ip_mc_leave_src(struct sock *sk, struct ip_mc_socklist *iml,
1811 struct in_device *in_dev) 1820 struct in_device *in_dev)
1812{ 1821{
1822 struct ip_sf_socklist *psf = iml->sflist;
1813 int err; 1823 int err;
1814 1824
1815 if (iml->sflist == NULL) { 1825 if (psf == NULL) {
1816 /* any-source empty exclude case */ 1826 /* any-source empty exclude case */
1817 return ip_mc_del_src(in_dev, &iml->multi.imr_multiaddr.s_addr, 1827 return ip_mc_del_src(in_dev, &iml->multi.imr_multiaddr.s_addr,
1818 iml->sfmode, 0, NULL, 0); 1828 iml->sfmode, 0, NULL, 0);
1819 } 1829 }
1820 err = ip_mc_del_src(in_dev, &iml->multi.imr_multiaddr.s_addr, 1830 err = ip_mc_del_src(in_dev, &iml->multi.imr_multiaddr.s_addr,
1821 iml->sfmode, iml->sflist->sl_count, 1831 iml->sfmode, psf->sl_count, psf->sl_addr, 0);
1822 iml->sflist->sl_addr, 0); 1832 rcu_assign_pointer(iml->sflist, NULL);
1823 sock_kfree_s(sk, iml->sflist, IP_SFLSIZE(iml->sflist->sl_max)); 1833 /* decrease mem now to avoid the memleak warning */
1824 iml->sflist = NULL; 1834 atomic_sub(IP_SFLSIZE(psf->sl_max), &sk->sk_omem_alloc);
1835 call_rcu(&psf->rcu, ip_sf_socklist_reclaim);
1825 return err; 1836 return err;
1826} 1837}
1827 1838
1839
1840static void ip_mc_socklist_reclaim(struct rcu_head *rp)
1841{
1842 struct ip_mc_socklist *iml;
1843
1844 iml = container_of(rp, struct ip_mc_socklist, rcu);
1845 /* sk_omem_alloc should have been decreased by the caller*/
1846 kfree(iml);
1847}
1848
1849
1828/* 1850/*
1829 * Ask a socket to leave a group. 1851 * Ask a socket to leave a group.
1830 */ 1852 */
@@ -1854,12 +1876,14 @@ int ip_mc_leave_group(struct sock *sk, struct ip_mreqn *imr)
1854 1876
1855 (void) ip_mc_leave_src(sk, iml, in_dev); 1877 (void) ip_mc_leave_src(sk, iml, in_dev);
1856 1878
1857 *imlp = iml->next; 1879 rcu_assign_pointer(*imlp, iml->next);
1858 1880
1859 if (in_dev) 1881 if (in_dev)
1860 ip_mc_dec_group(in_dev, group); 1882 ip_mc_dec_group(in_dev, group);
1861 rtnl_unlock(); 1883 rtnl_unlock();
1862 sock_kfree_s(sk, iml, sizeof(*iml)); 1884 /* decrease mem now to avoid the memleak warning */
1885 atomic_sub(sizeof(*iml), &sk->sk_omem_alloc);
1886 call_rcu(&iml->rcu, ip_mc_socklist_reclaim);
1863 return 0; 1887 return 0;
1864 } 1888 }
1865 if (!in_dev) 1889 if (!in_dev)
@@ -1974,9 +1998,12 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
1974 if (psl) { 1998 if (psl) {
1975 for (i=0; i<psl->sl_count; i++) 1999 for (i=0; i<psl->sl_count; i++)
1976 newpsl->sl_addr[i] = psl->sl_addr[i]; 2000 newpsl->sl_addr[i] = psl->sl_addr[i];
1977 sock_kfree_s(sk, psl, IP_SFLSIZE(psl->sl_max)); 2001 /* decrease mem now to avoid the memleak warning */
2002 atomic_sub(IP_SFLSIZE(psl->sl_max), &sk->sk_omem_alloc);
2003 call_rcu(&psl->rcu, ip_sf_socklist_reclaim);
1978 } 2004 }
1979 pmc->sflist = psl = newpsl; 2005 rcu_assign_pointer(pmc->sflist, newpsl);
2006 psl = newpsl;
1980 } 2007 }
1981 rv = 1; /* > 0 for insert logic below if sl_count is 0 */ 2008 rv = 1; /* > 0 for insert logic below if sl_count is 0 */
1982 for (i=0; i<psl->sl_count; i++) { 2009 for (i=0; i<psl->sl_count; i++) {
@@ -2072,11 +2099,13 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex)
2072 if (psl) { 2099 if (psl) {
2073 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode, 2100 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode,
2074 psl->sl_count, psl->sl_addr, 0); 2101 psl->sl_count, psl->sl_addr, 0);
2075 sock_kfree_s(sk, psl, IP_SFLSIZE(psl->sl_max)); 2102 /* decrease mem now to avoid the memleak warning */
2103 atomic_sub(IP_SFLSIZE(psl->sl_max), &sk->sk_omem_alloc);
2104 call_rcu(&psl->rcu, ip_sf_socklist_reclaim);
2076 } else 2105 } else
2077 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode, 2106 (void) ip_mc_del_src(in_dev, &msf->imsf_multiaddr, pmc->sfmode,
2078 0, NULL, 0); 2107 0, NULL, 0);
2079 pmc->sflist = newpsl; 2108 rcu_assign_pointer(pmc->sflist, newpsl);
2080 pmc->sfmode = msf->imsf_fmode; 2109 pmc->sfmode = msf->imsf_fmode;
2081 err = 0; 2110 err = 0;
2082done: 2111done:
@@ -2209,30 +2238,40 @@ int ip_mc_sf_allow(struct sock *sk, __be32 loc_addr, __be32 rmt_addr, int dif)
2209 struct ip_mc_socklist *pmc; 2238 struct ip_mc_socklist *pmc;
2210 struct ip_sf_socklist *psl; 2239 struct ip_sf_socklist *psl;
2211 int i; 2240 int i;
2241 int ret;
2212 2242
2243 ret = 1;
2213 if (!ipv4_is_multicast(loc_addr)) 2244 if (!ipv4_is_multicast(loc_addr))
2214 return 1; 2245 goto out;
2215 2246
2216 for (pmc=inet->mc_list; pmc; pmc=pmc->next) { 2247 rcu_read_lock();
2248 for (pmc=rcu_dereference(inet->mc_list); pmc; pmc=rcu_dereference(pmc->next)) {
2217 if (pmc->multi.imr_multiaddr.s_addr == loc_addr && 2249 if (pmc->multi.imr_multiaddr.s_addr == loc_addr &&
2218 pmc->multi.imr_ifindex == dif) 2250 pmc->multi.imr_ifindex == dif)
2219 break; 2251 break;
2220 } 2252 }
2253 ret = inet->mc_all;
2221 if (!pmc) 2254 if (!pmc)
2222 return inet->mc_all; 2255 goto unlock;
2223 psl = pmc->sflist; 2256 psl = pmc->sflist;
2257 ret = (pmc->sfmode == MCAST_EXCLUDE);
2224 if (!psl) 2258 if (!psl)
2225 return pmc->sfmode == MCAST_EXCLUDE; 2259 goto unlock;
2226 2260
2227 for (i=0; i<psl->sl_count; i++) { 2261 for (i=0; i<psl->sl_count; i++) {
2228 if (psl->sl_addr[i] == rmt_addr) 2262 if (psl->sl_addr[i] == rmt_addr)
2229 break; 2263 break;
2230 } 2264 }
2265 ret = 0;
2231 if (pmc->sfmode == MCAST_INCLUDE && i >= psl->sl_count) 2266 if (pmc->sfmode == MCAST_INCLUDE && i >= psl->sl_count)
2232 return 0; 2267 goto unlock;
2233 if (pmc->sfmode == MCAST_EXCLUDE && i < psl->sl_count) 2268 if (pmc->sfmode == MCAST_EXCLUDE && i < psl->sl_count)
2234 return 0; 2269 goto unlock;
2235 return 1; 2270 ret = 1;
2271unlock:
2272 rcu_read_unlock();
2273out:
2274 return ret;
2236} 2275}
2237 2276
2238/* 2277/*
@@ -2251,7 +2290,7 @@ void ip_mc_drop_socket(struct sock *sk)
2251 rtnl_lock(); 2290 rtnl_lock();
2252 while ((iml = inet->mc_list) != NULL) { 2291 while ((iml = inet->mc_list) != NULL) {
2253 struct in_device *in_dev; 2292 struct in_device *in_dev;
2254 inet->mc_list = iml->next; 2293 rcu_assign_pointer(inet->mc_list, iml->next);
2255 2294
2256 in_dev = inetdev_by_index(net, iml->multi.imr_ifindex); 2295 in_dev = inetdev_by_index(net, iml->multi.imr_ifindex);
2257 (void) ip_mc_leave_src(sk, iml, in_dev); 2296 (void) ip_mc_leave_src(sk, iml, in_dev);
@@ -2259,7 +2298,9 @@ void ip_mc_drop_socket(struct sock *sk)
2259 ip_mc_dec_group(in_dev, iml->multi.imr_multiaddr.s_addr); 2298 ip_mc_dec_group(in_dev, iml->multi.imr_multiaddr.s_addr);
2260 in_dev_put(in_dev); 2299 in_dev_put(in_dev);
2261 } 2300 }
2262 sock_kfree_s(sk, iml, sizeof(*iml)); 2301 /* decrease mem now to avoid the memleak warning */
2302 atomic_sub(sizeof(*iml), &sk->sk_omem_alloc);
2303 call_rcu(&iml->rcu, ip_mc_socklist_reclaim);
2263 } 2304 }
2264 rtnl_unlock(); 2305 rtnl_unlock();
2265} 2306}
@@ -2603,7 +2644,7 @@ static const struct file_operations igmp_mcf_seq_fops = {
2603 .release = seq_release_net, 2644 .release = seq_release_net,
2604}; 2645};
2605 2646
2606static int igmp_net_init(struct net *net) 2647static int __net_init igmp_net_init(struct net *net)
2607{ 2648{
2608 struct proc_dir_entry *pde; 2649 struct proc_dir_entry *pde;
2609 2650
@@ -2621,7 +2662,7 @@ out_igmp:
2621 return -ENOMEM; 2662 return -ENOMEM;
2622} 2663}
2623 2664
2624static void igmp_net_exit(struct net *net) 2665static void __net_exit igmp_net_exit(struct net *net)
2625{ 2666{
2626 proc_net_remove(net, "mcfilter"); 2667 proc_net_remove(net, "mcfilter");
2627 proc_net_remove(net, "igmp"); 2668 proc_net_remove(net, "igmp");
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index ee16475f8fc3..8da6429269dd 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -529,6 +529,8 @@ void inet_csk_reqsk_queue_prune(struct sock *parent,
529 syn_ack_recalc(req, thresh, max_retries, 529 syn_ack_recalc(req, thresh, max_retries,
530 queue->rskq_defer_accept, 530 queue->rskq_defer_accept,
531 &expire, &resend); 531 &expire, &resend);
532 if (req->rsk_ops->syn_ack_timeout)
533 req->rsk_ops->syn_ack_timeout(parent, req);
532 if (!expire && 534 if (!expire &&
533 (!resend || 535 (!resend ||
534 !req->rsk_ops->rtx_syn_ack(parent, req, NULL) || 536 !req->rsk_ops->rtx_syn_ack(parent, req, NULL) ||
diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c
index bdb78dd180ce..1aaa8110d84b 100644
--- a/net/ipv4/inet_diag.c
+++ b/net/ipv4/inet_diag.c
@@ -368,7 +368,7 @@ static int inet_diag_bc_run(const void *bc, int len,
368 yes = entry->sport >= op[1].no; 368 yes = entry->sport >= op[1].no;
369 break; 369 break;
370 case INET_DIAG_BC_S_LE: 370 case INET_DIAG_BC_S_LE:
371 yes = entry->dport <= op[1].no; 371 yes = entry->sport <= op[1].no;
372 break; 372 break;
373 case INET_DIAG_BC_D_GE: 373 case INET_DIAG_BC_D_GE:
374 yes = entry->dport >= op[1].no; 374 yes = entry->dport >= op[1].no;
diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index 86964b353c31..b59430bc041c 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -32,6 +32,8 @@
32#include <linux/netdevice.h> 32#include <linux/netdevice.h>
33#include <linux/jhash.h> 33#include <linux/jhash.h>
34#include <linux/random.h> 34#include <linux/random.h>
35#include <net/route.h>
36#include <net/dst.h>
35#include <net/sock.h> 37#include <net/sock.h>
36#include <net/ip.h> 38#include <net/ip.h>
37#include <net/icmp.h> 39#include <net/icmp.h>
@@ -205,11 +207,34 @@ static void ip_expire(unsigned long arg)
205 if ((qp->q.last_in & INET_FRAG_FIRST_IN) && qp->q.fragments != NULL) { 207 if ((qp->q.last_in & INET_FRAG_FIRST_IN) && qp->q.fragments != NULL) {
206 struct sk_buff *head = qp->q.fragments; 208 struct sk_buff *head = qp->q.fragments;
207 209
208 /* Send an ICMP "Fragment Reassembly Timeout" message. */
209 rcu_read_lock(); 210 rcu_read_lock();
210 head->dev = dev_get_by_index_rcu(net, qp->iif); 211 head->dev = dev_get_by_index_rcu(net, qp->iif);
211 if (head->dev) 212 if (!head->dev)
212 icmp_send(head, ICMP_TIME_EXCEEDED, ICMP_EXC_FRAGTIME, 0); 213 goto out_rcu_unlock;
214
215 /*
216 * Only search router table for the head fragment,
217 * when defraging timeout at PRE_ROUTING HOOK.
