Skip to contentSkip to conversation

tcp: add sysctl_tcp_rto_min_us

PATCH Superseded Kevin Yang 3 participants 6 messages May 30, 2024 – Jun 3, 2024

Subsystems (4)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Kevin Yang

author
#0 lore raw

v2 cover tcp: add sysctl_tcp_rto_min_us

Adding a sysctl knob to allow user to specify a default rto_min at socket init time.

After this patch series, the rto_min will has multiple sources:
route option has the highest precedence, followed by the TCP_BPF_RTO_MIN socket option, followed by this new tcp_rto_min_us sysctl.

v2:
fit line width to 80 column.

v1: https://lore.kernel.org/netdev/[email protected]/

Kevin Yang (2):
tcp: derive delack_max with tcp_rto_min helper
tcp: add sysctl_tcp_rto_min_us

 Documentation/networking/ip-sysctl.rst | 13 +++++++++++++
 include/net/netns/ipv4.h               |  1 +
 net/ipv4/sysctl_net_ipv4.c             |  8 ++++++++
 net/ipv4/tcp.c                         |  4 +++-
 net/ipv4/tcp_ipv4.c                    |  1 +
 net/ipv4/tcp_output.c                  | 11 ++---------
 6 files changed, 28 insertions(+), 10 deletions(-)
Signature · 1 line
2.45.1.288.g0e0cd299f1-goog
author
#1 lore raw

PATCH v2 1/2 tcp: derive delack_max with tcp_rto_min helper

Rto_min now has multiple souces, ordered by preprecedence high to
low: ip route option rto_min, icsk->icsk_rto_min.

When derive delack_max from rto_min, we should not only use ip route option, but should use tcp_rto_min helper to get the correct rto_min.

---
 net/ipv4/tcp_output.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
net/ipv4/tcp_output.c +2 -9
index f97e098f18a5..b44f639a9fa6 100644
@@ -4163,16 +4163,9 @@ EXPORT_SYMBOL(tcp_connect);
 
 u32 tcp_delack_max(const struct sock *sk)
 {
-	const struct dst_entry *dst = __sk_dst_get(sk);
-	u32 delack_max = inet_csk(sk)->icsk_delack_max;
-
-	if (dst && dst_metric_locked(dst, RTAX_RTO_MIN)) {
-		u32 rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN);
-		u32 delack_from_rto_min = max_t(int, 1, rto_min - 1);
+	u32 delack_from_rto_min = max_t(int, 1, tcp_rto_min(sk) - 1);
 
-		delack_max = min_t(u32, delack_max, delack_from_rto_min);
-	}
-	return delack_max;
+	return min_t(u32, inet_csk(sk)->icsk_delack_max, delack_from_rto_min);
 }
 
 /* Send out a delayed ack, the caller does the policy checking
Signature · 1 line
2.45.1.288.g0e0cd299f1-goog
reviewer
#2 lore raw

Re: [PATCH net-next v2 1/2] tcp: derive delack_max with tcp_rto_min helper

Quoted 2 lines · Kevin Yang

Rto_min now has multiple souces, ordered by preprecedence high to
low: ip route option rto_min, icsk->icsk_rto_min.

Hi Kevin,

If you have to respin for some other reason: souces -> sources

Flagged by checkpatch.pl --codespell

...

#3 lore raw

RE: [PATCH net-next v2 1/2] tcp: derive delack_max with tcp_rto_min helper

In reply to Kevin Yang

From: Kevin Yang

Quoted 9 lines

Sent: 30 May 2024 16:35
To: David Miller <[email protected]>; Eric Dumazet <[email protected]>; Jakub Kicinski

Rto_min now has multiple souces, ordered by preprecedence high to
low: ip route option rto_min, icsk->icsk_rto_min.

When derive delack_max from rto_min, we should not only use ip route option, but should use tcp_rto_min helper to get the correct rto_min.

...

