linux/net/netfilter/nfnetlink.c

469 lines
11 KiB
C
Raw Normal View History

/* Netfilter messages via netlink socket. Allows for user space
* protocol helpers and general trouble making from userspace.
*
* (C) 2001 by Jay Schulist <jschlst@samba.org>,
* (C) 2002-2005 by Harald Welte <laforge@gnumonks.org>
* (C) 2005,2007 by Pablo Neira Ayuso <pablo@netfilter.org>
*
* Initial netfilter messages via netlink development funded and
* generally made possible by Network Robots, Inc. (www.networkrobots.com)
*
* Further development of this code funded by Astaro AG (http://www.astaro.com)
*
* This software may be used and distributed according to the terms
* of the GNU General Public License, incorporated herein by reference.
*/
#include <linux/module.h>
#include <linux/types.h>
#include <linux/socket.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/sockios.h>
#include <linux/net.h>
#include <linux/skbuff.h>
#include <asm/uaccess.h>
#include <net/sock.h>
#include <linux/init.h>
#include <net/netlink.h>
#include <linux/netfilter/nfnetlink.h>
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Harald Welte <laforge@netfilter.org>");
MODULE_ALIAS_NET_PF_PROTO(PF_NETLINK, NETLINK_NETFILTER);
static char __initdata nfversion[] = "0.30";
static struct {
struct mutex mutex;
const struct nfnetlink_subsystem __rcu *subsys;
} table[NFNL_SUBSYS_COUNT];
static const int nfnl_group2type[NFNLGRP_MAX+1] = {
[NFNLGRP_CONNTRACK_NEW] = NFNL_SUBSYS_CTNETLINK,
[NFNLGRP_CONNTRACK_UPDATE] = NFNL_SUBSYS_CTNETLINK,
[NFNLGRP_CONNTRACK_DESTROY] = NFNL_SUBSYS_CTNETLINK,
[NFNLGRP_CONNTRACK_EXP_NEW] = NFNL_SUBSYS_CTNETLINK_EXP,
[NFNLGRP_CONNTRACK_EXP_UPDATE] = NFNL_SUBSYS_CTNETLINK_EXP,
[NFNLGRP_CONNTRACK_EXP_DESTROY] = NFNL_SUBSYS_CTNETLINK_EXP,
};
void nfnl_lock(__u8 subsys_id)
{
mutex_lock(&table[subsys_id].mutex);
}
EXPORT_SYMBOL_GPL(nfnl_lock);
void nfnl_unlock(__u8 subsys_id)
{
mutex_unlock(&table[subsys_id].mutex);
}
EXPORT_SYMBOL_GPL(nfnl_unlock);
#ifdef CONFIG_PROVE_LOCKING
int lockdep_nfnl_is_held(u8 subsys_id)
{
return lockdep_is_held(&table[subsys_id].mutex);
}
EXPORT_SYMBOL_GPL(lockdep_nfnl_is_held);
#endif
int nfnetlink_subsys_register(const struct nfnetlink_subsystem *n)
{
nfnl_lock(n->subsys_id);
if (table[n->subsys_id].subsys) {
nfnl_unlock(n->subsys_id);
return -EBUSY;
}
rcu_assign_pointer(table[n->subsys_id].subsys, n);
nfnl_unlock(n->subsys_id);
return 0;
}
EXPORT_SYMBOL_GPL(nfnetlink_subsys_register);
int nfnetlink_subsys_unregister(const struct nfnetlink_subsystem *n)
{
nfnl_lock(n->subsys_id);
table[n->subsys_id].subsys = NULL;
nfnl_unlock(n->subsys_id);
synchronize_rcu();
return 0;
}
EXPORT_SYMBOL_GPL(nfnetlink_subsys_unregister);
static inline const struct nfnetlink_subsystem *nfnetlink_get_subsys(u_int16_t type)
{
u_int8_t subsys_id = NFNL_SUBSYS_ID(type);
if (subsys_id >= NFNL_SUBSYS_COUNT)
return NULL;
return rcu_dereference(table[subsys_id].subsys);
}
static inline const struct nfnl_callback *
nfnetlink_find_client(u_int16_t type, const struct nfnetlink_subsystem *ss)
{
u_int8_t cb_id = NFNL_MSG_TYPE(type);
if (cb_id >= ss->cb_count)
return NULL;
return &ss->cb[cb_id];
}
int nfnetlink_has_listeners(struct net *net, unsigned int group)
{
return netlink_has_listeners(net->nfnl, group);
}
EXPORT_SYMBOL_GPL(nfnetlink_has_listeners);
struct sk_buff *nfnetlink_alloc_skb(struct net *net, unsigned int size,
u32 dst_portid, gfp_t gfp_mask)
{
return netlink_alloc_skb(net->nfnl, size, dst_portid, gfp_mask);
}
EXPORT_SYMBOL_GPL(nfnetlink_alloc_skb);
int nfnetlink_send(struct sk_buff *skb, struct net *net, u32 portid,
unsigned int group, int echo, gfp_t flags)
{
