[PATCH net-next] tcp: add ability to set a timestamp offset

[PATCH net-next] tcp: add ability to set a timestamp offset

[Andrey Vagin]

If a TCP socket will get live-migrated from one box to another the
timestamps (which are typically ON) will get screwed up -- the new
kernel will generate TS values that has nothing to do with what they
were on dump. The solution is to yet again fix the kernel and put a
"timestamp offset" on a socket.

A socket offset is added in places where externally visible tcp
timestamp option is parsed/initialized.

Connections in the SYN_RECV state are not supported, global
tcp_time_stamp is used for them, because repair mode doesn't support
this state. In a future it can be implemented by the similar way as for
TIME_WAIT sockets.

For time-wait sockets offset is inhereted by a proper tcp_sock.

A per-socket offset can be set only for sockets in repair mode.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>

Andrey Vagin (3):
  tcp: adding a per-socket timestamp offset
  tcp: set and get per-socket timestamp
  tcp: send packets with a socket timestamp

 include/linux/tcp.h      |  3 +++
 include/uapi/linux/tcp.h |  1 +
 net/ipv4/tcp.c           | 11 +++++++++++
 net/ipv4/tcp_input.c     |  8 +++++++-
 net/ipv4/tcp_ipv4.c      | 12 +++++++-----
 net/ipv4/tcp_minisocks.c |  3 +++
 net/ipv4/tcp_output.c    |  4 ++--
 net/ipv6/tcp_ipv6.c      | 22 +++++++++++++---------
 8 files changed, 47 insertions(+), 17 deletions(-)
-- 
1.7.11.7

[PATCH 1/3] tcp: adding a per-socket timestamp offset

[Andrey Vagin]

This functionality is used for restoring tcp sockets. A tcp timestamp
depends on how long a system has been running, so it's differ for each
host. The solution is to set a per-socket offset.

A per-socket offset for a TIME_WAIT socket is inherited from a proper
tcp socket.

tcp_request_sock doesn't have a timestamp offset, because the repair
mode for them are not implemented.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
---
 include/linux/tcp.h      | 3 +++
 net/ipv4/tcp.c           | 2 ++
 net/ipv4/tcp_minisocks.c | 2 ++
 3 files changed, 7 insertions(+)

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 4e1d228..974fb4a 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -162,6 +162,8 @@ struct tcp_sock {
 	u32	rcv_tstamp;	/* timestamp of last received ACK (for keepalives) */
 	u32	lsndtime;	/* timestamp of last sent data packet (for restart window) */
 
+	u32	tsoffset;	/* timestamp offset */
+
 	struct list_head tsq_node; /* anchor in tsq_tasklet.head list */
 	unsigned long	tsq_flags;
 
@@ -354,6 +356,7 @@ struct tcp_timewait_sock {
 	u32			  tw_rcv_nxt;
 	u32			  tw_snd_nxt;
 	u32			  tw_rcv_wnd;
+	u32			  tw_ts_offset;
 	u32			  tw_ts_recent;
 	long			  tw_ts_recent_stamp;
 #ifdef CONFIG_TCP_MD5SIG
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 1ca2536..60fab09 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -400,6 +400,8 @@ void tcp_init_sock(struct sock *sk)
 	tcp_enable_early_retrans(tp);
 	icsk->icsk_ca_ops = &tcp_init_congestion_ops;
 
+	tp->tsoffset = 0;
+
 	sk->sk_state = TCP_CLOSE;
 
 	sk->sk_write_space = sk_stream_write_space;
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index f35f2df..b94940e 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -288,6 +288,7 @@ void tcp_time_wait(struct sock *sk, int state, int timeo)
 		tcptw->tw_rcv_wnd	= tcp_receive_window(tp);
 		tcptw->tw_ts_recent	= tp->rx_opt.ts_recent;
 		tcptw->tw_ts_recent_stamp = tp->rx_opt.ts_recent_stamp;
+		tcptw->tw_ts_offset	= tp->tsoffset;
 
 #if IS_ENABLED(CONFIG_IPV6)
 		if (tw->tw_family == PF_INET6) {
@@ -500,6 +501,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
 			newtp->rx_opt.ts_recent_stamp = 0;
 			newtp->tcp_header_len = sizeof(struct tcphdr);
 		}
+		newtp->tsoffset = 0;
 #ifdef CONFIG_TCP_MD5SIG
 		newtp->md5sig_info = NULL;	/*XXX*/
 		if (newtp->af_specific->md5_lookup(sk, newsk))
-- 
1.7.11.7

[PATCH 2/3] tcp: set and get per-socket timestamp

[Andrey Vagin]

A timestamp can be set, only if a socket is in the repair mode.