218 */
219 if (qp->user == IP_DEFRAG_CONNTRACK_IN && !skb_dst(head)) {
220 const struct iphdr *iph = ip_hdr(head);
221 int err = ip_route_input(head, iph->daddr, iph->saddr,
222 iph->tos, head->dev);
223 if (unlikely(err))
224 goto out_rcu_unlock;
225
226 /*
227 * Only an end host needs to send an ICMP
228 * "Fragment Reassembly Timeout" message, per RFC792.
229 */
230 if (skb_rtable(head)->rt_type != RTN_LOCAL)
231 goto out_rcu_unlock;
232
233 }
234
235 /* Send an ICMP "Fragment Reassembly Timeout" message. */
236 icmp_send(head, ICMP_TIME_EXCEEDED, ICMP_EXC_FRAGTIME, 0);
237out_rcu_unlock:
213 rcu_read_unlock(); 238 rcu_read_unlock();
214 } 239 }
215out: 240out:
@@ -646,7 +671,7 @@ static struct ctl_table ip4_frags_ctl_table[] = {
646 { } 671 { }
647}; 672};
648 673
649static int ip4_frags_ns_ctl_register(struct net *net) 674static int __net_init ip4_frags_ns_ctl_register(struct net *net)
650{ 675{
651 struct ctl_table *table; 676 struct ctl_table *table;
652 struct ctl_table_header *hdr; 677 struct ctl_table_header *hdr;
@@ -676,7 +701,7 @@ err_alloc:
676 return -ENOMEM; 701 return -ENOMEM;
677} 702}
678 703
679static void ip4_frags_ns_ctl_unregister(struct net *net) 704static void __net_exit ip4_frags_ns_ctl_unregister(struct net *net)
680{ 705{
681 struct ctl_table *table; 706 struct ctl_table *table;
682 707
@@ -704,7 +729,7 @@ static inline void ip4_frags_ctl_register(void)
704} 729}
705#endif 730#endif
706 731
707static int ipv4_frags_init_net(struct net *net) 732static int __net_init ipv4_frags_init_net(struct net *net)
708{ 733{
709 /* 734 /*
710 * Fragment cache limits. We will commit 256K at one time. Should we 735 * Fragment cache limits. We will commit 256K at one time. Should we
@@ -726,7 +751,7 @@ static int ipv4_frags_init_net(struct net *net)
726 return ip4_frags_ns_ctl_register(net); 751 return ip4_frags_ns_ctl_register(net);
727} 752}
728 753
729static void ipv4_frags_exit_net(struct net *net) 754static void __net_exit ipv4_frags_exit_net(struct net *net)
730{ 755{
731 ip4_frags_ns_ctl_unregister(net); 756 ip4_frags_ns_ctl_unregister(net);
732 inet_frags_exit_net(&net->ipv4.frags, &ip4_frags); 757 inet_frags_exit_net(&net->ipv4.frags, &ip4_frags);
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index f36ce156cac6..7631b20490f5 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -1307,7 +1307,7 @@ static void ipgre_destroy_tunnels(struct ipgre_net *ign, struct list_head *head)
1307 } 1307 }
1308} 1308}
1309 1309
1310static int ipgre_init_net(struct net *net) 1310static int __net_init ipgre_init_net(struct net *net)
1311{ 1311{
1312 struct ipgre_net *ign = net_generic(net, ipgre_net_id); 1312 struct ipgre_net *ign = net_generic(net, ipgre_net_id);
1313 int err; 1313 int err;
@@ -1334,7 +1334,7 @@ err_alloc_dev:
1334 return err; 1334 return err;
1335} 1335}
1336 1336
1337static void ipgre_exit_net(struct net *net) 1337static void __net_exit ipgre_exit_net(struct net *net)
1338{ 1338{
1339 struct ipgre_net *ign; 1339 struct ipgre_net *ign;
1340 LIST_HEAD(list); 1340 LIST_HEAD(list);
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index e34013a78ef4..3451799e3dbf 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -254,7 +254,7 @@ int ip_mc_output(struct sk_buff *skb)
254 */ 254 */
255 255
256 if (rt->rt_flags&RTCF_MULTICAST) { 256 if (rt->rt_flags&RTCF_MULTICAST) {
257 if ((!sk || inet_sk(sk)->mc_loop) 257 if (sk_mc_loop(sk)
258#ifdef CONFIG_IP_MROUTE 258#ifdef CONFIG_IP_MROUTE
259 /* Small optimization: do not loopback not local frames, 259 /* Small optimization: do not loopback not local frames,
260 which returned after forwarding; they will be dropped 260 which returned after forwarding; they will be dropped
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index cafad9baff03..644dc43a55de 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -451,7 +451,8 @@ static int do_ip_setsockopt(struct sock *sk, int level,
451 (1<<IP_TTL) | (1<<IP_HDRINCL) | 451 (1<<IP_TTL) | (1<<IP_HDRINCL) |
452 (1<<IP_MTU_DISCOVER) | (1<<IP_RECVERR) | 452 (1<<IP_MTU_DISCOVER) | (1<<IP_RECVERR) |
453 (1<<IP_ROUTER_ALERT) | (1<<IP_FREEBIND) | 453 (1<<IP_ROUTER_ALERT) | (1<<IP_FREEBIND) |
454 (1<<IP_PASSSEC) | (1<<IP_TRANSPARENT))) || 454 (1<<IP_PASSSEC) | (1<<IP_TRANSPARENT) |
455 (1<<IP_MINTTL))) ||
455 optname == IP_MULTICAST_TTL || 456 optname == IP_MULTICAST_TTL ||
456 optname == IP_MULTICAST_ALL || 457 optname == IP_MULTICAST_ALL ||
457 optname == IP_MULTICAST_LOOP || 458 optname == IP_MULTICAST_LOOP ||
@@ -936,6 +937,14 @@ mc_msf_out:
936 inet->transparent = !!val; 937 inet->transparent = !!val;
937 break; 938 break;
938 939
940 case IP_MINTTL:
941 if (optlen < 1)
942 goto e_inval;
943 if (val < 0 || val > 255)
944 goto e_inval;
945 inet->min_ttl = val;
946 break;
947
939 default: 948 default:
940 err = -ENOPROTOOPT; 949 err = -ENOPROTOOPT;
941 break; 950 break;
@@ -1198,6 +1207,9 @@ static int do_ip_getsockopt(struct sock *sk, int level, int optname,
1198 case IP_TRANSPARENT: 1207 case IP_TRANSPARENT:
1199 val = inet->transparent; 1208 val = inet->transparent;
1200 break; 1209 break;
1210 case IP_MINTTL:
1211 val = inet->min_ttl;
1212 break;
1201 default: 1213 default:
1202 release_sock(sk); 1214 release_sock(sk);
1203 return -ENOPROTOOPT; 1215 return -ENOPROTOOPT;
diff --git a/net/ipv4/ipcomp.c b/net/ipv4/ipcomp.c
index 38fbf04150ae..b55a0c3df82f 100644
--- a/net/ipv4/ipcomp.c
+++ b/net/ipv4/ipcomp.c
@@ -25,6 +25,7 @@
25 25
26static void ipcomp4_err(struct sk_buff *skb, u32 info) 26static void ipcomp4_err(struct sk_buff *skb, u32 info)
27{ 27{
28 struct net *net = dev_net(skb->dev);
28 __be32 spi; 29 __be32 spi;
29 struct iphdr *iph = (struct iphdr *)skb->data; 30 struct iphdr *iph = (struct iphdr *)skb->data;
30 struct ip_comp_hdr *ipch = (struct ip_comp_hdr *)(skb->data+(iph->ihl<<2)); 31 struct ip_comp_hdr *ipch = (struct ip_comp_hdr *)(skb->data+(iph->ihl<<2));
@@ -35,7 +36,7 @@ static void ipcomp4_err(struct sk_buff *skb, u32 info)
35 return; 36 return;
36 37
37 spi = htonl(ntohs(ipch->cpi)); 38 spi = htonl(ntohs(ipch->cpi));
38 x = xfrm_state_lookup(&init_net, (xfrm_address_t *)&iph->daddr, 39 x = xfrm_state_lookup(net, (xfrm_address_t *)&iph->daddr,
39 spi, IPPROTO_COMP, AF_INET); 40 spi, IPPROTO_COMP, AF_INET);
40 if (!x) 41 if (!x)
41 return; 42 return;
@@ -47,9 +48,10 @@ static void ipcomp4_err(struct sk_buff *skb, u32 info)
47/* We always hold one tunnel user reference to indicate a tunnel */ 48/* We always hold one tunnel user reference to indicate a tunnel */
48static struct xfrm_state *ipcomp_tunnel_create(struct xfrm_state *x) 49static struct xfrm_state *ipcomp_tunnel_create(struct xfrm_state *x)
49{ 50{
51 struct net *net = xs_net(x);
50 struct xfrm_state *t; 52 struct xfrm_state *t;
51 53
52 t = xfrm_state_alloc(&init_net); 54 t = xfrm_state_alloc(net);
53 if (t == NULL) 55 if (t == NULL)
54 goto out; 56 goto out;
55 57
@@ -82,10 +84,11 @@ error:
82 */ 84 */
83static int ipcomp_tunnel_attach(struct xfrm_state *x) 85static int ipcomp_tunnel_attach(struct xfrm_state *x)
84{ 86{
87 struct net *net = xs_net(x);
85 int err = 0; 88 int err = 0;
86 struct xfrm_state *t; 89 struct xfrm_state *t;
87 90
88 t = xfrm_state_lookup(&init_net, (xfrm_address_t *)&x->id.daddr.a4, 91 t = xfrm_state_lookup(net, (xfrm_address_t *)&x->id.daddr.a4,
89 x->props.saddr.a4, IPPROTO_IPIP, AF_INET); 92 x->props.saddr.a4, IPPROTO_IPIP, AF_INET);
90 if (!t) { 93 if (!t) {
91 t = ipcomp_tunnel_create(x); 94 t = ipcomp_tunnel_create(x);
diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c
index 4e08b7f2331c..10a6a604bf32 100644
--- a/net/ipv4/ipconfig.c
+++ b/net/ipv4/ipconfig.c
@@ -1446,7 +1446,7 @@ late_initcall(ip_auto_config);
1446 1446
1447/* 1447/*
1448 * Decode any IP configuration options in the "ip=" or "nfsaddrs=" kernel 1448 * Decode any IP configuration options in the "ip=" or "nfsaddrs=" kernel
1449 * command line parameter. See Documentation/filesystems/nfsroot.txt. 1449 * command line parameter. See Documentation/filesystems/nfs/nfsroot.txt.
