[PATCH net-next 00/15] tcp: move 14 sysctls to namespaces

[PATCH net-next 00/15] tcp: move 14 sysctls to namespaces

[Eric Dumazet]

Ideally all TCP sysctls should be per netns.
This patch series takes care of 14 of sysctls.
More to come later.

Eric Dumazet (15):
  tcp: Namespace-ify sysctl_tcp_early_retrans
  tcp: Namespace-ify sysctl_tcp_recovery
  tcp: Namespace-ify sysctl_tcp_thin_linear_timeouts
  tcp: Namespace-ify sysctl_tcp_slow_start_after_idle
  tcp: Namespace-ify sysctl_tcp_retrans_collapse
  tcp: Namespace-ify sysctl_tcp_stdurg
  tcp: Namespace-ify sysctl_tcp_rfc1337
  tcp: Namespace-ify sysctl_tcp_abort_on_overflow
  tcp: Namespace-ify sysctl_tcp_fack
  tcp: remove stale sysctl_tcp_reordering
  tcp: Namespace-ify sysctl_tcp_max_reordering
  tcp: Namespace-ify sysctl_tcp_dsack
  tcp: Namespace-ify sysctl_tcp_app_win
  tcp: Namespace-ify sysctl_tcp_adv_win_scale
  tcp: Namespace-ify sysctl_tcp_frto

 include/net/netns/ipv4.h   |  14 ++++
 include/net/tcp.h          |  27 ++----
 net/ipv4/sysctl_net_ipv4.c | 204 ++++++++++++++++++++++-----------------------
 net/ipv4/tcp.c             |   2 +-
 net/ipv4/tcp_input.c       |  42 ++++------
 net/ipv4/tcp_ipv4.c        |   9 ++
 net/ipv4/tcp_minisocks.c   |   8 +-
 net/ipv4/tcp_output.c      |  14 ++--
 net/ipv4/tcp_recovery.c    |   2 -
 net/ipv4/tcp_timer.c       |   4 +-
 10 files changed, 157 insertions(+), 169 deletions(-)

-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 01/15] tcp: Namespace-ify sysctl_tcp_early_retrans

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 18 +++++++++---------
 net/ipv4/tcp_input.c       |  1 -
 net/ipv4/tcp_ipv4.c        |  1 +
 net/ipv4/tcp_output.c      |  4 +++-
 6 files changed, 14 insertions(+), 12 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 2c4222a5d1025f5928665e10edb70fad65352dba..a7f39e3ea666a835b6042e4008c86ccaadd14b46 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -128,6 +128,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_sack;
 	int sysctl_tcp_window_scaling;
 	int sysctl_tcp_timestamps;
+	int sysctl_tcp_early_retrans;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 2392f74074e71af6a6f8113f1ba9c13c7ee5c1bc..862cecd9424b30e637cfb717db519e2fe95d2c8e 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -262,7 +262,6 @@ extern int sysctl_tcp_workaround_signed_windows;
 extern int sysctl_tcp_slow_start_after_idle;
 extern int sysctl_tcp_thin_linear_timeouts;
 extern int sysctl_tcp_thin_dupack;
-extern int sysctl_tcp_early_retrans;
 extern int sysctl_tcp_recovery;
 #define TCP_RACK_LOSS_DETECTION  0x1 /* Use RACK to detect losses */
 
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 81d218346cf7a0f340f964c434a21cace5c41fa0..f0f650f020afd535f41943c6c9fb1483be7cfb8d 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -634,15 +634,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode           = 0644,
 		.proc_handler   = proc_dointvec
 	},
-	{
-		.procname	= "tcp_early_retrans",
-		.data		= &sysctl_tcp_early_retrans,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1		= &zero,
-		.extra2		= &four,
-	},
 	{
 		.procname	= "tcp_min_tso_segs",
 		.data		= &sysctl_tcp_min_tso_segs,
@@ -1145,6 +1136,15 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_early_retrans",
+		.data		= &init_net.ipv4.sysctl_tcp_early_retrans,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &four,
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 893286db4623773afe75f4e0cfea7ad0d24ea2fe..91b4f14f71fb6c63c97dde8ba6a5af61c5a1fc82 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -93,7 +93,6 @@ int sysctl_tcp_max_orphans __read_mostly = NR_FILE;
 int sysctl_tcp_frto __read_mostly = 2;
 int sysctl_tcp_min_rtt_wlen __read_mostly = 300;
 int sysctl_tcp_moderate_rcvbuf __read_mostly = 1;
-int sysctl_tcp_early_retrans __read_mostly = 3;
 int sysctl_tcp_invalid_ratelimit __read_mostly = HZ/2;
 
 #define FLAG_DATA		0x01 /* Incoming frame contained data.		*/
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 23a8100af5ad399d4fa2568f4cac19192a008055..7ab313f6768e234173d78f17cfb1f664b230e958 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2484,6 +2484,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_sack = 1;
 	net->ipv4.sysctl_tcp_window_scaling = 1;
 	net->ipv4.sysctl_tcp_timestamps = 1;
+	net->ipv4.sysctl_tcp_early_retrans = 3;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 1f01f4c9c738484822aee7a6810d90857e6a72dc..dd7c22b7d23b016b1da83138d6e8047abdef97e6 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2378,6 +2378,7 @@ bool tcp_schedule_loss_probe(struct sock *sk)
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct tcp_sock *tp = tcp_sk(sk);
 	u32 timeout, rto_delta_us;
+	int early_retrans;
 
 	/* Don't do any loss probe on a Fast Open connection before 3WHS
 	 * finishes.
@@ -2385,10 +2386,11 @@ bool tcp_schedule_loss_probe(struct sock *sk)
 	if (tp->fastopen_rsk)
 		return false;
 
+	early_retrans = sock_net(sk)->ipv4.sysctl_tcp_early_retrans;
 	/* Schedule a loss probe in 2*RTT for SACK capable connections
 	 * in Open state, that are either limited by cwnd or application.
 	 */
-	if ((sysctl_tcp_early_retrans != 3 && sysctl_tcp_early_retrans != 4) ||
+	if ((early_retrans != 3 && early_retrans != 4) ||
 	    !tp->packets_out || !tcp_is_sack(tp) ||
 	    icsk->icsk_ca_state != TCP_CA_Open)
 		return false;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 02/15] tcp: Namespace-ify sysctl_tcp_recovery

