[TCP]: Add IPv6 support to TCP SYN cookies

Updated to incorporate Eric's suggestion of using a per cpu buffer
rather than allocating on the stack.  Just a two line change, but will
resend in it's entirety.

Signed-off-by: Glenn Griffin <ggriffin.kernel@gmail.com>
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>

7 files changed, 330 insertions, 26 deletions

diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 177da14f6b14..4704f27f6c0b 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -10,8 +10,6 @@
  *      2 of the License, or (at your option) any later version.
  *
  *  $Id: syncookies.c,v 1.18 2002/02/01 22:01:04 davem Exp $
- *
- *  Missing: IPv6 support.
  */
 
 #include <linux/tcp.h>
@@ -23,14 +21,15 @@
 
 extern int sysctl_tcp_syncookies;
 
-static __u32 syncookie_secret[2][16-3+SHA_DIGEST_WORDS];
+__u32 syncookie_secret[2][16-3+SHA_DIGEST_WORDS];
+EXPORT_SYMBOL(syncookie_secret);
 
 static __init int init_syncookies(void)
 {
         get_random_bytes(syncookie_secret, sizeof(syncookie_secret));
         return 0;
 }
-module_init(init_syncookies);
+__initcall(init_syncookies);
 
 #define COOKIEBITS 24   /* Upper bits store count */
 #define COOKIEMASK (((__u32)1 << COOKIEBITS) - 1)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 19c449f62672..93e128c239c4 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5326,6 +5326,7 @@ discard:
 
 EXPORT_SYMBOL(sysctl_tcp_ecn);
 EXPORT_SYMBOL(sysctl_tcp_reordering);
+EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
 EXPORT_SYMBOL(tcp_parse_options);
 EXPORT_SYMBOL(tcp_rcv_established);
 EXPORT_SYMBOL(tcp_rcv_state_process);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 0fdd1db641ac..8245247a6ceb 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -35,6 +35,8 @@
 #endif
 
 int sysctl_tcp_syncookies __read_mostly = SYNC_INIT;
+EXPORT_SYMBOL(sysctl_tcp_syncookies);
+
 int sysctl_tcp_abort_on_overflow __read_mostly;
 
 struct inet_timewait_death_row tcp_death_row = {
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index ed750f9ceb07..cbfef8b1f5e8 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2560,6 +2560,7 @@ void tcp_send_probe0(struct sock *sk)
         }
 }
 
+EXPORT_SYMBOL(tcp_select_initial_window);
 EXPORT_SYMBOL(tcp_connect);
 EXPORT_SYMBOL(tcp_make_synack);
 EXPORT_SYMBOL(tcp_simple_retransmit);
diff --git a/net/ipv6/Makefile b/net/ipv6/Makefile
index 24f3aa0f2a35..ae14617e607f 100644
--- a/net/ipv6/Makefile
+++ b/net/ipv6/Makefile
@@ -16,6 +16,7 @@ ipv6-$(CONFIG_XFRM) += xfrm6_policy.o xfrm6_state.o xfrm6_input.o \
 ipv6-$(CONFIG_NETFILTER) += netfilter.o
 ipv6-$(CONFIG_IPV6_MULTIPLE_TABLES) += fib6_rules.o
 ipv6-$(CONFIG_PROC_FS) += proc.o
+ipv6-$(CONFIG_SYN_COOKIES) += syncookies.o
 
 ipv6-objs += $(ipv6-y)
 
diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c
new file mode 100644
index 000000000000..827c5aa7524c
--- /dev/null
+++ b/net/ipv6/syncookies.c
@@ -0,0 +1,267 @@
+/*
+ *  IPv6 Syncookies implementation for the Linux kernel
+ *
+ *  Authors:
+ *  Glenn Griffin       <ggriffin.kernel@gmail.com>
+ *
+ *  Based on IPv4 implementation by Andi Kleen
+ *  linux/net/ipv4/syncookies.c
+ *
+ *      This program is free software; you can redistribute it and/or
+ *      modify it under the terms of the GNU General Public License
+ *      as published by the Free Software Foundation; either version
+ *      2 of the License, or (at your option) any later version.
+ *
+ */
+
+#include <linux/tcp.h>
+#include <linux/random.h>
+#include <linux/cryptohash.h>
+#include <linux/kernel.h>
+#include <net/ipv6.h>
+#include <net/tcp.h>
+
+extern int sysctl_tcp_syncookies;
+extern __u32 syncookie_secret[2][16-3+SHA_DIGEST_WORDS];
+
+#define COOKIEBITS 24   /* Upper bits store count */
+#define COOKIEMASK (((__u32)1 << COOKIEBITS) - 1)
+
+/*
+ * This table has to be sorted and terminated with (__u16)-1.
+ * XXX generate a better table.
+ * Unresolved Issues: HIPPI with a 64k MSS is not well supported.
+ *
+ * Taken directly from ipv4 implementation.
+ * Should this list be modified for ipv6 use or is it close enough?
+ * rfc 2460 8.3 suggests mss values 20 bytes less than ipv4 counterpart
+ */
+static __u16 const msstab[] = {
+        64 - 1,
+        256 - 1,
+        512 - 1,
+        536 - 1,
+        1024 - 1,
+        1440 - 1,
+        1460 - 1,
+        4312 - 1,
+        (__u16)-1
+};
+/* The number doesn't include the -1 terminator */
+#define NUM_MSS (ARRAY_SIZE(msstab) - 1)
+
+/*
+ * This (misnamed) value is the age of syncookie which is permitted.
+ * Its ideal value should be dependent on TCP_TIMEOUT_INIT and
+ * sysctl_tcp_retries1. It's a rather complicated formula (exponential
+ * backoff) to compute at runtime so it's currently hardcoded here.
+ */
+#define COUNTER_TRIES 4
+
+static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
+                                           struct request_sock *req,
+                                           struct dst_entry *dst)
+{
+        struct inet_connection_sock *icsk = inet_csk(sk);
+        struct sock *child;
+
+        child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst);
+        if (child)
+                inet_csk_reqsk_queue_add(sk, req, child);
+        else
+                reqsk_free(req);
+
+        return child;
+}
+
+static DEFINE_PER_CPU(__u32, cookie_scratch)[16 + 5 + SHA_WORKSPACE_WORDS];
+
+static u32 cookie_hash(struct in6_addr *saddr, struct in6_addr *daddr,
+                       __be16 sport, __be16 dport, u32 count, int c)
+{
+        __u32 *tmp = __get_cpu_var(cookie_scratch);
+
+        /*
+         * we have 320 bits of information to hash, copy in the remaining
+         * 192 bits required for sha_transform, from the syncookie_secret
+         * and overwrite the digest with the secret
+         */
+        memcpy(tmp + 10, syncookie_secret[c], 44);
+        memcpy(tmp, saddr, 16);
+        memcpy(tmp + 4, daddr, 16);
+        tmp[8] = ((__force u32)sport << 16) + (__force u32)dport;
+        tmp[9] = count;
+        sha_transform(tmp + 16, (__u8 *)tmp, tmp + 16 + 5);