return nlmsg_notify(net->nfnl, skb, portid, group, echo, flags);
}
EXPORT_SYMBOL_GPL(nfnetlink_send);
int nfnetlink_set_err(struct net *net, u32 portid, u32 group, int error)
{
return netlink_set_err(net->nfnl, portid, group, error);
}
EXPORT_SYMBOL_GPL(nfnetlink_set_err);
int nfnetlink_unicast(struct sk_buff *skb, struct net *net, u32 portid,
int flags)
{
return netlink_unicast(net->nfnl, skb, portid, flags);
}
EXPORT_SYMBOL_GPL(nfnetlink_unicast);
/* Process one complete nfnetlink message. */
static int nfnetlink_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
{
struct net *net = sock_net(skb->sk);
const struct nfnl_callback *nc;
const struct nfnetlink_subsystem *ss;
int type, err;
/* All the messages must at least contain nfgenmsg */
if (nlmsg_len(nlh) < sizeof(struct nfgenmsg))
return 0;
type = nlh->nlmsg_type;
replay:
rcu_read_lock();
ss = nfnetlink_get_subsys(type);
if (!ss) {
#ifdef CONFIG_MODULES
rcu_read_unlock();
request_module("nfnetlink-subsys-%d", NFNL_SUBSYS_ID(type));
rcu_read_lock();
ss = nfnetlink_get_subsys(type);
if (!ss)
#endif
{
rcu_read_unlock();
return -EINVAL;
}
}
nc = nfnetlink_find_client(type, ss);
if (!nc) {
rcu_read_unlock();
return -EINVAL;
}
{
int min_len = nlmsg_total_size(sizeof(struct nfgenmsg));
u_int8_t cb_id = NFNL_MSG_TYPE(nlh->nlmsg_type);
struct nlattr *cda[ss->cb[cb_id].attr_count + 1];
struct nlattr *attr = (void *)nlh + min_len;
int attrlen = nlh->nlmsg_len - min_len;
__u8 subsys_id = NFNL_SUBSYS_ID(type);
err = nla_parse(cda, ss->cb[cb_id].attr_count,
attr, attrlen, ss->cb[cb_id].policy);
if (err < 0) {
rcu_read_unlock();
return err;
}
if (nc->call_rcu) {
err = nc->call_rcu(net->nfnl, skb, nlh,
(const struct nlattr **)cda);
rcu_read_unlock();
} else {
rcu_read_unlock();
nfnl_lock(subsys_id);
if (rcu_dereference_protected(table[subsys_id].subsys,
lockdep_is_held(&table[subsys_id].mutex)) != ss ||
nfnetlink_find_client(type, ss) != nc)
err = -EAGAIN;
else if (nc->call)
err = nc->call(net->nfnl, skb, nlh,
(const struct nlattr **)cda);
else
err = -EINVAL;
nfnl_unlock(subsys_id);
}
if (err == -EAGAIN)
goto replay;
return err;
}
}
netfilter: nfnetlink: add batch support and use it from nf_tables This patch adds a batch support to nfnetlink. Basically, it adds two new control messages: * NFNL_MSG_BATCH_BEGIN, that indicates the beginning of a batch, the nfgenmsg->res_id indicates the nfnetlink subsystem ID. * NFNL_MSG_BATCH_END, that results in the invocation of the ss->commit callback function. If not specified or an error ocurred in the batch, the ss->abort function is invoked instead. The end message represents the commit operation in nftables, the lack of end message results in an abort. This patch also adds the .call_batch function that is only called from the batch receival path. This patch adds atomic rule updates and dumps based on bitmask generations. This allows to atomically commit a set of rule-set updates incrementally without altering the internal state of existing nf_tables expressions/matches/targets. The idea consists of using a generation cursor of 1 bit and a bitmask of 2 bits per rule. Assuming the gencursor is 0, then the genmask (expressed as a bitmask) can be interpreted as: 00 active in the present, will be active in the next generation. 01 inactive in the present, will be active in the next generation. 10 active in the present, will be deleted in the next generation. ^ gencursor Once you invoke the transition to the next generation, the global gencursor is updated: 00 active in the present, will be active in the next generation. 01 active in the present, needs to zero its future, it becomes 00. 