summaryrefslogtreecommitdiffstats
path: root/net/ipv4
diff options
context:
space:
mode:
Diffstat (limited to 'net/ipv4')
-rw-r--r--net/ipv4/Kconfig154
-rw-r--r--net/ipv4/Makefile14
-rw-r--r--net/ipv4/af_inet.c23
-rw-r--r--net/ipv4/ah4.c2
-rw-r--r--net/ipv4/devinet.c11
-rw-r--r--net/ipv4/esp4.c2
-rw-r--r--net/ipv4/fib_frontend.c55
-rw-r--r--net/ipv4/fib_hash.c3
-rw-r--r--net/ipv4/fib_lookup.h3
-rw-r--r--net/ipv4/fib_rules.c7
-rw-r--r--net/ipv4/fib_semantics.c10
-rw-r--r--net/ipv4/fib_trie.c2610
-rw-r--r--net/ipv4/icmp.c3
-rw-r--r--net/ipv4/igmp.c96
-rw-r--r--net/ipv4/ip_input.c11
-rw-r--r--net/ipv4/ip_output.c30
-rw-r--r--net/ipv4/ip_sockglue.c12
-rw-r--r--net/ipv4/ipcomp.c11
-rw-r--r--net/ipv4/ipconfig.c4
-rw-r--r--net/ipv4/ipmr.c11
-rw-r--r--net/ipv4/ipvs/Kconfig4
-rw-r--r--net/ipv4/ipvs/ip_vs_conn.c31
-rw-r--r--net/ipv4/ipvs/ip_vs_ctl.c17
-rw-r--r--net/ipv4/ipvs/ip_vs_sync.c4
-rw-r--r--net/ipv4/ipvs/ip_vs_xmit.c1
-rw-r--r--net/ipv4/netfilter/arp_tables.c1
-rw-r--r--net/ipv4/netfilter/ip_conntrack_amanda.c7
-rw-r--r--net/ipv4/netfilter/ip_conntrack_core.c107
-rw-r--r--net/ipv4/netfilter/ip_conntrack_ftp.c7
-rw-r--r--net/ipv4/netfilter/ip_conntrack_irc.c7
-rw-r--r--net/ipv4/netfilter/ip_conntrack_proto_sctp.c23
-rw-r--r--net/ipv4/netfilter/ip_conntrack_proto_tcp.c27
-rw-r--r--net/ipv4/netfilter/ip_conntrack_proto_udp.c1
-rw-r--r--net/ipv4/netfilter/ip_conntrack_standalone.c29
-rw-r--r--net/ipv4/netfilter/ip_nat_core.c32
-rw-r--r--net/ipv4/netfilter/ip_nat_helper.c13
-rw-r--r--net/ipv4/netfilter/ip_nat_rule.c4
-rw-r--r--net/ipv4/netfilter/ip_nat_standalone.c5
-rw-r--r--net/ipv4/netfilter/ip_tables.c1
-rw-r--r--net/ipv4/netfilter/ipt_CLUSTERIP.c58
-rw-r--r--net/ipv4/netfilter/ipt_MASQUERADE.c10
-rw-r--r--net/ipv4/netfilter/ipt_REJECT.c13
-rw-r--r--net/ipv4/netfilter/ipt_ULOG.c15
-rw-r--r--net/ipv4/netfilter/ipt_hashlimit.c17
-rw-r--r--net/ipv4/netfilter/ipt_helper.c4
-rw-r--r--net/ipv4/raw.c22
-rw-r--r--net/ipv4/route.c156
-rw-r--r--net/ipv4/syncookies.c49
-rw-r--r--net/ipv4/sysctl_net_ipv4.c114
-rw-r--r--net/ipv4/tcp.c169
-rw-r--r--net/ipv4/tcp_bic.c331
-rw-r--r--net/ipv4/tcp_cong.c237
-rw-r--r--net/ipv4/tcp_diag.c71
-rw-r--r--net/ipv4/tcp_highspeed.c181
-rw-r--r--net/ipv4/tcp_htcp.c289
-rw-r--r--net/ipv4/tcp_hybla.c187
-rw-r--r--net/ipv4/tcp_input.c824
-rw-r--r--net/ipv4/tcp_ipv4.c177
-rw-r--r--net/ipv4/tcp_minisocks.c72
-rw-r--r--net/ipv4/tcp_output.c596
-rw-r--r--net/ipv4/tcp_scalable.c68
-rw-r--r--net/ipv4/tcp_timer.c23
-rw-r--r--net/ipv4/tcp_vegas.c411
-rw-r--r--net/ipv4/tcp_westwood.c259
-rw-r--r--net/ipv4/xfrm4_output.c8
-rw-r--r--net/ipv4/xfrm4_state.c9
-rw-r--r--net/ipv4/xfrm4_tunnel.c2
67 files changed, 6106 insertions, 1659 deletions
diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig
index 6d3e8b1bd1f..df5386885a9 100644
--- a/net/ipv4/Kconfig
+++ b/net/ipv4/Kconfig
@@ -3,7 +3,6 @@
#
config IP_MULTICAST
bool "IP: multicasting"
- depends on INET
help
This is code for addressing several networked computers at once,
enlarging your kernel by about 2 KB. You need multicasting if you
@@ -17,7 +16,6 @@ config IP_MULTICAST
config IP_ADVANCED_ROUTER
bool "IP: advanced router"
- depends on INET
---help---
If you intend to run your Linux box mostly as a router, i.e. as a
computer that forwards and redistributes network packets, say Y; you
@@ -53,6 +51,44 @@ config IP_ADVANCED_ROUTER
If unsure, say N here.
+choice
+ prompt "Choose IP: FIB lookup algorithm (choose FIB_HASH if unsure)"
+ depends on IP_ADVANCED_ROUTER
+ default IP_FIB_HASH
+
+config IP_FIB_HASH
+ bool "FIB_HASH"
+ ---help---
+ Current FIB is very proven and good enough for most users.
+
+config IP_FIB_TRIE
+ bool "FIB_TRIE"
+ ---help---
+ Use new experimental LC-trie as FIB lookup algoritm.
+ This improves lookup performance if you have a large
+ number of routes.
+
+ LC-trie is a longest matching prefix lookup algorithm which
+ performs better than FIB_HASH for large routing tables.
+ But, it consumes more memory and is more complex.
+
+ LC-trie is described in:
+
+ IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson
+ IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999
+ An experimental study of compression methods for dynamic tries
+ Stefan Nilsson and Matti Tikkanen. Algorithmica, 33(1):19-33, 2002.
+ http://www.nada.kth.se/~snilsson/public/papers/dyntrie2/
+
+endchoice
+
+# If the user does not enable advanced routing, he gets the safe
+# default of the fib-hash algorithm.
+config IP_FIB_HASH
+ bool
+ depends on !IP_ADVANCED_ROUTER
+ default y
+
config IP_MULTIPLE_TABLES
bool "IP: policy routing"
depends on IP_ADVANCED_ROUTER
@@ -145,7 +181,6 @@ config IP_ROUTE_VERBOSE
config IP_PNP
bool "IP: kernel level autoconfiguration"
- depends on INET
help
This enables automatic configuration of IP addresses of devices and
of the routing table during kernel boot, based on either information
@@ -204,7 +239,6 @@ config IP_PNP_RARP
# bool ' IP: ARP support' CONFIG_IP_PNP_ARP
config NET_IPIP
tristate "IP: tunneling"
- depends on INET
select INET_TUNNEL
---help---
Tunneling means encapsulating data of one protocol type within
@@ -222,7 +256,6 @@ config NET_IPIP
config NET_IPGRE
tristate "IP: GRE tunnels over IP"
- depends on INET
select XFRM
help
Tunneling means encapsulating data of one protocol type within
@@ -281,7 +314,7 @@ config IP_PIMSM_V2
config ARPD
bool "IP: ARP daemon support (EXPERIMENTAL)"
- depends on INET && EXPERIMENTAL
+ depends on EXPERIMENTAL
---help---
Normally, the kernel maintains an internal cache which maps IP
addresses to hardware addresses on the local network, so that
@@ -306,7 +339,6 @@ config ARPD
config SYN_COOKIES
bool "IP: TCP syncookie support (disabled per default)"
- depends on INET
---help---
Normal TCP/IP networking is open to an attack known as "SYN
flooding". This denial-of-service attack prevents legitimate remote
@@ -343,7 +375,6 @@ config SYN_COOKIES
config INET_AH
tristate "IP: AH transformation"
- depends on INET
select XFRM
select CRYPTO
select CRYPTO_HMAC
@@ -356,7 +387,6 @@ config INET_AH
config INET_ESP
tristate "IP: ESP transformation"
- depends on INET
select XFRM
select CRYPTO
select CRYPTO_HMAC
@@ -370,7 +400,6 @@ config INET_ESP
config INET_IPCOMP
tristate "IP: IPComp transformation"
- depends on INET
select XFRM
select INET_TUNNEL
select CRYPTO
@@ -383,7 +412,6 @@ config INET_IPCOMP
config INET_TUNNEL
tristate "IP: tunnel transformation"
- depends on INET
select XFRM
---help---
Support for generic IP tunnel transformation, which is required by
@@ -393,7 +421,6 @@ config INET_TUNNEL
config IP_TCPDIAG
tristate "IP: TCP socket monitoring interface"
- depends on INET
default y
---help---
Support for TCP socket monitoring interface used by native Linux
@@ -407,5 +434,108 @@ config IP_TCPDIAG
config IP_TCPDIAG_IPV6
def_bool (IP_TCPDIAG=y && IPV6=y) || (IP_TCPDIAG=m && IPV6)
+config TCP_CONG_ADVANCED
+ bool "TCP: advanced congestion control"
+ ---help---
+ Support for selection of various TCP congestion control
+ modules.
+
+ Nearly all users can safely say no here, and a safe default
+ selection will be made (BIC-TCP with new Reno as a fallback).
+
+ If unsure, say N.
+
+# TCP Reno is builtin (required as fallback)
+menu "TCP congestion control"
+ depends on TCP_CONG_ADVANCED
+
+config TCP_CONG_BIC
+ tristate "Binary Increase Congestion (BIC) control"
+ default y
+ ---help---
+ BIC-TCP is a sender-side only change that ensures a linear RTT
+ fairness under large windows while offering both scalability and
+ bounded TCP-friendliness. The protocol combines two schemes
+ called additive increase and binary search increase. When the
+ congestion window is large, additive increase with a large
+ increment ensures linear RTT fairness as well as good
+ scalability. Under small congestion windows, binary search
+ increase provides TCP friendliness.
+ See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/
+
+config TCP_CONG_WESTWOOD
+ tristate "TCP Westwood+"
+ default m
+ ---help---
+ TCP Westwood+ is a sender-side only modification of the TCP Reno
+ protocol stack that optimizes the performance of TCP congestion
+ control. It is based on end-to-end bandwidth estimation to set
+ congestion window and slow start threshold after a congestion
+ episode. Using this estimation, TCP Westwood+ adaptively sets a
+ slow start threshold and a congestion window which takes into
+ account the bandwidth used at the time congestion is experienced.
+ TCP Westwood+ significantly increases fairness wrt TCP Reno in
+ wired networks and throughput over wireless links.
+
+config TCP_CONG_HTCP
+ tristate "H-TCP"
+ default m
+ ---help---
+ H-TCP is a send-side only modifications of the TCP Reno
+ protocol stack that optimizes the performance of TCP
+ congestion control for high speed network links. It uses a
+ modeswitch to change the alpha and beta parameters of TCP Reno
+ based on network conditions and in a way so as to be fair with
+ other Reno and H-TCP flows.
+
+config TCP_CONG_HSTCP
+ tristate "High Speed TCP"
+ depends on EXPERIMENTAL
+ default n
+ ---help---
+ Sally Floyd's High Speed TCP (RFC 3649) congestion control.
+ A modification to TCP's congestion control mechanism for use
+ with large congestion windows. A table indicates how much to
+ increase the congestion window by when an ACK is received.
+ For more detail see http://www.icir.org/floyd/hstcp.html
+
+config TCP_CONG_HYBLA
+ tristate "TCP-Hybla congestion control algorithm"
+ depends on EXPERIMENTAL
+ default n
+ ---help---
+ TCP-Hybla is a sender-side only change that eliminates penalization of
+ long-RTT, large-bandwidth connections, like when satellite legs are
+ involved, expecially when sharing a common bottleneck with normal
+ terrestrial connections.
+
+config TCP_CONG_VEGAS
+ tristate "TCP Vegas"
+ depends on EXPERIMENTAL
+ default n
+ ---help---
+ TCP Vegas is a sender-side only change to TCP that anticipates
+ the onset of congestion by estimating the bandwidth. TCP Vegas
+ adjusts the sending rate by modifying the congestion
+ window. TCP Vegas should provide less packet loss, but it is
+ not as aggressive as TCP Reno.
+
+config TCP_CONG_SCALABLE
+ tristate "Scalable TCP"
+ depends on EXPERIMENTAL
+ default n
+ ---help---
+ Scalable TCP is a sender-side only change to TCP which uses a
+ MIMD congestion control algorithm which has some nice scaling
+ properties, though is known to have fairness issues.
+ See http://www-lce.eng.cam.ac.uk/~ctk21/scalable/
+
+endmenu
+
+config TCP_CONG_BIC
+ tristate
+ depends on !TCP_CONG_ADVANCED
+ default y
+
source "net/ipv4/ipvs/Kconfig"
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index 8b379627ebb..5718cdb3a61 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -5,10 +5,13 @@
obj-y := utils.o route.o inetpeer.o protocol.o \
ip_input.o ip_fragment.o ip_forward.o ip_options.o \
ip_output.o ip_sockglue.o \
- tcp.o tcp_input.o tcp_output.o tcp_timer.o tcp_ipv4.o tcp_minisocks.o \
+ tcp.o tcp_input.o tcp_output.o tcp_timer.o tcp_ipv4.o \
+ tcp_minisocks.o tcp_cong.o \
datagram.o raw.o udp.o arp.o icmp.o devinet.o af_inet.o igmp.o \
- sysctl_net_ipv4.o fib_frontend.o fib_semantics.o fib_hash.o
+ sysctl_net_ipv4.o fib_frontend.o fib_semantics.o
+obj-$(CONFIG_IP_FIB_HASH) += fib_hash.o
+obj-$(CONFIG_IP_FIB_TRIE) += fib_trie.o
obj-$(CONFIG_PROC_FS) += proc.o
obj-$(CONFIG_IP_MULTIPLE_TABLES) += fib_rules.o
obj-$(CONFIG_IP_MROUTE) += ipmr.o
@@ -28,6 +31,13 @@ obj-$(CONFIG_NETFILTER) += netfilter/
obj-$(CONFIG_IP_VS) += ipvs/
obj-$(CONFIG_IP_TCPDIAG) += tcp_diag.o
obj-$(CONFIG_IP_ROUTE_MULTIPATH_CACHED) += multipath.o
+obj-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o
+obj-$(CONFIG_TCP_CONG_WESTWOOD) += tcp_westwood.o
+obj-$(CONFIG_TCP_CONG_HSTCP) += tcp_highspeed.o
+obj-$(CONFIG_TCP_CONG_HYBLA) += tcp_hybla.o
+obj-$(CONFIG_TCP_CONG_HTCP) += tcp_htcp.o
+obj-$(CONFIG_TCP_CONG_VEGAS) += tcp_vegas.o
+obj-$(CONFIG_TCP_CONG_SCALABLE) += tcp_scalable.o
obj-$(CONFIG_XFRM) += xfrm4_policy.o xfrm4_state.o xfrm4_input.o \
xfrm4_output.o
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 03942f13394..ef7468376ae 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1009,6 +1009,15 @@ static int __init init_ipv4_mibs(void)
static int ipv4_proc_init(void);
extern void ipfrag_init(void);
+/*
+ * IP protocol layer initialiser
+ */
+
+static struct packet_type ip_packet_type = {
+ .type = __constant_htons(ETH_P_IP),
+ .func = ip_rcv,
+};
+
static int __init inet_init(void)
{
struct sk_buff *dummy_skb;
@@ -1102,6 +1111,8 @@ static int __init inet_init(void)
ipfrag_init();
+ dev_add_pack(&ip_packet_type);
+
rc = 0;
out:
return rc;
@@ -1119,6 +1130,10 @@ module_init(inet_init);
#ifdef CONFIG_PROC_FS
extern int fib_proc_init(void);
extern void fib_proc_exit(void);
+#ifdef CONFIG_IP_FIB_TRIE
+extern int fib_stat_proc_init(void);
+extern void fib_stat_proc_exit(void);
+#endif
extern int ip_misc_proc_init(void);
extern int raw_proc_init(void);
extern void raw_proc_exit(void);
@@ -1139,11 +1154,19 @@ static int __init ipv4_proc_init(void)
goto out_udp;
if (fib_proc_init())
goto out_fib;
+#ifdef CONFIG_IP_FIB_TRIE
+ if (fib_stat_proc_init())
+ goto out_fib_stat;
+ #endif
if (ip_misc_proc_init())
goto out_misc;
out:
return rc;
out_misc:
+#ifdef CONFIG_IP_FIB_TRIE
+ fib_stat_proc_exit();
+out_fib_stat:
+#endif
fib_proc_exit();
out_fib:
udp4_proc_exit();
diff --git a/net/ipv4/ah4.c b/net/ipv4/ah4.c
index 0e98f2235b6..514c85b2631 100644
--- a/net/ipv4/ah4.c
+++ b/net/ipv4/ah4.c
@@ -200,7 +200,7 @@ static void ah4_err(struct sk_buff *skb, u32 info)
xfrm_state_put(x);
}
-static int ah_init_state(struct xfrm_state *x, void *args)
+static int ah_init_state(struct xfrm_state *x)
{
struct ah_data *ahp = NULL;
struct xfrm_algo_desc *aalg_desc;
diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index 478a30179a5..d8a10e3dd77 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -1030,14 +1030,13 @@ static struct notifier_block ip_netdev_notifier = {
};
static int inet_fill_ifaddr(struct sk_buff *skb, struct in_ifaddr *ifa,
- u32 pid, u32 seq, int event)
+ u32 pid, u32 seq, int event, unsigned int flags)
{
struct ifaddrmsg *ifm;
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
- nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*ifm));
- if (pid) nlh->nlmsg_flags |= NLM_F_MULTI;
+ nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*ifm), flags);
ifm = NLMSG_DATA(nlh);
ifm->ifa_family = AF_INET;
ifm->ifa_prefixlen = ifa->ifa_prefixlen;
@@ -1090,7 +1089,7 @@ static int inet_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
continue;
if (inet_fill_ifaddr(skb, ifa, NETLINK_CB(cb->skb).pid,
cb->nlh->nlmsg_seq,
- RTM_NEWADDR) <= 0) {
+ RTM_NEWADDR, NLM_F_MULTI) <= 0) {
rcu_read_unlock();
goto done;
}
@@ -1113,7 +1112,7 @@ static void rtmsg_ifa(int event, struct in_ifaddr* ifa)
if (!skb)
netlink_set_err(rtnl, 0, RTMGRP_IPV4_IFADDR, ENOBUFS);
- else if (inet_fill_ifaddr(skb, ifa, 0, 0, event) < 0) {
+ else if (inet_fill_ifaddr(skb, ifa, current->pid, 0, event, 0) < 0) {
kfree_skb(skb);
netlink_set_err(rtnl, 0, RTMGRP_IPV4_IFADDR, EINVAL);
} else {
@@ -1472,7 +1471,7 @@ static void devinet_sysctl_register(struct in_device *in_dev,
* by sysctl and we wouldn't want anyone to change it under our feet
* (see SIOCSIFNAME).
*/
- dev_name = net_sysctl_strdup(dev_name);
+ dev_name = kstrdup(dev_name, GFP_KERNEL);
if (!dev_name)
goto free;
diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c
index eae84cc39d3..ba57446d5d1 100644
--- a/net/ipv4/esp4.c
+++ b/net/ipv4/esp4.c
@@ -362,7 +362,7 @@ static void esp_destroy(struct xfrm_state *x)
kfree(esp);
}
-static int esp_init_state(struct xfrm_state *x, void *args)
+static int esp_init_state(struct xfrm_state *x)
{
struct esp_data *esp = NULL;
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index 563e7d61270..cd8e45ab958 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -516,6 +516,60 @@ static void fib_del_ifaddr(struct in_ifaddr *ifa)
#undef BRD1_OK
}
+static void nl_fib_lookup(struct fib_result_nl *frn, struct fib_table *tb )
+{
+
+ struct fib_result res;
+ struct flowi fl = { .nl_u = { .ip4_u = { .daddr = frn->fl_addr,
+ .fwmark = frn->fl_fwmark,
+ .tos = frn->fl_tos,
+ .scope = frn->fl_scope } } };
+ if (tb) {
+ local_bh_disable();
+
+ frn->tb_id = tb->tb_id;
+ frn->err = tb->tb_lookup(tb, &fl, &res);
+
+ if (!frn->err) {
+ frn->prefixlen = res.prefixlen;
+ frn->nh_sel = res.nh_sel;
+ frn->type = res.type;
+ frn->scope = res.scope;
+ }
+ local_bh_enable();
+ }
+}
+
+static void nl_fib_input(struct sock *sk, int len)
+{
+ struct sk_buff *skb = NULL;
+ struct nlmsghdr *nlh = NULL;
+ struct fib_result_nl *frn;
+ int err;
+ u32 pid;
+ struct fib_table *tb;
+
+ skb = skb_recv_datagram(sk, 0, 0, &err);
+ nlh = (struct nlmsghdr *)skb->data;
+
+ frn = (struct fib_result_nl *) NLMSG_DATA(nlh);
+ tb = fib_get_table(frn->tb_id_in);
+
+ nl_fib_lookup(frn, tb);
+
+ pid = nlh->nlmsg_pid; /*pid of sending process */
+ NETLINK_CB(skb).groups = 0; /* not in mcast group */
+ NETLINK_CB(skb).pid = 0; /* from kernel */
+ NETLINK_CB(skb).dst_pid = pid;
+ NETLINK_CB(skb).dst_groups = 0; /* unicast */
+ netlink_unicast(sk, skb, pid, MSG_DONTWAIT);
+}
+
+static void nl_fib_lookup_init(void)
+{
+ netlink_kernel_create(NETLINK_FIB_LOOKUP, nl_fib_input);
+}
+
static void fib_disable_ip(struct net_device *dev, int force)
{
if (fib_sync_down(0, dev, force))
@@ -604,6 +658,7 @@ void __init ip_fib_init(void)
register_netdevice_notifier(&fib_netdev_notifier);
register_inetaddr_notifier(&fib_inetaddr_notifier);
+ nl_fib_lookup_init();
}
EXPORT_SYMBOL(inet_addr_type);
diff --git a/net/ipv4/fib_hash.c b/net/ipv4/fib_hash.c
index 6506dcc01b4..b10d6bb5ef3 100644
--- a/net/ipv4/fib_hash.c
+++ b/net/ipv4/fib_hash.c
@@ -703,7 +703,8 @@ fn_hash_dump_bucket(struct sk_buff *skb, struct netlink_callback *cb,
&f->fn_key,
fz->fz_order,
fa->fa_tos,
- fa->fa_info) < 0) {
+ fa->fa_info,
+ NLM_F_MULTI) < 0) {
cb->args[3] = i;
return -1;
}
diff --git a/net/ipv4/fib_lookup.h b/net/ipv4/fib_lookup.h
index ac4485f75e9..b729d97cfa9 100644
--- a/net/ipv4/fib_lookup.h
+++ b/net/ipv4/fib_lookup.h
@@ -30,7 +30,8 @@ extern int fib_nh_match(struct rtmsg *r, struct nlmsghdr *,
struct kern_rta *rta, struct fib_info *fi);
extern int fib_dump_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
u8 tb_id, u8 type, u8 scope, void *dst,
- int dst_len, u8 tos, struct fib_info *fi);
+ int dst_len, u8 tos, struct fib_info *fi,
+ unsigned int);
extern void rtmsg_fib(int event, u32 key, struct fib_alias *fa,
int z, int tb_id,
struct nlmsghdr *n, struct netlink_skb_parms *req);
diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c
index 39d0aadb9a2..0b298bbc151 100644
--- a/net/ipv4/fib_rules.c
+++ b/net/ipv4/fib_rules.c
@@ -367,13 +367,14 @@ static struct notifier_block fib_rules_notifier = {
static __inline__ int inet_fill_rule(struct sk_buff *skb,
struct fib_rule *r,
- struct netlink_callback *cb)
+ struct netlink_callback *cb,
+ unsigned int flags)
{
struct rtmsg *rtm;
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
- nlh = NLMSG_PUT(skb, NETLINK_CREDS(cb->skb)->pid, cb->nlh->nlmsg_seq, RTM_NEWRULE, sizeof(*rtm));
+ nlh = NLMSG_NEW_ANSWER(skb, cb, RTM_NEWRULE, sizeof(*rtm), flags);
rtm = NLMSG_DATA(nlh);
rtm->rtm_family = AF_INET;
rtm->rtm_dst_len = r->r_dst_len;
@@ -422,7 +423,7 @@ int inet_dump_rules(struct sk_buff *skb, struct netlink_callback *cb)
for (r=fib_rules, idx=0; r; r = r->r_next, idx++) {
if (idx < s_idx)
continue;
- if (inet_fill_rule(skb, r, cb) < 0)
+ if (inet_fill_rule(skb, r, cb, NLM_F_MULTI) < 0)
break;
}
read_unlock(&fib_rules_lock);
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
index 029362d6613..c886b28ba9f 100644
--- a/net/ipv4/fib_semantics.c
+++ b/net/ipv4/fib_semantics.c
@@ -276,7 +276,7 @@ void rtmsg_fib(int event, u32 key, struct fib_alias *fa,
struct nlmsghdr *n, struct netlink_skb_parms *req)
{
struct sk_buff *skb;
- u32 pid = req ? req->pid : 0;
+ u32 pid = req ? req->pid : n->nlmsg_pid;
int size = NLMSG_SPACE(sizeof(struct rtmsg)+256);
skb = alloc_skb(size, GFP_KERNEL);
@@ -286,7 +286,7 @@ void rtmsg_fib(int event, u32 key, struct fib_alias *fa,
if (fib_dump_info(skb, pid, n->nlmsg_seq, event, tb_id,
fa->fa_type, fa->fa_scope, &key, z,
fa->fa_tos,
- fa->fa_info) < 0) {
+ fa->fa_info, 0) < 0) {
kfree_skb(skb);
return;
}
@@ -932,13 +932,13 @@ u32 __fib_res_prefsrc(struct fib_result *res)
int
fib_dump_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
u8 tb_id, u8 type, u8 scope, void *dst, int dst_len, u8 tos,
- struct fib_info *fi)
+ struct fib_info *fi, unsigned int flags)
{
struct rtmsg *rtm;
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
- nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*rtm));
+ nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*rtm), flags);
rtm = NLMSG_DATA(nlh);
rtm->rtm_family = AF_INET;
rtm->rtm_dst_len = dst_len;
@@ -1035,7 +1035,7 @@ fib_convert_rtentry(int cmd, struct nlmsghdr *nl, struct rtmsg *rtm,
}
nl->nlmsg_flags = NLM_F_REQUEST;
- nl->nlmsg_pid = 0;
+ nl->nlmsg_pid = current->pid;
nl->nlmsg_seq = 0;
nl->nlmsg_len = NLMSG_LENGTH(sizeof(*rtm));
if (cmd == SIOCDELRT) {
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
new file mode 100644
index 00000000000..4be234c7d8c
--- /dev/null
+++ b/net/ipv4/fib_trie.c
@@ -0,0 +1,2610 @@
+/*
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ *
+ * Robert Olsson <robert.olsson@its.uu.se> Uppsala Universitet
+ * & Swedish University of Agricultural Sciences.
+ *
+ * Jens Laas <jens.laas@data.slu.se> Swedish University of
+ * Agricultural Sciences.
+ *
+ * Hans Liss <hans.liss@its.uu.se> Uppsala Universitet
+ *
+ * This work is based on the LPC-trie which is originally descibed in:
+ *
+ * An experimental study of compression methods for dynamic tries
+ * Stefan Nilsson and Matti Tikkanen. Algorithmica, 33(1):19-33, 2002.
+ * http://www.nada.kth.se/~snilsson/public/papers/dyntrie2/
+ *
+ *
+ * IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson
+ * IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999
+ *
+ * Version: $Id: fib_trie.c,v 1.3 2005/06/08 14:20:01 robert Exp $
+ *
+ *
+ * Code from fib_hash has been reused which includes the following header:
+ *
+ *
+ * INET An implementation of the TCP/IP protocol suite for the LINUX
+ * operating system. INET is implemented using the BSD Socket
+ * interface as the means of communication with the user level.
+ *
+ * IPv4 FIB: lookup engine and maintenance routines.
+ *
+ *
+ * Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ */
+
+#define VERSION "0.325"
+
+#include <linux/config.h>
+#include <asm/uaccess.h>
+#include <asm/system.h>
+#include <asm/bitops.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/errno.h>
+#include <linux/in.h>
+#include <linux/inet.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netlink.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <net/ip.h>
+#include <net/protocol.h>
+#include <net/route.h>
+#include <net/tcp.h>
+#include <net/sock.h>
+#include <net/ip_fib.h>
+#include "fib_lookup.h"
+
+#undef CONFIG_IP_FIB_TRIE_STATS
+#define MAX_CHILDS 16384
+
+#define EXTRACT(p, n, str) ((str)<<(p)>>(32-(n)))
+#define KEYLENGTH (8*sizeof(t_key))
+#define MASK_PFX(k, l) (((l)==0)?0:(k >> (KEYLENGTH-l)) << (KEYLENGTH-l))
+#define TKEY_GET_MASK(offset, bits) (((bits)==0)?0:((t_key)(-1) << (KEYLENGTH - bits) >> offset))
+
+static DEFINE_RWLOCK(fib_lock);
+
+typedef unsigned int t_key;
+
+#define T_TNODE 0
+#define T_LEAF 1
+#define NODE_TYPE_MASK 0x1UL
+#define NODE_PARENT(_node) \
+((struct tnode *)((_node)->_parent & ~NODE_TYPE_MASK))
+#define NODE_SET_PARENT(_node, _ptr) \
+((_node)->_parent = (((unsigned long)(_ptr)) | \
+ ((_node)->_parent & NODE_TYPE_MASK)))
+#define NODE_INIT_PARENT(_node, _type) \
+((_node)->_parent = (_type))
+#define NODE_TYPE(_node) \
+((_node)->_parent & NODE_TYPE_MASK)
+
+#define IS_TNODE(n) (!(n->_parent & T_LEAF))
+#define IS_LEAF(n) (n->_parent & T_LEAF)
+
+struct node {
+ t_key key;
+ unsigned long _parent;
+};
+
+struct leaf {
+ t_key key;
+ unsigned long _parent;
+ struct hlist_head list;
+};
+
+struct leaf_info {
+ struct hlist_node hlist;
+ int plen;
+ struct list_head falh;
+};
+
+struct tnode {
+ t_key key;
+ unsigned long _parent;
+ unsigned short pos:5; /* 2log(KEYLENGTH) bits needed */
+ unsigned short bits:5; /* 2log(KEYLENGTH) bits needed */
+ unsigned short full_children; /* KEYLENGTH bits needed */
+ unsigned short empty_children; /* KEYLENGTH bits needed */
+ struct node *child[0];
+};
+
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+struct trie_use_stats {
+ unsigned int gets;
+ unsigned int backtrack;
+ unsigned int semantic_match_passed;
+ unsigned int semantic_match_miss;
+ unsigned int null_node_hit;
+ unsigned int resize_node_skipped;
+};
+#endif
+
+struct trie_stat {
+ unsigned int totdepth;
+ unsigned int maxdepth;
+ unsigned int tnodes;
+ unsigned int leaves;
+ unsigned int nullpointers;
+ unsigned int nodesizes[MAX_CHILDS];
+};
+
+struct trie {
+ struct node *trie;
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ struct trie_use_stats stats;
+#endif
+ int size;
+ unsigned int revision;
+};
+
+static int trie_debug = 0;
+
+static int tnode_full(struct tnode *tn, struct node *n);
+static void put_child(struct trie *t, struct tnode *tn, int i, struct node *n);
+static void tnode_put_child_reorg(struct tnode *tn, int i, struct node *n, int wasfull);
+static int tnode_child_length(struct tnode *tn);
+static struct node *resize(struct trie *t, struct tnode *tn);
+static struct tnode *inflate(struct trie *t, struct tnode *tn, int *err);
+static struct tnode *halve(struct trie *t, struct tnode *tn, int *err);
+static void tnode_free(struct tnode *tn);
+static void trie_dump_seq(struct seq_file *seq, struct trie *t);
+extern struct fib_alias *fib_find_alias(struct list_head *fah, u8 tos, u32 prio);
+extern int fib_detect_death(struct fib_info *fi, int order,
+ struct fib_info **last_resort, int *last_idx, int *dflt);
+
+extern void rtmsg_fib(int event, u32 key, struct fib_alias *fa, int z, int tb_id,
+ struct nlmsghdr *n, struct netlink_skb_parms *req);
+
+static kmem_cache_t *fn_alias_kmem;
+static struct trie *trie_local = NULL, *trie_main = NULL;
+
+static void trie_bug(char *err)
+{
+ printk("Trie Bug: %s\n", err);
+ BUG();
+}
+
+static inline struct node *tnode_get_child(struct tnode *tn, int i)
+{
+ if (i >= 1<<tn->bits)
+ trie_bug("tnode_get_child");
+
+ return tn->child[i];
+}
+
+static inline int tnode_child_length(struct tnode *tn)
+{
+ return 1<<tn->bits;
+}
+
+/*
+ _________________________________________________________________
+ | i | i | i | i | i | i | i | N | N | N | S | S | S | S | S | C |
+ ----------------------------------------------------------------
+ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
+
+ _________________________________________________________________
+ | C | C | C | u | u | u | u | u | u | u | u | u | u | u | u | u |
+ -----------------------------------------------------------------
+ 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
+
+ tp->pos = 7
+ tp->bits = 3
+ n->pos = 15
+ n->bits=4
+ KEYLENGTH=32
+*/
+
+static inline t_key tkey_extract_bits(t_key a, int offset, int bits)
+{
+ if (offset < KEYLENGTH)
+ return ((t_key)(a << offset)) >> (KEYLENGTH - bits);
+ else
+ return 0;
+}
+
+static inline int tkey_equals(t_key a, t_key b)
+{
+ return a == b;
+}
+
+static inline int tkey_sub_equals(t_key a, int offset, int bits, t_key b)
+{
+ if (bits == 0 || offset >= KEYLENGTH)
+ return 1;
+ bits = bits > KEYLENGTH ? KEYLENGTH : bits;
+ return ((a ^ b) << offset) >> (KEYLENGTH - bits) == 0;
+}
+
+static inline int tkey_mismatch(t_key a, int offset, t_key b)
+{
+ t_key diff = a ^ b;
+ int i = offset;
+
+ if(!diff)
+ return 0;
+ while((diff << i) >> (KEYLENGTH-1) == 0)
+ i++;
+ return i;
+}
+
+/* Candiate for fib_semantics */
+
+static void fn_free_alias(struct fib_alias *fa)
+{
+ fib_release_info(fa->fa_info);
+ kmem_cache_free(fn_alias_kmem, fa);
+}
+
+/*
+ To understand this stuff, an understanding of keys and all their bits is
+ necessary. Every node in the trie has a key associated with it, but not
+ all of the bits in that key are significant.
+
+ Consider a node 'n' and its parent 'tp'.
+
+ If n is a leaf, every bit in its key is significant. Its presence is
+ necessitaded by path compression, since during a tree traversal (when
+ searching for a leaf - unless we are doing an insertion) we will completely
+ ignore all skipped bits we encounter. Thus we need to verify, at the end of
+ a potentially successful search, that we have indeed been walking the
+ correct key path.
+
+ Note that we can never "miss" the correct key in the tree if present by
+ following the wrong path. Path compression ensures that segments of the key
+ that are the same for all keys with a given prefix are skipped, but the
+ skipped part *is* identical for each node in the subtrie below the skipped
+ bit! trie_insert() in this implementation takes care of that - note the
+ call to tkey_sub_equals() in trie_insert().
+
+ if n is an internal node - a 'tnode' here, the various parts of its key
+ have many different meanings.
+
+ Example:
+ _________________________________________________________________
+ | i | i | i | i | i | i | i | N | N | N | S | S | S | S | S | C |
+ -----------------------------------------------------------------
+ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
+
+ _________________________________________________________________
+ | C | C | C | u | u | u | u | u | u | u | u | u | u | u | u | u |
+ -----------------------------------------------------------------
+ 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
+
+ tp->pos = 7
+ tp->bits = 3
+ n->pos = 15
+ n->bits=4
+
+ First, let's just ignore the bits that come before the parent tp, that is
+ the bits from 0 to (tp->pos-1). They are *known* but at this point we do
+ not use them for anything.
+
+ The bits from (tp->pos) to (tp->pos + tp->bits - 1) - "N", above - are the
+ index into the parent's child array. That is, they will be used to find
+ 'n' among tp's children.
+
+ The bits from (tp->pos + tp->bits) to (n->pos - 1) - "S" - are skipped bits
+ for the node n.
+
+ All the bits we have seen so far are significant to the node n. The rest
+ of the bits are really not needed or indeed known in n->key.
+
+ The bits from (n->pos) to (n->pos + n->bits - 1) - "C" - are the index into
+ n's child array, and will of course be different for each child.
+
+ The rest of the bits, from (n->pos + n->bits) onward, are completely unknown
+ at this point.
+
+*/
+
+static void check_tnode(struct tnode *tn)
+{
+ if(tn && tn->pos+tn->bits > 32) {
+ printk("TNODE ERROR tn=%p, pos=%d, bits=%d\n", tn, tn->pos, tn->bits);
+ }
+}
+
+static int halve_threshold = 25;
+static int inflate_threshold = 50;
+
+static struct leaf *leaf_new(void)
+{
+ struct leaf *l = kmalloc(sizeof(struct leaf), GFP_KERNEL);
+ if(l) {
+ NODE_INIT_PARENT(l, T_LEAF);
+ INIT_HLIST_HEAD(&l->list);
+ }
+ return l;
+}
+
+static struct leaf_info *leaf_info_new(int plen)
+{
+ struct leaf_info *li = kmalloc(sizeof(struct leaf_info), GFP_KERNEL);
+ if(li) {
+ li->plen = plen;
+ INIT_LIST_HEAD(&li->falh);
+ }
+ return li;
+}
+
+static inline void free_leaf(struct leaf *l)
+{
+ kfree(l);
+}
+
+static inline void free_leaf_info(struct leaf_info *li)
+{
+ kfree(li);
+}
+
+static struct tnode *tnode_alloc(unsigned int size)
+{
+ if (size <= PAGE_SIZE) {
+ return kmalloc(size, GFP_KERNEL);
+ } else {
+ return (struct tnode *)
+ __get_free_pages(GFP_KERNEL, get_order(size));
+ }
+}
+
+static void __tnode_free(struct tnode *tn)
+{
+ unsigned int size = sizeof(struct tnode) +
+ (1<<tn->bits) * sizeof(struct node *);
+
+ if (size <= PAGE_SIZE)
+ kfree(tn);
+ else
+ free_pages((unsigned long)tn, get_order(size));
+}
+
+static struct tnode* tnode_new(t_key key, int pos, int bits)
+{
+ int nchildren = 1<<bits;
+ int sz = sizeof(struct tnode) + nchildren * sizeof(struct node *);
+ struct tnode *tn = tnode_alloc(sz);
+
+ if(tn) {
+ memset(tn, 0, sz);
+ NODE_INIT_PARENT(tn, T_TNODE);
+ tn->pos = pos;
+ tn->bits = bits;
+ tn->key = key;
+ tn->full_children = 0;
+ tn->empty_children = 1<<bits;
+ }
+ if(trie_debug > 0)
+ printk("AT %p s=%u %u\n", tn, (unsigned int) sizeof(struct tnode),
+ (unsigned int) (sizeof(struct node) * 1<<bits));
+ return tn;
+}
+
+static void tnode_free(struct tnode *tn)
+{
+ if(!tn) {
+ trie_bug("tnode_free\n");
+ }
+ if(IS_LEAF(tn)) {
+ free_leaf((struct leaf *)tn);
+ if(trie_debug > 0 )
+ printk("FL %p \n", tn);
+ }
+ else if(IS_TNODE(tn)) {
+ __tnode_free(tn);
+ if(trie_debug > 0 )
+ printk("FT %p \n", tn);
+ }
+ else {
+ trie_bug("tnode_free\n");
+ }
+}
+
+/*
+ * Check whether a tnode 'n' is "full", i.e. it is an internal node
+ * and no bits are skipped. See discussion in dyntree paper p. 6
+ */
+
+static inline int tnode_full(struct tnode *tn, struct node *n)
+{
+ if(n == NULL || IS_LEAF(n))
+ return 0;
+
+ return ((struct tnode *) n)->pos == tn->pos + tn->bits;
+}
+
+static inline void put_child(struct trie *t, struct tnode *tn, int i, struct node *n)
+{
+ tnode_put_child_reorg(tn, i, n, -1);
+}
+
+ /*
+ * Add a child at position i overwriting the old value.
+ * Update the value of full_children and empty_children.
+ */
+
+static void tnode_put_child_reorg(struct tnode *tn, int i, struct node *n, int wasfull)
+{
+ struct node *chi;
+ int isfull;
+
+ if(i >= 1<<tn->bits) {
+ printk("bits=%d, i=%d\n", tn->bits, i);
+ trie_bug("tnode_put_child_reorg bits");
+ }
+ write_lock_bh(&fib_lock);
+ chi = tn->child[i];
+
+ /* update emptyChildren */
+ if (n == NULL && chi != NULL)
+ tn->empty_children++;
+ else if (n != NULL && chi == NULL)
+ tn->empty_children--;
+
+ /* update fullChildren */
+ if (wasfull == -1)
+ wasfull = tnode_full(tn, chi);
+
+ isfull = tnode_full(tn, n);
+ if (wasfull && !isfull)
+ tn->full_children--;
+
+ else if (!wasfull && isfull)
+ tn->full_children++;
+ if(n)
+ NODE_SET_PARENT(n, tn);
+
+ tn->child[i] = n;
+ write_unlock_bh(&fib_lock);
+}
+
+static struct node *resize(struct trie *t, struct tnode *tn)
+{
+ int i;
+ int err = 0;
+
+ if (!tn)
+ return NULL;
+
+ if(trie_debug)
+ printk("In tnode_resize %p inflate_threshold=%d threshold=%d\n",
+ tn, inflate_threshold, halve_threshold);
+
+ /* No children */
+ if (tn->empty_children == tnode_child_length(tn)) {
+ tnode_free(tn);
+ return NULL;
+ }
+ /* One child */
+ if (tn->empty_children == tnode_child_length(tn) - 1)
+ for (i = 0; i < tnode_child_length(tn); i++) {
+
+ write_lock_bh(&fib_lock);
+ if (tn->child[i] != NULL) {
+
+ /* compress one level */
+ struct node *n = tn->child[i];
+ if(n)
+ NODE_INIT_PARENT(n, NODE_TYPE(n));
+
+ write_unlock_bh(&fib_lock);
+ tnode_free(tn);
+ return n;
+ }
+ write_unlock_bh(&fib_lock);
+ }
+ /*
+ * Double as long as the resulting node has a number of
+ * nonempty nodes that are above the threshold.
+ */
+
+ /*
+ * From "Implementing a dynamic compressed trie" by Stefan Nilsson of
+ * the Helsinki University of Technology and Matti Tikkanen of Nokia
+ * Telecommunications, page 6:
+ * "A node is doubled if the ratio of non-empty children to all
+ * children in the *doubled* node is at least 'high'."
+ *
+ * 'high' in this instance is the variable 'inflate_threshold'. It
+ * is expressed as a percentage, so we multiply it with
+ * tnode_child_length() and instead of multiplying by 2 (since the
+ * child array will be doubled by inflate()) and multiplying
+ * the left-hand side by 100 (to handle the percentage thing) we
+ * multiply the left-hand side by 50.
+ *
+ * The left-hand side may look a bit weird: tnode_child_length(tn)
+ * - tn->empty_children is of course the number of non-null children
+ * in the current node. tn->full_children is the number of "full"
+ * children, that is non-null tnodes with a skip value of 0.
+ * All of those will be doubled in the resulting inflated tnode, so
+ * we just count them one extra time here.
+ *
+ * A clearer way to write this would be:
+ *
+ * to_be_doubled = tn->full_children;
+ * not_to_be_doubled = tnode_child_length(tn) - tn->empty_children -
+ * tn->full_children;
+ *
+ * new_child_length = tnode_child_length(tn) * 2;
+ *
+ * new_fill_factor = 100 * (not_to_be_doubled + 2*to_be_doubled) /
+ * new_child_length;
+ * if (new_fill_factor >= inflate_threshold)
+ *
+ * ...and so on, tho it would mess up the while() loop.
+ *
+ * anyway,
+ * 100 * (not_to_be_doubled + 2*to_be_doubled) / new_child_length >=
+ * inflate_threshold
+ *
+ * avoid a division:
+ * 100 * (not_to_be_doubled + 2*to_be_doubled) >=
+ * inflate_threshold * new_child_length
+ *
+ * expand not_to_be_doubled and to_be_doubled, and shorten:
+ * 100 * (tnode_child_length(tn) - tn->empty_children +
+ * tn->full_children ) >= inflate_threshold * new_child_length
+ *
+ * expand new_child_length:
+ * 100 * (tnode_child_length(tn) - tn->empty_children +
+ * tn->full_children ) >=
+ * inflate_threshold * tnode_child_length(tn) * 2
+ *
+ * shorten again:
+ * 50 * (tn->full_children + tnode_child_length(tn) -
+ * tn->empty_children ) >= inflate_threshold *
+ * tnode_child_length(tn)
+ *
+ */
+
+ check_tnode(tn);
+
+ err = 0;
+ while ((tn->full_children > 0 &&
+ 50 * (tn->full_children + tnode_child_length(tn) - tn->empty_children) >=
+ inflate_threshold * tnode_child_length(tn))) {
+
+ tn = inflate(t, tn, &err);
+
+ if(err) {
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.resize_node_skipped++;
+#endif
+ break;
+ }
+ }
+
+ check_tnode(tn);
+
+ /*
+ * Halve as long as the number of empty children in this
+ * node is above threshold.
+ */
+
+ err = 0;
+ while (tn->bits > 1 &&
+ 100 * (tnode_child_length(tn) - tn->empty_children) <
+ halve_threshold * tnode_child_length(tn)) {
+
+ tn = halve(t, tn, &err);
+
+ if(err) {
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.resize_node_skipped++;
+#endif
+ break;
+ }
+ }
+
+
+ /* Only one child remains */
+
+ if (tn->empty_children == tnode_child_length(tn) - 1)
+ for (i = 0; i < tnode_child_length(tn); i++) {
+
+ write_lock_bh(&fib_lock);
+ if (tn->child[i] != NULL) {
+ /* compress one level */
+ struct node *n = tn->child[i];
+
+ if(n)
+ NODE_INIT_PARENT(n, NODE_TYPE(n));
+
+ write_unlock_bh(&fib_lock);
+ tnode_free(tn);
+ return n;
+ }
+ write_unlock_bh(&fib_lock);
+ }
+
+ return (struct node *) tn;
+}
+
+static struct tnode *inflate(struct trie *t, struct tnode *tn, int *err)
+{
+ struct tnode *inode;
+ struct tnode *oldtnode = tn;
+ int olen = tnode_child_length(tn);
+ int i;
+
+ if(trie_debug)
+ printk("In inflate\n");
+
+ tn = tnode_new(oldtnode->key, oldtnode->pos, oldtnode->bits + 1);
+
+ if (!tn) {
+ *err = -ENOMEM;
+ return oldtnode;
+ }
+
+ /*
+ * Preallocate and store tnodes before the actual work so we
+ * don't get into an inconsistent state if memory allocation
+ * fails. In case of failure we return the oldnode and inflate
+ * of tnode is ignored.
+ */
+
+ for(i = 0; i < olen; i++) {
+ struct tnode *inode = (struct tnode *) tnode_get_child(oldtnode, i);
+
+ if (inode &&
+ IS_TNODE(inode) &&
+ inode->pos == oldtnode->pos + oldtnode->bits &&
+ inode->bits > 1) {
+ struct tnode *left, *right;
+
+ t_key m = TKEY_GET_MASK(inode->pos, 1);
+
+ left = tnode_new(inode->key&(~m), inode->pos + 1,
+ inode->bits - 1);
+
+ if(!left) {
+ *err = -ENOMEM;
+ break;
+ }
+
+ right = tnode_new(inode->key|m, inode->pos + 1,
+ inode->bits - 1);
+
+ if(!right) {
+ *err = -ENOMEM;
+ break;
+ }
+
+ put_child(t, tn, 2*i, (struct node *) left);
+ put_child(t, tn, 2*i+1, (struct node *) right);
+ }
+ }
+
+ if(*err) {
+ int size = tnode_child_length(tn);
+ int j;
+
+ for(j = 0; j < size; j++)
+ if( tn->child[j])
+ tnode_free((struct tnode *)tn->child[j]);
+
+ tnode_free(tn);
+
+ *err = -ENOMEM;
+ return oldtnode;
+ }
+
+ for(i = 0; i < olen; i++) {
+ struct node *node = tnode_get_child(oldtnode, i);
+
+ /* An empty child */
+ if (node == NULL)
+ continue;
+
+ /* A leaf or an internal node with skipped bits */
+
+ if(IS_LEAF(node) || ((struct tnode *) node)->pos >
+ tn->pos + tn->bits - 1) {
+ if(tkey_extract_bits(node->key, oldtnode->pos + oldtnode->bits,
+ 1) == 0)
+ put_child(t, tn, 2*i, node);
+ else
+ put_child(t, tn, 2*i+1, node);
+ continue;
+ }
+
+ /* An internal node with two children */
+ inode = (struct tnode *) node;
+
+ if (inode->bits == 1) {
+ put_child(t, tn, 2*i, inode->child[0]);
+ put_child(t, tn, 2*i+1, inode->child[1]);
+
+ tnode_free(inode);
+ }
+
+ /* An internal node with more than two children */
+ else {
+ struct tnode *left, *right;
+ int size, j;
+
+ /* We will replace this node 'inode' with two new
+ * ones, 'left' and 'right', each with half of the
+ * original children. The two new nodes will have
+ * a position one bit further down the key and this
+ * means that the "significant" part of their keys
+ * (see the discussion near the top of this file)
+ * will differ by one bit, which will be "0" in
+ * left's key and "1" in right's key. Since we are
+ * moving the key position by one step, the bit that
+ * we are moving away from - the bit at position
+ * (inode->pos) - is the one that will differ between
+ * left and right. So... we synthesize that bit in the
+ * two new keys.
+ * The mask 'm' below will be a single "one" bit at
+ * the position (inode->pos)
+ */
+
+ /* Use the old key, but set the new significant
+ * bit to zero.
+ */
+
+ left = (struct tnode *) tnode_get_child(tn, 2*i);
+ put_child(t, tn, 2*i, NULL);
+
+ if(!left)
+ BUG();
+
+ right = (struct tnode *) tnode_get_child(tn, 2*i+1);
+ put_child(t, tn, 2*i+1, NULL);
+
+ if(!right)
+ BUG();
+
+ size = tnode_child_length(left);
+ for(j = 0; j < size; j++) {
+ put_child(t, left, j, inode->child[j]);
+ put_child(t, right, j, inode->child[j + size]);
+ }
+ put_child(t, tn, 2*i, resize(t, left));
+ put_child(t, tn, 2*i+1, resize(t, right));
+
+ tnode_free(inode);
+ }
+ }
+ tnode_free(oldtnode);
+ return tn;
+}
+
+static struct tnode *halve(struct trie *t, struct tnode *tn, int *err)
+{
+ struct tnode *oldtnode = tn;
+ struct node *left, *right;
+ int i;
+ int olen = tnode_child_length(tn);
+
+ if(trie_debug) printk("In halve\n");
+
+ tn=tnode_new(oldtnode->key, oldtnode->pos, oldtnode->bits - 1);
+
+ if (!tn) {
+ *err = -ENOMEM;
+ return oldtnode;
+ }
+
+ /*
+ * Preallocate and store tnodes before the actual work so we
+ * don't get into an inconsistent state if memory allocation
+ * fails. In case of failure we return the oldnode and halve
+ * of tnode is ignored.
