[PATCH] tcp: expose listen socket backlog via SIOCGBACKLOG
DORMANTno replies
From: Derrick Pallas <hidden>
Date: 2015-05-07 02:42:19
Subsystem:
networking [general], networking [sockets], networking [tcp], the rest · Maintainers:
"David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Kuniyuki Iwashima, Willem de Bruijn, Neal Cardwell, Linus Torvalds
This IOCTL exposes the current backlog for a listen socket. This value can be useful in userland, e.g. in an overloaded transparent proxy that would rather push new connections into bypass than have then hang waiting for accept. Knowing the current backlog allows this process to make that decision and then begin intercepting connections again once the backlog falls to a reasonable level. Signed-off-by: Derrick Pallas <redacted> --- include/net/sock.h | 5 +++++ include/uapi/linux/sockios.h | 1 + net/ipv4/tcp.c | 8 ++++++++ 3 files changed, 14 insertions(+)
diff --git a/include/net/sock.h b/include/net/sock.h
index 3a4898e..7597671 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h@@ -772,6 +772,11 @@ static inline void sk_acceptq_added(struct sock *sk) sk->sk_ack_backlog++; } +static inline int sk_acceptq_backlog(struct sock *sk) +{ + return sk->sk_ack_backlog; +} + static inline bool sk_acceptq_is_full(const struct sock *sk) { return sk->sk_ack_backlog > sk->sk_max_ack_backlog;
diff --git a/include/uapi/linux/sockios.h b/include/uapi/linux/sockios.h
index e888b1a..9320493 100644
--- a/include/uapi/linux/sockios.h
+++ b/include/uapi/linux/sockios.h@@ -84,6 +84,7 @@ #define SIOCWANDEV 0x894A /* get/set netdev parameters */ #define SIOCOUTQNSD 0x894B /* output queue size (not sent only) */ +#define SIOCGBACKLOG 0x894C /* get listen socket backlog length */ /* ARP cache control calls. */ /* 0x8950 - 0x8952 * obsolete calls, don't re-use */
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 8c5cd9e..1aa509d 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c@@ -591,6 +591,14 @@ int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg) else answ = tp->write_seq - tp->snd_nxt; break; + + case SIOCGBACKLOG: + if (sk->sk_state != TCP_LISTEN) + return -EINVAL; + + answ = sk_acceptq_backlog(sk); + break; + default: return -ENOIOCTLCMD; }
--
1.9.1