10 inactive in the present, delete now. ^ gencursor If a dump is in progress and nf_tables enters a new generation, the dump will stop and return -EBUSY to let userspace know that it has to retry again. In order to invalidate dumps, a global genctr counter is increased everytime nf_tables enters a new generation. This new operation can be used from the user-space utility that controls the firewall, eg. nft -f restore The rule updates contained in `file' will be applied atomically. cat file ----- add filter INPUT ip saddr 1.1.1.1 counter accept #1 del filter INPUT ip daddr 2.2.2.2 counter drop #2 -EOF- Note that the rule 1 will be inactive until the transition to the next generation, the rule 2 will be evicted in the next generation. There is a penalty during the rule update due to the branch misprediction in the packet matching framework. But that should be quickly resolved once the iteration over the commit list that contain rules that require updates is finished. Event notification happens once the rule-set update has been committed. So we skip notifications is case the rule-set update is aborted, which can happen in case that the rule-set is tested to apply correctly. This patch squashed the following patches from Pablo: * nf_tables: atomic rule updates and dumps * nf_tables: get rid of per rule list_head for commits * nf_tables: use per netns commit list * nfnetlink: add batch support and use it from nf_tables * nf_tables: all rule updates are transactional * nf_tables: attach replacement rule after stale one * nf_tables: do not allow deletion/replacement of stale rules * nf_tables: remove unused NFTA_RULE_FLAGS Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2013-10-14 17:05:33 +08:00
static void nfnetlink_rcv_batch(struct sk_buff *skb, struct nlmsghdr *nlh,
u_int16_t subsys_id)
{
struct sk_buff *nskb, *oskb = skb;
struct net *net = sock_net(skb->sk);
const struct nfnetlink_subsystem *ss;
const struct nfnl_callback *nc;
bool success = true, done = false;
int err;
if (subsys_id >= NFNL_SUBSYS_COUNT)
return netlink_ack(skb, nlh, -EINVAL);
replay:
nskb = netlink_skb_clone(oskb, GFP_KERNEL);
if (!nskb)
return netlink_ack(oskb, nlh, -ENOMEM);
nskb->sk = oskb->sk;
skb = nskb;
nfnl_lock(subsys_id);
ss = rcu_dereference_protected(table[subsys_id].subsys,
lockdep_is_held(&table[subsys_id].mutex));
if (!ss) {
#ifdef CONFIG_MODULES
nfnl_unlock(subsys_id);
request_module("nfnetlink-subsys-%d", subsys_id);
nfnl_lock(subsys_id);
ss = rcu_dereference_protected(table[subsys_id].subsys,
lockdep_is_held(&table[subsys_id].mutex));
if (!ss)
#endif
{
nfnl_unlock(subsys_id);
netfilter: nfnetlink: Fix use after free when it fails to process batch This bug manifests when calling the nft command line tool without nf_tables kernel support. kernel message: [ 44.071555] Netfilter messages via NETLINK v0.30. [ 44.072253] BUG: unable to handle kernel NULL pointer dereference at 0000000000000119 [ 44.072264] IP: [<ffffffff8171db1f>] netlink_getsockbyportid+0xf/0x70 [ 44.072272] PGD 7f2b74067 PUD 7f2b73067 PMD 0 [ 44.072277] Oops: 0000 [#1] SMP [...] [ 44.072369] Call Trace: [ 44.072373] [<ffffffff8171fd81>] netlink_unicast+0x91/0x200 [ 44.072377] [<ffffffff817206c9>] netlink_ack+0x99/0x110 [ 44.072381] [<ffffffffa004b951>] nfnetlink_rcv+0x3c1/0x408 [nfnetlink] [ 44.072385] [<ffffffff8171fde3>] netlink_unicast+0xf3/0x200 [ 44.072389] [<ffffffff817201ef>] netlink_sendmsg+0x2ff/0x740 [ 44.072394] [<ffffffff81044752>] ? __mmdrop+0x62/0x90 [ 44.072398] [<ffffffff816dafdb>] sock_sendmsg+0x8b/0xc0 [ 44.072403] [<ffffffff812f1af5>] ? copy_user_enhanced_fast_string+0x5/0x10 [ 