+
+        return tmp[17];
+}
+
+static __u32 secure_tcp_syn_cookie(struct in6_addr *saddr, struct in6_addr *daddr,
+                                   __be16 sport, __be16 dport, __u32 sseq,
+                                   __u32 count, __u32 data)
+{
+        return (cookie_hash(saddr, daddr, sport, dport, 0, 0) +
+                sseq + (count << COOKIEBITS) +
+                ((cookie_hash(saddr, daddr, sport, dport, count, 1) + data)
+                & COOKIEMASK));
+}
+
+static __u32 check_tcp_syn_cookie(__u32 cookie, struct in6_addr *saddr,
+                                  struct in6_addr *daddr, __be16 sport,
+                                  __be16 dport, __u32 sseq, __u32 count,
+                                  __u32 maxdiff)
+{
+        __u32 diff;
+
+        cookie -= cookie_hash(saddr, daddr, sport, dport, 0, 0) + sseq;
+
+        diff = (count - (cookie >> COOKIEBITS)) & ((__u32) -1 >> COOKIEBITS);
+        if (diff >= maxdiff)
+                return (__u32)-1;
+
+        return (cookie -
+                cookie_hash(saddr, daddr, sport, dport, count - diff, 1))
+                & COOKIEMASK;
+}
+
+__u32 cookie_v6_init_sequence(struct sock *sk, struct sk_buff *skb, __u16 *mssp)
+{
+        struct ipv6hdr *iph = ipv6_hdr(skb);
+        const struct tcphdr *th = tcp_hdr(skb);
+        int mssind;
+        const __u16 mss = *mssp;
+
+        tcp_sk(sk)->last_synq_overflow = jiffies;
+
+        for (mssind = 0; mss > msstab[mssind + 1]; mssind++)
+                ;
+        *mssp = msstab[mssind] + 1;
+
+        NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESSENT);
+
+        return secure_tcp_syn_cookie(&iph->saddr, &iph->daddr, th->source,
+                                     th->dest, ntohl(th->seq),
+                                     jiffies / (HZ * 60), mssind);
+}
+
+static inline int cookie_check(struct sk_buff *skb, __u32 cookie)
+{
+        struct ipv6hdr *iph = ipv6_hdr(skb);
+        const struct tcphdr *th = tcp_hdr(skb);
+        __u32 seq = ntohl(th->seq) - 1;
+        __u32 mssind = check_tcp_syn_cookie(cookie, &iph->saddr, &iph->daddr,
+                                            th->source, th->dest, seq,
+                                            jiffies / (HZ * 60), COUNTER_TRIES);
+
+        return mssind < NUM_MSS ? msstab[mssind] + 1 : 0;
+}
+
+struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
+{
+        struct inet_request_sock *ireq;
+        struct inet6_request_sock *ireq6;
+        struct tcp_request_sock *treq;
+        struct ipv6_pinfo *np = inet6_sk(sk);
+        struct tcp_sock *tp = tcp_sk(sk);
+        const struct tcphdr *th = tcp_hdr(skb);
+        __u32 cookie = ntohl(th->ack_seq) - 1;
+        struct sock *ret = sk;
+        struct request_sock *req;
+        int mss;
+        struct dst_entry *dst;
+        __u8 rcv_wscale;
+
+        if (!sysctl_tcp_syncookies || !th->ack)
+                goto out;
+
+        if (time_after(jiffies, tp->last_synq_overflow + TCP_TIMEOUT_INIT) ||
+                (mss = cookie_check(skb, cookie)) == 0) {
+                NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESFAILED);
+                goto out;
+        }
+
+        NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV);
+
+        ret = NULL;