1450 */ 1450 */
1451static int __init ic_proto_name(char *name) 1451static int __init ic_proto_name(char *name)
1452{ 1452{
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index eda04fed3379..95db732e542b 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -130,7 +130,6 @@ struct ipip_net {
130 struct net_device *fb_tunnel_dev; 130 struct net_device *fb_tunnel_dev;
131}; 131};
132 132
133static void ipip_fb_tunnel_init(struct net_device *dev);
134static void ipip_tunnel_init(struct net_device *dev); 133static void ipip_tunnel_init(struct net_device *dev);
135static void ipip_tunnel_setup(struct net_device *dev); 134static void ipip_tunnel_setup(struct net_device *dev);
136 135
@@ -730,7 +729,7 @@ static void ipip_tunnel_init(struct net_device *dev)
730 ipip_tunnel_bind_dev(dev); 729 ipip_tunnel_bind_dev(dev);
731} 730}
732 731
733static void ipip_fb_tunnel_init(struct net_device *dev) 732static void __net_init ipip_fb_tunnel_init(struct net_device *dev)
734{ 733{
735 struct ip_tunnel *tunnel = netdev_priv(dev); 734 struct ip_tunnel *tunnel = netdev_priv(dev);
736 struct iphdr *iph = &tunnel->parms.iph; 735 struct iphdr *iph = &tunnel->parms.iph;
@@ -773,7 +772,7 @@ static void ipip_destroy_tunnels(struct ipip_net *ipn, struct list_head *head)
773 } 772 }
774} 773}
775 774
776static int ipip_init_net(struct net *net) 775static int __net_init ipip_init_net(struct net *net)
777{ 776{
778 struct ipip_net *ipn = net_generic(net, ipip_net_id); 777 struct ipip_net *ipn = net_generic(net, ipip_net_id);
779 int err; 778 int err;
@@ -806,7 +805,7 @@ err_alloc_dev:
806 return err; 805 return err;
807} 806}
808 807
809static void ipip_exit_net(struct net *net) 808static void __net_exit ipip_exit_net(struct net *net)
810{ 809{
811 struct ipip_net *ipn = net_generic(net, ipip_net_id); 810 struct ipip_net *ipn = net_generic(net, ipip_net_id);
812 LIST_HEAD(list); 811 LIST_HEAD(list);
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index 06632762ba5f..90203e1b9187 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -925,10 +925,10 @@ static int get_info(struct net *net, void __user *user, int *len, int compat)
925 if (t && !IS_ERR(t)) { 925 if (t && !IS_ERR(t)) {
926 struct arpt_getinfo info; 926 struct arpt_getinfo info;
927 const struct xt_table_info *private = t->private; 927 const struct xt_table_info *private = t->private;
928
929#ifdef CONFIG_COMPAT 928#ifdef CONFIG_COMPAT
929 struct xt_table_info tmp;
930
930 if (compat) { 931 if (compat) {
931 struct xt_table_info tmp;
932 ret = compat_table_info(private, &tmp); 932 ret = compat_table_info(private, &tmp);
933 xt_compat_flush_offsets(NFPROTO_ARP); 933 xt_compat_flush_offsets(NFPROTO_ARP);
934 private = &tmp; 934 private = &tmp;
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index 572330a552ef..3ce53cf13d5a 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -1132,10 +1132,10 @@ static int get_info(struct net *net, void __user *user, int *len, int compat)
1132 if (t && !IS_ERR(t)) { 1132 if (t && !IS_ERR(t)) {
1133 struct ipt_getinfo info; 1133 struct ipt_getinfo info;
1134 const struct xt_table_info *private = t->private; 1134 const struct xt_table_info *private = t->private;
1135
1136#ifdef CONFIG_COMPAT 1135#ifdef CONFIG_COMPAT
1136 struct xt_table_info tmp;
1137
1137 if (compat) { 1138 if (compat) {
1138 struct xt_table_info tmp;
1139 ret = compat_table_info(private, &tmp); 1139 ret = compat_table_info(private, &tmp);
1140 xt_compat_flush_offsets(AF_INET); 1140 xt_compat_flush_offsets(AF_INET);
1141 private = &tmp; 1141 private = &tmp;
diff --git a/net/ipv4/netfilter/ipt_ECN.c b/net/ipv4/netfilter/ipt_ECN.c
index 549e206cdd42..ea5cea2415c1 100644
--- a/net/ipv4/netfilter/ipt_ECN.c
+++ b/net/ipv4/netfilter/ipt_ECN.c
@@ -50,7 +50,7 @@ set_ect_tcp(struct sk_buff *skb, const struct ipt_ECN_info *einfo)
50 struct tcphdr _tcph, *tcph; 50 struct tcphdr _tcph, *tcph;
51 __be16 oldval; 51 __be16 oldval;
52 52
53 /* Not enought header? */ 53 /* Not enough header? */
54 tcph = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_tcph), &_tcph); 54 tcph = skb_header_pointer(skb, ip_hdrlen(skb), sizeof(_tcph), &_tcph);
55 if (!tcph) 55 if (!tcph)
56 return false; 56 return false;
diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
index d171b123a656..d1ea38a7c490 100644
--- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
+++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
@@ -210,7 +210,7 @@ static ctl_table ip_ct_sysctl_table[] = {
210 }, 210 },
211 { 211 {
212 .procname = "ip_conntrack_buckets", 212 .procname = "ip_conntrack_buckets",
213 .data = &nf_conntrack_htable_size, 213 .data = &init_net.ct.htable_size,
214 .maxlen = sizeof(unsigned int), 214 .maxlen = sizeof(unsigned int),
215 .mode = 0444, 215 .mode = 0444,
216 .proc_handler = proc_dointvec, 216 .proc_handler = proc_dointvec,
diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
index 8668a3defda6..2fb7b76da94f 100644
--- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
+++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
@@ -32,7 +32,7 @@ static struct hlist_nulls_node *ct_get_first(struct seq_file *seq)
32 struct hlist_nulls_node *n; 32 struct hlist_nulls_node *n;
33 33
34 for (st->bucket = 0; 34 for (st->bucket = 0;
35 st->bucket < nf_conntrack_htable_size; 35 st->bucket < net->ct.htable_size;
36 st->bucket++) { 36 st->bucket++) {
37 n = rcu_dereference(net->ct.hash[st->bucket].first); 37 n = rcu_dereference(net->ct.hash[st->bucket].first);
38 if (!is_a_nulls(n)) 38 if (!is_a_nulls(n))
@@ -50,7 +50,7 @@ static struct hlist_nulls_node *ct_get_next(struct seq_file *seq,
50 head = rcu_dereference(head->next); 50 head = rcu_dereference(head->next);
51 while (is_a_nulls(head)) { 51 while (is_a_nulls(head)) {
52 if (likely(get_nulls_value(head) == st->bucket)) { 52 if (likely(get_nulls_value(head) == st->bucket)) {
53 if (++st->bucket >= nf_conntrack_htable_size) 53 if (++st->bucket >= net->ct.htable_size)
54 return NULL; 54 return NULL;
55 } 55 }
56 head = rcu_dereference(net->ct.hash[st->bucket].first); 56 head = rcu_dereference(net->ct.hash[st->bucket].first);
diff --git a/net/ipv4/netfilter/nf_defrag_ipv4.c b/net/ipv4/netfilter/nf_defrag_ipv4.c
index fa2d6b6fc3e5..331ead3ebd1b 100644
--- a/net/ipv4/netfilter/nf_defrag_ipv4.c
+++ b/net/ipv4/netfilter/nf_defrag_ipv4.c
@@ -14,6 +14,7 @@
14#include <net/route.h> 14#include <net/route.h>
15#include <net/ip.h> 15#include <net/ip.h>
16 16
17#include <linux/netfilter_bridge.h>
17#include <linux/netfilter_ipv4.h> 18#include <linux/netfilter_ipv4.h>
18#include <net/netfilter/ipv4/nf_defrag_ipv4.h> 19#include <net/netfilter/ipv4/nf_defrag_ipv4.h>
19 20
@@ -34,6 +35,20 @@ static int nf_ct_ipv4_gather_frags(struct sk_buff *skb, u_int32_t user)
34 return err; 35 return err;
35} 36}
36 37
38static enum ip_defrag_users nf_ct_defrag_user(unsigned int hooknum,
39 struct sk_buff *skb)
40{
41#ifdef CONFIG_BRIDGE_NETFILTER
42 if (skb->nf_bridge &&
43 skb->nf_bridge->mask & BRNF_NF_BRIDGE_PREROUTING)
44 return IP_DEFRAG_CONNTRACK_BRIDGE_IN;
45#endif
46 if (hooknum == NF_INET_PRE_ROUTING)
47 return IP_DEFRAG_CONNTRACK_IN;
48 else
49 return IP_DEFRAG_CONNTRACK_OUT;
50}
51
37static unsigned int ipv4_conntrack_defrag(unsigned int hooknum, 52static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
38 struct sk_buff *skb, 53 struct sk_buff *skb,
39 const struct net_device *in, 54 const struct net_device *in,
@@ -50,10 +65,8 @@ static unsigned int ipv4_conntrack_defrag(unsigned int hooknum,
50#endif 65#endif
51 /* Gather fragments. */ 66 /* Gather fragments. */
52 if (ip_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) { 67 if (ip_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) {
53 if (nf_ct_ipv4_gather_frags(skb, 68 enum ip_defrag_users user = nf_ct_defrag_user(hooknum, skb);
54 hooknum == NF_INET_PRE_ROUTING ? 69 if (nf_ct_ipv4_gather_frags(skb, user))
55 IP_DEFRAG_CONNTRACK_IN :
56 IP_DEFRAG_CONNTRACK_OUT))
57 return NF_STOLEN; 70 return NF_STOLEN;
58 } 71 }
59 return NF_ACCEPT; 72 return NF_ACCEPT;
diff --git a/net/ipv4/netfilter/nf_nat_core.c b/net/ipv4/netfilter/nf_nat_core.c
index fe1a64479dd0..26066a2327ad 100644
--- a/net/ipv4/netfilter/nf_nat_core.c
+++ b/net/ipv4/netfilter/nf_nat_core.c
@@ -35,9 +35,6 @@ static DEFINE_SPINLOCK(nf_nat_lock);
35 35
36static struct nf_conntrack_l3proto *l3proto __read_mostly; 36static struct nf_conntrack_l3proto *l3proto __read_mostly;
37 37
38/* Calculated at init based on memory size */
39static unsigned int nf_nat_htable_size __read_mostly;
40
41#define MAX_IP_NAT_PROTO 256 38#define MAX_IP_NAT_PROTO 256
42static const struct nf_nat_protocol *nf_nat_protos[MAX_IP_NAT_PROTO] 39static const struct nf_nat_protocol *nf_nat_protos[MAX_IP_NAT_PROTO]
43 __read_mostly; 40 __read_mostly;
@@ -72,7 +69,7 @@ EXPORT_SYMBOL_GPL(nf_nat_proto_put);
72 69
73/* We keep an extra hash for each conntrack, for fast searching. */ 70/* We keep an extra hash for each conntrack, for fast searching. */
74static inline unsigned int 71static inline unsigned int