+ */
+
+ for(i = 0; i < olen; i += 2) {
+ left = tnode_get_child(oldtnode, i);
+ right = tnode_get_child(oldtnode, i+1);
+
+ /* Two nonempty children */
+ if( left && right) {
+ struct tnode *newBinNode =
+ tnode_new(left->key, tn->pos + tn->bits, 1);
+
+ if(!newBinNode) {
+ *err = -ENOMEM;
+ break;
+ }
+ put_child(t, tn, i/2, (struct node *)newBinNode);
+ }
+ }
+
+ if(*err) {
+ int size = tnode_child_length(tn);
+ int j;
+
+ for(j = 0; j < size; j++)
+ if( tn->child[j])
+ tnode_free((struct tnode *)tn->child[j]);
+
+ tnode_free(tn);
+
+ *err = -ENOMEM;
+ return oldtnode;
+ }
+
+ for(i = 0; i < olen; i += 2) {
+ left = tnode_get_child(oldtnode, i);
+ right = tnode_get_child(oldtnode, i+1);
+
+ /* At least one of the children is empty */
+ if (left == NULL) {
+ if (right == NULL) /* Both are empty */
+ continue;
+ put_child(t, tn, i/2, right);
+ } else if (right == NULL)
+ put_child(t, tn, i/2, left);
+
+ /* Two nonempty children */
+ else {
+ struct tnode *newBinNode =
+ (struct tnode *) tnode_get_child(tn, i/2);
+ put_child(t, tn, i/2, NULL);
+
+ if(!newBinNode)
+ BUG();
+
+ put_child(t, newBinNode, 0, left);
+ put_child(t, newBinNode, 1, right);
+ put_child(t, tn, i/2, resize(t, newBinNode));
+ }
+ }
+ tnode_free(oldtnode);
+ return tn;
+}
+
+static void *trie_init(struct trie *t)
+{
+ if(t) {
+ t->size = 0;
+ t->trie = NULL;
+ t->revision = 0;
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ memset(&t->stats, 0, sizeof(struct trie_use_stats));
+#endif
+ }
+ return t;
+}
+
+static struct leaf_info *find_leaf_info(struct hlist_head *head, int plen)
+{
+ struct hlist_node *node;
+ struct leaf_info *li;
+
+ hlist_for_each_entry(li, node, head, hlist) {
+
+ if ( li->plen == plen )
+ return li;
+ }
+ return NULL;
+}
+
+static inline struct list_head * get_fa_head(struct leaf *l, int plen)
+{
+ struct list_head *fa_head=NULL;
+ struct leaf_info *li = find_leaf_info(&l->list, plen);
+
+ if(li)
+ fa_head = &li->falh;
+
+ return fa_head;
+}
+
+static void insert_leaf_info(struct hlist_head *head, struct leaf_info *new)
+{
+ struct leaf_info *li=NULL, *last=NULL;
+ struct hlist_node *node, *tmp;
+
+ write_lock_bh(&fib_lock);
+
+ if(hlist_empty(head))
+ hlist_add_head(&new->hlist, head);
+ else {
+ hlist_for_each_entry_safe(li, node, tmp, head, hlist) {
+
+ if (new->plen > li->plen)
+ break;
+
+ last = li;
+ }
+ if(last)
+ hlist_add_after(&last->hlist, &new->hlist);
+ else
+ hlist_add_before(&new->hlist, &li->hlist);
+ }
+ write_unlock_bh(&fib_lock);
+}
+
+static struct leaf *
+fib_find_node(struct trie *t, u32 key)
+{
+ int pos;
+ struct tnode *tn;
+ struct node *n;
+
+ pos = 0;
+ n=t->trie;
+
+ while (n != NULL && NODE_TYPE(n) == T_TNODE) {
+ tn = (struct tnode *) n;
+
+ check_tnode(tn);
+
+ if(tkey_sub_equals(tn->key, pos, tn->pos-pos, key)) {
+ pos=tn->pos + tn->bits;
+ n = tnode_get_child(tn, tkey_extract_bits(key, tn->pos, tn->bits));
+ }
+ else
+ break;
+ }
+ /* Case we have found a leaf. Compare prefixes */
+
+ if (n != NULL && IS_LEAF(n) && tkey_equals(key, n->key)) {
+ struct leaf *l = (struct leaf *) n;
+ return l;
+ }
+ return NULL;
+}
+
+static struct node *trie_rebalance(struct trie *t, struct tnode *tn)
+{
+ int i = 0;
+ int wasfull;
+ t_key cindex, key;
+ struct tnode *tp = NULL;
+
+ if(!tn)
+ BUG();
+
+ key = tn->key;
+ i = 0;
+
+ while (tn != NULL && NODE_PARENT(tn) != NULL) {
+
+ if( i > 10 ) {
+ printk("Rebalance tn=%p \n", tn);
+ if(tn) printk("tn->parent=%p \n", NODE_PARENT(tn));
+
+ printk("Rebalance tp=%p \n", tp);
+ if(tp) printk("tp->parent=%p \n", NODE_PARENT(tp));
+ }
+
+ if( i > 12 ) BUG();
+ i++;
+
+ tp = NODE_PARENT(tn);
+ cindex = tkey_extract_bits(key, tp->pos, tp->bits);
+ wasfull = tnode_full(tp, tnode_get_child(tp, cindex));
+ tn = (struct tnode *) resize (t, (struct tnode *)tn);
+ tnode_put_child_reorg((struct tnode *)tp, cindex,(struct node*)tn, wasfull);
+
+ if(!NODE_PARENT(tn))
+ break;
+
+ tn = NODE_PARENT(tn);
+ }
+ /* Handle last (top) tnode */
+ if (IS_TNODE(tn))
+ tn = (struct tnode*) resize(t, (struct tnode *)tn);
+
+ return (struct node*) tn;
+}
+
+static struct list_head *
+fib_insert_node(struct trie *t, int *err, u32 key, int plen)
+{
+ int pos, newpos;
+ struct tnode *tp = NULL, *tn = NULL;
+ struct node *n;
+ struct leaf *l;
+ int missbit;
+ struct list_head *fa_head=NULL;
+ struct leaf_info *li;
+ t_key cindex;
+
+ pos = 0;
+ n=t->trie;
+
+ /* If we point to NULL, stop. Either the tree is empty and we should
+ * just put a new leaf in if, or we have reached an empty child slot,
+ * and we should just put our new leaf in that.
+ * If we point to a T_TNODE, check if it matches our key. Note that
+ * a T_TNODE might be skipping any number of bits - its 'pos' need
+ * not be the parent's 'pos'+'bits'!
+ *
+ * If it does match the current key, get pos/bits from it, extract
+ * the index from our key, push the T_TNODE and walk the tree.
+ *
+ * If it doesn't, we have to replace it with a new T_TNODE.
+ *
+ * If we point to a T_LEAF, it might or might not have the same key
+ * as we do. If it does, just change the value, update the T_LEAF's
+ * value, and return it.
+ * If it doesn't, we need to replace it with a T_TNODE.
+ */
+
+ while (n != NULL && NODE_TYPE(n) == T_TNODE) {
+ tn = (struct tnode *) n;
+
+ check_tnode(tn);
+
+ if(tkey_sub_equals(tn->key, pos, tn->pos-pos, key)) {
+ tp = tn;
+ pos=tn->pos + tn->bits;
+ n = tnode_get_child(tn, tkey_extract_bits(key, tn->pos, tn->bits));
+
+ if(n && NODE_PARENT(n) != tn) {
+ printk("BUG tn=%p, n->parent=%p\n", tn, NODE_PARENT(n));
+ BUG();
+ }
+ }
+ else
+ break;
+ }
+
+ /*
+ * n ----> NULL, LEAF or TNODE
+ *
+ * tp is n's (parent) ----> NULL or TNODE
+ */
+
+ if(tp && IS_LEAF(tp))
+ BUG();
+
+
+ /* Case 1: n is a leaf. Compare prefixes */
+
+ if (n != NULL && IS_LEAF(n) && tkey_equals(key, n->key)) {
+ struct leaf *l = ( struct leaf *) n;
+
+ li = leaf_info_new(plen);
+
+ if(! li) {
+ *err = -ENOMEM;
+ goto err;
+ }
+
+ fa_head = &li->falh;
+ insert_leaf_info(&l->list, li);
+ goto done;
+ }
+ t->size++;
+ l = leaf_new();
+
+ if(! l) {
+ *err = -ENOMEM;
+ goto err;
+ }
+
+ l->key = key;
+ li = leaf_info_new(plen);
+
+ if(! li) {
+ tnode_free((struct tnode *) l);
+ *err = -ENOMEM;
+ goto err;
+ }
+
+ fa_head = &li->falh;
+ insert_leaf_info(&l->list, li);
+
+ /* Case 2: n is NULL, and will just insert a new leaf */
+ if (t->trie && n == NULL) {
+
+ NODE_SET_PARENT(l, tp);
+
+ if (!tp)
+ BUG();
+
+ else {
+ cindex = tkey_extract_bits(key, tp->pos, tp->bits);
+ put_child(t, (struct tnode *)tp, cindex, (struct node *)l);
+ }
+ }
+ /* Case 3: n is a LEAF or a TNODE and the key doesn't match. */
+ else {
+ /*
+ * Add a new tnode here
+ * first tnode need some special handling
+ */
+
+ if (tp)
+ pos=tp->pos+tp->bits;
+ else
+ pos=0;
+ if(n) {
+ newpos = tkey_mismatch(key, pos, n->key);
+ tn = tnode_new(n->key, newpos, 1);
+ }
+ else {
+ newpos = 0;
+ tn = tnode_new(key, newpos, 1); /* First tnode */
+ }
+
+ if(!tn) {
+ free_leaf_info(li);
+ tnode_free((struct tnode *) l);
+ *err = -ENOMEM;
+ goto err;
+ }
+
+ NODE_SET_PARENT(tn, tp);
+
+ missbit=tkey_extract_bits(key, newpos, 1);
+ put_child(t, tn, missbit, (struct node *)l);
+ put_child(t, tn, 1-missbit, n);
+
+ if(tp) {
+ cindex = tkey_extract_bits(key, tp->pos, tp->bits);
+ put_child(t, (struct tnode *)tp, cindex, (struct node *)tn);
+ }
+ else {
+ t->trie = (struct node*) tn; /* First tnode */
+ tp = tn;
+ }
+ }
+ if(tp && tp->pos+tp->bits > 32) {
+ printk("ERROR tp=%p pos=%d, bits=%d, key=%0x plen=%d\n",
+ tp, tp->pos, tp->bits, key, plen);
+ }
+ /* Rebalance the trie */
+ t->trie = trie_rebalance(t, tp);
+done:
+ t->revision++;
+err:;
+ return fa_head;
+}
+
+static int
+fn_trie_insert(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta,
+ struct nlmsghdr *nlhdr, struct netlink_skb_parms *req)
+{
+ struct trie *t = (struct trie *) tb->tb_data;
+ struct fib_alias *fa, *new_fa;
+ struct list_head *fa_head=NULL;
+ struct fib_info *fi;
+ int plen = r->rtm_dst_len;
+ int type = r->rtm_type;
+ u8 tos = r->rtm_tos;
+ u32 key, mask;
+ int err;
+ struct leaf *l;
+
+ if (plen > 32)
+ return -EINVAL;
+
+ key = 0;
+ if (rta->rta_dst)
+ memcpy(&key, rta->rta_dst, 4);
+
+ key = ntohl(key);
+
+ if(trie_debug)
+ printk("Insert table=%d %08x/%d\n", tb->tb_id, key, plen);
+
+ mask = ntohl( inet_make_mask(plen) );
+
+ if(key & ~mask)
+ return -EINVAL;
+
+ key = key & mask;
+
+ if ((fi = fib_create_info(r, rta, nlhdr, &err)) == NULL)
+ goto err;
+
+ l = fib_find_node(t, key);
+ fa = NULL;
+
+ if(l) {
+ fa_head = get_fa_head(l, plen);
+ fa = fib_find_alias(fa_head, tos, fi->fib_priority);
+ }
+
+ /* Now fa, if non-NULL, points to the first fib alias
+ * with the same keys [prefix,tos,priority], if such key already
+ * exists or to the node before which we will insert new one.
+ *
+ * If fa is NULL, we will need to allocate a new one and
+ * insert to the head of f.
+ *
+ * If f is NULL, no fib node matched the destination key
+ * and we need to allocate a new one of those as well.
+ */
+
+ if (fa &&
+ fa->fa_info->fib_priority == fi->fib_priority) {
+ struct fib_alias *fa_orig;
+
+ err = -EEXIST;
+ if (nlhdr->nlmsg_flags & NLM_F_EXCL)
+ goto out;
+
+ if (nlhdr->nlmsg_flags & NLM_F_REPLACE) {
+ struct fib_info *fi_drop;
+ u8 state;
+
+ write_lock_bh(&fib_lock);
+
+ fi_drop = fa->fa_info;
+ fa->fa_info = fi;
+ fa->fa_type = type;
+ fa->fa_scope = r->rtm_scope;
+ state = fa->fa_state;
+ fa->fa_state &= ~FA_S_ACCESSED;
+
+ write_unlock_bh(&fib_lock);
+
+ fib_release_info(fi_drop);
+ if (state & FA_S_ACCESSED)
+ rt_cache_flush(-1);
+
+ goto succeeded;
+ }
+ /* Error if we find a perfect match which
+ * uses the same scope, type, and nexthop
+ * information.
+ */
+ fa_orig = fa;
+ list_for_each_entry(fa, fa_orig->fa_list.prev, fa_list) {
+ if (fa->fa_tos != tos)
+ break;
+ if (fa->fa_info->fib_priority != fi->fib_priority)
+ break;
+ if (fa->fa_type == type &&
+ fa->fa_scope == r->rtm_scope &&
+ fa->fa_info == fi) {
+ goto out;
+ }
+ }
+ if (!(nlhdr->nlmsg_flags & NLM_F_APPEND))
+ fa = fa_orig;
+ }
+ err = -ENOENT;
+ if (!(nlhdr->nlmsg_flags&NLM_F_CREATE))
+ goto out;
+
+ err = -ENOBUFS;
+ new_fa = kmem_cache_alloc(fn_alias_kmem, SLAB_KERNEL);
+ if (new_fa == NULL)
+ goto out;
+
+ new_fa->fa_info = fi;
+ new_fa->fa_tos = tos;
+ new_fa->fa_type = type;
+ new_fa->fa_scope = r->rtm_scope;
+ new_fa->fa_state = 0;
+#if 0
+ new_fa->dst = NULL;
+#endif
+ /*
+ * Insert new entry to the list.
+ */
+
+ if(!fa_head) {
+ fa_head = fib_insert_node(t, &err, key, plen);
+ err = 0;
+ if(err)
+ goto out_free_new_fa;
+ }
+
+ write_lock_bh(&fib_lock);
+
+ list_add_tail(&new_fa->fa_list,
+ (fa ? &fa->fa_list : fa_head));
+
+ write_unlock_bh(&fib_lock);
+
+ rt_cache_flush(-1);
+ rtmsg_fib(RTM_NEWROUTE, htonl(key), new_fa, plen, tb->tb_id, nlhdr, req);
+succeeded:
+ return 0;
+
+out_free_new_fa:
+ kmem_cache_free(fn_alias_kmem, new_fa);
+out:
+ fib_release_info(fi);
+err:;
+ return err;
+}
+
+static inline int check_leaf(struct trie *t, struct leaf *l, t_key key, int *plen, const struct flowi *flp,
+ struct fib_result *res, int *err)
+{
+ int i;
+ t_key mask;
+ struct leaf_info *li;
+ struct hlist_head *hhead = &l->list;
+ struct hlist_node *node;
+
+ hlist_for_each_entry(li, node, hhead, hlist) {
+
+ i = li->plen;
+ mask = ntohl(inet_make_mask(i));
+ if (l->key != (key & mask))
+ continue;
+
+ if (((*err) = fib_semantic_match(&li->falh, flp, res, l->key, mask, i)) == 0) {
+ *plen = i;
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.semantic_match_passed++;
+#endif
+ return 1;
+ }
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.semantic_match_miss++;
+#endif
+ }
+ return 0;
+}
+
+static int
+fn_trie_lookup(struct fib_table *tb, const struct flowi *flp, struct fib_result *res)
+{
+ struct trie *t = (struct trie *) tb->tb_data;
+ int plen, ret = 0;
+ struct node *n;
+ struct tnode *pn;
+ int pos, bits;
+ t_key key=ntohl(flp->fl4_dst);
+ int chopped_off;
+ t_key cindex = 0;
+ int current_prefix_length = KEYLENGTH;
+ n = t->trie;
+
+ read_lock(&fib_lock);
+ if(!n)
+ goto failed;
+
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.gets++;
+#endif
+
+ /* Just a leaf? */
+ if (IS_LEAF(n)) {
+ if( check_leaf(t, (struct leaf *)n, key, &plen, flp, res, &ret) )
+ goto found;
+ goto failed;
+ }
+ pn = (struct tnode *) n;
+ chopped_off = 0;
+
+ while (pn) {
+
+ pos = pn->pos;
+ bits = pn->bits;
+
+ if(!chopped_off)
+ cindex = tkey_extract_bits(MASK_PFX(key, current_prefix_length), pos, bits);
+
+ n = tnode_get_child(pn, cindex);
+
+ if (n == NULL) {
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.null_node_hit++;
+#endif
+ goto backtrace;
+ }
+
+ if (IS_TNODE(n)) {
+#define HL_OPTIMIZE
+#ifdef HL_OPTIMIZE
+ struct tnode *cn = (struct tnode *)n;
+ t_key node_prefix, key_prefix, pref_mismatch;
+ int mp;
+
+ /*
+ * It's a tnode, and we can do some extra checks here if we
+ * like, to avoid descending into a dead-end branch.
+ * This tnode is in the parent's child array at index
+ * key[p_pos..p_pos+p_bits] but potentially with some bits
+ * chopped off, so in reality the index may be just a
+ * subprefix, padded with zero at the end.
+ * We can also take a look at any skipped bits in this
+ * tnode - everything up to p_pos is supposed to be ok,
+ * and the non-chopped bits of the index (se previous
+ * paragraph) are also guaranteed ok, but the rest is
+ * considered unknown.
+ *
+ * The skipped bits are key[pos+bits..cn->pos].
+ */
+
+ /* If current_prefix_length < pos+bits, we are already doing
+ * actual prefix matching, which means everything from
+ * pos+(bits-chopped_off) onward must be zero along some
+ * branch of this subtree - otherwise there is *no* valid
+ * prefix present. Here we can only check the skipped
+ * bits. Remember, since we have already indexed into the
+ * parent's child array, we know that the bits we chopped of
+ * *are* zero.
+ */
+
+ /* NOTA BENE: CHECKING ONLY SKIPPED BITS FOR THE NEW NODE HERE */
+
+ if (current_prefix_length < pos+bits) {
+ if (tkey_extract_bits(cn->key, current_prefix_length,
+ cn->pos - current_prefix_length) != 0 ||
+ !(cn->child[0]))
+ goto backtrace;
+ }
+
+ /*
+ * If chopped_off=0, the index is fully validated and we
+ * only need to look at the skipped bits for this, the new,
+ * tnode. What we actually want to do is to find out if
+ * these skipped bits match our key perfectly, or if we will
+ * have to count on finding a matching prefix further down,
+ * because if we do, we would like to have some way of
+ * verifying the existence of such a prefix at this point.
+ */
+
+ /* The only thing we can do at this point is to verify that
+ * any such matching prefix can indeed be a prefix to our
+ * key, and if the bits in the node we are inspecting that
+ * do not match our key are not ZERO, this cannot be true.
+ * Thus, find out where there is a mismatch (before cn->pos)
+ * and verify that all the mismatching bits are zero in the
+ * new tnode's key.
+ */
+
+ /* Note: We aren't very concerned about the piece of the key
+ * that precede pn->pos+pn->bits, since these have already been
+ * checked. The bits after cn->pos aren't checked since these are
+ * by definition "unknown" at this point. Thus, what we want to
+ * see is if we are about to enter the "prefix matching" state,
+ * and in that case verify that the skipped bits that will prevail
+ * throughout this subtree are zero, as they have to be if we are
+ * to find a matching prefix.
+ */
+
+ node_prefix = MASK_PFX(cn->key, cn->pos);
+ key_prefix = MASK_PFX(key, cn->pos);
+ pref_mismatch = key_prefix^node_prefix;
+ mp = 0;
+
+ /* In short: If skipped bits in this node do not match the search
+ * key, enter the "prefix matching" state.directly.
+ */
+ if (pref_mismatch) {
+ while (!(pref_mismatch & (1<<(KEYLENGTH-1)))) {
+ mp++;
+ pref_mismatch = pref_mismatch <<1;
+ }
+ key_prefix = tkey_extract_bits(cn->key, mp, cn->pos-mp);
+
+ if (key_prefix != 0)
+ goto backtrace;
+
+ if (current_prefix_length >= cn->pos)
+ current_prefix_length=mp;
+ }
+#endif
+ pn = (struct tnode *)n; /* Descend */
+ chopped_off = 0;
+ continue;
+ }
+ if (IS_LEAF(n)) {
+ if( check_leaf(t, (struct leaf *)n, key, &plen, flp, res, &ret))
+ goto found;
+ }
+backtrace:
+ chopped_off++;
+
+ /* As zero don't change the child key (cindex) */
+ while ((chopped_off <= pn->bits) && !(cindex & (1<<(chopped_off-1)))) {
+ chopped_off++;
+ }
+
+ /* Decrease current_... with bits chopped off */
+ if (current_prefix_length > pn->pos + pn->bits - chopped_off)
+ current_prefix_length = pn->pos + pn->bits - chopped_off;
+
+ /*
+ * Either we do the actual chop off according or if we have
+ * chopped off all bits in this tnode walk up to our parent.
+ */
+
+ if(chopped_off <= pn->bits)
+ cindex &= ~(1 << (chopped_off-1));
+ else {
+ if( NODE_PARENT(pn) == NULL)
+ goto failed;
+
+ /* Get Child's index */
+ cindex = tkey_extract_bits(pn->key, NODE_PARENT(pn)->pos, NODE_PARENT(pn)->bits);
+ pn = NODE_PARENT(pn);
+ chopped_off = 0;
+
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ t->stats.backtrack++;
+#endif
+ goto backtrace;
+ }
+ }
+failed:
+ ret = 1;
+found:
+ read_unlock(&fib_lock);
+ return ret;
+}
+
+static int trie_leaf_remove(struct trie *t, t_key key)
+{
+ t_key cindex;
+ struct tnode *tp = NULL;
+ struct node *n = t->trie;
+ struct leaf *l;
+
+ if(trie_debug)
+ printk("entering trie_leaf_remove(%p)\n", n);
+
+ /* Note that in the case skipped bits, those bits are *not* checked!
+ * When we finish this, we will have NULL or a T_LEAF, and the
+ * T_LEAF may or may not match our key.
+ */
+
+ while (n != NULL && IS_TNODE(n)) {
+ struct tnode *tn = (struct tnode *) n;
+ check_tnode(tn);
+ n = tnode_get_child(tn ,tkey_extract_bits(key, tn->pos, tn->bits));
+
+ if(n && NODE_PARENT(n) != tn) {
+ printk("BUG tn=%p, n->parent=%p\n", tn, NODE_PARENT(n));
+ BUG();
+ }
+ }
+ l = (struct leaf *) n;
+
+ if(!n || !tkey_equals(l->key, key))
+ return 0;
+
+ /*
+ * Key found.
+ * Remove the leaf and rebalance the tree
+ */
+
+ t->revision++;
+ t->size--;
+
+ tp = NODE_PARENT(n);
+ tnode_free((struct tnode *) n);
+
+ if(tp) {
+ cindex = tkey_extract_bits(key, tp->pos, tp->bits);
+ put_child(t, (struct tnode *)tp, cindex, NULL);
+ t->trie = trie_rebalance(t, tp);
+ }
+ else
+ t->trie = NULL;
+
+ return 1;
+}
+
+static int
+fn_trie_delete(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta,
+ struct nlmsghdr *nlhdr, struct netlink_skb_parms *req)
+{
+ struct trie *t = (struct trie *) tb->tb_data;
+ u32 key, mask;
+ int plen = r->rtm_dst_len;
+ u8 tos = r->rtm_tos;
+ struct fib_alias *fa, *fa_to_delete;
+ struct list_head *fa_head;
+ struct leaf *l;
+
+ if (plen > 32)
+ return -EINVAL;
+
+ key = 0;
+ if (rta->rta_dst)
+ memcpy(&key, rta->rta_dst, 4);
+
+ key = ntohl(key);
+ mask = ntohl( inet_make_mask(plen) );
+
+ if(key & ~mask)
+ return -EINVAL;
+
+ key = key & mask;
+ l = fib_find_node(t, key);
+
+ if(!l)
+ return -ESRCH;
+
+ fa_head = get_fa_head(l, plen);
+ fa = fib_find_alias(fa_head, tos, 0);
+
+ if (!fa)
+ return -ESRCH;
+
+ if (trie_debug)
+ printk("Deleting %08x/%d tos=%d t=%p\n", key, plen, tos, t);
+
+ fa_to_delete = NULL;
+ fa_head = fa->fa_list.prev;
+ list_for_each_entry(fa, fa_head, fa_list) {
+ struct fib_info *fi = fa->fa_info;
+
+ if (fa->fa_tos != tos)
+ break;
+
+ if ((!r->rtm_type ||
+ fa->fa_type == r->rtm_type) &&
+ (r->rtm_scope == RT_SCOPE_NOWHERE ||
+ fa->fa_scope == r->rtm_scope) &&
+ (!r->rtm_protocol ||
+ fi->fib_protocol == r->rtm_protocol) &&
+ fib_nh_match(r, nlhdr, rta, fi) == 0) {
+ fa_to_delete = fa;
+ break;
+ }
+ }
+
+ if (fa_to_delete) {
+ int kill_li = 0;
+ struct leaf_info *li;
+
+ fa = fa_to_delete;
+ rtmsg_fib(RTM_DELROUTE, htonl(key), fa, plen, tb->tb_id, nlhdr, req);
+
+ l = fib_find_node(t, key);
+ li = find_leaf_info(&l->list, plen);
+
+ write_lock_bh(&fib_lock);
+
+ list_del(&fa->fa_list);
+
+ if(list_empty(fa_head)) {
+ hlist_del(&li->hlist);
+ kill_li = 1;
+ }
+ write_unlock_bh(&fib_lock);
+
+ if(kill_li)
+ free_leaf_info(li);
+
+ if(hlist_empty(&l->list))
+ trie_leaf_remove(t, key);
+
+ if (fa->fa_state & FA_S_ACCESSED)
+ rt_cache_flush(-1);
+
+ fn_free_alias(fa);
+ return 0;
+ }
+ return -ESRCH;
+}
+
+static int trie_flush_list(struct trie *t, struct list_head *head)
+{
+ struct fib_alias *fa, *fa_node;
+ int found = 0;
+
+ list_for_each_entry_safe(fa, fa_node, head, fa_list) {
+ struct fib_info *fi = fa->fa_info;
+
+ if (fi && (fi->fib_flags&RTNH_F_DEAD)) {
+
+ write_lock_bh(&fib_lock);
+ list_del(&fa->fa_list);
+ write_unlock_bh(&fib_lock);
+
+ fn_free_alias(fa);
+ found++;
+ }
+ }
+ return found;
+}
+
+static int trie_flush_leaf(struct trie *t, struct leaf *l)
+{
+ int found = 0;
+ struct hlist_head *lih = &l->list;
+ struct hlist_node *node, *tmp;
+ struct leaf_info *li = NULL;
+
+ hlist_for_each_entry_safe(li, node, tmp, lih, hlist) {
+
+ found += trie_flush_list(t, &li->falh);
+
+ if (list_empty(&li->falh)) {
+
+ write_lock_bh(&fib_lock);
+ hlist_del(&li->hlist);
+ write_unlock_bh(&fib_lock);
+
+ free_leaf_info(li);
+ }
+ }
+ return found;
+}
+
+static struct leaf *nextleaf(struct trie *t, struct leaf *thisleaf)
+{
+ struct node *c = (struct node *) thisleaf;
+ struct tnode *p;
+ int idx;
+
+ if(c == NULL) {
+ if(t->trie == NULL)
+ return NULL;
+
+ if (IS_LEAF(t->trie)) /* trie w. just a leaf */
+ return (struct leaf *) t->trie;
+
+ p = (struct tnode*) t->trie; /* Start */
+ }
+ else
+ p = (struct tnode *) NODE_PARENT(c);
+ while (p) {
+ int pos, last;
+
+ /* Find the next child of the parent */
+ if(c)
+ pos = 1 + tkey_extract_bits(c->key, p->pos, p->bits);
+ else
+ pos = 0;
+
+ last = 1 << p->bits;
+ for(idx = pos; idx < last ; idx++) {
+ if( p->child[idx]) {
+
+ /* Decend if tnode */
+
+ while (IS_TNODE(p->child[idx])) {
+ p = (struct tnode*) p->child[idx];
+ idx = 0;
+
+ /* Rightmost non-NULL branch */
+ if( p && IS_TNODE(p) )
+ while ( p->child[idx] == NULL && idx < (1 << p->bits) ) idx++;
+
+ /* Done with this tnode? */
+ if( idx >= (1 << p->bits) || p->child[idx] == NULL )
+ goto up;
+ }
+ return (struct leaf*) p->child[idx];
+ }
+ }
+up:
+ /* No more children go up one step */
+ c = (struct node*) p;
+ p = (struct tnode *) NODE_PARENT(p);
+ }
+ return NULL; /* Ready. Root of trie */
+}
+
+static int fn_trie_flush(struct fib_table *tb)
+{
+ struct trie *t = (struct trie *) tb->tb_data;
+ struct leaf *ll = NULL, *l = NULL;
+ int found = 0, h;
+
+ t->revision++;
+
+ for (h=0; (l = nextleaf(t, l)) != NULL; h++) {
+ found += trie_flush_leaf(t, l);
+
+ if (ll && hlist_empty(&ll->list))
+ trie_leaf_remove(t, ll->key);
+ ll = l;
+ }
+
+ if (ll && hlist_empty(&ll->list))
+ trie_leaf_remove(t, ll->key);
+
+ if(trie_debug)
+ printk("trie_flush found=%d\n", found);
+ return found;
+}
+
+static int trie_last_dflt=-1;
+
+static void
+fn_trie_select_default(struct fib_table *tb, const struct flowi *flp, struct fib_result *res)
+{
+ struct trie *t = (struct trie *) tb->tb_data;
+ int order, last_idx;
+ struct fib_info *fi = NULL;
+ struct fib_info *last_resort;
+ struct fib_alias *fa = NULL;
+ struct list_head *fa_head;
+ struct leaf *l;
+
+ last_idx = -1;
+ last_resort = NULL;
+ order = -1;
+
+ read_lock(&fib_lock);
+
+ l = fib_find_node(t, 0);
+ if(!l)
+ goto out;
+
+ fa_head = get_fa_head(l, 0);
+ if(!fa_head)
+ goto out;
+
+ if (list_empty(fa_head))
+ goto out;
+
+ list_for_each_entry(fa, fa_head, fa_list) {
+ struct fib_info *next_fi = fa->fa_info;
+
+ if (fa->fa_scope != res->scope ||
+ fa->fa_type != RTN_UNICAST)
+ continue;
+
+ if (next_fi->fib_priority > res->fi->fib_priority)
+ break;
+ if (!next_fi->fib_nh[0].nh_gw ||
+ next_fi->fib_nh[0].nh_scope != RT_SCOPE_LINK)
+ continue;
+ fa->fa_state |= FA_S_ACCESSED;
+
+ if (fi == NULL) {
+ if (next_fi != res->fi)
+ break;
+ } else if (!fib_detect_death(fi, order, &last_resort,
+ &last_idx, &trie_last_dflt)) {
+ if (res->fi)
+ fib_info_put(res->fi);
+ res->fi = fi;
+ atomic_inc(&fi->fib_clntref);
+ trie_last_dflt = order;
+ goto out;
+ }
+ fi = next_fi;
+ order++;
+ }
+ if (order <= 0 || fi == NULL) {
+ trie_last_dflt = -1;
+ goto out;
+ }
+
+ if (!fib_detect_death(fi, order, &last_resort, &last_idx, &trie_last_dflt)) {
+ if (res->fi)
+ fib_info_put(res->fi);
+ res->fi = fi;
+ atomic_inc(&fi->fib_clntref);
+ trie_last_dflt = order;
+ goto out;
+ }
+ if (last_idx >= 0) {
+ if (res->fi)
+ fib_info_put(res->fi);
+ res->fi = last_resort;
+ if (last_resort)
+ atomic_inc(&last_resort->fib_clntref);
+ }
+ trie_last_dflt = last_idx;
+ out:;
+ read_unlock(&fib_lock);
+}
+
+static int fn_trie_dump_fa(t_key key, int plen, struct list_head *fah, struct fib_table *tb,
+ struct sk_buff *skb, struct netlink_callback *cb)
+{
+ int i, s_i;
+ struct fib_alias *fa;
+
+ u32 xkey=htonl(key);
+
+ s_i=cb->args[3];
+ i = 0;
+
+ list_for_each_entry(fa, fah, fa_list) {
+ if (i < s_i) {
+ i++;
+ continue;
+ }
+ if (fa->fa_info->fib_nh == NULL) {
+ printk("Trie error _fib_nh=NULL in fa[%d] k=%08x plen=%d\n", i, key, plen);
+ i++;
+ continue;
+ }
+ if (fa->fa_info == NULL) {
+ printk("Trie error fa_info=NULL in fa[%d] k=%08x plen=%d\n", i, key, plen);
+ i++;
+ continue;
+ }
+
+ if (fib_dump_info(skb, NETLINK_CB(cb->skb).pid,
+ cb->nlh->nlmsg_seq,
+ RTM_NEWROUTE,
+ tb->tb_id,
+ fa->fa_type,
+ fa->fa_scope,
+ &xkey,
+ plen,
+ fa->fa_tos,
+ fa->fa_info, 0) < 0) {
+ cb->args[3] = i;
+ return -1;
+ }
+ i++;
+ }
+ cb->args[3]=i;
+ return skb->len;
+}
+
+static int fn_trie_dump_plen(struct trie *t, int plen, struct fib_table *tb, struct sk_buff *skb,
+ struct netlink_callback *cb)
+{
+ int h, s_h;
+ struct list_head *fa_head;
+ struct leaf *l = NULL;
+ s_h=cb->args[2];
+
+ for (h=0; (l = nextleaf(t, l)) != NULL; h++) {
+
+ if (h < s_h)
+ continue;
+ if (h > s_h)
+ memset(&cb->args[3], 0,
+ sizeof(cb->args) - 3*sizeof(cb->args[0]));
+
+ fa_head = get_fa_head(l, plen);
+
+ if(!fa_head)
+ continue;
+
+ if(list_empty(fa_head))
+ continue;
+
+ if (fn_trie_dump_fa(l->key, plen, fa_head, tb, skb, cb)<0) {
+ cb->args[2]=h;
+ return -1;
+ }
+ }
+ cb->args[2]=h;
+ return skb->len;
+}
+
+static int fn_trie_dump(struct fib_table *tb, struct sk_buff *skb, struct netlink_callback *cb)
+{
+ int m, s_m;
+ struct trie *t = (struct trie *) tb->tb_data;
+
+ s_m = cb->args[1];
+
+ read_lock(&fib_lock);
+ for (m=0; m<=32; m++) {
+
+ if (m < s_m)
+ continue;
+ if (m > s_m)
+ memset(&cb->args[2], 0,
+ sizeof(cb->args) - 2*sizeof(cb->args[0]));
+
+ if (fn_trie_dump_plen(t, 32-m, tb, skb, cb)<0) {
+ cb->args[1] = m;
+ goto out;
+ }
+ }
+ read_unlock(&fib_lock);
+ cb->args[1] = m;
+ return skb->len;
+ out:
+ read_unlock(&fib_lock);
+ return -1;
+}
+
+/* Fix more generic FIB names for init later */
+
+#ifdef CONFIG_IP_MULTIPLE_TABLES
+struct fib_table * fib_hash_init(int id)
+#else
+struct fib_table * __init fib_hash_init(int id)
+#endif
+{
+ struct fib_table *tb;
+ struct trie *t;
+
+ if (fn_alias_kmem == NULL)
+ fn_alias_kmem = kmem_cache_create("ip_fib_alias",
+ sizeof(struct fib_alias),
+ 0, SLAB_HWCACHE_ALIGN,
+ NULL, NULL);
+
+ tb = kmalloc(sizeof(struct fib_table) + sizeof(struct trie),
+ GFP_KERNEL);
+ if (tb == NULL)
+ return NULL;
+
+ tb->tb_id = id;
+ tb->tb_lookup = fn_trie_lookup;
+ tb->tb_insert = fn_trie_insert;
+ tb->tb_delete = fn_trie_delete;
+ tb->tb_flush = fn_trie_flush;
+ tb->tb_select_default = fn_trie_select_default;
+ tb->tb_dump = fn_trie_dump;
+ memset(tb->tb_data, 0, sizeof(struct trie));
+
+ t = (struct trie *) tb->tb_data;
+
+ trie_init(t);
+
+ if (id == RT_TABLE_LOCAL)
+ trie_local=t;
+ else if (id == RT_TABLE_MAIN)
+ trie_main=t;
+
+ if (id == RT_TABLE_LOCAL)
+ printk("IPv4 FIB: Using LC-trie version %s\n", VERSION);
+
+ return tb;
+}
+
+/* Trie dump functions */
+
+static void putspace_seq(struct seq_file *seq, int n)
+{
+ while (n--) seq_printf(seq, " ");
+}
+
+static void printbin_seq(struct seq_file *seq, unsigned int v, int bits)
+{
+ while (bits--)
+ seq_printf(seq, "%s", (v & (1<<bits))?"1":"0");
+}
+
+static void printnode_seq(struct seq_file *seq, int indent, struct node *n,
+ int pend, int cindex, int bits)
+{
+ putspace_seq(seq, indent);
+ if (IS_LEAF(n))
+ seq_printf(seq, "|");
+ else
+ seq_printf(seq, "+");
+ if (bits) {
+ seq_printf(seq, "%d/", cindex);
+ printbin_seq(seq, cindex, bits);
+ seq_printf(seq, ": ");
+ }
+ else
+ seq_printf(seq, "<root>: ");
+ seq_printf(seq, "%s:%p ", IS_LEAF(n)?"Leaf":"Internal node", n);
+
+ if (IS_LEAF(n))
+ seq_printf(seq, "key=%d.%d.%d.%d\n",
+ n->key >> 24, (n->key >> 16) % 256, (n->key >> 8) % 256, n->key % 256);
+ else {
+ int plen=((struct tnode *)n)->pos;
+ t_key prf=MASK_PFX(n->key, plen);
+ seq_printf(seq, "key=%d.%d.%d.%d/%d\n",
+ prf >> 24, (prf >> 16) % 256, (prf >> 8) % 256, prf % 256, plen);
+ }
+ if (IS_LEAF(n)) {
+ struct leaf *l=(struct leaf *)n;
+ struct fib_alias *fa;
+ int i;
+ for (i=32; i>=0; i--)
+ if(find_leaf_info(&l->list, i)) {
+
+ struct list_head *fa_head = get_fa_head(l, i);
+
+ if(!fa_head)
+ continue;
+
+ if(list_empty(fa_head))
+ continue;
+
+ putspace_seq(seq, indent+2);
+ seq_printf(seq, "{/%d...dumping}\n", i);
+
+
+ list_for_each_entry(fa, fa_head, fa_list) {
+ putspace_seq(seq, indent+2);
+ if (fa->fa_info->fib_nh == NULL) {
+ seq_printf(seq, "Error _fib_nh=NULL\n");
+ continue;
+ }
+ if (fa->fa_info == NULL) {
+ seq_printf(seq, "Error fa_info=NULL\n");
+ continue;
+ }
+
+ seq_printf(seq, "{type=%d scope=%d TOS=%d}\n",
+ fa->fa_type,
+ fa->fa_scope,
+ fa->fa_tos);
+ }
+ }
+ }
+ else if (IS_TNODE(n)) {
+ struct tnode *tn=(struct tnode *)n;
+ putspace_seq(seq, indent); seq_printf(seq, "| ");
+ seq_printf(seq, "{key prefix=%08x/", tn->key&TKEY_GET_MASK(0, tn->pos));
+ printbin_seq(seq, tkey_extract_bits(tn->key, 0, tn->pos), tn->pos);
+ seq_printf(seq, "}\n");
+ putspace_seq(seq, indent); seq_printf(seq, "| ");
+ seq_printf(seq, "{pos=%d", tn->pos);
+ seq_printf(seq, " (skip=%d bits)", tn->pos - pend);
+ seq_printf(seq, " bits=%d (%u children)}\n", tn->bits, (1 << tn->bits));
+ putspace_seq(seq, indent); seq_printf(seq, "| ");
+ seq_printf(seq, "{empty=%d full=%d}\n", tn->empty_children, tn->full_children);
+ }
+}
+
+static void trie_dump_seq(struct seq_file *seq, struct trie *t)
+{
+ struct node *n=t->trie;
+ int cindex=0;
+ int indent=1;
+ int pend=0;
+ int depth = 0;
+
+ read_lock(&fib_lock);
+
+ seq_printf(seq, "------ trie_dump of t=%p ------\n", t);
+ if (n) {
+ printnode_seq(seq, indent, n, pend, cindex, 0);
+ if (IS_TNODE(n)) {
+ struct tnode *tn=(struct tnode *)n;
+ pend = tn->pos+tn->bits;
+ putspace_seq(seq, indent); seq_printf(seq, "\\--\n");
+ indent += 3;
+ depth++;
+
+ while (tn && cindex < (1 << tn->bits)) {
+ if (tn->child[cindex]) {
+
+ /* Got a child */
+
+ printnode_seq(seq, indent, tn->child[cindex], pend, cindex, tn->bits);
+ if (IS_LEAF(tn->child[cindex])) {
+ cindex++;
+
+ }
+ else {
+ /*
+ * New tnode. Decend one level
+ */
+
+ depth++;
+ n=tn->child[cindex];
+ tn=(struct tnode *)n;
+ pend=tn->pos+tn->bits;
+ putspace_seq(seq, indent); seq_printf(seq, "\\--\n");
+ indent+=3;
+ cindex=0;
+ }
+ }
+ else
+ cindex++;
+
+ /*
+ * Test if we are done
+ */
+
+ while (cindex >= (1 << tn->bits)) {
+
+ /*
+ * Move upwards and test for root
+ * pop off all traversed nodes
+ */
+
+ if (NODE_PARENT(tn) == NULL) {
+ tn = NULL;
+ n = NULL;
+ break;
+ }
+ else {
+ cindex = tkey_extract_bits(tn->key, NODE_PARENT(tn)->pos, NODE_PARENT(tn)->bits);
+ tn = NODE_PARENT(tn);
+ cindex++;
+ n=(struct node *)tn;
+ pend=tn->pos+tn->bits;
+ indent-=3;
+ depth--;
+ }
+ }
+ }
+ }
+ else n = NULL;
+ }
+ else seq_printf(seq, "------ trie is empty\n");
+
+ read_unlock(&fib_lock);
+}
+
+static struct trie_stat *trie_stat_new(void)
+{
+ struct trie_stat *s = kmalloc(sizeof(struct trie_stat), GFP_KERNEL);
+ int i;
+
+ if(s) {
+ s->totdepth = 0;
+ s->maxdepth = 0;
+ s->tnodes = 0;
+ s->leaves = 0;
+ s->nullpointers = 0;
+
+ for(i=0; i< MAX_CHILDS; i++)
+ s->nodesizes[i] = 0;
+ }
+ return s;
+}
+
+static struct trie_stat *trie_collect_stats(struct trie *t)
+{
+ struct node *n=t->trie;
+ struct trie_stat *s = trie_stat_new();
+ int cindex = 0;
+ int indent = 1;
+ int pend = 0;
+ int depth = 0;
+
+ read_lock(&fib_lock);
+
+ if (s) {
+ if (n) {
+ if (IS_TNODE(n)) {
+ struct tnode *tn = (struct tnode *)n;
+ pend=tn->pos+tn->bits;
+ indent += 3;
+ s->nodesizes[tn->bits]++;
+ depth++;
+
+ while (tn && cindex < (1 << tn->bits)) {
+ if (tn->child[cindex]) {
+ /* Got a child */
+
+ if (IS_LEAF(tn->child[cindex])) {
+ cindex++;
+
+ /* stats */
+ if (depth > s->maxdepth)
+ s->maxdepth = depth;
+ s->totdepth += depth;
+ s->leaves++;
+ }
+
+ else {
+ /*
+ * New tnode. Decend one level
+ */
+
+ s->tnodes++;
+ s->nodesizes[tn->bits]++;
+ depth++;
+
+ n = tn->child[cindex];
+ tn = (struct tnode *)n;
+ pend = tn->pos+tn->bits;
+
+ indent += 3;
+ cindex = 0;
+ }
+ }
+ else {
+ cindex++;
+ s->nullpointers++;
+ }
+
+ /*
+ * Test if we are done
+ */
+
+ while (cindex >= (1 << tn->bits)) {
+
+ /*
+ * Move upwards and test for root
+ * pop off all traversed nodes
+ */
+
+
+ if (NODE_PARENT(tn) == NULL) {
+ tn = NULL;
+ n = NULL;
+ break;
+ }
+ else {
+ cindex = tkey_extract_bits(tn->key, NODE_PARENT(tn)->pos, NODE_PARENT(tn)->bits);
+ tn = NODE_PARENT(tn);
+ cindex++;
+ n = (struct node *)tn;
+ pend=tn->pos+tn->bits;
+ indent -= 3;
+ depth--;
+ }
+ }
+ }
+ }
+ else n = NULL;
+ }
+ }
+
+ read_unlock(&fib_lock);
+ return s;
+}
+
+#ifdef CONFIG_PROC_FS
+
+static struct fib_alias *fib_triestat_get_first(struct seq_file *seq)
+{
+ return NULL;
+}
+
+static struct fib_alias *fib_triestat_get_next(struct seq_file *seq)
+{
+ return NULL;
+}
+
+static void *fib_triestat_seq_start(struct seq_file *seq, loff_t *pos)
+{
+ void *v = NULL;
+
+ if (ip_fib_main_table)
+ v = *pos ? fib_triestat_get_next(seq) : SEQ_START_TOKEN;
+ return v;
+}
+
+static void *fib_triestat_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ ++*pos;
+ return v == SEQ_START_TOKEN ? fib_triestat_get_first(seq) : fib_triestat_get_next(seq);
+}
+
+static void fib_triestat_seq_stop(struct seq_file *seq, void *v)
+{
+
+}
+
+/*
+ * This outputs /proc/net/fib_triestats
+ *
+ * It always works in backward compatibility mode.
+ * The format of the file is not supposed to be changed.
+ */
+
+static void collect_and_show(struct trie *t, struct seq_file *seq)
+{
+ int bytes = 0; /* How many bytes are used, a ref is 4 bytes */
+ int i, max, pointers;
+ struct trie_stat *stat;
+ int avdepth;
+
+ stat = trie_collect_stats(t);
+
+ bytes=0;
+ seq_printf(seq, "trie=%p\n", t);
+
+ if (stat) {
+ if (stat->leaves)
+ avdepth=stat->totdepth*100 / stat->leaves;
+ else
+ avdepth=0;
+ seq_printf(seq, "Aver depth: %d.%02d\n", avdepth / 100, avdepth % 100 );
+ seq_printf(seq, "Max depth: %4d\n", stat->maxdepth);
+
+ seq_printf(seq, "Leaves: %d\n", stat->leaves);
+ bytes += sizeof(struct leaf) * stat->leaves;
+ seq_printf(seq, "Internal nodes: %d\n", stat->tnodes);
+ bytes += sizeof(struct tnode) * stat->tnodes;
+
+ max = MAX_CHILDS-1;
+
+ while (max >= 0 && stat->nodesizes[max] == 0)
+ max--;
+ pointers = 0;
+
+ for (i = 1; i <= max; i++)
+ if (stat->nodesizes[i] != 0) {
+ seq_printf(seq, " %d: %d", i, stat->nodesizes[i]);
+ pointers += (1<<i) * stat->nodesizes[i];
+ }
+ seq_printf(seq, "\n");
+ seq_printf(seq, "Pointers: %d\n", pointers);
+ bytes += sizeof(struct node *) * pointers;
+ seq_printf(seq, "Null ptrs: %d\n", stat->nullpointers);
+ seq_printf(seq, "Total size: %d kB\n", bytes / 1024);
+
+ kfree(stat);
+ }
+
+#ifdef CONFIG_IP_FIB_TRIE_STATS
+ seq_printf(seq, "Counters:\n---------\n");
+ seq_printf(seq,"gets = %d\n", t->stats.gets);
+ seq_printf(seq,"backtracks = %d\n", t->stats.backtrack);
+ seq_printf(seq,"semantic match passed = %d\n", t->stats.semantic_match_passed);
+ seq_printf(seq,"semantic match miss = %d\n", t->stats.semantic_match_miss);
+ seq_printf(seq,"null node hit= %d\n", t->stats.null_node_hit);
+ seq_printf(seq,"skipped node resize = %d\n", t->stats.resize_node_skipped);
+#ifdef CLEAR_STATS
+ memset(&(t->stats), 0, sizeof(t->stats));
+#endif
+#endif /* CONFIG_IP_FIB_TRIE_STATS */
+}
+
+static int fib_triestat_seq_show(struct seq_file *seq, void *v)
+{
+ char bf[128];
+
+ if (v == SEQ_START_TOKEN) {
+ seq_printf(seq, "Basic info: size of leaf: %Zd bytes, size of tnode: %Zd bytes.\n",
+ sizeof(struct leaf), sizeof(struct tnode));
+ if (trie_local)
+ collect_and_show(trie_local, seq);
+
+ if (trie_main)
+ collect_and_show(trie_main, seq);
+ }
+ else {
+ snprintf(bf, sizeof(bf),
+ "*\t%08X\t%08X", 200, 400);
+
+ seq_printf(seq, "%-127s\n", bf);
+ }
+ return 0;
+}
+
+static struct seq_operations fib_triestat_seq_ops = {
+ .start = fib_triestat_seq_start,
+ .next = fib_triestat_seq_next,
+ .stop = fib_triestat_seq_stop,
+ .show = fib_triestat_seq_show,
+};
+
+static int fib_triestat_seq_open(struct inode *inode, struct file *file)
+{
+ struct seq_file *seq;
+ int rc = -ENOMEM;
+
+ rc = seq_open(file, &fib_triestat_seq_ops);
+ if (rc)
+ goto out_kfree;
+
+ seq = file->private_data;
+out:
+ return rc;
+out_kfree:
+ goto out;
+}
+
+static struct file_operations fib_triestat_seq_fops = {
+ .owner = THIS_MODULE,
+ .open = fib_triestat_seq_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release_private,
+};
+
+int __init fib_stat_proc_init(void)
+{
+ if (!proc_net_fops_create("fib_triestat", S_IRUGO, &fib_triestat_seq_fops))
+ return -ENOMEM;
+ return 0;
+}
+
+void __init fib_stat_proc_exit(void)
+{
+ proc_net_remove("fib_triestat");
+}
+
+static struct fib_alias *fib_trie_get_first(struct seq_file *seq)
+{
+ return NULL;
+}
+
+static struct fib_alias *fib_trie_get_next(struct seq_file *seq)
+{
+ return NULL;
+}
+
+static void *fib_trie_seq_start(struct seq_file *seq, loff_t *pos)
+{
+ void *v = NULL;
+
+ if (ip_fib_main_table)
+ v = *pos ? fib_trie_get_next(seq) : SEQ_START_TOKEN;
+ return v;
+}
+
+static void *fib_trie_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+ ++*pos;
+ return v == SEQ_START_TOKEN ? fib_trie_get_first(seq) : fib_trie_get_next(seq);
+}
+
+static void fib_trie_seq_stop(struct seq_file *seq, void *v)
+{
+
+}
+
+/*
+ * This outputs /proc/net/fib_trie.
