From: Jason Xing <kernelxing@tencent.com> Taking CLOSE-WAIT sockets into CurrEstab counters is in accordance with RFC 1213, as suggested by Eric and Neal. Previous discussion Link: https://lore.kernel.org/all/20240529033104.33882-1-kerneljasonxing@gmail.com/ Jason Xing (2): tcp: count CLOSE-WAIT sockets for TCP_MIB_CURRESTAB mptcp: count CLOSE-WAIT sockets for MPTCP_MIB_CURRESTAB net/ipv4/tcp.c | 6 +++++- net/mptcp/protocol.c | 5 +++-- 2 files changed, 8 insertions(+), 3 deletions(-) -- 2.37.3 From: Jason Xing <kernelxing@tencent.com> According to RFC 1213, we should also take CLOSE-WAIT sockets into consideration: "tcpCurrEstab OBJECT-TYPE ... The number of TCP connections for which the current state is either ESTABLISHED or CLOSE- WAIT." After this, CurrEstab counter will display the total number of ESTABLISHED and CLOSE-WAIT sockets. The logic of counting When we increment the counter? a) if we change the state to ESTABLISHED. b) if we change the state from SYN-RECEIVED to CLOSE-WAIT. When we decrement the counter? a) if the socket leaves ESTABLISHED and will never go into CLOSE-WAIT, say, on the client side, changing from ESTABLISHED to FIN-WAIT-1. b) if the socket leaves CLOSE-WAIT, say, on the server side, changing from CLOSE-WAIT to LAST-ACK. Please note: there are two chances that old state of socket can be changed to CLOSE-WAIT in tcp_fin(). One is SYN-RECV, the other is ESTABLISHED. So we have to take care of the former case. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Jason Xing <kernelxing@tencent.com> --- previous discussion Link: https://lore.kernel.org/all/20240529033104.33882-1-kerneljasonxing@gmail.com/ 1. Chose to fix CurrEstab instead of introduing a new counter (Eric, Neal) --- net/ipv4/tcp.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index XXXXXXX..XXXXXXX 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -XXX,XX +XXX,XX @@ void tcp_set_state(struct sock *sk, int state) if (oldstate != TCP_ESTABLISHED) TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); break; + case TCP_CLOSE_WAIT: + if (oldstate == TCP_SYN_RECV) + TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); + break; case TCP_CLOSE: if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED) @@ -XXX,XX +XXX,XX @@ void tcp_set_state(struct sock *sk, int state) inet_put_port(sk); fallthrough; default: - if (oldstate == TCP_ESTABLISHED) + if (oldstate == TCP_ESTABLISHED || oldstate == TCP_CLOSE_WAIT) TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); } -- 2.37.3 From: Jason Xing <kernelxing@tencent.com> Like previous patch does in TCP, we need to adhere to RFC 1213: "tcpCurrEstab OBJECT-TYPE ... The number of TCP connections for which the current state is either ESTABLISHED or CLOSE- WAIT." So let's consider CLOSE-WAIT sockets. The logic of counting When we increment the counter? a) Only if we change the state to ESTABLISHED. When we decrement the counter? a) if the socket leaves ESTABLISHED and will never go into CLOSE-WAIT, say, on the client side, changing from ESTABLISHED to FIN-WAIT-1. b) if the socket leaves CLOSE-WAIT, say, on the server side, changing from CLOSE-WAIT to LAST-ACK. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Jason Xing <kernelxing@tencent.com> --- net/mptcp/protocol.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index XXXXXXX..XXXXXXX 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -XXX,XX +XXX,XX @@ void mptcp_set_state(struct sock *sk, int state) if (oldstate != TCP_ESTABLISHED) MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_CURRESTAB); break; - + case TCP_CLOSE_WAIT: + break; default: - if (oldstate == TCP_ESTABLISHED) + if (oldstate == TCP_ESTABLISHED || oldstate == TCP_CLOSE_WAIT) MPTCP_DEC_STATS(sock_net(sk), MPTCP_MIB_CURRESTAB); } -- 2.37.3 From: Jason Xing <kernelxing@tencent.com> Taking CLOSE-WAIT sockets into CurrEstab counters is in accordance with RFC 1213, as suggested by Eric and Neal. v5 Link: https://lore.kernel.org/all/20240531091753.75930-1-kerneljasonxing@gmail.com/ 1. add more detailed