Skip to contentSkip to conversation

make sysctl per namespace - V3

PATCH In review Daniel Lezcano 2 participants 13 messages Jan 9, 2008 – Jan 10, 2008

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover make sysctl per namespace - V3

The following patchset makes the ipv6 sysctl to handle multiple network namespaces. Each instance of a network namespace as its own set of sysctl values, that means the behavior of the ipv6 stack can be different depending on the sysctl values setup in the different network namespaces.

Changelog:
V3 : fixed compilation error when CONFIG_SYSCTL=n,
fixed missing initialization when CONFIG_SYSCTL=n

V2 : make the mld_max_msf variable readonly when we are
not in the initial network namespace

V1 : initial post

#1 lore raw

PATCH 1/10 make ipv6_sysctl_register to return a value

This patch makes the function ipv6_sysctl_register to return a value. The af_inet6 init function is now able to handle an error and catch it from the initialization of the sysctl.

---
 include/net/ipv6.h         |    2 +-
 net/ipv6/af_inet6.c        |    5 ++++-
 net/ipv6/sysctl_net_ipv6.c |    9 +++++++--
 3 files changed, 12 insertions(+), 4 deletions(-)
include/net/ipv6.h +1 -1
Index: net-2.6.25/include/net/ipv6.h
===================================================================
@@ -622,7 +622,7 @@ static inline int snmp6_unregister_dev(s
 extern ctl_table ipv6_route_table[];
 extern ctl_table ipv6_icmp_table[];
 
-extern void ipv6_sysctl_register(void);
+extern int ipv6_sysctl_register(void);
 extern void ipv6_sysctl_unregister(void);
 #endif
 
net/ipv6/af_inet6.c +4 -1
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -783,7 +783,9 @@ static int __init inet6_init(void)
 	 */
 
 #ifdef CONFIG_SYSCTL
-	ipv6_sysctl_register();
+	err = ipv6_sysctl_register();
+	if (err)
+		goto sysctl_fail;
 #endif
 	err = icmpv6_init(&inet6_family_ops);
 	if (err)
@@ -897,6 +899,7 @@ ndisc_fail:
 icmp_fail:
 #ifdef CONFIG_SYSCTL
 	ipv6_sysctl_unregister();
+sysctl_fail:
 #endif
 	cleanup_ipv6_mibs();
 out_unregister_sock:
net/ipv6/sysctl_net_ipv6.c +7 -2
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -91,10 +91,15 @@ EXPORT_SYMBOL_GPL(net_ipv6_ctl_path);
 
 static struct ctl_table_header *ipv6_sysctl_header;
 
-void ipv6_sysctl_register(void)
+int ipv6_sysctl_register(void)
 {
 	ipv6_sysctl_header = register_sysctl_paths(net_ipv6_ctl_path,
-			ipv6_table);
+						   ipv6_table);
+	if (!ipv6_sysctl_header)
+		return -ENOMEM;
+
+	return 0;
+
 }
 
 void ipv6_sysctl_unregister(void)
#2 lore raw

PATCH 2/10 make a subsystem for af_inet6

In reply to Daniel Lezcano

This patch add a network namespace subsystem for the af_inet6 module. It does nothing right now, but one of its purpose is to receive the different variables for sysctl in order to initialize them.

When the sysctl variable will be moved to the network namespace structure, they will be no longer initialized as global static variables, so we must find a place to initialize them. Because the sysctl can be disabled, it has no sense to store them in the sysctl_net_ipv6 file.

---
 net/ipv6/af_inet6.c |   22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
net/ipv6/af_inet6.c +22 -0
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -719,6 +719,21 @@ static void cleanup_ipv6_mibs(void)
 	snmp_mib_free((void **)udplite_stats_in6);
 }
 
+static int inet6_net_init(struct net *net)
+{
+	return 0;
+}
+
+static void inet6_net_exit(struct net *net)
+{
+	return;
+}
+
+static struct pernet_operations inet6_net_ops = {
+	.init = inet6_net_init,
+	.exit = inet6_net_exit,
+};
+
 static int __init inet6_init(void)
 {
 	struct sk_buff *dummy_skb;
@@ -782,6 +797,10 @@ static int __init inet6_init(void)
 	 *	able to communicate via both network protocols.
 	 */
 
+	err = register_pernet_subsys(&inet6_net_ops);
+	if (err)
+		goto register_pernet_fail;
+
 #ifdef CONFIG_SYSCTL
 	err = ipv6_sysctl_register();
 	if (err)
@@ -901,6 +920,8 @@ icmp_fail:
 	ipv6_sysctl_unregister();
 sysctl_fail:
 #endif
+	unregister_pernet_subsys(&inet6_net_ops);
+register_pernet_fail:
 	cleanup_ipv6_mibs();
 out_unregister_sock:
 	sock_unregister(PF_INET6);
@@ -956,6 +977,7 @@ static void __exit inet6_exit(void)
 #ifdef CONFIG_SYSCTL
 	ipv6_sysctl_unregister();
 #endif
+	unregister_pernet_subsys(&inet6_net_ops);
 	cleanup_ipv6_mibs();
 	proto_unregister(&rawv6_prot);
 	proto_unregister(&udplitev6_prot);
#3 lore raw

PATCH 3/10 add ipv6 structure for netns

In reply to Daniel Lezcano

Like the ipv4 part, this patch adds an ipv6 structure in the net structure to aggregate the different resources to make ipv6 per namespace.