+ *
+ * It always works in backward compatibility mode.
+ * The format of the file is not supposed to be changed.
+ */
+
+static int fib_trie_seq_show(struct seq_file *seq, void *v)
+{
+ char bf[128];
+
+ if (v == SEQ_START_TOKEN) {
+ if (trie_local)
+ trie_dump_seq(seq, trie_local);
+
+ if (trie_main)
+ trie_dump_seq(seq, trie_main);
+ }
+
+ else {
+ snprintf(bf, sizeof(bf),
+ "*\t%08X\t%08X", 200, 400);
+ seq_printf(seq, "%-127s\n", bf);
+ }
+
+ return 0;
+}
+
+static struct seq_operations fib_trie_seq_ops = {
+ .start = fib_trie_seq_start,
+ .next = fib_trie_seq_next,
+ .stop = fib_trie_seq_stop,
+ .show = fib_trie_seq_show,
+};
+
+static int fib_trie_seq_open(struct inode *inode, struct file *file)
+{
+ struct seq_file *seq;
+ int rc = -ENOMEM;
+
+ rc = seq_open(file, &fib_trie_seq_ops);
+ if (rc)
+ goto out_kfree;
+
+ seq = file->private_data;
+out:
+ return rc;
+out_kfree:
+ goto out;
+}
+
+static struct file_operations fib_trie_seq_fops = {
+ .owner = THIS_MODULE,
+ .open = fib_trie_seq_open,
+ .read = seq_read,
+ .llseek = seq_lseek,
+ .release = seq_release_private,
+};
+
+int __init fib_proc_init(void)
+{
+ if (!proc_net_fops_create("fib_trie", S_IRUGO, &fib_trie_seq_fops))
+ return -ENOMEM;
+ return 0;
+}
+
+void __init fib_proc_exit(void)
+{
+ proc_net_remove("fib_trie");
+}
+
+#endif /* CONFIG_PROC_FS */
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index cb759484979..279f57abfec 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -970,7 +970,8 @@ int icmp_rcv(struct sk_buff *skb)
* RFC 1122: 3.2.2.8 An ICMP_TIMESTAMP MAY be silently
* discarded if to broadcast/multicast.
*/
- if (icmph->type == ICMP_ECHO &&
+ if ((icmph->type == ICMP_ECHO ||
+ icmph->type == ICMP_TIMESTAMP) &&
sysctl_icmp_echo_ignore_broadcasts) {
goto error;
}
diff --git a/net/ipv4/igmp.c b/net/ipv4/igmp.c
index 1f3183168a9..5088f90835a 100644
--- a/net/ipv4/igmp.c
+++ b/net/ipv4/igmp.c
@@ -1615,9 +1615,10 @@ int ip_mc_join_group(struct sock *sk , struct ip_mreqn *imr)
{
int err;
u32 addr = imr->imr_multiaddr.s_addr;
- struct ip_mc_socklist *iml, *i;
+ struct ip_mc_socklist *iml=NULL, *i;
struct in_device *in_dev;
struct inet_sock *inet = inet_sk(sk);
+ int ifindex;
int count = 0;
if (!MULTICAST(addr))
@@ -1633,37 +1634,30 @@ int ip_mc_join_group(struct sock *sk , struct ip_mreqn *imr)
goto done;
}
- iml = (struct ip_mc_socklist *)sock_kmalloc(sk, sizeof(*iml), GFP_KERNEL);
-
err = -EADDRINUSE;
+ ifindex = imr->imr_ifindex;
for (i = inet->mc_list; i; i = i->next) {
- if (memcmp(&i->multi, imr, sizeof(*imr)) == 0) {
- /* New style additions are reference counted */
- if (imr->imr_address.s_addr == 0) {
- i->count++;
- err = 0;
- }
+ if (i->multi.imr_multiaddr.s_addr == addr &&
+ i->multi.imr_ifindex == ifindex)
goto done;
- }
count++;
}
err = -ENOBUFS;
- if (iml == NULL || count >= sysctl_igmp_max_memberships)
+ if (count >= sysctl_igmp_max_memberships)
+ goto done;
+ iml = (struct ip_mc_socklist *)sock_kmalloc(sk,sizeof(*iml),GFP_KERNEL);
+ if (iml == NULL)
goto done;
+
memcpy(&iml->multi, imr, sizeof(*imr));
iml->next = inet->mc_list;
- iml->count = 1;
iml->sflist = NULL;
iml->sfmode = MCAST_EXCLUDE;
inet->mc_list = iml;
ip_mc_inc_group(in_dev, addr);
- iml = NULL;
err = 0;
-
done:
rtnl_shunlock();
- if (iml)
- sock_kfree_s(sk, iml, sizeof(*iml));
return err;
}
@@ -1693,30 +1687,25 @@ int ip_mc_leave_group(struct sock *sk, struct ip_mreqn *imr)
{
struct inet_sock *inet = inet_sk(sk);
struct ip_mc_socklist *iml, **imlp;
+ struct in_device *in_dev;
+ u32 group = imr->imr_multiaddr.s_addr;
+ u32 ifindex;
rtnl_lock();
+ in_dev = ip_mc_find_dev(imr);
+ if (!in_dev) {
+ rtnl_unlock();
+ return -ENODEV;
+ }
+ ifindex = imr->imr_ifindex;
for (imlp = &inet->mc_list; (iml = *imlp) != NULL; imlp = &iml->next) {
- if (iml->multi.imr_multiaddr.s_addr==imr->imr_multiaddr.s_addr &&
- iml->multi.imr_address.s_addr==imr->imr_address.s_addr &&
- (!imr->imr_ifindex || iml->multi.imr_ifindex==imr->imr_ifindex)) {
- struct in_device *in_dev;
-
- in_dev = inetdev_by_index(iml->multi.imr_ifindex);
- if (in_dev)
- (void) ip_mc_leave_src(sk, iml, in_dev);
- if (--iml->count) {
- rtnl_unlock();
- if (in_dev)
- in_dev_put(in_dev);
- return 0;
- }
+ if (iml->multi.imr_multiaddr.s_addr == group &&
+ iml->multi.imr_ifindex == ifindex) {
+ (void) ip_mc_leave_src(sk, iml, in_dev);
*imlp = iml->next;
- if (in_dev) {
- ip_mc_dec_group(in_dev, imr->imr_multiaddr.s_addr);
- in_dev_put(in_dev);
- }
+ ip_mc_dec_group(in_dev, group);
rtnl_unlock();
sock_kfree_s(sk, iml, sizeof(*iml));
return 0;
@@ -1736,6 +1725,7 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
struct in_device *in_dev = NULL;
struct inet_sock *inet = inet_sk(sk);
struct ip_sf_socklist *psl;
+ int leavegroup = 0;
int i, j, rv;
if (!MULTICAST(addr))
@@ -1755,15 +1745,20 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
err = -EADDRNOTAVAIL;
for (pmc=inet->mc_list; pmc; pmc=pmc->next) {
- if (memcmp(&pmc->multi, mreqs, 2*sizeof(__u32)) == 0)
+ if (pmc->multi.imr_multiaddr.s_addr == imr.imr_multiaddr.s_addr
+ && pmc->multi.imr_ifindex == imr.imr_ifindex)
break;
}
- if (!pmc) /* must have a prior join */
+ if (!pmc) { /* must have a prior join */
+ err = -EINVAL;
goto done;
+ }
/* if a source filter was set, must be the same mode as before */
if (pmc->sflist) {
- if (pmc->sfmode != omode)
+ if (pmc->sfmode != omode) {
+ err = -EINVAL;
goto done;
+ }
} else if (pmc->sfmode != omode) {
/* allow mode switches for empty-set filters */
ip_mc_add_src(in_dev, &mreqs->imr_multiaddr, omode, 0, NULL, 0);
@@ -1775,7 +1770,7 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
psl = pmc->sflist;
if (!add) {
if (!psl)
- goto done;
+ goto done; /* err = -EADDRNOTAVAIL */
rv = !0;
for (i=0; i<psl->sl_count; i++) {
rv = memcmp(&psl->sl_addr[i], &mreqs->imr_sourceaddr,
@@ -1784,7 +1779,13 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
break;
}
if (rv) /* source not found */
+ goto done; /* err = -EADDRNOTAVAIL */
+
+ /* special case - (INCLUDE, empty) == LEAVE_GROUP */
+ if (psl->sl_count == 1 && omode == MCAST_INCLUDE) {
+ leavegroup = 1;
goto done;
+ }
/* update the interface filter */
ip_mc_del_src(in_dev, &mreqs->imr_multiaddr, omode, 1,
@@ -1842,18 +1843,21 @@ int ip_mc_source(int add, int omode, struct sock *sk, struct
&mreqs->imr_sourceaddr, 1);
done:
rtnl_shunlock();
+ if (leavegroup)
+ return ip_mc_leave_group(sk, &imr);
return err;
}
int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex)
{
- int err;
+ int err = 0;
struct ip_mreqn imr;
u32 addr = msf->imsf_multiaddr;
struct ip_mc_socklist *pmc;
struct in_device *in_dev;
struct inet_sock *inet = inet_sk(sk);
struct ip_sf_socklist *newpsl, *psl;
+ int leavegroup = 0;
if (!MULTICAST(addr))
return -EINVAL;
@@ -1872,15 +1876,22 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex)
err = -ENODEV;
goto done;
}
- err = -EADDRNOTAVAIL;
+
+ /* special case - (INCLUDE, empty) == LEAVE_GROUP */
+ if (msf->imsf_fmode == MCAST_INCLUDE && msf->imsf_numsrc == 0) {
+ leavegroup = 1;
+ goto done;
+ }
for (pmc=inet->mc_list; pmc; pmc=pmc->next) {
if (pmc->multi.imr_multiaddr.s_addr == msf->imsf_multiaddr &&
pmc->multi.imr_ifindex == imr.imr_ifindex)
break;
}
- if (!pmc) /* must have a prior join */
+ if (!pmc) { /* must have a prior join */
+ err = -EINVAL;
goto done;
+ }
if (msf->imsf_numsrc) {
newpsl = (struct ip_sf_socklist *)sock_kmalloc(sk,
IP_SFLSIZE(msf->imsf_numsrc), GFP_KERNEL);
@@ -1909,8 +1920,11 @@ int ip_mc_msfilter(struct sock *sk, struct ip_msfilter *msf, int ifindex)
0, NULL, 0);
pmc->sflist = newpsl;
pmc->sfmode = msf->imsf_fmode;
+ err = 0;
done:
rtnl_shunlock();
+ if (leavegroup)
+ err = ip_mc_leave_group(sk, &imr);
return err;
}
diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
index 4e47a2658c7..c703528e0bc 100644
--- a/net/ipv4/ip_input.c
+++ b/net/ipv4/ip_input.c
@@ -184,6 +184,7 @@ int ip_call_ra_chain(struct sk_buff *skb)
raw_rcv(last, skb2);
}
last = sk;
+ nf_reset(skb);
}
}
@@ -200,10 +201,6 @@ static inline int ip_local_deliver_finish(struct sk_buff *skb)
{
int ihl = skb->nh.iph->ihl*4;
-#ifdef CONFIG_NETFILTER_DEBUG
- nf_debug_ip_local_deliver(skb);
-#endif /*CONFIG_NETFILTER_DEBUG*/
-
__skb_pull(skb, ihl);
/* Free reference early: we don't need it any more, and it may
@@ -286,14 +283,18 @@ static inline int ip_rcv_finish(struct sk_buff *skb)
{
struct net_device *dev = skb->dev;
struct iphdr *iph = skb->nh.iph;
+ int err;
/*
* Initialise the virtual path cache for the packet. It describes
* how the packet travels inside Linux networking.
*/
if (skb->dst == NULL) {
- if (ip_route_input(skb, iph->daddr, iph->saddr, iph->tos, dev))
+ if ((err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos, dev))) {
+ if (err == -EHOSTUNREACH)
+ IP_INC_STATS_BH(IPSTATS_MIB_INADDRERRORS);
goto drop;
+ }
}
#ifdef CONFIG_NET_CLS_ROUTE
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 760dc8238d6..80d13103b2b 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -107,11 +107,6 @@ static int ip_dev_loopback_xmit(struct sk_buff *newskb)
newskb->pkt_type = PACKET_LOOPBACK;
newskb->ip_summed = CHECKSUM_UNNECESSARY;
BUG_TRAP(newskb->dst);
-
-#ifdef CONFIG_NETFILTER_DEBUG
- nf_debug_ip_loopback_xmit(newskb);
-#endif
- nf_reset(newskb);
netif_rx(newskb);
return 0;
}
@@ -192,12 +187,6 @@ static inline int ip_finish_output2(struct sk_buff *skb)
skb = skb2;
}
-#ifdef CONFIG_NETFILTER_DEBUG
- nf_debug_ip_finish_output2(skb);
-#endif /*CONFIG_NETFILTER_DEBUG*/
-
- nf_reset(skb);
-
if (hh) {
int hh_alen;
@@ -391,7 +380,6 @@ static void ip_copy_metadata(struct sk_buff *to, struct sk_buff *from)
to->pkt_type = from->pkt_type;
to->priority = from->priority;
to->protocol = from->protocol;
- to->security = from->security;
dst_release(to->dst);
to->dst = dst_clone(from->dst);
to->dev = from->dev;
@@ -415,9 +403,6 @@ static void ip_copy_metadata(struct sk_buff *to, struct sk_buff *from)
to->nf_bridge = from->nf_bridge;
nf_bridge_get(to->nf_bridge);
#endif
-#ifdef CONFIG_NETFILTER_DEBUG
- to->nf_debug = from->nf_debug;
-#endif
#endif
}
@@ -1334,23 +1319,8 @@ void ip_send_reply(struct sock *sk, struct sk_buff *skb, struct ip_reply_arg *ar
ip_rt_put(rt);
}
-/*
- * IP protocol layer initialiser
- */
-
-static struct packet_type ip_packet_type = {
- .type = __constant_htons(ETH_P_IP),
- .func = ip_rcv,
-};
-
-/*
- * IP registers the packet type and then calls the subprotocol initialisers
- */
-
void __init ip_init(void)
{
- dev_add_pack(&ip_packet_type);
-
ip_rt_init();
inet_initpeers();
diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
index 47012b93cad..fc7c481d0d7 100644
--- a/net/ipv4/ip_sockglue.c
+++ b/net/ipv4/ip_sockglue.c
@@ -360,14 +360,14 @@ int ip_recv_error(struct sock *sk, struct msghdr *msg, int len)
err = copied;
/* Reset and regenerate socket error */
- spin_lock_irq(&sk->sk_error_queue.lock);
+ spin_lock_bh(&sk->sk_error_queue.lock);
sk->sk_err = 0;
if ((skb2 = skb_peek(&sk->sk_error_queue)) != NULL) {
sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
- spin_unlock_irq(&sk->sk_error_queue.lock);
+ spin_unlock_bh(&sk->sk_error_queue.lock);
sk->sk_error_report(sk);
} else
- spin_unlock_irq(&sk->sk_error_queue.lock);
+ spin_unlock_bh(&sk->sk_error_queue.lock);
out_free_skb:
kfree_skb(skb);
@@ -677,11 +677,11 @@ int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
mreq.imr_address.s_addr = mreqs.imr_interface;
mreq.imr_ifindex = 0;
err = ip_mc_join_group(sk, &mreq);
- if (err)
+ if (err && err != -EADDRINUSE)
break;
omode = MCAST_INCLUDE;
add = 1;
- } else /*IP_DROP_SOURCE_MEMBERSHIP */ {
+ } else /* IP_DROP_SOURCE_MEMBERSHIP */ {
omode = MCAST_INCLUDE;
add = 0;
}
@@ -754,7 +754,7 @@ int ip_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
mreq.imr_address.s_addr = 0;
mreq.imr_ifindex = greqs.gsr_interface;
err = ip_mc_join_group(sk, &mreq);
- if (err)
+ if (err && err != -EADDRINUSE)
break;
greqs.gsr_interface = mreq.imr_ifindex;
omode = MCAST_INCLUDE;
diff --git a/net/ipv4/ipcomp.c b/net/ipv4/ipcomp.c
index 1a23c5263b9..2065944fd9e 100644
--- a/net/ipv4/ipcomp.c
+++ b/net/ipv4/ipcomp.c
@@ -236,15 +236,10 @@ static struct xfrm_state *ipcomp_tunnel_create(struct xfrm_state *x)
t->props.mode = 1;
t->props.saddr.a4 = x->props.saddr.a4;
t->props.flags = x->props.flags;
-
- t->type = xfrm_get_type(IPPROTO_IPIP, t->props.family);
- if (t->type == NULL)
- goto error;
-
- if (t->type->init_state(t, NULL))
+
+ if (xfrm_init_state(t))
goto error;
- t->km.state = XFRM_STATE_VALID;
atomic_set(&t->tunnel_users, 1);
out:
return t;
@@ -422,7 +417,7 @@ static void ipcomp_destroy(struct xfrm_state *x)
kfree(ipcd);
}
-static int ipcomp_init_state(struct xfrm_state *x, void *args)
+static int ipcomp_init_state(struct xfrm_state *x)
{
int err;
struct ipcomp_data *ipcd;
diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c
index f2509034ce7..d2bf8e1930a 100644
--- a/net/ipv4/ipconfig.c
+++ b/net/ipv4/ipconfig.c
@@ -1149,8 +1149,10 @@ static int __init ic_dynamic(void)
ic_rarp_cleanup();
#endif
- if (!ic_got_reply)
+ if (!ic_got_reply) {
+ ic_myaddr = INADDR_NONE;
return -1;
+ }
printk("IP-Config: Got %s answer from %u.%u.%u.%u, ",
((ic_got_reply & IC_RARP) ? "RARP"
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index e21c049ec62..7833d920bdb 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -297,6 +297,7 @@ static int vif_delete(int vifi)
static void ipmr_destroy_unres(struct mfc_cache *c)
{
struct sk_buff *skb;
+ struct nlmsgerr *e;
atomic_dec(&cache_resolve_queue_len);
@@ -306,7 +307,9 @@ static void ipmr_destroy_unres(struct mfc_cache *c)
nlh->nlmsg_type = NLMSG_ERROR;
nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
skb_trim(skb, nlh->nlmsg_len);
- ((struct nlmsgerr*)NLMSG_DATA(nlh))->error = -ETIMEDOUT;
+ e = NLMSG_DATA(nlh);
+ e->error = -ETIMEDOUT;
+ memset(&e->msg, 0, sizeof(e->msg));
netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
} else
kfree_skb(skb);
@@ -499,6 +502,7 @@ static struct mfc_cache *ipmr_cache_alloc_unres(void)
static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
{
struct sk_buff *skb;
+ struct nlmsgerr *e;
/*
* Play the pending entries through our router
@@ -515,7 +519,9 @@ static void ipmr_cache_resolve(struct mfc_cache *uc, struct mfc_cache *c)
nlh->nlmsg_type = NLMSG_ERROR;
nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
skb_trim(skb, nlh->nlmsg_len);
- ((struct nlmsgerr*)NLMSG_DATA(nlh))->error = -EMSGSIZE;
+ e = NLMSG_DATA(nlh);
+ e->error = -EMSGSIZE;
+ memset(&e->msg, 0, sizeof(e->msg));
}
err = netlink_unicast(rtnl, skb, NETLINK_CB(skb).dst_pid, MSG_DONTWAIT);
} else
@@ -1350,6 +1356,7 @@ int ip_mr_input(struct sk_buff *skb)
*/
read_lock(&mrt_lock);
if (mroute_socket) {
+ nf_reset(skb);
raw_rcv(mroute_socket, skb);
read_unlock(&mrt_lock);
return 0;
diff --git a/net/ipv4/ipvs/Kconfig b/net/ipv4/ipvs/Kconfig
index 63a82b4b64b..c9820bfc493 100644
--- a/net/ipv4/ipvs/Kconfig
+++ b/net/ipv4/ipvs/Kconfig
@@ -2,11 +2,11 @@
# IP Virtual Server configuration
#
menu "IP: Virtual Server Configuration"
- depends on INET && NETFILTER
+ depends on NETFILTER
config IP_VS
tristate "IP virtual server support (EXPERIMENTAL)"
- depends on INET && NETFILTER
+ depends on NETFILTER
---help---
IP Virtual Server support will let you build a high-performance
virtual server based on cluster of two or more real servers. This
diff --git a/net/ipv4/ipvs/ip_vs_conn.c b/net/ipv4/ipvs/ip_vs_conn.c
index fd6feb5499f..d0145a8b155 100644
--- a/net/ipv4/ipvs/ip_vs_conn.c
+++ b/net/ipv4/ipvs/ip_vs_conn.c
@@ -548,7 +548,6 @@ void ip_vs_conn_expire_now(struct ip_vs_conn *cp)
{
if (del_timer(&cp->timer))
mod_timer(&cp->timer, jiffies);
- __ip_vs_conn_put(cp);
}
@@ -759,12 +758,11 @@ static inline int todrop_entry(struct ip_vs_conn *cp)
return 1;
}
-
+/* Called from keventd and must protect itself from softirqs */
void ip_vs_random_dropentry(void)
{
int idx;
struct ip_vs_conn *cp;
- struct ip_vs_conn *ct;
/*
* Randomly scan 1/32 of the whole table every second
@@ -775,7 +773,7 @@ void ip_vs_random_dropentry(void)
/*
* Lock is actually needed in this loop.
*/
- ct_write_lock(hash);
+ ct_write_lock_bh(hash);
list_for_each_entry(cp, &ip_vs_conn_tab[hash], c_list) {
if (!cp->cport && !(cp->flags & IP_VS_CONN_F_NO_CPORT))
@@ -801,23 +799,14 @@ void ip_vs_random_dropentry(void)
continue;
}
- /*
- * Drop the entry, and drop its ct if not referenced
- */
- atomic_inc(&cp->refcnt);
- ct_write_unlock(hash);
-
- if ((ct = cp->control))
- atomic_inc(&ct->refcnt);
IP_VS_DBG(4, "del connection\n");
ip_vs_conn_expire_now(cp);
- if (ct) {
+ if (cp->control) {
IP_VS_DBG(4, "del conn template\n");
- ip_vs_conn_expire_now(ct);
+ ip_vs_conn_expire_now(cp->control);
}
- ct_write_lock(hash);
}
- ct_write_unlock(hash);
+ ct_write_unlock_bh(hash);
}
}
@@ -829,7 +818,6 @@ static void ip_vs_conn_flush(void)
{
int idx;
struct ip_vs_conn *cp;
- struct ip_vs_conn *ct;
flush_again:
for (idx=0; idx<IP_VS_CONN_TAB_SIZE; idx++) {
@@ -839,18 +827,13 @@ static void ip_vs_conn_flush(void)
ct_write_lock_bh(idx);
list_for_each_entry(cp, &ip_vs_conn_tab[idx], c_list) {
- atomic_inc(&cp->refcnt);
- ct_write_unlock(idx);
- if ((ct = cp->control))
- atomic_inc(&ct->refcnt);
IP_VS_DBG(4, "del connection\n");
ip_vs_conn_expire_now(cp);
- if (ct) {
+ if (cp->control) {
IP_VS_DBG(4, "del conn template\n");
- ip_vs_conn_expire_now(ct);
+ ip_vs_conn_expire_now(cp->control);
}
- ct_write_lock(idx);
}
ct_write_unlock_bh(idx);
}
diff --git a/net/ipv4/ipvs/ip_vs_ctl.c b/net/ipv4/ipvs/ip_vs_ctl.c
index 218d9701036..7d99ede2ef7 100644
--- a/net/ipv4/ipvs/ip_vs_ctl.c
+++ b/net/ipv4/ipvs/ip_vs_ctl.c
@@ -90,7 +90,8 @@ int ip_vs_get_debug_level(void)
#endif
/*
- * update_defense_level is called from keventd and from sysctl.
+ * update_defense_level is called from keventd and from sysctl,
+ * so it needs to protect itself from softirqs
*/
static void update_defense_level(void)
{
@@ -110,6 +111,8 @@ static void update_defense_level(void)
nomem = (availmem < sysctl_ip_vs_amemthresh);
+ local_bh_disable();
+
/* drop_entry */
spin_lock(&__ip_vs_dropentry_lock);
switch (sysctl_ip_vs_drop_entry) {
@@ -206,6 +209,8 @@ static void update_defense_level(void)
if (to_change >= 0)
ip_vs_protocol_timeout_change(sysctl_ip_vs_secure_tcp>1);
write_unlock(&__ip_vs_securetcp_lock);
+
+ local_bh_enable();
}
@@ -1360,9 +1365,7 @@ proc_do_defense_mode(ctl_table *table, int write, struct file * filp,
/* Restore the correct value */
*valp = val;
} else {
- local_bh_disable();
update_defense_level();
- local_bh_enable();
}
}
return rc;
@@ -2059,7 +2062,7 @@ ip_vs_copy_service(struct ip_vs_service_entry *dst, struct ip_vs_service *src)
dst->addr = src->addr;
dst->port = src->port;
dst->fwmark = src->fwmark;
- strcpy(dst->sched_name, src->scheduler->name);
+ strlcpy(dst->sched_name, src->scheduler->name, sizeof(dst->sched_name));
dst->flags = src->flags;
dst->timeout = src->timeout / HZ;
dst->netmask = src->netmask;
@@ -2080,6 +2083,7 @@ __ip_vs_get_service_entries(const struct ip_vs_get_services *get,
list_for_each_entry(svc, &ip_vs_svc_table[idx], s_list) {
if (count >= get->num_services)
goto out;
+ memset(&entry, 0, sizeof(entry));
ip_vs_copy_service(&entry, svc);
if (copy_to_user(&uptr->entrytable[count],
&entry, sizeof(entry))) {
@@ -2094,6 +2098,7 @@ __ip_vs_get_service_entries(const struct ip_vs_get_services *get,
list_for_each_entry(svc, &ip_vs_svc_fwm_table[idx], f_list) {
if (count >= get->num_services)
goto out;
+ memset(&entry, 0, sizeof(entry));
ip_vs_copy_service(&entry, svc);
if (copy_to_user(&uptr->entrytable[count],
&entry, sizeof(entry))) {
@@ -2304,12 +2309,12 @@ do_ip_vs_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
memset(&d, 0, sizeof(d));
if (ip_vs_sync_state & IP_VS_STATE_MASTER) {
d[0].state = IP_VS_STATE_MASTER;
- strcpy(d[0].mcast_ifn, ip_vs_master_mcast_ifn);
+ strlcpy(d[0].mcast_ifn, ip_vs_master_mcast_ifn, sizeof(d[0].mcast_ifn));
d[0].syncid = ip_vs_master_syncid;
}
if (ip_vs_sync_state & IP_VS_STATE_BACKUP) {
d[1].state = IP_VS_STATE_BACKUP;
- strcpy(d[1].mcast_ifn, ip_vs_backup_mcast_ifn);
+ strlcpy(d[1].mcast_ifn, ip_vs_backup_mcast_ifn, sizeof(d[1].mcast_ifn));
d[1].syncid = ip_vs_backup_syncid;
}
if (copy_to_user(user, &d, sizeof(d)) != 0)
diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/ipv4/ipvs/ip_vs_sync.c
index 25c479550a3..574d1f509b4 100644
--- a/net/ipv4/ipvs/ip_vs_sync.c
+++ b/net/ipv4/ipvs/ip_vs_sync.c
@@ -839,10 +839,10 @@ int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
ip_vs_sync_state |= state;
if (state == IP_VS_STATE_MASTER) {
- strcpy(ip_vs_master_mcast_ifn, mcast_ifn);
+ strlcpy(ip_vs_master_mcast_ifn, mcast_ifn, sizeof(ip_vs_master_mcast_ifn));
ip_vs_master_syncid = syncid;
} else {
- strcpy(ip_vs_backup_mcast_ifn, mcast_ifn);
+ strlcpy(ip_vs_backup_mcast_ifn, mcast_ifn, sizeof(ip_vs_backup_mcast_ifn));
ip_vs_backup_syncid = syncid;
}
diff --git a/net/ipv4/ipvs/ip_vs_xmit.c b/net/ipv4/ipvs/ip_vs_xmit.c
index de21da00057..a8512a3fd08 100644
--- a/net/ipv4/ipvs/ip_vs_xmit.c
+++ b/net/ipv4/ipvs/ip_vs_xmit.c
@@ -127,7 +127,6 @@ ip_vs_dst_reset(struct ip_vs_dest *dest)
#define IP_VS_XMIT(skb, rt) \
do { \
- nf_reset_debug(skb); \
(skb)->nfcache |= NFC_IPVS_PROPERTY; \
(skb)->ip_summed = CHECKSUM_NONE; \
NF_HOOK(PF_INET, NF_IP_LOCAL_OUT, (skb), NULL, \
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index df79f5ed6a0..fa163425668 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -60,7 +60,6 @@ static DECLARE_MUTEX(arpt_mutex);
#define ASSERT_READ_LOCK(x) ARP_NF_ASSERT(down_trylock(&arpt_mutex) != 0)
#define ASSERT_WRITE_LOCK(x) ARP_NF_ASSERT(down_trylock(&arpt_mutex) != 0)
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <linux/netfilter_ipv4/listhelp.h>
struct arpt_table_info {
diff --git a/net/ipv4/netfilter/ip_conntrack_amanda.c b/net/ipv4/netfilter/ip_conntrack_amanda.c
index 3dbddd06260..a78a320eee0 100644
--- a/net/ipv4/netfilter/ip_conntrack_amanda.c
+++ b/net/ipv4/netfilter/ip_conntrack_amanda.c
@@ -26,7 +26,6 @@
#include <net/checksum.h>
#include <net/udp.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <linux/netfilter_ipv4/ip_conntrack_helper.h>
#include <linux/netfilter_ipv4/ip_conntrack_amanda.h>
@@ -42,7 +41,7 @@ static char *conns[] = { "DATA ", "MESG ", "INDEX " };
/* This is slow, but it's simple. --RR */
static char amanda_buffer[65536];
-static DECLARE_LOCK(amanda_buffer_lock);
+static DEFINE_SPINLOCK(amanda_buffer_lock);
unsigned int (*ip_nat_amanda_hook)(struct sk_buff **pskb,
enum ip_conntrack_info ctinfo,
@@ -76,7 +75,7 @@ static int help(struct sk_buff **pskb,
return NF_ACCEPT;
}
- LOCK_BH(&amanda_buffer_lock);
+ spin_lock_bh(&amanda_buffer_lock);
skb_copy_bits(*pskb, dataoff, amanda_buffer, (*pskb)->len - dataoff);
data = amanda_buffer;
data_limit = amanda_buffer + (*pskb)->len - dataoff;
@@ -134,7 +133,7 @@ static int help(struct sk_buff **pskb,
}
out:
- UNLOCK_BH(&amanda_buffer_lock);
+ spin_unlock_bh(&amanda_buffer_lock);
return ret;
}
diff --git a/net/ipv4/netfilter/ip_conntrack_core.c b/net/ipv4/netfilter/ip_conntrack_core.c
index 09e82462297..4b78ebeb663 100644
--- a/net/ipv4/netfilter/ip_conntrack_core.c
+++ b/net/ipv4/netfilter/ip_conntrack_core.c
@@ -38,10 +38,10 @@
#include <linux/percpu.h>
#include <linux/moduleparam.h>
-/* This rwlock protects the main hash table, protocol/helper/expected
+/* ip_conntrack_lock protects the main hash table, protocol/helper/expected
registrations, conntrack timers*/
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_conntrack_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_conntrack_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
@@ -57,7 +57,7 @@
#define DEBUGP(format, args...)
#endif
-DECLARE_RWLOCK(ip_conntrack_lock);
+DEFINE_RWLOCK(ip_conntrack_lock);
/* ip_conntrack_standalone needs this */
atomic_t ip_conntrack_count = ATOMIC_INIT(0);
@@ -147,7 +147,7 @@ static void destroy_expect(struct ip_conntrack_expect *exp)
static void unlink_expect(struct ip_conntrack_expect *exp)
{
- MUST_BE_WRITE_LOCKED(&ip_conntrack_lock);
+ ASSERT_WRITE_LOCK(&ip_conntrack_lock);
list_del(&exp->list);
/* Logically in destroy_expect, but we hold the lock here. */
exp->master->expecting--;
@@ -157,9 +157,9 @@ static void expectation_timed_out(unsigned long ul_expect)
{
struct ip_conntrack_expect *exp = (void *)ul_expect;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
unlink_expect(exp);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
destroy_expect(exp);
}
@@ -209,7 +209,7 @@ clean_from_lists(struct ip_conntrack *ct)
unsigned int ho, hr;
DEBUGP("clean_from_lists(%p)\n", ct);
- MUST_BE_WRITE_LOCKED(&ip_conntrack_lock);
+ ASSERT_WRITE_LOCK(&ip_conntrack_lock);
ho = hash_conntrack(&ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
hr = hash_conntrack(&ct->tuplehash[IP_CT_DIR_REPLY].tuple);
@@ -240,7 +240,7 @@ destroy_conntrack(struct nf_conntrack *nfct)
if (ip_conntrack_destroyed)
ip_conntrack_destroyed(ct);
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* Expectations will have been removed in clean_from_lists,
* except TFTP can create an expectation on the first packet,
* before connection is in the list, so we need to clean here,
@@ -254,7 +254,7 @@ destroy_conntrack(struct nf_conntrack *nfct)
}
CONNTRACK_STAT_INC(delete);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
if (ct->master)
ip_conntrack_put(ct->master);
@@ -268,12 +268,12 @@ static void death_by_timeout(unsigned long ul_conntrack)
{
struct ip_conntrack *ct = (void *)ul_conntrack;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* Inside lock so preempt is disabled on module removal path.
* Otherwise we can get spurious warnings. */
CONNTRACK_STAT_INC(delete_list);
clean_from_lists(ct);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
ip_conntrack_put(ct);
}
@@ -282,7 +282,7 @@ conntrack_tuple_cmp(const struct ip_conntrack_tuple_hash *i,
const struct ip_conntrack_tuple *tuple,
const struct ip_conntrack *ignored_conntrack)
{
- MUST_BE_READ_LOCKED(&ip_conntrack_lock);
+ ASSERT_READ_LOCK(&ip_conntrack_lock);
return tuplehash_to_ctrack(i) != ignored_conntrack
&& ip_ct_tuple_equal(tuple, &i->tuple);
}
@@ -294,7 +294,7 @@ __ip_conntrack_find(const struct ip_conntrack_tuple *tuple,
struct ip_conntrack_tuple_hash *h;
unsigned int hash = hash_conntrack(tuple);
- MUST_BE_READ_LOCKED(&ip_conntrack_lock);
+ ASSERT_READ_LOCK(&ip_conntrack_lock);
list_for_each_entry(h, &ip_conntrack_hash[hash], list) {
if (conntrack_tuple_cmp(h, tuple, ignored_conntrack)) {
CONNTRACK_STAT_INC(found);
@@ -313,11 +313,11 @@ ip_conntrack_find_get(const struct ip_conntrack_tuple *tuple,
{
struct ip_conntrack_tuple_hash *h;
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
h = __ip_conntrack_find(tuple, ignored_conntrack);
if (h)
atomic_inc(&tuplehash_to_ctrack(h)->ct_general.use);
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
return h;
}
@@ -352,7 +352,7 @@ __ip_conntrack_confirm(struct sk_buff **pskb)
IP_NF_ASSERT(!is_confirmed(ct));
DEBUGP("Confirming conntrack %p\n", ct);
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* See if there's one in the list already, including reverse:
NAT could have grabbed it without realizing, since we're
@@ -380,12 +380,12 @@ __ip_conntrack_confirm(struct sk_buff **pskb)
atomic_inc(&ct->ct_general.use);
set_bit(IPS_CONFIRMED_BIT, &ct->status);
CONNTRACK_STAT_INC(insert);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return NF_ACCEPT;
}
CONNTRACK_STAT_INC(insert_failed);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return NF_DROP;
}
@@ -398,9 +398,9 @@ ip_conntrack_tuple_taken(const struct ip_conntrack_tuple *tuple,
{
struct ip_conntrack_tuple_hash *h;
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
h = __ip_conntrack_find(tuple, ignored_conntrack);
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
return h != NULL;
}
@@ -419,13 +419,13 @@ static int early_drop(struct list_head *chain)
struct ip_conntrack *ct = NULL;
int dropped = 0;
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
h = LIST_FIND_B(chain, unreplied, struct ip_conntrack_tuple_hash *);
if (h) {
ct = tuplehash_to_ctrack(h);
atomic_inc(&ct->ct_general.use);
}
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
if (!ct)
return dropped;
@@ -508,7 +508,7 @@ init_conntrack(const struct ip_conntrack_tuple *tuple,
conntrack->timeout.data = (unsigned long)conntrack;
conntrack->timeout.function = death_by_timeout;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
exp = find_expectation(tuple);
if (exp) {
@@ -532,7 +532,7 @@ init_conntrack(const struct ip_conntrack_tuple *tuple,
list_add(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL].list, &unconfirmed);
atomic_inc(&ip_conntrack_count);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
if (exp) {
if (exp->expectfn)
@@ -723,17 +723,17 @@ void ip_conntrack_unexpect_related(struct ip_conntrack_expect *exp)
{
struct ip_conntrack_expect *i;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* choose the the oldest expectation to evict */
list_for_each_entry_reverse(i, &ip_conntrack_expect_list, list) {
if (expect_matches(i, exp) && del_timer(&i->timeout)) {
unlink_expect(i);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
destroy_expect(i);
return;
}
}
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
}
struct ip_conntrack_expect *ip_conntrack_expect_alloc(void)
@@ -760,15 +760,11 @@ static void ip_conntrack_expect_insert(struct ip_conntrack_expect *exp)
exp->master->expecting++;
list_add(&exp->list, &ip_conntrack_expect_list);
- if (exp->master->helper->timeout) {
- init_timer(&exp->timeout);
- exp->timeout.data = (unsigned long)exp;
- exp->timeout.function = expectation_timed_out;
- exp->timeout.expires
- = jiffies + exp->master->helper->timeout * HZ;
- add_timer(&exp->timeout);
- } else
- exp->timeout.function = NULL;
+ init_timer(&exp->timeout);
+ exp->timeout.data = (unsigned long)exp;
+ exp->timeout.function = expectation_timed_out;
+ exp->timeout.expires = jiffies + exp->master->helper->timeout * HZ;
+ add_timer(&exp->timeout);
CONNTRACK_STAT_INC(expect_create);
}
@@ -808,7 +804,7 @@ int ip_conntrack_expect_related(struct ip_conntrack_expect *expect)
DEBUGP("tuple: "); DUMP_TUPLE(&expect->tuple);
DEBUGP("mask: "); DUMP_TUPLE(&expect->mask);
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
list_for_each_entry(i, &ip_conntrack_expect_list, list) {
if (expect_matches(i, expect)) {
/* Refresh timer: if it's dying, ignore.. */
@@ -832,7 +828,7 @@ int ip_conntrack_expect_related(struct ip_conntrack_expect *expect)
ip_conntrack_expect_insert(expect);
ret = 0;
out:
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return ret;
}
@@ -841,7 +837,7 @@ out:
void ip_conntrack_alter_reply(struct ip_conntrack *conntrack,
const struct ip_conntrack_tuple *newreply)
{
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* Should be unconfirmed, so not in hash table yet */
IP_NF_ASSERT(!is_confirmed(conntrack));
@@ -851,15 +847,15 @@ void ip_conntrack_alter_reply(struct ip_conntrack *conntrack,
conntrack->tuplehash[IP_CT_DIR_REPLY].tuple = *newreply;
if (!conntrack->master && conntrack->expecting == 0)
conntrack->helper = ip_ct_find_helper(newreply);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
}
int ip_conntrack_helper_register(struct ip_conntrack_helper *me)
{
BUG_ON(me->timeout == 0);
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
list_prepend(&helpers, me);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return 0;
}
@@ -878,7 +874,7 @@ void ip_conntrack_helper_unregister(struct ip_conntrack_helper *me)
struct ip_conntrack_expect *exp, *tmp;
/* Need write lock here, to delete helper. */
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
LIST_DELETE(&helpers, me);
/* Get rid of expectations */
@@ -893,7 +889,7 @@ void ip_conntrack_helper_unregister(struct ip_conntrack_helper *me)
for (i = 0; i < ip_conntrack_htable_size; i++)
LIST_FIND_W(&ip_conntrack_hash[i], unhelp,
struct ip_conntrack_tuple_hash *, me);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
/* Someone could be still looking at the helper in a bh. */
synchronize_net();
@@ -925,14 +921,14 @@ void ip_ct_refresh_acct(struct ip_conntrack *ct,
ct->timeout.expires = extra_jiffies;
ct_add_counters(ct, ctinfo, skb);
} else {
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
/* Need del_timer for race avoidance (may already be dying). */
if (del_timer(&ct->timeout)) {
ct->timeout.expires = jiffies + extra_jiffies;
add_timer(&ct->timeout);
}
ct_add_counters(ct, ctinfo, skb);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
}
}
@@ -940,10 +936,6 @@ void ip_ct_refresh_acct(struct ip_conntrack *ct,
struct sk_buff *
ip_ct_gather_frags(struct sk_buff *skb, u_int32_t user)
{
-#ifdef CONFIG_NETFILTER_DEBUG
- unsigned int olddebug = skb->nf_debug;
-#endif
-
skb_orphan(skb);
local_bh_disable();
@@ -953,12 +945,7 @@ ip_ct_gather_frags(struct sk_buff *skb, u_int32_t user)
if (skb) {
ip_send_check(skb->nh.iph);
skb->nfcache |= NFC_ALTERED;
-#ifdef CONFIG_NETFILTER_DEBUG
- /* Packet path as if nothing had happened. */
- skb->nf_debug = olddebug;
-#endif
}
-
return skb;
}
@@ -997,7 +984,7 @@ get_next_corpse(int (*iter)(struct ip_conntrack *i, void *data),
{
struct ip_conntrack_tuple_hash *h = NULL;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
for (; *bucket < ip_conntrack_htable_size; (*bucket)++) {
h = LIST_FIND_W(&ip_conntrack_hash[*bucket], do_iter,
struct ip_conntrack_tuple_hash *, iter, data);
@@ -1009,7 +996,7 @@ get_next_corpse(int (*iter)(struct ip_conntrack *i, void *data),
struct ip_conntrack_tuple_hash *, iter, data);
if (h)
atomic_inc(&tuplehash_to_ctrack(h)->ct_general.use);
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return h;
}
@@ -1201,14 +1188,14 @@ int __init ip_conntrack_init(void)
}
/* Don't NEED lock here, but good form anyway. */
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
for (i = 0; i < MAX_IP_CT_PROTO; i++)
ip_ct_protos[i] = &ip_conntrack_generic_protocol;
/* Sew in builtin protocols. */
ip_ct_protos[IPPROTO_TCP] = &ip_conntrack_protocol_tcp;
ip_ct_protos[IPPROTO_UDP] = &ip_conntrack_protocol_udp;
ip_ct_protos[IPPROTO_ICMP] = &ip_conntrack_protocol_icmp;
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
for (i = 0; i < ip_conntrack_htable_size; i++)
INIT_LIST_HEAD(&ip_conntrack_hash[i]);
diff --git a/net/ipv4/netfilter/ip_conntrack_ftp.c b/net/ipv4/netfilter/ip_conntrack_ftp.c
index dd86503aa78..fea6dd2a00b 100644
--- a/net/ipv4/netfilter/ip_conntrack_ftp.c
+++ b/net/ipv4/netfilter/ip_conntrack_ftp.c
@@ -16,7 +16,6 @@
#include <net/checksum.h>
#include <net/tcp.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <linux/netfilter_ipv4/ip_conntrack_helper.h>
#include <linux/netfilter_ipv4/ip_conntrack_ftp.h>
#include <linux/moduleparam.h>
@@ -28,7 +27,7 @@ MODULE_DESCRIPTION("ftp connection tracking helper");
/* This is slow, but it's simple. --RR */
static char ftp_buffer[65536];
-static DECLARE_LOCK(ip_ftp_lock);
+static DEFINE_SPINLOCK(ip_ftp_lock);
#define MAX_PORTS 8
static int ports[MAX_PORTS];
@@ -319,7 +318,7 @@ static int help(struct sk_buff **pskb,
}
datalen = (*pskb)->len - dataoff;
- LOCK_BH(&ip_ftp_lock);
+ spin_lock_bh(&ip_ftp_lock);
fb_ptr = skb_header_pointer(*pskb, dataoff,
(*pskb)->len - dataoff, ftp_buffer);
BUG_ON(fb_ptr == NULL);
@@ -442,7 +441,7 @@ out_update_nl:
if (ends_in_nl)
update_nl_seq(seq, ct_ftp_info,dir);
out:
- UNLOCK_BH(&ip_ftp_lock);
+ spin_unlock_bh(&ip_ftp_lock);
return ret;
}
diff --git a/net/ipv4/netfilter/ip_conntrack_irc.c b/net/ipv4/netfilter/ip_conntrack_irc.c
index 33cc7348b6e..cd98772cc33 100644
--- a/net/ipv4/netfilter/ip_conntrack_irc.c
+++ b/net/ipv4/netfilter/ip_conntrack_irc.c
@@ -29,7 +29,6 @@
#include <net/checksum.h>
#include <net/tcp.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <linux/netfilter_ipv4/ip_conntrack_helper.h>
#include <linux/netfilter_ipv4/ip_conntrack_irc.h>
#include <linux/moduleparam.h>
@@ -41,7 +40,7 @@ static int max_dcc_channels = 8;
static unsigned int dcc_timeout = 300;
/* This is slow, but it's simple. --RR */
static char irc_buffer[65536];
-static DECLARE_LOCK(irc_buffer_lock);
+static DEFINE_SPINLOCK(irc_buffer_lock);
unsigned int (*ip_nat_irc_hook)(struct sk_buff **pskb,
enum ip_conntrack_info ctinfo,
@@ -141,7 +140,7 @@ static int help(struct sk_buff **pskb,
if (dataoff >= (*pskb)->len)
return NF_ACCEPT;
- LOCK_BH(&irc_buffer_lock);
+ spin_lock_bh(&irc_buffer_lock);
ib_ptr = skb_header_pointer(*pskb, dataoff,
(*pskb)->len - dataoff, irc_buffer);
BUG_ON(ib_ptr == NULL);
@@ -237,7 +236,7 @@ static int help(struct sk_buff **pskb,
} /* while data < ... */
out:
- UNLOCK_BH(&irc_buffer_lock);
+ spin_unlock_bh(&irc_buffer_lock);
return ret;
}
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
index ff8c34a860f..31d75390bf1 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_sctp.c
@@ -26,7 +26,6 @@
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#if 0
#define DEBUGP(format, ...) printk(format, ## __VA_ARGS__)
@@ -35,7 +34,7 @@
#endif
/* Protects conntrack->proto.sctp */
-static DECLARE_RWLOCK(sctp_lock);
+static DEFINE_RWLOCK(sctp_lock);
/* FIXME: Examine ipfilter's timeouts and conntrack transitions more
closely. They're more complex. --RR
@@ -199,9 +198,9 @@ static int sctp_print_conntrack(struct seq_file *s,
DEBUGP(__FUNCTION__);
DEBUGP("\n");
- READ_LOCK(&sctp_lock);
+ read_lock_bh(&sctp_lock);
state = conntrack->proto.sctp.state;
- READ_UNLOCK(&sctp_lock);
+ read_unlock_bh(&sctp_lock);
return seq_printf(s, "%s ", sctp_conntrack_names[state]);
}
@@ -343,13 +342,13 @@ static int sctp_packet(struct ip_conntrack *conntrack,
oldsctpstate = newconntrack = SCTP_CONNTRACK_MAX;
for_each_sctp_chunk (skb, sch, _sch, offset, count) {
- WRITE_LOCK(&sctp_lock);
+ write_lock_bh(&sctp_lock);
/* Special cases of Verification tag check (Sec 8.5.1) */
if (sch->type == SCTP_CID_INIT) {
/* Sec 8.5.1 (A) */
if (sh->vtag != 0) {
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
} else if (sch->type == SCTP_CID_ABORT) {
@@ -357,7 +356,7 @@ static int sctp_packet(struct ip_conntrack *conntrack,
if (!(sh->vtag == conntrack->proto.sctp.vtag[CTINFO2DIR(ctinfo)])
&& !(sh->vtag == conntrack->proto.sctp.vtag
[1 - CTINFO2DIR(ctinfo)])) {
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
} else if (sch->type == SCTP_CID_SHUTDOWN_COMPLETE) {
@@ -366,13 +365,13 @@ static int sctp_packet(struct ip_conntrack *conntrack,
&& !(sh->vtag == conntrack->proto.sctp.vtag
[1 - CTINFO2DIR(ctinfo)]
&& (sch->flags & 1))) {
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
} else if (sch->type == SCTP_CID_COOKIE_ECHO) {
/* Sec 8.5.1 (D) */
if (!(sh->vtag == conntrack->proto.sctp.vtag[CTINFO2DIR(ctinfo)])) {
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
}
@@ -384,7 +383,7 @@ static int sctp_packet(struct ip_conntrack *conntrack,
if (newconntrack == SCTP_CONNTRACK_MAX) {
DEBUGP("ip_conntrack_sctp: Invalid dir=%i ctype=%u conntrack=%u\n",
CTINFO2DIR(ctinfo), sch->type, oldsctpstate);
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
@@ -396,7 +395,7 @@ static int sctp_packet(struct ip_conntrack *conntrack,
ih = skb_header_pointer(skb, offset + sizeof(sctp_chunkhdr_t),
sizeof(_inithdr), &_inithdr);
if (ih == NULL) {
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
return -1;
}
DEBUGP("Setting vtag %x for dir %d\n",
@@ -405,7 +404,7 @@ static int sctp_packet(struct ip_conntrack *conntrack,
}
conntrack->proto.sctp.state = newconntrack;
- WRITE_UNLOCK(&sctp_lock);
+ write_unlock_bh(&sctp_lock);
}
ip_ct_refresh_acct(conntrack, ctinfo, skb, *sctp_timeouts[newconntrack]);
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_tcp.c b/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
index 721ddbf522b..809dfed766d 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_tcp.c
@@ -36,7 +36,6 @@
#include <linux/netfilter_ipv4.h>
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#if 0
#define DEBUGP printk
@@ -46,7 +45,7 @@
#endif
/* Protects conntrack->proto.tcp */
-static DECLARE_RWLOCK(tcp_lock);
+static DEFINE_RWLOCK(tcp_lock);
/* "Be conservative in what you do,
be liberal in what you accept from others."
@@ -330,9 +329,9 @@ static int tcp_print_conntrack(struct seq_file *s,
{
enum tcp_conntrack state;
- READ_LOCK(&tcp_lock);
+ read_lock_bh(&tcp_lock);
state = conntrack->proto.tcp.state;
- READ_UNLOCK(&tcp_lock);
+ read_unlock_bh(&tcp_lock);
return seq_printf(s, "%s ", tcp_conntrack_names[state]);
}
@@ -738,14 +737,14 @@ void ip_conntrack_tcp_update(struct sk_buff *skb,
end = segment_seq_plus_len(ntohl(tcph->seq), skb->len, iph, tcph);
- WRITE_LOCK(&tcp_lock);
+ write_lock_bh(&tcp_lock);
/*
* We have to worry for the ack in the reply packet only...