44.072406] [<ffffffff816dbb6c>] ? move_addr_to_kernel+0x2c/0x50 [ 44.072410] [<ffffffff816db423>] ___sys_sendmsg+0x3c3/0x3d0 [ 44.072415] [<ffffffff811301ba>] ? handle_mm_fault+0xa9a/0xc60 [ 44.072420] [<ffffffff811362d6>] ? mmap_region+0x166/0x5a0 [ 44.072424] [<ffffffff817da84c>] ? __do_page_fault+0x1dc/0x510 [ 44.072428] [<ffffffff812b8b2c>] ? apparmor_capable+0x1c/0x60 [ 44.072435] [<ffffffff817d6e9a>] ? _raw_spin_unlock_bh+0x1a/0x20 [ 44.072439] [<ffffffff816dfc86>] ? release_sock+0x106/0x150 [ 44.072443] [<ffffffff816dc212>] __sys_sendmsg+0x42/0x80 [ 44.072446] [<ffffffff816dc262>] SyS_sendmsg+0x12/0x20 [ 44.072450] [<ffffffff817df616>] system_call_fastpath+0x1a/0x1f Signed-off-by: Denys Fedoryshchenko <nuclearcat@nuclearcat.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2014-05-04 19:35:37 +08:00
netlink_ack(skb, nlh, -EOPNOTSUPP);
return kfree_skb(nskb);
netfilter: nfnetlink: add batch support and use it from nf_tables This patch adds a batch support to nfnetlink. Basically, it adds two new control messages: * NFNL_MSG_BATCH_BEGIN, that indicates the beginning of a batch, the nfgenmsg->res_id indicates the nfnetlink subsystem ID. * NFNL_MSG_BATCH_END, that results in the invocation of the ss->commit callback function. If not specified or an error ocurred in the batch, the ss->abort function is invoked instead. The end message represents the commit operation in nftables, the lack of end message results in an abort. This patch also adds the .call_batch function that is only called from the batch receival path. This patch adds atomic rule updates and dumps based on bitmask generations. This allows to atomically commit a set of rule-set updates incrementally without altering the internal state of existing nf_tables expressions/matches/targets. The idea consists of using a generation cursor of 1 bit and a bitmask of 2 bits per rule. Assuming the gencursor is 0, then the genmask (expressed as a bitmask) can be interpreted as: 00 active in the present, will be active in the next generation. 01 inactive in the present, will be active in the next generation. 10 active in the present, will be deleted in the next generation. ^ gencursor Once you invoke the transition to the next generation, the global gencursor is updated: 00 active in the present, will be active in the next generation. 01 active in the present, needs to zero its future, it becomes 00. 10 inactive in the present, delete now. ^ gencursor If a dump is in progress and nf_tables enters a new generation, the dump will stop and return -EBUSY to let userspace know that it has to retry again. In order to invalidate dumps, a global genctr counter is increased everytime nf_tables enters a new generation. This new operation can be used from the user-space utility that controls the firewall, eg. nft -f restore The rule updates contained in `file' will be applied atomically. cat file ----- add filter INPUT ip saddr 1.1.1.1 counter accept #1 del filter INPUT ip daddr 2.2.2.2 counter drop #2 -EOF- Note that the rule 1 will be inactive until the transition to the next generation, the rule 2 will be evicted in the next generation. There is a penalty during the rule update due to the branch misprediction in the packet matching framework. But that should be quickly resolved once the iteration over the commit list that contain rules that require updates is finished. Event notification happens once the rule-set update has been committed. So we skip notifications is case the rule-set update is aborted, which can happen in case that the rule-set is tested to apply correctly. This patch squashed the following patches from Pablo: * nf_tables: atomic rule updates and dumps * nf_tables: get rid of per rule list_head for commits * nf_tables: use per netns commit list * nfnetlink: add batch support and use it from nf_tables * nf_tables: all rule updates are transactional * nf_tables: attach replacement rule after stale one * nf_tables: do not allow deletion/replacement of stale rules * nf_tables: remove unused NFTA_RULE_FLAGS Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2013-10-14 17:05:33 +08:00