75hash_by_src(const struct nf_conntrack_tuple *tuple) 72hash_by_src(const struct net *net, const struct nf_conntrack_tuple *tuple)
76{ 73{
77 unsigned int hash; 74 unsigned int hash;
78 75
@@ -80,7 +77,7 @@ hash_by_src(const struct nf_conntrack_tuple *tuple)
80 hash = jhash_3words((__force u32)tuple->src.u3.ip, 77 hash = jhash_3words((__force u32)tuple->src.u3.ip,
81 (__force u32)tuple->src.u.all, 78 (__force u32)tuple->src.u.all,
82 tuple->dst.protonum, 0); 79 tuple->dst.protonum, 0);
83 return ((u64)hash * nf_nat_htable_size) >> 32; 80 return ((u64)hash * net->ipv4.nat_htable_size) >> 32;
84} 81}
85 82
86/* Is this tuple already taken? (not by us) */ 83/* Is this tuple already taken? (not by us) */
@@ -147,7 +144,7 @@ find_appropriate_src(struct net *net,
147 struct nf_conntrack_tuple *result, 144 struct nf_conntrack_tuple *result,
148 const struct nf_nat_range *range) 145 const struct nf_nat_range *range)
149{ 146{
150 unsigned int h = hash_by_src(tuple); 147 unsigned int h = hash_by_src(net, tuple);
151 const struct nf_conn_nat *nat; 148 const struct nf_conn_nat *nat;
152 const struct nf_conn *ct; 149 const struct nf_conn *ct;
153 const struct hlist_node *n; 150 const struct hlist_node *n;
@@ -330,7 +327,7 @@ nf_nat_setup_info(struct nf_conn *ct,
330 if (have_to_hash) { 327 if (have_to_hash) {
331 unsigned int srchash; 328 unsigned int srchash;
332 329
333 srchash = hash_by_src(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple); 330 srchash = hash_by_src(net, &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
334 spin_lock_bh(&nf_nat_lock); 331 spin_lock_bh(&nf_nat_lock);
335 /* nf_conntrack_alter_reply might re-allocate exntension aera */ 332 /* nf_conntrack_alter_reply might re-allocate exntension aera */
336 nat = nfct_nat(ct); 333 nat = nfct_nat(ct);
@@ -679,8 +676,10 @@ nfnetlink_parse_nat_setup(struct nf_conn *ct,
679 676
680static int __net_init nf_nat_net_init(struct net *net) 677static int __net_init nf_nat_net_init(struct net *net)
681{ 678{
682 net->ipv4.nat_bysource = nf_ct_alloc_hashtable(&nf_nat_htable_size, 679 /* Leave them the same for the moment. */
683 &net->ipv4.nat_vmalloced, 0); 680 net->ipv4.nat_htable_size = net->ct.htable_size;
681 net->ipv4.nat_bysource = nf_ct_alloc_hashtable(&net->ipv4.nat_htable_size,
682 &net->ipv4.nat_vmalloced, 0);
684 if (!net->ipv4.nat_bysource) 683 if (!net->ipv4.nat_bysource)
685 return -ENOMEM; 684 return -ENOMEM;
686 return 0; 685 return 0;
@@ -703,7 +702,7 @@ static void __net_exit nf_nat_net_exit(struct net *net)
703 nf_ct_iterate_cleanup(net, &clean_nat, NULL); 702 nf_ct_iterate_cleanup(net, &clean_nat, NULL);
704 synchronize_rcu(); 703 synchronize_rcu();
705 nf_ct_free_hashtable(net->ipv4.nat_bysource, net->ipv4.nat_vmalloced, 704 nf_ct_free_hashtable(net->ipv4.nat_bysource, net->ipv4.nat_vmalloced,
706 nf_nat_htable_size); 705 net->ipv4.nat_htable_size);
707} 706}
708 707
709static struct pernet_operations nf_nat_net_ops = { 708static struct pernet_operations nf_nat_net_ops = {
@@ -724,9 +723,6 @@ static int __init nf_nat_init(void)
724 return ret; 723 return ret;
725 } 724 }
726 725
727 /* Leave them the same for the moment. */
728 nf_nat_htable_size = nf_conntrack_htable_size;
729
730 ret = register_pernet_subsys(&nf_nat_net_ops); 726 ret = register_pernet_subsys(&nf_nat_net_ops);
731 if (ret < 0) 727 if (ret < 0)
732 goto cleanup_extend; 728 goto cleanup_extend;
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index f25542c48b7d..1b09a6dde7c0 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -127,8 +127,8 @@ static const struct snmp_mib snmp4_ipextstats_list[] = {
127 SNMP_MIB_SENTINEL 127 SNMP_MIB_SENTINEL
128}; 128};
129 129
130static struct { 130static const struct {
131 char *name; 131 const char *name;
132 int index; 132 int index;
133} icmpmibmap[] = { 133} icmpmibmap[] = {
134 { "DestUnreachs", ICMP_DEST_UNREACH }, 134 { "DestUnreachs", ICMP_DEST_UNREACH },
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index e446496f564f..b16dfadbe6d6 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -586,7 +586,9 @@ static void __net_exit ip_rt_do_proc_exit(struct net *net)
586{ 586{
587 remove_proc_entry("rt_cache", net->proc_net_stat); 587 remove_proc_entry("rt_cache", net->proc_net_stat);
588 remove_proc_entry("rt_cache", net->proc_net); 588 remove_proc_entry("rt_cache", net->proc_net);
589#ifdef CONFIG_NET_CLS_ROUTE
589 remove_proc_entry("rt_acct", net->proc_net); 590 remove_proc_entry("rt_acct", net->proc_net);
591#endif
590} 592}
591 593
592static struct pernet_operations ip_rt_proc_ops __net_initdata = { 594static struct pernet_operations ip_rt_proc_ops __net_initdata = {
@@ -1988,8 +1990,13 @@ static int __mkroute_input(struct sk_buff *skb,
1988 if (skb->protocol != htons(ETH_P_IP)) { 1990 if (skb->protocol != htons(ETH_P_IP)) {
1989 /* Not IP (i.e. ARP). Do not create route, if it is 1991 /* Not IP (i.e. ARP). Do not create route, if it is
1990 * invalid for proxy arp. DNAT routes are always valid. 1992 * invalid for proxy arp. DNAT routes are always valid.
1993 *
1994 * Proxy arp feature have been extended to allow, ARP
1995 * replies back to the same interface, to support
1996 * Private VLAN switch technologies. See arp.c.
1991 */ 1997 */
1992 if (out_dev == in_dev) { 1998 if (out_dev == in_dev &&
1999 IN_DEV_PROXY_ARP_PVLAN(in_dev) == 0) {
1993 err = -EINVAL; 2000 err = -EINVAL;
1994 goto cleanup; 2001 goto cleanup;
1995 } 2002 }
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 26399ad2a289..5c24db4a3c91 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -277,6 +277,13 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
277 277
278 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV); 278 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_SYNCOOKIESRECV);
279 279
280 /* check for timestamp cookie support */
281 memset(&tcp_opt, 0, sizeof(tcp_opt));
282 tcp_parse_options(skb, &tcp_opt, &hash_location, 0);
283
284 if (tcp_opt.saw_tstamp)
285 cookie_check_timestamp(&tcp_opt);
286
280 ret = NULL; 287 ret = NULL;
281 req = inet_reqsk_alloc(&tcp_request_sock_ops); /* for safety */ 288 req = inet_reqsk_alloc(&tcp_request_sock_ops); /* for safety */
282 if (!req) 289 if (!req)
@@ -292,6 +299,12 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
292 ireq->loc_addr = ip_hdr(skb)->daddr; 299 ireq->loc_addr = ip_hdr(skb)->daddr;
293 ireq->rmt_addr = ip_hdr(skb)->saddr; 300 ireq->rmt_addr = ip_hdr(skb)->saddr;
294 ireq->ecn_ok = 0; 301 ireq->ecn_ok = 0;
302 ireq->snd_wscale = tcp_opt.snd_wscale;
303 ireq->rcv_wscale = tcp_opt.rcv_wscale;
304 ireq->sack_ok = tcp_opt.sack_ok;
305 ireq->wscale_ok = tcp_opt.wscale_ok;
306 ireq->tstamp_ok = tcp_opt.saw_tstamp;
307 req->ts_recent = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
295 308
296 /* We throwed the options of the initial SYN away, so we hope 309 /* We throwed the options of the initial SYN away, so we hope
297 * the ACK carries the same options again (see RFC1122 4.2.3.8) 310 * the ACK carries the same options again (see RFC1122 4.2.3.8)
@@ -340,26 +353,13 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
340 } 353 }
341 } 354 }
342 355
343 /* check for timestamp cookie support */
344 memset(&tcp_opt, 0, sizeof(tcp_opt));
345 tcp_parse_options(skb, &tcp_opt, &hash_location, 0, &rt->u.dst);
346
347 if (tcp_opt.saw_tstamp)
348 cookie_check_timestamp(&tcp_opt);
349
350 ireq->snd_wscale = tcp_opt.snd_wscale;
351 ireq->rcv_wscale = tcp_opt.rcv_wscale;
352 ireq->sack_ok = tcp_opt.sack_ok;
353 ireq->wscale_ok = tcp_opt.wscale_ok;
354 ireq->tstamp_ok = tcp_opt.saw_tstamp;
355 req->ts_recent = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
356
357 /* Try to redo what tcp_v4_send_synack did. */ 356 /* Try to redo what tcp_v4_send_synack did. */
358 req->window_clamp = tp->window_clamp ? :dst_metric(&rt->u.dst, RTAX_WINDOW); 357 req->window_clamp = tp->window_clamp ? :dst_metric(&rt->u.dst, RTAX_WINDOW);
359 358
360 tcp_select_initial_window(tcp_full_space(sk), req->mss, 359 tcp_select_initial_window(tcp_full_space(sk), req->mss,
361 &req->rcv_wnd, &req->window_clamp, 360 &req->rcv_wnd, &req->window_clamp,
362 ireq->wscale_ok, &rcv_wscale); 361 ireq->wscale_ok, &rcv_wscale,
362 dst_metric(&rt->u.dst, RTAX_INITRWND));
363 363
364 ireq->rcv_wscale = rcv_wscale; 364 ireq->rcv_wscale = rcv_wscale;
365 365
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index b0a26bb25e2e..d5d69ea8f249 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -536,8 +536,7 @@ static inline void skb_entail(struct sock *sk, struct sk_buff *skb)
536 tp->nonagle &= ~TCP_NAGLE_PUSH; 536 tp->nonagle &= ~TCP_NAGLE_PUSH;
537} 537}
538 538
539static inline void tcp_mark_urg(struct tcp_sock *tp, int flags, 539static inline void tcp_mark_urg(struct tcp_sock *tp, int flags)
540 struct sk_buff *skb)
541{ 540{
542 if (flags & MSG_OOB) 541 if (flags & MSG_OOB)
543 tp->snd_up = tp->write_seq; 542 tp->snd_up = tp->write_seq;
@@ -546,13 +545,13 @@ static inline void tcp_mark_urg(struct tcp_sock *tp, int flags,
546static inline void tcp_push(struct sock *sk, int flags, int mss_now, 545static inline void tcp_push(struct sock *sk, int flags, int mss_now,