*/
if (after(end, conntrack->proto.tcp.seen[dir].td_end))
conntrack->proto.tcp.seen[dir].td_end = end;
conntrack->proto.tcp.last_end = end;
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
DEBUGP("tcp_update: sender end=%u maxend=%u maxwin=%u scale=%i "
"receiver end=%u maxend=%u maxwin=%u scale=%i\n",
sender->td_end, sender->td_maxend, sender->td_maxwin,
@@ -857,7 +856,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
sizeof(_tcph), &_tcph);
BUG_ON(th == NULL);
- WRITE_LOCK(&tcp_lock);
+ write_lock_bh(&tcp_lock);
old_state = conntrack->proto.tcp.state;
dir = CTINFO2DIR(ctinfo);
index = get_conntrack_index(th);
@@ -879,7 +878,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
* that the client cannot but retransmit its SYN and
* thus initiate a clean new session.
*/
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (LOG_INVALID(IPPROTO_TCP))
nf_log_packet(PF_INET, 0, skb, NULL, NULL,
"ip_ct_tcp: killing out of sync session ");
@@ -894,7 +893,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
conntrack->proto.tcp.last_end =
segment_seq_plus_len(ntohl(th->seq), skb->len, iph, th);
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (LOG_INVALID(IPPROTO_TCP))
nf_log_packet(PF_INET, 0, skb, NULL, NULL,
"ip_ct_tcp: invalid packet ignored ");
@@ -904,7 +903,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
DEBUGP("ip_ct_tcp: Invalid dir=%i index=%u ostate=%u\n",
dir, get_conntrack_index(th),
old_state);
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (LOG_INVALID(IPPROTO_TCP))
nf_log_packet(PF_INET, 0, skb, NULL, NULL,
"ip_ct_tcp: invalid state ");
@@ -918,13 +917,13 @@ static int tcp_packet(struct ip_conntrack *conntrack,
conntrack->proto.tcp.seen[dir].td_end)) {
/* Attempt to reopen a closed connection.
* Delete this connection and look up again. */
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (del_timer(&conntrack->timeout))
conntrack->timeout.function((unsigned long)
conntrack);
return -NF_REPEAT;
} else {
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (LOG_INVALID(IPPROTO_TCP))
nf_log_packet(PF_INET, 0, skb, NULL, NULL,
"ip_ct_tcp: invalid SYN");
@@ -949,7 +948,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
if (!tcp_in_window(&conntrack->proto.tcp, dir, index,
skb, iph, th)) {
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
return -NF_ACCEPT;
}
in_window:
@@ -972,7 +971,7 @@ static int tcp_packet(struct ip_conntrack *conntrack,
timeout = conntrack->proto.tcp.retrans >= ip_ct_tcp_max_retrans
&& *tcp_timeouts[new_state] > ip_ct_tcp_timeout_max_retrans
? ip_ct_tcp_timeout_max_retrans : *tcp_timeouts[new_state];
- WRITE_UNLOCK(&tcp_lock);
+ write_unlock_bh(&tcp_lock);
if (!test_bit(IPS_SEEN_REPLY_BIT, &conntrack->status)) {
/* If only reply is a RST, we can consider ourselves not to
diff --git a/net/ipv4/netfilter/ip_conntrack_proto_udp.c b/net/ipv4/netfilter/ip_conntrack_proto_udp.c
index 5bc28a22462..8c1eaba098d 100644
--- a/net/ipv4/netfilter/ip_conntrack_proto_udp.c
+++ b/net/ipv4/netfilter/ip_conntrack_proto_udp.c
@@ -120,6 +120,7 @@ static int udp_error(struct sk_buff *skb, enum ip_conntrack_info *ctinfo,
* and moreover root might send raw packets.
* FIXME: Source route IP option packets --RR */
if (hooknum == NF_IP_PRE_ROUTING
+ && skb->ip_summed != CHECKSUM_UNNECESSARY
&& csum_tcpudp_magic(iph->saddr, iph->daddr, udplen, IPPROTO_UDP,
skb->ip_summed == CHECKSUM_HW ? skb->csum
: skb_checksum(skb, iph->ihl*4, udplen, 0))) {
diff --git a/net/ipv4/netfilter/ip_conntrack_standalone.c b/net/ipv4/netfilter/ip_conntrack_standalone.c
index bc59f7b3980..1dd824f3cf0 100644
--- a/net/ipv4/netfilter/ip_conntrack_standalone.c
+++ b/net/ipv4/netfilter/ip_conntrack_standalone.c
@@ -28,8 +28,8 @@
#include <net/checksum.h>
#include <net/ip.h>
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_conntrack_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_conntrack_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_protocol.h>
@@ -119,7 +119,7 @@ static struct list_head *ct_get_idx(struct seq_file *seq, loff_t pos)
static void *ct_seq_start(struct seq_file *seq, loff_t *pos)
{
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
return ct_get_idx(seq, *pos);
}
@@ -131,7 +131,7 @@ static void *ct_seq_next(struct seq_file *s, void *v, loff_t *pos)
static void ct_seq_stop(struct seq_file *s, void *v)
{
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
}
static int ct_seq_show(struct seq_file *s, void *v)
@@ -140,7 +140,7 @@ static int ct_seq_show(struct seq_file *s, void *v)
const struct ip_conntrack *conntrack = tuplehash_to_ctrack(hash);
struct ip_conntrack_protocol *proto;
- MUST_BE_READ_LOCKED(&ip_conntrack_lock);
+ ASSERT_READ_LOCK(&ip_conntrack_lock);
IP_NF_ASSERT(conntrack);
/* we only want to print DIR_ORIGINAL */
@@ -239,7 +239,7 @@ static void *exp_seq_start(struct seq_file *s, loff_t *pos)
/* strange seq_file api calls stop even if we fail,
* thus we need to grab lock since stop unlocks */
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
if (list_empty(e))
return NULL;
@@ -267,7 +267,7 @@ static void *exp_seq_next(struct seq_file *s, void *v, loff_t *pos)
static void exp_seq_stop(struct seq_file *s, void *v)
{
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
}
static int exp_seq_show(struct seq_file *s, void *v)
@@ -432,6 +432,13 @@ static unsigned int ip_conntrack_defrag(unsigned int hooknum,
const struct net_device *out,
int (*okfn)(struct sk_buff *))
{
+#if !defined(CONFIG_IP_NF_NAT) && !defined(CONFIG_IP_NF_NAT_MODULE)
+ /* Previously seen (loopback)? Ignore. Do this before
+ fragment check. */
+ if ((*pskb)->nfct)
+ return NF_ACCEPT;
+#endif
+
/* Gather fragments. */
if ((*pskb)->nh.iph->frag_off & htons(IP_MF|IP_OFFSET)) {
*pskb = ip_ct_gather_frags(*pskb,
@@ -921,22 +928,22 @@ int ip_conntrack_protocol_register(struct ip_conntrack_protocol *proto)
{
int ret = 0;
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
if (ip_ct_protos[proto->proto] != &ip_conntrack_generic_protocol) {
ret = -EBUSY;
goto out;
}
ip_ct_protos[proto->proto] = proto;
out:
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
return ret;
}
void ip_conntrack_protocol_unregister(struct ip_conntrack_protocol *proto)
{
- WRITE_LOCK(&ip_conntrack_lock);
+ write_lock_bh(&ip_conntrack_lock);
ip_ct_protos[proto->proto] = &ip_conntrack_generic_protocol;
- WRITE_UNLOCK(&ip_conntrack_lock);
+ write_unlock_bh(&ip_conntrack_lock);
/* Somebody could be still looking at the proto in bh. */
synchronize_net();
diff --git a/net/ipv4/netfilter/ip_nat_core.c b/net/ipv4/netfilter/ip_nat_core.c
index 9fc6f93af0d..739b6dde1c8 100644
--- a/net/ipv4/netfilter/ip_nat_core.c
+++ b/net/ipv4/netfilter/ip_nat_core.c
@@ -22,8 +22,8 @@
#include <linux/udp.h>
#include <linux/jhash.h>
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_nat_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_nat_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_core.h>
@@ -41,7 +41,7 @@
#define DEBUGP(format, args...)
#endif
-DECLARE_RWLOCK(ip_nat_lock);
+DEFINE_RWLOCK(ip_nat_lock);
/* Calculated at init based on memory size */
static unsigned int ip_nat_htable_size;
@@ -65,9 +65,9 @@ static void ip_nat_cleanup_conntrack(struct ip_conntrack *conn)
if (!(conn->status & IPS_NAT_DONE_MASK))
return;
- WRITE_LOCK(&ip_nat_lock);
+ write_lock_bh(&ip_nat_lock);
list_del(&conn->nat.info.bysource);
- WRITE_UNLOCK(&ip_nat_lock);
+ write_unlock_bh(&ip_nat_lock);
}
/* We do checksum mangling, so if they were wrong before they're still
@@ -142,7 +142,7 @@ find_appropriate_src(const struct ip_conntrack_tuple *tuple,
unsigned int h = hash_by_src(tuple);
struct ip_conntrack *ct;
- READ_LOCK(&ip_nat_lock);
+ read_lock_bh(&ip_nat_lock);
list_for_each_entry(ct, &bysource[h], nat.info.bysource) {
if (same_src(ct, tuple)) {
/* Copy source part from reply tuple. */
@@ -151,12 +151,12 @@ find_appropriate_src(const struct ip_conntrack_tuple *tuple,
result->dst = tuple->dst;
if (in_range(result, range)) {
- READ_UNLOCK(&ip_nat_lock);
+ read_unlock_bh(&ip_nat_lock);
return 1;
}
}
}
- READ_UNLOCK(&ip_nat_lock);
+ read_unlock_bh(&ip_nat_lock);
return 0;
}
@@ -297,9 +297,9 @@ ip_nat_setup_info(struct ip_conntrack *conntrack,
unsigned int srchash
= hash_by_src(&conntrack->tuplehash[IP_CT_DIR_ORIGINAL]
.tuple);
- WRITE_LOCK(&ip_nat_lock);
+ write_lock_bh(&ip_nat_lock);
list_add(&info->bysource, &bysource[srchash]);
- WRITE_UNLOCK(&ip_nat_lock);
+ write_unlock_bh(&ip_nat_lock);
}
/* It's done. */
@@ -474,23 +474,23 @@ int ip_nat_protocol_register(struct ip_nat_protocol *proto)
{
int ret = 0;
- WRITE_LOCK(&ip_nat_lock);
+ write_lock_bh(&ip_nat_lock);
if (ip_nat_protos[proto->protonum] != &ip_nat_unknown_protocol) {
ret = -EBUSY;
goto out;
}
ip_nat_protos[proto->protonum] = proto;
out:
- WRITE_UNLOCK(&ip_nat_lock);
+ write_unlock_bh(&ip_nat_lock);
return ret;
}
/* Noone stores the protocol anywhere; simply delete it. */
void ip_nat_protocol_unregister(struct ip_nat_protocol *proto)
{
- WRITE_LOCK(&ip_nat_lock);
+ write_lock_bh(&ip_nat_lock);
ip_nat_protos[proto->protonum] = &ip_nat_unknown_protocol;
- WRITE_UNLOCK(&ip_nat_lock);
+ write_unlock_bh(&ip_nat_lock);
/* Someone could be still looking at the proto in a bh. */
synchronize_net();
@@ -509,13 +509,13 @@ int __init ip_nat_init(void)
return -ENOMEM;
/* Sew in builtin protocols. */
- WRITE_LOCK(&ip_nat_lock);
+ write_lock_bh(&ip_nat_lock);
for (i = 0; i < MAX_IP_NAT_PROTO; i++)
ip_nat_protos[i] = &ip_nat_unknown_protocol;
ip_nat_protos[IPPROTO_TCP] = &ip_nat_protocol_tcp;
ip_nat_protos[IPPROTO_UDP] = &ip_nat_protocol_udp;
ip_nat_protos[IPPROTO_ICMP] = &ip_nat_protocol_icmp;
- WRITE_UNLOCK(&ip_nat_lock);
+ write_unlock_bh(&ip_nat_lock);
for (i = 0; i < ip_nat_htable_size; i++) {
INIT_LIST_HEAD(&bysource[i]);
diff --git a/net/ipv4/netfilter/ip_nat_helper.c b/net/ipv4/netfilter/ip_nat_helper.c
index 1637b96d8c0..158f34f32c0 100644
--- a/net/ipv4/netfilter/ip_nat_helper.c
+++ b/net/ipv4/netfilter/ip_nat_helper.c
@@ -28,8 +28,8 @@
#include <net/tcp.h>
#include <net/udp.h>
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_nat_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_nat_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_conntrack.h>
#include <linux/netfilter_ipv4/ip_conntrack_helper.h>
@@ -47,7 +47,7 @@
#define DUMP_OFFSET(x)
#endif
-static DECLARE_LOCK(ip_nat_seqofs_lock);
+static DEFINE_SPINLOCK(ip_nat_seqofs_lock);
/* Setup TCP sequence correction given this change at this sequence */
static inline void
@@ -70,7 +70,7 @@ adjust_tcp_sequence(u32 seq,
DEBUGP("ip_nat_resize_packet: Seq_offset before: ");
DUMP_OFFSET(this_way);
- LOCK_BH(&ip_nat_seqofs_lock);
+ spin_lock_bh(&ip_nat_seqofs_lock);
/* SYN adjust. If it's uninitialized, or this is after last
* correction, record it: we don't handle more than one
@@ -82,7 +82,7 @@ adjust_tcp_sequence(u32 seq,
this_way->offset_before = this_way->offset_after;
this_way->offset_after += sizediff;
}
- UNLOCK_BH(&ip_nat_seqofs_lock);
+ spin_unlock_bh(&ip_nat_seqofs_lock);
DEBUGP("ip_nat_resize_packet: Seq_offset after: ");
DUMP_OFFSET(this_way);
@@ -142,9 +142,6 @@ static int enlarge_skb(struct sk_buff **pskb, unsigned int extra)
/* Transfer socket to new skb. */
if ((*pskb)->sk)
skb_set_owner_w(nskb, (*pskb)->sk);
-#ifdef CONFIG_NETFILTER_DEBUG
- nskb->nf_debug = (*pskb)->nf_debug;
-#endif
kfree_skb(*pskb);
*pskb = nskb;
return 1;
diff --git a/net/ipv4/netfilter/ip_nat_rule.c b/net/ipv4/netfilter/ip_nat_rule.c
index 581f097f5a2..60d70fa41a1 100644
--- a/net/ipv4/netfilter/ip_nat_rule.c
+++ b/net/ipv4/netfilter/ip_nat_rule.c
@@ -19,8 +19,8 @@
#include <net/route.h>
#include <linux/bitops.h>
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_nat_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_nat_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_tables.h>
#include <linux/netfilter_ipv4/ip_nat.h>
diff --git a/net/ipv4/netfilter/ip_nat_standalone.c b/net/ipv4/netfilter/ip_nat_standalone.c
index 79f56f662b3..bc59d0d6e89 100644
--- a/net/ipv4/netfilter/ip_nat_standalone.c
+++ b/net/ipv4/netfilter/ip_nat_standalone.c
@@ -31,8 +31,8 @@
#include <net/checksum.h>
#include <linux/spinlock.h>
-#define ASSERT_READ_LOCK(x) MUST_BE_READ_LOCKED(&ip_nat_lock)
-#define ASSERT_WRITE_LOCK(x) MUST_BE_WRITE_LOCKED(&ip_nat_lock)
+#define ASSERT_READ_LOCK(x)
+#define ASSERT_WRITE_LOCK(x)
#include <linux/netfilter_ipv4/ip_nat.h>
#include <linux/netfilter_ipv4/ip_nat_rule.h>
@@ -373,7 +373,6 @@ static int init_or_cleanup(int init)
cleanup_rule_init:
ip_nat_rule_cleanup();
cleanup_nothing:
- MUST_BE_READ_WRITE_UNLOCKED(&ip_nat_lock);
return ret;
}
diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c
index 8a54f92b849..c88dfcd38c5 100644
--- a/net/ipv4/netfilter/ip_tables.c
+++ b/net/ipv4/netfilter/ip_tables.c
@@ -67,7 +67,6 @@ static DECLARE_MUTEX(ipt_mutex);
/* Must have mutex */
#define ASSERT_READ_LOCK(x) IP_NF_ASSERT(down_trylock(&ipt_mutex) != 0)
#define ASSERT_WRITE_LOCK(x) IP_NF_ASSERT(down_trylock(&ipt_mutex) != 0)
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <linux/netfilter_ipv4/listhelp.h>
#if 0
diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 0f12e3a3dc7..6706d3a1bc4 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -29,9 +29,8 @@
#include <linux/netfilter_ipv4/ip_tables.h>
#include <linux/netfilter_ipv4/ipt_CLUSTERIP.h>
#include <linux/netfilter_ipv4/ip_conntrack.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
-#define CLUSTERIP_VERSION "0.6"
+#define CLUSTERIP_VERSION "0.7"
#define DEBUG_CLUSTERIP
@@ -41,6 +40,8 @@
#define DEBUGP
#endif
+#define ASSERT_READ_LOCK(x)
+
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Harald Welte <laforge@netfilter.org>");
MODULE_DESCRIPTION("iptables target for CLUSTERIP");
@@ -67,7 +68,7 @@ static LIST_HEAD(clusterip_configs);
/* clusterip_lock protects the clusterip_configs list _AND_ the configurable
* data within all structurses (num_local_nodes, local_nodes[]) */
-static DECLARE_RWLOCK(clusterip_lock);
+static DEFINE_RWLOCK(clusterip_lock);
#ifdef CONFIG_PROC_FS
static struct file_operations clusterip_proc_fops;
@@ -82,9 +83,9 @@ clusterip_config_get(struct clusterip_config *c) {
static inline void
clusterip_config_put(struct clusterip_config *c) {
if (atomic_dec_and_test(&c->refcount)) {
- WRITE_LOCK(&clusterip_lock);
+ write_lock_bh(&clusterip_lock);
list_del(&c->list);
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
dev_mc_delete(c->dev, c->clustermac, ETH_ALEN, 0);
dev_put(c->dev);
kfree(c);
@@ -97,7 +98,7 @@ __clusterip_config_find(u_int32_t clusterip)
{
struct list_head *pos;
- MUST_BE_READ_LOCKED(&clusterip_lock);
+ ASSERT_READ_LOCK(&clusterip_lock);
list_for_each(pos, &clusterip_configs) {
struct clusterip_config *c = list_entry(pos,
struct clusterip_config, list);
@@ -114,14 +115,14 @@ clusterip_config_find_get(u_int32_t clusterip)
{
struct clusterip_config *c;
- READ_LOCK(&clusterip_lock);
+ read_lock_bh(&clusterip_lock);
c = __clusterip_config_find(clusterip);
if (!c) {
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
return NULL;
}
atomic_inc(&c->refcount);
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
return c;
}
@@ -160,9 +161,9 @@ clusterip_config_init(struct ipt_clusterip_tgt_info *i, u_int32_t ip,
c->pde->data = c;
#endif
- WRITE_LOCK(&clusterip_lock);
+ write_lock_bh(&clusterip_lock);
list_add(&c->list, &clusterip_configs);
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return c;
}
@@ -172,25 +173,25 @@ clusterip_add_node(struct clusterip_config *c, u_int16_t nodenum)
{
int i;
- WRITE_LOCK(&clusterip_lock);
+ write_lock_bh(&clusterip_lock);
if (c->num_local_nodes >= CLUSTERIP_MAX_NODES
|| nodenum > CLUSTERIP_MAX_NODES) {
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 1;
}
/* check if we alrady have this number in our array */
for (i = 0; i < c->num_local_nodes; i++) {
if (c->local_nodes[i] == nodenum) {
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 1;
}
}
c->local_nodes[c->num_local_nodes++] = nodenum;
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 0;
}
@@ -199,10 +200,10 @@ clusterip_del_node(struct clusterip_config *c, u_int16_t nodenum)
{
int i;
- WRITE_LOCK(&clusterip_lock);
+ write_lock_bh(&clusterip_lock);
if (c->num_local_nodes <= 1 || nodenum > CLUSTERIP_MAX_NODES) {
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 1;
}
@@ -211,12 +212,12 @@ clusterip_del_node(struct clusterip_config *c, u_int16_t nodenum)
int size = sizeof(u_int16_t)*(c->num_local_nodes-(i+1));
memmove(&c->local_nodes[i], &c->local_nodes[i+1], size);
c->num_local_nodes--;
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 0;
}
}
- WRITE_UNLOCK(&clusterip_lock);
+ write_unlock_bh(&clusterip_lock);
return 1;
}
@@ -286,21 +287,21 @@ clusterip_responsible(struct clusterip_config *config, u_int32_t hash)
{
int i;
- READ_LOCK(&clusterip_lock);
+ read_lock_bh(&clusterip_lock);
if (config->num_local_nodes == 0) {
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
return 0;
}
for (i = 0; i < config->num_local_nodes; i++) {
if (config->local_nodes[i] == hash) {
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
return 1;
}
}
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
return 0;
}
@@ -338,7 +339,7 @@ target(struct sk_buff **pskb,
* error messages (RELATED) and information requests (see below) */
if ((*pskb)->nh.iph->protocol == IPPROTO_ICMP
&& (ctinfo == IP_CT_RELATED
- || ctinfo == IP_CT_IS_REPLY+IP_CT_IS_REPLY))
+ || ctinfo == IP_CT_RELATED+IP_CT_IS_REPLY))
return IPT_CONTINUE;
/* ip_conntrack_icmp guarantees us that we only have ICMP_ECHO,
@@ -523,8 +524,9 @@ arp_mangle(unsigned int hook,
|| arp->ar_pln != 4 || arp->ar_hln != ETH_ALEN)
return NF_ACCEPT;
- /* we only want to mangle arp replies */
- if (arp->ar_op != htons(ARPOP_REPLY))
+ /* we only want to mangle arp requests and replies */
+ if (arp->ar_op != htons(ARPOP_REPLY)
+ && arp->ar_op != htons(ARPOP_REQUEST))
return NF_ACCEPT;
payload = (void *)(arp+1);
@@ -578,7 +580,7 @@ static void *clusterip_seq_start(struct seq_file *s, loff_t *pos)
struct clusterip_config *c = pde->data;
unsigned int *nodeidx;
- READ_LOCK(&clusterip_lock);
+ read_lock_bh(&clusterip_lock);
if (*pos >= c->num_local_nodes)
return NULL;
@@ -608,7 +610,7 @@ static void clusterip_seq_stop(struct seq_file *s, void *v)
{
kfree(v);
- READ_UNLOCK(&clusterip_lock);
+ read_unlock_bh(&clusterip_lock);
}
static int clusterip_seq_show(struct seq_file *s, void *v)
diff --git a/net/ipv4/netfilter/ipt_MASQUERADE.c b/net/ipv4/netfilter/ipt_MASQUERADE.c
index 57e9f6cf1c3..91e74502c3d 100644
--- a/net/ipv4/netfilter/ipt_MASQUERADE.c
+++ b/net/ipv4/netfilter/ipt_MASQUERADE.c
@@ -33,7 +33,7 @@ MODULE_DESCRIPTION("iptables MASQUERADE target module");
#endif
/* Lock protects masq region inside conntrack */
-static DECLARE_RWLOCK(masq_lock);
+static DEFINE_RWLOCK(masq_lock);
/* FIXME: Multiple targets. --RR */
static int
@@ -103,9 +103,9 @@ masquerade_target(struct sk_buff **pskb,
return NF_DROP;
}
- WRITE_LOCK(&masq_lock);
+ write_lock_bh(&masq_lock);
ct->nat.masq_index = out->ifindex;
- WRITE_UNLOCK(&masq_lock);
+ write_unlock_bh(&masq_lock);
/* Transfer from original range. */
newrange = ((struct ip_nat_range)
@@ -122,9 +122,9 @@ device_cmp(struct ip_conntrack *i, void *ifindex)
{
int ret;
- READ_LOCK(&masq_lock);
+ read_lock_bh(&masq_lock);
ret = (i->nat.masq_index == (int)(long)ifindex);
- READ_UNLOCK(&masq_lock);
+ read_unlock_bh(&masq_lock);
return ret;
}
diff --git a/net/ipv4/netfilter/ipt_REJECT.c b/net/ipv4/netfilter/ipt_REJECT.c
index 266d6497928..91569644602 100644
--- a/net/ipv4/netfilter/ipt_REJECT.c
+++ b/net/ipv4/netfilter/ipt_REJECT.c
@@ -104,10 +104,12 @@ static inline struct rtable *route_reverse(struct sk_buff *skb,
static void send_reset(struct sk_buff *oldskb, int hook)
{
struct sk_buff *nskb;
+ struct iphdr *iph = oldskb->nh.iph;
struct tcphdr _otcph, *oth, *tcph;
struct rtable *rt;
u_int16_t tmp_port;
u_int32_t tmp_addr;
+ unsigned int tcplen;
int needs_ack;
int hh_len;
@@ -124,7 +126,16 @@ static void send_reset(struct sk_buff *oldskb, int hook)
if (oth->rst)
return;
- /* FIXME: Check checksum --RR */
+ /* Check checksum */
+ tcplen = oldskb->len - iph->ihl * 4;
+ if (((hook != NF_IP_LOCAL_IN && oldskb->ip_summed != CHECKSUM_HW) ||
+ (hook == NF_IP_LOCAL_IN &&
+ oldskb->ip_summed != CHECKSUM_UNNECESSARY)) &&
+ csum_tcpudp_magic(iph->saddr, iph->daddr, tcplen, IPPROTO_TCP,
+ oldskb->ip_summed == CHECKSUM_HW ? oldskb->csum :
+ skb_checksum(oldskb, iph->ihl * 4, tcplen, 0)))
+ return;
+
if ((rt = route_reverse(oldskb, oth, hook)) == NULL)
return;
diff --git a/net/ipv4/netfilter/ipt_ULOG.c b/net/ipv4/netfilter/ipt_ULOG.c
index 6f2cefbe16c..52a0076302a 100644
--- a/net/ipv4/netfilter/ipt_ULOG.c
+++ b/net/ipv4/netfilter/ipt_ULOG.c
@@ -56,7 +56,6 @@
#include <linux/netfilter.h>
#include <linux/netfilter_ipv4/ip_tables.h>
#include <linux/netfilter_ipv4/ipt_ULOG.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
#include <net/sock.h>
#include <linux/bitops.h>
@@ -99,8 +98,8 @@ typedef struct {
static ulog_buff_t ulog_buffers[ULOG_MAXNLGROUPS]; /* array of buffers */
-static struct sock *nflognl; /* our socket */
-static DECLARE_LOCK(ulog_lock); /* spinlock */
+static struct sock *nflognl; /* our socket */
+static DEFINE_SPINLOCK(ulog_lock); /* spinlock */
/* send one ulog_buff_t to userspace */
static void ulog_send(unsigned int nlgroupnum)
@@ -135,9 +134,9 @@ static void ulog_timer(unsigned long data)
/* lock to protect against somebody modifying our structure
* from ipt_ulog_target at the same time */
- LOCK_BH(&ulog_lock);
+ spin_lock_bh(&ulog_lock);
ulog_send(data);
- UNLOCK_BH(&ulog_lock);
+ spin_unlock_bh(&ulog_lock);
}
static struct sk_buff *ulog_alloc_skb(unsigned int size)
@@ -193,7 +192,7 @@ static void ipt_ulog_packet(unsigned int hooknum,
ub = &ulog_buffers[groupnum];
- LOCK_BH(&ulog_lock);
+ spin_lock_bh(&ulog_lock);
if (!ub->skb) {
if (!(ub->skb = ulog_alloc_skb(size)))
@@ -278,7 +277,7 @@ static void ipt_ulog_packet(unsigned int hooknum,
ulog_send(groupnum);
}
- UNLOCK_BH(&ulog_lock);
+ spin_unlock_bh(&ulog_lock);
return;
@@ -288,7 +287,7 @@ nlmsg_failure:
alloc_failure:
PRINTR("ipt_ULOG: Error building netlink message\n");
- UNLOCK_BH(&ulog_lock);
+ spin_unlock_bh(&ulog_lock);
}
static unsigned int ipt_ulog_target(struct sk_buff **pskb,
diff --git a/net/ipv4/netfilter/ipt_hashlimit.c b/net/ipv4/netfilter/ipt_hashlimit.c
index f1937190cd7..564b49bfebc 100644
--- a/net/ipv4/netfilter/ipt_hashlimit.c
+++ b/net/ipv4/netfilter/ipt_hashlimit.c
@@ -37,7 +37,6 @@
#include <linux/netfilter_ipv4/ip_tables.h>
#include <linux/netfilter_ipv4/ipt_hashlimit.h>
-#include <linux/netfilter_ipv4/lockhelp.h>
/* FIXME: this is just for IP_NF_ASSERRT */
#include <linux/netfilter_ipv4/ip_conntrack.h>
@@ -92,7 +91,7 @@ struct ipt_hashlimit_htable {
struct hlist_head hash[0]; /* hashtable itself */
};
-static DECLARE_LOCK(hashlimit_lock); /* protects htables list */
+static DEFINE_SPINLOCK(hashlimit_lock); /* protects htables list */
static DECLARE_MUTEX(hlimit_mutex); /* additional checkentry protection */
static HLIST_HEAD(hashlimit_htables);
static kmem_cache_t *hashlimit_cachep;
@@ -233,9 +232,9 @@ static int htable_create(struct ipt_hashlimit_info *minfo)
hinfo->timer.function = htable_gc;
add_timer(&hinfo->timer);
- LOCK_BH(&hashlimit_lock);
+ spin_lock_bh(&hashlimit_lock);
hlist_add_head(&hinfo->node, &hashlimit_htables);
- UNLOCK_BH(&hashlimit_lock);
+ spin_unlock_bh(&hashlimit_lock);
return 0;
}
@@ -301,15 +300,15 @@ static struct ipt_hashlimit_htable *htable_find_get(char *name)
struct ipt_hashlimit_htable *hinfo;
struct hlist_node *pos;
- LOCK_BH(&hashlimit_lock);
+ spin_lock_bh(&hashlimit_lock);
hlist_for_each_entry(hinfo, pos, &hashlimit_htables, node) {
if (!strcmp(name, hinfo->pde->name)) {
atomic_inc(&hinfo->use);
- UNLOCK_BH(&hashlimit_lock);
+ spin_unlock_bh(&hashlimit_lock);
return hinfo;
}
}
- UNLOCK_BH(&hashlimit_lock);
+ spin_unlock_bh(&hashlimit_lock);
return NULL;
}
@@ -317,9 +316,9 @@ static struct ipt_hashlimit_htable *htable_find_get(char *name)
static void htable_put(struct ipt_hashlimit_htable *hinfo)
{
if (atomic_dec_and_test(&hinfo->use)) {
- LOCK_BH(&hashlimit_lock);
+ spin_lock_bh(&hashlimit_lock);
hlist_del(&hinfo->node);
- UNLOCK_BH(&hashlimit_lock);
+ spin_unlock_bh(&hashlimit_lock);
htable_destroy(hinfo);
}
}
diff --git a/net/ipv4/netfilter/ipt_helper.c b/net/ipv4/netfilter/ipt_helper.c
index 33fdf364d3d..3e7dd014de4 100644
--- a/net/ipv4/netfilter/ipt_helper.c
+++ b/net/ipv4/netfilter/ipt_helper.c
@@ -53,7 +53,7 @@ match(const struct sk_buff *skb,
return ret;
}
- READ_LOCK(&ip_conntrack_lock);
+ read_lock_bh(&ip_conntrack_lock);
if (!ct->master->helper) {
DEBUGP("ipt_helper: master ct %p has no helper\n",
exp->expectant);
@@ -69,7 +69,7 @@ match(const struct sk_buff *skb,
ret ^= !strncmp(ct->master->helper->name, info->name,
strlen(ct->master->helper->name));
out_unlock:
- READ_UNLOCK(&ip_conntrack_lock);
+ read_unlock_bh(&ip_conntrack_lock);
return ret;
}
diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index 5b1ec586bae..d1835b1bc8c 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c
@@ -259,7 +259,7 @@ int raw_rcv(struct sock *sk, struct sk_buff *skb)
return 0;
}
-static int raw_send_hdrinc(struct sock *sk, void *from, int length,
+static int raw_send_hdrinc(struct sock *sk, void *from, size_t length,
struct rtable *rt,
unsigned int flags)
{
@@ -298,7 +298,7 @@ static int raw_send_hdrinc(struct sock *sk, void *from, int length,
goto error_fault;
/* We don't modify invalid header */
- if (length >= sizeof(*iph) && iph->ihl * 4 <= length) {
+ if (length >= sizeof(*iph) && iph->ihl * 4U <= length) {
if (!iph->saddr)
iph->saddr = rt->rt_src;
iph->check = 0;
@@ -332,7 +332,7 @@ static void raw_probe_proto_opt(struct flowi *fl, struct msghdr *msg)
u8 __user *type = NULL;
u8 __user *code = NULL;
int probed = 0;
- int i;
+ unsigned int i;
if (!msg->msg_iov)
return;
@@ -384,7 +384,7 @@ static int raw_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
int err;
err = -EMSGSIZE;
- if (len < 0 || len > 0xFFFF)
+ if (len > 0xFFFF)
goto out;
/*
@@ -514,7 +514,10 @@ done:
kfree(ipc.opt);
ip_rt_put(rt);
-out: return err < 0 ? err : len;
+out:
+ if (err < 0)
+ return err;
+ return len;
do_confirm:
dst_confirm(&rt->u.dst);
@@ -610,7 +613,10 @@ static int raw_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
copied = skb->len;
done:
skb_free_datagram(sk, skb);
-out: return err ? err : copied;
+out:
+ if (err)
+ return err;
+ return copied;
}
static int raw_init(struct sock *sk)
@@ -691,11 +697,11 @@ static int raw_ioctl(struct sock *sk, int cmd, unsigned long arg)
struct sk_buff *skb;
int amount = 0;
- spin_lock_irq(&sk->sk_receive_queue.lock);
+ spin_lock_bh(&sk->sk_receive_queue.lock);
skb = skb_peek(&sk->sk_receive_queue);
if (skb != NULL)
amount = skb->len;
- spin_unlock_irq(&sk->sk_receive_queue.lock);
+ spin_unlock_bh(&sk->sk_receive_queue.lock);
return put_user(amount, (int __user *)arg);
}
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index a682d28e247..d675ff80b04 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -54,6 +54,7 @@
* Marc Boucher : routing by fwmark
* Robert Olsson : Added rt_cache statistics
* Arnaldo C. Melo : Convert proc stuff to seq_file
+ * Eric Dumazet : hashed spinlocks and rt_check_expire() fixes.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
@@ -70,6 +71,7 @@
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/mm.h>
+#include <linux/bootmem.h>
#include <linux/string.h>
#include <linux/socket.h>
#include <linux/sockios.h>
@@ -201,8 +203,37 @@ __u8 ip_tos2prio[16] = {
struct rt_hash_bucket {
struct rtable *chain;
- spinlock_t lock;
-} __attribute__((__aligned__(8)));
+};
+#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK)
+/*
+ * Instead of using one spinlock for each rt_hash_bucket, we use a table of spinlocks
+ * The size of this table is a power of two and depends on the number of CPUS.
+ */
+#if NR_CPUS >= 32
+#define RT_HASH_LOCK_SZ 4096
+#elif NR_CPUS >= 16
+#define RT_HASH_LOCK_SZ 2048
+#elif NR_CPUS >= 8
+#define RT_HASH_LOCK_SZ 1024
+#elif NR_CPUS >= 4
+#define RT_HASH_LOCK_SZ 512
+#else
+#define RT_HASH_LOCK_SZ 256
+#endif
+
+static spinlock_t *rt_hash_locks;
+# define rt_hash_lock_addr(slot) &rt_hash_locks[(slot) & (RT_HASH_LOCK_SZ - 1)]
+# define rt_hash_lock_init() { \
+ int i; \
+ rt_hash_locks = kmalloc(sizeof(spinlock_t) * RT_HASH_LOCK_SZ, GFP_KERNEL); \
+ if (!rt_hash_locks) panic("IP: failed to allocate rt_hash_locks\n"); \
+ for (i = 0; i < RT_HASH_LOCK_SZ; i++) \
+ spin_lock_init(&rt_hash_locks[i]); \
+ }
+#else
+# define rt_hash_lock_addr(slot) NULL
+# define rt_hash_lock_init()
+#endif
static struct rt_hash_bucket *rt_hash_table;
static unsigned rt_hash_mask;
@@ -575,19 +606,26 @@ static struct rtable **rt_remove_balanced_route(struct rtable **chain_head,
/* This runs via a timer and thus is always in BH context. */
static void rt_check_expire(unsigned long dummy)
{
- static int rover;
- int i = rover, t;
+ static unsigned int rover;
+ unsigned int i = rover, goal;
struct rtable *rth, **rthp;
unsigned long now = jiffies;
-
- for (t = ip_rt_gc_interval << rt_hash_log; t >= 0;
- t -= ip_rt_gc_timeout) {
+ u64 mult;
+
+ mult = ((u64)ip_rt_gc_interval) << rt_hash_log;
+ if (ip_rt_gc_timeout > 1)
+ do_div(mult, ip_rt_gc_timeout);
+ goal = (unsigned int)mult;
+ if (goal > rt_hash_mask) goal = rt_hash_mask + 1;
+ for (; goal > 0; goal--) {
unsigned long tmo = ip_rt_gc_timeout;
i = (i + 1) & rt_hash_mask;
rthp = &rt_hash_table[i].chain;
- spin_lock(&rt_hash_table[i].lock);
+ if (*rthp == 0)
+ continue;
+ spin_lock(rt_hash_lock_addr(i));
while ((rth = *rthp) != NULL) {
if (rth->u.dst.expires) {
/* Entry is expired even if it is in use */
@@ -620,14 +658,14 @@ static void rt_check_expire(unsigned long dummy)
rt_free(rth);
#endif /* CONFIG_IP_ROUTE_MULTIPATH_CACHED */
}
- spin_unlock(&rt_hash_table[i].lock);
+ spin_unlock(rt_hash_lock_addr(i));
/* Fallback loop breaker. */
if (time_after(jiffies, now))
break;
}
rover = i;
- mod_timer(&rt_periodic_timer, now + ip_rt_gc_interval);
+ mod_timer(&rt_periodic_timer, jiffies + ip_rt_gc_interval);
}
/* This can run from both BH and non-BH contexts, the latter
@@ -643,11 +681,11 @@ static void rt_run_flush(unsigned long dummy)
get_random_bytes(&rt_hash_rnd, 4);
for (i = rt_hash_mask; i >= 0; i--) {
- spin_lock_bh(&rt_hash_table[i].lock);
+ spin_lock_bh(rt_hash_lock_addr(i));
rth = rt_hash_table[i].chain;
if (rth)
rt_hash_table[i].chain = NULL;
- spin_unlock_bh(&rt_hash_table[i].lock);
+ spin_unlock_bh(rt_hash_lock_addr(i));
for (; rth; rth = next) {
next = rth->u.rt_next;
@@ -780,7 +818,7 @@ static int rt_garbage_collect(void)
k = (k + 1) & rt_hash_mask;
rthp = &rt_hash_table[k].chain;
- spin_lock_bh(&rt_hash_table[k].lock);
+ spin_lock_bh(rt_hash_lock_addr(k));
while ((rth = *rthp) != NULL) {
if (!rt_may_expire(rth, tmo, expire)) {
tmo >>= 1;
@@ -812,7 +850,7 @@ static int rt_garbage_collect(void)
goal--;
#endif /* CONFIG_IP_ROUTE_MULTIPATH_CACHED */
}
- spin_unlock_bh(&rt_hash_table[k].lock);
+ spin_unlock_bh(rt_hash_lock_addr(k));
if (goal <= 0)
break;
}
@@ -882,7 +920,7 @@ restart:
rthp = &rt_hash_table[hash].chain;
- spin_lock_bh(&rt_hash_table[hash].lock);
+ spin_lock_bh(rt_hash_lock_addr(hash));
while ((rth = *rthp) != NULL) {
#ifdef CONFIG_IP_ROUTE_MULTIPATH_CACHED
if (!(rth->u.dst.flags & DST_BALANCED) &&
@@ -908,7 +946,7 @@ restart:
rth->u.dst.__use++;
dst_hold(&rth->u.dst);
rth->u.dst.lastuse = now;
- spin_unlock_bh(&rt_hash_table[hash].lock);
+ spin_unlock_bh(rt_hash_lock_addr(hash));
rt_drop(rt);
*rp = rth;
@@ -949,7 +987,7 @@ restart:
if (rt->rt_type == RTN_UNICAST || rt->fl.iif == 0) {
int err = arp_bind_neighbour(&rt->u.dst);
if (err) {
- spin_unlock_bh(&rt_hash_table[hash].lock);
+ spin_unlock_bh(rt_hash_lock_addr(hash));
if (err != -ENOBUFS) {
rt_drop(rt);
@@ -990,7 +1028,7 @@ restart:
}
#endif
rt_hash_table[hash].chain = rt;
- spin_unlock_bh(&rt_hash_table[hash].lock);
+ spin_unlock_bh(rt_hash_lock_addr(hash));
*rp = rt;
return 0;
}
@@ -1058,7 +1096,7 @@ static void rt_del(unsigned hash, struct rtable *rt)
{
struct rtable **rthp;
- spin_lock_bh(&rt_hash_table[hash].lock);
+ spin_lock_bh(rt_hash_lock_addr(hash));
ip_rt_put(rt);
for (rthp = &rt_hash_table[hash].chain; *rthp;
rthp = &(*rthp)->u.rt_next)
@@ -1067,7 +1105,7 @@ static void rt_del(unsigned hash, struct rtable *rt)
rt_free(rt);
break;
}
- spin_unlock_bh(&rt_hash_table[hash].lock);
+ spin_unlock_bh(rt_hash_lock_addr(hash));
}
void ip_rt_redirect(u32 old_gw, u32 daddr, u32 new_gw,
@@ -1647,7 +1685,7 @@ static void ip_handle_martian_source(struct net_device *dev,
printk(KERN_WARNING "martian source %u.%u.%u.%u from "
"%u.%u.%u.%u, on dev %s\n",
NIPQUAD(daddr), NIPQUAD(saddr), dev->name);
- if (dev->hard_header_len) {
+ if (dev->hard_header_len && skb->mac.raw) {
int i;
unsigned char *p = skb->mac.raw;
printk(KERN_WARNING "ll header: ");
@@ -1767,7 +1805,7 @@ static inline int ip_mkroute_input_def(struct sk_buff *skb,
struct in_device *in_dev,
u32 daddr, u32 saddr, u32 tos)
{
- struct rtable* rth;
+ struct rtable* rth = NULL;
int err;
unsigned hash;
@@ -1794,7 +1832,7 @@ static inline int ip_mkroute_input(struct sk_buff *skb,
u32 daddr, u32 saddr, u32 tos)
{
#ifdef CONFIG_IP_ROUTE_MULTIPATH_CACHED
- struct rtable* rth;
+ struct rtable* rth = NULL;
unsigned char hop, hopcount, lasthop;
int err = -EINVAL;
unsigned int hash;
@@ -1909,7 +1947,7 @@ static int ip_route_input_slow(struct sk_buff *skb, u32 daddr, u32 saddr,
*/
if ((err = fib_lookup(&fl, &res)) != 0) {
if (!IN_DEV_FORWARD(in_dev))
- goto e_inval;
+ goto e_hostunreach;
goto no_route;
}
free_res = 1;
@@ -1933,7 +1971,7 @@ static int ip_route_input_slow(struct sk_buff *skb, u32 daddr, u32 saddr,
}
if (!IN_DEV_FORWARD(in_dev))
- goto e_inval;
+ goto e_hostunreach;
if (res.type != RTN_UNICAST)
goto martian_destination;
@@ -2025,6 +2063,11 @@ martian_destination:
"%u.%u.%u.%u, dev %s\n",
NIPQUAD(daddr), NIPQUAD(saddr), dev->name);
#endif
+
+e_hostunreach:
+ err = -EHOSTUNREACH;
+ goto done;
+
e_inval:
err = -EINVAL;
goto done;
@@ -2239,7 +2282,7 @@ static inline int ip_mkroute_output_def(struct rtable **rp,
struct net_device *dev_out,
unsigned flags)
{
- struct rtable *rth;
+ struct rtable *rth = NULL;
int err = __mkroute_output(&rth, res, fl, oldflp, dev_out, flags);
unsigned hash;
if (err == 0) {
@@ -2267,7 +2310,7 @@ static inline int ip_mkroute_output(struct rtable** rp,
unsigned char hop;
unsigned hash;
int err = -EINVAL;
- struct rtable *rth;
+ struct rtable *rth = NULL;
if (res->fi && res->fi->fib_nhs > 1) {
unsigned char hopcount = res->fi->fib_nhs;
@@ -2581,7 +2624,7 @@ int ip_route_output_key(struct rtable **rp, struct flowi *flp)
}
static int rt_fill_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
- int nowait)
+ int nowait, unsigned int flags)
{
struct rtable *rt = (struct rtable*)skb->dst;
struct rtmsg *r;
@@ -2591,9 +2634,8 @@ static int rt_fill_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
#ifdef CONFIG_IP_MROUTE
struct rtattr *eptr;
#endif
- nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*r));
+ nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*r), flags);
r = NLMSG_DATA(nlh);
- nlh->nlmsg_flags = (nowait && pid) ? NLM_F_MULTI : 0;
r->rtm_family = AF_INET;
r->rtm_dst_len = 32;
r->rtm_src_len = 0;
@@ -2744,7 +2786,7 @@ int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr* nlh, void *arg)
NETLINK_CB(skb).dst_pid = NETLINK_CB(in_skb).pid;
err = rt_fill_info(skb, NETLINK_CB(in_skb).pid, nlh->nlmsg_seq,
- RTM_NEWROUTE, 0);
+ RTM_NEWROUTE, 0, 0);
if (!err)
goto out_free;
if (err < 0) {
@@ -2781,8 +2823,8 @@ int ip_rt_dump(struct sk_buff *skb, struct netlink_callback *cb)
continue;
skb->dst = dst_clone(&rt->u.dst);
if (rt_fill_info(skb, NETLINK_CB(cb->skb).pid,
- cb->nlh->nlmsg_seq,
- RTM_NEWROUTE, 1) <= 0) {
+ cb->nlh->nlmsg_seq, RTM_NEWROUTE,
+ 1, NLM_F_MULTI) <= 0) {
dst_release(xchg(&skb->dst, NULL));
rcu_read_unlock_bh();
goto done;
@@ -3069,12 +3111,14 @@ __setup("rhash_entries=", set_rhash_entries);
int __init ip_rt_init(void)
{
- int i, order, goal, rc = 0;
+ int rc = 0;
rt_hash_rnd = (int) ((num_physpages ^ (num_physpages>>8)) ^
(jiffies ^ (jiffies >> 7)));
#ifdef CONFIG_NET_CLS_ROUTE
+ {
+ int order;
for (order = 0;
(PAGE_SIZE << order) < 256 * sizeof(struct ip_rt_acct) * NR_CPUS; order++)
/* NOTHING */;
@@ -3082,6 +3126,7 @@ int __init ip_rt_init(void)
if (!ip_rt_acct)
panic("IP: failed to allocate ip_rt_acct\n");
memset(ip_rt_acct, 0, PAGE_SIZE << order);
+ }
#endif
ipv4_dst_ops.kmem_cachep = kmem_cache_create("ip_dst_cache",
@@ -3092,36 +3137,19 @@ int __init ip_rt_init(void)
if (!ipv4_dst_ops.kmem_cachep)
panic("IP: failed to allocate ip_dst_cache\n");
- goal = num_physpages >> (26 - PAGE_SHIFT);
- if (rhash_entries)
- goal = (rhash_entries * sizeof(struct rt_hash_bucket)) >> PAGE_SHIFT;
- for (order = 0; (1UL << order) < goal; order++)
- /* NOTHING */;
-
- do {
- rt_hash_mask = (1UL << order) * PAGE_SIZE /
- sizeof(struct rt_hash_bucket);
- while (rt_hash_mask & (rt_hash_mask - 1))
- rt_hash_mask--;
- rt_hash_table = (struct rt_hash_bucket *)
- __get_free_pages(GFP_ATOMIC, order);
- } while (rt_hash_table == NULL && --order > 0);
-
- if (!rt_hash_table)
- panic("Failed to allocate IP route cache hash table\n");
-
- printk(KERN_INFO "IP: routing cache hash table of %u buckets, %ldKbytes\n",
- rt_hash_mask,
- (long) (rt_hash_mask * sizeof(struct rt_hash_bucket)) / 1024);
-
- for (rt_hash_log = 0; (1 << rt_hash_log) != rt_hash_mask; rt_hash_log++)
- /* NOTHING */;
-
- rt_hash_mask--;
- for (i = 0; i <= rt_hash_mask; i++) {
- spin_lock_init(&rt_hash_table[i].lock);
- rt_hash_table[i].chain = NULL;
- }
+ rt_hash_table = (struct rt_hash_bucket *)
+ alloc_large_system_hash("IP route cache",
+ sizeof(struct rt_hash_bucket),
+ rhash_entries,
+ (num_physpages >= 128 * 1024) ?