---
 include/net/net_namespace.h |    4 ++++
 include/net/netns/ipv6.h    |   10 ++++++++++
 2 files changed, 14 insertions(+)
include/net/net_namespace.h +4 -0
Index: net-2.6.25/include/net/net_namespace.h
===================================================================
@@ -11,6 +11,7 @@
 #include <net/netns/unix.h>
 #include <net/netns/packet.h>
 #include <net/netns/ipv4.h>
+#include <net/netns/ipv6.h>
 
 struct proc_dir_entry;
 struct net_device;
@@ -48,6 +49,9 @@ struct net {
 	struct netns_packet	packet;
 	struct netns_unix	unx;
 	struct netns_ipv4	ipv4;
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+	struct netns_ipv6	ipv6;
+#endif
 };
 
 #ifdef CONFIG_NET
include/net/netns/ipv6.h new file +10 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -0,0 +1,10 @@
+/*
+ * ipv6 in net namespaces
+ */
+
+#ifndef __NETNS_IPV6_H__
+#define __NETNS_IPV6_H__
+
+struct netns_ipv6 {
+};
+#endif
#4 lore raw

PATCH 4/10 make the ipv6 sysctl to be a netns subsystem

In reply to Daniel Lezcano

The initialization of the sysctl for the ipv6 protocol is changed to a network namespace subsystem. That means when a new network namespace is created the initialization function for the sysctl will be called.

That do not change the behavior of the sysctl in case of the kernel with the network namespace disabled.

---
 net/ipv6/sysctl_net_ipv6.c |   23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)
net/ipv6/sysctl_net_ipv6.c +19 -4
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -91,10 +91,10 @@ EXPORT_SYMBOL_GPL(net_ipv6_ctl_path);
 
 static struct ctl_table_header *ipv6_sysctl_header;
 
-int ipv6_sysctl_register(void)
+static int ipv6_sysctl_net_init(struct net *net)
 {
-	ipv6_sysctl_header = register_sysctl_paths(net_ipv6_ctl_path,
-						   ipv6_table);
+	ipv6_sysctl_header = register_net_sysctl_table(net, net_ipv6_ctl_path,
+						       ipv6_table);
 	if (!ipv6_sysctl_header)
 		return -ENOMEM;
 
@@ -102,7 +102,22 @@ int ipv6_sysctl_register(void)
 
 }
 
+static void ipv6_sysctl_net_exit(struct net *net)
+{
+	unregister_net_sysctl_table(ipv6_sysctl_header);
+}
+
+static struct pernet_operations ipv6_sysctl_net_ops = {
+	.init = ipv6_sysctl_net_init,
+	.exit = ipv6_sysctl_net_exit,
+};
+
+int ipv6_sysctl_register(void)
+{
+	return register_pernet_subsys(&ipv6_sysctl_net_ops);
+}
+
 void ipv6_sysctl_unregister(void)
 {
-	unregister_sysctl_table(ipv6_sysctl_header);
+	unregister_pernet_subsys(&ipv6_sysctl_net_ops);
 }
#5 lore raw

PATCH 5/10 make multiple instance of sysctl tables

In reply to Daniel Lezcano

Each network namespace wants its own set of sysctl value, eg. we should not be able from a namespace to set a sysctl value for another namespace , especially for the initial network namespace.

This patch duplicates the sysctl table when we register a new network namespace for ipv6. The duplicated table are postfixed with the "template" word to notify the developper the table is cloned.

---
 include/net/ipv6.h         |    4 +-
 include/net/netns/ipv6.h   |    9 ++++++
 net/ipv6/icmp.c            |   12 +++++++-
 net/ipv6/route.c           |   11 ++++++-
 net/ipv6/sysctl_net_ipv6.c |   67 ++++++++++++++++++++++++++++++++++++++-------
 5 files changed, 89 insertions(+), 14 deletions(-)
net/ipv6/sysctl_net_ipv6.c +57 -10
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -14,20 +14,23 @@
 #include <net/addrconf.h>
 #include <net/inet_frag.h>
 
-static ctl_table ipv6_table[] = {
+extern struct ctl_table *ipv6_route_sysctl_init(struct net *net);
+extern struct ctl_table *ipv6_icmp_sysctl_init(struct net *net);
+
+static ctl_table ipv6_table_template[] = {
 	{
 		.ctl_name	= NET_IPV6_ROUTE,
 		.procname	= "route",
 		.maxlen		= 0,
 		.mode		= 0555,
-		.child		= ipv6_route_table
+		.child		= ipv6_route_table_template
 	},
 	{
 		.ctl_name	= NET_IPV6_ICMP,
 		.procname	= "icmp",
 		.maxlen		= 0,
 		.mode		= 0555,
-		.child		= ipv6_icmp_table
+		.child		= ipv6_icmp_table_template
 	},
 	{
 		.ctl_name	= NET_IPV6_BINDV6ONLY,
@@ -89,22 +92,66 @@ struct ctl_path net_ipv6_ctl_path[] = {
 };
 EXPORT_SYMBOL_GPL(net_ipv6_ctl_path);
 