+        req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
+        if (!req)
+                goto out;
+
+        ireq = inet_rsk(req);
+        ireq6 = inet6_rsk(req);
+        treq = tcp_rsk(req);
+        ireq6->pktopts = NULL;
+
+        if (security_inet_conn_request(sk, skb, req)) {
+                reqsk_free(req);
+                goto out;
+        }
+
+        req->mss = mss;
+        ireq->rmt_port = th->source;
+        ipv6_addr_copy(&ireq6->rmt_addr, &ipv6_hdr(skb)->saddr);
+        ipv6_addr_copy(&ireq6->loc_addr, &ipv6_hdr(skb)->daddr);
+        if (ipv6_opt_accepted(sk, skb) ||
+            np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||
+            np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim) {
+                atomic_inc(&skb->users);
+                ireq6->pktopts = skb;
+        }
+
+        ireq6->iif = sk->sk_bound_dev_if;
+        /* So that link locals have meaning */
+        if (!sk->sk_bound_dev_if &&
+            ipv6_addr_type(&ireq6->rmt_addr) & IPV6_ADDR_LINKLOCAL)
+                ireq6->iif = inet6_iif(skb);
+
+        req->expires = 0UL;
+        req->retrans = 0;
+        ireq->snd_wscale = ireq->rcv_wscale = ireq->tstamp_ok = 0;
+        ireq->wscale_ok = ireq->sack_ok = 0;
+        treq->rcv_isn = ntohl(th->seq) - 1;
+        treq->snt_isn = cookie;
+
+        /*
+         * We need to lookup the dst_entry to get the correct window size.
+         * This is taken from tcp_v6_syn_recv_sock.  Somebody please enlighten
+         * me if there is a preferred way.
+         */
+        {
+                struct in6_addr *final_p = NULL, final;
+                struct flowi fl;
+                memset(&fl, 0, sizeof(fl));
+                fl.proto = IPPROTO_TCP;
+                ipv6_addr_copy(&fl.fl6_dst, &ireq6->rmt_addr);
+                if (np->opt && np->opt->srcrt) {
+                        struct rt0_hdr *rt0 = (struct rt0_hdr *) np->opt->srcrt;
+                        ipv6_addr_copy(&final, &fl.fl6_dst);
+                        ipv6_addr_copy(&fl.fl6_dst, rt0->addr);
+                        final_p = &final;
+                }
+                ipv6_addr_copy(&fl.fl6_src, &ireq6->loc_addr);
+                fl.oif = sk->sk_bound_dev_if;
+                fl.fl_ip_dport = inet_rsk(req)->rmt_port;
+                fl.fl_ip_sport = inet_sk(sk)->sport;
+                security_req_classify_flow(req, &fl);
+                if (ip6_dst_lookup(sk, &dst, &fl)) {
+                        reqsk_free(req);
+                        goto out;
+                }
+                if (final_p)
+                        ipv6_addr_copy(&fl.fl6_dst, final_p);
+                if ((xfrm_lookup(&dst, &fl, sk, 0)) < 0)
+                        goto out;
+        }
+
+        req->window_clamp = dst_metric(dst, RTAX_WINDOW);
+        tcp_select_initial_window(tcp_full_space(sk), req->mss,
+                                  &req->rcv_wnd, &req->window_clamp,
+                                  0, &rcv_wscale);
+
+        ireq->rcv_wscale = rcv_wscale;
+
+        ret = get_cookie_sock(sk, skb, req, dst);
+
+out:    return ret;
+}
+
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 1cbbb87dbad2..fd773ac7531a 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -512,6 +512,20 @@ done:
         return err;
 }
 
