Make ipv4_devconf (all and default) live in net namespaces

PATCH Applied Pavel Emelyanov 3 participants 15 messages Dec 11, 2007 – Dec 16, 2007 1 message not in the archive

Subsystems (3)

Networking Network drivers IPv4/IPv6

lore

Share & tools

lore

Pavel Emelyanov

author

2007-12-11 17:44 UTC #0 lore raw

cover Make ipv4_devconf (all and default) live in net namespaces

The ipv4_devconf_(all) and ipv4_devconf_dflt are currently global, but should be per-namespace.

This set moves them on the struct net. Or, more precisely, on the struct netns_ipv4, which in turn is on the struct net.

There are two minor things that are to be done additionally to this set:

1. The snmp_seq_show() needs the IPV4_DEVCONF_ALL(FORWARDING) value, but since this entry is still global no valid struct net can be get in it, so I use the init_net's one. After snmp is made per-namespace, this will be fixed easily.

2. The rt_fill_info() needs the IPV4_DEVCONF_ALL(MC_FORWARDING), but the routing code is not tuned to work inside namespaces yet, so I use the init_net in it as well. Denis is currently working on ipv4 routing, so this will be prepared shortly.

Happily, all the other places of devconf-s usage can provide a struct net pointer.

Signed-off-by Pavel Emelyanov <[email protected]>

Pavel Emelyanov

author

2007-12-11 17:45 UTC #1 lore raw

PATCH 1/7 Add the netns_ipv4 struct

The ipv4 will store its parameters inside this structure. This one is empty now, but it will be eventually filled.

Signed-off-by Pavel Emelyanov <[email protected]>

---

include/net/net_namespace.h +2 -0

index 18da0af..d04ddf2 100644
@@ -10,6 +10,7 @@
 
 #include <net/netns/unix.h>
 #include <net/netns/packet.h>
+#include <net/netns/ipv4.h>
 
 struct proc_dir_entry;
 struct net_device;

@@ -46,6 +47,7 @@ struct net {
 
 	struct netns_packet	packet;
 	struct netns_unix	unx;
+	struct netns_ipv4	ipv4;
 };
 
 #ifdef CONFIG_NET

include/net/netns/ipv4.h new file +9 -0

new file mode 100644
index 0000000..ce830d5
@@ -0,0 +1,9 @@
+/*
+ * ipv4 in net namespaces
+ */
+
+#ifndef __NETNS_IPV4_H__
+#define __NETNS_IPV4_H__
+struct netns_ipv4 {
+};
+#endif

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:29 UTC #2 lore raw

Re: [PATCH net-2.6.25 1/7] Add the netns_ipv4 struct

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:45:58 +0300

Quoted 4 lines

The ipv4 will store its parameters inside this structure. This one is empty now, but it will be eventually filled.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 17:48 UTC #3 lore raw

PATCH 2/7 Make __devinet_sysctl_register return an error

In reply to Pavel Emelyanov

Currently, this function is void, so failures in creating sysctls for new/renamed devices are not reported to anywhere.

Fixing this is another complex (needed?) task, but this return value is needed during the namespaces creation to handle the case, when we failed to create "all" and "default" entries.

Signed-off-by Pavel Emelyanov <[email protected]>

---

net/ipv4/devinet.c +4 -4

index 872883e..bfb0fb0 100644
@@ -1455,8 +1460,8 @@ static struct devinet_sysctl_table {
 	},
 };
 
-static void __devinet_sysctl_register(char *dev_name, int ctl_name,
-		struct ipv4_devconf *p)
+static int __devinet_sysctl_register(struct net *net, char *dev_name,
+		int ctl_name, struct ipv4_devconf *p)
 {
 	int i;
 	struct devinet_sysctl_table *t;

@@ -1498,14 +1504,14 @@ static void __devinet_sysctl_register(char *dev_name, int ctl_name,
 		goto free_procname;
 
 	p->sysctl = t;
-	return;
+	return 0;
 
 free_procname:
 	kfree(t->dev_name);
 free:
 	kfree(t);
 out:
-	return;
+	return -ENOBUFS;
 }
 
 static void __devinet_sysctl_unregister(struct ipv4_devconf *cnf)