-static struct ctl_table_header *ipv6_sysctl_header;
-
 static int ipv6_sysctl_net_init(struct net *net)
 {
-	ipv6_sysctl_header = register_net_sysctl_table(net, net_ipv6_ctl_path,
-						       ipv6_table);
-	if (!ipv6_sysctl_header)
+     	struct ctl_table *ipv6_table;
+     	struct ctl_table *ipv6_route_table;
+     	struct ctl_table *ipv6_icmp_table;
+     	int err;
+
+     	err = -ENOMEM;
+     	ipv6_table = kmemdup(ipv6_table_template, sizeof(ipv6_table_template),
+     			     GFP_KERNEL);
+     	if (!ipv6_table)
+     		goto out;
+
+  	ipv6_route_table = ipv6_route_sysctl_init(net);
+  	if (!ipv6_route_table)
+  		goto out_ipv6_table;
+
+     	ipv6_icmp_table = ipv6_icmp_sysctl_init(net);
+     	if (!ipv6_icmp_table)
+     		goto out_ipv6_route_table;
+
+     	ipv6_table[0].child = ipv6_route_table;
+     	ipv6_table[1].child = ipv6_icmp_table;
+
+	net->ipv6.sysctl.table = register_net_sysctl_table(net, net_ipv6_ctl_path,
+							   ipv6_table);
+	if (!net->ipv6.sysctl.table)
 		return -ENOMEM;
 
-	return 0;
+     	if (!net->ipv6.sysctl.table)
+     		goto out_ipv6_icmp_table;
 
+     	err = 0;
+out:
+     	return err;
+
+out_ipv6_icmp_table:
+     	kfree(ipv6_icmp_table);
+out_ipv6_route_table:
+     	kfree(ipv6_route_table);
+out_ipv6_table:
+     	kfree(ipv6_table);
+     	goto out;
 }
 
 static void ipv6_sysctl_net_exit(struct net *net)
 {
-	unregister_net_sysctl_table(ipv6_sysctl_header);
+     	struct ctl_table *ipv6_table;
+     	struct ctl_table *ipv6_route_table;
+     	struct ctl_table *ipv6_icmp_table;
+
+     	ipv6_table = net->ipv6.sysctl.table->ctl_table_arg;
+     	ipv6_route_table = ipv6_table[0].child;
+     	ipv6_icmp_table = ipv6_table[1].child;
+
+     	unregister_net_sysctl_table(net->ipv6.sysctl.table);
+
+     	kfree(ipv6_table);
+     	kfree(ipv6_route_table);
+     	kfree(ipv6_icmp_table);
 }
 
 static struct pernet_operations ipv6_sysctl_net_ops = {
include/net/netns/ipv6.h +9 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -5,6 +5,15 @@
 #ifndef __NETNS_IPV6_H__
 #define __NETNS_IPV6_H__
 
+struct ctl_table_header;
+
+struct netns_sysctl_ipv6 {
+#ifdef CONFIG_SYSCTL
+	struct ctl_table_header *table;
+#endif
+};
+
 struct netns_ipv6 {
+	struct netns_sysctl_ipv6 sysctl;
 };
 #endif
include/net/ipv6.h +2 -2
Index: net-2.6.25/include/net/ipv6.h
===================================================================
@@ -619,8 +619,8 @@ static inline int snmp6_unregister_dev(s
 #endif
 
 #ifdef CONFIG_SYSCTL
-extern ctl_table ipv6_route_table[];
-extern ctl_table ipv6_icmp_table[];
+extern ctl_table ipv6_route_table_template[];
+extern ctl_table ipv6_icmp_table_template[];
 
 extern int ipv6_sysctl_register(void);
 extern void ipv6_sysctl_unregister(void);
net/ipv6/icmp.c +11 -1
Index: net-2.6.25/net/ipv6/icmp.c
===================================================================
@@ -909,7 +909,7 @@ int icmpv6_err_convert(int type, int cod
 EXPORT_SYMBOL(icmpv6_err_convert);
 
 #ifdef CONFIG_SYSCTL
-ctl_table ipv6_icmp_table[] = {
+ctl_table ipv6_icmp_table_template[] = {
 	{
 		.ctl_name	= NET_IPV6_ICMP_RATELIMIT,
 		.procname	= "ratelimit",
@@ -920,5 +920,15 @@ ctl_table ipv6_icmp_table[] = {
 	},
 	{ .ctl_name = 0 },
 };
+
+struct ctl_table *ipv6_icmp_sysctl_init(struct net *net)
+{
+	struct ctl_table *table;
+
+   	table = kmemdup(ipv6_icmp_table_template,
+			sizeof(ipv6_icmp_table_template),
+			GFP_KERNEL);
+	return table;
+}
 #endif
 
net/ipv6/route.c +10 -1
Index: net-2.6.25/net/ipv6/route.c
===================================================================
@@ -2404,7 +2404,7 @@ int ipv6_sysctl_rtcache_flush(ctl_table 
 		return -EINVAL;
 }
 