547 int nonagle) 546 int nonagle)
548{ 547{
549 struct tcp_sock *tp = tcp_sk(sk);
550
551 if (tcp_send_head(sk)) { 548 if (tcp_send_head(sk)) {
552 struct sk_buff *skb = tcp_write_queue_tail(sk); 549 struct tcp_sock *tp = tcp_sk(sk);
550
553 if (!(flags & MSG_MORE) || forced_push(tp)) 551 if (!(flags & MSG_MORE) || forced_push(tp))
554 tcp_mark_push(tp, skb); 552 tcp_mark_push(tp, tcp_write_queue_tail(sk));
555 tcp_mark_urg(tp, flags, skb); 553
554 tcp_mark_urg(tp, flags);
556 __tcp_push_pending_frames(sk, mss_now, 555 __tcp_push_pending_frames(sk, mss_now,
557 (flags & MSG_MORE) ? TCP_NAGLE_CORK : nonagle); 556 (flags & MSG_MORE) ? TCP_NAGLE_CORK : nonagle);
558 } 557 }
@@ -877,12 +876,12 @@ ssize_t tcp_sendpage(struct socket *sock, struct page *page, int offset,
877#define TCP_PAGE(sk) (sk->sk_sndmsg_page) 876#define TCP_PAGE(sk) (sk->sk_sndmsg_page)
878#define TCP_OFF(sk) (sk->sk_sndmsg_off) 877#define TCP_OFF(sk) (sk->sk_sndmsg_off)
879 878
880static inline int select_size(struct sock *sk) 879static inline int select_size(struct sock *sk, int sg)
881{ 880{
882 struct tcp_sock *tp = tcp_sk(sk); 881 struct tcp_sock *tp = tcp_sk(sk);
883 int tmp = tp->mss_cache; 882 int tmp = tp->mss_cache;
884 883
885 if (sk->sk_route_caps & NETIF_F_SG) { 884 if (sg) {
886 if (sk_can_gso(sk)) 885 if (sk_can_gso(sk))
887 tmp = 0; 886 tmp = 0;
888 else { 887 else {
@@ -906,7 +905,7 @@ int tcp_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
906 struct sk_buff *skb; 905 struct sk_buff *skb;
907 int iovlen, flags; 906 int iovlen, flags;
908 int mss_now, size_goal; 907 int mss_now, size_goal;
909 int err, copied; 908 int sg, err, copied;
910 long timeo; 909 long timeo;
911 910
912 lock_sock(sk); 911 lock_sock(sk);
@@ -934,6 +933,8 @@ int tcp_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
934 if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN)) 933 if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
935 goto out_err; 934 goto out_err;
936 935
936 sg = sk->sk_route_caps & NETIF_F_SG;
937
937 while (--iovlen >= 0) { 938 while (--iovlen >= 0) {
938 int seglen = iov->iov_len; 939 int seglen = iov->iov_len;
939 unsigned char __user *from = iov->iov_base; 940 unsigned char __user *from = iov->iov_base;
@@ -959,8 +960,9 @@ new_segment:
959 if (!sk_stream_memory_free(sk)) 960 if (!sk_stream_memory_free(sk))
960 goto wait_for_sndbuf; 961 goto wait_for_sndbuf;
961 962
962 skb = sk_stream_alloc_skb(sk, select_size(sk), 963 skb = sk_stream_alloc_skb(sk,
963 sk->sk_allocation); 964 select_size(sk, sg),
965 sk->sk_allocation);
964 if (!skb) 966 if (!skb)
965 goto wait_for_memory; 967 goto wait_for_memory;
966 968
@@ -997,9 +999,7 @@ new_segment:
997 /* We can extend the last page 999 /* We can extend the last page
998 * fragment. */ 1000 * fragment. */
999 merge = 1; 1001 merge = 1;
1000 } else if (i == MAX_SKB_FRAGS || 1002 } else if (i == MAX_SKB_FRAGS || !sg) {
1001 (!i &&
1002 !(sk->sk_route_caps & NETIF_F_SG))) {
1003 /* Need to add new fragment and cannot 1003 /* Need to add new fragment and cannot
1004 * do this because interface is non-SG, 1004 * do this because interface is non-SG,
1005 * or because all the page slots are 1005 * or because all the page slots are
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 12cab7d74dba..28e029632493 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3727,7 +3727,7 @@ old_ack:
3727 * the fast version below fails. 3727 * the fast version below fails.
3728 */ 3728 */
3729void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx, 3729void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
3730 u8 **hvpp, int estab, struct dst_entry *dst) 3730 u8 **hvpp, int estab)
3731{ 3731{
3732 unsigned char *ptr; 3732 unsigned char *ptr;
3733 struct tcphdr *th = tcp_hdr(skb); 3733 struct tcphdr *th = tcp_hdr(skb);
@@ -3766,8 +3766,7 @@ void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
3766 break; 3766 break;
3767 case TCPOPT_WINDOW: 3767 case TCPOPT_WINDOW:
3768 if (opsize == TCPOLEN_WINDOW && th->syn && 3768 if (opsize == TCPOLEN_WINDOW && th->syn &&
3769 !estab && sysctl_tcp_window_scaling && 3769 !estab && sysctl_tcp_window_scaling) {
3770 !dst_feature(dst, RTAX_FEATURE_NO_WSCALE)) {
3771 __u8 snd_wscale = *(__u8 *)ptr; 3770 __u8 snd_wscale = *(__u8 *)ptr;
3772 opt_rx->wscale_ok = 1; 3771 opt_rx->wscale_ok = 1;
3773 if (snd_wscale > 14) { 3772 if (snd_wscale > 14) {
@@ -3783,8 +3782,7 @@ void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
3783 case TCPOPT_TIMESTAMP: 3782 case TCPOPT_TIMESTAMP:
3784 if ((opsize == TCPOLEN_TIMESTAMP) && 3783 if ((opsize == TCPOLEN_TIMESTAMP) &&
3785 ((estab && opt_rx->tstamp_ok) || 3784 ((estab && opt_rx->tstamp_ok) ||
3786 (!estab && sysctl_tcp_timestamps && 3785 (!estab && sysctl_tcp_timestamps))) {
3787 !dst_feature(dst, RTAX_FEATURE_NO_TSTAMP)))) {
3788 opt_rx->saw_tstamp = 1; 3786 opt_rx->saw_tstamp = 1;
3789 opt_rx->rcv_tsval = get_unaligned_be32(ptr); 3787 opt_rx->rcv_tsval = get_unaligned_be32(ptr);
3790 opt_rx->rcv_tsecr = get_unaligned_be32(ptr + 4); 3788 opt_rx->rcv_tsecr = get_unaligned_be32(ptr + 4);
@@ -3792,8 +3790,7 @@ void tcp_parse_options(struct sk_buff *skb, struct tcp_options_received *opt_rx,
3792 break; 3790 break;
3793 case TCPOPT_SACK_PERM: 3791 case TCPOPT_SACK_PERM:
3794 if (opsize == TCPOLEN_SACK_PERM && th->syn && 3792 if (opsize == TCPOLEN_SACK_PERM && th->syn &&
3795 !estab && sysctl_tcp_sack && 3793 !estab && sysctl_tcp_sack) {
3796 !dst_feature(dst, RTAX_FEATURE_NO_SACK)) {
3797 opt_rx->sack_ok = 1; 3794 opt_rx->sack_ok = 1;
3798 tcp_sack_reset(opt_rx); 3795 tcp_sack_reset(opt_rx);
3799 } 3796 }
@@ -3878,7 +3875,7 @@ static int tcp_fast_parse_options(struct sk_buff *skb, struct tcphdr *th,
3878 if (tcp_parse_aligned_timestamp(tp, th)) 3875 if (tcp_parse_aligned_timestamp(tp, th))
3879 return 1; 3876 return 1;
3880 } 3877 }
3881 tcp_parse_options(skb, &tp->rx_opt, hvpp, 1, NULL); 3878 tcp_parse_options(skb, &tp->rx_opt, hvpp, 1);
3882 return 1; 3879 return 1;
3883} 3880}
3884 3881
@@ -4133,10 +4130,8 @@ static inline int tcp_sack_extend(struct tcp_sack_block *sp, u32 seq,
4133static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq) 4130static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq)
4134{ 4131{
4135 struct tcp_sock *tp = tcp_sk(sk); 4132 struct tcp_sock *tp = tcp_sk(sk);
4136 struct dst_entry *dst = __sk_dst_get(sk);
4137 4133
4138 if (tcp_is_sack(tp) && sysctl_tcp_dsack && 4134 if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
4139 !dst_feature(dst, RTAX_FEATURE_NO_DSACK)) {
4140 int mib_idx; 4135 int mib_idx;
4141 4136
4142 if (before(seq, tp->rcv_nxt)) 4137 if (before(seq, tp->rcv_nxt))
@@ -4165,15 +4160,13 @@ static void tcp_dsack_extend(struct sock *sk, u32 seq, u32 end_seq)
4165static void tcp_send_dupack(struct sock *sk, struct sk_buff *skb) 4160static void tcp_send_dupack(struct sock *sk, struct sk_buff *skb)
4166{ 4161{
4167 struct tcp_sock *tp = tcp_sk(sk); 4162 struct tcp_sock *tp = tcp_sk(sk);
4168 struct dst_entry *dst = __sk_dst_get(sk);
4169 4163
4170 if (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq && 4164 if (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq &&
4171 before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) { 4165 before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) {
4172 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOST); 4166 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOST);
4173 tcp_enter_quickack_mode(sk); 4167 tcp_enter_quickack_mode(sk);
4174 4168
4175 if (tcp_is_sack(tp) && sysctl_tcp_dsack && 4169 if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
4176 !dst_feature(dst, RTAX_FEATURE_NO_DSACK)) {
4177 u32 end_seq = TCP_SKB_CB(skb)->end_seq; 4170 u32 end_seq = TCP_SKB_CB(skb)->end_seq;
4178 4171
4179 if (after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) 4172 if (after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt))
@@ -5428,11 +5421,10 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
5428 u8 *hash_location; 5421 u8 *hash_location;
5429 struct inet_connection_sock *icsk = inet_csk(sk); 5422 struct inet_connection_sock *icsk = inet_csk(sk);
5430 struct tcp_sock *tp = tcp_sk(sk); 5423 struct tcp_sock *tp = tcp_sk(sk);
5431 struct dst_entry *dst = __sk_dst_get(sk);
5432 struct tcp_cookie_values *cvp = tp->cookie_values; 5424 struct tcp_cookie_values *cvp = tp->cookie_values;
5433 int saved_clamp = tp->rx_opt.mss_clamp; 5425 int saved_clamp = tp->rx_opt.mss_clamp;
5434 5426
5435 tcp_parse_options(skb, &tp->rx_opt, &hash_location, 0, dst); 5427 tcp_parse_options(skb, &tp->rx_opt, &hash_location, 0);
5436 5428
5437 if (th->ack) { 5429 if (th->ack) {
5438 /* rfc793: 5430 /* rfc793:
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 15e96030ce47..c3588b4fd979 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -742,9 +742,9 @@ static void tcp_v4_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
742 * This still operates on a request_sock only, not on a big 742 * This still operates on a request_sock only, not on a big
743 * socket. 743 * socket.