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:30 UTC #4 lore raw

Re: [PATCH net-2.6.25 2/7] Make __devinet_sysctl_register return an error

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:48:44 +0300

Quoted 9 lines

Currently, this function is void, so failures in creating sysctls for new/renamed devices are not reported to anywhere.
Fixing this is another complex (needed?) task, but this return value is needed during the namespaces creation to handle the case, when we failed to create "all" and "default" entries.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 17:50 UTC #5 lore raw

PATCH 3/7 Pass the net pointer to the arp_req_set_proxy()

In reply to Pavel Emelyanov

This one will need to set the IPV4_DEVCONF_ALL(PROXY_ARP), but there's no ways to get the net right in place, so we have to pull one from the inet_ioctl's struct sock.

Signed-off-by Pavel Emelyanov <[email protected]>

---

include/net/arp.h +1 -1

index f026645..36482bf 100644
@@ -11,7 +11,7 @@ extern struct neigh_table arp_tbl;
 
 extern void	arp_init(void);
 extern int	arp_find(unsigned char *haddr, struct sk_buff *skb);
-extern int	arp_ioctl(unsigned int cmd, void __user *arg);
+extern int	arp_ioctl(struct net *net, unsigned int cmd, void __user *arg);
 extern void     arp_send(int type, int ptype, __be32 dest_ip,
 			 struct net_device *dev, __be32 src_ip,
 			 unsigned char *dest_hw, unsigned char *src_hw, unsigned char *th);

net/ipv4/af_inet.c +1 -1

index 7f8b27f..a69b0c6 100644
@@ -798,7 +798,7 @@ int inet_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
 		case SIOCDARP:
 		case SIOCGARP:
 		case SIOCSARP:
-			err = arp_ioctl(cmd, (void __user *)arg);
+			err = arp_ioctl(sk->sk_net, cmd, (void __user *)arg);
 			break;
 		case SIOCGIFADDR:
 		case SIOCSIFADDR:

net/ipv4/arp.c +16 ~~-12~~

index 48e7bf6..41a4d73 100644
@@ -953,7 +953,7 @@ out_of_mem:
  *	Set (create) an ARP cache entry.
  */
 
-static int arp_req_set_proxy(struct net_device *dev, int on)
+static int arp_req_set_proxy(struct net *net, struct net_device *dev, int on)
 {
 	if (dev == NULL) {
 		IPV4_DEVCONF_ALL(PROXY_ARP) = on;

@@ -966,7 +966,8 @@ static int arp_req_set_proxy(struct net_device *dev, int on)
 	return -ENXIO;
 }
 
-static int arp_req_set_public(struct arpreq *r, struct net_device *dev)
+static int arp_req_set_public(struct net *net, struct arpreq *r,
+		struct net_device *dev)
 {
 	__be32 ip = ((struct sockaddr_in *)&r->arp_pa)->sin_addr.s_addr;
 	__be32 mask = ((struct sockaddr_in *)&r->arp_netmask)->sin_addr.s_addr;

@@ -985,17 +986,18 @@ static int arp_req_set_public(struct arpreq *r, struct net_device *dev)
 		return 0;
 	}
 
-	return arp_req_set_proxy(dev, 1);
+	return arp_req_set_proxy(net, dev, 1);
 }
 
-static int arp_req_set(struct arpreq *r, struct net_device * dev)
+static int arp_req_set(struct net *net, struct arpreq *r,
+		struct net_device * dev)
 {
 	__be32 ip;
 	struct neighbour *neigh;
 	int err;
 
 	if (r->arp_flags & ATF_PUBL)
-		return arp_req_set_public(r, dev);
+		return arp_req_set_public(net, r, dev);
 
 	ip = ((struct sockaddr_in *)&r->arp_pa)->sin_addr.s_addr;
 	if (r->arp_flags & ATF_PERM)