-ctl_table ipv6_route_table[] = {
+ctl_table ipv6_route_table_template[] = {
 	{
 		.procname	=	"flush",
 		.data		=	&flush_delay,
@@ -2494,6 +2494,15 @@ ctl_table ipv6_route_table[] = {
 	{ .ctl_name = 0 }
 };
 
+struct ctl_table *ipv6_route_sysctl_init(struct net *net)
+{
+	struct ctl_table *table;
+
+   	table = kmemdup(ipv6_route_table_template,
+			sizeof(ipv6_route_table_template),
+			GFP_KERNEL);
+	return table;
+}
 #endif
 
 int __init ip6_route_init(void)
#6 lore raw

PATCH 6/10 make bindv6only sysctl per namespace

In reply to Daniel Lezcano

This patch moves the bindv6only sysctl to the network namespace structure. Until the ipv6 protocol is not per namespace, the sysctl variable is always from the initial network namespace.

---
 include/net/ipv6.h         |    1 -
 include/net/netns/ipv6.h   |    1 +
 net/ipv6/af_inet6.c        |    5 ++---
 net/ipv6/sysctl_net_ipv6.c |    4 +++-
 4 files changed, 6 insertions(+), 5 deletions(-)
include/net/ipv6.h +0 -1
Index: net-2.6.25/include/net/ipv6.h
===================================================================
@@ -109,7 +109,6 @@ struct frag_hdr {
 #include <net/sock.h>
 
 /* sysctls */
-extern int sysctl_ipv6_bindv6only;
 extern int sysctl_mld_max_msf;
 
 extern struct ctl_path net_ipv6_ctl_path[];
include/net/netns/ipv6.h +1 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -11,6 +11,7 @@ struct netns_sysctl_ipv6 {
 #ifdef CONFIG_SYSCTL
 	struct ctl_table_header *table;
 #endif
+	int bindv6only;
 };
 
 struct netns_ipv6 {
net/ipv6/af_inet6.c +2 -3
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -66,8 +66,6 @@ MODULE_AUTHOR("Cast of dozens");
 MODULE_DESCRIPTION("IPv6 protocol stack for Linux");
 MODULE_LICENSE("GPL");
 
-int sysctl_ipv6_bindv6only __read_mostly;
-
 /* The inetsw6 table contains everything that inet6_create needs to
  * build a new socket.
  */
@@ -193,7 +191,7 @@ lookup_protocol:
 	np->mcast_hops	= -1;
 	np->mc_loop	= 1;
 	np->pmtudisc	= IPV6_PMTUDISC_WANT;
-	np->ipv6only	= sysctl_ipv6_bindv6only;
+	np->ipv6only	= init_net.ipv6.sysctl.bindv6only;
 
 	/* Init the ipv4 part of the socket since we can have sockets
 	 * using v6 API for ipv4.
@@ -721,6 +719,7 @@ static void cleanup_ipv6_mibs(void)
 
 static int inet6_net_init(struct net *net)
 {
+	net->ipv6.sysctl.bindv6only = 0;
 	return 0;
 }
 
net/ipv6/sysctl_net_ipv6.c +3 -1
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -35,7 +35,7 @@ static ctl_table ipv6_table_template[] =
 	{
 		.ctl_name	= NET_IPV6_BINDV6ONLY,
 		.procname	= "bindv6only",
-		.data		= &sysctl_ipv6_bindv6only,
+		.data		= &init_net.ipv6.sysctl.bindv6only,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec
@@ -116,6 +116,8 @@ static int ipv6_sysctl_net_init(struct n
      	ipv6_table[0].child = ipv6_route_table;
      	ipv6_table[1].child = ipv6_icmp_table;
 
+	ipv6_table[2].data = &net->ipv6.sysctl.bindv6only;
+
 	net->ipv6.sysctl.table = register_net_sysctl_table(net, net_ipv6_ctl_path,
 							   ipv6_table);
 	if (!net->ipv6.sysctl.table)
#7 lore raw

PATCH 7/10 make ip6_frags per namespace

In reply to Daniel Lezcano

The ip6_frags is moved to the network namespace structure. Because there can be multiple instances of the network namespaces, and the ip6_frags is no longer a global static variable, a helper function has been added to facilitate the initialization of the variables.

Until the ipv6 protocol is not per namespace, the variables are accessed relatively from the initial network namespace.