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  2 +-
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  2 +-
 net/ipv4/tcp_ipv4.c        |  1 +
 net/ipv4/tcp_recovery.c    |  2 --
 6 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index a7f39e3ea666a835b6042e4008c86ccaadd14b46..d6ed718075d44cfc0e60995c1e938d588ad261a8 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -129,6 +129,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_window_scaling;
 	int sysctl_tcp_timestamps;
 	int sysctl_tcp_early_retrans;
+	int sysctl_tcp_recovery;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 862cecd9424b30e637cfb717db519e2fe95d2c8e..469b4954beee8571ba8d930e852db8b1fef5eed3 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -262,7 +262,7 @@ extern int sysctl_tcp_workaround_signed_windows;
 extern int sysctl_tcp_slow_start_after_idle;
 extern int sysctl_tcp_thin_linear_timeouts;
 extern int sysctl_tcp_thin_dupack;
-extern int sysctl_tcp_recovery;
+
 #define TCP_RACK_LOSS_DETECTION  0x1 /* Use RACK to detect losses */
 
 extern int sysctl_tcp_limit_output_bytes;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index f0f650f020afd535f41943c6c9fb1483be7cfb8d..78019adcae875a438264ee47723670f6b54cacf9 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -449,13 +449,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
-	{
-		.procname	= "tcp_recovery",
-		.data		= &sysctl_tcp_recovery,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec,
-	},
 	{
 		.procname	= "tcp_max_reordering",
 		.data		= &sysctl_tcp_max_reordering,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.extra1		= &zero,
 		.extra2		= &four,
 	},
+	{
+		.procname	= "tcp_recovery",
+		.data		= &init_net.ipv4.sysctl_tcp_recovery,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec,
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 91b4f14f71fb6c63c97dde8ba6a5af61c5a1fc82..2cc56fd57b751dd6b457c15067aa9309683a04a8 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2786,7 +2786,7 @@ static void tcp_rack_identify_loss(struct sock *sk, int *ack_flag)
 	struct tcp_sock *tp = tcp_sk(sk);
 
 	/* Use RACK to detect loss */
-	if (sysctl_tcp_recovery & TCP_RACK_LOSS_DETECTION) {
+	if (sock_net(sk)->ipv4.sysctl_tcp_recovery & TCP_RACK_LOSS_DETECTION) {
 		u32 prior_retrans = tp->retrans_out;
 
 		tcp_rack_mark_lost(sk);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 7ab313f6768e234173d78f17cfb1f664b230e958..517ff1948a71287b06ea0859e1f25a15119a3dd9 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2485,6 +2485,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_window_scaling = 1;
 	net->ipv4.sysctl_tcp_timestamps = 1;
 	net->ipv4.sysctl_tcp_early_retrans = 3;
+	net->ipv4.sysctl_tcp_recovery = TCP_RACK_LOSS_DETECTION;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
diff --git a/net/ipv4/tcp_recovery.c b/net/ipv4/tcp_recovery.c
index cda6074a429a24b7be600065d95600f4f9810ee4..d3603a9e24eae8649edd12d3f0678015b09b2037 100644
--- a/net/ipv4/tcp_recovery.c
+++ b/net/ipv4/tcp_recovery.c
@@ -1,8 +1,6 @@
 #include <linux/tcp.h>
 #include <net/tcp.h>
 
-int sysctl_tcp_recovery __read_mostly = TCP_RACK_LOSS_DETECTION;
-
 static void tcp_rack_mark_skb_lost(struct sock *sk, struct sk_buff *skb)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 03/15] tcp: Namespace-ify sysctl_tcp_thin_linear_timeouts

[Eric Dumazet]

Note that sysctl_tcp_thin_dupack was not used, I deleted it.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  2 --
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_timer.c       |  4 +---
 4 files changed, 9 insertions(+), 12 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index d6ed718075d44cfc0e60995c1e938d588ad261a8..2a9f37b39c45fe451e45025790a4e5c45ece5cbc 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -130,6 +130,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_timestamps;
 	int sysctl_tcp_early_retrans;
 	int sysctl_tcp_recovery;
+	int sysctl_tcp_thin_linear_timeouts;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 469b4954beee8571ba8d930e852db8b1fef5eed3..8ef12bfe1a3a7083f1bbc924bc43a022a4057768 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -260,8 +260,6 @@ extern int sysctl_tcp_moderate_rcvbuf;
 extern int sysctl_tcp_tso_win_divisor;
 extern int sysctl_tcp_workaround_signed_windows;
 extern int sysctl_tcp_slow_start_after_idle;
-extern int sysctl_tcp_thin_linear_timeouts;
-extern int sysctl_tcp_thin_dupack;
 
 #define TCP_RACK_LOSS_DETECTION  0x1 /* Use RACK to detect losses */
 
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 78019adcae875a438264ee47723670f6b54cacf9..12003214f4d80b38d5f754ddd91be8a990168ade 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -620,13 +620,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler   = proc_allowed_congestion_control,
 	},
-	{
-		.procname       = "tcp_thin_linear_timeouts",
-		.data           = &sysctl_tcp_thin_linear_timeouts,
-		.maxlen         = sizeof(int),
-		.mode           = 0644,
-		.proc_handler   = proc_dointvec
-	},
 	{
 		.procname	= "tcp_min_tso_segs",
 		.data		= &sysctl_tcp_min_tso_segs,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec,
 	},
+	{
+		.procname       = "tcp_thin_linear_timeouts",
+		.data           = &init_net.ipv4.sysctl_tcp_thin_linear_timeouts,
+		.maxlen         = sizeof(int),
+		.mode           = 0644,
+		.proc_handler   = proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 804a8d34ce86cc17472c918c00c25de88b85184f..035a1ef1f2d8462c1d19f364b599ffac538ef688 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -22,8 +22,6 @@
 #include <linux/gfp.h>
 #include <net/tcp.h>
 
-int sysctl_tcp_thin_linear_timeouts __read_mostly;
-
 /**
  *  tcp_write_err() - close socket and save error info
  *  @sk:  The socket the error has appeared on.
@@ -522,7 +520,7 @@ void tcp_retransmit_timer(struct sock *sk)
 	 * linear-timeout retransmissions into a black hole
 	 */
 	if (sk->sk_state == TCP_ESTABLISHED &&
-	    (tp->thin_lto || sysctl_tcp_thin_linear_timeouts) &&
+	    (tp->thin_lto || net->ipv4.sysctl_tcp_thin_linear_timeouts) &&
 	    tcp_stream_is_thin(tp) &&
 	    icsk->icsk_retransmits <= TCP_THIN_LINEAR_RETRIES) {
 		icsk->icsk_backoff = 0;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 04/15] tcp: Namespace-ify sysctl_tcp_slow_start_after_idle