@@ -1081,7 +1083,8 @@ static int arp_req_get(struct arpreq *r, struct net_device *dev)
 	return err;
 }
 
-static int arp_req_delete_public(struct arpreq *r, struct net_device *dev)
+static int arp_req_delete_public(struct net *net, struct arpreq *r,
+		struct net_device *dev)
 {
 	__be32 ip = ((struct sockaddr_in *) &r->arp_pa)->sin_addr.s_addr;
 	__be32 mask = ((struct sockaddr_in *)&r->arp_netmask)->sin_addr.s_addr;

@@ -1092,17 +1095,18 @@ static int arp_req_delete_public(struct arpreq *r, struct net_device *dev)
 	if (mask)
 		return -EINVAL;
 
-	return arp_req_set_proxy(dev, 0);
+	return arp_req_set_proxy(net, dev, 0);
 }
 
-static int arp_req_delete(struct arpreq *r, struct net_device * dev)
+static int arp_req_delete(struct net *net, struct arpreq *r,
+		struct net_device * dev)
 {
 	int err;
 	__be32 ip;
 	struct neighbour *neigh;
 
 	if (r->arp_flags & ATF_PUBL)
-		return arp_req_delete_public(r, dev);
+		return arp_req_delete_public(net, r, dev);
 
 	ip = ((struct sockaddr_in *)&r->arp_pa)->sin_addr.s_addr;
 	if (dev == NULL) {

@@ -1132,7 +1136,7 @@ static int arp_req_delete(struct arpreq *r, struct net_device * dev)
  *	Handle an ARP layer I/O control request.
  */
 
-int arp_ioctl(unsigned int cmd, void __user *arg)
+int arp_ioctl(struct net *net, unsigned int cmd, void __user *arg)
 {
 	int err;
 	struct arpreq r;

@@ -1180,10 +1184,10 @@ int arp_ioctl(unsigned int cmd, void __user *arg)
 
 	switch (cmd) {
 	case SIOCDARP:
-		err = arp_req_delete(&r, dev);
+		err = arp_req_delete(net, &r, dev);
 		break;
 	case SIOCSARP:
-		err = arp_req_set(&r, dev);
+		err = arp_req_set(net, &r, dev);
 		break;
 	case SIOCGARP:
 		err = arp_req_get(&r, dev);

Signature · 1 line

1.5.3.4

A message that is not in the archive (sent off-list or not yet received): [email protected]

David Miller

Applied

2007-12-16 21:30 UTC #7 lore raw

Re: [PATCH net-2.6.25 3/7] Pass the net pointer to the arp_req_set_proxy()

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:50:59 +0300

Quoted 5 lines

This one will need to set the IPV4_DEVCONF_ALL(PROXY_ARP), but there's no ways to get the net right in place, so we have to pull one from the inet_ioctl's struct sock.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 17:53 UTC #8 lore raw

PATCH 4/7 Store the net pointer on devinet's ctl tables

In reply to Pavel Emelyanov

Some handers and strategies of devinet sysctl tables need to know the net to propagate the ctl change to all the net devices.

I use the (currently unused) extra2 pointer on the tables to get it.

Holding the reference on the struct net is not possible, because otherwise we'll get a net->ctl_table->net circular dependency. But since the ctl tables are unregistered during the net destruction, this is safe to get it w/o additional protection.

Signed-off-by Pavel Emelyanov <[email protected]>

---

net/ipv4/devinet.c +18 ~~-11~~

index 872883e..bfb0fb0 100644
@@ -1237,12 +1237,12 @@ errout:
 
 #ifdef CONFIG_SYSCTL
 
-static void devinet_copy_dflt_conf(int i)
+static void devinet_copy_dflt_conf(struct net *net, int i)
 {
 	struct net_device *dev;
 
 	read_lock(&dev_base_lock);
-	for_each_netdev(&init_net, dev) {
+	for_each_netdev(net, dev) {
 		struct in_device *in_dev;
 		rcu_read_lock();
 		in_dev = __in_dev_get_rcu(dev);