---
 include/net/ipv6.h         |    3 ---
 include/net/netns/ipv6.h   |    3 +++
 net/ipv6/af_inet6.c        |    8 ++++++++
 net/ipv6/reassembly.c      |   16 +++++++---------
 net/ipv6/sysctl_net_ipv6.c |   12 ++++++++----
 5 files changed, 26 insertions(+), 16 deletions(-)
include/net/ipv6.h +0 -3
Index: net-2.6.25/include/net/ipv6.h
===================================================================
@@ -572,9 +572,6 @@ extern int inet6_hash_connect(struct ine
 /*
  * reassembly.c
  */
-struct inet_frags_ctl;
-extern struct inet_frags_ctl ip6_frags_ctl;
-
 extern const struct proto_ops inet6_stream_ops;
 extern const struct proto_ops inet6_dgram_ops;
 
include/net/netns/ipv6.h +3 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -2,6 +2,8 @@
  * ipv6 in net namespaces
  */
 
+#include <net/inet_frag.h>
+
 #ifndef __NETNS_IPV6_H__
 #define __NETNS_IPV6_H__
 
@@ -11,6 +13,7 @@ struct netns_sysctl_ipv6 {
 #ifdef CONFIG_SYSCTL
 	struct ctl_table_header *table;
 #endif
+	struct inet_frags_ctl frags;
 	int bindv6only;
 };
 
net/ipv6/reassembly.c +7 -9
Index: net-2.6.25/net/ipv6/reassembly.c
===================================================================
@@ -82,13 +82,6 @@ struct frag_queue
 	__u16			nhoffset;
 };
 
-struct inet_frags_ctl ip6_frags_ctl __read_mostly = {
-	.high_thresh 	 = 256 * 1024,
-	.low_thresh	 = 192 * 1024,
-	.timeout	 = IPV6_FRAG_TIMEOUT,
-	.secret_interval = 10 * 60 * HZ,
-};
-
 static struct inet_frags ip6_frags;
 
 int ip6_frag_nqueues(void)
@@ -605,7 +598,7 @@ static int ipv6_frag_rcv(struct sk_buff 
 		return 1;
 	}
 
-	if (atomic_read(&ip6_frags.mem) > ip6_frags_ctl.high_thresh)
+	if (atomic_read(&ip6_frags.mem) > init_net.ipv6.sysctl.frags.high_thresh)
 		ip6_evictor(ip6_dst_idev(skb->dst));
 
 	if ((fq = fq_find(fhdr->identification, &hdr->saddr, &hdr->daddr,
@@ -632,6 +625,11 @@ static struct inet6_protocol frag_protoc
 	.flags		=	INET6_PROTO_NOPOLICY,
 };
 
+void ipv6_frag_sysctl_init(struct net *net)
+{
+	ip6_frags.ctl = &net->ipv6.sysctl.frags;
+}
+
 int __init ipv6_frag_init(void)
 {
 	int ret;
@@ -639,7 +637,7 @@ int __init ipv6_frag_init(void)
 	ret = inet6_add_protocol(&frag_protocol, IPPROTO_FRAGMENT);
 	if (ret)
 		goto out;
-	ip6_frags.ctl = &ip6_frags_ctl;
+
 	ip6_frags.hashfn = ip6_hashfn;
 	ip6_frags.constructor = ip6_frag_init;
 	ip6_frags.destructor = NULL;
net/ipv6/sysctl_net_ipv6.c +8 -4
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -43,7 +43,7 @@ static ctl_table ipv6_table_template[] =
 	{
 		.ctl_name	= NET_IPV6_IP6FRAG_HIGH_THRESH,
 		.procname	= "ip6frag_high_thresh",
-		.data		= &ip6_frags_ctl.high_thresh,
+		.data		= &init_net.ipv6.sysctl.frags.high_thresh,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec
@@ -51,7 +51,7 @@ static ctl_table ipv6_table_template[] =
 	{
 		.ctl_name	= NET_IPV6_IP6FRAG_LOW_THRESH,
 		.procname	= "ip6frag_low_thresh",
-		.data		= &ip6_frags_ctl.low_thresh,
+		.data		= &init_net.ipv6.sysctl.frags.low_thresh,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec
@@ -59,7 +59,7 @@ static ctl_table ipv6_table_template[] =
 	{
 		.ctl_name	= NET_IPV6_IP6FRAG_TIME,
 		.procname	= "ip6frag_time",
-		.data		= &ip6_frags_ctl.timeout,
+		.data		= &init_net.ipv6.sysctl.frags.timeout,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec_jiffies,
@@ -68,7 +68,7 @@ static ctl_table ipv6_table_template[] =
 	{
 		.ctl_name	= NET_IPV6_IP6FRAG_SECRET_INTERVAL,
 		.procname	= "ip6frag_secret_interval",
-		.data		= &ip6_frags_ctl.secret_interval,
+		.data		= &init_net.ipv6.sysctl.frags.secret_interval,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec_jiffies,
@@ -117,6 +117,10 @@ static int ipv6_sysctl_net_init(struct n
      	ipv6_table[1].child = ipv6_icmp_table;
 
 	ipv6_table[2].data = &net->ipv6.sysctl.bindv6only;
+      	ipv6_table[3].data = &net->ipv6.sysctl.frags.high_thresh;
+      	ipv6_table[4].data = &net->ipv6.sysctl.frags.low_thresh;
+      	ipv6_table[5].data = &net->ipv6.sysctl.frags.timeout;
+    	ipv6_table[6].data = &net->ipv6.sysctl.frags.secret_interval;
 
 	net->ipv6.sysctl.table = register_net_sysctl_table(net, net_ipv6_ctl_path,
 							   ipv6_table);
net/ipv6/af_inet6.c +8 -0
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -72,6 +72,8 @@ MODULE_LICENSE("GPL");
 static struct list_head inetsw6[SOCK_MAX];
 static DEFINE_SPINLOCK(inetsw6_lock);
 
