tcp: add TCPMemoryPressuresChrono counter

DRAM supply shortage and poor memory pressure tracking in TCP
stack makes any change in SO_SNDBUF/SO_RCVBUF (or equivalent autotuning
limits) and tcp_mem[] quite hazardous.

TCPMemoryPressures SNMP counter is an indication of tcp_mem sysctl
limits being hit, but only tracking number of transitions.

If TCP stack behavior under stress was perfect :
1) It would maintain memory usage close to the limit.
2) Memory pressure state would be entered for short times.

We certainly prefer 100 events lasting 10ms compared to one event
lasting 200 seconds.

This patch adds a new SNMP counter tracking cumulative duration of
memory pressure events, given in ms units.

$ cat /proc/sys/net/ipv4/tcp_mem
3088    4117    6176
$ grep TCP /proc/net/sockstat
TCP: inuse 180 orphan 0 tw 2 alloc 234 mem 4140
$ nstat -n ; sleep 10 ; nstat |grep Pressure
TcpExtTCPMemoryPressures        1700
TcpExtTCPMemoryPressuresChrono  5209

v2: Used EXPORT_SYMBOL_GPL() instead of EXPORT_SYMBOL() as David
instructed.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

7 files changed, 50 insertions, 8 deletions

diff --git a/net/core/sock.c b/net/core/sock.c
index bef844127e01..ad8a4bc84126 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -2076,6 +2076,26 @@ int sock_cmsg_send(struct sock *sk, struct msghdr *msg,
 }
 EXPORT_SYMBOL(sock_cmsg_send);
 
+static void sk_enter_memory_pressure(struct sock *sk)
+{
+	if (!sk->sk_prot->enter_memory_pressure)
+		return;
+
+	sk->sk_prot->enter_memory_pressure(sk);
+}
+
+static void sk_leave_memory_pressure(struct sock *sk)
+{
+	if (sk->sk_prot->leave_memory_pressure) {
+		sk->sk_prot->leave_memory_pressure(sk);
+	} else {
+		unsigned long *memory_pressure = sk->sk_prot->memory_pressure;
+
+		if (memory_pressure && *memory_pressure)
+			*memory_pressure = 0;
+	}
+}
+
 /* On 32bit arches, an skb frag is limited to 2^15 */
 #define SKB_FRAG_PAGE_ORDER	get_order(32768)
 
diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c
index 405483a07efc..73a0399dc7a2 100644
--- a/net/decnet/af_decnet.c
+++ b/net/decnet/af_decnet.c
@@ -447,7 +447,7 @@ static void dn_destruct(struct sock *sk)
 	dst_release(rcu_dereference_check(sk->sk_dst_cache, 1));
 }
 
-static int dn_memory_pressure;
+static unsigned long dn_memory_pressure;
 
 static void dn_enter_memory_pressure(struct sock *sk)
 {
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index fa44e752a9a3..43eb6567b3a0 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -250,6 +250,7 @@ static const struct snmp_mib snmp4_net_list[] = {
 	SNMP_MIB_ITEM("TCPAbortOnLinger", LINUX_MIB_TCPABORTONLINGER),
 	SNMP_MIB_ITEM("TCPAbortFailed", LINUX_MIB_TCPABORTFAILED),
 	SNMP_MIB_ITEM("TCPMemoryPressures", LINUX_MIB_TCPMEMORYPRESSURES),
+	SNMP_MIB_ITEM("TCPMemoryPressuresChrono", LINUX_MIB_TCPMEMORYPRESSURESCHRONO),
 	SNMP_MIB_ITEM("TCPSACKDiscard", LINUX_MIB_TCPSACKDISCARD),
 	SNMP_MIB_ITEM("TCPDSACKIgnoredOld", LINUX_MIB_TCPDSACKIGNOREDOLD),
 	SNMP_MIB_ITEM("TCPDSACKIgnoredNoUndo", LINUX_MIB_TCPDSACKIGNOREDNOUNDO),
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 87981fcdfcf2..cc8fd8b747a4 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -320,17 +320,36 @@ struct tcp_splice_state {
  * All the __sk_mem_schedule() is of this nature: accounting
  * is strict, actions are advisory and have some latency.
  */
-int tcp_memory_pressure __read_mostly;
-EXPORT_SYMBOL(tcp_memory_pressure);
+unsigned long tcp_memory_pressure __read_mostly;
+EXPORT_SYMBOL_GPL(tcp_memory_pressure);
 
 void tcp_enter_memory_pressure(struct sock *sk)
 {
-	if (!tcp_memory_pressure) {
+	unsigned long val;
+
+	if (tcp_memory_pressure)
+		return;
+	val = jiffies;
+
+	if (!val)
+		val--;
+	if (!cmpxchg(&tcp_memory_pressure, 0, val))
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURES);
-		tcp_memory_pressure = 1;
-	}
 }
-EXPORT_SYMBOL(tcp_enter_memory_pressure);
+EXPORT_SYMBOL_GPL(tcp_enter_memory_pressure);
+
+void tcp_leave_memory_pressure(struct sock *sk)
+{
+	unsigned long val;
+
+	if (!tcp_memory_pressure)
+		return;
+	val = xchg(&tcp_memory_pressure, 0);
+	if (val)
+		NET_ADD_STATS(sock_net(sk), LINUX_MIB_TCPMEMORYPRESSURESCHRONO,
+			      jiffies_to_msecs(jiffies - val));
+}
+EXPORT_SYMBOL_GPL(tcp_leave_memory_pressure);
 
 /* Convert seconds to retransmits based on initial and max timeout */
 static u8 secs_to_retrans(int seconds, int timeout, int rto_max)
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 13c7ae7d4504..1dc8c449e16a 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2387,6 +2387,7 @@ struct proto tcp_prot = {
 	.unhash			= inet_unhash,
 	.get_port		= inet_csk_get_port,
 	.enter_memory_pressure	= tcp_enter_memory_pressure,
+	.leave_memory_pressure	= tcp_leave_memory_pressure,
 	.stream_memory_free	= tcp_stream_memory_free,
 	.sockets_allocated	= &tcp_sockets_allocated,
 	.orphan_count		= &tcp_orphan_count,
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 5a525426fe93..0840543fc245 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1910,6 +1910,7 @@ struct proto tcpv6_prot = {
 	.unhash			= inet_unhash,
 	.get_port		= inet_csk_get_port,
 	.enter_memory_pressure	= tcp_enter_memory_pressure,
+	.leave_memory_pressure	= tcp_leave_memory_pressure,
 	.stream_memory_free	= tcp_stream_memory_free,
 	.sockets_allocated	= &tcp_sockets_allocated,
 	.memory_allocated	= &tcp_memory_allocated,
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index 0822046e4f3f..5f58dd03e3ac 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -103,7 +103,7 @@ static int sctp_autobind(struct sock *sk);
 static void sctp_sock_migrate(struct sock *, struct sock *,
 			      struct sctp_association *, sctp_socket_type_t);
 
-static int sctp_memory_pressure;
+static unsigned long sctp_memory_pressure;
 static atomic_long_t sctp_memory_allocated;
 struct percpu_counter sctp_sockets_allocated;