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  3 +--
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_ipv4.c        |  1 +
 net/ipv4/tcp_output.c      |  5 +----
 5 files changed, 11 insertions(+), 13 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 2a9f37b39c45fe451e45025790a4e5c45ece5cbc..8662692686b3af98a94a176230b9ed147881d87a 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -131,6 +131,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_early_retrans;
 	int sysctl_tcp_recovery;
 	int sysctl_tcp_thin_linear_timeouts;
+	int sysctl_tcp_slow_start_after_idle;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 8ef12bfe1a3a7083f1bbc924bc43a022a4057768..09c79705a0742aa8c22b3b7795d01b6c685d32e2 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -259,7 +259,6 @@ extern int sysctl_tcp_nometrics_save;
 extern int sysctl_tcp_moderate_rcvbuf;
 extern int sysctl_tcp_tso_win_divisor;
 extern int sysctl_tcp_workaround_signed_windows;
-extern int sysctl_tcp_slow_start_after_idle;
 
 #define TCP_RACK_LOSS_DETECTION  0x1 /* Use RACK to detect losses */
 
@@ -1305,7 +1304,7 @@ static inline void tcp_slow_start_after_idle_check(struct sock *sk)
 	struct tcp_sock *tp = tcp_sk(sk);
 	s32 delta;
 
-	if (!sysctl_tcp_slow_start_after_idle || tp->packets_out ||
+	if (!sock_net(sk)->ipv4.sysctl_tcp_slow_start_after_idle || tp->packets_out ||
 	    ca_ops->cong_control)
 		return;
 	delta = tcp_jiffies32 - tp->lsndtime;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 12003214f4d80b38d5f754ddd91be8a990168ade..40d69af8b363bc236e23879973872d8f9346d85e 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -571,13 +571,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
-	{
-		.procname	= "tcp_slow_start_after_idle",
-		.data		= &sysctl_tcp_slow_start_after_idle,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 #ifdef CONFIG_NETLABEL
 	{
 		.procname	= "cipso_cache_enable",
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode           = 0644,
 		.proc_handler   = proc_dointvec
 	},
+	{
+		.procname	= "tcp_slow_start_after_idle",
+		.data		= &init_net.ipv4.sysctl_tcp_slow_start_after_idle,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 517ff1948a71287b06ea0859e1f25a15119a3dd9..cea63a4b59655823def7a423d27191003c7f084c 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2486,6 +2486,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_timestamps = 1;
 	net->ipv4.sysctl_tcp_early_retrans = 3;
 	net->ipv4.sysctl_tcp_recovery = TCP_RACK_LOSS_DETECTION;
+	net->ipv4.sysctl_tcp_slow_start_after_idle = 1; /* By default, RFC2861 behavior.  */
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index dd7c22b7d23b016b1da83138d6e8047abdef97e6..bc93a346e6e91289c2aeb2f2d2522b809da12dd6 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -61,9 +61,6 @@ int sysctl_tcp_limit_output_bytes __read_mostly = 262144;
  */
 int sysctl_tcp_tso_win_divisor __read_mostly = 3;
 
-/* By default, RFC2861 behavior.  */
-int sysctl_tcp_slow_start_after_idle __read_mostly = 1;
-
 static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle,
 			   int push_one, gfp_t gfp);
 
@@ -1633,7 +1630,7 @@ static void tcp_cwnd_validate(struct sock *sk, bool is_cwnd_limited)
 		if (tp->packets_out > tp->snd_cwnd_used)
 			tp->snd_cwnd_used = tp->packets_out;
 
-		if (sysctl_tcp_slow_start_after_idle &&
+		if (sock_net(sk)->ipv4.sysctl_tcp_slow_start_after_idle &&
 		    (s32)(tcp_jiffies32 - tp->snd_cwnd_stamp) >= inet_csk(sk)->icsk_rto &&
 		    !ca_ops->cong_control)
 			tcp_cwnd_application_limited(sk);
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 05/15] tcp: Namespace-ify sysctl_tcp_retrans_collapse

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_ipv4.c        |  2 +-
 net/ipv4/tcp_output.c      |  5 +----
 5 files changed, 10 insertions(+), 13 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 8662692686b3af98a94a176230b9ed147881d87a..b28c172b10e497f235b51aae0fc2d3bbf7cc51f3 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -132,6 +132,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_recovery;
 	int sysctl_tcp_thin_linear_timeouts;
 	int sysctl_tcp_slow_start_after_idle;
+	int sysctl_tcp_retrans_collapse;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 09c79705a0742aa8c22b3b7795d01b6c685d32e2..085848e4de38058bb09f025387c713ade32b263e 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -240,7 +240,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 
 /* sysctl variables for tcp */
-extern int sysctl_tcp_retrans_collapse;
 extern int sysctl_tcp_stdurg;
 extern int sysctl_tcp_rfc1337;
 extern int sysctl_tcp_abort_on_overflow;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 40d69af8b363bc236e23879973872d8f9346d85e..533b92ad39dd0cada542028fe2f276d9eebcd2c8 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -386,13 +386,6 @@ static int proc_tcp_available_ulp(struct ctl_table *ctl,
 }
 
 static struct ctl_table ipv4_table[] = {
-	{
-		.procname	= "tcp_retrans_collapse",
-		.data		= &sysctl_tcp_retrans_collapse,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_max_orphans",
 		.data		= &sysctl_tcp_max_orphans,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_retrans_collapse",
+		.data		= &init_net.ipv4.sysctl_tcp_retrans_collapse,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index cea63a4b59655823def7a423d27191003c7f084c..2bc6ba2059d32aa848dbc415b4b0e194b61b0268 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2487,7 +2487,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_early_retrans = 3;
 	net->ipv4.sysctl_tcp_recovery = TCP_RACK_LOSS_DETECTION;
 	net->ipv4.sysctl_tcp_slow_start_after_idle = 1; /* By default, RFC2861 behavior.  */
-
+	net->ipv4.sysctl_tcp_retrans_collapse = 1;
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
 	net->ipv4.sysctl_tcp_fastopen_blackhole_timeout = 60 * 60;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index bc93a346e6e91289c2aeb2f2d2522b809da12dd6..735fff44aaca3d5afbb0ac55ebdb9898b6c44ae6 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -44,9 +44,6 @@
 
 #include <trace/events/tcp.h>
 
-/* People can turn this off for buggy TCP's found in printers etc. */
-int sysctl_tcp_retrans_collapse __read_mostly = 1;
-
 /* People can turn this on to work with those rare, broken TCPs that
  * interpret the window field as a signed quantity.
  */
@@ -2747,7 +2744,7 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *to,
 	struct sk_buff *skb = to, *tmp;
 	bool first = true;
 