This patch adds a new socket option TCP_TIMESTAMP, which allows to
get and set current tcp times stamp.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
---
 include/uapi/linux/tcp.h | 1 +
 net/ipv4/tcp.c           | 9 +++++++++
 2 files changed, 10 insertions(+)

diff --git a/include/uapi/linux/tcp.h b/include/uapi/linux/tcp.h
index e962faa..6b1ead0 100644
--- a/include/uapi/linux/tcp.h
+++ b/include/uapi/linux/tcp.h
@@ -111,6 +111,7 @@ enum {
 #define TCP_QUEUE_SEQ		21
 #define TCP_REPAIR_OPTIONS	22
 #define TCP_FASTOPEN		23	/* Enable FastOpen on listeners */
+#define TCP_TIMESTAMP		24
 
 struct tcp_repair_opt {
 	__u32	opt_code;
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 60fab09..9e484fc 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2706,6 +2706,12 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
 		else
 			err = -EINVAL;
 		break;
+	case TCP_TIMESTAMP:
+		if (!tp->repair)
+			err = -EPERM;
+		else
+			tp->tsoffset = val - tcp_time_stamp;
+		break;
 	default:
 		err = -ENOPROTOOPT;
 		break;
@@ -2954,6 +2960,9 @@ static int do_tcp_getsockopt(struct sock *sk, int level,
 	case TCP_USER_TIMEOUT:
 		val = jiffies_to_msecs(icsk->icsk_user_timeout);
 		break;
+	case TCP_TIMESTAMP:
+		val = tcp_time_stamp + tp->tsoffset;
+		break;
 	default:
 		return -ENOPROTOOPT;
 	}
-- 
1.7.11.7

[PATCH 3/3] tcp: send packets with a socket timestamp

[Andrey Vagin]

A socket timestamp is a sum of the global tcp_time_stamp and
a per-socket offset.

A socket offset is added in places where externally visible
tcp timestamp option is parsed/initialized.

Connections in the SYN_RECV state are not supported, global
tcp_time_stamp is used for them, because repair mode doesn't support
this state. In a future it can be implemented by the similar way
as for TIME_WAIT sockets.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
---
 net/ipv4/tcp_input.c     |  8 +++++++-
 net/ipv4/tcp_ipv4.c      | 12 +++++++-----
 net/ipv4/tcp_minisocks.c |  1 +
 net/ipv4/tcp_output.c    |  4 ++--
 net/ipv6/tcp_ipv6.c      | 22 +++++++++++++---------
 5 files changed, 30 insertions(+), 17 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index a28e4db..dcb7a52 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -3872,7 +3872,7 @@ static bool tcp_parse_aligned_timestamp(struct tcp_sock *tp, const struct tcphdr
 		++ptr;
 		tp->rx_opt.rcv_tsval = ntohl(*ptr);
 		++ptr;
-		tp->rx_opt.rcv_tsecr = ntohl(*ptr);
+		tp->rx_opt.rcv_tsecr = ntohl(*ptr) - tp->tsoffset;
 		return true;
 	}
 	return false;
@@ -3896,7 +3896,11 @@ static bool tcp_fast_parse_options(const struct sk_buff *skb,
 		if (tcp_parse_aligned_timestamp(tp, th))
 			return true;
 	}
+
 	tcp_parse_options(skb, &tp->rx_opt, hvpp, 1, NULL);
+	if (tp->rx_opt.saw_tstamp)
+		tp->rx_opt.rcv_tsecr -= tp->tsoffset;
+
 	return true;
 }
 