+ (27 - PAGE_SHIFT) :
+ (29 - PAGE_SHIFT),
+ HASH_HIGHMEM,
+ &rt_hash_log,
+ &rt_hash_mask,
+ 0);
+ memset(rt_hash_table, 0, (rt_hash_mask + 1) * sizeof(struct rt_hash_bucket));
+ rt_hash_lock_init();
ipv4_dst_ops.gc_thresh = (rt_hash_mask + 1);
ip_rt_max_size = (rt_hash_mask + 1) * 16;
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index e923d2f021a..72d01444218 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -169,10 +169,10 @@ static inline int cookie_check(struct sk_buff *skb, __u32 cookie)
return mssind < NUM_MSS ? msstab[mssind] + 1 : 0;
}
-extern struct or_calltable or_ipv4;
+extern struct request_sock_ops tcp_request_sock_ops;
static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
- struct open_request *req,
+ struct request_sock *req,
struct dst_entry *dst)
{
struct tcp_sock *tp = tcp_sk(sk);
@@ -182,7 +182,7 @@ static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
if (child)
tcp_acceptq_queue(sk, req, child);
else
- tcp_openreq_free(req);
+ reqsk_free(req);
return child;
}
@@ -190,10 +190,12 @@ static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
struct ip_options *opt)
{
+ struct inet_request_sock *ireq;
+ struct tcp_request_sock *treq;
struct tcp_sock *tp = tcp_sk(sk);
__u32 cookie = ntohl(skb->h.th->ack_seq) - 1;
struct sock *ret = sk;
- struct open_request *req;
+ struct request_sock *req;
int mss;
struct rtable *rt;
__u8 rcv_wscale;
@@ -209,19 +211,20 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV);
- req = tcp_openreq_alloc();
ret = NULL;
+ req = reqsk_alloc(&tcp_request_sock_ops); /* for safety */
if (!req)
goto out;
- req->rcv_isn = htonl(skb->h.th->seq) - 1;
- req->snt_isn = cookie;
+ ireq = inet_rsk(req);
+ treq = tcp_rsk(req);
+ treq->rcv_isn = htonl(skb->h.th->seq) - 1;
+ treq->snt_isn = cookie;
req->mss = mss;
- req->rmt_port = skb->h.th->source;
- req->af.v4_req.loc_addr = skb->nh.iph->daddr;
- req->af.v4_req.rmt_addr = skb->nh.iph->saddr;
- req->class = &or_ipv4; /* for savety */
- req->af.v4_req.opt = NULL;
+ ireq->rmt_port = skb->h.th->source;
+ ireq->loc_addr = skb->nh.iph->daddr;
+ ireq->rmt_addr = skb->nh.iph->saddr;
+ ireq->opt = NULL;
/* We throwed the options of the initial SYN away, so we hope
* the ACK carries the same options again (see RFC1122 4.2.3.8)
@@ -229,17 +232,15 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
if (opt && opt->optlen) {
int opt_size = sizeof(struct ip_options) + opt->optlen;
- req->af.v4_req.opt = kmalloc(opt_size, GFP_ATOMIC);
- if (req->af.v4_req.opt) {
- if (ip_options_echo(req->af.v4_req.opt, skb)) {
- kfree(req->af.v4_req.opt);
- req->af.v4_req.opt = NULL;
- }
+ ireq->opt = kmalloc(opt_size, GFP_ATOMIC);
+ if (ireq->opt != NULL && ip_options_echo(ireq->opt, skb)) {
+ kfree(ireq->opt);
+ ireq->opt = NULL;
}
}
- req->snd_wscale = req->rcv_wscale = req->tstamp_ok = 0;
- req->wscale_ok = req->sack_ok = 0;
+ ireq->snd_wscale = ireq->rcv_wscale = ireq->tstamp_ok = 0;
+ ireq->wscale_ok = ireq->sack_ok = 0;
req->expires = 0UL;
req->retrans = 0;
@@ -253,15 +254,15 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
struct flowi fl = { .nl_u = { .ip4_u =
{ .daddr = ((opt && opt->srr) ?
opt->faddr :
- req->af.v4_req.rmt_addr),
- .saddr = req->af.v4_req.loc_addr,
+ ireq->rmt_addr),
+ .saddr = ireq->loc_addr,
.tos = RT_CONN_FLAGS(sk) } },
.proto = IPPROTO_TCP,
.uli_u = { .ports =
{ .sport = skb->h.th->dest,
.dport = skb->h.th->source } } };
if (ip_route_output_key(&rt, &fl)) {
- tcp_openreq_free(req);
+ reqsk_free(req);
goto out;
}
}
@@ -272,7 +273,7 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
&req->rcv_wnd, &req->window_clamp,
0, &rcv_wscale);
/* BTW win scale with syncookies is 0 by definition */
- req->rcv_wscale = rcv_wscale;
+ ireq->rcv_wscale = rcv_wscale;
ret = get_cookie_sock(sk, skb, req, &rt->u.dst);
out: return ret;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 23068bddbf0..e3289453241 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -118,6 +118,45 @@ static int ipv4_sysctl_forward_strategy(ctl_table *table,
return 1;
}
+static int proc_tcp_congestion_control(ctl_table *ctl, int write, struct file * filp,
+ void __user *buffer, size_t *lenp, loff_t *ppos)
+{
+ char val[TCP_CA_NAME_MAX];
+ ctl_table tbl = {
+ .data = val,
+ .maxlen = TCP_CA_NAME_MAX,
+ };
+ int ret;
+
+ tcp_get_default_congestion_control(val);
+
+ ret = proc_dostring(&tbl, write, filp, buffer, lenp, ppos);
+ if (write && ret == 0)
+ ret = tcp_set_default_congestion_control(val);
+ return ret;
+}
+
+int sysctl_tcp_congestion_control(ctl_table *table, int __user *name, int nlen,
+ void __user *oldval, size_t __user *oldlenp,
+ void __user *newval, size_t newlen,
+ void **context)
+{
+ char val[TCP_CA_NAME_MAX];
+ ctl_table tbl = {
+ .data = val,
+ .maxlen = TCP_CA_NAME_MAX,
+ };
+ int ret;
+
+ tcp_get_default_congestion_control(val);
+ ret = sysctl_string(&tbl, name, nlen, oldval, oldlenp, newval, newlen,
+ context);
+ if (ret == 0 && newval && newlen)
+ ret = tcp_set_default_congestion_control(val);
+ return ret;
+}
+
+
ctl_table ipv4_table[] = {
{
.ctl_name = NET_IPV4_TCP_TIMESTAMPS,
@@ -612,70 +651,6 @@ ctl_table ipv4_table[] = {
.proc_handler = &proc_dointvec,
},
{
- .ctl_name = NET_TCP_WESTWOOD,
- .procname = "tcp_westwood",
- .data = &sysctl_tcp_westwood,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_VEGAS,
- .procname = "tcp_vegas_cong_avoid",
- .data = &sysctl_tcp_vegas_cong_avoid,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_VEGAS_ALPHA,
- .procname = "tcp_vegas_alpha",
- .data = &sysctl_tcp_vegas_alpha,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_VEGAS_BETA,
- .procname = "tcp_vegas_beta",
- .data = &sysctl_tcp_vegas_beta,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_VEGAS_GAMMA,
- .procname = "tcp_vegas_gamma",
- .data = &sysctl_tcp_vegas_gamma,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_BIC,
- .procname = "tcp_bic",
- .data = &sysctl_tcp_bic,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_BIC_FAST_CONVERGENCE,
- .procname = "tcp_bic_fast_convergence",
- .data = &sysctl_tcp_bic_fast_convergence,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
- .ctl_name = NET_TCP_BIC_LOW_WINDOW,
- .procname = "tcp_bic_low_window",
- .data = &sysctl_tcp_bic_low_window,
- .maxlen = sizeof(int),
- .mode = 0644,
- .proc_handler = &proc_dointvec,
- },
- {
.ctl_name = NET_TCP_MODERATE_RCVBUF,
.procname = "tcp_moderate_rcvbuf",
.data = &sysctl_tcp_moderate_rcvbuf,
@@ -692,13 +667,14 @@ ctl_table ipv4_table[] = {
.proc_handler = &proc_dointvec,
},
{
- .ctl_name = NET_TCP_BIC_BETA,
- .procname = "tcp_bic_beta",
- .data = &sysctl_tcp_bic_beta,
- .maxlen = sizeof(int),
+ .ctl_name = NET_TCP_CONG_CONTROL,
+ .procname = "tcp_congestion_control",
.mode = 0644,
- .proc_handler = &proc_dointvec,
+ .maxlen = TCP_CA_NAME_MAX,
+ .proc_handler = &proc_tcp_congestion_control,
+ .strategy = &sysctl_tcp_congestion_control,
},
+
{ .ctl_name = 0 }
};
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 0d9a4fd5f1a..ddb6ce4ecff 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -271,7 +271,6 @@ int sysctl_tcp_fin_timeout = TCP_FIN_TIMEOUT;
DEFINE_SNMP_STAT(struct tcp_mib, tcp_statistics);
-kmem_cache_t *tcp_openreq_cachep;
kmem_cache_t *tcp_bucket_cachep;
kmem_cache_t *tcp_timewait_cachep;
@@ -317,7 +316,7 @@ EXPORT_SYMBOL(tcp_enter_memory_pressure);
static __inline__ unsigned int tcp_listen_poll(struct sock *sk,
poll_table *wait)
{
- return tcp_sk(sk)->accept_queue ? (POLLIN | POLLRDNORM) : 0;
+ return !reqsk_queue_empty(&tcp_sk(sk)->accept_queue) ? (POLLIN | POLLRDNORM) : 0;
}
/*
@@ -463,28 +462,15 @@ int tcp_listen_start(struct sock *sk)
{
struct inet_sock *inet = inet_sk(sk);
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt;
+ int rc = reqsk_queue_alloc(&tp->accept_queue, TCP_SYNQ_HSIZE);
+
+ if (rc != 0)
+ return rc;
sk->sk_max_ack_backlog = 0;
sk->sk_ack_backlog = 0;
- tp->accept_queue = tp->accept_queue_tail = NULL;
- rwlock_init(&tp->syn_wait_lock);
tcp_delack_init(tp);
- lopt = kmalloc(sizeof(struct tcp_listen_opt), GFP_KERNEL);
- if (!lopt)
- return -ENOMEM;
-
- memset(lopt, 0, sizeof(struct tcp_listen_opt));
- for (lopt->max_qlen_log = 6; ; lopt->max_qlen_log++)
- if ((1 << lopt->max_qlen_log) >= sysctl_max_syn_backlog)
- break;
- get_random_bytes(&lopt->hash_rnd, 4);
-
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = lopt;
- write_unlock_bh(&tp->syn_wait_lock);
-
/* There is race window here: we announce ourselves listening,
* but this transition is still not validated by get_port().
* It is OK, because this socket enters to hash table only
@@ -501,10 +487,7 @@ int tcp_listen_start(struct sock *sk)
}
sk->sk_state = TCP_CLOSE;
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = NULL;
- write_unlock_bh(&tp->syn_wait_lock);
- kfree(lopt);
+ reqsk_queue_destroy(&tp->accept_queue);
return -EADDRINUSE;
}
@@ -516,25 +499,23 @@ int tcp_listen_start(struct sock *sk)
static void tcp_listen_stop (struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
- struct open_request *acc_req = tp->accept_queue;
- struct open_request *req;
+ struct listen_sock *lopt;
+ struct request_sock *acc_req;
+ struct request_sock *req;
int i;
tcp_delete_keepalive_timer(sk);
/* make all the listen_opt local to us */
- write_lock_bh(&tp->syn_wait_lock);
- tp->listen_opt = NULL;
- write_unlock_bh(&tp->syn_wait_lock);
- tp->accept_queue = tp->accept_queue_tail = NULL;
+ lopt = reqsk_queue_yank_listen_sk(&tp->accept_queue);
+ acc_req = reqsk_queue_yank_acceptq(&tp->accept_queue);
if (lopt->qlen) {
for (i = 0; i < TCP_SYNQ_HSIZE; i++) {
while ((req = lopt->syn_table[i]) != NULL) {
lopt->syn_table[i] = req->dl_next;
lopt->qlen--;
- tcp_openreq_free(req);
+ reqsk_free(req);
/* Following specs, it would be better either to send FIN
* (and enter FIN-WAIT-1, it is normal close)
@@ -574,7 +555,7 @@ static void tcp_listen_stop (struct sock *sk)
sock_put(child);
sk_acceptq_removed(sk);
- tcp_openreq_fastfree(req);
+ __reqsk_free(req);
}
BUG_TRAP(!sk->sk_ack_backlog);
}
@@ -634,7 +615,7 @@ static ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffse
size_t psize, int flags)
{
struct tcp_sock *tp = tcp_sk(sk);
- int mss_now;
+ int mss_now, size_goal;
int err;
ssize_t copied;
long timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
@@ -647,6 +628,7 @@ static ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffse
clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
mss_now = tcp_current_mss(sk, !(flags&MSG_OOB));
+ size_goal = tp->xmit_size_goal;
copied = 0;
err = -EPIPE;
@@ -660,7 +642,7 @@ static ssize_t do_tcp_sendpages(struct sock *sk, struct page **pages, int poffse
int offset = poffset % PAGE_SIZE;
int size = min_t(size_t, psize, PAGE_SIZE - offset);
- if (!sk->sk_send_head || (copy = mss_now - skb->len) <= 0) {
+ if (!sk->sk_send_head || (copy = size_goal - skb->len) <= 0) {
new_segment:
if (!sk_stream_memory_free(sk))
goto wait_for_sndbuf;
@@ -671,7 +653,7 @@ new_segment:
goto wait_for_memory;
skb_entail(sk, tp, skb);
- copy = mss_now;
+ copy = size_goal;
}
if (copy > size)
@@ -712,7 +694,7 @@ new_segment:
if (!(psize -= copy))
goto out;
- if (skb->len != mss_now || (flags & MSG_OOB))
+ if (skb->len < mss_now || (flags & MSG_OOB))
continue;
if (forced_push(tp)) {
@@ -732,6 +714,7 @@ wait_for_memory:
goto do_error;
mss_now = tcp_current_mss(sk, !(flags&MSG_OOB));
+ size_goal = tp->xmit_size_goal;
}
out:
@@ -773,15 +756,20 @@ ssize_t tcp_sendpage(struct socket *sock, struct page *page, int offset,
static inline int select_size(struct sock *sk, struct tcp_sock *tp)
{
- int tmp = tp->mss_cache_std;
+ int tmp = tp->mss_cache;
if (sk->sk_route_caps & NETIF_F_SG) {
- int pgbreak = SKB_MAX_HEAD(MAX_TCP_HEADER);
+ if (sk->sk_route_caps & NETIF_F_TSO)
+ tmp = 0;
+ else {
+ int pgbreak = SKB_MAX_HEAD(MAX_TCP_HEADER);
- if (tmp >= pgbreak &&
- tmp <= pgbreak + (MAX_SKB_FRAGS - 1) * PAGE_SIZE)
- tmp = pgbreak;
+ if (tmp >= pgbreak &&
+ tmp <= pgbreak + (MAX_SKB_FRAGS - 1) * PAGE_SIZE)
+ tmp = pgbreak;
+ }
}
+
return tmp;
}
@@ -792,7 +780,7 @@ int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
int iovlen, flags;
- int mss_now;
+ int mss_now, size_goal;
int err, copied;
long timeo;
@@ -811,6 +799,7 @@ int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
clear_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
mss_now = tcp_current_mss(sk, !(flags&MSG_OOB));
+ size_goal = tp->xmit_size_goal;
/* Ok commence sending. */
iovlen = msg->msg_iovlen;
@@ -833,7 +822,7 @@ int tcp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
skb = sk->sk_write_queue.prev;
if (!sk->sk_send_head ||
- (copy = mss_now - skb->len) <= 0) {
+ (copy = size_goal - skb->len) <= 0) {
new_segment:
/* Allocate new segment. If the interface is SG,
@@ -856,7 +845,7 @@ new_segment:
skb->ip_summed = CHECKSUM_HW;
skb_entail(sk, tp, skb);
- copy = mss_now;
+ copy = size_goal;
}
/* Try to append data to the end of skb. */
@@ -891,11 +880,6 @@ new_segment:
tcp_mark_push(tp, skb);
goto new_segment;
} else if (page) {
- /* If page is cached, align
- * offset to L1 cache boundary
- */
- off = (off + L1_CACHE_BYTES - 1) &
- ~(L1_CACHE_BYTES - 1);
if (off == PAGE_SIZE) {
put_page(page);
TCP_PAGE(sk) = page = NULL;
@@ -956,7 +940,7 @@ new_segment:
if ((seglen -= copy) == 0 && iovlen == 0)
goto out;
- if (skb->len != mss_now || (flags & MSG_OOB))
+ if (skb->len < mss_now || (flags & MSG_OOB))
continue;
if (forced_push(tp)) {
@@ -976,6 +960,7 @@ wait_for_memory:
goto do_error;
mss_now = tcp_current_mss(sk, !(flags&MSG_OOB));
+ size_goal = tp->xmit_size_goal;
}
}
@@ -1120,7 +1105,7 @@ static void tcp_prequeue_process(struct sock *sk)
struct sk_buff *skb;
struct tcp_sock *tp = tcp_sk(sk);
- NET_ADD_STATS_USER(LINUX_MIB_TCPPREQUEUED, skb_queue_len(&tp->ucopy.prequeue));
+ NET_INC_STATS_USER(LINUX_MIB_TCPPREQUEUED);
/* RX process wants to run with disabled BHs, though it is not
* necessary */
@@ -1345,7 +1330,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
cleanup_rbuf(sk, copied);
- if (tp->ucopy.task == user_recv) {
+ if (!sysctl_tcp_low_latency && tp->ucopy.task == user_recv) {
/* Install new reader */
if (!user_recv && !(flags & (MSG_TRUNC | MSG_PEEK))) {
user_recv = current;
@@ -1384,7 +1369,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
* is not empty. It is more elegant, but eats cycles,
* unfortunately.
*/
- if (skb_queue_len(&tp->ucopy.prequeue))
+ if (!skb_queue_empty(&tp->ucopy.prequeue))
goto do_prequeue;
/* __ Set realtime policy in scheduler __ */
@@ -1409,7 +1394,7 @@ int tcp_recvmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
}
if (tp->rcv_nxt == tp->copied_seq &&
- skb_queue_len(&tp->ucopy.prequeue)) {
+ !skb_queue_empty(&tp->ucopy.prequeue)) {
do_prequeue:
tcp_prequeue_process(sk);
@@ -1491,7 +1476,7 @@ skip_copy:
} while (len > 0);
if (user_recv) {
- if (skb_queue_len(&tp->ucopy.prequeue)) {
+ if (!skb_queue_empty(&tp->ucopy.prequeue)) {
int chunk;
tp->ucopy.len = copied > 0 ? len : 0;
@@ -1868,11 +1853,11 @@ static int wait_for_connect(struct sock *sk, long timeo)
prepare_to_wait_exclusive(sk->sk_sleep, &wait,
TASK_INTERRUPTIBLE);
release_sock(sk);
- if (!tp->accept_queue)
+ if (reqsk_queue_empty(&tp->accept_queue))
timeo = schedule_timeout(timeo);
lock_sock(sk);
err = 0;
- if (tp->accept_queue)
+ if (!reqsk_queue_empty(&tp->accept_queue))
break;
err = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
@@ -1895,7 +1880,6 @@ static int wait_for_connect(struct sock *sk, long timeo)
struct sock *tcp_accept(struct sock *sk, int flags, int *err)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct open_request *req;
struct sock *newsk;
int error;
@@ -1906,37 +1890,31 @@ struct sock *tcp_accept(struct sock *sk, int flags, int *err)
*/
error = -EINVAL;
if (sk->sk_state != TCP_LISTEN)
- goto out;
+ goto out_err;
/* Find already established connection */
- if (!tp->accept_queue) {
+ if (reqsk_queue_empty(&tp->accept_queue)) {
long timeo = sock_rcvtimeo(sk, flags & O_NONBLOCK);
/* If this is a non blocking socket don't sleep */
error = -EAGAIN;
if (!timeo)
- goto out;
+ goto out_err;
error = wait_for_connect(sk, timeo);
if (error)
- goto out;
+ goto out_err;
}
- req = tp->accept_queue;
- if ((tp->accept_queue = req->dl_next) == NULL)
- tp->accept_queue_tail = NULL;
-
- newsk = req->sk;
- sk_acceptq_removed(sk);
- tcp_openreq_fastfree(req);
+ newsk = reqsk_queue_get_child(&tp->accept_queue, sk);
BUG_TRAP(newsk->sk_state != TCP_SYN_RECV);
- release_sock(sk);
- return newsk;
-
out:
release_sock(sk);
+ return newsk;
+out_err:
+ newsk = NULL;
*err = error;
- return NULL;
+ goto out;
}
/*
@@ -1953,6 +1931,25 @@ int tcp_setsockopt(struct sock *sk, int level, int optname, char __user *optval,
return tp->af_specific->setsockopt(sk, level, optname,
optval, optlen);
+ /* This is a string value all the others are int's */
+ if (optname == TCP_CONGESTION) {
+ char name[TCP_CA_NAME_MAX];
+
+ if (optlen < 1)
+ return -EINVAL;
+
+ val = strncpy_from_user(name, optval,
+ min(TCP_CA_NAME_MAX-1, optlen));
+ if (val < 0)
+ return -EFAULT;
+ name[val] = 0;
+
+ lock_sock(sk);
+ err = tcp_set_congestion_control(tp, name);
+ release_sock(sk);
+ return err;
+ }
+
if (optlen < sizeof(int))
return -EINVAL;
@@ -2135,7 +2132,7 @@ void tcp_get_info(struct sock *sk, struct tcp_info *info)
info->tcpi_rto = jiffies_to_usecs(tp->rto);
info->tcpi_ato = jiffies_to_usecs(tp->ack.ato);
- info->tcpi_snd_mss = tp->mss_cache_std;
+ info->tcpi_snd_mss = tp->mss_cache;
info->tcpi_rcv_mss = tp->ack.rcv_mss;
info->tcpi_unacked = tp->packets_out;
@@ -2185,7 +2182,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
switch (optname) {
case TCP_MAXSEG:
- val = tp->mss_cache_std;
+ val = tp->mss_cache;
if (!val && ((1 << sk->sk_state) & (TCPF_CLOSE | TCPF_LISTEN)))
val = tp->rx_opt.user_mss;
break;
@@ -2237,6 +2234,16 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
case TCP_QUICKACK:
val = !tp->ack.pingpong;
break;
+
+ case TCP_CONGESTION:
+ if (get_user(len, optlen))
+ return -EFAULT;
+ len = min_t(unsigned int, len, TCP_CA_NAME_MAX);
+ if (put_user(len, optlen))
+ return -EFAULT;
+ if (copy_to_user(optval, tp->ca_ops->name, len))
+ return -EFAULT;
+ return 0;
default:
return -ENOPROTOOPT;
};
@@ -2250,7 +2257,7 @@ int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
extern void __skb_cb_too_small_for_tcp(int, int);
-extern void tcpdiag_init(void);
+extern struct tcp_congestion_ops tcp_reno;
static __initdata unsigned long thash_entries;
static int __init set_thash_entries(char *str)
@@ -2271,13 +2278,6 @@ void __init tcp_init(void)
__skb_cb_too_small_for_tcp(sizeof(struct tcp_skb_cb),
sizeof(skb->cb));
- tcp_openreq_cachep = kmem_cache_create("tcp_open_request",
- sizeof(struct open_request),
- 0, SLAB_HWCACHE_ALIGN,
- NULL, NULL);
- if (!tcp_openreq_cachep)
- panic("tcp_init: Cannot alloc open_request cache.");
-
tcp_bucket_cachep = kmem_cache_create("tcp_bind_bucket",
sizeof(struct tcp_bind_bucket),
0, SLAB_HWCACHE_ALIGN,
@@ -2366,6 +2366,8 @@ void __init tcp_init(void)
printk(KERN_INFO "TCP: Hash tables configured "
"(established %d bind %d)\n",
tcp_ehash_size << 1, tcp_bhash_size);
+
+ tcp_register_congestion_control(&tcp_reno);
}
EXPORT_SYMBOL(tcp_accept);
@@ -2374,7 +2376,6 @@ EXPORT_SYMBOL(tcp_destroy_sock);
EXPORT_SYMBOL(tcp_disconnect);
EXPORT_SYMBOL(tcp_getsockopt);
EXPORT_SYMBOL(tcp_ioctl);
-EXPORT_SYMBOL(tcp_openreq_cachep);
EXPORT_SYMBOL(tcp_poll);
EXPORT_SYMBOL(tcp_read_sock);
EXPORT_SYMBOL(tcp_recvmsg);
diff --git a/net/ipv4/tcp_bic.c b/net/ipv4/tcp_bic.c
new file mode 100644
index 00000000000..ec38d45d664
--- /dev/null
+++ b/net/ipv4/tcp_bic.c
@@ -0,0 +1,331 @@
+/*
+ * Binary Increase Congestion control for TCP
+ *
+ * This is from the implementation of BICTCP in
+ * Lison-Xu, Kahaled Harfoush, and Injong Rhee.
+ * "Binary Increase Congestion Control for Fast, Long Distance
+ * Networks" in InfoComm 2004
+ * Available from:
+ * http://www.csc.ncsu.edu/faculty/rhee/export/bitcp.pdf
+ *
+ * Unless BIC is enabled and congestion window is large
+ * this behaves the same as the original Reno.
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <net/tcp.h>
+
+
+#define BICTCP_BETA_SCALE 1024 /* Scale factor beta calculation
+ * max_cwnd = snd_cwnd * beta
+ */
+#define BICTCP_B 4 /*
+ * In binary search,
+ * go to point (max+min)/N
+ */
+
+static int fast_convergence = 1;
+static int max_increment = 32;
+static int low_window = 14;
+static int beta = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */
+static int low_utilization_threshold = 153;
+static int low_utilization_period = 2;
+static int initial_ssthresh = 100;
+static int smooth_part = 20;
+
+module_param(fast_convergence, int, 0644);
+MODULE_PARM_DESC(fast_convergence, "turn on/off fast convergence");
+module_param(max_increment, int, 0644);
+MODULE_PARM_DESC(max_increment, "Limit on increment allowed during binary search");
+module_param(low_window, int, 0644);
+MODULE_PARM_DESC(low_window, "lower bound on congestion window (for TCP friendliness)");
+module_param(beta, int, 0644);
+MODULE_PARM_DESC(beta, "beta for multiplicative increase");
+module_param(low_utilization_threshold, int, 0644);
+MODULE_PARM_DESC(low_utilization_threshold, "percent (scaled by 1024) for low utilization mode");
+module_param(low_utilization_period, int, 0644);
+MODULE_PARM_DESC(low_utilization_period, "if average delay exceeds then goto to low utilization mode (seconds)");
+module_param(initial_ssthresh, int, 0644);
+MODULE_PARM_DESC(initial_ssthresh, "initial value of slow start threshold");
+module_param(smooth_part, int, 0644);
+MODULE_PARM_DESC(smooth_part, "log(B/(B*Smin))/log(B/(B-1))+B, # of RTT from Wmax-B to Wmax");
+
+
+/* BIC TCP Parameters */
+struct bictcp {
+ u32 cnt; /* increase cwnd by 1 after ACKs */
+ u32 last_max_cwnd; /* last maximum snd_cwnd */
+ u32 loss_cwnd; /* congestion window at last loss */
+ u32 last_cwnd; /* the last snd_cwnd */
+ u32 last_time; /* time when updated last_cwnd */
+ u32 delay_min; /* min delay */
+ u32 delay_max; /* max delay */
+ u32 last_delay;
+ u8 low_utilization;/* 0: high; 1: low */
+ u32 low_utilization_start; /* starting time of low utilization detection*/
+ u32 epoch_start; /* beginning of an epoch */
+#define ACK_RATIO_SHIFT 4
+ u32 delayed_ack; /* estimate the ratio of Packets/ACKs << 4 */
+};
+
+static inline void bictcp_reset(struct bictcp *ca)
+{
+ ca->cnt = 0;
+ ca->last_max_cwnd = 0;
+ ca->loss_cwnd = 0;
+ ca->last_cwnd = 0;
+ ca->last_time = 0;
+ ca->delay_min = 0;
+ ca->delay_max = 0;
+ ca->last_delay = 0;
+ ca->low_utilization = 0;
+ ca->low_utilization_start = 0;
+ ca->epoch_start = 0;
+ ca->delayed_ack = 2 << ACK_RATIO_SHIFT;
+}
+
+static void bictcp_init(struct tcp_sock *tp)
+{
+ bictcp_reset(tcp_ca(tp));
+ if (initial_ssthresh)
+ tp->snd_ssthresh = initial_ssthresh;
+}
+
+/*
+ * Compute congestion window to use.
+ */
+static inline void bictcp_update(struct bictcp *ca, u32 cwnd)
+{
+ if (ca->last_cwnd == cwnd &&
+ (s32)(tcp_time_stamp - ca->last_time) <= HZ / 32)
+ return;
+
+ ca->last_cwnd = cwnd;
+ ca->last_time = tcp_time_stamp;
+
+ if (ca->epoch_start == 0) /* record the beginning of an epoch */
+ ca->epoch_start = tcp_time_stamp;
+
+ /* start off normal */
+ if (cwnd <= low_window) {
+ ca->cnt = cwnd;
+ return;
+ }
+
+ /* binary increase */
+ if (cwnd < ca->last_max_cwnd) {
+ __u32 dist = (ca->last_max_cwnd - cwnd)
+ / BICTCP_B;
+
+ if (dist > max_increment)
+ /* linear increase */
+ ca->cnt = cwnd / max_increment;
+ else if (dist <= 1U)
+ /* binary search increase */
+ ca->cnt = (cwnd * smooth_part) / BICTCP_B;
+ else
+ /* binary search increase */
+ ca->cnt = cwnd / dist;
+ } else {
+ /* slow start AMD linear increase */
+ if (cwnd < ca->last_max_cwnd + BICTCP_B)
+ /* slow start */
+ ca->cnt = (cwnd * smooth_part) / BICTCP_B;
+ else if (cwnd < ca->last_max_cwnd + max_increment*(BICTCP_B-1))
+ /* slow start */
+ ca->cnt = (cwnd * (BICTCP_B-1))
+ / cwnd-ca->last_max_cwnd;
+ else
+ /* linear increase */
+ ca->cnt = cwnd / max_increment;
+ }
+
+ /* if in slow start or link utilization is very low */
+ if ( ca->loss_cwnd == 0 ||
+ (cwnd > ca->loss_cwnd && ca->low_utilization)) {
+ if (ca->cnt > 20) /* increase cwnd 5% per RTT */
+ ca->cnt = 20;
+ }
+
+ ca->cnt = (ca->cnt << ACK_RATIO_SHIFT) / ca->delayed_ack;
+ if (ca->cnt == 0) /* cannot be zero */
+ ca->cnt = 1;
+}
+
+
+/* Detect low utilization in congestion avoidance */
+static inline void bictcp_low_utilization(struct tcp_sock *tp, int flag)
+{
+ struct bictcp *ca = tcp_ca(tp);
+ u32 dist, delay;
+
+ /* No time stamp */
+ if (!(tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr) ||
+ /* Discard delay samples right after fast recovery */
+ tcp_time_stamp < ca->epoch_start + HZ ||
+ /* this delay samples may not be accurate */
+ flag == 0) {
+ ca->last_delay = 0;
+ goto notlow;
+ }
+
+ delay = ca->last_delay<<3; /* use the same scale as tp->srtt*/
+ ca->last_delay = tcp_time_stamp - tp->rx_opt.rcv_tsecr;
+ if (delay == 0) /* no previous delay sample */
+ goto notlow;
+
+ /* first time call or link delay decreases */
+ if (ca->delay_min == 0 || ca->delay_min > delay) {
+ ca->delay_min = ca->delay_max = delay;
+ goto notlow;
+ }
+
+ if (ca->delay_max < delay)
+ ca->delay_max = delay;
+
+ /* utilization is low, if avg delay < dist*threshold
+ for checking_period time */
+ dist = ca->delay_max - ca->delay_min;
+ if (dist <= ca->delay_min>>6 ||
+ tp->srtt - ca->delay_min >= (dist*low_utilization_threshold)>>10)
+ goto notlow;
+
+ if (ca->low_utilization_start == 0) {
+ ca->low_utilization = 0;
+ ca->low_utilization_start = tcp_time_stamp;
+ } else if ((s32)(tcp_time_stamp - ca->low_utilization_start)
+ > low_utilization_period*HZ) {
+ ca->low_utilization = 1;
+ }
+
+ return;
+
+ notlow:
+ ca->low_utilization = 0;
+ ca->low_utilization_start = 0;
+
+}
+
+static void bictcp_cong_avoid(struct tcp_sock *tp, u32 ack,
+ u32 seq_rtt, u32 in_flight, int data_acked)
+{
+ struct bictcp *ca = tcp_ca(tp);
+
+ bictcp_low_utilization(tp, data_acked);
+
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (tp->snd_cwnd <= tp->snd_ssthresh) {
+ /* In "safe" area, increase. */
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ } else {
+ bictcp_update(ca, tp->snd_cwnd);
+
+ /* In dangerous area, increase slowly.
+ * In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd
+ */
+ if (tp->snd_cwnd_cnt >= ca->cnt) {
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ tp->snd_cwnd_cnt = 0;
+ } else
+ tp->snd_cwnd_cnt++;
+ }
+
+}
+
+/*
+ * behave like Reno until low_window is reached,
+ * then increase congestion window slowly
+ */
+static u32 bictcp_recalc_ssthresh(struct tcp_sock *tp)
+{
+ struct bictcp *ca = tcp_ca(tp);
+
+ ca->epoch_start = 0; /* end of epoch */
+
+ /* in case of wrong delay_max*/
+ if (ca->delay_min > 0 && ca->delay_max > ca->delay_min)
+ ca->delay_max = ca->delay_min
+ + ((ca->delay_max - ca->delay_min)* 90) / 100;
+
+ /* Wmax and fast convergence */
+ if (tp->snd_cwnd < ca->last_max_cwnd && fast_convergence)
+ ca->last_max_cwnd = (tp->snd_cwnd * (BICTCP_BETA_SCALE + beta))
+ / (2 * BICTCP_BETA_SCALE);
+ else
+ ca->last_max_cwnd = tp->snd_cwnd;
+
+ ca->loss_cwnd = tp->snd_cwnd;
+
+
+ if (tp->snd_cwnd <= low_window)
+ return max(tp->snd_cwnd >> 1U, 2U);
+ else
+ return max((tp->snd_cwnd * beta) / BICTCP_BETA_SCALE, 2U);
+}
+
+static u32 bictcp_undo_cwnd(struct tcp_sock *tp)
+{
+ struct bictcp *ca = tcp_ca(tp);
+
+ return max(tp->snd_cwnd, ca->last_max_cwnd);
+}
+
+static u32 bictcp_min_cwnd(struct tcp_sock *tp)
+{
+ return tp->snd_ssthresh;
+}
+
+static void bictcp_state(struct tcp_sock *tp, u8 new_state)
+{
+ if (new_state == TCP_CA_Loss)
+ bictcp_reset(tcp_ca(tp));
+}
+
+/* Track delayed acknowledgement ratio using sliding window
+ * ratio = (15*ratio + sample) / 16
+ */
+static void bictcp_acked(struct tcp_sock *tp, u32 cnt)
+{
+ if (cnt > 0 && tp->ca_state == TCP_CA_Open) {
+ struct bictcp *ca = tcp_ca(tp);
+ cnt -= ca->delayed_ack >> ACK_RATIO_SHIFT;
+ ca->delayed_ack += cnt;
+ }
+}
+
+
+static struct tcp_congestion_ops bictcp = {
+ .init = bictcp_init,
+ .ssthresh = bictcp_recalc_ssthresh,
+ .cong_avoid = bictcp_cong_avoid,
+ .set_state = bictcp_state,
+ .undo_cwnd = bictcp_undo_cwnd,
+ .min_cwnd = bictcp_min_cwnd,
+ .pkts_acked = bictcp_acked,
+ .owner = THIS_MODULE,
+ .name = "bic",
+};
+
+static int __init bictcp_register(void)
+{
+ BUG_ON(sizeof(struct bictcp) > TCP_CA_PRIV_SIZE);
+ return tcp_register_congestion_control(&bictcp);
+}
+
+static void __exit bictcp_unregister(void)
+{
+ tcp_unregister_congestion_control(&bictcp);
+}
+
+module_init(bictcp_register);
+module_exit(bictcp_unregister);
+
+MODULE_AUTHOR("Stephen Hemminger");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("BIC TCP");
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
new file mode 100644
index 00000000000..4970d10a778
--- /dev/null
+++ b/net/ipv4/tcp_cong.c
@@ -0,0 +1,237 @@
+/*
+ * Plugable TCP congestion control support and newReno
+ * congestion control.
+ * Based on ideas from I/O scheduler suport and Web100.
+ *
+ * Copyright (C) 2005 Stephen Hemminger <shemminger@osdl.org>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/types.h>
+#include <linux/list.h>
+#include <net/tcp.h>
+
+static DEFINE_SPINLOCK(tcp_cong_list_lock);
+static LIST_HEAD(tcp_cong_list);
+
+/* Simple linear search, don't expect many entries! */
+static struct tcp_congestion_ops *tcp_ca_find(const char *name)
+{
+ struct tcp_congestion_ops *e;
+
+ list_for_each_entry_rcu(e, &tcp_cong_list, list) {
+ if (strcmp(e->name, name) == 0)
+ return e;
+ }
+
+ return NULL;
+}
+
+/*
+ * Attach new congestion control algorthim to the list
+ * of available options.
+ */
+int tcp_register_congestion_control(struct tcp_congestion_ops *ca)
+{
+ int ret = 0;
+
+ /* all algorithms must implement ssthresh and cong_avoid ops */
+ if (!ca->ssthresh || !ca->cong_avoid || !ca->min_cwnd) {
+ printk(KERN_ERR "TCP %s does not implement required ops\n",
+ ca->name);
+ return -EINVAL;
+ }
+
+ spin_lock(&tcp_cong_list_lock);
+ if (tcp_ca_find(ca->name)) {
+ printk(KERN_NOTICE "TCP %s already registered\n", ca->name);
+ ret = -EEXIST;
+ } else {
+ list_add_rcu(&ca->list, &tcp_cong_list);
+ printk(KERN_INFO "TCP %s registered\n", ca->name);
+ }
+ spin_unlock(&tcp_cong_list_lock);
+
+ return ret;
+}
+EXPORT_SYMBOL_GPL(tcp_register_congestion_control);
+
+/*
+ * Remove congestion control algorithm, called from
+ * the module's remove function. Module ref counts are used
+ * to ensure that this can't be done till all sockets using
+ * that method are closed.
+ */
+void tcp_unregister_congestion_control(struct tcp_congestion_ops *ca)
+{
+ spin_lock(&tcp_cong_list_lock);
+ list_del_rcu(&ca->list);
+ spin_unlock(&tcp_cong_list_lock);
+}
+EXPORT_SYMBOL_GPL(tcp_unregister_congestion_control);
+
+/* Assign choice of congestion control. */
+void tcp_init_congestion_control(struct tcp_sock *tp)
+{
+ struct tcp_congestion_ops *ca;
+
+ if (tp->ca_ops != &tcp_init_congestion_ops)
+ return;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(ca, &tcp_cong_list, list) {
+ if (try_module_get(ca->owner)) {
+ tp->ca_ops = ca;
+ break;
+ }
+
+ }
+ rcu_read_unlock();
+
+ if (tp->ca_ops->init)
+ tp->ca_ops->init(tp);
+}
+
+/* Manage refcounts on socket close. */
+void tcp_cleanup_congestion_control(struct tcp_sock *tp)
+{
+ if (tp->ca_ops->release)
+ tp->ca_ops->release(tp);
+ module_put(tp->ca_ops->owner);
+}
+
+/* Used by sysctl to change default congestion control */
+int tcp_set_default_congestion_control(const char *name)
+{
+ struct tcp_congestion_ops *ca;
+ int ret = -ENOENT;
+
+ spin_lock(&tcp_cong_list_lock);
+ ca = tcp_ca_find(name);
+#ifdef CONFIG_KMOD
+ if (!ca) {
+ spin_unlock(&tcp_cong_list_lock);
+
+ request_module("tcp_%s", name);
+ spin_lock(&tcp_cong_list_lock);
+ ca = tcp_ca_find(name);
+ }
+#endif
+
+ if (ca) {
+ list_move(&ca->list, &tcp_cong_list);
+ ret = 0;
+ }
+ spin_unlock(&tcp_cong_list_lock);
+
+ return ret;
+}
+
+/* Get current default congestion control */
+void tcp_get_default_congestion_control(char *name)
+{
+ struct tcp_congestion_ops *ca;
+ /* We will always have reno... */
+ BUG_ON(list_empty(&tcp_cong_list));
+
+ rcu_read_lock();
+ ca = list_entry(tcp_cong_list.next, struct tcp_congestion_ops, list);
+ strncpy(name, ca->name, TCP_CA_NAME_MAX);
+ rcu_read_unlock();
+}
+
+/* Change congestion control for socket */
+int tcp_set_congestion_control(struct tcp_sock *tp, const char *name)
+{
+ struct tcp_congestion_ops *ca;
+ int err = 0;
+
+ rcu_read_lock();
+ ca = tcp_ca_find(name);
+ if (ca == tp->ca_ops)
+ goto out;
+
+ if (!ca)
+ err = -ENOENT;
+
+ else if (!try_module_get(ca->owner))
+ err = -EBUSY;
+
+ else {
+ tcp_cleanup_congestion_control(tp);
+ tp->ca_ops = ca;
+ if (tp->ca_ops->init)
+ tp->ca_ops->init(tp);
+ }
+ out:
+ rcu_read_unlock();
+ return err;
+}
+
+/*
+ * TCP Reno congestion control
+ * This is special case used for fallback as well.
+ */
+/* This is Jacobson's slow start and congestion avoidance.
+ * SIGCOMM '88, p. 328.
+ */
+void tcp_reno_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt, u32 in_flight,
+ int flag)
+{
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (tp->snd_cwnd <= tp->snd_ssthresh) {
+ /* In "safe" area, increase. */
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ } else {
+ /* In dangerous area, increase slowly.
+ * In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd
+ */
+ if (tp->snd_cwnd_cnt >= tp->snd_cwnd) {
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ tp->snd_cwnd_cnt = 0;
+ } else
+ tp->snd_cwnd_cnt++;
+ }
+}
+EXPORT_SYMBOL_GPL(tcp_reno_cong_avoid);
+
+/* Slow start threshold is half the congestion window (min 2) */
+u32 tcp_reno_ssthresh(struct tcp_sock *tp)
+{
+ return max(tp->snd_cwnd >> 1U, 2U);
+}
+EXPORT_SYMBOL_GPL(tcp_reno_ssthresh);
+
+/* Lower bound on congestion window. */
+u32 tcp_reno_min_cwnd(struct tcp_sock *tp)
+{
+ return tp->snd_ssthresh/2;
+}
+EXPORT_SYMBOL_GPL(tcp_reno_min_cwnd);
+
+struct tcp_congestion_ops tcp_reno = {
+ .name = "reno",
+ .owner = THIS_MODULE,
+ .ssthresh = tcp_reno_ssthresh,
+ .cong_avoid = tcp_reno_cong_avoid,
+ .min_cwnd = tcp_reno_min_cwnd,
+};
+
+/* Initial congestion control used (until SYN)
+ * really reno under another name so we can tell difference
+ * during tcp_set_default_congestion_control
+ */
+struct tcp_congestion_ops tcp_init_congestion_ops = {
+ .name = "",
+ .owner = THIS_MODULE,
+ .ssthresh = tcp_reno_ssthresh,
+ .cong_avoid = tcp_reno_cong_avoid,
+ .min_cwnd = tcp_reno_min_cwnd,
+};
+EXPORT_SYMBOL_GPL(tcp_init_congestion_ops);
diff --git a/net/ipv4/tcp_diag.c b/net/ipv4/tcp_diag.c
index 8faa8948f75..f66945cb158 100644
--- a/net/ipv4/tcp_diag.c
+++ b/net/ipv4/tcp_diag.c
@@ -42,15 +42,8 @@ struct tcpdiag_entry
static struct sock *tcpnl;
-
#define TCPDIAG_PUT(skb, attrtype, attrlen) \
-({ int rtalen = RTA_LENGTH(attrlen); \
- struct rtattr *rta; \
- if (skb_tailroom(skb) < RTA_ALIGN(rtalen)) goto nlmsg_failure; \
- rta = (void*)__skb_put(skb, RTA_ALIGN(rtalen)); \
- rta->rta_type = attrtype; \
- rta->rta_len = rtalen; \
- RTA_DATA(rta); })
+ RTA_DATA(__RTA_PUT(skb, attrtype, attrlen))
static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk,
int ext, u32 pid, u32 seq, u16 nlmsg_flags)
@@ -61,7 +54,6 @@ static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk,
struct nlmsghdr *nlh;
struct tcp_info *info = NULL;
struct tcpdiag_meminfo *minfo = NULL;
- struct tcpvegas_info *vinfo = NULL;
unsigned char *b = skb->tail;
nlh = NLMSG_PUT(skb, pid, seq, TCPDIAG_GETSOCK, sizeof(*r));
@@ -73,9 +65,11 @@ static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk,
if (ext & (1<<(TCPDIAG_INFO-1)))
info = TCPDIAG_PUT(skb, TCPDIAG_INFO, sizeof(*info));
- if ((tcp_is_westwood(tp) || tcp_is_vegas(tp))
- && (ext & (1<<(TCPDIAG_VEGASINFO-1))))
- vinfo = TCPDIAG_PUT(skb, TCPDIAG_VEGASINFO, sizeof(*vinfo));
+ if (ext & (1<<(TCPDIAG_CONG-1))) {
+ size_t len = strlen(tp->ca_ops->name);
+ strcpy(TCPDIAG_PUT(skb, TCPDIAG_CONG, len+1),
+ tp->ca_ops->name);
+ }
}
r->tcpdiag_family = sk->sk_family;
r->tcpdiag_state = sk->sk_state;
@@ -166,23 +160,13 @@ static int tcpdiag_fill(struct sk_buff *skb, struct sock *sk,
if (info)
tcp_get_info(sk, info);
- if (vinfo) {
- if (tcp_is_vegas(tp)) {
- vinfo->tcpv_enabled = tp->vegas.doing_vegas_now;
- vinfo->tcpv_rttcnt = tp->vegas.cntRTT;
- vinfo->tcpv_rtt = jiffies_to_usecs(tp->vegas.baseRTT);
- vinfo->tcpv_minrtt = jiffies_to_usecs(tp->vegas.minRTT);
- } else {
- vinfo->tcpv_enabled = 0;
- vinfo->tcpv_rttcnt = 0;
- vinfo->tcpv_rtt = jiffies_to_usecs(tp->westwood.rtt);
- vinfo->tcpv_minrtt = jiffies_to_usecs(tp->westwood.rtt_min);
- }
- }
+ if (sk->sk_state < TCP_TIME_WAIT && tp->ca_ops->get_info)
+ tp->ca_ops->get_info(tp, ext, skb);
nlh->nlmsg_len = skb->tail - b;
return skb->len;
+rtattr_failure:
nlmsg_failure:
skb_trim(skb, b - skb->data);
return -1;
@@ -455,9 +439,10 @@ static int tcpdiag_dump_sock(struct sk_buff *skb, struct sock *sk,
}
static int tcpdiag_fill_req(struct sk_buff *skb, struct sock *sk,
- struct open_request *req,
+ struct request_sock *req,
u32 pid, u32 seq)
{
+ const struct inet_request_sock *ireq = inet_rsk(req);
struct inet_sock *inet = inet_sk(sk);
unsigned char *b = skb->tail;
struct tcpdiagmsg *r;
@@ -482,9 +467,9 @@ static int tcpdiag_fill_req(struct sk_buff *skb, struct sock *sk,
tmo = 0;
r->id.tcpdiag_sport = inet->sport;
- r->id.tcpdiag_dport = req->rmt_port;
- r->id.tcpdiag_src[0] = req->af.v4_req.loc_addr;
- r->id.tcpdiag_dst[0] = req->af.v4_req.rmt_addr;
+ r->id.tcpdiag_dport = ireq->rmt_port;
+ r->id.tcpdiag_src[0] = ireq->loc_addr;
+ r->id.tcpdiag_dst[0] = ireq->rmt_addr;
r->tcpdiag_expires = jiffies_to_msecs(tmo),
r->tcpdiag_rqueue = 0;
r->tcpdiag_wqueue = 0;
@@ -493,9 +478,9 @@ static int tcpdiag_fill_req(struct sk_buff *skb, struct sock *sk,
#ifdef CONFIG_IP_TCPDIAG_IPV6
if (r->tcpdiag_family == AF_INET6) {
ipv6_addr_copy((struct in6_addr *)r->id.tcpdiag_src,
- &req->af.v6_req.loc_addr);
+ &tcp6_rsk(req)->loc_addr);
ipv6_addr_copy((struct in6_addr *)r->id.tcpdiag_dst,
- &req->af.v6_req.rmt_addr);
+ &tcp6_rsk(req)->rmt_addr);
}
#endif
nlh->nlmsg_len = skb->tail - b;
@@ -513,7 +498,7 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk,
struct tcpdiag_entry entry;
struct tcpdiagreq *r = NLMSG_DATA(cb->nlh);
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt;
+ struct listen_sock *lopt;
struct rtattr *bc = NULL;
struct inet_sock *inet = inet_sk(sk);
int j, s_j;
@@ -528,9 +513,9 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk,
entry.family = sk->sk_family;
- read_lock_bh(&tp->syn_wait_lock);
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
- lopt = tp->listen_opt;
+ lopt = tp->accept_queue.listen_opt;
if (!lopt || !lopt->qlen)
goto out;
@@ -541,13 +526,15 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk,
}
for (j = s_j; j < TCP_SYNQ_HSIZE; j++) {
- struct open_request *req, *head = lopt->syn_table[j];
+ struct request_sock *req, *head = lopt->syn_table[j];
reqnum = 0;
for (req = head; req; reqnum++, req = req->dl_next) {
+ struct inet_request_sock *ireq = inet_rsk(req);
+
if (reqnum < s_reqnum)
continue;
- if (r->id.tcpdiag_dport != req->rmt_port &&
+ if (r->id.tcpdiag_dport != ireq->rmt_port &&
r->id.tcpdiag_dport)
continue;
@@ -555,16 +542,16 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk,
entry.saddr =
#ifdef CONFIG_IP_TCPDIAG_IPV6
(entry.family == AF_INET6) ?
- req->af.v6_req.loc_addr.s6_addr32 :
+ tcp6_rsk(req)->loc_addr.s6_addr32 :
#endif
- &req->af.v4_req.loc_addr;
+ &ireq->loc_addr;
entry.daddr =
#ifdef CONFIG_IP_TCPDIAG_IPV6
(entry.family == AF_INET6) ?
- req->af.v6_req.rmt_addr.s6_addr32 :
+ tcp6_rsk(req)->rmt_addr.s6_addr32 :
#endif
- &req->af.v4_req.rmt_addr;
- entry.dport = ntohs(req->rmt_port);
+ &ireq->rmt_addr;
+ entry.dport = ntohs(ireq->rmt_port);
if (!tcpdiag_bc_run(RTA_DATA(bc),
RTA_PAYLOAD(bc), &entry))
@@ -585,7 +572,7 @@ static int tcpdiag_dump_reqs(struct sk_buff *skb, struct sock *sk,
}
out:
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
return err;
}
diff --git a/net/ipv4/tcp_highspeed.c b/net/ipv4/tcp_highspeed.c
new file mode 100644
index 00000000000..36c51f8136b
--- /dev/null
+++ b/net/ipv4/tcp_highspeed.c
@@ -0,0 +1,181 @@
+/*
+ * Sally Floyd's High Speed TCP (RFC 3649) congestion control
+ *
+ * See http://www.icir.org/floyd/hstcp.html
+ *
+ * John Heffner <jheffner@psc.edu>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <net/tcp.h>
+
+
+/* From AIMD tables from RFC 3649 appendix B,
+ * with fixed-point MD scaled <<8.