@@ -1253,7 +1253,7 @@ static void devinet_copy_dflt_conf(int i)
 	read_unlock(&dev_base_lock);
 }
 
-static void inet_forward_change(void)
+static void inet_forward_change(struct net *net)
 {
 	struct net_device *dev;
 	int on = IPV4_DEVCONF_ALL(FORWARDING);

@@ -1262,7 +1262,7 @@ static void inet_forward_change(void)
 	IPV4_DEVCONF_DFLT(FORWARDING) = on;
 
 	read_lock(&dev_base_lock);
-	for_each_netdev(&init_net, dev) {
+	for_each_netdev(net, dev) {
 		struct in_device *in_dev;
 		rcu_read_lock();
 		in_dev = __in_dev_get_rcu(dev);

@@ -1283,12 +1283,13 @@ static int devinet_conf_proc(ctl_table *ctl, int write,
 
 	if (write) {
 		struct ipv4_devconf *cnf = ctl->extra1;
+		struct net *net = ctl->extra2;
 		int i = (int *)ctl->data - cnf->data;
 
 		set_bit(i, cnf->state);
 
 		if (cnf == &ipv4_devconf_dflt)
-			devinet_copy_dflt_conf(i);
+			devinet_copy_dflt_conf(net, i);
 	}
 
 	return ret;

@@ -1299,6 +1300,7 @@ static int devinet_conf_sysctl(ctl_table *table, int __user *name, int nlen,
 			       void __user *newval, size_t newlen)
 {
 	struct ipv4_devconf *cnf;
+	struct net *net;
 	int *valp = table->data;
 	int new;
 	int i;

@@ -1334,12 +1336,13 @@ static int devinet_conf_sysctl(ctl_table *table, int __user *name, int nlen,
 	*valp = new;
 
 	cnf = table->extra1;
+	net = table->extra2;
 	i = (int *)table->data - cnf->data;
 
 	set_bit(i, cnf->state);
 
 	if (cnf == &ipv4_devconf_dflt)
-		devinet_copy_dflt_conf(i);
+		devinet_copy_dflt_conf(net, i);
 
 	return 1;
 }

@@ -1353,8 +1356,10 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write,
 	int ret = proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
 
 	if (write && *valp != val) {
+		struct net *net = ctl->extra2;
+
 		if (valp == &IPV4_DEVCONF_ALL(FORWARDING))
-			inet_forward_change();
+			inet_forward_change(net);
 		else if (valp != &IPV4_DEVCONF_DFLT(FORWARDING))
 			rt_cache_flush(0);
 	}

@@ -1478,6 +1483,7 @@ static void __devinet_sysctl_register(char *dev_name, int ctl_name,
 	for (i = 0; i < ARRAY_SIZE(t->devinet_vars) - 1; i++) {
 		t->devinet_vars[i].data += (char *)p - (char *)&ipv4_devconf;
 		t->devinet_vars[i].extra1 = p;
+		t->devinet_vars[i].extra2 = net;
 	}
 
 	/*

@@ -1525,8 +1531,8 @@ static void devinet_sysctl_register(struct in_device *idev)
 {
 	neigh_sysctl_register(idev->dev, idev->arp_parms, NET_IPV4,
 			NET_IPV4_NEIGH, "ipv4", NULL, NULL);
-	__devinet_sysctl_register(idev->dev->name, idev->dev->ifindex,
-			&idev->cnf);
+	__devinet_sysctl_register(idev->dev->nd_net, idev->dev->name,
+			idev->dev->ifindex, &idev->cnf);
 }
 
 static void devinet_sysctl_unregister(struct in_device *idev)

@@ -1547,6 +1553,7 @@ static struct ctl_table ctl_forward_entry[] = {
 		.proc_handler	= devinet_sysctl_forward,
 		.strategy	= devinet_conf_sysctl,
 		.extra1		= &ipv4_devconf,
+		.extra2		= &init_net,
 	},
 	{ },
 };