}
}
if (!ss->commit || !ss->abort) {
nfnl_unlock(subsys_id);
netfilter: nfnetlink: Fix use after free when it fails to process batch This bug manifests when calling the nft command line tool without nf_tables kernel support. kernel message: [ 44.071555] Netfilter messages via NETLINK v0.30. [ 44.072253] BUG: unable to handle kernel NULL pointer dereference at 0000000000000119 [ 44.072264] IP: [<ffffffff8171db1f>] netlink_getsockbyportid+0xf/0x70 [ 44.072272] PGD 7f2b74067 PUD 7f2b73067 PMD 0 [ 44.072277] Oops: 0000 [#1] SMP [...] [ 44.072369] Call Trace: [ 44.072373] [<ffffffff8171fd81>] netlink_unicast+0x91/0x200 [ 44.072377] [<ffffffff817206c9>] netlink_ack+0x99/0x110 [ 44.072381] [<ffffffffa004b951>] nfnetlink_rcv+0x3c1/0x408 [nfnetlink] [ 44.072385] [<ffffffff8171fde3>] netlink_unicast+0xf3/0x200 [ 44.072389] [<ffffffff817201ef>] netlink_sendmsg+0x2ff/0x740 [ 44.072394] [<ffffffff81044752>] ? __mmdrop+0x62/0x90 [ 44.072398] [<ffffffff816dafdb>] sock_sendmsg+0x8b/0xc0 [ 44.072403] [<ffffffff812f1af5>] ? copy_user_enhanced_fast_string+0x5/0x10 [ 44.072406] [<ffffffff816dbb6c>] ? move_addr_to_kernel+0x2c/0x50 [ 44.072410] [<ffffffff816db423>] ___sys_sendmsg+0x3c3/0x3d0 [ 44.072415] [<ffffffff811301ba>] ? handle_mm_fault+0xa9a/0xc60 [ 44.072420] [<ffffffff811362d6>] ? mmap_region+0x166/0x5a0 [ 44.072424] [<ffffffff817da84c>] ? __do_page_fault+0x1dc/0x510 [ 44.072428] [<ffffffff812b8b2c>] ? apparmor_capable+0x1c/0x60 [ 44.072435] [<ffffffff817d6e9a>] ? _raw_spin_unlock_bh+0x1a/0x20 [ 44.072439] [<ffffffff816dfc86>] ? release_sock+0x106/0x150 [ 44.072443] [<ffffffff816dc212>] __sys_sendmsg+0x42/0x80 [ 44.072446] [<ffffffff816dc262>] SyS_sendmsg+0x12/0x20 [ 44.072450] [<ffffffff817df616>] system_call_fastpath+0x1a/0x1f Signed-off-by: Denys Fedoryshchenko <nuclearcat@nuclearcat.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2014-05-04 19:35:37 +08:00
netlink_ack(skb, nlh, -EOPNOTSUPP);
return kfree_skb(skb);
netfilter: nfnetlink: add batch support and use it from nf_tables This patch adds a batch support to nfnetlink. Basically, it adds two new control messages: * NFNL_MSG_BATCH_BEGIN, that indicates the beginning of a batch, the nfgenmsg->res_id indicates the nfnetlink subsystem ID. * NFNL_MSG_BATCH_END, that results in the invocation of the ss->commit callback function. If not specified or an error ocurred in the batch, the ss->abort function is invoked instead. The end message represents the commit operation in nftables, the lack of end message results in an abort. This patch also adds the .call_batch function that is only called from the batch receival path. This patch adds atomic rule updates and dumps based on bitmask generations. This allows to atomically commit a set of rule-set updates incrementally without altering the internal state of existing nf_tables expressions/matches/targets. The idea consists of using a generation cursor of 1 bit and a bitmask of 2 bits per rule. Assuming the gencursor is 0, then the genmask (expressed as a bitmask) can be interpreted as: 00 active in the present, will be active in the next generation. 