From: jolsa@redhat.com <jolsa@redhat.com>
Date: Mon, 24 Aug 2009 14:36:21 +0200
Subject: [net] syncookies: support for TCP options via timestamps
Message-id: 1251117381-6002-3-git-send-email-jolsa@redhat.com
O-Subject: [RHEL5 PATCH 2/2] BZ 509062 [Syncookies]: Add support for TCP options via timestamps
Bugzilla: 509062
RH-Acked-by: Neil Horman <nhorman@redhat.com>
RH-Acked-by: Thomas Graf <tgraf@redhat.com>

Bugzilla: 509062
https://bugzilla.redhat.com/show_bug.cgi?id=509062

Upstream status:
================
[Syncookies]: Add support for TCP options via timestamps.
commit 4dfc2817025965a2fc78a18c50f540736a6b5c24

wbr,
jirka

diff --git a/include/net/request_sock.h b/include/net/request_sock.h
index f743a94..c548750 100644
--- a/include/net/request_sock.h
+++ b/include/net/request_sock.h
@@ -45,6 +45,9 @@ struct request_sock {
 	struct request_sock		*dl_next; /* Must be first member! */
 	u16				mss;
 	u8				retrans;
+	/* The following field is used under the name 'cookie_ts' by
+	   the define below. The reason is KABI compliance.
+	*/
 	u8				__pad;
 	/* The following two fields can be easily recomputed I think -AK */
 	u32				window_clamp; /* window clamp at creation time */
@@ -57,6 +60,9 @@ struct request_sock {
 	u32				peer_secid;
 };
 
+/* syncookie: encode tcpopts in timestamp in struct request_sock */
+#define cookie_ts __pad
+
 static inline struct request_sock *reqsk_alloc(struct request_sock_ops *ops)
 {
 	struct request_sock *req = kmem_cache_alloc(ops->slab, SLAB_ATOMIC);
diff --git a/include/net/tcp.h b/include/net/tcp.h
index df07c61..92466dd 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -424,6 +424,9 @@ extern struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb);
 extern __u32 cookie_v6_init_sequence(struct sock *sk, struct sk_buff *skb,
 				     __u16 *mss);
 
+extern __u32 cookie_init_timestamp(struct request_sock *req);
+extern void cookie_check_timestamp(struct tcp_options_received *tcp_opt);
+
 /* tcp_output.c */
 
 extern void __tcp_push_pending_frames(struct sock *sk, struct tcp_sock *tp,
@@ -976,6 +979,7 @@ static inline void tcp_openreq_init(struct request_sock *req,
 	struct inet_request_sock *ireq = inet_rsk(req);
 
 	req->rcv_wnd = 0;		/* So that tcp_send_synack() knows! */
+	req->cookie_ts = 0;
 	tcp_rsk(req)->rcv_isn = TCP_SKB_CB(skb)->seq;
 	req->mss = rx_opt->mss_clamp;
 	req->ts_recent = rx_opt->saw_tstamp ? rx_opt->rcv_tsval : 0;
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 38e8aa8..99d9c02 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -19,6 +19,10 @@
 #include <linux/kernel.h>
 #include <net/tcp.h>
 
+/* Timestamps: lowest 9 bits store TCP options */
+#define TSBITS 9
+#define TSMASK (((__u32)1 << TSBITS) - 1)
+
 extern int sysctl_tcp_syncookies;
 
 __u32 syncookie_secret[2][16-3+SHA_DIGEST_WORDS];
@@ -49,6 +53,37 @@ static u32 cookie_hash(u32 saddr, u32 daddr, u32 sport, u32 dport,
 	return tmp[17];
 }
 