744 */ 744 */
745static int __tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst, 745static int tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst,
746 struct request_sock *req, 746 struct request_sock *req,
747 struct request_values *rvp) 747 struct request_values *rvp)
748{ 748{
749 const struct inet_request_sock *ireq = inet_rsk(req); 749 const struct inet_request_sock *ireq = inet_rsk(req);
750 int err = -1; 750 int err = -1;
@@ -775,10 +775,11 @@ static int __tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst,
775 return err; 775 return err;
776} 776}
777 777
778static int tcp_v4_send_synack(struct sock *sk, struct request_sock *req, 778static int tcp_v4_rtx_synack(struct sock *sk, struct request_sock *req,
779 struct request_values *rvp) 779 struct request_values *rvp)
780{ 780{
781 return __tcp_v4_send_synack(sk, NULL, req, rvp); 781 TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_RETRANSSEGS);
782 return tcp_v4_send_synack(sk, NULL, req, rvp);
782} 783}
783 784
784/* 785/*
@@ -1192,10 +1193,11 @@ static int tcp_v4_inbound_md5_hash(struct sock *sk, struct sk_buff *skb)
1192struct request_sock_ops tcp_request_sock_ops __read_mostly = { 1193struct request_sock_ops tcp_request_sock_ops __read_mostly = {
1193 .family = PF_INET, 1194 .family = PF_INET,
1194 .obj_size = sizeof(struct tcp_request_sock), 1195 .obj_size = sizeof(struct tcp_request_sock),
1195 .rtx_syn_ack = tcp_v4_send_synack, 1196 .rtx_syn_ack = tcp_v4_rtx_synack,
1196 .send_ack = tcp_v4_reqsk_send_ack, 1197 .send_ack = tcp_v4_reqsk_send_ack,
1197 .destructor = tcp_v4_reqsk_destructor, 1198 .destructor = tcp_v4_reqsk_destructor,
1198 .send_reset = tcp_v4_send_reset, 1199 .send_reset = tcp_v4_send_reset,
1200 .syn_ack_timeout = tcp_syn_ack_timeout,
1199}; 1201};
1200 1202
1201#ifdef CONFIG_TCP_MD5SIG 1203#ifdef CONFIG_TCP_MD5SIG
@@ -1262,20 +1264,10 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
1262 tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops; 1264 tcp_rsk(req)->af_specific = &tcp_request_sock_ipv4_ops;
1263#endif 1265#endif
1264 1266
1265 ireq = inet_rsk(req);
1266 ireq->loc_addr = daddr;
1267 ireq->rmt_addr = saddr;
1268 ireq->no_srccheck = inet_sk(sk)->transparent;
1269 ireq->opt = tcp_v4_save_options(sk, skb);
1270
1271 dst = inet_csk_route_req(sk, req);
1272 if(!dst)
1273 goto drop_and_free;
1274
1275 tcp_clear_options(&tmp_opt); 1267 tcp_clear_options(&tmp_opt);
1276 tmp_opt.mss_clamp = TCP_MSS_DEFAULT; 1268 tmp_opt.mss_clamp = TCP_MSS_DEFAULT;
1277 tmp_opt.user_mss = tp->rx_opt.user_mss; 1269 tmp_opt.user_mss = tp->rx_opt.user_mss;
1278 tcp_parse_options(skb, &tmp_opt, &hash_location, 0, dst); 1270 tcp_parse_options(skb, &tmp_opt, &hash_location, 0);
1279 1271
1280 if (tmp_opt.cookie_plus > 0 && 1272 if (tmp_opt.cookie_plus > 0 &&
1281 tmp_opt.saw_tstamp && 1273 tmp_opt.saw_tstamp &&
@@ -1319,8 +1311,14 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
1319 tmp_opt.tstamp_ok = tmp_opt.saw_tstamp; 1311 tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
1320 tcp_openreq_init(req, &tmp_opt, skb); 1312 tcp_openreq_init(req, &tmp_opt, skb);
1321 1313
1314 ireq = inet_rsk(req);
1315 ireq->loc_addr = daddr;
1316 ireq->rmt_addr = saddr;
1317 ireq->no_srccheck = inet_sk(sk)->transparent;
1318 ireq->opt = tcp_v4_save_options(sk, skb);
1319
1322 if (security_inet_conn_request(sk, skb, req)) 1320 if (security_inet_conn_request(sk, skb, req))
1323 goto drop_and_release; 1321 goto drop_and_free;
1324 1322
1325 if (!want_cookie) 1323 if (!want_cookie)
1326 TCP_ECN_create_request(req, tcp_hdr(skb)); 1324 TCP_ECN_create_request(req, tcp_hdr(skb));
@@ -1345,6 +1343,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
1345 */ 1343 */
1346 if (tmp_opt.saw_tstamp && 1344 if (tmp_opt.saw_tstamp &&
1347 tcp_death_row.sysctl_tw_recycle && 1345 tcp_death_row.sysctl_tw_recycle &&
1346 (dst = inet_csk_route_req(sk, req)) != NULL &&
1348 (peer = rt_get_peer((struct rtable *)dst)) != NULL && 1347 (peer = rt_get_peer((struct rtable *)dst)) != NULL &&
1349 peer->v4daddr == saddr) { 1348 peer->v4daddr == saddr) {
1350 if ((u32)get_seconds() - peer->tcp_ts_stamp < TCP_PAWS_MSL && 1349 if ((u32)get_seconds() - peer->tcp_ts_stamp < TCP_PAWS_MSL &&
@@ -1376,8 +1375,8 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
1376 } 1375 }
1377 tcp_rsk(req)->snt_isn = isn; 1376 tcp_rsk(req)->snt_isn = isn;
1378 1377
1379 if (__tcp_v4_send_synack(sk, dst, req, 1378 if (tcp_v4_send_synack(sk, dst, req,
1380 (struct request_values *)&tmp_ext) || 1379 (struct request_values *)&tmp_ext) ||
1381 want_cookie) 1380 want_cookie)
1382 goto drop_and_free; 1381 goto drop_and_free;
1383 1382
@@ -1652,6 +1651,9 @@ int tcp_v4_rcv(struct sk_buff *skb)
1652 if (!sk) 1651 if (!sk)
1653 goto no_tcp_socket; 1652 goto no_tcp_socket;
1654 1653
1654 if (iph->ttl < inet_sk(sk)->min_ttl)
1655 goto discard_and_relse;
1656
1655process: 1657process:
1656 if (sk->sk_state == TCP_TIME_WAIT) 1658 if (sk->sk_state == TCP_TIME_WAIT)
1657 goto do_time_wait; 1659 goto do_time_wait;
@@ -2428,12 +2430,12 @@ static struct tcp_seq_afinfo tcp4_seq_afinfo = {
2428 }, 2430 },
2429}; 2431};
2430 2432
2431static int tcp4_proc_init_net(struct net *net) 2433static int __net_init tcp4_proc_init_net(struct net *net)
2432{ 2434{
2433 return tcp_proc_register(net, &tcp4_seq_afinfo); 2435 return tcp_proc_register(net, &tcp4_seq_afinfo);
2434} 2436}
2435 2437
2436static void tcp4_proc_exit_net(struct net *net) 2438static void __net_exit tcp4_proc_exit_net(struct net *net)
2437{ 2439{
2438 tcp_proc_unregister(net, &tcp4_seq_afinfo); 2440 tcp_proc_unregister(net, &tcp4_seq_afinfo);
2439} 2441}
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 87accec8d097..f206ee5dda80 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -95,9 +95,9 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
95 struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw); 95 struct tcp_timewait_sock *tcptw = tcp_twsk((struct sock *)tw);
96 int paws_reject = 0; 96 int paws_reject = 0;
97 97
98 tmp_opt.saw_tstamp = 0;
98 if (th->doff > (sizeof(*th) >> 2) && tcptw->tw_ts_recent_stamp) { 99 if (th->doff > (sizeof(*th) >> 2) && tcptw->tw_ts_recent_stamp) {
99 tmp_opt.tstamp_ok = 1; 100 tcp_parse_options(skb, &tmp_opt, &hash_location, 0);
100 tcp_parse_options(skb, &tmp_opt, &hash_location, 1, NULL);
101 101
102 if (tmp_opt.saw_tstamp) { 102 if (tmp_opt.saw_tstamp) {
103 tmp_opt.ts_recent = tcptw->tw_ts_recent; 103 tmp_opt.ts_recent = tcptw->tw_ts_recent;
@@ -526,9 +526,9 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
526 __be32 flg = tcp_flag_word(th) & (TCP_FLAG_RST|TCP_FLAG_SYN|TCP_FLAG_ACK); 526 __be32 flg = tcp_flag_word(th) & (TCP_FLAG_RST|TCP_FLAG_SYN|TCP_FLAG_ACK);
527 int paws_reject = 0; 527 int paws_reject = 0;
528 528
529 if ((th->doff > (sizeof(*th) >> 2)) && (req->ts_recent)) { 529 tmp_opt.saw_tstamp = 0;
530 tmp_opt.tstamp_ok = 1; 530 if (th->doff > (sizeof(struct tcphdr)>>2)) {
531 tcp_parse_options(skb, &tmp_opt, &hash_location, 1, NULL); 531 tcp_parse_options(skb, &tmp_opt, &hash_location, 0);
532 532
533 if (tmp_opt.saw_tstamp) { 533 if (tmp_opt.saw_tstamp) {
534 tmp_opt.ts_recent = req->ts_recent; 534 tmp_opt.ts_recent = req->ts_recent;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 93316a96d820..4a1605d3f909 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -183,7 +183,8 @@ static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
183 */ 183 */
184void tcp_select_initial_window(int __space, __u32 mss, 184void tcp_select_initial_window(int __space, __u32 mss,
185 __u32 *rcv_wnd, __u32 *window_clamp, 185 __u32 *rcv_wnd, __u32 *window_clamp,
186 int wscale_ok, __u8 *rcv_wscale) 186 int wscale_ok, __u8 *rcv_wscale,
187 __u32 init_rcv_wnd)
187{ 188{
188 unsigned int space = (__space < 0 ? 0 : __space); 189 unsigned int space = (__space < 0 ? 0 : __space);
189 190
@@ -232,7 +233,13 @@ void tcp_select_initial_window(int __space, __u32 mss,
232 init_cwnd = 2; 233 init_cwnd = 2;
233 else if (mss > 1460) 234 else if (mss > 1460)
234 init_cwnd = 3; 235 init_cwnd = 3;
235 if (*rcv_wnd > init_cwnd * mss) 236 /* when initializing use the value from init_rcv_wnd
237 * rather than the default from above
238 */
239 if (init_rcv_wnd &&
240 (*rcv_wnd > init_rcv_wnd * mss))
241 *rcv_wnd = init_rcv_wnd * mss;
242 else if (*rcv_wnd > init_cwnd * mss)
236 *rcv_wnd = init_cwnd * mss; 243 *rcv_wnd = init_cwnd * mss;
237 } 244 }
238 245
@@ -553,7 +560,6 @@ static unsigned tcp_syn_options(struct sock *sk, struct sk_buff *skb,
553 struct tcp_md5sig_key **md5) { 560 struct tcp_md5sig_key **md5) {
554 struct tcp_sock *tp = tcp_sk(sk); 561 struct tcp_sock *tp = tcp_sk(sk);
555 struct tcp_cookie_values *cvp = tp->cookie_values; 562 struct tcp_cookie_values *cvp = tp->cookie_values;
556 struct dst_entry *dst = __sk_dst_get(sk);
557 unsigned remaining = MAX_TCP_OPTION_SPACE; 563 unsigned remaining = MAX_TCP_OPTION_SPACE;
558 u8 cookie_size = (!tp->rx_opt.cookie_out_never && cvp != NULL) ? 564 u8 cookie_size = (!tp->rx_opt.cookie_out_never && cvp != NULL) ?