comment (Matthieu) v4 Link: https://lore.kernel.org/all/20240530131308.59737-1-kerneljasonxing@gmail.com/ 1. correct the Fixes: tag in patch [2/2]. (Eric) Previous discussion Link: https://lore.kernel.org/all/20240529033104.33882-1-kerneljasonxing@gmail.com/ Jason Xing (2): tcp: count CLOSE-WAIT sockets for TCP_MIB_CURRESTAB mptcp: count CLOSE-WAIT sockets for MPTCP_MIB_CURRESTAB net/ipv4/tcp.c | 6 +++++- net/mptcp/protocol.c | 9 +++++++-- 2 files changed, 12 insertions(+), 3 deletions(-) -- 2.37.3 From: Jason Xing <kernelxing@tencent.com> According to RFC 1213, we should also take CLOSE-WAIT sockets into consideration: "tcpCurrEstab OBJECT-TYPE ... The number of TCP connections for which the current state is either ESTABLISHED or CLOSE- WAIT." After this, CurrEstab counter will display the total number of ESTABLISHED and CLOSE-WAIT sockets. The logic of counting When we increment the counter? a) if we change the state to ESTABLISHED. b) if we change the state from SYN-RECEIVED to CLOSE-WAIT. When we decrement the counter? a) if the socket leaves ESTABLISHED and will never go into CLOSE-WAIT, say, on the client side, changing from ESTABLISHED to FIN-WAIT-1. b) if the socket leaves CLOSE-WAIT, say, on the server side, changing from CLOSE-WAIT to LAST-ACK. Please note: there are two chances that old state of socket can be changed to CLOSE-WAIT in tcp_fin(). One is SYN-RECV, the other is ESTABLISHED. So we have to take care of the former case. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Jason Xing <kernelxing@tencent.com> Reviewed-by: Eric Dumazet <edumazet@google.com> --- previous discussion Link: https://lore.kernel.org/all/20240529033104.33882-1-kerneljasonxing@gmail.com/ 1. Chose to fix CurrEstab instead of introduing a new counter (Eric, Neal) --- net/ipv4/tcp.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index XXXXXXX..XXXXXXX 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -XXX,XX +XXX,XX @@ void tcp_set_state(struct sock *sk, int state) if (oldstate != TCP_ESTABLISHED) TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); break; + case TCP_CLOSE_WAIT: + if (oldstate == TCP_SYN_RECV) + TCP_INC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); + break; case TCP_CLOSE: if (oldstate == TCP_CLOSE_WAIT || oldstate == TCP_ESTABLISHED) @@ -XXX,XX +XXX,XX @@ void tcp_set_state(struct sock *sk, int state) inet_put_port(sk); fallthrough; default: - if (oldstate == TCP_ESTABLISHED) + if (oldstate == TCP_ESTABLISHED || oldstate == TCP_CLOSE_WAIT) TCP_DEC_STATS(sock_net(sk), TCP_MIB_CURRESTAB); } -- 2.37.3 From: Jason Xing <kernelxing@tencent.com> Like previous patch does in TCP, we need to adhere to RFC 1213: "tcpCurrEstab OBJECT-TYPE ... The number of TCP connections for which the current state is either ESTABLISHED or CLOSE- WAIT." So let's consider CLOSE-WAIT sockets. The logic of counting When we increment the counter? a) Only if we change the state to ESTABLISHED. When we decrement the counter? a) if the socket leaves ESTABLISHED and will never go into CLOSE-WAIT, say, on the client side, changing from ESTABLISHED to FIN-WAIT-1. b) if the socket leaves CLOSE-WAIT, say, on the server side, changing from CLOSE-WAIT to LAST-ACK. Fixes: d9cd27b8cd19 ("mptcp: add CurrEstab MIB counter support") Signed-off-by: Jason Xing <kernelxing@tencent.com> --- net/mptcp/protocol.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index XXXXXXX..XXXXXXX 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -XXX,XX +XXX,XX @@ void mptcp_set_state(struct sock *sk, int state) if (oldstate != TCP_ESTABLISHED) MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_CURRESTAB); break; - + case TCP_CLOSE_WAIT: + /* Unlike TCP, MPTCP sk would not have the TCP_SYN_RECV state: + * MPTCP "accepted" sockets will be created later on. So no + * transition from TCP_SYN_RECV to TCP_CLOSE_WAIT. + */ + break; default: - if (oldstate == TCP_ESTABLISHED) + if (oldstate == TCP_ESTABLISHED || oldstate == TCP_CLOSE_WAIT) MPTCP_DEC_STATS(sock_net(sk), MPTCP_MIB_CURRESTAB); } -- 2.37.3