01 inactive in the present, will be active in the next generation. 10 active in the present, will be deleted in the next generation. ^ gencursor Once you invoke the transition to the next generation, the global gencursor is updated: 00 active in the present, will be active in the next generation. 01 active in the present, needs to zero its future, it becomes 00. 10 inactive in the present, delete now. ^ gencursor If a dump is in progress and nf_tables enters a new generation, the dump will stop and return -EBUSY to let userspace know that it has to retry again. In order to invalidate dumps, a global genctr counter is increased everytime nf_tables enters a new generation. This new operation can be used from the user-space utility that controls the firewall, eg. nft -f restore The rule updates contained in `file' will be applied atomically. cat file ----- add filter INPUT ip saddr 1.1.1.1 counter accept #1 del filter INPUT ip daddr 2.2.2.2 counter drop #2 -EOF- Note that the rule 1 will be inactive until the transition to the next generation, the rule 2 will be evicted in the next generation. There is a penalty during the rule update due to the branch misprediction in the packet matching framework. But that should be quickly resolved once the iteration over the commit list that contain rules that require updates is finished. Event notification happens once the rule-set update has been committed. So we skip notifications is case the rule-set update is aborted, which can happen in case that the rule-set is tested to apply correctly. This patch squashed the following patches from Pablo: * nf_tables: atomic rule updates and dumps * nf_tables: get rid of per rule list_head for commits * nf_tables: use per netns commit list * nfnetlink: add batch support and use it from nf_tables * nf_tables: all rule updates are transactional * nf_tables: attach replacement rule after stale one * nf_tables: do not allow deletion/replacement of stale rules * nf_tables: remove unused NFTA_RULE_FLAGS Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2013-10-14 17:05:33 +08:00
}
while (skb->len >= nlmsg_total_size(0)) {
int msglen, type;
nlh = nlmsg_hdr(skb);
err = 0;
if (nlh->nlmsg_len < NLMSG_HDRLEN) {
err = -EINVAL;
goto ack;
}
/* Only requests are handled by the kernel */
if (!(nlh->nlmsg_flags & NLM_F_REQUEST)) {
err = -EINVAL;
goto ack;
}
type = nlh->nlmsg_type;
if (type == NFNL_MSG_BATCH_BEGIN) {
/* Malformed: Batch begin twice */
success = false;
goto done;
} else if (type == NFNL_MSG_BATCH_END) {
done = true;
goto done;
} else if (type < NLMSG_MIN_TYPE) {
err = -EINVAL;
goto ack;
}
/* We only accept a batch with messages for the same
* subsystem.
*/
if (NFNL_SUBSYS_ID(type) != subsys_id) {
err = -EINVAL;
goto ack;
}
nc = nfnetlink_find_client(type, ss);
if (!nc) {
err = -EINVAL;
goto ack;
}
{
int min_len = nlmsg_total_size(sizeof(struct nfgenmsg));
u_int8_t cb_id = NFNL_MSG_TYPE(nlh->nlmsg_type);
struct nlattr *cda[ss->cb[cb_id].attr_count + 1];
struct nlattr *attr = (void *)nlh + min_len;
int attrlen = nlh->nlmsg_len - min_len;
err = nla_parse(cda, ss->cb[cb_id].attr_count,
attr, attrlen, ss->cb[cb_id].policy);
if (err < 0)
goto ack;
if (nc->call_batch) {
err = nc->call_batch(net->nfnl, skb, nlh,
(const struct nlattr **)cda);
}
/* The lock was released to autoload some module, we
* have to abort and start from scratch using the
* original skb.
*/
if (err == -EAGAIN) {
ss->abort(skb);
nfnl_unlock(subsys_id);
kfree_skb(nskb);
goto replay;
}
}
ack:
if (nlh->nlmsg_flags & NLM_F_ACK || err) {
/* We don't stop processing the batch on errors, thus,
* userspace gets all the errors that the batch
* triggers.