+void ipv6_frag_sysctl_init(struct net *net);
+
 static __inline__ struct ipv6_pinfo *inet6_sk_generic(struct sock *sk)
 {
 	const int offset = sk->sk_prot->obj_size - sizeof(struct ipv6_pinfo);
@@ -720,6 +722,12 @@ static void cleanup_ipv6_mibs(void)
 static int inet6_net_init(struct net *net)
 {
 	net->ipv6.sysctl.bindv6only = 0;
+	net->ipv6.sysctl.frags.high_thresh = 256 * 1024;
+	net->ipv6.sysctl.frags.low_thresh = 192 * 1024;
+	net->ipv6.sysctl.frags.timeout = IPV6_FRAG_TIMEOUT;
+	net->ipv6.sysctl.frags.secret_interval = 10 * 60 * HZ;
+	ipv6_frag_sysctl_init(net);
+
 	return 0;
 }
 
#8 lore raw

PATCH 8/10 make mld_max_msf readonly in other namespaces

In reply to Daniel Lezcano

The mld_max_msf protects the system with a maximum allowed multicast source filters. Making this variable per namespace can be potentially an problem if someone inside a namespace set it to a big value, that will impact the whole system including other namespaces.

I don't see any benefits to have it per namespace for now, so in order to keep a directory entry in a newly created namespace, I make it read-only when we are not in the initial network namespace.

---
 net/ipv6/sysctl_net_ipv6.c |    6 ++++++
 1 file changed, 6 insertions(+)
net/ipv6/sysctl_net_ipv6.c +6 -0
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -122,6 +122,12 @@ static int ipv6_sysctl_net_init(struct n
       	ipv6_table[5].data = &net->ipv6.sysctl.frags.timeout;
     	ipv6_table[6].data = &net->ipv6.sysctl.frags.secret_interval;
 
+ 	/* We don't want this value to be per namespace, it should be global
+	   to all namespaces, so make it read-only when we are not in the
+	   init network namespace */
+    	if (net != &init_net)
+    		ipv6_table[7].mode = 0444;
+
 	net->ipv6.sysctl.table = register_net_sysctl_table(net, net_ipv6_ctl_path,
 							   ipv6_table);
 	if (!net->ipv6.sysctl.table)
#9 lore raw

PATCH 9/10 make sysctls route per namespace

In reply to Daniel Lezcano

All the sysctl concerning the routes are moved to the network namespace structure. A helper function is called to initialize the variables.

Because the ipv6 protocol is not yet per namespace, the variables are accessed relatively from the network namespace.

---
 include/net/ip6_route.h    |    2 -
 include/net/netns/ipv6.h   |    8 +++++++
 net/ipv6/af_inet6.c        |    8 +++++++
 net/ipv6/ip6_fib.c         |   14 ++++++++----
 net/ipv6/route.c           |   49 ++++++++++++++++++---------------------------
 net/ipv6/sysctl_net_ipv6.c |   11 ++++++++++
 6 files changed, 56 insertions(+), 36 deletions(-)
include/net/netns/ipv6.h +8 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -15,6 +15,14 @@ struct netns_sysctl_ipv6 {
 #endif
 	struct inet_frags_ctl frags;
 	int bindv6only;
+   	int flush_delay;
+   	int ip6_rt_max_size;
+   	int ip6_rt_gc_min_interval;
+   	int ip6_rt_gc_timeout;
+   	int ip6_rt_gc_interval;
+   	int ip6_rt_gc_elasticity;
+   	int ip6_rt_mtu_expires;
+   	int ip6_rt_min_advmss;
 };
 
 struct netns_ipv6 {
net/ipv6/route.c +20 -29

153 lines not on this page · raw message

net/ipv6/sysctl_net_ipv6.c +11 -0
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -113,7 +113,18 @@ static int ipv6_sysctl_net_init(struct n
      	if (!ipv6_icmp_table)
      		goto out_ipv6_route_table;
 