+static inline void syn_flood_warning(struct sk_buff *skb)
+{
+#ifdef CONFIG_SYN_COOKIES
+        if (sysctl_tcp_syncookies)
+                printk(KERN_INFO
+                       "TCPv6: Possible SYN flooding on port %d. "
+                       "Sending cookies.\n", ntohs(tcp_hdr(skb)->dest));
+        else
+#endif
+                printk(KERN_INFO
+                       "TCPv6: Possible SYN flooding on port %d. "
+                       "Dropping request.\n", ntohs(tcp_hdr(skb)->dest));
+}
+
 static void tcp_v6_reqsk_destructor(struct request_sock *req)
 {
         if (inet6_rsk(req)->pktopts)
@@ -915,7 +929,7 @@ done_opts:
 }
 #endif
 
-static struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
+struct request_sock_ops tcp6_request_sock_ops __read_mostly = {
         .family         =       AF_INET6,
         .obj_size       =       sizeof(struct tcp6_request_sock),
         .rtx_syn_ack    =       tcp_v6_send_synack,
@@ -1213,9 +1227,9 @@ static struct sock *tcp_v6_hnd_req(struct sock *sk,struct sk_buff *skb)
                 return NULL;
         }
 
-#if 0 /*def CONFIG_SYN_COOKIES*/
+#ifdef CONFIG_SYN_COOKIES
         if (!th->rst && !th->syn && th->ack)
-                sk = cookie_v6_check(sk, skb, &(IPCB(skb)->opt));
+                sk = cookie_v6_check(sk, skb);
 #endif
         return sk;
 }
@@ -1231,6 +1245,11 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
         struct tcp_sock *tp = tcp_sk(sk);
         struct request_sock *req = NULL;
         __u32 isn = TCP_SKB_CB(skb)->when;
+#ifdef CONFIG_SYN_COOKIES
+        int want_cookie = 0;
+#else
+#define want_cookie 0
+#endif
 
         if (skb->protocol == htons(ETH_P_IP))
                 return tcp_v4_conn_request(sk, skb);
@@ -1238,12 +1257,14 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
         if (!ipv6_unicast_destination(skb))
                 goto drop;
 
-        /*
-         *      There are no SYN attacks on IPv6, yet...
-         */
         if (inet_csk_reqsk_queue_is_full(sk) && !isn) {
                 if (net_ratelimit())
-                        printk(KERN_INFO "TCPv6: dropping request, synflood is possible\n");
+                        syn_flood_warning(skb);
+#ifdef CONFIG_SYN_COOKIES
+                if (sysctl_tcp_syncookies)
+                        want_cookie = 1;
+                else
+#endif
                 goto drop;
         }
 
@@ -1264,29 +1285,39 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
 
         tcp_parse_options(skb, &tmp_opt, 0);
 
+        if (want_cookie) {
+                tcp_clear_options(&tmp_opt);
+                tmp_opt.saw_tstamp = 0;
+        }
+
         tmp_opt.tstamp_ok = tmp_opt.saw_tstamp;
         tcp_openreq_init(req, &tmp_opt, skb);
 
         treq = inet6_rsk(req);
         ipv6_addr_copy(&treq->rmt_addr, &ipv6_hdr(skb)->saddr);
         ipv6_addr_copy(&treq->loc_addr, &ipv6_hdr(skb)->daddr);
-        TCP_ECN_create_request(req, tcp_hdr(skb));
         treq->pktopts = NULL;
-        if (ipv6_opt_accepted(sk, skb) ||
-            np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||
-            np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim) {
-                atomic_inc(&skb->users);
-                treq->pktopts = skb;
-        }
-        treq->iif = sk->sk_bound_dev_if;
+        if (!want_cookie)
+                TCP_ECN_create_request(req, tcp_hdr(skb));
+
+        if (want_cookie) {
+                isn = cookie_v6_init_sequence(sk, skb, &req->mss);
+        } else if (!isn) {
+                if (ipv6_opt_accepted(sk, skb) ||
+                    np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||
+                    np->rxopt.bits.rxhlim || np->rxopt.bits.rxohlim) {
+                        atomic_inc(&skb->users);
+                        treq->pktopts = skb;
+                }
+                treq->iif = sk->sk_bound_dev_if;
 
-        /* So that link locals have meaning */
-        if (!sk->sk_bound_dev_if &&
-            ipv6_addr_type(&treq->rmt_addr) & IPV6_ADDR_LINKLOCAL)
-                treq->iif = inet6_iif(skb);
+                /* So that link locals have meaning */
+                if (!sk->sk_bound_dev_if &&
+                    ipv6_addr_type(&treq->rmt_addr) & IPV6_ADDR_LINKLOCAL)
+                        treq->iif = inet6_iif(skb);
 
-        if (isn == 0)
                 isn = tcp_v6_init_sequence(skb);
+        }
 
         tcp_rsk(req)->snt_isn = isn;
 
@@ -1295,8 +1326,10 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
         if (tcp_v6_send_synack(sk, req))
                 goto drop;
 
-        inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
-        return 0;
+        if (!want_cookie) {
+                inet6_csk_reqsk_queue_hash_add(sk, req, TCP_TIMEOUT_INIT);
+                return 0;
+        }
 
 drop:
         if (req)