-	if (!sysctl_tcp_retrans_collapse)
+	if (!sock_net(sk)->ipv4.sysctl_tcp_retrans_collapse)
 		return;
 	if (TCP_SKB_CB(skb)->tcp_flags & TCPHDR_SYN)
 		return;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 06/15] tcp: Namespace-ify sysctl_tcp_stdurg

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  3 +--
 4 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index b28c172b10e497f235b51aae0fc2d3bbf7cc51f3..ffa2cf3dc747ca9443df3927dc7928c18357f872 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -133,6 +133,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_thin_linear_timeouts;
 	int sysctl_tcp_slow_start_after_idle;
 	int sysctl_tcp_retrans_collapse;
+	int sysctl_tcp_stdurg;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 085848e4de38058bb09f025387c713ade32b263e..0cf9de8506916c67369ce78833207ba648f34a10 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -240,7 +240,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 
 /* sysctl variables for tcp */
-extern int sysctl_tcp_stdurg;
 extern int sysctl_tcp_rfc1337;
 extern int sysctl_tcp_abort_on_overflow;
 extern int sysctl_tcp_max_orphans;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 533b92ad39dd0cada542028fe2f276d9eebcd2c8..a34bb75815c15afc077ba7ff36939b5abc9229f6 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -400,13 +400,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
-	{
-		.procname	= "tcp_stdurg",
-		.data		= &sysctl_tcp_stdurg,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_rfc1337",
 		.data		= &sysctl_tcp_rfc1337,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_stdurg",
+		.data		= &init_net.ipv4.sysctl_tcp_stdurg,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 2cc56fd57b751dd6b457c15067aa9309683a04a8..6c3655b538f6b2315af7dc611acc574f7489bde6 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -87,7 +87,6 @@ EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
 /* rfc5961 challenge ack rate limiting */
 int sysctl_tcp_challenge_ack_limit = 1000;
 
-int sysctl_tcp_stdurg __read_mostly;
 int sysctl_tcp_rfc1337 __read_mostly;
 int sysctl_tcp_max_orphans __read_mostly = NR_FILE;
 int sysctl_tcp_frto __read_mostly = 2;
@@ -5103,7 +5102,7 @@ static void tcp_check_urg(struct sock *sk, const struct tcphdr *th)
 	struct tcp_sock *tp = tcp_sk(sk);
 	u32 ptr = ntohs(th->urg_ptr);
 
-	if (ptr && !sysctl_tcp_stdurg)
+	if (ptr && !sock_net(sk)->ipv4.sysctl_tcp_stdurg)
 		ptr--;
 	ptr += ntohl(th->seq);
 
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 07/15] tcp: Namespace-ify sysctl_tcp_rfc1337

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  1 -
 net/ipv4/tcp_minisocks.c   |  2 +-
 5 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index ffa2cf3dc747ca9443df3927dc7928c18357f872..968edce38eb5d3399724b3142277eab44f19f2fb 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -134,6 +134,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_slow_start_after_idle;
 	int sysctl_tcp_retrans_collapse;
 	int sysctl_tcp_stdurg;
+	int sysctl_tcp_rfc1337;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 0cf9de8506916c67369ce78833207ba648f34a10..7f88987bc62dd76206c15eb91f2990d4469e5421 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -240,7 +240,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 
 /* sysctl variables for tcp */
-extern int sysctl_tcp_rfc1337;
 extern int sysctl_tcp_abort_on_overflow;
 extern int sysctl_tcp_max_orphans;
 extern int sysctl_tcp_fack;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index a34bb75815c15afc077ba7ff36939b5abc9229f6..832e554235df37770809541ad8f9f1ca2f201739 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -400,13 +400,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
-	{
-		.procname	= "tcp_rfc1337",
-		.data		= &sysctl_tcp_rfc1337,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "inet_peer_threshold",
 		.data		= &inet_peer_threshold,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_rfc1337",
+		.data		= &init_net.ipv4.sysctl_tcp_rfc1337,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 6c3655b538f6b2315af7dc611acc574f7489bde6..d2d3f62387a98d7f955f8c9e27320b9722035b2a 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -87,7 +87,6 @@ EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
 /* rfc5961 challenge ack rate limiting */
 int sysctl_tcp_challenge_ack_limit = 1000;
 
-int sysctl_tcp_rfc1337 __read_mostly;
 int sysctl_tcp_max_orphans __read_mostly = NR_FILE;
 int sysctl_tcp_frto __read_mostly = 2;
 int sysctl_tcp_min_rtt_wlen __read_mostly = 300;
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index a952357054f4ddfbb98746ebcf323d1c45f7e951..2abaa4c1fe0108f2645d8e783ae6b48e87a82fb3 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -180,7 +180,7 @@ tcp_timewait_state_process(struct inet_timewait_sock *tw, struct sk_buff *skb,
 			 * Oh well... nobody has a sufficient solution to this
 			 * protocol bug yet.
 			 */
-			if (sysctl_tcp_rfc1337 == 0) {
+			if (twsk_net(tw)->ipv4.sysctl_tcp_rfc1337 == 0) {
 kill:
 				inet_twsk_deschedule_put(tw);
 				return TCP_TW_SUCCESS;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 08/15] tcp: Namespace-ify sysctl_tcp_abort_on_overflow

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_minisocks.c   |  4 +---
 4 files changed, 9 insertions(+), 11 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 968edce38eb5d3399724b3142277eab44f19f2fb..3875fdf6b18653477408beb25176eac849e65ba4 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -135,6 +135,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_retrans_collapse;
 	int sysctl_tcp_stdurg;
 	int sysctl_tcp_rfc1337;
+	int sysctl_tcp_abort_on_overflow;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 7f88987bc62dd76206c15eb91f2990d4469e5421..e28471ce52bd815346676931075588d59306a441 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -240,7 +240,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 
 /* sysctl variables for tcp */
-extern int sysctl_tcp_abort_on_overflow;
 extern int sysctl_tcp_max_orphans;
 extern int sysctl_tcp_fack;
 extern int sysctl_tcp_reordering;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 832e554235df37770809541ad8f9f1ca2f201739..ffd1fd769bba7c3524aa6dfac734e1de0cad1506 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -393,13 +393,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
-	{
-		.procname	= "tcp_abort_on_overflow",
-		.data		= &sysctl_tcp_abort_on_overflow,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "inet_peer_threshold",
 		.data		= &inet_peer_threshold,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_abort_on_overflow",
+		.data		= &init_net.ipv4.sysctl_tcp_abort_on_overflow,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 2abaa4c1fe0108f2645d8e783ae6b48e87a82fb3..a79a89fdb558a5d66ec5241fdb8bfcab196c744d 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -28,8 +28,6 @@
 #include <net/xfrm.h>
 #include <net/busy_poll.h>
 