+     	ipv6_route_table[0].data = &net->ipv6.sysctl.flush_delay;
+     	/* ipv6_route_table[1].data will be handled when we have
+	   routes per namespace */
+     	ipv6_route_table[2].data = &net->ipv6.sysctl.ip6_rt_max_size;
+     	ipv6_route_table[3].data = &net->ipv6.sysctl.ip6_rt_gc_min_interval;
+     	ipv6_route_table[4].data = &net->ipv6.sysctl.ip6_rt_gc_timeout;
+     	ipv6_route_table[5].data = &net->ipv6.sysctl.ip6_rt_gc_interval;
+     	ipv6_route_table[6].data = &net->ipv6.sysctl.ip6_rt_gc_elasticity;
+     	ipv6_route_table[7].data = &net->ipv6.sysctl.ip6_rt_mtu_expires;
+     	ipv6_route_table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
      	ipv6_table[0].child = ipv6_route_table;
+
      	ipv6_table[1].child = ipv6_icmp_table;
 
 	ipv6_table[2].data = &net->ipv6.sysctl.bindv6only;
include/net/ip6_route.h +0 -2
Index: net-2.6.25/include/net/ip6_route.h
===================================================================
@@ -43,8 +43,6 @@ extern struct rt6_info	ip6_prohibit_entr
 extern struct rt6_info	ip6_blk_hole_entry;
 #endif
 
-extern int ip6_rt_gc_interval;
-
 extern void			ip6_route_input(struct sk_buff *skb);
 
 extern struct dst_entry *	ip6_route_output(struct sock *sk,
net/ipv6/ip6_fib.c +9 -5
Index: net-2.6.25/net/ipv6/ip6_fib.c
===================================================================
@@ -681,13 +681,15 @@ static __inline__ void fib6_start_gc(str
 {
 	if (ip6_fib_timer.expires == 0 &&
 	    (rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE)))
-		mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
+		mod_timer(&ip6_fib_timer, jiffies +
+			  init_net.ipv6.sysctl.ip6_rt_gc_interval);
 }
 
 void fib6_force_start_gc(void)
 {
 	if (ip6_fib_timer.expires == 0)
-		mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
+		mod_timer(&ip6_fib_timer, jiffies +
+			  init_net.ipv6.sysctl.ip6_rt_gc_interval);
 }
 
 /*
@@ -1447,7 +1449,8 @@ void fib6_run_gc(unsigned long dummy)
 {
 	if (dummy != ~0UL) {
 		spin_lock_bh(&fib6_gc_lock);
-		gc_args.timeout = dummy ? (int)dummy : ip6_rt_gc_interval;
+		gc_args.timeout = dummy ? (int)dummy :
+			init_net.ipv6.sysctl.ip6_rt_gc_interval;
 	} else {
 		local_bh_disable();
 		if (!spin_trylock(&fib6_gc_lock)) {
@@ -1455,7 +1458,7 @@ void fib6_run_gc(unsigned long dummy)
 			local_bh_enable();
 			return;
 		}
-		gc_args.timeout = ip6_rt_gc_interval;
+		gc_args.timeout = init_net.ipv6.sysctl.ip6_rt_gc_interval;
 	}
 	gc_args.more = 0;
 
@@ -1463,7 +1466,8 @@ void fib6_run_gc(unsigned long dummy)
 	fib6_clean_all(fib6_age, 0, NULL);
 
 	if (gc_args.more)
-		mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
+		mod_timer(&ip6_fib_timer, jiffies +
+			  init_net.ipv6.sysctl.ip6_rt_gc_interval);
 	else {
 		del_timer(&ip6_fib_timer);
 		ip6_fib_timer.expires = 0;
net/ipv6/af_inet6.c +8 -0
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -726,6 +726,14 @@ static int inet6_net_init(struct net *ne
 	net->ipv6.sysctl.frags.low_thresh = 192 * 1024;
 	net->ipv6.sysctl.frags.timeout = IPV6_FRAG_TIMEOUT;
 	net->ipv6.sysctl.frags.secret_interval = 10 * 60 * HZ;
+ 	net->ipv6.sysctl.flush_delay = 0;
+ 	net->ipv6.sysctl.ip6_rt_max_size = 4096;
+ 	net->ipv6.sysctl.ip6_rt_gc_min_interval = HZ / 2;
+ 	net->ipv6.sysctl.ip6_rt_gc_timeout = 60*HZ;
+ 	net->ipv6.sysctl.ip6_rt_gc_interval = 30*HZ;
+ 	net->ipv6.sysctl.ip6_rt_gc_elasticity = 9;
+ 	net->ipv6.sysctl.ip6_rt_mtu_expires = 10*60*HZ;
+ 	net->ipv6.sysctl.ip6_rt_min_advmss = IPV6_MIN_MTU - 20 - 40;
 	ipv6_frag_sysctl_init(net);
 
 	return 0;
#10 lore raw

PATCH 10/10 make icmpv6_time sysctl per namespace

In reply to Daniel Lezcano

This patch moves the icmpv6_time sysctl to the network namespace structure.

Because the ipv6 protocol is not yet per namespace, the variable is accessed relatively to the initial network namespace.