+ */
+static const struct hstcp_aimd_val {
+ unsigned int cwnd;
+ unsigned int md;
+} hstcp_aimd_vals[] = {
+ { 38, 128, /* 0.50 */ },
+ { 118, 112, /* 0.44 */ },
+ { 221, 104, /* 0.41 */ },
+ { 347, 98, /* 0.38 */ },
+ { 495, 93, /* 0.37 */ },
+ { 663, 89, /* 0.35 */ },
+ { 851, 86, /* 0.34 */ },
+ { 1058, 83, /* 0.33 */ },
+ { 1284, 81, /* 0.32 */ },
+ { 1529, 78, /* 0.31 */ },
+ { 1793, 76, /* 0.30 */ },
+ { 2076, 74, /* 0.29 */ },
+ { 2378, 72, /* 0.28 */ },
+ { 2699, 71, /* 0.28 */ },
+ { 3039, 69, /* 0.27 */ },
+ { 3399, 68, /* 0.27 */ },
+ { 3778, 66, /* 0.26 */ },
+ { 4177, 65, /* 0.26 */ },
+ { 4596, 64, /* 0.25 */ },
+ { 5036, 62, /* 0.25 */ },
+ { 5497, 61, /* 0.24 */ },
+ { 5979, 60, /* 0.24 */ },
+ { 6483, 59, /* 0.23 */ },
+ { 7009, 58, /* 0.23 */ },
+ { 7558, 57, /* 0.22 */ },
+ { 8130, 56, /* 0.22 */ },
+ { 8726, 55, /* 0.22 */ },
+ { 9346, 54, /* 0.21 */ },
+ { 9991, 53, /* 0.21 */ },
+ { 10661, 52, /* 0.21 */ },
+ { 11358, 52, /* 0.20 */ },
+ { 12082, 51, /* 0.20 */ },
+ { 12834, 50, /* 0.20 */ },
+ { 13614, 49, /* 0.19 */ },
+ { 14424, 48, /* 0.19 */ },
+ { 15265, 48, /* 0.19 */ },
+ { 16137, 47, /* 0.19 */ },
+ { 17042, 46, /* 0.18 */ },
+ { 17981, 45, /* 0.18 */ },
+ { 18955, 45, /* 0.18 */ },
+ { 19965, 44, /* 0.17 */ },
+ { 21013, 43, /* 0.17 */ },
+ { 22101, 43, /* 0.17 */ },
+ { 23230, 42, /* 0.17 */ },
+ { 24402, 41, /* 0.16 */ },
+ { 25618, 41, /* 0.16 */ },
+ { 26881, 40, /* 0.16 */ },
+ { 28193, 39, /* 0.16 */ },
+ { 29557, 39, /* 0.15 */ },
+ { 30975, 38, /* 0.15 */ },
+ { 32450, 38, /* 0.15 */ },
+ { 33986, 37, /* 0.15 */ },
+ { 35586, 36, /* 0.14 */ },
+ { 37253, 36, /* 0.14 */ },
+ { 38992, 35, /* 0.14 */ },
+ { 40808, 35, /* 0.14 */ },
+ { 42707, 34, /* 0.13 */ },
+ { 44694, 33, /* 0.13 */ },
+ { 46776, 33, /* 0.13 */ },
+ { 48961, 32, /* 0.13 */ },
+ { 51258, 32, /* 0.13 */ },
+ { 53677, 31, /* 0.12 */ },
+ { 56230, 30, /* 0.12 */ },
+ { 58932, 30, /* 0.12 */ },
+ { 61799, 29, /* 0.12 */ },
+ { 64851, 28, /* 0.11 */ },
+ { 68113, 28, /* 0.11 */ },
+ { 71617, 27, /* 0.11 */ },
+ { 75401, 26, /* 0.10 */ },
+ { 79517, 26, /* 0.10 */ },
+ { 84035, 25, /* 0.10 */ },
+ { 89053, 24, /* 0.10 */ },
+};
+
+#define HSTCP_AIMD_MAX ARRAY_SIZE(hstcp_aimd_vals)
+
+struct hstcp {
+ u32 ai;
+};
+
+static void hstcp_init(struct tcp_sock *tp)
+{
+ struct hstcp *ca = tcp_ca(tp);
+
+ ca->ai = 0;
+
+ /* Ensure the MD arithmetic works. This is somewhat pedantic,
+ * since I don't think we will see a cwnd this large. :) */
+ tp->snd_cwnd_clamp = min_t(u32, tp->snd_cwnd_clamp, 0xffffffff/128);
+}
+
+static void hstcp_cong_avoid(struct tcp_sock *tp, u32 adk, u32 rtt,
+ u32 in_flight, int good)
+{
+ struct hstcp *ca = tcp_ca(tp);
+
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (tp->snd_cwnd <= tp->snd_ssthresh) {
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ } else {
+ /* Update AIMD parameters */
+ if (tp->snd_cwnd > hstcp_aimd_vals[ca->ai].cwnd) {
+ while (tp->snd_cwnd > hstcp_aimd_vals[ca->ai].cwnd &&
+ ca->ai < HSTCP_AIMD_MAX)
+ ca->ai++;
+ } else if (tp->snd_cwnd < hstcp_aimd_vals[ca->ai].cwnd) {
+ while (tp->snd_cwnd > hstcp_aimd_vals[ca->ai].cwnd &&
+ ca->ai > 0)
+ ca->ai--;
+ }
+
+ /* Do additive increase */
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp) {
+ tp->snd_cwnd_cnt += ca->ai;
+ if (tp->snd_cwnd_cnt >= tp->snd_cwnd) {
+ tp->snd_cwnd++;
+ tp->snd_cwnd_cnt -= tp->snd_cwnd;
+ }
+ }
+ }
+}
+
+static u32 hstcp_ssthresh(struct tcp_sock *tp)
+{
+ struct hstcp *ca = tcp_ca(tp);
+
+ /* Do multiplicative decrease */
+ return max(tp->snd_cwnd - ((tp->snd_cwnd * hstcp_aimd_vals[ca->ai].md) >> 8), 2U);
+}
+
+
+static struct tcp_congestion_ops tcp_highspeed = {
+ .init = hstcp_init,
+ .ssthresh = hstcp_ssthresh,
+ .cong_avoid = hstcp_cong_avoid,
+ .min_cwnd = tcp_reno_min_cwnd,
+
+ .owner = THIS_MODULE,
+ .name = "highspeed"
+};
+
+static int __init hstcp_register(void)
+{
+ BUG_ON(sizeof(struct hstcp) > TCP_CA_PRIV_SIZE);
+ return tcp_register_congestion_control(&tcp_highspeed);
+}
+
+static void __exit hstcp_unregister(void)
+{
+ tcp_unregister_congestion_control(&tcp_highspeed);
+}
+
+module_init(hstcp_register);
+module_exit(hstcp_unregister);
+
+MODULE_AUTHOR("John Heffner");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("High Speed TCP");
diff --git a/net/ipv4/tcp_htcp.c b/net/ipv4/tcp_htcp.c
new file mode 100644
index 00000000000..40168275acf
--- /dev/null
+++ b/net/ipv4/tcp_htcp.c
@@ -0,0 +1,289 @@
+/*
+ * H-TCP congestion control. The algorithm is detailed in:
+ * R.N.Shorten, D.J.Leith:
+ * "H-TCP: TCP for high-speed and long-distance networks"
+ * Proc. PFLDnet, Argonne, 2004.
+ * http://www.hamilton.ie/net/htcp3.pdf
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <net/tcp.h>
+
+#define ALPHA_BASE (1<<7) /* 1.0 with shift << 7 */
+#define BETA_MIN (1<<6) /* 0.5 with shift << 7 */
+#define BETA_MAX 102 /* 0.8 with shift << 7 */
+
+static int use_rtt_scaling = 1;
+module_param(use_rtt_scaling, int, 0644);
+MODULE_PARM_DESC(use_rtt_scaling, "turn on/off RTT scaling");
+
+static int use_bandwidth_switch = 1;
+module_param(use_bandwidth_switch, int, 0644);
+MODULE_PARM_DESC(use_bandwidth_switch, "turn on/off bandwidth switcher");
+
+struct htcp {
+ u16 alpha; /* Fixed point arith, << 7 */
+ u8 beta; /* Fixed point arith, << 7 */
+ u8 modeswitch; /* Delay modeswitch until we had at least one congestion event */
+ u8 ccount; /* Number of RTTs since last congestion event */
+ u8 undo_ccount;
+ u16 packetcount;
+ u32 minRTT;
+ u32 maxRTT;
+ u32 snd_cwnd_cnt2;
+
+ u32 undo_maxRTT;
+ u32 undo_old_maxB;
+
+ /* Bandwidth estimation */
+ u32 minB;
+ u32 maxB;
+ u32 old_maxB;
+ u32 Bi;
+ u32 lasttime;
+};
+
+static inline void htcp_reset(struct htcp *ca)
+{
+ ca->undo_ccount = ca->ccount;
+ ca->undo_maxRTT = ca->maxRTT;
+ ca->undo_old_maxB = ca->old_maxB;
+
+ ca->ccount = 0;
+ ca->snd_cwnd_cnt2 = 0;
+}
+
+static u32 htcp_cwnd_undo(struct tcp_sock *tp)
+{
+ struct htcp *ca = tcp_ca(tp);
+ ca->ccount = ca->undo_ccount;
+ ca->maxRTT = ca->undo_maxRTT;
+ ca->old_maxB = ca->undo_old_maxB;
+ return max(tp->snd_cwnd, (tp->snd_ssthresh<<7)/ca->beta);
+}
+
+static inline void measure_rtt(struct tcp_sock *tp)
+{
+ struct htcp *ca = tcp_ca(tp);
+ u32 srtt = tp->srtt>>3;
+
+ /* keep track of minimum RTT seen so far, minRTT is zero at first */
+ if (ca->minRTT > srtt || !ca->minRTT)
+ ca->minRTT = srtt;
+
+ /* max RTT */
+ if (tp->ca_state == TCP_CA_Open && tp->snd_ssthresh < 0xFFFF && ca->ccount > 3) {
+ if (ca->maxRTT < ca->minRTT)
+ ca->maxRTT = ca->minRTT;
+ if (ca->maxRTT < srtt && srtt <= ca->maxRTT+HZ/50)
+ ca->maxRTT = srtt;
+ }
+}
+
+static void measure_achieved_throughput(struct tcp_sock *tp, u32 pkts_acked)
+{
+ struct htcp *ca = tcp_ca(tp);
+ u32 now = tcp_time_stamp;
+
+ /* achieved throughput calculations */
+ if (tp->ca_state != TCP_CA_Open && tp->ca_state != TCP_CA_Disorder) {
+ ca->packetcount = 0;
+ ca->lasttime = now;
+ return;
+ }
+
+ ca->packetcount += pkts_acked;
+
+ if (ca->packetcount >= tp->snd_cwnd - (ca->alpha>>7? : 1)
+ && now - ca->lasttime >= ca->minRTT
+ && ca->minRTT > 0) {
+ __u32 cur_Bi = ca->packetcount*HZ/(now - ca->lasttime);
+ if (ca->ccount <= 3) {
+ /* just after backoff */
+ ca->minB = ca->maxB = ca->Bi = cur_Bi;
+ } else {
+ ca->Bi = (3*ca->Bi + cur_Bi)/4;
+ if (ca->Bi > ca->maxB)
+ ca->maxB = ca->Bi;
+ if (ca->minB > ca->maxB)
+ ca->minB = ca->maxB;
+ }
+ ca->packetcount = 0;
+ ca->lasttime = now;
+ }
+}
+
+static inline void htcp_beta_update(struct htcp *ca, u32 minRTT, u32 maxRTT)
+{
+ if (use_bandwidth_switch) {
+ u32 maxB = ca->maxB;
+ u32 old_maxB = ca->old_maxB;
+ ca->old_maxB = ca->maxB;
+
+ if (!between(5*maxB, 4*old_maxB, 6*old_maxB)) {
+ ca->beta = BETA_MIN;
+ ca->modeswitch = 0;
+ return;
+ }
+ }
+
+ if (ca->modeswitch && minRTT > max(HZ/100, 1) && maxRTT) {
+ ca->beta = (minRTT<<7)/maxRTT;
+ if (ca->beta < BETA_MIN)
+ ca->beta = BETA_MIN;
+ else if (ca->beta > BETA_MAX)
+ ca->beta = BETA_MAX;
+ } else {
+ ca->beta = BETA_MIN;
+ ca->modeswitch = 1;
+ }
+}
+
+static inline void htcp_alpha_update(struct htcp *ca)
+{
+ u32 minRTT = ca->minRTT;
+ u32 factor = 1;
+ u32 diff = ca->ccount * minRTT; /* time since last backoff */
+
+ if (diff > HZ) {
+ diff -= HZ;
+ factor = 1+ ( 10*diff + ((diff/2)*(diff/2)/HZ) )/HZ;
+ }
+
+ if (use_rtt_scaling && minRTT) {
+ u32 scale = (HZ<<3)/(10*minRTT);
+ scale = min(max(scale, 1U<<2), 10U<<3); /* clamping ratio to interval [0.5,10]<<3 */
+ factor = (factor<<3)/scale;
+ if (!factor)
+ factor = 1;
+ }
+
+ ca->alpha = 2*factor*((1<<7)-ca->beta);
+ if (!ca->alpha)
+ ca->alpha = ALPHA_BASE;
+}
+
+/* After we have the rtt data to calculate beta, we'd still prefer to wait one
+ * rtt before we adjust our beta to ensure we are working from a consistent
+ * data.
+ *
+ * This function should be called when we hit a congestion event since only at
+ * that point do we really have a real sense of maxRTT (the queues en route
+ * were getting just too full now).
+ */
+static void htcp_param_update(struct tcp_sock *tp)
+{
+ struct htcp *ca = tcp_ca(tp);
+ u32 minRTT = ca->minRTT;
+ u32 maxRTT = ca->maxRTT;
+
+ htcp_beta_update(ca, minRTT, maxRTT);
+ htcp_alpha_update(ca);
+
+ /* add slowly fading memory for maxRTT to accommodate routing changes etc */
+ if (minRTT > 0 && maxRTT > minRTT)
+ ca->maxRTT = minRTT + ((maxRTT-minRTT)*95)/100;
+}
+
+static u32 htcp_recalc_ssthresh(struct tcp_sock *tp)
+{
+ struct htcp *ca = tcp_ca(tp);
+ htcp_param_update(tp);
+ return max((tp->snd_cwnd * ca->beta) >> 7, 2U);
+}
+
+static void htcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt,
+ u32 in_flight, int data_acked)
+{
+ struct htcp *ca = tcp_ca(tp);
+
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (tp->snd_cwnd <= tp->snd_ssthresh) {
+ /* In "safe" area, increase. */
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ } else {
+ measure_rtt(tp);
+
+ /* keep track of number of round-trip times since last backoff event */
+ if (ca->snd_cwnd_cnt2++ > tp->snd_cwnd) {
+ ca->ccount++;
+ ca->snd_cwnd_cnt2 = 0;
+ htcp_alpha_update(ca);
+ }
+
+ /* In dangerous area, increase slowly.
+ * In theory this is tp->snd_cwnd += alpha / tp->snd_cwnd
+ */
+ if ((tp->snd_cwnd_cnt++ * ca->alpha)>>7 >= tp->snd_cwnd) {
+ if (tp->snd_cwnd < tp->snd_cwnd_clamp)
+ tp->snd_cwnd++;
+ tp->snd_cwnd_cnt = 0;
+ ca->ccount++;
+ }
+ }
+}
+
+/* Lower bound on congestion window. */
+static u32 htcp_min_cwnd(struct tcp_sock *tp)
+{
+ return tp->snd_ssthresh;
+}
+
+
+static void htcp_init(struct tcp_sock *tp)
+{
+ struct htcp *ca = tcp_ca(tp);
+
+ memset(ca, 0, sizeof(struct htcp));
+ ca->alpha = ALPHA_BASE;
+ ca->beta = BETA_MIN;
+}
+
+static void htcp_state(struct tcp_sock *tp, u8 new_state)
+{
+ switch (new_state) {
+ case TCP_CA_CWR:
+ case TCP_CA_Recovery:
+ case TCP_CA_Loss:
+ htcp_reset(tcp_ca(tp));
+ break;
+ }
+}
+
+static struct tcp_congestion_ops htcp = {
+ .init = htcp_init,
+ .ssthresh = htcp_recalc_ssthresh,
+ .min_cwnd = htcp_min_cwnd,
+ .cong_avoid = htcp_cong_avoid,
+ .set_state = htcp_state,
+ .undo_cwnd = htcp_cwnd_undo,
+ .pkts_acked = measure_achieved_throughput,
+ .owner = THIS_MODULE,
+ .name = "htcp",
+};
+
+static int __init htcp_register(void)
+{
+ BUG_ON(sizeof(struct htcp) > TCP_CA_PRIV_SIZE);
+ BUILD_BUG_ON(BETA_MIN >= BETA_MAX);
+ if (!use_bandwidth_switch)
+ htcp.pkts_acked = NULL;
+ return tcp_register_congestion_control(&htcp);
+}
+
+static void __exit htcp_unregister(void)
+{
+ tcp_unregister_congestion_control(&htcp);
+}
+
+module_init(htcp_register);
+module_exit(htcp_unregister);
+
+MODULE_AUTHOR("Baruch Even");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("H-TCP");
diff --git a/net/ipv4/tcp_hybla.c b/net/ipv4/tcp_hybla.c
new file mode 100644
index 00000000000..13a66342c30
--- /dev/null
+++ b/net/ipv4/tcp_hybla.c
@@ -0,0 +1,187 @@
+/*
+ * TCP HYBLA
+ *
+ * TCP-HYBLA Congestion control algorithm, based on:
+ * C.Caini, R.Firrincieli, "TCP-Hybla: A TCP Enhancement
+ * for Heterogeneous Networks",
+ * International Journal on satellite Communications,
+ * September 2004
+ * Daniele Lacamera
+ * root at danielinux.net
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <net/tcp.h>
+
+/* Tcp Hybla structure. */
+struct hybla {
+ u8 hybla_en;
+ u32 snd_cwnd_cents; /* Keeps increment values when it is <1, <<7 */
+ u32 rho; /* Rho parameter, integer part */
+ u32 rho2; /* Rho * Rho, integer part */
+ u32 rho_3ls; /* Rho parameter, <<3 */
+ u32 rho2_7ls; /* Rho^2, <<7 */
+ u32 minrtt; /* Minimum smoothed round trip time value seen */
+};
+
+/* Hybla reference round trip time (default= 1/40 sec = 25 ms),
+ expressed in jiffies */
+static int rtt0 = 25;
+module_param(rtt0, int, 0644);
+MODULE_PARM_DESC(rtt0, "reference rout trip time (ms)");
+
+
+/* This is called to refresh values for hybla parameters */
+static inline void hybla_recalc_param (struct tcp_sock *tp)
+{
+ struct hybla *ca = tcp_ca(tp);
+
+ ca->rho_3ls = max_t(u32, tp->srtt / msecs_to_jiffies(rtt0), 8);
+ ca->rho = ca->rho_3ls >> 3;
+ ca->rho2_7ls = (ca->rho_3ls * ca->rho_3ls) << 1;
+ ca->rho2 = ca->rho2_7ls >>7;
+}
+
+static void hybla_init(struct tcp_sock *tp)
+{
+ struct hybla *ca = tcp_ca(tp);
+
+ ca->rho = 0;
+ ca->rho2 = 0;
+ ca->rho_3ls = 0;
+ ca->rho2_7ls = 0;
+ ca->snd_cwnd_cents = 0;
+ ca->hybla_en = 1;
+ tp->snd_cwnd = 2;
+ tp->snd_cwnd_clamp = 65535;
+
+ /* 1st Rho measurement based on initial srtt */
+ hybla_recalc_param(tp);
+
+ /* set minimum rtt as this is the 1st ever seen */
+ ca->minrtt = tp->srtt;
+ tp->snd_cwnd = ca->rho;
+}
+
+static void hybla_state(struct tcp_sock *tp, u8 ca_state)
+{
+ struct hybla *ca = tcp_ca(tp);
+
+ ca->hybla_en = (ca_state == TCP_CA_Open);
+}
+
+static inline u32 hybla_fraction(u32 odds)
+{
+ static const u32 fractions[] = {
+ 128, 139, 152, 165, 181, 197, 215, 234,
+ };
+
+ return (odds < ARRAY_SIZE(fractions)) ? fractions[odds] : 128;
+}
+
+/* TCP Hybla main routine.
+ * This is the algorithm behavior:
+ * o Recalc Hybla parameters if min_rtt has changed
+ * o Give cwnd a new value based on the model proposed
+ * o remember increments <1
+ */
+static void hybla_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt,
+ u32 in_flight, int flag)
+{
+ struct hybla *ca = tcp_ca(tp);
+ u32 increment, odd, rho_fractions;
+ int is_slowstart = 0;
+
+ /* Recalculate rho only if this srtt is the lowest */
+ if (tp->srtt < ca->minrtt){
+ hybla_recalc_param(tp);
+ ca->minrtt = tp->srtt;
+ }
+
+ if (!ca->hybla_en)
+ return tcp_reno_cong_avoid(tp, ack, rtt, in_flight, flag);
+
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (ca->rho == 0)
+ hybla_recalc_param(tp);
+
+ rho_fractions = ca->rho_3ls - (ca->rho << 3);
+
+ if (tp->snd_cwnd < tp->snd_ssthresh) {
+ /*
+ * slow start
+ * INC = 2^RHO - 1
+ * This is done by splitting the rho parameter
+ * into 2 parts: an integer part and a fraction part.
+ * Inrement<<7 is estimated by doing:
+ * [2^(int+fract)]<<7
+ * that is equal to:
+ * (2^int) * [(2^fract) <<7]
+ * 2^int is straightly computed as 1<<int,
+ * while we will use hybla_slowstart_fraction_increment() to
+ * calculate 2^fract in a <<7 value.
+ */
+ is_slowstart = 1;
+ increment = ((1 << ca->rho) * hybla_fraction(rho_fractions))
+ - 128;
+ } else {
+ /*
+ * congestion avoidance
+ * INC = RHO^2 / W
+ * as long as increment is estimated as (rho<<7)/window
+ * it already is <<7 and we can easily count its fractions.
+ */
+ increment = ca->rho2_7ls / tp->snd_cwnd;
+ if (increment < 128)
+ tp->snd_cwnd_cnt++;
+ }
+
+ odd = increment % 128;
+ tp->snd_cwnd += increment >> 7;
+ ca->snd_cwnd_cents += odd;
+
+ /* check when fractions goes >=128 and increase cwnd by 1. */
+ while(ca->snd_cwnd_cents >= 128) {
+ tp->snd_cwnd++;
+ ca->snd_cwnd_cents -= 128;
+ tp->snd_cwnd_cnt = 0;
+ }
+
+ /* clamp down slowstart cwnd to ssthresh value. */
+ if (is_slowstart)
+ tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
+
+ tp->snd_cwnd = min_t(u32, tp->snd_cwnd, tp->snd_cwnd_clamp);
+}
+
+static struct tcp_congestion_ops tcp_hybla = {
+ .init = hybla_init,
+ .ssthresh = tcp_reno_ssthresh,
+ .min_cwnd = tcp_reno_min_cwnd,
+ .cong_avoid = hybla_cong_avoid,
+ .set_state = hybla_state,
+
+ .owner = THIS_MODULE,
+ .name = "hybla"
+};
+
+static int __init hybla_register(void)
+{
+ BUG_ON(sizeof(struct hybla) > TCP_CA_PRIV_SIZE);
+ return tcp_register_congestion_control(&tcp_hybla);
+}
+
+static void __exit hybla_unregister(void)
+{
+ tcp_unregister_congestion_control(&tcp_hybla);
+}
+
+module_init(hybla_register);
+module_exit(hybla_unregister);
+
+MODULE_AUTHOR("Daniele Lacamera");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("TCP Hybla");
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 5bad504630a..53a8a5399f1 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -61,7 +61,6 @@
* Panu Kuhlberg: Experimental audit of TCP (re)transmission
* engine. Lots of bugs are found.
* Pasi Sarolahti: F-RTO for dealing with spurious RTOs
- * Angelo Dell'Aera: TCP Westwood+ support
*/
#include <linux/config.h>
@@ -88,23 +87,9 @@ int sysctl_tcp_rfc1337;
int sysctl_tcp_max_orphans = NR_FILE;
int sysctl_tcp_frto;
int sysctl_tcp_nometrics_save;
-int sysctl_tcp_westwood;
-int sysctl_tcp_vegas_cong_avoid;
int sysctl_tcp_moderate_rcvbuf = 1;
-/* Default values of the Vegas variables, in fixed-point representation
- * with V_PARAM_SHIFT bits to the right of the binary point.
- */
-#define V_PARAM_SHIFT 1
-int sysctl_tcp_vegas_alpha = 1<<V_PARAM_SHIFT;
-int sysctl_tcp_vegas_beta = 3<<V_PARAM_SHIFT;
-int sysctl_tcp_vegas_gamma = 1<<V_PARAM_SHIFT;
-int sysctl_tcp_bic = 1;
-int sysctl_tcp_bic_fast_convergence = 1;
-int sysctl_tcp_bic_low_window = 14;
-int sysctl_tcp_bic_beta = 819; /* = 819/1024 (BICTCP_BETA_SCALE) */
-
#define FLAG_DATA 0x01 /* Incoming frame contained data. */
#define FLAG_WIN_UPDATE 0x02 /* Incoming ACK was a window update. */
#define FLAG_DATA_ACKED 0x04 /* This ACK acknowledged new data. */
@@ -333,15 +318,6 @@ static void tcp_init_buffer_space(struct sock *sk)
tp->snd_cwnd_stamp = tcp_time_stamp;
}
-static void init_bictcp(struct tcp_sock *tp)
-{
- tp->bictcp.cnt = 0;
-
- tp->bictcp.last_max_cwnd = 0;
- tp->bictcp.last_cwnd = 0;
- tp->bictcp.last_stamp = 0;
-}
-
/* 5. Recalculate window clamp after socket hit its memory bounds. */
static void tcp_clamp_window(struct sock *sk, struct tcp_sock *tp)
{
@@ -558,45 +534,6 @@ static void tcp_event_data_recv(struct sock *sk, struct tcp_sock *tp, struct sk_
tcp_grow_window(sk, tp, skb);
}
-/* When starting a new connection, pin down the current choice of
- * congestion algorithm.
- */
-void tcp_ca_init(struct tcp_sock *tp)
-{
- if (sysctl_tcp_westwood)
- tp->adv_cong = TCP_WESTWOOD;
- else if (sysctl_tcp_bic)
- tp->adv_cong = TCP_BIC;
- else if (sysctl_tcp_vegas_cong_avoid) {
- tp->adv_cong = TCP_VEGAS;
- tp->vegas.baseRTT = 0x7fffffff;
- tcp_vegas_enable(tp);
- }
-}
-
-/* Do RTT sampling needed for Vegas.
- * Basically we:
- * o min-filter RTT samples from within an RTT to get the current
- * propagation delay + queuing delay (we are min-filtering to try to
- * avoid the effects of delayed ACKs)
- * o min-filter RTT samples from a much longer window (forever for now)
- * to find the propagation delay (baseRTT)
- */
-static inline void vegas_rtt_calc(struct tcp_sock *tp, __u32 rtt)
-{
- __u32 vrtt = rtt + 1; /* Never allow zero rtt or baseRTT */
-
- /* Filter to find propagation delay: */
- if (vrtt < tp->vegas.baseRTT)
- tp->vegas.baseRTT = vrtt;
-
- /* Find the min RTT during the last RTT to find
- * the current prop. delay + queuing delay:
- */
- tp->vegas.minRTT = min(tp->vegas.minRTT, vrtt);
- tp->vegas.cntRTT++;
-}
-
/* Called to compute a smoothed rtt estimate. The data fed to this
* routine either comes from timestamps, or from segments that were
* known _not_ to have been retransmitted [see Karn/Partridge
@@ -606,13 +543,10 @@ static inline void vegas_rtt_calc(struct tcp_sock *tp, __u32 rtt)
* To save cycles in the RFC 1323 implementation it was better to break
* it up into three procedures. -- erics
*/
-static void tcp_rtt_estimator(struct tcp_sock *tp, __u32 mrtt)
+static void tcp_rtt_estimator(struct tcp_sock *tp, __u32 mrtt, u32 *usrtt)
{
long m = mrtt; /* RTT */
- if (tcp_vegas_enabled(tp))
- vegas_rtt_calc(tp, mrtt);
-
/* The following amusing code comes from Jacobson's
* article in SIGCOMM '88. Note that rtt and mdev
* are scaled versions of rtt and mean deviation.
@@ -670,7 +604,8 @@ static void tcp_rtt_estimator(struct tcp_sock *tp, __u32 mrtt)
tp->rtt_seq = tp->snd_nxt;
}
- tcp_westwood_update_rtt(tp, tp->srtt >> 3);
+ if (tp->ca_ops->rtt_sample)
+ tp->ca_ops->rtt_sample(tp, *usrtt);
}
/* Calculate rto without backoff. This is the second half of Van Jacobson's
@@ -805,10 +740,10 @@ __u32 tcp_init_cwnd(struct tcp_sock *tp, struct dst_entry *dst)
__u32 cwnd = (dst ? dst_metric(dst, RTAX_INITCWND) : 0);
if (!cwnd) {
- if (tp->mss_cache_std > 1460)
+ if (tp->mss_cache > 1460)
cwnd = 2;
else
- cwnd = (tp->mss_cache_std > 1095) ? 3 : 4;
+ cwnd = (tp->mss_cache > 1095) ? 3 : 4;
}
return min_t(__u32, cwnd, tp->snd_cwnd_clamp);
}
@@ -979,7 +914,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_
if (sk->sk_route_caps & NETIF_F_TSO) {
sk->sk_route_caps &= ~NETIF_F_TSO;
sock_set_flag(sk, SOCK_NO_LARGESEND);
- tp->mss_cache = tp->mss_cache_std;
+ tp->mss_cache = tp->mss_cache;
}
if (!tp->sacked_out)
@@ -1142,7 +1077,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_
(IsFack(tp) ||
!before(lost_retrans,
TCP_SKB_CB(skb)->ack_seq + tp->reordering *
- tp->mss_cache_std))) {
+ tp->mss_cache))) {
TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
tp->retrans_out -= tcp_skb_pcount(skb);
@@ -1185,8 +1120,8 @@ void tcp_enter_frto(struct sock *sk)
tp->snd_una == tp->high_seq ||
(tp->ca_state == TCP_CA_Loss && !tp->retransmits)) {
tp->prior_ssthresh = tcp_current_ssthresh(tp);
- if (!tcp_westwood_ssthresh(tp))
- tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
+ tp->snd_ssthresh = tp->ca_ops->ssthresh(tp);
+ tcp_ca_event(tp, CA_EVENT_FRTO);
}
/* Have to clear retransmission markers here to keep the bookkeeping
@@ -1252,8 +1187,6 @@ static void tcp_enter_frto_loss(struct sock *sk)
tcp_set_ca_state(tp, TCP_CA_Loss);
tp->high_seq = tp->frto_highmark;
TCP_ECN_queue_cwr(tp);
-
- init_bictcp(tp);
}
void tcp_clear_retrans(struct tcp_sock *tp)
@@ -1283,7 +1216,8 @@ void tcp_enter_loss(struct sock *sk, int how)
if (tp->ca_state <= TCP_CA_Disorder || tp->snd_una == tp->high_seq ||
(tp->ca_state == TCP_CA_Loss && !tp->retransmits)) {
tp->prior_ssthresh = tcp_current_ssthresh(tp);
- tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
+ tp->snd_ssthresh = tp->ca_ops->ssthresh(tp);
+ tcp_ca_event(tp, CA_EVENT_LOSS);
}
tp->snd_cwnd = 1;
tp->snd_cwnd_cnt = 0;
@@ -1596,28 +1530,14 @@ static inline void tcp_moderate_cwnd(struct tcp_sock *tp)
}
/* Decrease cwnd each second ack. */
-
static void tcp_cwnd_down(struct tcp_sock *tp)
{
int decr = tp->snd_cwnd_cnt + 1;
- __u32 limit;
-
- /*
- * TCP Westwood
- * Here limit is evaluated as BWestimation*RTTmin (for obtaining it
- * in packets we use mss_cache). If sysctl_tcp_westwood is off
- * tcp_westwood_bw_rttmin() returns 0. In such case snd_ssthresh is
- * still used as usual. It prevents other strange cases in which
- * BWE*RTTmin could assume value 0. It should not happen but...
- */
-
- if (!(limit = tcp_westwood_bw_rttmin(tp)))
- limit = tp->snd_ssthresh/2;
tp->snd_cwnd_cnt = decr&1;
decr >>= 1;
- if (decr && tp->snd_cwnd > limit)
+ if (decr && tp->snd_cwnd > tp->ca_ops->min_cwnd(tp))
tp->snd_cwnd -= decr;
tp->snd_cwnd = min(tp->snd_cwnd, tcp_packets_in_flight(tp)+1);
@@ -1654,8 +1574,8 @@ static void DBGUNDO(struct sock *sk, struct tcp_sock *tp, const char *msg)
static void tcp_undo_cwr(struct tcp_sock *tp, int undo)
{
if (tp->prior_ssthresh) {
- if (tcp_is_bic(tp))
- tp->snd_cwnd = max(tp->snd_cwnd, tp->bictcp.last_max_cwnd);
+ if (tp->ca_ops->undo_cwnd)
+ tp->snd_cwnd = tp->ca_ops->undo_cwnd(tp);
else
tp->snd_cwnd = max(tp->snd_cwnd, tp->snd_ssthresh<<1);
@@ -1767,11 +1687,9 @@ static int tcp_try_undo_loss(struct sock *sk, struct tcp_sock *tp)
static inline void tcp_complete_cwr(struct tcp_sock *tp)
{
- if (tcp_westwood_cwnd(tp))
- tp->snd_ssthresh = tp->snd_cwnd;
- else
- tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
+ tp->snd_cwnd = min(tp->snd_cwnd, tp->snd_ssthresh);
tp->snd_cwnd_stamp = tcp_time_stamp;
+ tcp_ca_event(tp, CA_EVENT_COMPLETE_CWR);
}
static void tcp_try_to_open(struct sock *sk, struct tcp_sock *tp, int flag)
@@ -1946,7 +1864,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una,
if (tp->ca_state < TCP_CA_CWR) {
if (!(flag&FLAG_ECE))
tp->prior_ssthresh = tcp_current_ssthresh(tp);
- tp->snd_ssthresh = tcp_recalc_ssthresh(tp);
+ tp->snd_ssthresh = tp->ca_ops->ssthresh(tp);
TCP_ECN_queue_cwr(tp);
}
@@ -1963,7 +1881,7 @@ tcp_fastretrans_alert(struct sock *sk, u32 prior_snd_una,
/* Read draft-ietf-tcplw-high-performance before mucking
* with this code. (Superceeds RFC1323)
*/
-static void tcp_ack_saw_tstamp(struct tcp_sock *tp, int flag)
+static void tcp_ack_saw_tstamp(struct tcp_sock *tp, u32 *usrtt, int flag)
{
__u32 seq_rtt;
@@ -1983,13 +1901,13 @@ static void tcp_ack_saw_tstamp(struct tcp_sock *tp, int flag)
* in window is lost... Voila. --ANK (010210)
*/
seq_rtt = tcp_time_stamp - tp->rx_opt.rcv_tsecr;
- tcp_rtt_estimator(tp, seq_rtt);
+ tcp_rtt_estimator(tp, seq_rtt, usrtt);
tcp_set_rto(tp);
tp->backoff = 0;
tcp_bound_rto(tp);
}
-static void tcp_ack_no_tstamp(struct tcp_sock *tp, u32 seq_rtt, int flag)
+static void tcp_ack_no_tstamp(struct tcp_sock *tp, u32 seq_rtt, u32 *usrtt, int flag)
{
/* We don't have a timestamp. Can only use
* packets that are not retransmitted to determine
@@ -2003,338 +1921,29 @@ static void tcp_ack_no_tstamp(struct tcp_sock *tp, u32 seq_rtt, int flag)
if (flag & FLAG_RETRANS_DATA_ACKED)
return;
- tcp_rtt_estimator(tp, seq_rtt);
+ tcp_rtt_estimator(tp, seq_rtt, usrtt);
tcp_set_rto(tp);
tp->backoff = 0;
tcp_bound_rto(tp);
}
static inline void tcp_ack_update_rtt(struct tcp_sock *tp,
- int flag, s32 seq_rtt)
+ int flag, s32 seq_rtt, u32 *usrtt)
{
/* Note that peer MAY send zero echo. In this case it is ignored. (rfc1323) */
if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr)
- tcp_ack_saw_tstamp(tp, flag);
+ tcp_ack_saw_tstamp(tp, usrtt, flag);
else if (seq_rtt >= 0)
- tcp_ack_no_tstamp(tp, seq_rtt, flag);
-}
-
-/*
- * Compute congestion window to use.
- *
- * This is from the implementation of BICTCP in
- * Lison-Xu, Kahaled Harfoush, and Injog Rhee.
- * "Binary Increase Congestion Control for Fast, Long Distance
- * Networks" in InfoComm 2004
- * Available from:
- * http://www.csc.ncsu.edu/faculty/rhee/export/bitcp.pdf
- *
- * Unless BIC is enabled and congestion window is large
- * this behaves the same as the original Reno.
- */
-static inline __u32 bictcp_cwnd(struct tcp_sock *tp)
-{
- /* orignal Reno behaviour */
- if (!tcp_is_bic(tp))
- return tp->snd_cwnd;
-
- if (tp->bictcp.last_cwnd == tp->snd_cwnd &&
- (s32)(tcp_time_stamp - tp->bictcp.last_stamp) <= (HZ>>5))
- return tp->bictcp.cnt;
-
- tp->bictcp.last_cwnd = tp->snd_cwnd;
- tp->bictcp.last_stamp = tcp_time_stamp;
-
- /* start off normal */
- if (tp->snd_cwnd <= sysctl_tcp_bic_low_window)
- tp->bictcp.cnt = tp->snd_cwnd;
-
- /* binary increase */
- else if (tp->snd_cwnd < tp->bictcp.last_max_cwnd) {
- __u32 dist = (tp->bictcp.last_max_cwnd - tp->snd_cwnd)
- / BICTCP_B;
-
- if (dist > BICTCP_MAX_INCREMENT)
- /* linear increase */
- tp->bictcp.cnt = tp->snd_cwnd / BICTCP_MAX_INCREMENT;
- else if (dist <= 1U)
- /* binary search increase */
- tp->bictcp.cnt = tp->snd_cwnd * BICTCP_FUNC_OF_MIN_INCR
- / BICTCP_B;
- else
- /* binary search increase */
- tp->bictcp.cnt = tp->snd_cwnd / dist;
- } else {
- /* slow start amd linear increase */
- if (tp->snd_cwnd < tp->bictcp.last_max_cwnd + BICTCP_B)
- /* slow start */
- tp->bictcp.cnt = tp->snd_cwnd * BICTCP_FUNC_OF_MIN_INCR
- / BICTCP_B;
- else if (tp->snd_cwnd < tp->bictcp.last_max_cwnd
- + BICTCP_MAX_INCREMENT*(BICTCP_B-1))
- /* slow start */
- tp->bictcp.cnt = tp->snd_cwnd * (BICTCP_B-1)
- / (tp->snd_cwnd-tp->bictcp.last_max_cwnd);
- else
- /* linear increase */
- tp->bictcp.cnt = tp->snd_cwnd / BICTCP_MAX_INCREMENT;
- }
- return tp->bictcp.cnt;
+ tcp_ack_no_tstamp(tp, seq_rtt, usrtt, flag);
}
-/* This is Jacobson's slow start and congestion avoidance.
- * SIGCOMM '88, p. 328.
- */
-static inline void reno_cong_avoid(struct tcp_sock *tp)
+static inline void tcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt,
+ u32 in_flight, int good)
{
- if (tp->snd_cwnd <= tp->snd_ssthresh) {
- /* In "safe" area, increase. */
- if (tp->snd_cwnd < tp->snd_cwnd_clamp)
- tp->snd_cwnd++;
- } else {
- /* In dangerous area, increase slowly.
- * In theory this is tp->snd_cwnd += 1 / tp->snd_cwnd
- */
- if (tp->snd_cwnd_cnt >= bictcp_cwnd(tp)) {
- if (tp->snd_cwnd < tp->snd_cwnd_clamp)
- tp->snd_cwnd++;
- tp->snd_cwnd_cnt=0;
- } else
- tp->snd_cwnd_cnt++;
- }
+ tp->ca_ops->cong_avoid(tp, ack, rtt, in_flight, good);
tp->snd_cwnd_stamp = tcp_time_stamp;
}
-/* This is based on the congestion detection/avoidance scheme described in
- * Lawrence S. Brakmo and Larry L. Peterson.
- * "TCP Vegas: End to end congestion avoidance on a global internet."
- * IEEE Journal on Selected Areas in Communication, 13(8):1465--1480,
- * October 1995. Available from:
- * ftp://ftp.cs.arizona.edu/xkernel/Papers/jsac.ps
- *
- * See http://www.cs.arizona.edu/xkernel/ for their implementation.
- * The main aspects that distinguish this implementation from the
- * Arizona Vegas implementation are:
- * o We do not change the loss detection or recovery mechanisms of
- * Linux in any way. Linux already recovers from losses quite well,
- * using fine-grained timers, NewReno, and FACK.
- * o To avoid the performance penalty imposed by increasing cwnd
- * only every-other RTT during slow start, we increase during
- * every RTT during slow start, just like Reno.
- * o Largely to allow continuous cwnd growth during slow start,
- * we use the rate at which ACKs come back as the "actual"
- * rate, rather than the rate at which data is sent.
- * o To speed convergence to the right rate, we set the cwnd
- * to achieve the right ("actual") rate when we exit slow start.
- * o To filter out the noise caused by delayed ACKs, we use the
- * minimum RTT sample observed during the last RTT to calculate
- * the actual rate.
- * o When the sender re-starts from idle, it waits until it has
- * received ACKs for an entire flight of new data before making
- * a cwnd adjustment decision. The original Vegas implementation
- * assumed senders never went idle.
- */
-static void vegas_cong_avoid(struct tcp_sock *tp, u32 ack, u32 seq_rtt)
-{
- /* The key players are v_beg_snd_una and v_beg_snd_nxt.
- *
- * These are so named because they represent the approximate values
- * of snd_una and snd_nxt at the beginning of the current RTT. More
- * precisely, they represent the amount of data sent during the RTT.
- * At the end of the RTT, when we receive an ACK for v_beg_snd_nxt,
- * we will calculate that (v_beg_snd_nxt - v_beg_snd_una) outstanding
- * bytes of data have been ACKed during the course of the RTT, giving
- * an "actual" rate of:
- *
- * (v_beg_snd_nxt - v_beg_snd_una) / (rtt duration)
- *
- * Unfortunately, v_beg_snd_una is not exactly equal to snd_una,
- * because delayed ACKs can cover more than one segment, so they
- * don't line up nicely with the boundaries of RTTs.
- *
- * Another unfortunate fact of life is that delayed ACKs delay the
- * advance of the left edge of our send window, so that the number
- * of bytes we send in an RTT is often less than our cwnd will allow.
- * So we keep track of our cwnd separately, in v_beg_snd_cwnd.
- */
-
- if (after(ack, tp->vegas.beg_snd_nxt)) {
- /* Do the Vegas once-per-RTT cwnd adjustment. */
- u32 old_wnd, old_snd_cwnd;
-
-
- /* Here old_wnd is essentially the window of data that was
- * sent during the previous RTT, and has all
- * been acknowledged in the course of the RTT that ended
- * with the ACK we just received. Likewise, old_snd_cwnd
- * is the cwnd during the previous RTT.
- */
- old_wnd = (tp->vegas.beg_snd_nxt - tp->vegas.beg_snd_una) /
- tp->mss_cache_std;
- old_snd_cwnd = tp->vegas.beg_snd_cwnd;
-
- /* Save the extent of the current window so we can use this
- * at the end of the next RTT.
- */
- tp->vegas.beg_snd_una = tp->vegas.beg_snd_nxt;
- tp->vegas.beg_snd_nxt = tp->snd_nxt;
- tp->vegas.beg_snd_cwnd = tp->snd_cwnd;
-
- /* Take into account the current RTT sample too, to
- * decrease the impact of delayed acks. This double counts
- * this sample since we count it for the next window as well,
- * but that's not too awful, since we're taking the min,
- * rather than averaging.
- */
- vegas_rtt_calc(tp, seq_rtt);
-
- /* We do the Vegas calculations only if we got enough RTT
- * samples that we can be reasonably sure that we got
- * at least one RTT sample that wasn't from a delayed ACK.
- * If we only had 2 samples total,
- * then that means we're getting only 1 ACK per RTT, which
- * means they're almost certainly delayed ACKs.
- * If we have 3 samples, we should be OK.
- */
-
- if (tp->vegas.cntRTT <= 2) {
- /* We don't have enough RTT samples to do the Vegas
- * calculation, so we'll behave like Reno.
- */
- if (tp->snd_cwnd > tp->snd_ssthresh)
- tp->snd_cwnd++;
- } else {
- u32 rtt, target_cwnd, diff;
-
- /* We have enough RTT samples, so, using the Vegas
- * algorithm, we determine if we should increase or
- * decrease cwnd, and by how much.
- */
-
- /* Pluck out the RTT we are using for the Vegas
- * calculations. This is the min RTT seen during the
- * last RTT. Taking the min filters out the effects
- * of delayed ACKs, at the cost of noticing congestion
- * a bit later.
- */
- rtt = tp->vegas.minRTT;
-
- /* Calculate the cwnd we should have, if we weren't
- * going too fast.
- *
- * This is:
- * (actual rate in segments) * baseRTT
- * We keep it as a fixed point number with
- * V_PARAM_SHIFT bits to the right of the binary point.
- */
- target_cwnd = ((old_wnd * tp->vegas.baseRTT)
- << V_PARAM_SHIFT) / rtt;
-
- /* Calculate the difference between the window we had,
- * and the window we would like to have. This quantity
- * is the "Diff" from the Arizona Vegas papers.
- *
- * Again, this is a fixed point number with
- * V_PARAM_SHIFT bits to the right of the binary
- * point.
- */
- diff = (old_wnd << V_PARAM_SHIFT) - target_cwnd;
-
- if (tp->snd_cwnd < tp->snd_ssthresh) {
- /* Slow start. */
- if (diff > sysctl_tcp_vegas_gamma) {
- /* Going too fast. Time to slow down
- * and switch to congestion avoidance.
- */
- tp->snd_ssthresh = 2;
-
- /* Set cwnd to match the actual rate
- * exactly:
- * cwnd = (actual rate) * baseRTT
- * Then we add 1 because the integer
- * truncation robs us of full link
- * utilization.
- */
- tp->snd_cwnd = min(tp->snd_cwnd,
- (target_cwnd >>
- V_PARAM_SHIFT)+1);
-
- }
- } else {
- /* Congestion avoidance. */
- u32 next_snd_cwnd;
-
- /* Figure out where we would like cwnd
- * to be.
- */
- if (diff > sysctl_tcp_vegas_beta) {
- /* The old window was too fast, so
- * we slow down.
- */
- next_snd_cwnd = old_snd_cwnd - 1;
- } else if (diff < sysctl_tcp_vegas_alpha) {
- /* We don't have enough extra packets
- * in the network, so speed up.
- */
- next_snd_cwnd = old_snd_cwnd + 1;
- } else {
- /* Sending just as fast as we
- * should be.
- */
- next_snd_cwnd = old_snd_cwnd;
- }
-
- /* Adjust cwnd upward or downward, toward the
- * desired value.
- */
- if (next_snd_cwnd > tp->snd_cwnd)
- tp->snd_cwnd++;
- else if (next_snd_cwnd < tp->snd_cwnd)
- tp->snd_cwnd--;
- }
- }
-
- /* Wipe the slate clean for the next RTT. */
- tp->vegas.cntRTT = 0;
- tp->vegas.minRTT = 0x7fffffff;
- }
-
- /* The following code is executed for every ack we receive,
- * except for conditions checked in should_advance_cwnd()
- * before the call to tcp_cong_avoid(). Mainly this means that
- * we only execute this code if the ack actually acked some
- * data.
- */
-
- /* If we are in slow start, increase our cwnd in response to this ACK.
- * (If we are not in slow start then we are in congestion avoidance,
- * and adjust our congestion window only once per RTT. See the code
- * above.)
- */
- if (tp->snd_cwnd <= tp->snd_ssthresh)
- tp->snd_cwnd++;
-
- /* to keep cwnd from growing without bound */
- tp->snd_cwnd = min_t(u32, tp->snd_cwnd, tp->snd_cwnd_clamp);
-
- /* Make sure that we are never so timid as to reduce our cwnd below
- * 2 MSS.
- *
- * Going below 2 MSS would risk huge delayed ACKs from our receiver.
- */
- tp->snd_cwnd = max(tp->snd_cwnd, 2U);
-
- tp->snd_cwnd_stamp = tcp_time_stamp;
-}
-
-static inline void tcp_cong_avoid(struct tcp_sock *tp, u32 ack, u32 seq_rtt)
-{
- if (tcp_vegas_enabled(tp))
- vegas_cong_avoid(tp, ack, seq_rtt);
- else
- reno_cong_avoid(tp);
-}
-
/* Restart timer after forward progress on connection.
* RFC2988 recommends to restart timer to now+rto.
*/
@@ -2348,15 +1957,6 @@ static inline void tcp_ack_packets_out(struct sock *sk, struct tcp_sock *tp)
}
}
-/* There is one downside to this scheme. Although we keep the
- * ACK clock ticking, adjusting packet counters and advancing
- * congestion window, we do not liberate socket send buffer
- * space.
- *
- * Mucking with skb->truesize and sk->sk_wmem_alloc et al.
- * then making a write space wakeup callback is a possible
- * future enhancement. WARNING: it is not trivial to make.
- */
static int tcp_tso_acked(struct sock *sk, struct sk_buff *skb,
__u32 now, __s32 *seq_rtt)
{
@@ -2415,13 +2015,18 @@ static int tcp_tso_acked(struct sock *sk, struct sk_buff *skb,
/* Remove acknowledged frames from the retransmission queue. */
-static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
+static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p, s32 *seq_usrtt)
{
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
__u32 now = tcp_time_stamp;
int acked = 0;
__s32 seq_rtt = -1;
+ struct timeval usnow;
+ u32 pkts_acked = 0;
+
+ if (seq_usrtt)
+ do_gettimeofday(&usnow);
while ((skb = skb_peek(&sk->sk_write_queue)) &&
skb != sk->sk_send_head) {
@@ -2433,7 +2038,8 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
* the other end.
*/
if (after(scb->end_seq, tp->snd_una)) {
- if (tcp_skb_pcount(skb) > 1)
+ if (tcp_skb_pcount(skb) > 1 &&
+ after(tp->snd_una, scb->seq))
acked |= tcp_tso_acked(sk, skb,
now, &seq_rtt);
break;
@@ -2448,6 +2054,7 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
*/
if (!(scb->flags & TCPCB_FLAG_SYN)) {
acked |= FLAG_DATA_ACKED;
+ ++pkts_acked;
} else {
acked |= FLAG_SYN_ACKED;
tp->retrans_stamp = 0;
@@ -2461,6 +2068,10 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
seq_rtt = -1;
} else if (seq_rtt < 0)
seq_rtt = now - scb->when;
+ if (seq_usrtt)
+ *seq_usrtt = (usnow.tv_sec - skb->stamp.tv_sec) * 1000000
+ + (usnow.tv_usec - skb->stamp.tv_usec);
+
if (sacked & TCPCB_SACKED_ACKED)
tp->sacked_out -= tcp_skb_pcount(skb);
if (sacked & TCPCB_LOST)
@@ -2479,8 +2090,11 @@ static int tcp_clean_rtx_queue(struct sock *sk, __s32 *seq_rtt_p)
}
if (acked&FLAG_ACKED) {
- tcp_ack_update_rtt(tp, acked, seq_rtt);
+ tcp_ack_update_rtt(tp, acked, seq_rtt, seq_usrtt);
tcp_ack_packets_out(sk, tp);
+
+ if (tp->ca_ops->pkts_acked)
+ tp->ca_ops->pkts_acked(tp, pkts_acked);
}
#if FASTRETRANS_DEBUG > 0
@@ -2624,257 +2238,6 @@ static void tcp_process_frto(struct sock *sk, u32 prior_snd_una)
tp->frto_counter = (tp->frto_counter + 1) % 3;
}
-/*
- * TCP Westwood+
- */
-
-/*
- * @init_westwood
- * This function initializes fields used in TCP Westwood+. We can't
- * get no information about RTTmin at this time so we simply set it to
- * TCP_WESTWOOD_INIT_RTT. This value was chosen to be too conservative
- * since in this way we're sure it will be updated in a consistent
- * way as soon as possible. It will reasonably happen within the first
- * RTT period of the connection lifetime.
- */
-
-static void init_westwood(struct sock *sk)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- tp->westwood.bw_ns_est = 0;
- tp->westwood.bw_est = 0;
- tp->westwood.accounted = 0;
- tp->westwood.cumul_ack = 0;
- tp->westwood.rtt_win_sx = tcp_time_stamp;
- tp->westwood.rtt = TCP_WESTWOOD_INIT_RTT;
- tp->westwood.rtt_min = TCP_WESTWOOD_INIT_RTT;
- tp->westwood.snd_una = tp->snd_una;
-}
-
-/*
- * @westwood_do_filter
- * Low-pass filter. Implemented using constant coeffients.
- */
-
-static inline __u32 westwood_do_filter(__u32 a, __u32 b)
-{
- return (((7 * a) + b) >> 3);
-}
-
-static void westwood_filter(struct sock *sk, __u32 delta)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- tp->westwood.bw_ns_est =
- westwood_do_filter(tp->westwood.bw_ns_est,
- tp->westwood.bk / delta);
- tp->westwood.bw_est =
- westwood_do_filter(tp->westwood.bw_est,
- tp->westwood.bw_ns_est);
-}
-
-/*
- * @westwood_update_rttmin
- * It is used to update RTTmin. In this case we MUST NOT use
- * WESTWOOD_RTT_MIN minimum bound since we could be on a LAN!
- */
-
-static inline __u32 westwood_update_rttmin(const struct sock *sk)
-{
- const struct tcp_sock *tp = tcp_sk(sk);
- __u32 rttmin = tp->westwood.rtt_min;
-
- if (tp->westwood.rtt != 0 &&
- (tp->westwood.rtt < tp->westwood.rtt_min || !rttmin))
- rttmin = tp->westwood.rtt;
-
- return rttmin;
-}
-
-/*
- * @westwood_acked
- * Evaluate increases for dk.