@@ -1566,9 +1573,9 @@ void __init devinet_init(void)
 	rtnl_register(PF_INET, RTM_DELADDR, inet_rtm_deladdr, NULL);
 	rtnl_register(PF_INET, RTM_GETADDR, NULL, inet_dump_ifaddr);
 #ifdef CONFIG_SYSCTL
-	__devinet_sysctl_register("all", NET_PROTO_CONF_ALL,
+	__devinet_sysctl_register(&init_net, "all", NET_PROTO_CONF_ALL,
 			&ipv4_devconf);
-	__devinet_sysctl_register("default", NET_PROTO_CONF_DEFAULT,
+	__devinet_sysctl_register(&init_net, "default", NET_PROTO_CONF_DEFAULT,
 			&ipv4_devconf_dflt);
 	register_sysctl_paths(net_ipv4_path, ctl_forward_entry);
 #endif

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:31 UTC #9 lore raw

Re: [PATCH net-2.6.25 4/7] Store the net pointer on devinet's ctl tables

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:53:55 +0300

Quoted 7 lines

Some handers and strategies of devinet sysctl tables need to know the net to propagate the ctl change to all the net devices.
I use the (currently unused) extra2 pointer on the tables to get it.

Quoted 7 lines

Holding the reference on the struct net is not possible, because otherwise we'll get a net->ctl_table->net circular dependency. But since the ctl tables are unregistered during the net destruction, this is safe to get it w/o additional protection.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 17:57 UTC #10 lore raw

PATCH 5/7 Move the devinet pointers on the struct net

In reply to Pavel Emelyanov

This is the core.

Add all and default pointers on the netns_ipv4 and register a new pernet subsys to initialize them.

Also add the ctl_table_header to register the
net.ipv4.ip_forward ctl.

I don't allocate additional memory for init_net, but use global devinets.

Signed-off-by Pavel Emelyanov <[email protected]>

---

include/net/netns/ipv4.h +6 -0

index ce830d5..e06d7cf 100644
@@ -4,6 +4,12 @@
 
 #ifndef __NETNS_IPV4_H__
 #define __NETNS_IPV4_H__
+struct ctl_table_header;
+struct ipv4_devconf;
+
 struct netns_ipv4 {
+	struct ctl_table_header	*forw_hdr;
+	struct ipv4_devconf	*devconf_all;
+	struct ipv4_devconf	*devconf_dflt;
 };
 #endif

net/ipv4/devinet.c +96 -9

136 lines not on this page · raw message

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:31 UTC #11 lore raw

Re: [PATCH net-2.6.25 5/7] Move the devinet pointers on the struct net

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:57:18 +0300

Quoted 5 lines

This is the core.
Add all and default pointers on the netns_ipv4 and register a new pernet subsys to initialize them.

Quoted 7 lines

Also add the ctl_table_header to register the
net.ipv4.ip_forward ctl.
I don't allocate additional memory for init_net, but use global devinets.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 17:59 UTC #12 lore raw

PATCH 6/7 Switch users of ipv4_devconf_dflt to use the pernet one

In reply to Pavel Emelyanov

They are all collected in the net/ipv4/devinet.c file and mostly use the IPV4_DEVCONF_DFLT macro.

So I add the net parameter to it and patch users accordingly.

Signed-off-by Pavel Emelyanov <[email protected]>

---

net/ipv4/devinet.c +9 -7

index 996f07e..cd957f4 100644
@@ -83,7 +83,8 @@ static struct ipv4_devconf ipv4_devconf_dflt = {
 	},
 };
 
-#define IPV4_DEVCONF_DFLT(attr) IPV4_DEVCONF(ipv4_devconf_dflt, attr)
+#define IPV4_DEVCONF_DFLT(net, attr) \
+	IPV4_DEVCONF((*net->ipv4.devconf_dflt), attr)
 
 static const struct nla_policy ifa_ipv4_policy[IFA_MAX+1] = {
 	[IFA_LOCAL]     	= { .type = NLA_U32 },

