Re: [RFC PATCH net-next 1/4] txhash: Make rethinking txhash behavior configurable via sysctl

Add a per ns sysctl that controls the txhash rethink behavior,
sk_rethink_txhash. When enabled, the same behavior is retained, when
disabled, rethink is not performed. Sysctl is disabled by default.

Signed-off-by: Akhmat Karakotov <redacted>
---
 include/net/netns/core.h    |  2 ++
 include/net/sock.h          | 34 +++++++++++++++++++++-------------
 include/uapi/linux/socket.h |  3 +++
 net/core/net_namespace.c    |  3 +++
 net/core/sysctl_net_core.c  |  7 +++++++
 5 files changed, 36 insertions(+), 13 deletions(-)

diff --git a/include/net/netns/core.h b/include/net/netns/core.h
index 36c2d998a43c..177980b46ed7 100644
--- a/include/net/netns/core.h
+++ b/include/net/netns/core.h

@@ -11,6 +11,8 @@ struct netns_core {
 
 	int	sysctl_somaxconn;
 
+	unsigned int sysctl_txrehash;

+
 #ifdef CONFIG_PROC_FS
 	int __percpu *sock_inuse;
 	struct prot_inuse __percpu *prot_inuse;

diff --git a/include/net/sock.h b/include/net/sock.h
index 66a9a90f9558..d8a73edb1629 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h

@@ -577,6 +577,18 @@ static inline bool sk_user_data_is_nocopy(const struct sock *sk)
 			   __tmp | SK_USER_DATA_NOCOPY);		\
 })
 
+static inline
+struct net *sock_net(const struct sock *sk)
+{
+	return read_pnet(&sk->sk_net);
+}
+
+static inline
+void sock_net_set(struct sock *sk, struct net *net)
+{
+	write_pnet(&sk->sk_net, net);
+}
+
 /*
  * SK_CAN_REUSE and SK_NO_REUSE on a socket mean that the socket is OK
  * or not whether his port will be reused by someone else. SK_FORCE_REUSE

@@ -1942,10 +1954,18 @@ static inline void sk_set_txhash(struct sock *sk)
 
 static inline bool sk_rethink_txhash(struct sock *sk)
 {
-	if (sk->sk_txhash) {
+	unsigned int rehash;
+
+	if (!sk->sk_txhash)
+		return false;
+
+	rehash = READ_ONCE(sock_net(sk)->core.sysctl_txrehash);
+
+	if (rehash) {
 		sk_set_txhash(sk);
 		return true;
 	}
+
 	return false;
 }
 

@@ -2596,18 +2616,6 @@ static inline void sk_eat_skb(struct sock *sk, struct sk_buff *skb)
 	__kfree_skb(skb);
 }
 
-static inline
-struct net *sock_net(const struct sock *sk)
-{
-	return read_pnet(&sk->sk_net);
-}
-
-static inline
-void sock_net_set(struct sock *sk, struct net *net)
-{
-	write_pnet(&sk->sk_net, net);
-}
-
 static inline bool
 skb_sk_is_prefetched(struct sk_buff *skb)
 {

diff --git a/include/uapi/linux/socket.h b/include/uapi/linux/socket.h
index eb0a9a5b6e71..0accd6102ece 100644
--- a/include/uapi/linux/socket.h
+++ b/include/uapi/linux/socket.h

@@ -31,4 +31,7 @@ struct __kernel_sockaddr_storage {
 
 #define SOCK_BUF_LOCK_MASK (SOCK_SNDBUF_LOCK | SOCK_RCVBUF_LOCK)
 
+#define SOCK_TXREHASH_DISABLED	0
+#define SOCK_TXREHASH_ENABLED	1
+
 #endif /* _UAPI_LINUX_SOCKET_H */

diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index a448a9b5bb2d..0d833b861f00 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c

@@ -359,6 +359,9 @@ static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
 static int __net_init net_defaults_init_net(struct net *net)
 {
 	net->core.sysctl_somaxconn = SOMAXCONN;
+
+	net->core.sysctl_txrehash = SOCK_TXREHASH_DISABLED;
+
 	return 0;
 }
 

diff --git a/net/core/sysctl_net_core.c b/net/core/sysctl_net_core.c
index c8496c1142c9..34144abbb6a0 100644
--- a/net/core/sysctl_net_core.c
+++ b/net/core/sysctl_net_core.c

@@ -592,6 +592,13 @@ static struct ctl_table netns_core_table[] = {
 		.extra1		= SYSCTL_ZERO,
 		.proc_handler	= proc_dointvec_minmax
 	},
+	{
+		.procname	= "txrehash",
+		.data		= &init_net.core.sysctl_txrehash,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_do_intvec