@@ -5678,6 +5682,8 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
 	int saved_clamp = tp->rx_opt.mss_clamp;
 
 	tcp_parse_options(skb, &tp->rx_opt, &hash_location, 0, &foc);
+	if (tp->rx_opt.saw_tstamp)
+		tp->rx_opt.rcv_tsecr -= tp->tsoffset;
 
 	if (th->ack) {
 		/* rfc793:
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 54139fa..0679fb7 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -725,7 +725,7 @@ release_sk1:
  */
 
 static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
-			    u32 win, u32 ts, int oif,
+			    u32 win, u32 tsval, u32 tsecr, int oif,
 			    struct tcp_md5sig_key *key,
 			    int reply_flags, u8 tos)
 {
@@ -746,12 +746,12 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
 
 	arg.iov[0].iov_base = (unsigned char *)&rep;
 	arg.iov[0].iov_len  = sizeof(rep.th);
-	if (ts) {
+	if (tsecr) {
 		rep.opt[0] = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
 				   (TCPOPT_TIMESTAMP << 8) |
 				   TCPOLEN_TIMESTAMP);
-		rep.opt[1] = htonl(tcp_time_stamp);
-		rep.opt[2] = htonl(ts);
+		rep.opt[1] = htonl(tsval);
+		rep.opt[2] = htonl(tsecr);
 		arg.iov[0].iov_len += TCPOLEN_TSTAMP_ALIGNED;
 	}
 
@@ -766,7 +766,7 @@ static void tcp_v4_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
 
 #ifdef CONFIG_TCP_MD5SIG
 	if (key) {
-		int offset = (ts) ? 3 : 0;
+		int offset = (tsecr) ? 3 : 0;
 
 		rep.opt[offset++] = htonl((TCPOPT_NOP << 24) |
 					  (TCPOPT_NOP << 16) |
@@ -801,6 +801,7 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
 
 	tcp_v4_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
 			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
+			tcp_time_stamp + tcptw->tw_ts_offset,
 			tcptw->tw_ts_recent,
 			tw->tw_bound_dev_if,
 			tcp_twsk_md5_key(tcptw),
@@ -820,6 +821,7 @@ static void tcp_v4_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
 	tcp_v4_send_ack(skb, (sk->sk_state == TCP_LISTEN) ?
 			tcp_rsk(req)->snt_isn + 1 : tcp_sk(sk)->snd_nxt,
 			tcp_rsk(req)->rcv_nxt, req->rcv_wnd,
+			tcp_time_stamp,
 			req->ts_recent,
 			0,
 			tcp_md5_do_lookup(sk, (union tcp_md5_addr *)&ip_hdr(skb)->daddr,
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index b94940e..9f26eb3 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -102,6 +102,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
 		tcp_parse_options(skb, &tmp_opt, &hash_location, 0, NULL);
 
 		if (tmp_opt.saw_tstamp) {
+			tmp_opt.rcv_tsecr	-= tcptw->tw_ts_offset;
 			tmp_opt.ts_recent	= tcptw->tw_ts_recent;
 			tmp_opt.ts_recent_stamp	= tcptw->tw_ts_recent_stamp;
 			paws_reject = tcp_paws_reject(&tmp_opt, th->rst);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 5d45159..5c6fc14 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -622,7 +622,7 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb,
 
 	if (likely(sysctl_tcp_timestamps && *md5 == NULL)) {
 		opts->options |= OPTION_TS;
-		opts->tsval = TCP_SKB_CB(skb)->when;
+		opts->tsval = TCP_SKB_CB(skb)->when + tp->tsoffset;
 		opts->tsecr = tp->rx_opt.ts_recent;
 		remaining -= TCPOLEN_TSTAMP_ALIGNED;
 	}
@@ -806,7 +806,7 @@ static unsigned int tcp_established_options(struct sock *sk, struct sk_buff *skb
 
 	if (likely(tp->rx_opt.tstamp_ok)) {
 		opts->options |= OPTION_TS;
-		opts->tsval = tcb ? tcb->when : 0;
+		opts->tsval = tcb ? tcb->when + tp->tsoffset : 0;
 		opts->tsecr = tp->rx_opt.ts_recent;
 		size += TCPOLEN_TSTAMP_ALIGNED;
 	}
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 93825dd..0aa0d1d 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -712,7 +712,8 @@ static const struct tcp_request_sock_ops tcp_request_sock_ipv6_ops = {
 #endif
 
 static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
-				 u32 ts, struct tcp_md5sig_key *key, int rst, u8 tclass)
+				 u32 tsval, u32 tsecr,
+				 struct tcp_md5sig_key *key, int rst, u8 tclass)
 {
 	const struct tcphdr *th = tcp_hdr(skb);
 	struct tcphdr *t1;
@@ -724,7 +725,7 @@ static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
 	struct dst_entry *dst;
 	__be32 *topt;
 