-int sysctl_tcp_abort_on_overflow __read_mostly;
-
 static bool tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
 {
 	if (seq == s_win)
@@ -764,7 +762,7 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
 	return inet_csk_complete_hashdance(sk, child, req, own_req);
 
 listen_overflow:
-	if (!sysctl_tcp_abort_on_overflow) {
+	if (!sock_net(sk)->ipv4.sysctl_tcp_abort_on_overflow) {
 		inet_rsk(req)->acked = 1;
 		return NULL;
 	}
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 09/15] tcp: Namespace-ify sysctl_tcp_fack

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp.c             |  2 +-
 net/ipv4/tcp_input.c       |  3 +--
 net/ipv4/tcp_minisocks.c   |  2 +-
 6 files changed, 11 insertions(+), 12 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 3875fdf6b18653477408beb25176eac849e65ba4..f0e792beeea974b0850090d7624a3d7490124067 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -136,6 +136,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_stdurg;
 	int sysctl_tcp_rfc1337;
 	int sysctl_tcp_abort_on_overflow;
+	int sysctl_tcp_fack;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index e28471ce52bd815346676931075588d59306a441..38504d5ab109454219ac9570c3b11e02733384c1 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -241,7 +241,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 /* sysctl variables for tcp */
 extern int sysctl_tcp_max_orphans;
-extern int sysctl_tcp_fack;
 extern int sysctl_tcp_reordering;
 extern int sysctl_tcp_max_reordering;
 extern int sysctl_tcp_dsack;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index ffd1fd769bba7c3524aa6dfac734e1de0cad1506..1f23be13ce7be8b2a12b82aada36c6351fdfb70a 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -414,13 +414,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec_jiffies,
 	},
-	{
-		.procname	= "tcp_fack",
-		.data		= &sysctl_tcp_fack,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_max_reordering",
 		.data		= &sysctl_tcp_max_reordering,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_fack",
+		.data		= &init_net.ipv4.sysctl_tcp_fack,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 8f36277e82e9dbea750ce66b73018a81b30b5156..4a777ba113b9afe118e3020da65878d85848e1cb 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2511,7 +2511,7 @@ static int tcp_repair_options_est(struct sock *sk,
 				return -EINVAL;
 
 			tp->rx_opt.sack_ok |= TCP_SACK_SEEN;
-			if (sysctl_tcp_fack)
+			if (sock_net(sk)->ipv4.sysctl_tcp_fack)
 				tcp_enable_fack(tp);
 			break;
 		case TCPOPT_TIMESTAMP:
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index d2d3f62387a98d7f955f8c9e27320b9722035b2a..8941fc32072b69fedcb01afbe837f4d7791dd28d 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -77,7 +77,6 @@
 #include <linux/errqueue.h>
 #include <trace/events/tcp.h>
 
-int sysctl_tcp_fack __read_mostly;
 int sysctl_tcp_max_reordering __read_mostly = 300;
 int sysctl_tcp_dsack __read_mostly = 1;
 int sysctl_tcp_app_win __read_mostly = 31;
@@ -5690,7 +5689,7 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
 			tp->tcp_header_len = sizeof(struct tcphdr);
 		}
 
-		if (tcp_is_sack(tp) && sysctl_tcp_fack)
+		if (tcp_is_sack(tp) && sock_net(sk)->ipv4.sysctl_tcp_fack)
 			tcp_enable_fack(tp);
 
 		tcp_sync_mss(sk, icsk->icsk_pmtu_cookie);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index a79a89fdb558a5d66ec5241fdb8bfcab196c744d..eba61f77bc36e4f49580d15840c15af565b8b479 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -491,7 +491,7 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
 
 		newtp->rx_opt.tstamp_ok = ireq->tstamp_ok;
 		if ((newtp->rx_opt.sack_ok = ireq->sack_ok) != 0) {
-			if (sysctl_tcp_fack)
+			if (sock_net(sk)->ipv4.sysctl_tcp_fack)
 				tcp_enable_fack(newtp);
 		}
 		newtp->window_clamp = req->rsk_window_clamp;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 10/15] tcp: remove stale sysctl_tcp_reordering

[Eric Dumazet]

This extern is no longer used.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/tcp.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/include/net/tcp.h b/include/net/tcp.h
index 38504d5ab109454219ac9570c3b11e02733384c1..c912d63839e7cd4b6ad009344e8017de4c0b1483 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -241,7 +241,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 /* sysctl variables for tcp */
 extern int sysctl_tcp_max_orphans;
-extern int sysctl_tcp_reordering;
 extern int sysctl_tcp_max_reordering;
 extern int sysctl_tcp_dsack;
 extern long sysctl_tcp_mem[3];
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 11/15] tcp: Namespace-ify sysctl_tcp_max_reordering

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  3 +--
 net/ipv4/tcp_ipv4.c        |  2 ++
 5 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index f0e792beeea974b0850090d7624a3d7490124067..3f6844665a2fbe66fc0c91bd13e057ac2e03007a 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -137,6 +137,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_rfc1337;
 	int sysctl_tcp_abort_on_overflow;
 	int sysctl_tcp_fack;
+	int sysctl_tcp_max_reordering;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index c912d63839e7cd4b6ad009344e8017de4c0b1483..2b559c7bf16c70864e77a34c78479d01f538d6cd 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -241,7 +241,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 /* sysctl variables for tcp */
 extern int sysctl_tcp_max_orphans;
-extern int sysctl_tcp_max_reordering;
 extern int sysctl_tcp_dsack;
 extern long sysctl_tcp_mem[3];
 extern int sysctl_tcp_wmem[3];
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 1f23be13ce7be8b2a12b82aada36c6351fdfb70a..18cd228a20690541936dd6b3d9bb02cb283a9740 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -414,13 +414,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec_jiffies,
 	},
-	{
-		.procname	= "tcp_max_reordering",
-		.data		= &sysctl_tcp_max_reordering,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_dsack",
 		.data		= &sysctl_tcp_dsack,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_max_reordering",
+		.data		= &init_net.ipv4.sysctl_tcp_max_reordering,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 8941fc32072b69fedcb01afbe837f4d7791dd28d..bd6abf9a6d5a0f7a85384a259d61a34c4170eb50 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -77,7 +77,6 @@
 #include <linux/errqueue.h>
 #include <trace/events/tcp.h>
 
-int sysctl_tcp_max_reordering __read_mostly = 300;
 int sysctl_tcp_dsack __read_mostly = 1;
 int sysctl_tcp_app_win __read_mostly = 31;
 int sysctl_tcp_adv_win_scale __read_mostly = 1;
@@ -887,7 +886,7 @@ static void tcp_update_reordering(struct sock *sk, const int metric,
 		return;
 
 	if (metric > tp->reordering) {
-		tp->reordering = min(sysctl_tcp_max_reordering, metric);
+		tp->reordering = min(sock_net(sk)->ipv4.sysctl_tcp_max_reordering, metric);
 