@@ -164,7 +165,8 @@ static struct in_device *inetdev_init(struct net_device *dev)
 	if (!in_dev)
 		goto out;
 	INIT_RCU_HEAD(&in_dev->rcu_head);
-	memcpy(&in_dev->cnf, &ipv4_devconf_dflt, sizeof(in_dev->cnf));
+	memcpy(&in_dev->cnf, dev->nd_net->ipv4.devconf_dflt,
+			sizeof(in_dev->cnf));
 	in_dev->cnf.sysctl = NULL;
 	in_dev->dev = dev;
 	if ((in_dev->arp_parms = neigh_parms_alloc(dev, &arp_tbl)) == NULL)

@@ -1248,7 +1250,7 @@ static void devinet_copy_dflt_conf(struct net *net, int i)
 		rcu_read_lock();
 		in_dev = __in_dev_get_rcu(dev);
 		if (in_dev && !test_bit(i, in_dev->cnf.state))
-			in_dev->cnf.data[i] = ipv4_devconf_dflt.data[i];
+			in_dev->cnf.data[i] = net->ipv4.devconf_dflt->data[i];
 		rcu_read_unlock();
 	}
 	read_unlock(&dev_base_lock);

@@ -1260,7 +1262,7 @@ static void inet_forward_change(struct net *net)
 	int on = IPV4_DEVCONF_ALL(FORWARDING);
 
 	IPV4_DEVCONF_ALL(ACCEPT_REDIRECTS) = !on;
-	IPV4_DEVCONF_DFLT(FORWARDING) = on;
+	IPV4_DEVCONF_DFLT(net, FORWARDING) = on;
 
 	read_lock(&dev_base_lock);
 	for_each_netdev(net, dev) {

@@ -1289,7 +1291,7 @@ static int devinet_conf_proc(ctl_table *ctl, int write,
 
 		set_bit(i, cnf->state);
 
-		if (cnf == &ipv4_devconf_dflt)
+		if (cnf == net->ipv4.devconf_dflt)
 			devinet_copy_dflt_conf(net, i);
 	}

@@ -1342,7 +1344,7 @@ static int devinet_conf_sysctl(ctl_table *table, int __user *name, int nlen,
 
 	set_bit(i, cnf->state);
 
-	if (cnf == &ipv4_devconf_dflt)
+	if (cnf == net->ipv4.devconf_dflt)
 		devinet_copy_dflt_conf(net, i);
 
 	return 1;

@@ -1361,7 +1363,7 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write,
 
 		if (valp == &IPV4_DEVCONF_ALL(FORWARDING))
 			inet_forward_change(net);
-		else if (valp != &IPV4_DEVCONF_DFLT(FORWARDING))
+		else if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING))
 			rt_cache_flush(0);
 	}

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:32 UTC #13 lore raw

Re: [PATCH net-2.6.25 6/7] Switch users of ipv4_devconf_dflt to use the pernet one

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 20:59:58 +0300

Quoted 6 lines

They are all collected in the net/ipv4/devinet.c file and mostly use the IPV4_DEVCONF_DFLT macro.
So I add the net parameter to it and patch users accordingly.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.

Pavel Emelyanov

author

2007-12-11 18:02 UTC #14 lore raw

PATCH 7/7 Switch users of ipv4_devconf(_all) to use the pernet one

In reply to Pavel Emelyanov

These are scattered over the code, but almost all the "critical" places already have the proper struct net at hand except for snmp proc showing function and routing rtnl handler.

Signed-off-by Pavel Emelyanov <[email protected]>

---

include/linux/inetdevice.h +8 -4

index 962a062..b3c5081 100644
@@ -44,7 +44,8 @@ struct in_device
 };
 
 #define IPV4_DEVCONF(cnf, attr) ((cnf).data[NET_IPV4_CONF_ ## attr - 1])
-#define IPV4_DEVCONF_ALL(attr) IPV4_DEVCONF(ipv4_devconf, attr)
+#define IPV4_DEVCONF_ALL(net, attr) \
+	IPV4_DEVCONF((*(net)->ipv4.devconf_all), attr)
 
 static inline int ipv4_devconf_get(struct in_device *in_dev, int index)
 {

@@ -71,11 +72,14 @@ static inline void ipv4_devconf_setall(struct in_device *in_dev)
 	ipv4_devconf_set((in_dev), NET_IPV4_CONF_ ## attr, (val))
 