-	if (ts)
+	if (tsecr)
 		tot_len += TCPOLEN_TSTAMP_ALIGNED;
 #ifdef CONFIG_TCP_MD5SIG
 	if (key)
@@ -754,11 +755,11 @@ static void tcp_v6_send_response(struct sk_buff *skb, u32 seq, u32 ack, u32 win,
 
 	topt = (__be32 *)(t1 + 1);
 
-	if (ts) {
+	if (tsecr) {
 		*topt++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) |
 				(TCPOPT_TIMESTAMP << 8) | TCPOLEN_TIMESTAMP);
-		*topt++ = htonl(tcp_time_stamp);
-		*topt++ = htonl(ts);
+		*topt++ = htonl(tsval);
+		*topt++ = htonl(tsecr);
 	}
 
 #ifdef CONFIG_TCP_MD5SIG
@@ -858,7 +859,7 @@ static void tcp_v6_send_reset(struct sock *sk, struct sk_buff *skb)
 		ack_seq = ntohl(th->seq) + th->syn + th->fin + skb->len -
 			  (th->doff << 2);
 
-	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, key, 1, 0);
+	tcp_v6_send_response(skb, seq, ack_seq, 0, 0, 0, key, 1, 0);
 
 #ifdef CONFIG_TCP_MD5SIG
 release_sk1:
@@ -869,10 +870,11 @@ release_sk1:
 #endif
 }
 
-static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack, u32 win, u32 ts,
+static void tcp_v6_send_ack(struct sk_buff *skb, u32 seq, u32 ack,
+			    u32 win, u32 tsval, u32 tsecr,
 			    struct tcp_md5sig_key *key, u8 tclass)
 {
-	tcp_v6_send_response(skb, seq, ack, win, ts, key, 0, tclass);
+	tcp_v6_send_response(skb, seq, ack, win, tsval, tsecr, key, 0, tclass);
 }
 
 static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
@@ -882,6 +884,7 @@ static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
 
 	tcp_v6_send_ack(skb, tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
 			tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
+			tcp_time_stamp + tcptw->tw_ts_offset,
 			tcptw->tw_ts_recent, tcp_twsk_md5_key(tcptw),
 			tw->tw_tclass);
 
@@ -891,7 +894,8 @@ static void tcp_v6_timewait_ack(struct sock *sk, struct sk_buff *skb)
 static void tcp_v6_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
 				  struct request_sock *req)
 {
-	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1, req->rcv_wnd, req->ts_recent,
+	tcp_v6_send_ack(skb, tcp_rsk(req)->snt_isn + 1, tcp_rsk(req)->rcv_isn + 1,
+			req->rcv_wnd, tcp_time_stamp, req->ts_recent,
 			tcp_v6_md5_do_lookup(sk, &ipv6_hdr(skb)->daddr), 0);
 }
 
-- 
1.7.11.7

Re: [PATCH net-next] tcp: add ability to set a timestamp offset

[David Miller]

From: Andrey Vagin <avagin@openvz.org>
Date: Mon, 11 Feb 2013 19:50:16 +0400

> If a TCP socket will get live-migrated from one box to another the
> timestamps (which are typically ON) will get screwed up -- the new
> kernel will generate TS values that has nothing to do with what they
> were on dump. The solution is to yet again fix the kernel and put a
> "timestamp offset" on a socket.
> 
> A socket offset is added in places where externally visible tcp
> timestamp option is parsed/initialized.
> 
> Connections in the SYN_RECV state are not supported, global
> tcp_time_stamp is used for them, because repair mode doesn't support
> this state. In a future it can be implemented by the similar way as for
> TIME_WAIT sockets.
> 
> For time-wait sockets offset is inhereted by a proper tcp_sock.
> 
> A per-socket offset can be set only for sockets in repair mode.

Series applied, thanks.