 #if FASTRETRANS_DEBUG > 1
 		pr_debug("Disorder%d %d %u f%u s%u rr%d\n",
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 2bc6ba2059d32aa848dbc415b4b0e194b61b0268..c379a242abb3546044da9a3ef032f6f68acafe88 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2488,6 +2488,8 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_recovery = TCP_RACK_LOSS_DETECTION;
 	net->ipv4.sysctl_tcp_slow_start_after_idle = 1; /* By default, RFC2861 behavior.  */
 	net->ipv4.sysctl_tcp_retrans_collapse = 1;
+	net->ipv4.sysctl_tcp_max_reordering = 300;
+
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
 	net->ipv4.sysctl_tcp_fastopen_blackhole_timeout = 60 * 60;
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 12/15] tcp: Namespace-ify sysctl_tcp_dsack

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  5 ++---
 net/ipv4/tcp_ipv4.c        |  1 +
 5 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 3f6844665a2fbe66fc0c91bd13e057ac2e03007a..956957a77db96ad3d231cc018c13503d615d8d2e 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -138,6 +138,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_abort_on_overflow;
 	int sysctl_tcp_fack;
 	int sysctl_tcp_max_reordering;
+	int sysctl_tcp_dsack;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 2b559c7bf16c70864e77a34c78479d01f538d6cd..aba20c4828ee912d9f0f3ef49f3de5376729c022 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -241,7 +241,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 
 /* sysctl variables for tcp */
 extern int sysctl_tcp_max_orphans;
-extern int sysctl_tcp_dsack;
 extern long sysctl_tcp_mem[3];
 extern int sysctl_tcp_wmem[3];
 extern int sysctl_tcp_rmem[3];
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 18cd228a20690541936dd6b3d9bb02cb283a9740..7652a9c2a65d3f1cfa0a75d1198e1d9d56761c35 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -414,13 +414,6 @@ static struct ctl_table ipv4_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec_jiffies,
 	},
-	{
-		.procname	= "tcp_dsack",
-		.data		= &sysctl_tcp_dsack,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_mem",
 		.maxlen		= sizeof(sysctl_tcp_mem),
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_dsack",
+		.data		= &init_net.ipv4.sysctl_tcp_dsack,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index bd6abf9a6d5a0f7a85384a259d61a34c4170eb50..db4798f99bb093b6d5a3e0fdd76efb83b88da49e 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -77,7 +77,6 @@
 #include <linux/errqueue.h>
 #include <trace/events/tcp.h>
 
-int sysctl_tcp_dsack __read_mostly = 1;
 int sysctl_tcp_app_win __read_mostly = 31;
 int sysctl_tcp_adv_win_scale __read_mostly = 1;
 EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
@@ -4130,7 +4129,7 @@ static void tcp_dsack_set(struct sock *sk, u32 seq, u32 end_seq)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 
-	if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
+	if (tcp_is_sack(tp) && sock_net(sk)->ipv4.sysctl_tcp_dsack) {
 		int mib_idx;
 
 		if (before(seq, tp->rcv_nxt))
@@ -4165,7 +4164,7 @@ static void tcp_send_dupack(struct sock *sk, const struct sk_buff *skb)
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_DELAYEDACKLOST);
 		tcp_enter_quickack_mode(sk);
 
-		if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
+		if (tcp_is_sack(tp) && sock_net(sk)->ipv4.sysctl_tcp_dsack) {
 			u32 end_seq = TCP_SKB_CB(skb)->end_seq;
 
 			if (after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt))
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index c379a242abb3546044da9a3ef032f6f68acafe88..d9d4d191e8f3c962a6ee68015ffe5a6e7fb8e9c1 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2489,6 +2489,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_slow_start_after_idle = 1; /* By default, RFC2861 behavior.  */
 	net->ipv4.sysctl_tcp_retrans_collapse = 1;
 	net->ipv4.sysctl_tcp_max_reordering = 300;
+	net->ipv4.sysctl_tcp_dsack = 1;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 13/15] tcp: Namespace-ify sysctl_tcp_app_win

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  8 ++++----
 net/ipv4/tcp_ipv4.c        |  1 +
 5 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 956957a77db96ad3d231cc018c13503d615d8d2e..63f91d52cbc0ad35d8e04a8da0d9f57aa960bcb0 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -139,6 +139,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_fack;
 	int sysctl_tcp_max_reordering;
 	int sysctl_tcp_dsack;
+	int sysctl_tcp_app_win;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index aba20c4828ee912d9f0f3ef49f3de5376729c022..c6bee85a3dec0dea6d4402d89184ade02a637a2e 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -244,7 +244,6 @@ extern int sysctl_tcp_max_orphans;
 extern long sysctl_tcp_mem[3];
 extern int sysctl_tcp_wmem[3];
 extern int sysctl_tcp_rmem[3];
-extern int sysctl_tcp_app_win;
 extern int sysctl_tcp_adv_win_scale;
 extern int sysctl_tcp_frto;
 extern int sysctl_tcp_nometrics_save;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 7652a9c2a65d3f1cfa0a75d1198e1d9d56761c35..e057788834a99cf99e141a602ddbe19b8e6fce3c 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -437,13 +437,6 @@ static struct ctl_table ipv4_table[] = {
 		.proc_handler	= proc_dointvec_minmax,
 		.extra1		= &one,
 	},
-	{
-		.procname	= "tcp_app_win",
-		.data		= &sysctl_tcp_app_win,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_adv_win_scale",
 		.data		= &sysctl_tcp_adv_win_scale,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_app_win",
+		.data		= &init_net.ipv4.sysctl_tcp_app_win,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index db4798f99bb093b6d5a3e0fdd76efb83b88da49e..06a8c27e1a690e3b26cb6773320bafa31b06d3b3 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -77,7 +77,6 @@
 #include <linux/errqueue.h>
 #include <trace/events/tcp.h>
 
-int sysctl_tcp_app_win __read_mostly = 31;
 int sysctl_tcp_adv_win_scale __read_mostly = 1;
 EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
 
@@ -426,6 +425,7 @@ static void tcp_fixup_rcvbuf(struct sock *sk)
  */
 void tcp_init_buffer_space(struct sock *sk)
 {
+	int tcp_app_win = sock_net(sk)->ipv4.sysctl_tcp_app_win;
 	struct tcp_sock *tp = tcp_sk(sk);
 	int maxwin;
 