 #define IN_DEV_ANDCONF(in_dev, attr) \
-	(IPV4_DEVCONF_ALL(attr) && IN_DEV_CONF_GET((in_dev), attr))
+	(IPV4_DEVCONF_ALL(in_dev->dev->nd_net, attr) && \
+	 IN_DEV_CONF_GET((in_dev), attr))
 #define IN_DEV_ORCONF(in_dev, attr) \
-	(IPV4_DEVCONF_ALL(attr) || IN_DEV_CONF_GET((in_dev), attr))
+	(IPV4_DEVCONF_ALL(in_dev->dev->nd_net, attr) || \
+	 IN_DEV_CONF_GET((in_dev), attr))
 #define IN_DEV_MAXCONF(in_dev, attr) \
-	(max(IPV4_DEVCONF_ALL(attr), IN_DEV_CONF_GET((in_dev), attr)))
+	(max(IPV4_DEVCONF_ALL(in_dev->dev->nd_net, attr), \
+	     IN_DEV_CONF_GET((in_dev), attr)))
 
 #define IN_DEV_FORWARD(in_dev)		IN_DEV_CONF_GET((in_dev), FORWARDING)
 #define IN_DEV_MFORWARD(in_dev)		IN_DEV_ANDCONF((in_dev), MC_FORWARDING)

net/ipv4/arp.c +2 -2

index 41a4d73..cd4c95a 100644
@@ -861,7 +861,7 @@ static int arp_process(struct sk_buff *skb)
 
 	n = __neigh_lookup(&arp_tbl, &sip, dev, 0);
 
-	if (IPV4_DEVCONF_ALL(ARP_ACCEPT)) {
+	if (IPV4_DEVCONF_ALL(dev->nd_net, ARP_ACCEPT)) {
 		/* Unsolicited ARP is not accepted by default.
 		   It is possible, that this option should be enabled for some
 		   devices (strip is candidate)

@@ -956,7 +956,7 @@ out_of_mem:
 static int arp_req_set_proxy(struct net *net, struct net_device *dev, int on)
 {
 	if (dev == NULL) {
-		IPV4_DEVCONF_ALL(PROXY_ARP) = on;
+		IPV4_DEVCONF_ALL(net, PROXY_ARP) = on;
 		return 0;
 	}
 	if (__in_dev_get_rtnl(dev)) {

net/ipv4/devinet.c +3 -3

index cd957f4..e94de54 100644
@@ -1259,9 +1259,9 @@ static void devinet_copy_dflt_conf(struct net *net, int i)
 static void inet_forward_change(struct net *net)
 {
 	struct net_device *dev;
-	int on = IPV4_DEVCONF_ALL(FORWARDING);
+	int on = IPV4_DEVCONF_ALL(net, FORWARDING);
 
-	IPV4_DEVCONF_ALL(ACCEPT_REDIRECTS) = !on;
+	IPV4_DEVCONF_ALL(net, ACCEPT_REDIRECTS) = !on;
 	IPV4_DEVCONF_DFLT(net, FORWARDING) = on;
 
 	read_lock(&dev_base_lock);

@@ -1361,7 +1361,7 @@ static int devinet_sysctl_forward(ctl_table *ctl, int write,
 	if (write && *valp != val) {
 		struct net *net = ctl->extra2;
 