+/*
+ * when syncookies are in effect and tcp timestamps are enabled we encode
+ * tcp options in the lowest 9 bits of the timestamp value that will be
+ * sent in the syn-ack.
+ * Since subsequent timestamps use the normal tcp_time_stamp value, we
+ * must make sure that the resulting initial timestamp is <= tcp_time_stamp.
+ */
+__u32 cookie_init_timestamp(struct request_sock *req)
+{
+	struct inet_request_sock *ireq;
+	u32 ts, ts_now = tcp_time_stamp;
+	u32 options = 0;
+
+	ireq = inet_rsk(req);
+	if (ireq->wscale_ok) {
+		options = ireq->snd_wscale;
+		options |= ireq->rcv_wscale << 4;
+	}
+	options |= ireq->sack_ok << 8;
+
+	ts = ts_now & ~TSMASK;
+	ts |= options;
+	if (ts > ts_now) {
+		ts >>= TSBITS;
+		ts--;
+		ts <<= TSBITS;
+		ts |= options;
+	}
+	return ts;
+}
+
 static __u32 secure_tcp_syn_cookie(__u32 saddr, __u32 daddr, __u16 sport,
 				   __u16 dport, __u32 sseq, __u32 count,
 				   __u32 data)
@@ -184,6 +219,35 @@ static inline struct sock *get_cookie_sock(struct sock *sk, struct sk_buff *skb,
 	return child;
 }
 
+
+/*
+ * when syncookies are in effect and tcp timestamps are enabled we stored
+ * additional tcp options in the timestamp.
+ * This extracts these options from the timestamp echo.
+ *
+ * The lowest 4 bits are for snd_wscale
+ * The next 4 lsb are for rcv_wscale
+ * The next lsb is for sack_ok
+ */
+void cookie_check_timestamp(struct tcp_options_received *tcp_opt)
+{
+	/* echoed timestamp, 9 lowest bits contain options */
+	u32 options = tcp_opt->rcv_tsecr & TSMASK;
+
+	tcp_opt->snd_wscale = options & 0xf;
+	options >>= 4;
+	tcp_opt->rcv_wscale = options & 0xf;
+
+	tcp_opt->sack_ok = (options >> 4) & 0x1;
+
+	if (tcp_opt->sack_ok)
+		tcp_sack_reset(tcp_opt);
+
+	if (tcp_opt->snd_wscale || tcp_opt->rcv_wscale)
+		tcp_opt->wscale_ok = 1;
+}
+EXPORT_SYMBOL(cookie_check_timestamp);
+
 struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 			     struct ip_options *opt)
 {
@@ -196,6 +260,7 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 	int mss; 
 	struct rtable *rt; 
 	__u8 rcv_wscale;
+	struct tcp_options_received tcp_opt;
 
 	if (!sysctl_tcp_syncookies || !skb->h.th->ack)
 		goto out;
@@ -208,6 +273,13 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 
 	NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV);
 
+	/* check for timestamp cookie support */
+	memset(&tcp_opt, 0, sizeof(tcp_opt));
+	tcp_parse_options(skb, &tcp_opt, 0);
+
+	if (tcp_opt.saw_tstamp)
+		cookie_check_timestamp(&tcp_opt);
+
 	ret = NULL;
 	req = reqsk_alloc(&tcp_request_sock_ops); /* for safety */
 	if (!req)
@@ -226,6 +298,12 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 	ireq->loc_addr		= skb->nh.iph->daddr;
 	ireq->rmt_addr		= skb->nh.iph->saddr;
 	ireq->opt		= NULL;
+	ireq->snd_wscale	= tcp_opt.snd_wscale;
+	ireq->rcv_wscale	= tcp_opt.rcv_wscale;
+	ireq->sack_ok		= tcp_opt.sack_ok;
+	ireq->wscale_ok		= tcp_opt.wscale_ok;
+	ireq->tstamp_ok		= tcp_opt.saw_tstamp;
+	req->ts_recent		= tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
 
 	/* 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)
@@ -240,8 +318,6 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 		}
 	}
 
-	ireq->snd_wscale = ireq->rcv_wscale = ireq->tstamp_ok = 0;
-	ireq->wscale_ok	 = ireq->sack_ok = 0; 
 	req->expires	= 0UL; 
 	req->retrans	= 0; 
 	
@@ -270,11 +346,12 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
 	}
 
 	/* Try to redo what tcp_v4_send_synack did. */
-	req->window_clamp = dst_metric(&rt->u.dst, RTAX_WINDOW);
+	req->window_clamp = tp->window_clamp ? :dst_metric(&rt->u.dst, RTAX_WINDOW);
+
 	tcp_select_initial_window(tcp_full_space(sk), req->mss,
 				  &req->rcv_wnd, &req->window_clamp, 
-				  0, &rcv_wscale);
-	/* BTW win scale with syncookies is 0 by definition */
+				  ireq->wscale_ok, &rcv_wscale);
+
 	ireq->rcv_wscale  = rcv_wscale; 
 
 	ret = get_cookie_sock(sk, skb, req, &rt->u.dst);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 01583da..ec8e0a7 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -783,10 +783,8 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
 
 	tcp_parse_options(skb, &tmp_opt, 0);
 