*/
netlink_ack(skb, nlh, err);
if (err)
success = false;
}
msglen = NLMSG_ALIGN(nlh->nlmsg_len);
if (msglen > skb->len)
msglen = skb->len;
skb_pull(skb, msglen);
}
done:
if (success && done)
ss->commit(skb);
else
ss->abort(skb);
nfnl_unlock(subsys_id);
kfree_skb(nskb);
}
static void nfnetlink_rcv(struct sk_buff *skb)
{
netfilter: nfnetlink: add batch support and use it from nf_tables This patch adds a batch support to nfnetlink. Basically, it adds two new control messages: * NFNL_MSG_BATCH_BEGIN, that indicates the beginning of a batch, the nfgenmsg->res_id indicates the nfnetlink subsystem ID. * NFNL_MSG_BATCH_END, that results in the invocation of the ss->commit callback function. If not specified or an error ocurred in the batch, the ss->abort function is invoked instead. The end message represents the commit operation in nftables, the lack of end message results in an abort. This patch also adds the .call_batch function that is only called from the batch receival path. This patch adds atomic rule updates and dumps based on bitmask generations. This allows to atomically commit a set of rule-set updates incrementally without altering the internal state of existing nf_tables expressions/matches/targets. The idea consists of using a generation cursor of 1 bit and a bitmask of 2 bits per rule. Assuming the gencursor is 0, then the genmask (expressed as a bitmask) can be interpreted as: 00 active in the present, will be active in the next generation. 01 inactive in the present, will be active in the next generation. 10 active in the present, will be deleted in the next generation. ^ gencursor Once you invoke the transition to the next generation, the global gencursor is updated: 00 active in the present, will be active in the next generation. 01 active in the present, needs to zero its future, it becomes 00. 10 inactive in the present, delete now. ^ gencursor If a dump is in progress and nf_tables enters a new generation, the dump will stop and return -EBUSY to let userspace know that it has to retry again. In order to invalidate dumps, a global genctr counter is increased everytime nf_tables enters a new generation. This new operation can be used from the user-space utility that controls the firewall, eg. nft -f restore The rule updates contained in `file' will be applied atomically. cat file ----- add filter INPUT ip saddr 1.1.1.1 counter accept #1 del filter INPUT ip daddr 2.2.2.2 counter drop #2 -EOF- Note that the rule 1 will be inactive until the transition to the next generation, the rule 2 will be evicted in the next generation. There is a penalty during the rule update due to the branch misprediction in the packet matching framework. But that should be quickly resolved once the iteration over the commit list that contain rules that require updates is finished. Event notification happens once the rule-set update has been committed. So we skip notifications is case the rule-set update is aborted, which can happen in case that the rule-set is tested to apply correctly. This patch squashed the following patches from Pablo: * nf_tables: atomic rule updates and dumps * nf_tables: get rid of per rule list_head for commits * nf_tables: use per netns commit list * nfnetlink: add batch support and use it from nf_tables * nf_tables: all rule updates are transactional * nf_tables: attach replacement rule after stale one * nf_tables: do not allow deletion/replacement of stale rules * nf_tables: remove unused NFTA_RULE_FLAGS Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2013-10-14 17:05:33 +08:00
struct nlmsghdr *nlh = nlmsg_hdr(skb);
int msglen;
if (nlh->nlmsg_len < NLMSG_HDRLEN ||
skb->len < nlh->nlmsg_len)
return;
if (!netlink_net_capable(skb, CAP_NET_ADMIN)) {
netlink_ack(skb, nlh, -EPERM);
return;
}
netfilter: nfnetlink: add batch support and use it from nf_tables This patch adds a batch support to nfnetlink. Basically, it adds two new control messages: * NFNL_MSG_BATCH_BEGIN, that indicates the beginning of a batch, the nfgenmsg->res_id indicates the nfnetlink subsystem ID. * NFNL_MSG_BATCH_END, that results in the invocation of the ss->commit callback function. If not specified or an error ocurred in the batch, the ss->abort function is invoked instead. The end message represents the commit operation in nftables, the lack of end message results in an abort. This patch also adds the .call_batch function that is only called from the batch receival path. This patch adds atomic rule updates and dumps based on bitmask generations. This allows to atomically commit a set of rule-set updates incrementally without altering the internal state of existing nf_tables expressions/matches/targets. The idea consists of using a generation cursor of 1 bit and a bitmask of 2 bits per rule. Assuming the gencursor is 0, then the genmask (expressed as a bitmask) can be interpreted as: 00 active in the present, will be active in the next generation. 