-		if (valp == &IPV4_DEVCONF_ALL(FORWARDING))
+		if (valp == &IPV4_DEVCONF_ALL(net, FORWARDING))
 			inet_forward_change(net);
 		else if (valp != &IPV4_DEVCONF_DFLT(net, FORWARDING))
 			rt_cache_flush(0);

net/ipv4/igmp.c +2 -2

index c560a93..d3d5906 100644
@@ -130,12 +130,12 @@
  */
 
 #define IGMP_V1_SEEN(in_dev) \
-	(IPV4_DEVCONF_ALL(FORCE_IGMP_VERSION) == 1 || \
+	(IPV4_DEVCONF_ALL(in_dev->dev->nd_net, FORCE_IGMP_VERSION) == 1 || \
 	 IN_DEV_CONF_GET((in_dev), FORCE_IGMP_VERSION) == 1 || \
 	 ((in_dev)->mr_v1_seen && \
 	  time_before(jiffies, (in_dev)->mr_v1_seen)))
 #define IGMP_V2_SEEN(in_dev) \
-	(IPV4_DEVCONF_ALL(FORCE_IGMP_VERSION) == 2 || \
+	(IPV4_DEVCONF_ALL(in_dev->dev->nd_net, FORCE_IGMP_VERSION) == 2 || \
 	 IN_DEV_CONF_GET((in_dev), FORCE_IGMP_VERSION) == 2 || \
 	 ((in_dev)->mr_v2_seen && \
 	  time_before(jiffies, (in_dev)->mr_v2_seen)))

net/ipv4/ipmr.c +2 -2

index 1187928..9947f52 100644
@@ -849,7 +849,7 @@ static void mrtsock_destruct(struct sock *sk)
 {
 	rtnl_lock();
 	if (sk == mroute_socket) {
-		IPV4_DEVCONF_ALL(MC_FORWARDING)--;
+		IPV4_DEVCONF_ALL(sk->sk_net, MC_FORWARDING)--;
 
 		write_lock_bh(&mrt_lock);
 		mroute_socket=NULL;

@@ -898,7 +898,7 @@ int ip_mroute_setsockopt(struct sock *sk,int optname,char __user *optval,int opt
 			mroute_socket=sk;
 			write_unlock_bh(&mrt_lock);
 
-			IPV4_DEVCONF_ALL(MC_FORWARDING)++;
+			IPV4_DEVCONF_ALL(sk->sk_net, MC_FORWARDING)++;
 		}
 		rtnl_unlock();
 		return ret;

net/ipv4/proc.c +2 -1

index ce34b28..41734db 100644
@@ -309,7 +309,8 @@ static int snmp_seq_show(struct seq_file *seq, void *v)
 		seq_printf(seq, " %s", snmp4_ipstats_list[i].name);
 
 	seq_printf(seq, "\nIp: %d %d",
-		   IPV4_DEVCONF_ALL(FORWARDING) ? 1 : 2, sysctl_ip_default_ttl);
+		   IPV4_DEVCONF_ALL(&init_net, FORWARDING) ? 1 : 2,
+		   sysctl_ip_default_ttl);
 
 	for (i = 0; snmp4_ipstats_list[i].name != NULL; i++)
 		seq_printf(seq, " %lu",

net/ipv4/route.c +1 -1

index cc064e6..499e1e3 100644
@@ -2615,7 +2615,7 @@ static int rt_fill_info(struct sk_buff *skb, u32 pid, u32 seq, int event,
 		__be32 dst = rt->rt_dst;
 
 		if (MULTICAST(dst) && !LOCAL_MCAST(dst) &&
-		    IPV4_DEVCONF_ALL(MC_FORWARDING)) {
+		    IPV4_DEVCONF_ALL(&init_net, MC_FORWARDING)) {
 			int err = ipmr_get_route(skb, r, nowait);
 			if (err <= 0) {
 				if (!nowait) {

Signature · 1 line

1.5.3.4

David Miller

maintainer Applied

2007-12-16 21:32 UTC #15 lore raw

Re: [PATCH net-2.6.25 7/7] Switch users of ipv4_devconf(_all) to use the pernet one

From: Pavel Emelyanov <[email protected]>
Date: Tue, 11 Dec 2007 21:02:18 +0300

Quoted 6 lines

These are scattered over the code, but almost all the "critical" places already have the proper struct net at hand except for snmp proc showing function and routing rtnl handler.
Signed-off-by Pavel Emelyanov <[email protected]>

Applied.