@@ -444,14 +444,14 @@ void tcp_init_buffer_space(struct sock *sk)
 	if (tp->window_clamp >= maxwin) {
 		tp->window_clamp = maxwin;
 
-		if (sysctl_tcp_app_win && maxwin > 4 * tp->advmss)
+		if (tcp_app_win && maxwin > 4 * tp->advmss)
 			tp->window_clamp = max(maxwin -
-					       (maxwin >> sysctl_tcp_app_win),
+					       (maxwin >> tcp_app_win),
 					       4 * tp->advmss);
 	}
 
 	/* Force reservation of one segment. */
-	if (sysctl_tcp_app_win &&
+	if (tcp_app_win &&
 	    tp->window_clamp > 2 * tp->advmss &&
 	    tp->window_clamp + tp->advmss > maxwin)
 		tp->window_clamp = max(2 * tp->advmss, maxwin - tp->advmss);
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index d9d4d191e8f3c962a6ee68015ffe5a6e7fb8e9c1..189664ebd28e4cda7ef40a47591c3bd8cac3574b 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2490,6 +2490,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_retrans_collapse = 1;
 	net->ipv4.sysctl_tcp_max_reordering = 300;
 	net->ipv4.sysctl_tcp_dsack = 1;
+	net->ipv4.sysctl_tcp_app_win = 31;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 14/15] tcp: Namespace-ify sysctl_tcp_adv_win_scale

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  9 ++++-----
 net/ipv4/sysctl_net_ipv4.c | 18 +++++++++---------
 net/ipv4/tcp_input.c       | 13 +++++--------
 net/ipv4/tcp_ipv4.c        |  1 +
 5 files changed, 20 insertions(+), 22 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 63f91d52cbc0ad35d8e04a8da0d9f57aa960bcb0..9dbb07d4eff465428817831e55c6a4922b7208fb 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -140,6 +140,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_max_reordering;
 	int sysctl_tcp_dsack;
 	int sysctl_tcp_app_win;
+	int sysctl_tcp_adv_win_scale;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index c6bee85a3dec0dea6d4402d89184ade02a637a2e..2572b57682987dd5f3700ed47d63e7238946b9a8 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -244,7 +244,6 @@ extern int sysctl_tcp_max_orphans;
 extern long sysctl_tcp_mem[3];
 extern int sysctl_tcp_wmem[3];
 extern int sysctl_tcp_rmem[3];
-extern int sysctl_tcp_adv_win_scale;
 extern int sysctl_tcp_frto;
 extern int sysctl_tcp_nometrics_save;
 extern int sysctl_tcp_moderate_rcvbuf;
@@ -1308,9 +1307,9 @@ void tcp_select_initial_window(int __space, __u32 mss, __u32 *rcv_wnd,
 			       __u32 *window_clamp, int wscale_ok,
 			       __u8 *rcv_wscale, __u32 init_rcv_wnd);
 
-static inline int tcp_win_from_space(int space)
+static inline int tcp_win_from_space(const struct sock *sk, int space)
 {
-	int tcp_adv_win_scale = sysctl_tcp_adv_win_scale;
+	int tcp_adv_win_scale = sock_net(sk)->ipv4.sysctl_tcp_adv_win_scale;
 
 	return tcp_adv_win_scale <= 0 ?
 		(space>>(-tcp_adv_win_scale)) :
@@ -1320,13 +1319,13 @@ static inline int tcp_win_from_space(int space)
 /* Note: caller must be prepared to deal with negative returns */
 static inline int tcp_space(const struct sock *sk)
 {
-	return tcp_win_from_space(sk->sk_rcvbuf -
+	return tcp_win_from_space(sk, sk->sk_rcvbuf -
 				  atomic_read(&sk->sk_rmem_alloc));
 }
 
 static inline int tcp_full_space(const struct sock *sk)
 {
-	return tcp_win_from_space(sk->sk_rcvbuf);
+	return tcp_win_from_space(sk, sk->sk_rcvbuf);
 }
 
 extern void tcp_openreq_init_rwin(struct request_sock *req,
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index e057788834a99cf99e141a602ddbe19b8e6fce3c..a95123e1e7da706c88bf5553b7d8ef6c2653ab50 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -437,15 +437,6 @@ static struct ctl_table ipv4_table[] = {
 		.proc_handler	= proc_dointvec_minmax,
 		.extra1		= &one,
 	},
-	{
-		.procname	= "tcp_adv_win_scale",
-		.data		= &sysctl_tcp_adv_win_scale,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec_minmax,
-		.extra1		= &tcp_adv_win_scale_min,
-		.extra2		= &tcp_adv_win_scale_max,
-	},
 	{
 		.procname	= "tcp_frto",
 		.data		= &sysctl_tcp_frto,
@@ -1145,6 +1136,15 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec
 	},
+	{
+		.procname	= "tcp_adv_win_scale",
+		.data		= &init_net.ipv4.sysctl_tcp_adv_win_scale,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &tcp_adv_win_scale_min,
+		.extra2		= &tcp_adv_win_scale_max,
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 06a8c27e1a690e3b26cb6773320bafa31b06d3b3..d01f390da23dcd4100271b150bd8bc143f7328cf 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -77,9 +77,6 @@
 #include <linux/errqueue.h>
 #include <trace/events/tcp.h>
 
-int sysctl_tcp_adv_win_scale __read_mostly = 1;
-EXPORT_SYMBOL(sysctl_tcp_adv_win_scale);
-
 /* rfc5961 challenge ack rate limiting */
 int sysctl_tcp_challenge_ack_limit = 1000;
 
@@ -361,8 +358,8 @@ static int __tcp_grow_window(const struct sock *sk, const struct sk_buff *skb)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
 	/* Optimize this! */
-	int truesize = tcp_win_from_space(skb->truesize) >> 1;
-	int window = tcp_win_from_space(sysctl_tcp_rmem[2]) >> 1;
+	int truesize = tcp_win_from_space(sk, skb->truesize) >> 1;
+	int window = tcp_win_from_space(sk, sysctl_tcp_rmem[2]) >> 1;
 
 	while (tp->rcv_ssthresh <= window) {
 		if (truesize <= skb->len)
@@ -387,7 +384,7 @@ static void tcp_grow_window(struct sock *sk, const struct sk_buff *skb)
 		/* Check #2. Increase window, if skb with such overhead
 		 * will fit to rcvbuf in future.
 		 */
-		if (tcp_win_from_space(skb->truesize) <= skb->len)
+		if (tcp_win_from_space(sk, skb->truesize) <= skb->len)
 			incr = 2 * tp->advmss;
 		else
 			incr = __tcp_grow_window(sk, skb);
@@ -628,7 +625,7 @@ void tcp_rcv_space_adjust(struct sock *sk)
 		}
 
