1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
|
/*
* (C) 2008-2009 Pablo Neira Ayuso <pablo@netfilter.org>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/jhash.h>
#include <linux/ip.h>
#include <net/ipv6.h>
#include <linux/netfilter/x_tables.h>
#include <net/netfilter/nf_conntrack.h>
#include <linux/netfilter/xt_cluster.h>
static inline u_int32_t nf_ct_orig_ipv4_src(const struct nf_conn *ct)
{
return ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3.ip;
}
static inline const void *nf_ct_orig_ipv6_src(const struct nf_conn *ct)
{
return ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple.src.u3.ip6;
}
static inline u_int32_t
xt_cluster_hash_ipv4(u_int32_t ip, const struct xt_cluster_match_info *info)
{
return jhash_1word(ip, info->hash_seed);
}
static inline u_int32_t
xt_cluster_hash_ipv6(const void *ip, const struct xt_cluster_match_info *info)
{
return jhash2(ip, NF_CT_TUPLE_L3SIZE / sizeof(__u32), info->hash_seed);
}
static inline u_int32_t
xt_cluster_hash(const struct nf_conn *ct,
const struct xt_cluster_match_info *info)
{
u_int32_t hash = 0;
switch(nf_ct_l3num(ct)) {
case AF_INET:
hash = xt_cluster_hash_ipv4(nf_ct_orig_ipv4_src(ct), info);
break;
case AF_INET6:
hash = xt_cluster_hash_ipv6(nf_ct_orig_ipv6_src(ct), info);
break;
default:
WARN_ON(1);
break;
}
return (((u64)hash * info->total_nodes) >> 32);
}
static inline bool
xt_cluster_is_multicast_addr(const struct sk_buff *skb, u_int8_t family)
{
bool is_multicast = false;
switch(family) {
case NFPROTO_IPV4:
is_multicast = ipv4_is_multicast(ip_hdr(skb)->daddr);
break;
case NFPROTO_IPV6:
is_multicast = ipv6_addr_type(&ipv6_hdr(skb)->daddr) &
IPV6_ADDR_MULTICAST;
break;
default:
WARN_ON(1);
break;
}
return is_multicast;
}
static bool
xt_cluster_mt(const struct sk_buff *skb, const struct xt_match_param *par)
{
struct sk_buff *pskb = (struct sk_buff *)skb;
const struct xt_cluster_match_info *info = par->matchinfo;
const struct nf_conn *ct;
enum ip_conntrack_info ctinfo;
unsigned long hash;
/* This match assumes that all nodes see the same packets. This can be
* achieved if the switch that connects the cluster nodes support some
* sort of 'port mirroring'. However, if your switch does not support
* this, your cluster nodes can reply ARP request using a multicast MAC
* address. Thus, your switch will flood the same packets to the
* cluster nodes with the same multicast MAC address. Using a multicast
* link address is a RFC 1812 (section 3.3.2) violation, but this works
* fine in practise.
*
* Unfortunately, if you use the multicast MAC address, the link layer
* sets skbuff's pkt_type to PACKET_MULTICAST, which is not accepted
* by TCP and others for packets coming to this node. For that reason,
* this match mangles skbuff's pkt_type if it detects a packet
* addressed to a unicast address but using PACKET_MULTICAST. Yes, I
* know, matches should not alter packets, but we are doing this here
* because we would need to add a PKTTYPE target for this sole purpose.
*/
if (!xt_cluster_is_multicast_addr(skb, par->family) &&
skb->pkt_type == PACKET_MULTICAST) {
pskb->pkt_type = PACKET_HOST;
}
ct = nf_ct_get(skb, &ctinfo);
if (ct == NULL)
return false;
if (ct == &nf_conntrack_untracked)
return false;
if (ct->master)
hash = xt_cluster_hash(ct->master, info);
else
hash = xt_cluster_hash(ct, info);
return !!((1 << hash) & info->node_mask) ^
!!(info->flags & XT_CLUSTER_F_INV);
}
static bool xt_cluster_mt_checkentry(const struct xt_mtchk_param *par)
{
struct xt_cluster_match_info *info = par->matchinfo;
if (info->node_mask >= (1 << info->total_nodes)) {
printk(KERN_ERR "xt_cluster: this node mask cannot be "
"higher than the total number of nodes\n");
return false;
}
return true;
}
static struct xt_match xt_cluster_match __read_mostly = {
.name = "cluster",
.family = NFPROTO_UNSPEC,
.match = xt_cluster_mt,
.checkentry = xt_cluster_mt_checkentry,
.matchsize = sizeof(struct xt_cluster_match_info),
.me = THIS_MODULE,
};
static int __init xt_cluster_mt_init(void)
{
return xt_register_match(&xt_cluster_match);
}
static void __exit xt_cluster_mt_fini(void)
{
xt_unregister_match(&xt_cluster_match);
}
MODULE_AUTHOR("Pablo Neira Ayuso <pablo@netfilter.org>");
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("Xtables: hash-based cluster match");
MODULE_ALIAS("ipt_cluster");
MODULE_ALIAS("ip6t_cluster");
module_init(xt_cluster_mt_init);
module_exit(xt_cluster_mt_fini);
|