01 inactive in the present, will be active in the next generation. 10 active in the present, will be deleted in the next generation. ^ gencursor Once you invoke the transition to the next generation, the global gencursor is updated: 00 active in the present, will be active in the next generation. 01 active in the present, needs to zero its future, it becomes 00. 10 inactive in the present, delete now. ^ gencursor If a dump is in progress and nf_tables enters a new generation, the dump will stop and return -EBUSY to let userspace know that it has to retry again. In order to invalidate dumps, a global genctr counter is increased everytime nf_tables enters a new generation. This new operation can be used from the user-space utility that controls the firewall, eg. nft -f restore The rule updates contained in `file' will be applied atomically. cat file ----- add filter INPUT ip saddr 1.1.1.1 counter accept #1 del filter INPUT ip daddr 2.2.2.2 counter drop #2 -EOF- Note that the rule 1 will be inactive until the transition to the next generation, the rule 2 will be evicted in the next generation. There is a penalty during the rule update due to the branch misprediction in the packet matching framework. But that should be quickly resolved once the iteration over the commit list that contain rules that require updates is finished. Event notification happens once the rule-set update has been committed. So we skip notifications is case the rule-set update is aborted, which can happen in case that the rule-set is tested to apply correctly. This patch squashed the following patches from Pablo: * nf_tables: atomic rule updates and dumps * nf_tables: get rid of per rule list_head for commits * nf_tables: use per netns commit list * nfnetlink: add batch support and use it from nf_tables * nf_tables: all rule updates are transactional * nf_tables: attach replacement rule after stale one * nf_tables: do not allow deletion/replacement of stale rules * nf_tables: remove unused NFTA_RULE_FLAGS Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2013-10-14 17:05:33 +08:00
if (nlh->nlmsg_type == NFNL_MSG_BATCH_BEGIN) {
struct nfgenmsg *nfgenmsg;
msglen = NLMSG_ALIGN(nlh->nlmsg_len);
if (msglen > skb->len)
msglen = skb->len;
if (nlh->nlmsg_len < NLMSG_HDRLEN ||
skb->len < NLMSG_HDRLEN + sizeof(struct nfgenmsg))
return;
nfgenmsg = nlmsg_data(nlh);
skb_pull(skb, msglen);
nfnetlink_rcv_batch(skb, nlh, nfgenmsg->res_id);
} else {
netlink_rcv_skb(skb, &nfnetlink_rcv_msg);
}
}
#ifdef CONFIG_MODULES
static int nfnetlink_bind(int group)
{
const struct nfnetlink_subsystem *ss;
int type = nfnl_group2type[group];
rcu_read_lock();
ss = nfnetlink_get_subsys(type);
rcu_read_unlock();
if (!ss)
request_module("nfnetlink-subsys-%d", type);
return 0;
}
#endif
static int __net_init nfnetlink_net_init(struct net *net)
{
struct sock *nfnl;
struct netlink_kernel_cfg cfg = {
.groups = NFNLGRP_MAX,
.input = nfnetlink_rcv,
#ifdef CONFIG_MODULES
.bind = nfnetlink_bind,
#endif
};
nfnl = netlink_kernel_create(net, NETLINK_NETFILTER, &cfg);
if (!nfnl)
return -ENOMEM;
net->nfnl_stash = nfnl;
rcu_assign_pointer(net->nfnl, nfnl);
return 0;
}
static void __net_exit nfnetlink_net_exit_batch(struct list_head *net_exit_list)
{
struct net *net;
list_for_each_entry(net, net_exit_list, exit_list)
RCU_INIT_POINTER(net->nfnl, NULL);
synchronize_net();
list_for_each_entry(net, net_exit_list, exit_list)
netlink_kernel_release(net->nfnl_stash);
}
static struct pernet_operations nfnetlink_net_ops = {
.init = nfnetlink_net_init,
.exit_batch = nfnetlink_net_exit_batch,
};
static int __init nfnetlink_init(void)
{
int i;
for (i=0; i<NFNL_SUBSYS_COUNT; i++)
mutex_init(&table[i].mutex);
pr_info("Netfilter messages via NETLINK v%s.\n", nfversion);
return register_pernet_subsys(&nfnetlink_net_ops);
}
static void __exit nfnetlink_exit(void)
{
pr_info("Removing netfilter NETLINK layer.\n");
unregister_pernet_subsys(&nfnetlink_net_ops);
}
module_init(nfnetlink_init);
module_exit(nfnetlink_exit);