- */
-
-static inline __u32 westwood_acked(const struct sock *sk)
-{
- const struct tcp_sock *tp = tcp_sk(sk);
-
- return tp->snd_una - tp->westwood.snd_una;
-}
-
-/*
- * @westwood_new_window
- * It evaluates if we are receiving data inside the same RTT window as
- * when we started.
- * Return value:
- * It returns 0 if we are still evaluating samples in the same RTT
- * window, 1 if the sample has to be considered in the next window.
- */
-
-static int westwood_new_window(const struct sock *sk)
-{
- const struct tcp_sock *tp = tcp_sk(sk);
- __u32 left_bound;
- __u32 rtt;
- int ret = 0;
-
- left_bound = tp->westwood.rtt_win_sx;
- rtt = max(tp->westwood.rtt, (u32) TCP_WESTWOOD_RTT_MIN);
-
- /*
- * A RTT-window has passed. Be careful since if RTT is less than
- * 50ms we don't filter but we continue 'building the sample'.
- * This minimum limit was choosen since an estimation on small
- * time intervals is better to avoid...
- * Obvioulsy on a LAN we reasonably will always have
- * right_bound = left_bound + WESTWOOD_RTT_MIN
- */
-
- if ((left_bound + rtt) < tcp_time_stamp)
- ret = 1;
-
- return ret;
-}
-
-/*
- * @westwood_update_window
- * It updates RTT evaluation window if it is the right moment to do
- * it. If so it calls filter for evaluating bandwidth.
- */
-
-static void __westwood_update_window(struct sock *sk, __u32 now)
-{
- struct tcp_sock *tp = tcp_sk(sk);
- __u32 delta = now - tp->westwood.rtt_win_sx;
-
- if (delta) {
- if (tp->westwood.rtt)
- westwood_filter(sk, delta);
-
- tp->westwood.bk = 0;
- tp->westwood.rtt_win_sx = tcp_time_stamp;
- }
-}
-
-
-static void westwood_update_window(struct sock *sk, __u32 now)
-{
- if (westwood_new_window(sk))
- __westwood_update_window(sk, now);
-}
-
-/*
- * @__tcp_westwood_fast_bw
- * It is called when we are in fast path. In particular it is called when
- * header prediction is successfull. In such case infact update is
- * straight forward and doesn't need any particular care.
- */
-
-static void __tcp_westwood_fast_bw(struct sock *sk, struct sk_buff *skb)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- westwood_update_window(sk, tcp_time_stamp);
-
- tp->westwood.bk += westwood_acked(sk);
- tp->westwood.snd_una = tp->snd_una;
- tp->westwood.rtt_min = westwood_update_rttmin(sk);
-}
-
-static inline void tcp_westwood_fast_bw(struct sock *sk, struct sk_buff *skb)
-{
- if (tcp_is_westwood(tcp_sk(sk)))
- __tcp_westwood_fast_bw(sk, skb);
-}
-
-
-/*
- * @westwood_dupack_update
- * It updates accounted and cumul_ack when receiving a dupack.
- */
-
-static void westwood_dupack_update(struct sock *sk)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- tp->westwood.accounted += tp->mss_cache_std;
- tp->westwood.cumul_ack = tp->mss_cache_std;
-}
-
-static inline int westwood_may_change_cumul(struct tcp_sock *tp)
-{
- return (tp->westwood.cumul_ack > tp->mss_cache_std);
-}
-
-static inline void westwood_partial_update(struct tcp_sock *tp)
-{
- tp->westwood.accounted -= tp->westwood.cumul_ack;
- tp->westwood.cumul_ack = tp->mss_cache_std;
-}
-
-static inline void westwood_complete_update(struct tcp_sock *tp)
-{
- tp->westwood.cumul_ack -= tp->westwood.accounted;
- tp->westwood.accounted = 0;
-}
-
-/*
- * @westwood_acked_count
- * This function evaluates cumul_ack for evaluating dk in case of
- * delayed or partial acks.
- */
-
-static inline __u32 westwood_acked_count(struct sock *sk)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- tp->westwood.cumul_ack = westwood_acked(sk);
-
- /* If cumul_ack is 0 this is a dupack since it's not moving
- * tp->snd_una.
- */
- if (!(tp->westwood.cumul_ack))
- westwood_dupack_update(sk);
-
- if (westwood_may_change_cumul(tp)) {
- /* Partial or delayed ack */
- if (tp->westwood.accounted >= tp->westwood.cumul_ack)
- westwood_partial_update(tp);
- else
- westwood_complete_update(tp);
- }
-
- tp->westwood.snd_una = tp->snd_una;
-
- return tp->westwood.cumul_ack;
-}
-
-
-/*
- * @__tcp_westwood_slow_bw
- * It is called when something is going wrong..even if there could
- * be no problems! Infact a simple delayed packet may trigger a
- * dupack. But we need to be careful in such case.
- */
-
-static void __tcp_westwood_slow_bw(struct sock *sk, struct sk_buff *skb)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- westwood_update_window(sk, tcp_time_stamp);
-
- tp->westwood.bk += westwood_acked_count(sk);
- tp->westwood.rtt_min = westwood_update_rttmin(sk);
-}
-
-static inline void tcp_westwood_slow_bw(struct sock *sk, struct sk_buff *skb)
-{
- if (tcp_is_westwood(tcp_sk(sk)))
- __tcp_westwood_slow_bw(sk, skb);
-}
-
/* This routine deals with incoming acks, but not outgoing ones. */
static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
{
@@ -2884,6 +2247,7 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
u32 ack = TCP_SKB_CB(skb)->ack_seq;
u32 prior_in_flight;
s32 seq_rtt;
+ s32 seq_usrtt = 0;
int prior_packets;
/* If the ack is newer than sent or older than previous acks
@@ -2902,9 +2266,10 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
*/
tcp_update_wl(tp, ack, ack_seq);
tp->snd_una = ack;
- tcp_westwood_fast_bw(sk, skb);
flag |= FLAG_WIN_UPDATE;
+ tcp_ca_event(tp, CA_EVENT_FAST_ACK);
+
NET_INC_STATS_BH(LINUX_MIB_TCPHPACKS);
} else {
if (ack_seq != TCP_SKB_CB(skb)->end_seq)
@@ -2920,7 +2285,7 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
if (TCP_ECN_rcv_ecn_echo(tp, skb->h.th))
flag |= FLAG_ECE;
- tcp_westwood_slow_bw(sk,skb);
+ tcp_ca_event(tp, CA_EVENT_SLOW_ACK);
}
/* We passed data and got it acked, remove any soft error
@@ -2935,22 +2300,20 @@ static int tcp_ack(struct sock *sk, struct sk_buff *skb, int flag)
prior_in_flight = tcp_packets_in_flight(tp);
/* See if we can take anything off of the retransmit queue. */
- flag |= tcp_clean_rtx_queue(sk, &seq_rtt);
+ flag |= tcp_clean_rtx_queue(sk, &seq_rtt,
+ tp->ca_ops->rtt_sample ? &seq_usrtt : NULL);
if (tp->frto_counter)
tcp_process_frto(sk, prior_snd_una);
if (tcp_ack_is_dubious(tp, flag)) {
/* Advanve CWND, if state allows this. */
- if ((flag & FLAG_DATA_ACKED) &&
- (tcp_vegas_enabled(tp) || prior_in_flight >= tp->snd_cwnd) &&
- tcp_may_raise_cwnd(tp, flag))
- tcp_cong_avoid(tp, ack, seq_rtt);
+ if ((flag & FLAG_DATA_ACKED) && tcp_may_raise_cwnd(tp, flag))
+ tcp_cong_avoid(tp, ack, seq_rtt, prior_in_flight, 0);
tcp_fastretrans_alert(sk, prior_snd_una, prior_packets, flag);
} else {
- if ((flag & FLAG_DATA_ACKED) &&
- (tcp_vegas_enabled(tp) || prior_in_flight >= tp->snd_cwnd))
- tcp_cong_avoid(tp, ack, seq_rtt);
+ if ((flag & FLAG_DATA_ACKED))
+ tcp_cong_avoid(tp, ack, seq_rtt, prior_in_flight, 1);
}
if ((flag & FLAG_FORWARD_PROGRESS) || !(flag&FLAG_NOT_DUP))
@@ -3439,7 +2802,7 @@ static void tcp_sack_remove(struct tcp_sock *tp)
int this_sack;
/* Empty ofo queue, hence, all the SACKs are eaten. Clear. */
- if (skb_queue_len(&tp->out_of_order_queue) == 0) {
+ if (skb_queue_empty(&tp->out_of_order_queue)) {
tp->rx_opt.num_sacks = 0;
tp->rx_opt.eff_sacks = tp->rx_opt.dsack;
return;
@@ -3572,13 +2935,13 @@ queue_and_out:
if(th->fin)
tcp_fin(skb, sk, th);
- if (skb_queue_len(&tp->out_of_order_queue)) {
+ if (!skb_queue_empty(&tp->out_of_order_queue)) {
tcp_ofo_queue(sk);
/* RFC2581. 4.2. SHOULD send immediate ACK, when
* gap in queue is filled.
*/
- if (!skb_queue_len(&tp->out_of_order_queue))
+ if (skb_queue_empty(&tp->out_of_order_queue))
tp->ack.pingpong = 0;
}
@@ -3886,9 +3249,8 @@ static int tcp_prune_queue(struct sock *sk)
* This must not ever occur. */
/* First, purge the out_of_order queue. */
- if (skb_queue_len(&tp->out_of_order_queue)) {
- NET_ADD_STATS_BH(LINUX_MIB_OFOPRUNED,
- skb_queue_len(&tp->out_of_order_queue));
+ if (!skb_queue_empty(&tp->out_of_order_queue)) {
+ NET_INC_STATS_BH(LINUX_MIB_OFOPRUNED);
__skb_queue_purge(&tp->out_of_order_queue);
/* Reset SACK state. A conforming SACK implementation will
@@ -3937,6 +3299,28 @@ void tcp_cwnd_application_limited(struct sock *sk)
tp->snd_cwnd_stamp = tcp_time_stamp;
}
+static inline int tcp_should_expand_sndbuf(struct sock *sk, struct tcp_sock *tp)
+{
+ /* If the user specified a specific send buffer setting, do
+ * not modify it.
+ */
+ if (sk->sk_userlocks & SOCK_SNDBUF_LOCK)
+ return 0;
+
+ /* If we are under global TCP memory pressure, do not expand. */
+ if (tcp_memory_pressure)
+ return 0;
+
+ /* If we are under soft global TCP memory pressure, do not expand. */
+ if (atomic_read(&tcp_memory_allocated) >= sysctl_tcp_mem[0])
+ return 0;
+
+ /* If we filled the congestion window, do not expand. */
+ if (tp->packets_out >= tp->snd_cwnd)
+ return 0;
+
+ return 1;
+}
/* When incoming ACK allowed to free some skb from write_queue,
* we remember this event in flag SOCK_QUEUE_SHRUNK and wake up socket
@@ -3948,11 +3332,8 @@ static void tcp_new_space(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- if (tp->packets_out < tp->snd_cwnd &&
- !(sk->sk_userlocks & SOCK_SNDBUF_LOCK) &&
- !tcp_memory_pressure &&
- atomic_read(&tcp_memory_allocated) < sysctl_tcp_mem[0]) {
- int sndmem = max_t(u32, tp->rx_opt.mss_clamp, tp->mss_cache_std) +
+ if (tcp_should_expand_sndbuf(sk, tp)) {
+ int sndmem = max_t(u32, tp->rx_opt.mss_clamp, tp->mss_cache) +
MAX_TCP_HEADER + 16 + sizeof(struct sk_buff),
demanded = max_t(unsigned int, tp->snd_cwnd,
tp->reordering + 1);
@@ -3975,22 +3356,9 @@ static inline void tcp_check_space(struct sock *sk)
}
}
-static void __tcp_data_snd_check(struct sock *sk, struct sk_buff *skb)
+static __inline__ void tcp_data_snd_check(struct sock *sk, struct tcp_sock *tp)
{
- struct tcp_sock *tp = tcp_sk(sk);
-
- if (after(TCP_SKB_CB(skb)->end_seq, tp->snd_una + tp->snd_wnd) ||
- tcp_packets_in_flight(tp) >= tp->snd_cwnd ||
- tcp_write_xmit(sk, tp->nonagle))
- tcp_check_probe_timer(sk, tp);
-}
-
-static __inline__ void tcp_data_snd_check(struct sock *sk)
-{
- struct sk_buff *skb = sk->sk_send_head;
-
- if (skb != NULL)
- __tcp_data_snd_check(sk, skb);
+ tcp_push_pending_frames(sk, tp);
tcp_check_space(sk);
}
@@ -4284,7 +3652,7 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
*/
tcp_ack(sk, skb, 0);
__kfree_skb(skb);
- tcp_data_snd_check(sk);
+ tcp_data_snd_check(sk, tp);
return 0;
} else { /* Header too small */
TCP_INC_STATS_BH(TCP_MIB_INERRS);
@@ -4350,7 +3718,7 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
if (TCP_SKB_CB(skb)->ack_seq != tp->snd_una) {
/* Well, only one small jumplet in fast path... */
tcp_ack(sk, skb, FLAG_DATA);
- tcp_data_snd_check(sk);
+ tcp_data_snd_check(sk, tp);
if (!tcp_ack_scheduled(tp))
goto no_ack;
}
@@ -4428,7 +3796,7 @@ step5:
/* step 7: process the segment text */
tcp_data_queue(sk, skb);
- tcp_data_snd_check(sk);
+ tcp_data_snd_check(sk, tp);
tcp_ack_snd_check(sk);
return 0;
@@ -4552,6 +3920,8 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
tcp_init_metrics(sk);
+ tcp_init_congestion_control(tp);
+
/* Prevent spurious tcp_cwnd_restart() on first data
* packet.
*/
@@ -4708,9 +4078,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
if(tp->af_specific->conn_request(sk, skb) < 0)
return 1;
- init_westwood(sk);
- init_bictcp(tp);
-
/* Now we have several options: In theory there is
* nothing else in the frame. KA9Q has an option to
* send data with the syn, BSD accepts data with the
@@ -4732,9 +4099,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
goto discard;
case TCP_SYN_SENT:
- init_westwood(sk);
- init_bictcp(tp);
-
queued = tcp_rcv_synsent_state_process(sk, skb, th, len);
if (queued >= 0)
return queued;
@@ -4742,7 +4106,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
/* Do step6 onward by hand. */
tcp_urg(sk, skb, th);
__kfree_skb(skb);
- tcp_data_snd_check(sk);
+ tcp_data_snd_check(sk, tp);
return 0;
}
@@ -4816,7 +4180,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
*/
if (tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr &&
!tp->srtt)
- tcp_ack_saw_tstamp(tp, 0);
+ tcp_ack_saw_tstamp(tp, 0, 0);
if (tp->rx_opt.tstamp_ok)
tp->advmss -= TCPOLEN_TSTAMP_ALIGNED;
@@ -4828,6 +4192,8 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
tcp_init_metrics(sk);
+ tcp_init_congestion_control(tp);
+
/* Prevent spurious tcp_cwnd_restart() on
* first data packet.
*/
@@ -4931,7 +4297,7 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
/* tcp_data could move socket to TIME-WAIT */
if (sk->sk_state != TCP_CLOSE) {
- tcp_data_snd_check(sk);
+ tcp_data_snd_check(sk, tp);
tcp_ack_snd_check(sk);
}
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index dad98e4a504..62f62bb05c2 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -36,7 +36,7 @@
* ACK bit.
* Andi Kleen : Implemented fast path mtu discovery.
* Fixed many serious bugs in the
- * open_request handling and moved
+ * request_sock handling and moved
* most of it into the af independent code.
* Added tail drop and some other bugfixes.
* Added new listen sematics.
@@ -869,21 +869,23 @@ static __inline__ u32 tcp_v4_synq_hash(u32 raddr, u16 rport, u32 rnd)
return (jhash_2words(raddr, (u32) rport, rnd) & (TCP_SYNQ_HSIZE - 1));
}
-static struct open_request *tcp_v4_search_req(struct tcp_sock *tp,
- struct open_request ***prevp,
+static struct request_sock *tcp_v4_search_req(struct tcp_sock *tp,
+ struct request_sock ***prevp,
__u16 rport,
__u32 raddr, __u32 laddr)
{
- struct tcp_listen_opt *lopt = tp->listen_opt;
- struct open_request *req, **prev;
+ struct listen_sock *lopt = tp->accept_queue.listen_opt;
+ struct request_sock *req, **prev;
for (prev = &lopt->syn_table[tcp_v4_synq_hash(raddr, rport, lopt->hash_rnd)];
(req = *prev) != NULL;
prev = &req->dl_next) {
- if (req->rmt_port == rport &&
- req->af.v4_req.rmt_addr == raddr &&
- req->af.v4_req.loc_addr == laddr &&
- TCP_INET_FAMILY(req->class->family)) {
+ const struct inet_request_sock *ireq = inet_rsk(req);
+
+ if (ireq->rmt_port == rport &&
+ ireq->rmt_addr == raddr &&
+ ireq->loc_addr == laddr &&
+ TCP_INET_FAMILY(req->rsk_ops->family)) {
BUG_TRAP(!req->sk);
*prevp = prev;
break;
@@ -893,21 +895,13 @@ static struct open_request *tcp_v4_search_req(struct tcp_sock *tp,
return req;
}
-static void tcp_v4_synq_add(struct sock *sk, struct open_request *req)
+static void tcp_v4_synq_add(struct sock *sk, struct request_sock *req)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
- u32 h = tcp_v4_synq_hash(req->af.v4_req.rmt_addr, req->rmt_port, lopt->hash_rnd);
-
- req->expires = jiffies + TCP_TIMEOUT_INIT;
- req->retrans = 0;
- req->sk = NULL;
- req->dl_next = lopt->syn_table[h];
-
- write_lock(&tp->syn_wait_lock);
- lopt->syn_table[h] = req;
- write_unlock(&tp->syn_wait_lock);
+ struct listen_sock *lopt = tp->accept_queue.listen_opt;
+ u32 h = tcp_v4_synq_hash(inet_rsk(req)->rmt_addr, inet_rsk(req)->rmt_port, lopt->hash_rnd);
+ reqsk_queue_hash_req(&tp->accept_queue, h, req, TCP_TIMEOUT_INIT);
tcp_synq_added(sk);
}
@@ -1050,7 +1044,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
}
switch (sk->sk_state) {
- struct open_request *req, **prev;
+ struct request_sock *req, **prev;
case TCP_LISTEN:
if (sock_owned_by_user(sk))
goto out;
@@ -1065,7 +1059,7 @@ void tcp_v4_err(struct sk_buff *skb, u32 info)
*/
BUG_TRAP(!req->sk);
- if (seq != req->snt_isn) {
+ if (seq != tcp_rsk(req)->snt_isn) {
NET_INC_STATS_BH(LINUX_MIB_OUTOFWINDOWICMPS);
goto out;
}
@@ -1254,28 +1248,29 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
tcp_tw_put(tw);
}
-static void tcp_v4_or_send_ack(struct sk_buff *skb, struct open_request *req)
+static void tcp_v4_reqsk_send_ack(struct sk_buff *skb, struct request_sock *req)
{
- tcp_v4_send_ack(skb, req->snt_isn + 1, req->rcv_isn + 1, req->rcv_wnd,
+ tcp_v4_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd,
req->ts_recent);
}
static struct dst_entry* tcp_v4_route_req(struct sock *sk,
- struct open_request *req)
+ struct request_sock *req)
{
struct rtable *rt;
- struct ip_options *opt = req->af.v4_req.opt;
+ const struct inet_request_sock *ireq = inet_rsk(req);
+ struct ip_options *opt = inet_rsk(req)->opt;
struct flowi fl = { .oif = sk->sk_bound_dev_if,
.nl_u = { .ip4_u =
{ .daddr = ((opt && opt->srr) ?
opt->faddr :
- req->af.v4_req.rmt_addr),
- .saddr = req->af.v4_req.loc_addr,
+ ireq->rmt_addr),
+ .saddr = ireq->loc_addr,
.tos = RT_CONN_FLAGS(sk) } },
.proto = IPPROTO_TCP,
.uli_u = { .ports =
{ .sport = inet_sk(sk)->sport,
- .dport = req->rmt_port } } };
+ .dport = ireq->rmt_port } } };
if (ip_route_output_flow(&rt, &fl, sk, 0)) {
IP_INC_STATS_BH(IPSTATS_MIB_OUTNOROUTES);
@@ -1291,12 +1286,13 @@ static struct dst_entry* tcp_v4_route_req(struct sock *sk,
/*
* Send a SYN-ACK after having received an ACK.
- * This still operates on a open_request only, not on a big
+ * This still operates on a request_sock only, not on a big
* socket.
*/
-static int tcp_v4_send_synack(struct sock *sk, struct open_request *req,
+static int tcp_v4_send_synack(struct sock *sk, struct request_sock *req,
struct dst_entry *dst)
{
+ const struct inet_request_sock *ireq = inet_rsk(req);
int err = -1;
struct sk_buff * skb;
@@ -1310,14 +1306,14 @@ static int tcp_v4_send_synack(struct sock *sk, struct open_request *req,
struct tcphdr *th = skb->h.th;
th->check = tcp_v4_check(th, skb->len,
- req->af.v4_req.loc_addr,
- req->af.v4_req.rmt_addr,
+ ireq->loc_addr,
+ ireq->rmt_addr,
csum_partial((char *)th, skb->len,
skb->csum));
- err = ip_build_and_send_pkt(skb, sk, req->af.v4_req.loc_addr,
- req->af.v4_req.rmt_addr,
- req->af.v4_req.opt);
+ err = ip_build_and_send_pkt(skb, sk, ireq->loc_addr,
+ ireq->rmt_addr,
+ ireq->opt);
if (err == NET_XMIT_CN)
err = 0;
}
@@ -1328,12 +1324,12 @@ out:
}
/*
- * IPv4 open_request destructor.
+ * IPv4 request_sock destructor.
*/
-static void tcp_v4_or_free(struct open_request *req)
+static void tcp_v4_reqsk_destructor(struct request_sock *req)
{
- if (req->af.v4_req.opt)
- kfree(req->af.v4_req.opt);
+ if (inet_rsk(req)->opt)
+ kfree(inet_rsk(req)->opt);
}
static inline void syn_flood_warning(struct sk_buff *skb)
@@ -1349,7 +1345,7 @@ static inline void syn_flood_warning(struct sk_buff *skb)
}
/*
- * Save and compile IPv4 options into the open_request if needed.
+ * Save and compile IPv4 options into the request_sock if needed.
*/
static inline struct ip_options *tcp_v4_save_options(struct sock *sk,
struct sk_buff *skb)
@@ -1370,33 +1366,20 @@ static inline struct ip_options *tcp_v4_save_options(struct sock *sk,
return dopt;
}
-/*
- * Maximum number of SYN_RECV sockets in queue per LISTEN socket.
- * One SYN_RECV socket costs about 80bytes on a 32bit machine.
- * It would be better to replace it with a global counter for all sockets
- * but then some measure against one socket starving all other sockets
- * would be needed.
- *
- * It was 128 by default. Experiments with real servers show, that
- * it is absolutely not enough even at 100conn/sec. 256 cures most
- * of problems. This value is adjusted to 128 for very small machines
- * (<=32Mb of memory) and to 1024 on normal or better ones (>=256Mb).
- * Further increasing requires to change hash table size.
- */
-int sysctl_max_syn_backlog = 256;
-
-struct or_calltable or_ipv4 = {
+struct request_sock_ops tcp_request_sock_ops = {
.family = PF_INET,
+ .obj_size = sizeof(struct tcp_request_sock),
.rtx_syn_ack = tcp_v4_send_synack,
- .send_ack = tcp_v4_or_send_ack,
- .destructor = tcp_v4_or_free,
+ .send_ack = tcp_v4_reqsk_send_ack,
+ .destructor = tcp_v4_reqsk_destructor,
.send_reset = tcp_v4_send_reset,
};
int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
{
+ struct inet_request_sock *ireq;
struct tcp_options_received tmp_opt;
- struct open_request *req;
+ struct request_sock *req;
__u32 saddr = skb->nh.iph->saddr;
__u32 daddr = skb->nh.iph->daddr;
__u32 isn = TCP_SKB_CB(skb)->when;
@@ -1433,7 +1416,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
if (sk_acceptq_is_full(sk) && tcp_synq_young(sk) > 1)
goto drop;
- req = tcp_openreq_alloc();
+ req = reqsk_alloc(&tcp_request_sock_ops);
if (!req)
goto drop;
@@ -1461,10 +1444,10 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
tcp_openreq_init(req, &tmp_opt, skb);
- req->af.v4_req.loc_addr = daddr;
- req->af.v4_req.rmt_addr = saddr;
- req->af.v4_req.opt = tcp_v4_save_options(sk, skb);
- req->class = &or_ipv4;
+ ireq = inet_rsk(req);
+ ireq->loc_addr = daddr;
+ ireq->rmt_addr = saddr;
+ ireq->opt = tcp_v4_save_options(sk, skb);
if (!want_cookie)
TCP_ECN_create_request(req, skb->h.th);
@@ -1523,20 +1506,20 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
isn = tcp_v4_init_sequence(sk, skb);
}
- req->snt_isn = isn;
+ tcp_rsk(req)->snt_isn = isn;
if (tcp_v4_send_synack(sk, req, dst))
goto drop_and_free;
if (want_cookie) {
- tcp_openreq_free(req);
+ reqsk_free(req);
} else {
tcp_v4_synq_add(sk, req);
}
return 0;
drop_and_free:
- tcp_openreq_free(req);
+ reqsk_free(req);
drop:
TCP_INC_STATS_BH(TCP_MIB_ATTEMPTFAILS);
return 0;
@@ -1548,9 +1531,10 @@ drop:
* now create the new socket.
*/
struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
- struct open_request *req,
+ struct request_sock *req,
struct dst_entry *dst)
{
+ struct inet_request_sock *ireq;
struct inet_sock *newinet;
struct tcp_sock *newtp;
struct sock *newsk;
@@ -1570,11 +1554,12 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
newtp = tcp_sk(newsk);
newinet = inet_sk(newsk);
- newinet->daddr = req->af.v4_req.rmt_addr;
- newinet->rcv_saddr = req->af.v4_req.loc_addr;
- newinet->saddr = req->af.v4_req.loc_addr;
- newinet->opt = req->af.v4_req.opt;
- req->af.v4_req.opt = NULL;
+ ireq = inet_rsk(req);
+ newinet->daddr = ireq->rmt_addr;
+ newinet->rcv_saddr = ireq->loc_addr;
+ newinet->saddr = ireq->loc_addr;
+ newinet->opt = ireq->opt;
+ ireq->opt = NULL;
newinet->mc_index = tcp_v4_iif(skb);
newinet->mc_ttl = skb->nh.iph->ttl;
newtp->ext_header_len = 0;
@@ -1605,9 +1590,9 @@ static struct sock *tcp_v4_hnd_req(struct sock *sk, struct sk_buff *skb)
struct iphdr *iph = skb->nh.iph;
struct tcp_sock *tp = tcp_sk(sk);
struct sock *nsk;
- struct open_request **prev;
+ struct request_sock **prev;
/* Find possible connection requests. */
- struct open_request *req = tcp_v4_search_req(tp, &prev, th->source,
+ struct request_sock *req = tcp_v4_search_req(tp, &prev, th->source,
iph->saddr, iph->daddr);
if (req)
return tcp_check_req(sk, skb, req, prev);
@@ -2060,9 +2045,10 @@ static int tcp_v4_init_sock(struct sock *sk)
*/
tp->snd_ssthresh = 0x7fffffff; /* Infinity */
tp->snd_cwnd_clamp = ~0;
- tp->mss_cache_std = tp->mss_cache = 536;
+ tp->mss_cache = 536;
tp->reordering = sysctl_tcp_reordering;
+ tp->ca_ops = &tcp_init_congestion_ops;
sk->sk_state = TCP_CLOSE;
@@ -2085,6 +2071,8 @@ int tcp_v4_destroy_sock(struct sock *sk)
tcp_clear_xmit_timers(sk);
+ tcp_cleanup_congestion_control(tp);
+
/* Cleanup up the write buffer. */
sk_stream_writequeue_purge(sk);
@@ -2144,13 +2132,13 @@ static void *listening_get_next(struct seq_file *seq, void *cur)
++st->num;
if (st->state == TCP_SEQ_STATE_OPENREQ) {
- struct open_request *req = cur;
+ struct request_sock *req = cur;
tp = tcp_sk(st->syn_wait_sk);
req = req->dl_next;
while (1) {
while (req) {
- if (req->class->family == st->family) {
+ if (req->rsk_ops->family == st->family) {
cur = req;
goto out;
}
@@ -2159,17 +2147,17 @@ static void *listening_get_next(struct seq_file *seq, void *cur)
if (++st->sbucket >= TCP_SYNQ_HSIZE)
break;
get_req:
- req = tp->listen_opt->syn_table[st->sbucket];
+ req = tp->accept_queue.listen_opt->syn_table[st->sbucket];
}
sk = sk_next(st->syn_wait_sk);
st->state = TCP_SEQ_STATE_LISTENING;
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
} else {
tp = tcp_sk(sk);
- read_lock_bh(&tp->syn_wait_lock);
- if (tp->listen_opt && tp->listen_opt->qlen)
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
+ if (reqsk_queue_len(&tp->accept_queue))
goto start_req;
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
sk = sk_next(sk);
}
get_sk:
@@ -2179,8 +2167,8 @@ get_sk:
goto out;
}
tp = tcp_sk(sk);
- read_lock_bh(&tp->syn_wait_lock);
- if (tp->listen_opt && tp->listen_opt->qlen) {
+ read_lock_bh(&tp->accept_queue.syn_wait_lock);
+ if (reqsk_queue_len(&tp->accept_queue)) {
start_req:
st->uid = sock_i_uid(sk);
st->syn_wait_sk = sk;
@@ -2188,7 +2176,7 @@ start_req:
st->sbucket = 0;
goto get_req;
}
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
}
if (++st->bucket < TCP_LHTABLE_SIZE) {
sk = sk_head(&tcp_listening_hash[st->bucket]);
@@ -2375,7 +2363,7 @@ static void tcp_seq_stop(struct seq_file *seq, void *v)
case TCP_SEQ_STATE_OPENREQ:
if (v) {
struct tcp_sock *tp = tcp_sk(st->syn_wait_sk);
- read_unlock_bh(&tp->syn_wait_lock);
+ read_unlock_bh(&tp->accept_queue.syn_wait_lock);
}
case TCP_SEQ_STATE_LISTENING:
if (v != SEQ_START_TOKEN)
@@ -2451,18 +2439,19 @@ void tcp_proc_unregister(struct tcp_seq_afinfo *afinfo)
memset(afinfo->seq_fops, 0, sizeof(*afinfo->seq_fops));
}
-static void get_openreq4(struct sock *sk, struct open_request *req,
+static void get_openreq4(struct sock *sk, struct request_sock *req,
char *tmpbuf, int i, int uid)
{
+ const struct inet_request_sock *ireq = inet_rsk(req);
int ttd = req->expires - jiffies;
sprintf(tmpbuf, "%4d: %08X:%04X %08X:%04X"
" %02X %08X:%08X %02X:%08lX %08X %5d %8d %u %d %p",
i,
- req->af.v4_req.loc_addr,
+ ireq->loc_addr,
ntohs(inet_sk(sk)->sport),
- req->af.v4_req.rmt_addr,
- ntohs(req->rmt_port),
+ ireq->rmt_addr,
+ ntohs(ireq->rmt_port),
TCP_SYN_RECV,
0, 0, /* could print option size, but that is af dependent. */
1, /* timers active (only the expire timer) */
@@ -2618,6 +2607,7 @@ struct proto tcp_prot = {
.sysctl_rmem = sysctl_tcp_rmem,
.max_header = MAX_TCP_HEADER,
.obj_size = sizeof(struct tcp_sock),
+ .rsk_prot = &tcp_request_sock_ops,
};
@@ -2660,7 +2650,6 @@ EXPORT_SYMBOL(tcp_proc_register);
EXPORT_SYMBOL(tcp_proc_unregister);
#endif
EXPORT_SYMBOL(sysctl_local_port_range);
-EXPORT_SYMBOL(sysctl_max_syn_backlog);
EXPORT_SYMBOL(sysctl_tcp_low_latency);
EXPORT_SYMBOL(sysctl_tcp_tw_reuse);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index eea1a17a9ac..f42a284164b 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -684,7 +684,7 @@ out:
* Actually, we could lots of memory writes here. tp of listening
* socket contains all necessary default parameters.
*/
-struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req, struct sk_buff *skb)
+struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req, struct sk_buff *skb)
{
/* allocate the newsk from the same slab of the master sock,
* if not, at sk_free time we'll try to free it from the wrong
@@ -692,6 +692,8 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
struct sock *newsk = sk_alloc(PF_INET, GFP_ATOMIC, sk->sk_prot, 0);
if(newsk != NULL) {
+ struct inet_request_sock *ireq = inet_rsk(req);
+ struct tcp_request_sock *treq = tcp_rsk(req);
struct tcp_sock *newtp;
struct sk_filter *filter;
@@ -703,7 +705,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
tcp_sk(newsk)->bind_hash = NULL;
/* Clone the TCP header template */
- inet_sk(newsk)->dport = req->rmt_port;
+ inet_sk(newsk)->dport = ireq->rmt_port;
sock_lock_init(newsk);
bh_lock_sock(newsk);
@@ -739,14 +741,14 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
/* Now setup tcp_sock */
newtp = tcp_sk(newsk);
newtp->pred_flags = 0;
- newtp->rcv_nxt = req->rcv_isn + 1;
- newtp->snd_nxt = req->snt_isn + 1;
- newtp->snd_una = req->snt_isn + 1;
- newtp->snd_sml = req->snt_isn + 1;
+ newtp->rcv_nxt = treq->rcv_isn + 1;
+ newtp->snd_nxt = treq->snt_isn + 1;
+ newtp->snd_una = treq->snt_isn + 1;
+ newtp->snd_sml = treq->snt_isn + 1;
tcp_prequeue_init(newtp);
- tcp_init_wl(newtp, req->snt_isn, req->rcv_isn);
+ tcp_init_wl(newtp, treq->snt_isn, treq->rcv_isn);
newtp->retransmits = 0;
newtp->backoff = 0;
@@ -772,13 +774,15 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
newtp->frto_counter = 0;
newtp->frto_highmark = 0;
+ newtp->ca_ops = &tcp_reno;
+
tcp_set_ca_state(newtp, TCP_CA_Open);
tcp_init_xmit_timers(newsk);
skb_queue_head_init(&newtp->out_of_order_queue);
- newtp->rcv_wup = req->rcv_isn + 1;
- newtp->write_seq = req->snt_isn + 1;
+ newtp->rcv_wup = treq->rcv_isn + 1;
+ newtp->write_seq = treq->snt_isn + 1;
newtp->pushed_seq = newtp->write_seq;
- newtp->copied_seq = req->rcv_isn + 1;
+ newtp->copied_seq = treq->rcv_isn + 1;
newtp->rx_opt.saw_tstamp = 0;
@@ -788,10 +792,8 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
newtp->probes_out = 0;
newtp->rx_opt.num_sacks = 0;
newtp->urg_data = 0;
- newtp->listen_opt = NULL;
- newtp->accept_queue = newtp->accept_queue_tail = NULL;
- /* Deinitialize syn_wait_lock to trap illegal accesses. */
- memset(&newtp->syn_wait_lock, 0, sizeof(newtp->syn_wait_lock));
+ /* Deinitialize accept_queue to trap illegal accesses. */
+ memset(&newtp->accept_queue, 0, sizeof(newtp->accept_queue));
/* Back to base struct sock members. */
newsk->sk_err = 0;
@@ -808,18 +810,18 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
newsk->sk_socket = NULL;
newsk->sk_sleep = NULL;
- newtp->rx_opt.tstamp_ok = req->tstamp_ok;
- if((newtp->rx_opt.sack_ok = req->sack_ok) != 0) {
+ newtp->rx_opt.tstamp_ok = ireq->tstamp_ok;
+ if((newtp->rx_opt.sack_ok = ireq->sack_ok) != 0) {
if (sysctl_tcp_fack)
newtp->rx_opt.sack_ok |= 2;
}
newtp->window_clamp = req->window_clamp;
newtp->rcv_ssthresh = req->rcv_wnd;
newtp->rcv_wnd = req->rcv_wnd;
- newtp->rx_opt.wscale_ok = req->wscale_ok;
+ newtp->rx_opt.wscale_ok = ireq->wscale_ok;
if (newtp->rx_opt.wscale_ok) {
- newtp->rx_opt.snd_wscale = req->snd_wscale;
- newtp->rx_opt.rcv_wscale = req->rcv_wscale;
+ newtp->rx_opt.snd_wscale = ireq->snd_wscale;
+ newtp->rx_opt.rcv_wscale = ireq->rcv_wscale;
} else {
newtp->rx_opt.snd_wscale = newtp->rx_opt.rcv_wscale = 0;
newtp->window_clamp = min(newtp->window_clamp, 65535U);
@@ -842,8 +844,6 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
if (newtp->ecn_flags&TCP_ECN_OK)
sock_set_flag(newsk, SOCK_NO_LARGESEND);
- tcp_ca_init(newtp);
-
TCP_INC_STATS_BH(TCP_MIB_PASSIVEOPENS);
}
return newsk;
@@ -851,12 +851,12 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct open_request *req,
/*
* Process an incoming packet for SYN_RECV sockets represented
- * as an open_request.
+ * as a request_sock.
*/
struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
- struct open_request *req,
- struct open_request **prev)
+ struct request_sock *req,
+ struct request_sock **prev)
{
struct tcphdr *th = skb->h.th;
struct tcp_sock *tp = tcp_sk(sk);
@@ -881,7 +881,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
}
/* Check for pure retransmitted SYN. */
- if (TCP_SKB_CB(skb)->seq == req->rcv_isn &&
+ if (TCP_SKB_CB(skb)->seq == tcp_rsk(req)->rcv_isn &&
flg == TCP_FLAG_SYN &&
!paws_reject) {
/*
@@ -901,7 +901,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
* Enforce "SYN-ACK" according to figure 8, figure 6
* of RFC793, fixed by RFC1122.
*/
- req->class->rtx_syn_ack(sk, req, NULL);
+ req->rsk_ops->rtx_syn_ack(sk, req, NULL);
return NULL;
}
@@ -959,7 +959,7 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
* Invalid ACK: reset will be sent by listening socket
*/
if ((flg & TCP_FLAG_ACK) &&
- (TCP_SKB_CB(skb)->ack_seq != req->snt_isn+1))
+ (TCP_SKB_CB(skb)->ack_seq != tcp_rsk(req)->snt_isn + 1))
return sk;
/* Also, it would be not so bad idea to check rcv_tsecr, which
@@ -970,10 +970,10 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
/* RFC793: "first check sequence number". */
if (paws_reject || !tcp_in_window(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq,
- req->rcv_isn+1, req->rcv_isn+1+req->rcv_wnd)) {
+ tcp_rsk(req)->rcv_isn + 1, tcp_rsk(req)->rcv_isn + 1 + req->rcv_wnd)) {
/* Out of window: send ACK and drop. */
if (!(flg & TCP_FLAG_RST))
- req->class->send_ack(skb, req);
+ req->rsk_ops->send_ack(skb, req);
if (paws_reject)
NET_INC_STATS_BH(LINUX_MIB_PAWSESTABREJECTED);
return NULL;
@@ -981,12 +981,12 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
/* In sequence, PAWS is OK. */
- if (tmp_opt.saw_tstamp && !after(TCP_SKB_CB(skb)->seq, req->rcv_isn+1))
+ if (tmp_opt.saw_tstamp && !after(TCP_SKB_CB(skb)->seq, tcp_rsk(req)->rcv_isn + 1))
req->ts_recent = tmp_opt.rcv_tsval;
- if (TCP_SKB_CB(skb)->seq == req->rcv_isn) {
+ if (TCP_SKB_CB(skb)->seq == tcp_rsk(req)->rcv_isn) {
/* Truncate SYN, it is out of window starting
- at req->rcv_isn+1. */
+ at tcp_rsk(req)->rcv_isn + 1. */
flg &= ~TCP_FLAG_SYN;
}
@@ -1003,8 +1003,8 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
return NULL;
/* If TCP_DEFER_ACCEPT is set, drop bare ACK. */
- if (tp->defer_accept && TCP_SKB_CB(skb)->end_seq == req->rcv_isn+1) {
- req->acked = 1;
+ if (tp->defer_accept && TCP_SKB_CB(skb)->end_seq == tcp_rsk(req)->rcv_isn + 1) {
+ inet_rsk(req)->acked = 1;
return NULL;
}
@@ -1026,14 +1026,14 @@ struct sock *tcp_check_req(struct sock *sk,struct sk_buff *skb,
listen_overflow:
if (!sysctl_tcp_abort_on_overflow) {
- req->acked = 1;
+ inet_rsk(req)->acked = 1;
return NULL;
}
embryonic_reset:
NET_INC_STATS_BH(LINUX_MIB_EMBRYONICRSTS);
if (!(flg & TCP_FLAG_RST))
- req->class->send_reset(skb);
+ req->rsk_ops->send_reset(skb);
tcp_synq_drop(sk, req, prev);
return NULL;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index fa24e7ae1f4..e3f8ea1bfa9 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -49,7 +49,7 @@ int sysctl_tcp_retrans_collapse = 1;
* will allow a single TSO frame to consume. Building TSO frames
* which are too large can cause TCP streams to be bursty.
*/
-int sysctl_tcp_tso_win_divisor = 8;
+int sysctl_tcp_tso_win_divisor = 3;
static inline void update_send_head(struct sock *sk, struct tcp_sock *tp,
struct sk_buff *skb)
@@ -111,8 +111,7 @@ static void tcp_cwnd_restart(struct tcp_sock *tp, struct dst_entry *dst)
u32 restart_cwnd = tcp_init_cwnd(tp, dst);
u32 cwnd = tp->snd_cwnd;
- if (tcp_is_vegas(tp))
- tcp_vegas_enable(tp);
+ tcp_ca_event(tp, CA_EVENT_CWND_RESTART);
tp->snd_ssthresh = tcp_current_ssthresh(tp);
restart_cwnd = min(restart_cwnd, cwnd);
@@ -141,11 +140,11 @@ static inline void tcp_event_data_sent(struct tcp_sock *tp,
tp->ack.pingpong = 1;
}
-static __inline__ void tcp_event_ack_sent(struct sock *sk)
+static __inline__ void tcp_event_ack_sent(struct sock *sk, unsigned int pkts)
{
struct tcp_sock *tp = tcp_sk(sk);
- tcp_dec_quickack_mode(tp);
+ tcp_dec_quickack_mode(tp, pkts);
tcp_clear_xmit_timer(sk, TCP_TIME_DACK);
}
@@ -280,6 +279,10 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb)
#define SYSCTL_FLAG_WSCALE 0x2
#define SYSCTL_FLAG_SACK 0x4
+ /* If congestion control is doing timestamping */
+ if (tp->ca_ops->rtt_sample)
+ do_gettimeofday(&skb->stamp);
+
sysctl_flags = 0;
if (tcb->flags & TCPCB_FLAG_SYN) {
tcp_header_size = sizeof(struct tcphdr) + TCPOLEN_MSS;
@@ -304,17 +307,8 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb)
(tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK));
}
- /*
- * If the connection is idle and we are restarting,
- * then we don't want to do any Vegas calculations
- * until we get fresh RTT samples. So when we
- * restart, we reset our Vegas state to a clean
- * slate. After we get acks for this flight of
- * packets, _then_ we can make Vegas calculations
- * again.
- */
- if (tcp_is_vegas(tp) && tcp_packets_in_flight(tp) == 0)
- tcp_vegas_enable(tp);
+ if (tcp_packets_in_flight(tp) == 0)
+ tcp_ca_event(tp, CA_EVENT_TX_START);
th = (struct tcphdr *) skb_push(skb, tcp_header_size);
skb->h.th = th;
@@ -361,7 +355,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb)
tp->af_specific->send_check(sk, th, skb->len, skb);
if (tcb->flags & TCPCB_FLAG_ACK)
- tcp_event_ack_sent(sk);
+ tcp_event_ack_sent(sk, tcp_skb_pcount(skb));
if (skb->len != tcp_header_size)
tcp_event_data_sent(tp, skb, sk);
@@ -409,42 +403,11 @@ static void tcp_queue_skb(struct sock *sk, struct sk_buff *skb)
sk->sk_send_head = skb;
}
-static inline void tcp_tso_set_push(struct sk_buff *skb)
-{
- /* Force push to be on for any TSO frames to workaround
- * problems with busted implementations like Mac OS-X that
- * hold off socket receive wakeups until push is seen.
- */
- if (tcp_skb_pcount(skb) > 1)
- TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
-}
-
-/* Send _single_ skb sitting at the send head. This function requires
- * true push pending frames to setup probe timer etc.
- */
-void tcp_push_one(struct sock *sk, unsigned cur_mss)
+static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct sk_buff *skb = sk->sk_send_head;
- if (tcp_snd_test(sk, skb, cur_mss, TCP_NAGLE_PUSH)) {
- /* Send it out now. */
- TCP_SKB_CB(skb)->when = tcp_time_stamp;
- tcp_tso_set_push(skb);
- if (!tcp_transmit_skb(sk, skb_clone(skb, sk->sk_allocation))) {
- sk->sk_send_head = NULL;
- tp->snd_nxt = TCP_SKB_CB(skb)->end_seq;
- tcp_packets_out_inc(sk, tp, skb);
- return;
- }
- }
-}
-
-void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
-{
- struct tcp_sock *tp = tcp_sk(sk);
-
- if (skb->len <= tp->mss_cache_std ||
+ if (skb->len <= tp->mss_cache ||
!(sk->sk_route_caps & NETIF_F_TSO)) {
/* Avoid the costly divide in the normal
* non-TSO case.
@@ -454,10 +417,10 @@ void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb)
} else {
unsigned int factor;
- factor = skb->len + (tp->mss_cache_std - 1);
- factor /= tp->mss_cache_std;
+ factor = skb->len + (tp->mss_cache - 1);
+ factor /= tp->mss_cache;
skb_shinfo(skb)->tso_segs = factor;
- skb_shinfo(skb)->tso_size = tp->mss_cache_std;
+ skb_shinfo(skb)->tso_size = tp->mss_cache;
}
}
@@ -521,6 +484,7 @@ static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len)
* skbs, which it never sent before. --ANK
*/
TCP_SKB_CB(buff)->when = TCP_SKB_CB(skb)->when;
+ buff->stamp = skb->stamp;
if (TCP_SKB_CB(skb)->sacked & TCPCB_LOST) {
tp->lost_out -= tcp_skb_pcount(skb);
@@ -542,6 +506,7 @@ static int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len)
}
/* Link BUFF into the send queue. */
+ skb_header_release(buff);
__skb_append(skb, buff);
return 0;
@@ -662,7 +627,7 @@ unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu)
/* And store cached results */
tp->pmtu_cookie = pmtu;
- tp->mss_cache = tp->mss_cache_std = mss_now;
+ tp->mss_cache = mss_now;
return mss_now;
}
@@ -674,57 +639,316 @@ unsigned int tcp_sync_mss(struct sock *sk, u32 pmtu)
* cannot be large. However, taking into account rare use of URG, this
* is not a big flaw.
*/
-
-unsigned int tcp_current_mss(struct sock *sk, int large)
+unsigned int tcp_current_mss(struct sock *sk, int large_allowed)
{
struct tcp_sock *tp = tcp_sk(sk);
struct dst_entry *dst = __sk_dst_get(sk);
- unsigned int do_large, mss_now;
+ u32 mss_now;
+ u16 xmit_size_goal;
+ int doing_tso = 0;
+
+ mss_now = tp->mss_cache;
+
+ if (large_allowed &&
+ (sk->sk_route_caps & NETIF_F_TSO) &&
+ !tp->urg_mode)
+ doing_tso = 1;
- mss_now = tp->mss_cache_std;
if (dst) {
u32 mtu = dst_mtu(dst);
if (mtu != tp->pmtu_cookie)
mss_now = tcp_sync_mss(sk, mtu);
}
- do_large = (large &&
- (sk->sk_route_caps & NETIF_F_TSO) &&
- !tp->urg_mode);
+ if (tp->rx_opt.eff_sacks)
+ mss_now -= (TCPOLEN_SACK_BASE_ALIGNED +
+ (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK));
- if (do_large) {
- unsigned int large_mss, factor, limit;
+ xmit_size_goal = mss_now;
- large_mss = 65535 - tp->af_specific->net_header_len -
+ if (doing_tso) {
+ xmit_size_goal = 65535 -
+ tp->af_specific->net_header_len -
tp->ext_header_len - tp->tcp_header_len;
- if (tp->max_window && large_mss > (tp->max_window>>1))
- large_mss = max((tp->max_window>>1),
- 68U - tp->tcp_header_len);
+ if (tp->max_window &&
+ (xmit_size_goal > (tp->max_window >> 1)))
+ xmit_size_goal = max((tp->max_window >> 1),
+ 68U - tp->tcp_header_len);
- factor = large_mss / mss_now;
+ xmit_size_goal -= (xmit_size_goal % mss_now);
+ }
+ tp->xmit_size_goal = xmit_size_goal;
- /* Always keep large mss multiple of real mss, but
- * do not exceed 1/tso_win_divisor of the congestion window
- * so we can keep the ACK clock ticking and minimize
- * bursting.
- */
- limit = tp->snd_cwnd;
- if (sysctl_tcp_tso_win_divisor)
- limit /= sysctl_tcp_tso_win_divisor;
- limit = max(1U, limit);
- if (factor > limit)
- factor = limit;
+ return mss_now;
+}
+
+/* Congestion window validation. (RFC2861) */
+
+static inline void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp)
+{
+ __u32 packets_out = tp->packets_out;
- tp->mss_cache = mss_now * factor;
+ if (packets_out >= tp->snd_cwnd) {
+ /* Network is feed fully. */
+ tp->snd_cwnd_used = 0;
+ tp->snd_cwnd_stamp = tcp_time_stamp;
+ } else {
+ /* Network starves. */
+ if (tp->packets_out > tp->snd_cwnd_used)
+ tp->snd_cwnd_used = tp->packets_out;
- mss_now = tp->mss_cache;
+ if ((s32)(tcp_time_stamp - tp->snd_cwnd_stamp) >= tp->rto)
+ tcp_cwnd_application_limited(sk);
}
+}
- if (tp->rx_opt.eff_sacks)
- mss_now -= (TCPOLEN_SACK_BASE_ALIGNED +
- (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK));
- return mss_now;
+static unsigned int tcp_window_allows(struct tcp_sock *tp, struct sk_buff *skb, unsigned int mss_now, unsigned int cwnd)
+{
+ u32 window, cwnd_len;
+
+ window = (tp->snd_una + tp->snd_wnd - TCP_SKB_CB(skb)->seq);
+ cwnd_len = mss_now * cwnd;
+ return min(window, cwnd_len);
+}
+
+/* Can at least one segment of SKB be sent right now, according to the
+ * congestion window rules? If so, return how many segments are allowed.
+ */
+static inline unsigned int tcp_cwnd_test(struct tcp_sock *tp, struct sk_buff *skb)
+{
+ u32 in_flight, cwnd;
+
+ /* Don't be strict about the congestion window for the final FIN. */
+ if (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN)
+ return 1;
+
+ in_flight = tcp_packets_in_flight(tp);
+ cwnd = tp->snd_cwnd;
+ if (in_flight < cwnd)
+ return (cwnd - in_flight);
+
+ return 0;
+}
+
+/* This must be invoked the first time we consider transmitting
+ * SKB onto the wire.
+ */
+static inline int tcp_init_tso_segs(struct sock *sk, struct sk_buff *skb)
+{
+ int tso_segs = tcp_skb_pcount(skb);
+
+ if (!tso_segs) {
+ tcp_set_skb_tso_segs(sk, skb);
+ tso_segs = tcp_skb_pcount(skb);
+ }
+ return tso_segs;
+}
+
+static inline int tcp_minshall_check(const struct tcp_sock *tp)
+{
+ return after(tp->snd_sml,tp->snd_una) &&
+ !after(tp->snd_sml, tp->snd_nxt);
+}
+
+/* Return 0, if packet can be sent now without violation Nagle's rules:
+ * 1. It is full sized.