Quoted 5 lines
net/ipv4/tcp_output.c +0 -0
index f97e098f18a5..b44f639a9fa6 100644

net/ipv4/tcp_output.c EXPORT_SYMBOL(tcp_connect); · line 4166

Quoted 10 lines
net/ipv4/tcp_output.c +1 -6
@@ -4163,16 +4163,9 @@ EXPORT_SYMBOL(tcp_connect);
 
 u32 tcp_delack_max(const struct sock *sk)
 {
-	const struct dst_entry *dst = __sk_dst_get(sk);
-	u32 delack_max = inet_csk(sk)->icsk_delack_max;
-
-	if (dst && dst_metric_locked(dst, RTAX_RTO_MIN)) {
-		u32 rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN);
-		u32 delack_from_rto_min = max_t(int, 1, rto_min - 1);
+	u32 delack_from_rto_min = max_t(int, 1, tcp_rto_min(sk) - 1);

That max_t() is more horrid than most.
Perhaps:

		= max(tcp_rto_min(sk), 2) - 1;

net/ipv4/tcp_output.c EXPORT_SYMBOL(tcp_connect); · line 4168

Quoted 5 lines
net/ipv4/tcp_output.c +1 -3
-		delack_max = min_t(u32, delack_max, delack_from_rto_min);
-	}
-	return delack_max;
+	return min_t(u32, inet_csk(sk)->icsk_delack_max, delack_from_rto_min);

Can that just be a min() ??

David

Quoted 6 lines
net/ipv4/tcp_output.c +0 -1
 }
 
 /* Send out a delayed ack, the caller does the policy checking
--

2.45.1.288.g0e0cd299f1-goog

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)

author
#4 lore raw

Re: [PATCH net-next v2 1/2] tcp: derive delack_max with tcp_rto_min helper

thanks for the nice suggestions, sent v3
https://lore.kernel.org/netdev/[email protected]/

Quoted 52 lines · David Laight

From: Kevin Yang

Quoted 9 lines

Sent: 30 May 2024 16:35
To: David Miller <[email protected]>; Eric Dumazet <[email protected]>; Jakub Kicinski

Rto_min now has multiple souces, ordered by preprecedence high to
low: ip route option rto_min, icsk->icsk_rto_min.

When derive delack_max from rto_min, we should not only use ip route option, but should use tcp_rto_min helper to get the correct rto_min.

...

Quoted 5 lines
net/ipv4/tcp_output.c +0 -0
index f97e098f18a5..b44f639a9fa6 100644

net/ipv4/tcp_output.c EXPORT_SYMBOL(tcp_connect); · line 4166

Quoted 10 lines
net/ipv4/tcp_output.c +1 -6
@@ -4163,16 +4163,9 @@ EXPORT_SYMBOL(tcp_connect);
 
 u32 tcp_delack_max(const struct sock *sk)
 {
-     const struct dst_entry *dst = __sk_dst_get(sk);
-     u32 delack_max = inet_csk(sk)->icsk_delack_max;
-
-     if (dst && dst_metric_locked(dst, RTAX_RTO_MIN)) {
-             u32 rto_min = dst_metric_rtt(dst, RTAX_RTO_MIN);
-             u32 delack_from_rto_min = max_t(int, 1, rto_min - 1);
+     u32 delack_from_rto_min = max_t(int, 1, tcp_rto_min(sk) - 1);

That max_t() is more horrid than most.
Perhaps:

                = max(tcp_rto_min(sk), 2) - 1;

net/ipv4/tcp_output.c EXPORT_SYMBOL(tcp_connect); · line 4168

Quoted 5 lines
net/ipv4/tcp_output.c +1 -3
-             delack_max = min_t(u32, delack_max, delack_from_rto_min);
-     }
-     return delack_max;
+     return min_t(u32, inet_csk(sk)->icsk_delack_max, delack_from_rto_min);

Can that just be a min() ??

David

Quoted 6 lines
net/ipv4/tcp_output.c +0 -1
 }
 
 /* Send out a delayed ack, the caller does the policy checking
--

2.45.1.288.g0e0cd299f1-goog

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)

author
#5 lore raw

PATCH v2 2/2 tcp: add sysctl_tcp_rto_min_us

In reply to Kevin Yang

Adding a sysctl knob to allow user to specify a default rto_min at socket init time, other than using the hard coded 200ms default rto_min.