---
 include/net/netns/ipv6.h   |    1 +
 net/ipv6/af_inet6.c        |    1 +
 net/ipv6/icmp.c            |    6 ++----
 net/ipv6/sysctl_net_ipv6.c |    1 +
 4 files changed, 5 insertions(+), 4 deletions(-)
include/net/netns/ipv6.h +1 -0
Index: net-2.6.25/include/net/netns/ipv6.h
===================================================================
@@ -23,6 +23,7 @@ struct netns_sysctl_ipv6 {
    	int ip6_rt_gc_elasticity;
    	int ip6_rt_mtu_expires;
    	int ip6_rt_min_advmss;
+	int icmpv6_time;
 };
 
 struct netns_ipv6 {
net/ipv6/icmp.c +2 -4
Index: net-2.6.25/net/ipv6/icmp.c
===================================================================
@@ -154,8 +154,6 @@ static int is_ineligible(struct sk_buff 
 	return 0;
 }
 
-static int sysctl_icmpv6_time __read_mostly = 1*HZ;
-
 /*
  * Check the ICMP output rate limit
  */
@@ -186,7 +184,7 @@ static inline int icmpv6_xrlim_allow(str
 		res = 1;
 	} else {
 		struct rt6_info *rt = (struct rt6_info *)dst;
-		int tmo = sysctl_icmpv6_time;
+		int tmo = init_net.ipv6.sysctl.icmpv6_time;
 
 		/* Give more bandwidth to wider prefixes. */
 		if (rt->rt6i_dst.plen < 128)
@@ -913,7 +911,7 @@ ctl_table ipv6_icmp_table_template[] = {
 	{
 		.ctl_name	= NET_IPV6_ICMP_RATELIMIT,
 		.procname	= "ratelimit",
-		.data		= &sysctl_icmpv6_time,
+		.data		= &init_net.ipv6.sysctl.icmpv6_time,
 		.maxlen		= sizeof(int),
 		.mode		= 0644,
 		.proc_handler	= &proc_dointvec
net/ipv6/sysctl_net_ipv6.c +1 -0
Index: net-2.6.25/net/ipv6/sysctl_net_ipv6.c
===================================================================
@@ -125,6 +125,7 @@ static int ipv6_sysctl_net_init(struct n
      	ipv6_route_table[8].data = &net->ipv6.sysctl.ip6_rt_min_advmss;
      	ipv6_table[0].child = ipv6_route_table;
 
+ 	ipv6_icmp_table[0].data = &net->ipv6.sysctl.icmpv6_time;
      	ipv6_table[1].child = ipv6_icmp_table;
 
 	ipv6_table[2].data = &net->ipv6.sysctl.bindv6only;
net/ipv6/af_inet6.c +1 -0
Index: net-2.6.25/net/ipv6/af_inet6.c
===================================================================
@@ -734,6 +734,7 @@ static int inet6_net_init(struct net *ne
  	net->ipv6.sysctl.ip6_rt_gc_elasticity = 9;
  	net->ipv6.sysctl.ip6_rt_mtu_expires = 10*60*HZ;
  	net->ipv6.sysctl.ip6_rt_min_advmss = IPV6_MIN_MTU - 20 - 40;
+ 	net->ipv6.sysctl.icmpv6_time = 1*HZ;
 	ipv6_frag_sysctl_init(net);
 
 	return 0;
maintainer
#11 lore raw
In reply to Daniel Lezcano

From: Daniel Lezcano <[email protected]>
Date: Wed, 09 Jan 2008 17:45:33 +0100

Quoted 5 lines

The following patchset makes the ipv6 sysctl to handle multiple network namespaces. Each instance of a network namespace as its own set of sysctl values, that means the behavior of the ipv6 stack can be different depending on the sysctl values setup in the different network namespaces.

I applied all of this to net-2.6.25 but what a rough half hour it was :-/

Starting at patch #5 there were tons of "space before tab" errors. And as I fixed them up, this made subsequent patches need rediffing since the contextual lines in patches after #5 needed the whitespace fixed up as well.

I didn't push this back to you because this was already the 3rd round, but please show me some love and check this stuff out before submission. GIT gives you effective ways to verify the whitespace without even applying the patch.

~davem/bin/pcheck:

#!/bin/sh
set -x
git apply --check --whitespace=error-all $1

#12 lore raw
Quoted 22 lines · David Miller

From: Daniel Lezcano <[email protected]>
Date: Wed, 09 Jan 2008 17:45:33 +0100

Quoted 5 lines

The following patchset makes the ipv6 sysctl to handle multiple network namespaces. Each instance of a network namespace as its own set of sysctl values, that means the behavior of the ipv6 stack can be different depending on the sysctl values setup in the different network namespaces.

I applied all of this to net-2.6.25 but what a rough half hour it was :-/

Starting at patch #5 there were tons of "space before tab" errors. And as I fixed them up, this made subsequent patches need rediffing since the contextual lines in patches after #5 needed the whitespace fixed up as well.

I didn't push this back to you because this was already the 3rd round, but please show me some love and check this stuff out before submission. GIT gives you effective ways to verify the whitespace without even applying the patch.

Quoted 5 lines

~davem/bin/pcheck:

#!/bin/sh
set -x
git apply --check --whitespace=error-all $1

Sorry, I will check that in the future :|
Many thanks for taking the time to fix that.

-- Daniel.