+ * 2. Or it contains FIN. (already checked by caller)
+ * 3. Or TCP_NODELAY was set.
+ * 4. Or TCP_CORK is not set, and all sent packets are ACKed.
+ * With Minshall's modification: all sent small packets are ACKed.
+ */
+
+static inline int tcp_nagle_check(const struct tcp_sock *tp,
+ const struct sk_buff *skb,
+ unsigned mss_now, int nonagle)
+{
+ return (skb->len < mss_now &&
+ ((nonagle&TCP_NAGLE_CORK) ||
+ (!nonagle &&
+ tp->packets_out &&
+ tcp_minshall_check(tp))));
+}
+
+/* Return non-zero if the Nagle test allows this packet to be
+ * sent now.
+ */
+static inline int tcp_nagle_test(struct tcp_sock *tp, struct sk_buff *skb,
+ unsigned int cur_mss, int nonagle)
+{
+ /* Nagle rule does not apply to frames, which sit in the middle of the
+ * write_queue (they have no chances to get new data).
+ *
+ * This is implemented in the callers, where they modify the 'nonagle'
+ * argument based upon the location of SKB in the send queue.
+ */
+ if (nonagle & TCP_NAGLE_PUSH)
+ return 1;
+
+ /* Don't use the nagle rule for urgent data (or for the final FIN). */
+ if (tp->urg_mode ||
+ (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN))
+ return 1;
+
+ if (!tcp_nagle_check(tp, skb, cur_mss, nonagle))
+ return 1;
+
+ return 0;
+}
+
+/* Does at least the first segment of SKB fit into the send window? */
+static inline int tcp_snd_wnd_test(struct tcp_sock *tp, struct sk_buff *skb, unsigned int cur_mss)
+{
+ u32 end_seq = TCP_SKB_CB(skb)->end_seq;
+
+ if (skb->len > cur_mss)
+ end_seq = TCP_SKB_CB(skb)->seq + cur_mss;
+
+ return !after(end_seq, tp->snd_una + tp->snd_wnd);
+}
+
+/* This checks if the data bearing packet SKB (usually sk->sk_send_head)
+ * should be put on the wire right now. If so, it returns the number of
+ * packets allowed by the congestion window.
+ */
+static unsigned int tcp_snd_test(struct sock *sk, struct sk_buff *skb,
+ unsigned int cur_mss, int nonagle)
+{
+ struct tcp_sock *tp = tcp_sk(sk);
+ unsigned int cwnd_quota;
+
+ tcp_init_tso_segs(sk, skb);
+
+ if (!tcp_nagle_test(tp, skb, cur_mss, nonagle))
+ return 0;
+
+ cwnd_quota = tcp_cwnd_test(tp, skb);
+ if (cwnd_quota &&
+ !tcp_snd_wnd_test(tp, skb, cur_mss))
+ cwnd_quota = 0;
+
+ return cwnd_quota;
+}
+
+static inline int tcp_skb_is_last(const struct sock *sk,
+ const struct sk_buff *skb)
+{
+ return skb->next == (struct sk_buff *)&sk->sk_write_queue;
+}
+
+int tcp_may_send_now(struct sock *sk, struct tcp_sock *tp)
+{
+ struct sk_buff *skb = sk->sk_send_head;
+
+ return (skb &&
+ tcp_snd_test(sk, skb, tcp_current_mss(sk, 1),
+ (tcp_skb_is_last(sk, skb) ?
+ TCP_NAGLE_PUSH :
+ tp->nonagle)));
+}
+
+/* Trim TSO SKB to LEN bytes, put the remaining data into a new packet
+ * which is put after SKB on the list. It is very much like
+ * tcp_fragment() except that it may make several kinds of assumptions
+ * in order to speed up the splitting operation. In particular, we
+ * know that all the data is in scatter-gather pages, and that the
+ * packet has never been sent out before (and thus is not cloned).
+ */
+static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len)
+{
+ struct sk_buff *buff;
+ int nlen = skb->len - len;
+ u16 flags;
+
+ /* All of a TSO frame must be composed of paged data. */
+ BUG_ON(skb->len != skb->data_len);
+
+ buff = sk_stream_alloc_pskb(sk, 0, 0, GFP_ATOMIC);
+ if (unlikely(buff == NULL))
+ return -ENOMEM;
+
+ buff->truesize = nlen;
+ skb->truesize -= nlen;
+
+ /* Correct the sequence numbers. */
+ TCP_SKB_CB(buff)->seq = TCP_SKB_CB(skb)->seq + len;
+ TCP_SKB_CB(buff)->end_seq = TCP_SKB_CB(skb)->end_seq;
+ TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(buff)->seq;
+
+ /* PSH and FIN should only be set in the second packet. */
+ flags = TCP_SKB_CB(skb)->flags;
+ TCP_SKB_CB(skb)->flags = flags & ~(TCPCB_FLAG_FIN|TCPCB_FLAG_PSH);
+ TCP_SKB_CB(buff)->flags = flags;
+
+ /* This packet was never sent out yet, so no SACK bits. */
+ TCP_SKB_CB(buff)->sacked = 0;
+
+ buff->ip_summed = skb->ip_summed = CHECKSUM_HW;
+ skb_split(skb, buff, len);
+
+ /* Fix up tso_factor for both original and new SKB. */
+ tcp_set_skb_tso_segs(sk, skb);
+ tcp_set_skb_tso_segs(sk, buff);
+
+ /* Link BUFF into the send queue. */
+ skb_header_release(buff);
+ __skb_append(skb, buff);
+
+ return 0;
+}
+
+/* Try to defer sending, if possible, in order to minimize the amount
+ * of TSO splitting we do. View it as a kind of TSO Nagle test.
+ *
+ * This algorithm is from John Heffner.
+ */
+static int tcp_tso_should_defer(struct sock *sk, struct tcp_sock *tp, struct sk_buff *skb)
+{
+ u32 send_win, cong_win, limit, in_flight;
+
+ if (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN)
+ return 0;
+
+ if (tp->ca_state != TCP_CA_Open)
+ return 0;
+
+ in_flight = tcp_packets_in_flight(tp);
+
+ BUG_ON(tcp_skb_pcount(skb) <= 1 ||
+ (tp->snd_cwnd <= in_flight));
+
+ send_win = (tp->snd_una + tp->snd_wnd) - TCP_SKB_CB(skb)->seq;
+
+ /* From in_flight test above, we know that cwnd > in_flight. */
+ cong_win = (tp->snd_cwnd - in_flight) * tp->mss_cache;
+
+ limit = min(send_win, cong_win);
+
+ /* If sk_send_head can be sent fully now, just do it. */
+ if (skb->len <= limit)
+ return 0;
+
+ if (sysctl_tcp_tso_win_divisor) {
+ u32 chunk = min(tp->snd_wnd, tp->snd_cwnd * tp->mss_cache);
+
+ /* If at least some fraction of a window is available,
+ * just use it.
+ */
+ chunk /= sysctl_tcp_tso_win_divisor;
+ if (limit >= chunk)
+ return 0;
+ } else {
+ /* Different approach, try not to defer past a single
+ * ACK. Receiver should ACK every other full sized
+ * frame, so if we have space for more than 3 frames
+ * then send now.
+ */
+ if (limit > tcp_max_burst(tp) * tp->mss_cache)
+ return 0;
+ }
+
+ /* Ok, it looks like it is advisable to defer. */
+ return 1;
}
/* This routine writes packets to the network. It advances the
@@ -734,57 +958,158 @@ unsigned int tcp_current_mss(struct sock *sk, int large)
* Returns 1, if no segments are in flight and we have queued segments, but
* cannot send anything now because of SWS or another problem.
*/
-int tcp_write_xmit(struct sock *sk, int nonagle)
+static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle)
{
struct tcp_sock *tp = tcp_sk(sk);
- unsigned int mss_now;
+ struct sk_buff *skb;
+ unsigned int tso_segs, sent_pkts;
+ int cwnd_quota;
/* If we are closed, the bytes will have to remain here.
* In time closedown will finish, we empty the write queue and all
* will be happy.
*/
- if (sk->sk_state != TCP_CLOSE) {
- struct sk_buff *skb;
- int sent_pkts = 0;
+ if (unlikely(sk->sk_state == TCP_CLOSE))
+ return 0;
+
+ skb = sk->sk_send_head;
+ if (unlikely(!skb))
+ return 0;
+
+ tso_segs = tcp_init_tso_segs(sk, skb);
+ cwnd_quota = tcp_cwnd_test(tp, skb);
+ if (unlikely(!cwnd_quota))
+ goto out;
+
+ sent_pkts = 0;
+ while (likely(tcp_snd_wnd_test(tp, skb, mss_now))) {
+ BUG_ON(!tso_segs);
+
+ if (tso_segs == 1) {
+ if (unlikely(!tcp_nagle_test(tp, skb, mss_now,
+ (tcp_skb_is_last(sk, skb) ?
+ nonagle : TCP_NAGLE_PUSH))))
+ break;
+ } else {
+ if (tcp_tso_should_defer(sk, tp, skb))
+ break;
+ }
- /* Account for SACKS, we may need to fragment due to this.
- * It is just like the real MSS changing on us midstream.
- * We also handle things correctly when the user adds some
- * IP options mid-stream. Silly to do, but cover it.
- */
- mss_now = tcp_current_mss(sk, 1);
-
- while ((skb = sk->sk_send_head) &&
- tcp_snd_test(sk, skb, mss_now,
- tcp_skb_is_last(sk, skb) ? nonagle :
- TCP_NAGLE_PUSH)) {
- if (skb->len > mss_now) {
- if (tcp_fragment(sk, skb, mss_now))
+ if (tso_segs > 1) {
+ u32 limit = tcp_window_allows(tp, skb,
+ mss_now, cwnd_quota);
+
+ if (skb->len < limit) {
+ unsigned int trim = skb->len % mss_now;
+
+ if (trim)
+ limit = skb->len - trim;
+ }
+ if (skb->len > limit) {
+ if (tso_fragment(sk, skb, limit))
break;
}
-
- TCP_SKB_CB(skb)->when = tcp_time_stamp;
- tcp_tso_set_push(skb);
- if (tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC)))
+ } else if (unlikely(skb->len > mss_now)) {
+ if (unlikely(tcp_fragment(sk, skb, mss_now)))
break;
+ }
- /* Advance the send_head. This one is sent out.
- * This call will increment packets_out.
- */
- update_send_head(sk, tp, skb);
+ TCP_SKB_CB(skb)->when = tcp_time_stamp;
+
+ if (unlikely(tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC))))
+ break;
+
+ /* Advance the send_head. This one is sent out.
+ * This call will increment packets_out.
+ */
+ update_send_head(sk, tp, skb);
+
+ tcp_minshall_update(tp, mss_now, skb);
+ sent_pkts++;
+
+ /* Do not optimize this to use tso_segs. If we chopped up
+ * the packet above, tso_segs will no longer be valid.
+ */
+ cwnd_quota -= tcp_skb_pcount(skb);
+
+ BUG_ON(cwnd_quota < 0);
+ if (!cwnd_quota)
+ break;
+
+ skb = sk->sk_send_head;
+ if (!skb)
+ break;
+ tso_segs = tcp_init_tso_segs(sk, skb);
+ }
+
+ if (likely(sent_pkts)) {
+ tcp_cwnd_validate(sk, tp);
+ return 0;
+ }
+out:
+ return !tp->packets_out && sk->sk_send_head;
+}
+
+/* Push out any pending frames which were held back due to
+ * TCP_CORK or attempt at coalescing tiny packets.
+ * The socket must be locked by the caller.
+ */
+void __tcp_push_pending_frames(struct sock *sk, struct tcp_sock *tp,
+ unsigned int cur_mss, int nonagle)
+{
+ struct sk_buff *skb = sk->sk_send_head;
+
+ if (skb) {
+ if (tcp_write_xmit(sk, cur_mss, nonagle))
+ tcp_check_probe_timer(sk, tp);
+ }
+}
+
+/* Send _single_ skb sitting at the send head. This function requires
+ * true push pending frames to setup probe timer etc.
+ */
+void tcp_push_one(struct sock *sk, unsigned int mss_now)
+{
+ struct tcp_sock *tp = tcp_sk(sk);
+ struct sk_buff *skb = sk->sk_send_head;
+ unsigned int tso_segs, cwnd_quota;
+
+ BUG_ON(!skb || skb->len < mss_now);
+
+ tso_segs = tcp_init_tso_segs(sk, skb);
+ cwnd_quota = tcp_snd_test(sk, skb, mss_now, TCP_NAGLE_PUSH);
+
+ if (likely(cwnd_quota)) {
+ BUG_ON(!tso_segs);
- tcp_minshall_update(tp, mss_now, skb);
- sent_pkts = 1;
+ if (tso_segs > 1) {
+ u32 limit = tcp_window_allows(tp, skb,
+ mss_now, cwnd_quota);
+
+ if (skb->len < limit) {
+ unsigned int trim = skb->len % mss_now;
+
+ if (trim)
+ limit = skb->len - trim;
+ }
+ if (skb->len > limit) {
+ if (unlikely(tso_fragment(sk, skb, limit)))
+ return;
+ }
+ } else if (unlikely(skb->len > mss_now)) {
+ if (unlikely(tcp_fragment(sk, skb, mss_now)))
+ return;
}
- if (sent_pkts) {
+ /* Send it out now. */
+ TCP_SKB_CB(skb)->when = tcp_time_stamp;
+
+ if (likely(!tcp_transmit_skb(sk, skb_clone(skb, sk->sk_allocation)))) {
+ update_send_head(sk, tp, skb);
tcp_cwnd_validate(sk, tp);
- return 0;
+ return;
}
-
- return !tp->packets_out && sk->sk_send_head;
}
- return 0;
}
/* This function returns the amount that we can raise the
@@ -1044,7 +1369,6 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
if (sk->sk_route_caps & NETIF_F_TSO) {
sk->sk_route_caps &= ~NETIF_F_TSO;
sock_set_flag(sk, SOCK_NO_LARGESEND);
- tp->mss_cache = tp->mss_cache_std;
}
if (tcp_trim_head(sk, skb, tp->snd_una - TCP_SKB_CB(skb)->seq))
@@ -1106,7 +1430,6 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
* is still in somebody's hands, else make a clone.
*/
TCP_SKB_CB(skb)->when = tcp_time_stamp;
- tcp_tso_set_push(skb);
err = tcp_transmit_skb(sk, (skb_cloned(skb) ?
pskb_copy(skb, GFP_ATOMIC):
@@ -1290,7 +1613,7 @@ void tcp_send_fin(struct sock *sk)
* was unread data in the receive queue. This behavior is recommended
* by draft-ietf-tcpimpl-prob-03.txt section 3.10. -DaveM
*/
-void tcp_send_active_reset(struct sock *sk, int priority)
+void tcp_send_active_reset(struct sock *sk, unsigned int __nocast priority)
{
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
@@ -1356,8 +1679,9 @@ int tcp_send_synack(struct sock *sk)
* Prepare a SYN-ACK.
*/
struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
- struct open_request *req)
+ struct request_sock *req)
{
+ struct inet_request_sock *ireq = inet_rsk(req);
struct tcp_sock *tp = tcp_sk(sk);
struct tcphdr *th;
int tcp_header_size;
@@ -1373,47 +1697,47 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
skb->dst = dst_clone(dst);
tcp_header_size = (sizeof(struct tcphdr) + TCPOLEN_MSS +
- (req->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0) +
- (req->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) +
+ (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0) +
+ (ireq->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) +
/* SACK_PERM is in the place of NOP NOP of TS */
- ((req->sack_ok && !req->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0));
+ ((ireq->sack_ok && !ireq->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0));
skb->h.th = th = (struct tcphdr *) skb_push(skb, tcp_header_size);
memset(th, 0, sizeof(struct tcphdr));
th->syn = 1;
th->ack = 1;
if (dst->dev->features&NETIF_F_TSO)
- req->ecn_ok = 0;
+ ireq->ecn_ok = 0;
TCP_ECN_make_synack(req, th);
th->source = inet_sk(sk)->sport;
- th->dest = req->rmt_port;
- TCP_SKB_CB(skb)->seq = req->snt_isn;
+ th->dest = ireq->rmt_port;
+ TCP_SKB_CB(skb)->seq = tcp_rsk(req)->snt_isn;
TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq + 1;
TCP_SKB_CB(skb)->sacked = 0;
skb_shinfo(skb)->tso_segs = 1;
skb_shinfo(skb)->tso_size = 0;
th->seq = htonl(TCP_SKB_CB(skb)->seq);
- th->ack_seq = htonl(req->rcv_isn + 1);
+ th->ack_seq = htonl(tcp_rsk(req)->rcv_isn + 1);
if (req->rcv_wnd == 0) { /* ignored for retransmitted syns */
__u8 rcv_wscale;
/* Set this up on the first call only */
req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW);
/* tcp_full_space because it is guaranteed to be the first packet */
tcp_select_initial_window(tcp_full_space(sk),
- dst_metric(dst, RTAX_ADVMSS) - (req->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0),
+ dst_metric(dst, RTAX_ADVMSS) - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0),
&req->rcv_wnd,
&req->window_clamp,
- req->wscale_ok,
+ ireq->wscale_ok,
&rcv_wscale);
- req->rcv_wscale = rcv_wscale;
+ ireq->rcv_wscale = rcv_wscale;
}
/* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */
th->window = htons(req->rcv_wnd);
TCP_SKB_CB(skb)->when = tcp_time_stamp;
- tcp_syn_build_options((__u32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), req->tstamp_ok,
- req->sack_ok, req->wscale_ok, req->rcv_wscale,
+ tcp_syn_build_options((__u32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok,
+ ireq->sack_ok, ireq->wscale_ok, ireq->rcv_wscale,
TCP_SKB_CB(skb)->when,
req->ts_recent);
@@ -1448,7 +1772,6 @@ static inline void tcp_connect_init(struct sock *sk)
tp->window_clamp = dst_metric(dst, RTAX_WINDOW);
tp->advmss = dst_metric(dst, RTAX_ADVMSS);
tcp_initialize_rcv_mss(sk);
- tcp_ca_init(tp);
tcp_select_initial_window(tcp_full_space(sk),
tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0),
@@ -1502,7 +1825,6 @@ int tcp_connect(struct sock *sk)
TCP_SKB_CB(buff)->end_seq = tp->write_seq;
tp->snd_nxt = tp->write_seq;
tp->pushed_seq = tp->write_seq;
- tcp_ca_init(tp);
/* Send it off. */
TCP_SKB_CB(buff)->when = tcp_time_stamp;
@@ -1676,14 +1998,12 @@ int tcp_write_wakeup(struct sock *sk)
if (sk->sk_route_caps & NETIF_F_TSO) {
sock_set_flag(sk, SOCK_NO_LARGESEND);
sk->sk_route_caps &= ~NETIF_F_TSO;
- tp->mss_cache = tp->mss_cache_std;
}
} else if (!tcp_skb_pcount(skb))
tcp_set_skb_tso_segs(sk, skb);
TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
TCP_SKB_CB(skb)->when = tcp_time_stamp;
- tcp_tso_set_push(skb);
err = tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC));
if (!err) {
update_send_head(sk, tp, skb);
diff --git a/net/ipv4/tcp_scalable.c b/net/ipv4/tcp_scalable.c
new file mode 100644
index 00000000000..70e108e15c7
--- /dev/null
+++ b/net/ipv4/tcp_scalable.c
@@ -0,0 +1,68 @@
+/* Tom Kelly's Scalable TCP
+ *
+ * See htt://www-lce.eng.cam.ac.uk/~ctk21/scalable/
+ *
+ * John Heffner <jheffner@sc.edu>
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <net/tcp.h>
+
+/* These factors derived from the recommended values in the aer:
+ * .01 and and 7/8. We use 50 instead of 100 to account for
+ * delayed ack.
+ */
+#define TCP_SCALABLE_AI_CNT 50U
+#define TCP_SCALABLE_MD_SCALE 3
+
+static void tcp_scalable_cong_avoid(struct tcp_sock *tp, u32 ack, u32 rtt,
+ u32 in_flight, int flag)
+{
+ if (in_flight < tp->snd_cwnd)
+ return;
+
+ if (tp->snd_cwnd <= tp->snd_ssthresh) {
+ tp->snd_cwnd++;
+ } else {
+ tp->snd_cwnd_cnt++;
+ if (tp->snd_cwnd_cnt > min(tp->snd_cwnd, TCP_SCALABLE_AI_CNT)){
+ tp->snd_cwnd++;
+ tp->snd_cwnd_cnt = 0;
+ }
+ }
+ tp->snd_cwnd = min_t(u32, tp->snd_cwnd, tp->snd_cwnd_clamp);
+ tp->snd_cwnd_stamp = tcp_time_stamp;
+}
+
+static u32 tcp_scalable_ssthresh(struct tcp_sock *tp)
+{
+ return max(tp->snd_cwnd - (tp->snd_cwnd>>TCP_SCALABLE_MD_SCALE), 2U);
+}
+
+
+static struct tcp_congestion_ops tcp_scalable = {
+ .ssthresh = tcp_scalable_ssthresh,
+ .cong_avoid = tcp_scalable_cong_avoid,
+ .min_cwnd = tcp_reno_min_cwnd,
+
+ .owner = THIS_MODULE,
+ .name = "scalable",
+};
+
+static int __init tcp_scalable_register(void)
+{
+ return tcp_register_congestion_control(&tcp_scalable);
+}
+
+static void __exit tcp_scalable_unregister(void)
+{
+ tcp_unregister_congestion_control(&tcp_scalable);
+}
+
+module_init(tcp_scalable_register);
+module_exit(tcp_scalable_unregister);
+
+MODULE_AUTHOR("John Heffner");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Scalable TCP");
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 799ebe061e2..0084227438c 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -231,11 +231,10 @@ static void tcp_delack_timer(unsigned long data)
}
tp->ack.pending &= ~TCP_ACK_TIMER;
- if (skb_queue_len(&tp->ucopy.prequeue)) {
+ if (!skb_queue_empty(&tp->ucopy.prequeue)) {
struct sk_buff *skb;
- NET_ADD_STATS_BH(LINUX_MIB_TCPSCHEDULERFAILED,
- skb_queue_len(&tp->ucopy.prequeue));
+ NET_INC_STATS_BH(LINUX_MIB_TCPSCHEDULERFAILED);
while ((skb = __skb_dequeue(&tp->ucopy.prequeue)) != NULL)
sk->sk_backlog_rcv(sk, skb);
@@ -464,11 +463,11 @@ out_unlock:
static void tcp_synack_timer(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- struct tcp_listen_opt *lopt = tp->listen_opt;
+ struct listen_sock *lopt = tp->accept_queue.listen_opt;
int max_retries = tp->syn_retries ? : sysctl_tcp_synack_retries;
int thresh = max_retries;
unsigned long now = jiffies;
- struct open_request **reqp, *req;
+ struct request_sock **reqp, *req;
int i, budget;
if (lopt == NULL || lopt->qlen == 0)
@@ -513,8 +512,8 @@ static void tcp_synack_timer(struct sock *sk)
while ((req = *reqp) != NULL) {
if (time_after_eq(now, req->expires)) {
if ((req->retrans < thresh ||
- (req->acked && req->retrans < max_retries))
- && !req->class->rtx_syn_ack(sk, req, NULL)) {
+ (inet_rsk(req)->acked && req->retrans < max_retries))
+ && !req->rsk_ops->rtx_syn_ack(sk, req, NULL)) {
unsigned long timeo;
if (req->retrans++ == 0)
@@ -527,13 +526,9 @@ static void tcp_synack_timer(struct sock *sk)
}
/* Drop this request */
- write_lock(&tp->syn_wait_lock);
- *reqp = req->dl_next;
- write_unlock(&tp->syn_wait_lock);
- lopt->qlen--;
- if (req->retrans == 0)
- lopt->qlen_young--;
- tcp_openreq_free(req);
+ tcp_synq_unlink(tp, req, reqp);
+ reqsk_queue_removed(&tp->accept_queue, req);
+ reqsk_free(req);
continue;
}
reqp = &req->dl_next;
diff --git a/net/ipv4/tcp_vegas.c b/net/ipv4/tcp_vegas.c
new file mode 100644
index 00000000000..9bd443db519
--- /dev/null
+++ b/net/ipv4/tcp_vegas.c
@@ -0,0 +1,411 @@
+/*
+ * TCP Vegas congestion control
+ *
+ * This is based on the congestion detection/avoidance scheme described in
+ * Lawrence S. Brakmo and Larry L. Peterson.
+ * "TCP Vegas: End to end congestion avoidance on a global internet."
+ * IEEE Journal on Selected Areas in Communication, 13(8):1465--1480,
+ * October 1995. Available from:
+ * ftp://ftp.cs.arizona.edu/xkernel/Papers/jsac.ps
+ *
+ * See http://www.cs.arizona.edu/xkernel/ for their implementation.
+ * The main aspects that distinguish this implementation from the
+ * Arizona Vegas implementation are:
+ * o We do not change the loss detection or recovery mechanisms of
+ * Linux in any way. Linux already recovers from losses quite well,
+ * using fine-grained timers, NewReno, and FACK.
+ * o To avoid the performance penalty imposed by increasing cwnd
+ * only every-other RTT during slow start, we increase during
+ * every RTT during slow start, just like Reno.
+ * o Largely to allow continuous cwnd growth during slow start,
+ * we use the rate at which ACKs come back as the "actual"
+ * rate, rather than the rate at which data is sent.
+ * o To speed convergence to the right rate, we set the cwnd
+ * to achieve the right ("actual") rate when we exit slow start.
+ * o To filter out the noise caused by delayed ACKs, we use the
+ * minimum RTT sample observed during the last RTT to calculate
+ * the actual rate.
+ * o When the sender re-starts from idle, it waits until it has
+ * received ACKs for an entire flight of new data before making
+ * a cwnd adjustment decision. The original Vegas implementation
+ * assumed senders never went idle.
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/tcp_diag.h>
+
+#include <net/tcp.h>
+
+/* Default values of the Vegas variables, in fixed-point representation
+ * with V_PARAM_SHIFT bits to the right of the binary point.
+ */
+#define V_PARAM_SHIFT 1
+static int alpha = 1<<V_PARAM_SHIFT;
+static int beta = 3<<V_PARAM_SHIFT;
+static int gamma = 1<<V_PARAM_SHIFT;
+
+module_param(alpha, int, 0644);
+MODULE_PARM_DESC(alpha, "lower bound of packets in network (scale by 2)");
+module_param(beta, int, 0644);
+MODULE_PARM_DESC(beta, "upper bound of packets in network (scale by 2)");
+module_param(gamma, int, 0644);
+MODULE_PARM_DESC(gamma, "limit on increase (scale by 2)");
+
+
+/* Vegas variables */
+struct vegas {
+ u32 beg_snd_nxt; /* right edge during last RTT */
+ u32 beg_snd_una; /* left edge during last RTT */
+ u32 beg_snd_cwnd; /* saves the size of the cwnd */
+ u8 doing_vegas_now;/* if true, do vegas for this RTT */
+ u16 cntRTT; /* # of RTTs measured within last RTT */
+ u32 minRTT; /* min of RTTs measured within last RTT (in usec) */
+ u32 baseRTT; /* the min of all Vegas RTT measurements seen (in usec) */
+};
+
+/* There are several situations when we must "re-start" Vegas:
+ *
+ * o when a connection is established
+ * o after an RTO
+ * o after fast recovery
+ * o when we send a packet and there is no outstanding
+ * unacknowledged data (restarting an idle connection)
+ *
+ * In these circumstances we cannot do a Vegas calculation at the
+ * end of the first RTT, because any calculation we do is using
+ * stale info -- both the saved cwnd and congestion feedback are
+ * stale.
+ *
+ * Instead we must wait until the completion of an RTT during
+ * which we actually receive ACKs.
+ */
+static inline void vegas_enable(struct tcp_sock *tp)
+{
+ struct vegas *vegas = tcp_ca(tp);
+
+ /* Begin taking Vegas samples next time we send something. */
+ vegas->doing_vegas_now = 1;
+
+ /* Set the beginning of the next send window. */
+ vegas->beg_snd_nxt = tp->snd_nxt;
+
+ vegas->cntRTT = 0;
+ vegas->minRTT = 0x7fffffff;
+}
+
+/* Stop taking Vegas samples for now. */
+static inline void vegas_disable(struct tcp_sock *tp)
+{
+ struct vegas *vegas = tcp_ca(tp);
+
+ vegas->doing_vegas_now = 0;
+}
+
+static void tcp_vegas_init(struct tcp_sock *tp)
+{
+ struct vegas *vegas = tcp_ca(tp);
+
+ vegas->baseRTT = 0x7fffffff;
+ vegas_enable(tp);
+}
+
+/* Do RTT sampling needed for Vegas.
+ * Basically we:
+ * o min-filter RTT samples from within an RTT to get the current
+ * propagation delay + queuing delay (we are min-filtering to try to
+ * avoid the effects of delayed ACKs)
+ * o min-filter RTT samples from a much longer window (forever for now)
+ * to find the propagation delay (baseRTT)
+ */
+static void tcp_vegas_rtt_calc(struct tcp_sock *tp, u32 usrtt)
+{
+ struct vegas *vegas = tcp_ca(tp);
+ u32 vrtt = usrtt + 1; /* Never allow zero rtt or baseRTT */
+
+ /* Filter to find propagation delay: */
+ if (vrtt < vegas->baseRTT)
+ vegas->baseRTT = vrtt;
+
+ /* Find the min RTT during the last RTT to find
+ * the current prop. delay + queuing delay:
+ */
+ vegas->minRTT = min(vegas->minRTT, vrtt);
+ vegas->cntRTT++;
+}
+
+static void tcp_vegas_state(struct tcp_sock *tp, u8 ca_state)
+{
+
+ if (ca_state == TCP_CA_Open)
+ vegas_enable(tp);
+ else
+ vegas_disable(tp);
+}
+
+/*
+ * If the connection is idle and we are restarting,
+ * then we don't want to do any Vegas calculations
+ * until we get fresh RTT samples. So when we
+ * restart, we reset our Vegas state to a clean
+ * slate. After we get acks for this flight of
+ * packets, _then_ we can make Vegas calculations
+ * again.
+ */
+static void tcp_vegas_cwnd_event(struct tcp_sock *tp, enum tcp_ca_event event)
+{
+ if (event == CA_EVENT_CWND_RESTART ||
+ event == CA_EVENT_TX_START)
+ tcp_vegas_init(tp);
+}
+
+static void tcp_vegas_cong_avoid(struct tcp_sock *tp, u32 ack,
+ u32 seq_rtt, u32 in_flight, int flag)
+{
+ struct vegas *vegas = tcp_ca(tp);
+
+ if (!vegas->doing_vegas_now)
+ return tcp_reno_cong_avoid(tp, ack, seq_rtt, in_flight, flag);
+
+ /* The key players are v_beg_snd_una and v_beg_snd_nxt.
+ *
+ * These are so named because they represent the approximate values
+ * of snd_una and snd_nxt at the beginning of the current RTT. More
+ * precisely, they represent the amount of data sent during the RTT.
+ * At the end of the RTT, when we receive an ACK for v_beg_snd_nxt,
+ * we will calculate that (v_beg_snd_nxt - v_beg_snd_una) outstanding
+ * bytes of data have been ACKed during the course of the RTT, giving
+ * an "actual" rate of:
+ *
+ * (v_beg_snd_nxt - v_beg_snd_una) / (rtt duration)
+ *
+ * Unfortunately, v_beg_snd_una is not exactly equal to snd_una,
+ * because delayed ACKs can cover more than one segment, so they
+ * don't line up nicely with the boundaries of RTTs.
+ *
+ * Another unfortunate fact of life is that delayed ACKs delay the
+ * advance of the left edge of our send window, so that the number
+ * of bytes we send in an RTT is often less than our cwnd will allow.
+ * So we keep track of our cwnd separately, in v_beg_snd_cwnd.
+ */
+
+ if (after(ack, vegas->beg_snd_nxt)) {
+ /* Do the Vegas once-per-RTT cwnd adjustment. */
+ u32 old_wnd, old_snd_cwnd;
+
+
+ /* Here old_wnd is essentially the window of data that was
+ * sent during the previous RTT, and has all
+ * been acknowledged in the course of the RTT that ended
+ * with the ACK we just received. Likewise, old_snd_cwnd
+ * is the cwnd during the previous RTT.
+ */
+ old_wnd = (vegas->beg_snd_nxt - vegas->beg_snd_una) /
+ tp->mss_cache;
+ old_snd_cwnd = vegas->beg_snd_cwnd;
+
+ /* Save the extent of the current window so we can use this
+ * at the end of the next RTT.
+ */
+ vegas->beg_snd_una = vegas->beg_snd_nxt;
+ vegas->beg_snd_nxt = tp->snd_nxt;
+ vegas->beg_snd_cwnd = tp->snd_cwnd;
+
+ /* Take into account the current RTT sample too, to
+ * decrease the impact of delayed acks. This double counts
+ * this sample since we count it for the next window as well,
+ * but that's not too awful, since we're taking the min,
+ * rather than averaging.
+ */
+ tcp_vegas_rtt_calc(tp, seq_rtt*1000);
+
+ /* We do the Vegas calculations only if we got enough RTT
+ * samples that we can be reasonably sure that we got
+ * at least one RTT sample that wasn't from a delayed ACK.
+ * If we only had 2 samples total,
+ * then that means we're getting only 1 ACK per RTT, which
+ * means they're almost certainly delayed ACKs.
+ * If we have 3 samples, we should be OK.
+ */
+
+ if (vegas->cntRTT <= 2) {
+ /* We don't have enough RTT samples to do the Vegas
+ * calculation, so we'll behave like Reno.
+ */
+ if (tp->snd_cwnd > tp->snd_ssthresh)
+ tp->snd_cwnd++;
+ } else {
+ u32 rtt, target_cwnd, diff;
+
+ /* We have enough RTT samples, so, using the Vegas
+ * algorithm, we determine if we should increase or
+ * decrease cwnd, and by how much.
+ */
+
+ /* Pluck out the RTT we are using for the Vegas
+ * calculations. This is the min RTT seen during the
+ * last RTT. Taking the min filters out the effects
+ * of delayed ACKs, at the cost of noticing congestion
+ * a bit later.
+ */
+ rtt = vegas->minRTT;
+
+ /* Calculate the cwnd we should have, if we weren't
+ * going too fast.
+ *
+ * This is:
+ * (actual rate in segments) * baseRTT
+ * We keep it as a fixed point number with
+ * V_PARAM_SHIFT bits to the right of the binary point.
+ */
+ target_cwnd = ((old_wnd * vegas->baseRTT)
+ << V_PARAM_SHIFT) / rtt;
+
+ /* Calculate the difference between the window we had,
+ * and the window we would like to have. This quantity
+ * is the "Diff" from the Arizona Vegas papers.
+ *
+ * Again, this is a fixed point number with
+ * V_PARAM_SHIFT bits to the right of the binary
+ * point.
+ */
+ diff = (old_wnd << V_PARAM_SHIFT) - target_cwnd;
+
+ if (tp->snd_cwnd < tp->snd_ssthresh) {
+ /* Slow start. */
+ if (diff > gamma) {
+ /* Going too fast. Time to slow down
+ * and switch to congestion avoidance.
+ */
+ tp->snd_ssthresh = 2;
+
+ /* Set cwnd to match the actual rate
+ * exactly:
+ * cwnd = (actual rate) * baseRTT
+ * Then we add 1 because the integer
+ * truncation robs us of full link
+ * utilization.
+ */
+ tp->snd_cwnd = min(tp->snd_cwnd,
+ (target_cwnd >>
+ V_PARAM_SHIFT)+1);
+
+ }
+ } else {
+ /* Congestion avoidance. */
+ u32 next_snd_cwnd;
+
+ /* Figure out where we would like cwnd
+ * to be.
+ */
+ if (diff > beta) {
+ /* The old window was too fast, so
+ * we slow down.
+ */
+ next_snd_cwnd = old_snd_cwnd - 1;
+ } else if (diff < alpha) {
+ /* We don't have enough extra packets
+ * in the network, so speed up.
+ */
+ next_snd_cwnd = old_snd_cwnd + 1;
+ } else {
+ /* Sending just as fast as we
+ * should be.
+ */
+ next_snd_cwnd = old_snd_cwnd;
+ }
+
+ /* Adjust cwnd upward or downward, toward the
+ * desired value.
+ */
+ if (next_snd_cwnd > tp->snd_cwnd)
+ tp->snd_cwnd++;
+ else if (next_snd_cwnd < tp->snd_cwnd)
+ tp->snd_cwnd--;
+ }
+ }
+
+ /* Wipe the slate clean for the next RTT. */
+ vegas->cntRTT = 0;
+ vegas->minRTT = 0x7fffffff;
+ }
+
+ /* The following code is executed for every ack we receive,
+ * except for conditions checked in should_advance_cwnd()
+ * before the call to tcp_cong_avoid(). Mainly this means that
+ * we only execute this code if the ack actually acked some
+ * data.
+ */
+
+ /* If we are in slow start, increase our cwnd in response to this ACK.
+ * (If we are not in slow start then we are in congestion avoidance,
+ * and adjust our congestion window only once per RTT. See the code
+ * above.)
+ */
+ if (tp->snd_cwnd <= tp->snd_ssthresh)
+ tp->snd_cwnd++;
+
+ /* to keep cwnd from growing without bound */
+ tp->snd_cwnd = min_t(u32, tp->snd_cwnd, tp->snd_cwnd_clamp);
+
+ /* Make sure that we are never so timid as to reduce our cwnd below
+ * 2 MSS.
+ *
+ * Going below 2 MSS would risk huge delayed ACKs from our receiver.
+ */
+ tp->snd_cwnd = max(tp->snd_cwnd, 2U);
+}
+
+/* Extract info for Tcp socket info provided via netlink. */
+static void tcp_vegas_get_info(struct tcp_sock *tp, u32 ext,
+ struct sk_buff *skb)
+{
+ const struct vegas *ca = tcp_ca(tp);
+ if (ext & (1<<(TCPDIAG_VEGASINFO-1))) {
+ struct tcpvegas_info *info;
+
+ info = RTA_DATA(__RTA_PUT(skb, TCPDIAG_VEGASINFO,
+ sizeof(*info)));
+
+ info->tcpv_enabled = ca->doing_vegas_now;
+ info->tcpv_rttcnt = ca->cntRTT;
+ info->tcpv_rtt = ca->baseRTT;
+ info->tcpv_minrtt = ca->minRTT;
+ rtattr_failure: ;
+ }
+}
+
+static struct tcp_congestion_ops tcp_vegas = {
+ .init = tcp_vegas_init,
+ .ssthresh = tcp_reno_ssthresh,
+ .cong_avoid = tcp_vegas_cong_avoid,
+ .min_cwnd = tcp_reno_min_cwnd,
+ .rtt_sample = tcp_vegas_rtt_calc,
+ .set_state = tcp_vegas_state,
+ .cwnd_event = tcp_vegas_cwnd_event,
+ .get_info = tcp_vegas_get_info,
+
+ .owner = THIS_MODULE,
+ .name = "vegas",
+};
+
+static int __init tcp_vegas_register(void)
+{
+ BUG_ON(sizeof(struct vegas) > TCP_CA_PRIV_SIZE);
+ tcp_register_congestion_control(&tcp_vegas);
+ return 0;
+}
+
+static void __exit tcp_vegas_unregister(void)
+{
+ tcp_unregister_congestion_control(&tcp_vegas);
+}
+
+module_init(tcp_vegas_register);
+module_exit(tcp_vegas_unregister);
+
+MODULE_AUTHOR("Stephen Hemminger");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("TCP Vegas");
diff --git a/net/ipv4/tcp_westwood.c b/net/ipv4/tcp_westwood.c
new file mode 100644
index 00000000000..ef827242c94
--- /dev/null
+++ b/net/ipv4/tcp_westwood.c
@@ -0,0 +1,259 @@
+/*
+ * TCP Westwood+
+ *
+ * Angelo Dell'Aera: TCP Westwood+ support
+ */
+
+#include <linux/config.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/tcp_diag.h>
+#include <net/tcp.h>
+
+/* TCP Westwood structure */
+struct westwood {
+ u32 bw_ns_est; /* first bandwidth estimation..not too smoothed 8) */
+ u32 bw_est; /* bandwidth estimate */
+ u32 rtt_win_sx; /* here starts a new evaluation... */
+ u32 bk;
+ u32 snd_una; /* used for evaluating the number of acked bytes */
+ u32 cumul_ack;
+ u32 accounted;
+ u32 rtt;
+ u32 rtt_min; /* minimum observed RTT */
+};
+
+
+/* TCP Westwood functions and constants */
+#define TCP_WESTWOOD_RTT_MIN (HZ/20) /* 50ms */
+#define TCP_WESTWOOD_INIT_RTT (20*HZ) /* maybe too conservative?! */
+
+/*
+ * @tcp_westwood_create
+ * This function initializes fields used in TCP Westwood+,
+ * it is called after the initial SYN, so the sequence numbers
+ * are correct but new passive connections we have no
+ * information about RTTmin at this time so we simply set it to
+ * TCP_WESTWOOD_INIT_RTT. This value was chosen to be too conservative
+ * since in this way we're sure it will be updated in a consistent
+ * way as soon as possible. It will reasonably happen within the first
+ * RTT period of the connection lifetime.
+ */
+static void tcp_westwood_init(struct tcp_sock *tp)
+{
+ struct westwood *w = tcp_ca(tp);
+
+ w->bk = 0;
+ w->bw_ns_est = 0;
+ w->bw_est = 0;
+ w->accounted = 0;
+ w->cumul_ack = 0;
+ w->rtt_min = w->rtt = TCP_WESTWOOD_INIT_RTT;
+ w->rtt_win_sx = tcp_time_stamp;
+ w->snd_una = tp->snd_una;
+}
+
+/*
+ * @westwood_do_filter
+ * Low-pass filter. Implemented using constant coefficients.
+ */
+static inline u32 westwood_do_filter(u32 a, u32 b)
+{
+ return (((7 * a) + b) >> 3);
+}
+
+static inline void westwood_filter(struct westwood *w, u32 delta)
+{
+ w->bw_ns_est = westwood_do_filter(w->bw_ns_est, w->bk / delta);
+ w->bw_est = westwood_do_filter(w->bw_est, w->bw_ns_est);
+}
+
+/*
+ * @westwood_pkts_acked
+ * Called after processing group of packets.
+ * but all westwood needs is the last sample of srtt.
+ */
+static void tcp_westwood_pkts_acked(struct tcp_sock *tp, u32 cnt)
+{
+ struct westwood *w = tcp_ca(tp);
+ if (cnt > 0)
+ w->rtt = tp->srtt >> 3;
+}
+
+/*
+ * @westwood_update_window
+ * It updates RTT evaluation window if it is the right moment to do
+ * it. If so it calls filter for evaluating bandwidth.
+ */
+static void westwood_update_window(struct tcp_sock *tp)
+{
+ struct westwood *w = tcp_ca(tp);
+ s32 delta = tcp_time_stamp - w->rtt_win_sx;
+
+ /*
+ * See if a RTT-window has passed.
+ * Be careful since if RTT is less than
+ * 50ms we don't filter but we continue 'building the sample'.
+ * This minimum limit was chosen since an estimation on small
+ * time intervals is better to avoid...
+ * Obviously on a LAN we reasonably will always have
+ * right_bound = left_bound + WESTWOOD_RTT_MIN
+ */
+ if (w->rtt && delta > max_t(u32, w->rtt, TCP_WESTWOOD_RTT_MIN)) {
+ westwood_filter(w, delta);
+
+ w->bk = 0;
+ w->rtt_win_sx = tcp_time_stamp;
+ }
+}
+
+/*
+ * @westwood_fast_bw
+ * It is called when we are in fast path. In particular it is called when
+ * header prediction is successful. In such case in fact update is
+ * straight forward and doesn't need any particular care.
+ */
+static inline void westwood_fast_bw(struct tcp_sock *tp)
+{
+ struct westwood *w = tcp_ca(tp);
+
+ westwood_update_window(tp);
+
+ w->bk += tp->snd_una - w->snd_una;
+ w->snd_una = tp->snd_una;
+ w->rtt_min = min(w->rtt, w->rtt_min);
+}
+
+/*
+ * @westwood_acked_count
+ * This function evaluates cumul_ack for evaluating bk in case of
+ * delayed or partial acks.
+ */
+static inline u32 westwood_acked_count(struct tcp_sock *tp)
+{
+ struct westwood *w = tcp_ca(tp);
+
+ w->cumul_ack = tp->snd_una - w->snd_una;
+
+ /* If cumul_ack is 0 this is a dupack since it's not moving
+ * tp->snd_una.
+ */
+ if (!w->cumul_ack) {
+ w->accounted += tp->mss_cache;
+ w->cumul_ack = tp->mss_cache;
+ }
+
+ if (w->cumul_ack > tp->mss_cache) {
+ /* Partial or delayed ack */
+ if (w->accounted >= w->cumul_ack) {
+ w->accounted -= w->cumul_ack;
+ w->cumul_ack = tp->mss_cache;
+ } else {
+ w->cumul_ack -= w->accounted;
+ w->accounted = 0;
+ }
+ }
+
+ w->snd_una = tp->snd_una;
+
+ return w->cumul_ack;
+}
+
+static inline u32 westwood_bw_rttmin(const struct tcp_sock *tp)
+{
+ struct westwood *w = tcp_ca(tp);
+ return max_t(u32, (w->bw_est * w->rtt_min) / tp->mss_cache, 2);
+}
+
+/*
+ * TCP Westwood
+ * Here limit is evaluated as Bw estimation*RTTmin (for obtaining it
+ * in packets we use mss_cache). Rttmin is guaranteed to be >= 2
+ * so avoids ever returning 0.
+ */
+static u32 tcp_westwood_cwnd_min(struct tcp_sock *tp)
+{
+ return westwood_bw_rttmin(tp);
+}
+
+static void tcp_westwood_event(struct tcp_sock *tp, enum tcp_ca_event event)
+{
+ struct westwood *w = tcp_ca(tp);
+
+ switch(event) {
+ case CA_EVENT_FAST_ACK:
+ westwood_fast_bw(tp);
+ break;
+
+ case CA_EVENT_COMPLETE_CWR:
+ tp->snd_cwnd = tp->snd_ssthresh = westwood_bw_rttmin(tp);
+ break;
+
+ case CA_EVENT_FRTO:
+ tp->snd_ssthresh = westwood_bw_rttmin(tp);
+ break;
+
+ case CA_EVENT_SLOW_ACK:
+ westwood_update_window(tp);
+ w->bk += westwood_acked_count(tp);
+ w->rtt_min = min(w->rtt, w->rtt_min);
+ break;
+
+ default:
+ /* don't care */
+ break;
+ }
+}
+
+
+/* Extract info for Tcp socket info provided via netlink. */
+static void tcp_westwood_info(struct tcp_sock *tp, u32 ext,
+ struct sk_buff *skb)
+{
+ const struct westwood *ca = tcp_ca(tp);
+ if (ext & (1<<(TCPDIAG_VEGASINFO-1))) {
+ struct rtattr *rta;
+ struct tcpvegas_info *info;
+
+ rta = __RTA_PUT(skb, TCPDIAG_VEGASINFO, sizeof(*info));
+ info = RTA_DATA(rta);
+ info->tcpv_enabled = 1;
+ info->tcpv_rttcnt = 0;
+ info->tcpv_rtt = jiffies_to_usecs(ca->rtt);
+ info->tcpv_minrtt = jiffies_to_usecs(ca->rtt_min);
+ rtattr_failure: ;
+ }
+}
+
+
+static struct tcp_congestion_ops tcp_westwood = {
+ .init = tcp_westwood_init,
+ .ssthresh = tcp_reno_ssthresh,
+ .cong_avoid = tcp_reno_cong_avoid,
+ .min_cwnd = tcp_westwood_cwnd_min,
+ .cwnd_event = tcp_westwood_event,
+ .get_info = tcp_westwood_info,
+ .pkts_acked = tcp_westwood_pkts_acked,
+
+ .owner = THIS_MODULE,
+ .name = "westwood"
+};
+
+static int __init tcp_westwood_register(void)
+{
+ BUG_ON(sizeof(struct westwood) > TCP_CA_PRIV_SIZE);
+ return tcp_register_congestion_control(&tcp_westwood);
+}
+
+static void __exit tcp_westwood_unregister(void)
+{
+ tcp_unregister_congestion_control(&tcp_westwood);
+}
+
+module_init(tcp_westwood_register);
+module_exit(tcp_westwood_unregister);
+
+MODULE_AUTHOR("Stephen Hemminger, Angelo Dell'Aera");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("TCP Westwood+");
diff --git a/net/ipv4/xfrm4_output.c b/net/ipv4/xfrm4_output.c
index af2392ae576..66620a95942 100644
--- a/net/ipv4/xfrm4_output.c
+++ b/net/ipv4/xfrm4_output.c
@@ -33,6 +33,7 @@ static void xfrm4_encap(struct sk_buff *skb)
struct dst_entry *dst = skb->dst;
struct xfrm_state *x = dst->xfrm;
struct iphdr *iph, *top_iph;
+ int flags;
iph = skb->nh.iph;
skb->h.ipiph = iph;
@@ -51,10 +52,13 @@ static void xfrm4_encap(struct sk_buff *skb)
/* DS disclosed */
top_iph->tos = INET_ECN_encapsulate(iph->tos, iph->tos);
- if (x->props.flags & XFRM_STATE_NOECN)
+
+ flags = x->props.flags;
+ if (flags & XFRM_STATE_NOECN)
IP_ECN_clear(top_iph);
- top_iph->frag_off = iph->frag_off & htons(IP_DF);
+ top_iph->frag_off = (flags & XFRM_STATE_NOPMTUDISC) ?
+ 0 : (iph->frag_off & htons(IP_DF));
if (!top_iph->frag_off)
__ip_select_ident(top_iph, dst, 0);
diff --git a/net/ipv4/xfrm4_state.c b/net/ipv4/xfrm4_state.c
index 223a2e83853..050611d7a96 100644
--- a/net/ipv4/xfrm4_state.c
+++ b/net/ipv4/xfrm4_state.c
@@ -7,12 +7,20 @@
*
*/
+#include <net/ip.h>
#include <net/xfrm.h>
#include <linux/pfkeyv2.h>
#include <linux/ipsec.h>
static struct xfrm_state_afinfo xfrm4_state_afinfo;
+static int xfrm4_init_flags(struct xfrm_state *x)
+{
+ if (ipv4_config.no_pmtu_disc)
+ x->props.flags |= XFRM_STATE_NOPMTUDISC;
+ return 0;
+}
+
static void
__xfrm4_init_tempsel(struct xfrm_state *x, struct flowi *fl,
struct xfrm_tmpl *tmpl,
@@ -109,6 +117,7 @@ __xfrm4_find_acq(u8 mode, u32 reqid, u8 proto,
static struct xfrm_state_afinfo xfrm4_state_afinfo = {
.family = AF_INET,
.lock = RW_LOCK_UNLOCKED,
+ .init_flags = xfrm4_init_flags,
.init_tempsel = __xfrm4_init_tempsel,
.state_lookup = __xfrm4_state_lookup,
.find_acq = __xfrm4_find_acq,
diff --git a/net/ipv4/xfrm4_tunnel.c b/net/ipv4/xfrm4_tunnel.c
index 413191f585f..e1fe360ed27 100644
--- a/net/ipv4/xfrm4_tunnel.c
+++ b/net/ipv4/xfrm4_tunnel.c
@@ -84,7 +84,7 @@ static void ipip_err(struct sk_buff *skb, u32 info)
handler->err_handler(skb, &arg);
}
-static int ipip_init_state(struct xfrm_state *x, void *args)
+static int ipip_init_state(struct xfrm_state *x)
{
if (!x->props.mode)
return -EINVAL;