Note that the rto_min route option has the highest precedence for configuring this setting, followed by the TCP_BPF_RTO_MIN socket option, followed by the tcp_rto_min_us sysctl.

---
 Documentation/networking/ip-sysctl.rst | 13 +++++++++++++
 include/net/netns/ipv4.h               |  1 +
 net/ipv4/sysctl_net_ipv4.c             |  8 ++++++++
 net/ipv4/tcp.c                         |  4 +++-
 net/ipv4/tcp_ipv4.c                    |  1 +
 5 files changed, 26 insertions(+), 1 deletion(-)
Documentation/networking/ip-sysctl.rst +13 -0
index bd50df6a5a42..6e99eccdb837 100644
@@ -1196,6 +1196,19 @@ tcp_pingpong_thresh - INTEGER
 
 	Default: 1
 
+tcp_rto_min_us - INTEGER
+	Minimal TCP retransmission timeout (in microseconds). Note that the
+	rto_min route option has the highest precedence for configuring this
+	setting, followed by the TCP_BPF_RTO_MIN socket option, followed by
+	this tcp_rto_min_us sysctl.
+
+	The recommended practice is to use a value less or equal to 200000
+	microseconds.
+
+	Possible Values: 1 - INT_MAX
+
+	Default: 200000
+
 UDP variables
 =============
 
include/net/netns/ipv4.h +1 -0
index c356c458b340..a91bb971f901 100644
@@ -170,6 +170,7 @@ struct netns_ipv4 {
 	u8 sysctl_tcp_sack;
 	u8 sysctl_tcp_window_scaling;
 	u8 sysctl_tcp_timestamps;
+	int sysctl_tcp_rto_min_us;
 	u8 sysctl_tcp_recovery;
 	u8 sysctl_tcp_thin_linear_timeouts;
 	u8 sysctl_tcp_slow_start_after_idle;
net/ipv4/sysctl_net_ipv4.c +8 -0
index d7892f34a15b..bb64c0ef092d 100644
@@ -1503,6 +1503,14 @@ static struct ctl_table ipv4_net_table[] = {
 		.proc_handler	= proc_dou8vec_minmax,
 		.extra1		= SYSCTL_ONE,
 	},
+	{
+		.procname	= "tcp_rto_min_us",
+		.data		= &init_net.ipv4.sysctl_tcp_rto_min_us,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= SYSCTL_ONE,
+	},
 };
 
 static __net_init int ipv4_sysctl_init_net(struct net *net)
net/ipv4/tcp.c +3 -1
index 5fa68e7f6ddb..fa43aaacd92b 100644
@@ -420,6 +420,7 @@ void tcp_init_sock(struct sock *sk)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct tcp_sock *tp = tcp_sk(sk);
+	int rto_min_us;
 
 	tp->out_of_order_queue = RB_ROOT;
 	sk->tcp_rtx_queue = RB_ROOT;
@@ -428,7 +429,8 @@ void tcp_init_sock(struct sock *sk)
 	INIT_LIST_HEAD(&tp->tsorted_sent_queue);
 
 	icsk->icsk_rto = TCP_TIMEOUT_INIT;
-	icsk->icsk_rto_min = TCP_RTO_MIN;
+	rto_min_us = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rto_min_us);
+	icsk->icsk_rto_min = usecs_to_jiffies(rto_min_us);
 	icsk->icsk_delack_max = TCP_DELACK_MAX;
 	tp->mdev_us = jiffies_to_usecs(TCP_TIMEOUT_INIT);
 	minmax_reset(&tp->rtt_min, tcp_jiffies32, ~0U);
net/ipv4/tcp_ipv4.c +1 -0
index 041c7eda9abe..49a5e2c4ec18 100644
@@ -3506,6 +3506,7 @@ static int __net_init tcp_sk_init(struct net *net)
 	net->ipv4.sysctl_tcp_shrink_window = 0;
 
 	net->ipv4.sysctl_tcp_pingpong_thresh = 1;
+	net->ipv4.sysctl_tcp_rto_min_us = jiffies_to_usecs(TCP_RTO_MIN);
 
 	return 0;
 }
Signature · 1 line
2.45.1.288.g0e0cd299f1-goog