 		rcvmem = SKB_TRUESIZE(tp->advmss + MAX_TCP_HEADER);
-		while (tcp_win_from_space(rcvmem) < tp->advmss)
+		while (tcp_win_from_space(sk, rcvmem) < tp->advmss)
 			rcvmem += 128;
 
 		rcvbuf = min(rcvwin / tp->advmss * rcvmem, sysctl_tcp_rmem[2]);
@@ -4789,7 +4786,7 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct rb_root *root,
 		 *   overlaps to the next one.
 		 */
 		if (!(TCP_SKB_CB(skb)->tcp_flags & (TCPHDR_SYN | TCPHDR_FIN)) &&
-		    (tcp_win_from_space(skb->truesize) > skb->len ||
+		    (tcp_win_from_space(sk, skb->truesize) > skb->len ||
 		     before(TCP_SKB_CB(skb)->seq, start))) {
 			end_of_skbs = false;
 			break;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 189664ebd28e4cda7ef40a47591c3bd8cac3574b..1fe30fb99308b3e3fd07509b509b0e3727cc5d44 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2491,6 +2491,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_max_reordering = 300;
 	net->ipv4.sysctl_tcp_dsack = 1;
 	net->ipv4.sysctl_tcp_app_win = 31;
+	net->ipv4.sysctl_tcp_adv_win_scale = 1;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
-- 
2.15.0.rc2.357.g7e34df9404-goog

[PATCH net-next 15/15] tcp: Namespace-ify sysctl_tcp_frto

[Eric Dumazet]

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/netns/ipv4.h   |  1 +
 include/net/tcp.h          |  1 -
 net/ipv4/sysctl_net_ipv4.c | 14 +++++++-------
 net/ipv4/tcp_input.c       |  3 +--
 net/ipv4/tcp_ipv4.c        |  1 +
 5 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 9dbb07d4eff465428817831e55c6a4922b7208fb..f4622e28db3a1484553f51709b144ee769766a28 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -141,6 +141,7 @@ struct netns_ipv4 {
 	int sysctl_tcp_dsack;
 	int sysctl_tcp_app_win;
 	int sysctl_tcp_adv_win_scale;
+	int sysctl_tcp_frto;
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 	int sysctl_tcp_fastopen;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 2572b57682987dd5f3700ed47d63e7238946b9a8..19006a5d073c202995ba63199ab8cde814d6d869 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -244,7 +244,6 @@ extern int sysctl_tcp_max_orphans;
 extern long sysctl_tcp_mem[3];
 extern int sysctl_tcp_wmem[3];
 extern int sysctl_tcp_rmem[3];
-extern int sysctl_tcp_frto;
 extern int sysctl_tcp_nometrics_save;
 extern int sysctl_tcp_moderate_rcvbuf;
 extern int sysctl_tcp_tso_win_divisor;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index a95123e1e7da706c88bf5553b7d8ef6c2653ab50..f1bcb9b7e082c6688fad12e15be9b872ebed8151 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -437,13 +437,6 @@ static struct ctl_table ipv4_table[] = {
 		.proc_handler	= proc_dointvec_minmax,
 		.extra1		= &one,
 	},
-	{
-		.procname	= "tcp_frto",
-		.data		= &sysctl_tcp_frto,
-		.maxlen		= sizeof(int),
-		.mode		= 0644,
-		.proc_handler	= proc_dointvec
-	},
 	{
 		.procname	= "tcp_min_rtt_wlen",
 		.data		= &sysctl_tcp_min_rtt_wlen,
@@ -1145,6 +1138,13 @@ static struct ctl_table ipv4_net_table[] = {
 		.extra1		= &tcp_adv_win_scale_min,
 		.extra2		= &tcp_adv_win_scale_max,
 	},
+	{
+		.procname	= "tcp_frto",
+		.data		= &init_net.ipv4.sysctl_tcp_frto,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec
+	},
 	{ }
 };
 
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index d01f390da23dcd4100271b150bd8bc143f7328cf..24950ea3094288cad8d9cd9eb0e0698d6f50e989 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -81,7 +81,6 @@
 int sysctl_tcp_challenge_ack_limit = 1000;
 
 int sysctl_tcp_max_orphans __read_mostly = NR_FILE;
-int sysctl_tcp_frto __read_mostly = 2;
 int sysctl_tcp_min_rtt_wlen __read_mostly = 300;
 int sysctl_tcp_moderate_rcvbuf __read_mostly = 1;
 int sysctl_tcp_invalid_ratelimit __read_mostly = HZ/2;
@@ -2024,7 +2023,7 @@ void tcp_enter_loss(struct sock *sk)
 	 * falsely raise the receive window, which results in repeated
 	 * timeouts and stop-and-go behavior.
 	 */
-	tp->frto = sysctl_tcp_frto &&
+	tp->frto = net->ipv4.sysctl_tcp_frto &&
 		   (new_recovery || icsk->icsk_retransmits) &&
 		   !inet_csk(sk)->icsk_mtup.probe_size;
 }
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 1fe30fb99308b3e3fd07509b509b0e3727cc5d44..49757c75821111c6d2cf413415be2c1b58482659 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2492,6 +2492,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_dsack = 1;
 	net->ipv4.sysctl_tcp_app_win = 31;
 	net->ipv4.sysctl_tcp_adv_win_scale = 1;
+	net->ipv4.sysctl_tcp_frto = 2;
 
 	net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
 	spin_lock_init(&net->ipv4.tcp_fastopen_ctx_lock);
-- 
2.15.0.rc2.357.g7e34df9404-goog

Re: [PATCH net-next 00/15] tcp: move 14 sysctls to namespaces

[David Miller]

From: Eric Dumazet <edumazet@google.com>
Date: Thu, 26 Oct 2017 16:35:11 -0700

> Ideally all TCP sysctls should be per netns.
> This patch series takes care of 14 of sysctls.
> More to come later.

The tcp-fack patch doesn't apply cleanly, please respin.

Thank you.

Re: [PATCH net-next 00/15] tcp: move 14 sysctls to namespaces

[Eric Dumazet]

On Fri, 2017-10-27 at 13:46 +0900, David Miller wrote:
> From: Eric Dumazet <edumazet@google.com>
> Date: Thu, 26 Oct 2017 16:35:11 -0700
> 
> > Ideally all TCP sysctls should be per netns.
> > This patch series takes care of 14 of sysctls.
> > More to come later.
> 
> The tcp-fack patch doesn't apply cleanly, please respin.
> 
> Thank you.

Sure, I did a git rebase that went well.

v2 is coming right away, thanks.