559 tcp_cookie_size_check(cvp->cookie_desired) : 565 tcp_cookie_size_check(cvp->cookie_desired) :
@@ -581,22 +587,18 @@ static unsigned tcp_syn_options(struct sock *sk, struct sk_buff *skb,
581 opts->mss = tcp_advertise_mss(sk); 587 opts->mss = tcp_advertise_mss(sk);
582 remaining -= TCPOLEN_MSS_ALIGNED; 588 remaining -= TCPOLEN_MSS_ALIGNED;
583 589
584 if (likely(sysctl_tcp_timestamps && 590 if (likely(sysctl_tcp_timestamps && *md5 == NULL)) {
585 !dst_feature(dst, RTAX_FEATURE_NO_TSTAMP) &&
586 *md5 == NULL)) {
587 opts->options |= OPTION_TS; 591 opts->options |= OPTION_TS;
588 opts->tsval = TCP_SKB_CB(skb)->when; 592 opts->tsval = TCP_SKB_CB(skb)->when;
589 opts->tsecr = tp->rx_opt.ts_recent; 593 opts->tsecr = tp->rx_opt.ts_recent;
590 remaining -= TCPOLEN_TSTAMP_ALIGNED; 594 remaining -= TCPOLEN_TSTAMP_ALIGNED;
591 } 595 }
592 if (likely(sysctl_tcp_window_scaling && 596 if (likely(sysctl_tcp_window_scaling)) {
593 !dst_feature(dst, RTAX_FEATURE_NO_WSCALE))) {
594 opts->ws = tp->rx_opt.rcv_wscale; 597 opts->ws = tp->rx_opt.rcv_wscale;
595 opts->options |= OPTION_WSCALE; 598 opts->options |= OPTION_WSCALE;
596 remaining -= TCPOLEN_WSCALE_ALIGNED; 599 remaining -= TCPOLEN_WSCALE_ALIGNED;
597 } 600 }
598 if (likely(sysctl_tcp_sack && 601 if (likely(sysctl_tcp_sack)) {
599 !dst_feature(dst, RTAX_FEATURE_NO_SACK))) {
600 opts->options |= OPTION_SACK_ADVERTISE; 602 opts->options |= OPTION_SACK_ADVERTISE;
601 if (unlikely(!(OPTION_TS & opts->options))) 603 if (unlikely(!(OPTION_TS & opts->options)))
602 remaining -= TCPOLEN_SACKPERM_ALIGNED; 604 remaining -= TCPOLEN_SACKPERM_ALIGNED;
@@ -1799,11 +1801,6 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
1799void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss, 1801void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss,
1800 int nonagle) 1802 int nonagle)
1801{ 1803{
1802 struct sk_buff *skb = tcp_send_head(sk);
1803
1804 if (!skb)
1805 return;
1806
1807 /* If we are closed, the bytes will have to remain here. 1804 /* If we are closed, the bytes will have to remain here.
1808 * In time closedown will finish, we empty the write queue and 1805 * In time closedown will finish, we empty the write queue and
1809 * all will be happy. 1806 * all will be happy.
@@ -2427,7 +2424,8 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
2427 &req->rcv_wnd, 2424 &req->rcv_wnd,
2428 &req->window_clamp, 2425 &req->window_clamp,
2429 ireq->wscale_ok, 2426 ireq->wscale_ok,
2430 &rcv_wscale); 2427 &rcv_wscale,
2428 dst_metric(dst, RTAX_INITRWND));
2431 ireq->rcv_wscale = rcv_wscale; 2429 ireq->rcv_wscale = rcv_wscale;
2432 } 2430 }
2433 2431
@@ -2527,9 +2525,7 @@ static void tcp_connect_init(struct sock *sk)
2527 * See tcp_input.c:tcp_rcv_state_process case TCP_SYN_SENT. 2525 * See tcp_input.c:tcp_rcv_state_process case TCP_SYN_SENT.
2528 */ 2526 */
2529 tp->tcp_header_len = sizeof(struct tcphdr) + 2527 tp->tcp_header_len = sizeof(struct tcphdr) +
2530 (sysctl_tcp_timestamps && 2528 (sysctl_tcp_timestamps ? TCPOLEN_TSTAMP_ALIGNED : 0);
2531 (!dst_feature(dst, RTAX_FEATURE_NO_TSTAMP) ?
2532 TCPOLEN_TSTAMP_ALIGNED : 0));
2533 2529
2534#ifdef CONFIG_TCP_MD5SIG 2530#ifdef CONFIG_TCP_MD5SIG
2535 if (tp->af_specific->md5_lookup(sk, sk) != NULL) 2531 if (tp->af_specific->md5_lookup(sk, sk) != NULL)
@@ -2555,9 +2551,9 @@ static void tcp_connect_init(struct sock *sk)
2555 tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0), 2551 tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0),
2556 &tp->rcv_wnd, 2552 &tp->rcv_wnd,
2557 &tp->window_clamp, 2553 &tp->window_clamp,
2558 (sysctl_tcp_window_scaling && 2554 sysctl_tcp_window_scaling,
2559 !dst_feature(dst, RTAX_FEATURE_NO_WSCALE)), 2555 &rcv_wscale,
2560 &rcv_wscale); 2556 dst_metric(dst, RTAX_INITRWND));
2561 2557
2562 tp->rx_opt.rcv_wscale = rcv_wscale; 2558 tp->rx_opt.rcv_wscale = rcv_wscale;
2563 tp->rcv_ssthresh = tp->rcv_wnd; 2559 tp->rcv_ssthresh = tp->rcv_wnd;
diff --git a/net/ipv4/tcp_probe.c b/net/ipv4/tcp_probe.c
index bb110c5ce1d2..9bc805df95d2 100644
--- a/net/ipv4/tcp_probe.c
+++ b/net/ipv4/tcp_probe.c
@@ -39,9 +39,9 @@ static int port __read_mostly = 0;
39MODULE_PARM_DESC(port, "Port to match (0=all)"); 39MODULE_PARM_DESC(port, "Port to match (0=all)");
40module_param(port, int, 0); 40module_param(port, int, 0);
41 41
42static int bufsize __read_mostly = 4096; 42static unsigned int bufsize __read_mostly = 4096;
43MODULE_PARM_DESC(bufsize, "Log buffer size in packets (4096)"); 43MODULE_PARM_DESC(bufsize, "Log buffer size in packets (4096)");
44module_param(bufsize, int, 0); 44module_param(bufsize, uint, 0);
45 45
46static int full __read_mostly; 46static int full __read_mostly;
47MODULE_PARM_DESC(full, "Full log (1=every ack packet received, 0=only cwnd changes)"); 47MODULE_PARM_DESC(full, "Full log (1=every ack packet received, 0=only cwnd changes)");
@@ -75,12 +75,12 @@ static struct {
75 75
76static inline int tcp_probe_used(void) 76static inline int tcp_probe_used(void)
77{ 77{
78 return (tcp_probe.head - tcp_probe.tail) % bufsize; 78 return (tcp_probe.head - tcp_probe.tail) & (bufsize - 1);
79} 79}
80 80
81static inline int tcp_probe_avail(void) 81static inline int tcp_probe_avail(void)
82{ 82{
83 return bufsize - tcp_probe_used(); 83 return bufsize - tcp_probe_used() - 1;
84} 84}
85 85
86/* 86/*
@@ -116,7 +116,7 @@ static int jtcp_rcv_established(struct sock *sk, struct sk_buff *skb,
116 p->ssthresh = tcp_current_ssthresh(sk); 116 p->ssthresh = tcp_current_ssthresh(sk);
117 p->srtt = tp->srtt >> 3; 117 p->srtt = tp->srtt >> 3;
118 118
119 tcp_probe.head = (tcp_probe.head + 1) % bufsize; 119 tcp_probe.head = (tcp_probe.head + 1) & (bufsize - 1);
120 } 120 }
121 tcp_probe.lastcwnd = tp->snd_cwnd; 121 tcp_probe.lastcwnd = tp->snd_cwnd;
122 spin_unlock(&tcp_probe.lock); 122 spin_unlock(&tcp_probe.lock);
@@ -149,7 +149,7 @@ static int tcpprobe_open(struct inode * inode, struct file * file)
149static int tcpprobe_sprint(char *tbuf, int n) 149static int tcpprobe_sprint(char *tbuf, int n)
150{ 150{
151 const struct tcp_log *p 151 const struct tcp_log *p
152 = tcp_probe.log + tcp_probe.tail % bufsize; 152 = tcp_probe.log + tcp_probe.tail;
153 struct timespec tv 153 struct timespec tv
154 = ktime_to_timespec(ktime_sub(p->tstamp, tcp_probe.start)); 154 = ktime_to_timespec(ktime_sub(p->tstamp, tcp_probe.start));
155 155
@@ -192,7 +192,7 @@ static ssize_t tcpprobe_read(struct file *file, char __user *buf,
192 width = tcpprobe_sprint(tbuf, sizeof(tbuf)); 192 width = tcpprobe_sprint(tbuf, sizeof(tbuf));
193 193
194 if (cnt + width < len) 194 if (cnt + width < len)
195 tcp_probe.tail = (tcp_probe.tail + 1) % bufsize; 195 tcp_probe.tail = (tcp_probe.tail + 1) & (bufsize - 1);
196 196
197 spin_unlock_bh(&tcp_probe.lock); 197 spin_unlock_bh(&tcp_probe.lock);
198 198
@@ -222,9 +222,10 @@ static __init int tcpprobe_init(void)
222 init_waitqueue_head(&tcp_probe.wait); 222 init_waitqueue_head(&tcp_probe.wait);
223 spin_lock_init(&tcp_probe.lock); 223 spin_lock_init(&tcp_probe.lock);
224 224
225 if (bufsize < 0) 225 if (bufsize == 0)
226 return -EINVAL; 226 return -EINVAL;
227 227
228 bufsize = roundup_pow_of_two(bufsize);
228 tcp_probe.log = kcalloc(bufsize, sizeof(struct tcp_log), GFP_KERNEL); 229 tcp_probe.log = kcalloc(bufsize, sizeof(struct tcp_log), GFP_KERNEL);
229 if (!tcp_probe.log) 230 if (!tcp_probe.log)
230 goto err0; 231 goto err0;
@@ -236,7 +237,7 @@ static __init int tcpprobe_init(void)
236 if (ret) 237 if (ret)
237 goto err1; 238 goto err1;
238 239
239 pr_info("TCP probe registered (port=%d)\n", port); 240 pr_info("TCP probe registered (port=%d) bufsize=%u\n", port, bufsize);
240 return 0; 241 return 0;
241 err1: 242 err1:
242 proc_net_remove(&init_net, procname); 243 proc_net_remove(&init_net, procname);
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 8816a20c2597..de7d1bf9114f 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -474,6 +474,12 @@ static void tcp_synack_timer(struct sock *sk)
474 TCP_TIMEOUT_INIT, TCP_RTO_MAX); 474 TCP_TIMEOUT_INIT, TCP_RTO_MAX);
475} 475}
476 476
