Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 010b64f7 authored by David S. Miller's avatar David S. Miller
Browse files

Merge branch 'tcp-move-14-sysctls-to-namespaces' 



Eric Dumazet says:

====================
tcp: move 14 sysctls to namespaces

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

Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parents 4dc12ffe af9b69a7

include/net/netns/ipv4.h

+14 −0
Original line number Diff line number Diff line
@@ -128,6 +128,20 @@ struct netns_ipv4 { 
	int sysctl_tcp_sack;

	int sysctl_tcp_window_scaling;

	int sysctl_tcp_timestamps;

	int sysctl_tcp_early_retrans;

	int sysctl_tcp_recovery;

	int sysctl_tcp_thin_linear_timeouts;

	int sysctl_tcp_slow_start_after_idle;

	int sysctl_tcp_retrans_collapse;

	int sysctl_tcp_stdurg;

	int sysctl_tcp_rfc1337;

	int sysctl_tcp_abort_on_overflow;

	int sysctl_tcp_fack;

	int sysctl_tcp_max_reordering;

	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;

include/net/tcp.h

+6 −21
Original line number Diff line number Diff line
@@ -243,30 +243,15 @@ 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;

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;

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;

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;

extern int sysctl_tcp_early_retrans;

extern int sysctl_tcp_recovery;



#define TCP_RACK_LOSS_DETECTION  0x1 /* Use RACK to detect losses */



extern int sysctl_tcp_limit_output_bytes;
@@ -1311,7 +1296,7 @@ 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;
@@ -1324,9 +1309,9 @@ 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)) :
@@ -1336,13 +1321,13 @@ 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,

net/ipv4/sysctl_net_ipv4.c

+102 −102
Original line number Diff line number Diff line
@@ -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,
@@ -400,27 +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	= "tcp_stdurg",

		.data		= &sysctl_tcp_stdurg,

		.maxlen		= sizeof(int),

		.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,
@@ -442,34 +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_recovery",

		.data		= &sysctl_tcp_recovery,

		.maxlen		= sizeof(int),

		.mode		= 0644,

		.proc_handler	= proc_dointvec,

	},

	{

		.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,

		.maxlen		= sizeof(int),

		.mode		= 0644,

		.proc_handler	= proc_dointvec

	},

	{

		.procname	= "tcp_mem",

		.maxlen		= sizeof(sysctl_tcp_mem),
@@ -493,29 +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,

		.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,

		.maxlen		= sizeof(int),

		.mode		= 0644,

		.proc_handler	= proc_dointvec

	},

	{

		.procname	= "tcp_min_rtt_wlen",

		.data		= &sysctl_tcp_min_rtt_wlen,
@@ -578,13 +499,6 @@ 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",
@@ -627,22 +541,6 @@ 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_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 +1043,108 @@ 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,

	},

	{

		.procname	= "tcp_recovery",

		.data		= &init_net.ipv4.sysctl_tcp_recovery,

		.maxlen		= sizeof(int),

		.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

	},

	{

		.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

	},

	{

		.procname	= "tcp_retrans_collapse",

		.data		= &init_net.ipv4.sysctl_tcp_retrans_collapse,

		.maxlen		= sizeof(int),

		.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

	},

	{

		.procname	= "tcp_rfc1337",

		.data		= &init_net.ipv4.sysctl_tcp_rfc1337,

		.maxlen		= sizeof(int),

		.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

	},

	{

		.procname	= "tcp_fack",

		.data		= &init_net.ipv4.sysctl_tcp_fack,

		.maxlen		= sizeof(int),

		.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

	},

	{

		.procname	= "tcp_dsack",

		.data		= &init_net.ipv4.sysctl_tcp_dsack,

		.maxlen		= sizeof(int),

		.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

	},

	{

		.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,

	},

	{

		.procname	= "tcp_frto",

		.data		= &init_net.ipv4.sysctl_tcp_frto,

		.maxlen		= sizeof(int),

		.mode		= 0644,

		.proc_handler	= proc_dointvec

	},

	{ }

};

net/ipv4/tcp.c

+1 −1
Original line number Diff line number Diff line
@@ -2517,7 +2517,7 @@ 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:

net/ipv4/tcp_input.c

+16 −26
Original line number Diff line number Diff line
@@ -79,23 +79,12 @@ 
#include <linux/unaligned/access_ok.h>

#include <linux/static_key.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;

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;



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;

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.		*/
@@ -370,8 +359,8 @@ 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)
@@ -396,7 +385,7 @@ 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);
@@ -434,6 +423,7 @@ 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;
@@ -452,14 +442,14 @@ 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);
@@ -636,7 +626,7 @@ 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]);
@@ -893,7 +883,7 @@ 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",
@@ -2035,7 +2025,7 @@ 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;

}
@@ -2789,7 +2779,7 @@ 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);
@@ -4155,7 +4145,7 @@ 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))
@@ -4190,7 +4180,7 @@ 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))
@@ -4815,7 +4805,7 @@ 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;
@@ -5124,7 +5114,7 @@ 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);
@@ -5723,7 +5713,7 @@ 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);