-	if (want_cookie) {
+	if (want_cookie && !tmp_opt.saw_tstamp)
 		tcp_clear_options(&tmp_opt);
-		tmp_opt.saw_tstamp = 0;
-	}
 
 	if (tmp_opt.saw_tstamp && !tmp_opt.rcv_tsval) {
 		/* Some OSes (unknown ones, but I see them on web server, which
@@ -814,6 +812,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
 	if (want_cookie) {
 #ifdef CONFIG_SYN_COOKIES
 		syn_flood_warning(skb);
+		req->cookie_ts = tmp_opt.tstamp_ok;
 #endif
 		isn = cookie_v4_init_sequence(sk, skb, &req->mss);
 	} else if (!isn) {
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 673d2e8..9c6fbbf 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2082,7 +2082,11 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 
 	/* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */
 	th->window = htons(req->rcv_wnd);
-
+#ifdef CONFIG_SYN_COOKIES
+	if (unlikely(req->cookie_ts))
+		TCP_SKB_CB(skb)->when = cookie_init_timestamp(req);
+	else
+#endif
 	TCP_SKB_CB(skb)->when = tcp_time_stamp;
 	tcp_syn_build_options((__u32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok,
 			      ireq->sack_ok, ireq->wscale_ok, ireq->rcv_wscale,
diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c
index 827c5aa..e75a458 100644
--- a/net/ipv6/syncookies.c
+++ b/net/ipv6/syncookies.c
@@ -170,6 +170,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 	int mss;
 	struct dst_entry *dst;
 	__u8 rcv_wscale;
+	struct tcp_options_received tcp_opt;
 
 	if (!sysctl_tcp_syncookies || !th->ack)
 		goto out;
@@ -182,6 +183,13 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 
 	NET_INC_STATS_BH(LINUX_MIB_SYNCOOKIESRECV);
 
+	/* check for timestamp cookie support */
+	memset(&tcp_opt, 0, sizeof(tcp_opt));
+	tcp_parse_options(skb, &tcp_opt, 0);
+
+	if (tcp_opt.saw_tstamp)
+		cookie_check_timestamp(&tcp_opt);
+
 	ret = NULL;
 	req = inet6_reqsk_alloc(&tcp6_request_sock_ops);
 	if (!req)
@@ -216,8 +224,12 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 
 	req->expires = 0UL;
 	req->retrans = 0;
-	ireq->snd_wscale = ireq->rcv_wscale = ireq->tstamp_ok = 0;
-	ireq->wscale_ok = ireq->sack_ok = 0;
+	ireq->snd_wscale	= tcp_opt.snd_wscale;
+	ireq->rcv_wscale	= tcp_opt.rcv_wscale;
+	ireq->sack_ok		= tcp_opt.sack_ok;
+	ireq->wscale_ok		= tcp_opt.wscale_ok;
+	ireq->tstamp_ok		= tcp_opt.saw_tstamp;
+	req->ts_recent		= tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
 	treq->rcv_isn = ntohl(th->seq) - 1;
 	treq->snt_isn = cookie;
 
@@ -253,10 +265,10 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
 			goto out;
 	}
 
-	req->window_clamp = dst_metric(dst, RTAX_WINDOW);
+	req->window_clamp = tp->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->wscale_ok, &rcv_wscale);
 
 	ireq->rcv_wscale = rcv_wscale;
 
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 4f82238..6cfb9bd 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -859,10 +859,8 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
 
 	tcp_parse_options(skb, &tmp_opt, 0);
 
-	if (want_cookie) {
+	if (want_cookie && !tmp_opt.saw_tstamp)
 		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);
@@ -876,6 +874,7 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
 
 	if (want_cookie) {
 		isn = cookie_v6_init_sequence(sk, skb, &req->mss);
+		req->cookie_ts = tmp_opt.tstamp_ok;
 	} else if (!isn) {
 		if (ipv6_opt_accepted(sk, skb) ||
 		    np->rxopt.bits.rxinfo || np->rxopt.bits.rxoinfo ||