477void tcp_syn_ack_timeout(struct sock *sk, struct request_sock *req)
478{
479 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPTIMEOUTS);
480}
481EXPORT_SYMBOL(tcp_syn_ack_timeout);
482
477void tcp_set_keepalive(struct sock *sk, int val) 483void tcp_set_keepalive(struct sock *sk, int val)
478{ 484{
479 if ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)) 485 if ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN))
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 1f9534846ca9..608a5446d05b 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -216,9 +216,8 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum,
216 * force rand to be an odd multiple of UDP_HTABLE_SIZE 216 * force rand to be an odd multiple of UDP_HTABLE_SIZE
217 */ 217 */
218 rand = (rand | 1) * (udptable->mask + 1); 218 rand = (rand | 1) * (udptable->mask + 1);
219 for (last = first + udptable->mask + 1; 219 last = first + udptable->mask + 1;
220 first != last; 220 do {
221 first++) {
222 hslot = udp_hashslot(udptable, net, first); 221 hslot = udp_hashslot(udptable, net, first);
223 bitmap_zero(bitmap, PORTS_PER_CHAIN); 222 bitmap_zero(bitmap, PORTS_PER_CHAIN);
224 spin_lock_bh(&hslot->lock); 223 spin_lock_bh(&hslot->lock);
@@ -238,7 +237,7 @@ int udp_lib_get_port(struct sock *sk, unsigned short snum,
238 snum += rand; 237 snum += rand;
239 } while (snum != first); 238 } while (snum != first);
240 spin_unlock_bh(&hslot->lock); 239 spin_unlock_bh(&hslot->lock);
241 } 240 } while (++first != last);
242 goto fail; 241 goto fail;
243 } else { 242 } else {
244 hslot = udp_hashslot(udptable, net, snum); 243 hslot = udp_hashslot(udptable, net, snum);
@@ -1118,7 +1117,7 @@ int udp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
1118 struct inet_sock *inet = inet_sk(sk); 1117 struct inet_sock *inet = inet_sk(sk);
1119 struct sockaddr_in *sin = (struct sockaddr_in *)msg->msg_name; 1118 struct sockaddr_in *sin = (struct sockaddr_in *)msg->msg_name;
1120 struct sk_buff *skb; 1119 struct sk_buff *skb;
1121 unsigned int ulen, copied; 1120 unsigned int ulen;
1122 int peeked; 1121 int peeked;
1123 int err; 1122 int err;
1124 int is_udplite = IS_UDPLITE(sk); 1123 int is_udplite = IS_UDPLITE(sk);
@@ -1139,10 +1138,9 @@ try_again:
1139 goto out; 1138 goto out;
1140 1139
1141 ulen = skb->len - sizeof(struct udphdr); 1140 ulen = skb->len - sizeof(struct udphdr);
1142 copied = len; 1141 if (len > ulen)
1143 if (copied > ulen) 1142 len = ulen;
1144 copied = ulen; 1143 else if (len < ulen)
1145 else if (copied < ulen)
1146 msg->msg_flags |= MSG_TRUNC; 1144 msg->msg_flags |= MSG_TRUNC;
1147 1145
1148 /* 1146 /*
@@ -1151,14 +1149,14 @@ try_again:
1151 * coverage checksum (UDP-Lite), do it before the copy. 1149 * coverage checksum (UDP-Lite), do it before the copy.
1152 */ 1150 */
1153 1151
1154 if (copied < ulen || UDP_SKB_CB(skb)->partial_cov) { 1152 if (len < ulen || UDP_SKB_CB(skb)->partial_cov) {
1155 if (udp_lib_checksum_complete(skb)) 1153 if (udp_lib_checksum_complete(skb))
1156 goto csum_copy_err; 1154 goto csum_copy_err;
1157 } 1155 }
1158 1156
1159 if (skb_csum_unnecessary(skb)) 1157 if (skb_csum_unnecessary(skb))
1160 err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr), 1158 err = skb_copy_datagram_iovec(skb, sizeof(struct udphdr),
1161 msg->msg_iov, copied); 1159 msg->msg_iov, len);
1162 else { 1160 else {
1163 err = skb_copy_and_csum_datagram_iovec(skb, 1161 err = skb_copy_and_csum_datagram_iovec(skb,
1164 sizeof(struct udphdr), 1162 sizeof(struct udphdr),
@@ -1187,7 +1185,7 @@ try_again:
1187 if (inet->cmsg_flags) 1185 if (inet->cmsg_flags)
1188 ip_cmsg_recv(msg, skb); 1186 ip_cmsg_recv(msg, skb);
1189 1187
1190 err = copied; 1188 err = len;
1191 if (flags & MSG_TRUNC) 1189 if (flags & MSG_TRUNC)
1192 err = ulen; 1190 err = ulen;
1193 1191
@@ -2028,12 +2026,12 @@ static struct udp_seq_afinfo udp4_seq_afinfo = {
2028 }, 2026 },
2029}; 2027};
2030 2028
2031static int udp4_proc_init_net(struct net *net) 2029static int __net_init udp4_proc_init_net(struct net *net)
2032{ 2030{
2033 return udp_proc_register(net, &udp4_seq_afinfo); 2031 return udp_proc_register(net, &udp4_seq_afinfo);
2034} 2032}
2035 2033
2036static void udp4_proc_exit_net(struct net *net) 2034static void __net_exit udp4_proc_exit_net(struct net *net)
2037{ 2035{
2038 udp_proc_unregister(net, &udp4_seq_afinfo); 2036 udp_proc_unregister(net, &udp4_seq_afinfo);
2039} 2037}
diff --git a/net/ipv4/udplite.c b/net/ipv4/udplite.c
index 66f79513f4a5..6610bf76369f 100644
--- a/net/ipv4/udplite.c
+++ b/net/ipv4/udplite.c
@@ -81,12 +81,12 @@ static struct udp_seq_afinfo udplite4_seq_afinfo = {
81 }, 81 },
82}; 82};
83 83
84static int udplite4_proc_init_net(struct net *net) 84static int __net_init udplite4_proc_init_net(struct net *net)
85{ 85{
86 return udp_proc_register(net, &udplite4_seq_afinfo); 86 return udp_proc_register(net, &udplite4_seq_afinfo);
87} 87}
88 88
89static void udplite4_proc_exit_net(struct net *net) 89static void __net_exit udplite4_proc_exit_net(struct net *net)
90{ 90{
91 udp_proc_unregister(net, &udplite4_seq_afinfo); 91 udp_proc_unregister(net, &udplite4_seq_afinfo);
92} 92}
diff --git a/net/ipv4/xfrm4_policy.c b/net/ipv4/xfrm4_policy.c
index 8c08a28d8f83..67107d63c1cd 100644
--- a/net/ipv4/xfrm4_policy.c
+++ b/net/ipv4/xfrm4_policy.c
@@ -15,7 +15,6 @@
15#include <net/xfrm.h> 15#include <net/xfrm.h>
16#include <net/ip.h> 16#include <net/ip.h>
17 17
18static struct dst_ops xfrm4_dst_ops;
19static struct xfrm_policy_afinfo xfrm4_policy_afinfo; 18static struct xfrm_policy_afinfo xfrm4_policy_afinfo;
20 19
21static struct dst_entry *xfrm4_dst_lookup(struct net *net, int tos, 20static struct dst_entry *xfrm4_dst_lookup(struct net *net, int tos,
@@ -190,8 +189,10 @@ _decode_session4(struct sk_buff *skb, struct flowi *fl, int reverse)
190 189
191static inline int xfrm4_garbage_collect(struct dst_ops *ops) 190static inline int xfrm4_garbage_collect(struct dst_ops *ops)
192{ 191{
193 xfrm4_policy_afinfo.garbage_collect(&init_net); 192 struct net *net = container_of(ops, struct net, xfrm.xfrm4_dst_ops);
194 return (atomic_read(&xfrm4_dst_ops.entries) > xfrm4_dst_ops.gc_thresh*2); 193
194 xfrm4_policy_afinfo.garbage_collect(net);
195 return (atomic_read(&ops->entries) > ops->gc_thresh * 2);
195} 196}
196 197
197static void xfrm4_update_pmtu(struct dst_entry *dst, u32 mtu) 198static void xfrm4_update_pmtu(struct dst_entry *dst, u32 mtu)
@@ -268,7 +269,7 @@ static struct xfrm_policy_afinfo xfrm4_policy_afinfo = {
268static struct ctl_table xfrm4_policy_table[] = { 269static struct ctl_table xfrm4_policy_table[] = {
269 { 270 {
270 .procname = "xfrm4_gc_thresh", 271 .procname = "xfrm4_gc_thresh",
271 .data = &xfrm4_dst_ops.gc_thresh, 272 .data = &init_net.xfrm.xfrm4_dst_ops.gc_thresh,
272 .maxlen = sizeof(int), 273 .maxlen = sizeof(int),
273 .mode = 0644, 274 .mode = 0644,
274 .proc_handler = proc_dointvec, 275 .proc_handler = proc_dointvec,
@@ -295,8 +296,6 @@ static void __exit xfrm4_policy_fini(void)
295 296
296void __init xfrm4_init(int rt_max_size) 297void __init xfrm4_init(int rt_max_size)
297{ 298{
298 xfrm4_state_init();
299 xfrm4_policy_init();
300 /* 299 /*
301 * Select a default value for the gc_thresh based on the main route 300 * Select a default value for the gc_thresh based on the main route
302 * table hash size. It seems to me the worst case scenario is when 301 * table hash size. It seems to me the worst case scenario is when
@@ -308,6 +307,9 @@ void __init xfrm4_init(int rt_max_size)
308 * and start cleaning when were 1/2 full 307 * and start cleaning when were 1/2 full
309 */ 308 */
310 xfrm4_dst_ops.gc_thresh = rt_max_size/2; 309 xfrm4_dst_ops.gc_thresh = rt_max_size/2;
310
311 xfrm4_state_init();
312 xfrm4_policy_init();
311#ifdef CONFIG_SYSCTL 313#ifdef CONFIG_SYSCTL
312 sysctl_hdr = register_net_sysctl_table(&init_net, net_ipv4_ctl_path, 314 sysctl_hdr = register_net_sysctl_table(&init_net, net_ipv4_ctl_path,
313 xfrm4_policy_table); 315 xfrm4_policy_table);
generated by cgit v1.2.2 (git 2.25.0) at 2026-01-05 23:01:04 -0500