diff options
author | Patrick McHardy <kaber@trash.net> | 2010-04-13 01:03:23 -0400 |
---|---|---|
committer | David S. Miller <davem@davemloft.net> | 2010-04-13 17:49:34 -0400 |
commit | f0ad0860d01e47a3ffd220564c5c653b3afbe962 (patch) | |
tree | 91b69423f472b934daa2c18ae3b7ba065b7c7898 /net/ipv4/ipmr.c | |
parent | 0c12295a741d3186987f96f518cfbdaf01abb087 (diff) |
ipv4: ipmr: support multiple tables
This patch adds support for multiple independant multicast routing instances,
named "tables".
Userspace multicast routing daemons can bind to a specific table instance by
issuing a setsockopt call using a new option MRT_TABLE. The table number is
stored in the raw socket data and affects all following ipmr setsockopt(),
getsockopt() and ioctl() calls. By default, a single table (RT_TABLE_DEFAULT)
is created with a default routing rule pointing to it. Newly created pimreg
devices have the table number appended ("pimregX"), with the exception of
devices created in the default table, which are named just "pimreg" for
compatibility reasons.
Packets are directed to a specific table instance using routing rules,
similar to how regular routing rules work. Currently iif, oif and mark
are supported as keys, source and destination addresses could be supported
additionally.
Example usage:
- bind pimd/xorp/... to a specific table:
uint32_t table = 123;
setsockopt(fd, IPPROTO_IP, MRT_TABLE, &table, sizeof(table));
- create routing rules directing packets to the new table:
# ip mrule add iif eth0 lookup 123
# ip mrule add oif eth0 lookup 123
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4/ipmr.c')
-rw-r--r-- | net/ipv4/ipmr.c | 399 |
1 files changed, 338 insertions, 61 deletions
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c index 498f4e907d52..5df5fd74c6d1 100644 --- a/net/ipv4/ipmr.c +++ b/net/ipv4/ipmr.c | |||
@@ -63,12 +63,15 @@ | |||
63 | #include <net/ipip.h> | 63 | #include <net/ipip.h> |
64 | #include <net/checksum.h> | 64 | #include <net/checksum.h> |
65 | #include <net/netlink.h> | 65 | #include <net/netlink.h> |
66 | #include <net/fib_rules.h> | ||
66 | 67 | ||
67 | #if defined(CONFIG_IP_PIMSM_V1) || defined(CONFIG_IP_PIMSM_V2) | 68 | #if defined(CONFIG_IP_PIMSM_V1) || defined(CONFIG_IP_PIMSM_V2) |
68 | #define CONFIG_IP_PIMSM 1 | 69 | #define CONFIG_IP_PIMSM 1 |
69 | #endif | 70 | #endif |
70 | 71 | ||
71 | struct mr_table { | 72 | struct mr_table { |
73 | struct list_head list; | ||
74 | u32 id; | ||
72 | struct sock *mroute_sk; | 75 | struct sock *mroute_sk; |
73 | struct timer_list ipmr_expire_timer; | 76 | struct timer_list ipmr_expire_timer; |
74 | struct list_head mfc_unres_queue; | 77 | struct list_head mfc_unres_queue; |
@@ -83,6 +86,14 @@ struct mr_table { | |||
83 | #endif | 86 | #endif |
84 | }; | 87 | }; |
85 | 88 | ||
89 | struct ipmr_rule { | ||
90 | struct fib_rule common; | ||
91 | }; | ||
92 | |||
93 | struct ipmr_result { | ||
94 | struct mr_table *mrt; | ||
95 | }; | ||
96 | |||
86 | /* Big lock, protecting vif table, mrt cache and mroute socket state. | 97 | /* Big lock, protecting vif table, mrt cache and mroute socket state. |
87 | Note that the changes are semaphored via rtnl_lock. | 98 | Note that the changes are semaphored via rtnl_lock. |
88 | */ | 99 | */ |
@@ -108,6 +119,7 @@ static DEFINE_SPINLOCK(mfc_unres_lock); | |||
108 | 119 | ||
109 | static struct kmem_cache *mrt_cachep __read_mostly; | 120 | static struct kmem_cache *mrt_cachep __read_mostly; |
110 | 121 | ||
122 | static struct mr_table *ipmr_new_table(struct net *net, u32 id); | ||
111 | static int ip_mr_forward(struct net *net, struct mr_table *mrt, | 123 | static int ip_mr_forward(struct net *net, struct mr_table *mrt, |
112 | struct sk_buff *skb, struct mfc_cache *cache, | 124 | struct sk_buff *skb, struct mfc_cache *cache, |
113 | int local); | 125 | int local); |
@@ -115,6 +127,206 @@ static int ipmr_cache_report(struct mr_table *mrt, | |||
115 | struct sk_buff *pkt, vifi_t vifi, int assert); | 127 | struct sk_buff *pkt, vifi_t vifi, int assert); |
116 | static int ipmr_fill_mroute(struct mr_table *mrt, struct sk_buff *skb, | 128 | static int ipmr_fill_mroute(struct mr_table *mrt, struct sk_buff *skb, |
117 | struct mfc_cache *c, struct rtmsg *rtm); | 129 | struct mfc_cache *c, struct rtmsg *rtm); |
130 | static void ipmr_expire_process(unsigned long arg); | ||
131 | |||
132 | #ifdef CONFIG_IP_MROUTE_MULTIPLE_TABLES | ||
133 | #define ipmr_for_each_table(mrt, net) \ | ||
134 | list_for_each_entry_rcu(mrt, &net->ipv4.mr_tables, list) | ||
135 | |||
136 | static struct mr_table *ipmr_get_table(struct net *net, u32 id) | ||
137 | { | ||
138 | struct mr_table *mrt; | ||
139 | |||
140 | ipmr_for_each_table(mrt, net) { | ||
141 | if (mrt->id == id) | ||
142 | return mrt; | ||
143 | } | ||
144 | return NULL; | ||
145 | } | ||
146 | |||
147 | static int ipmr_fib_lookup(struct net *net, struct flowi *flp, | ||
148 | struct mr_table **mrt) | ||
149 | { | ||
150 | struct ipmr_result res; | ||
151 | struct fib_lookup_arg arg = { .result = &res, }; | ||
152 | int err; | ||
153 | |||
154 | err = fib_rules_lookup(net->ipv4.mr_rules_ops, flp, 0, &arg); | ||
155 | if (err < 0) | ||
156 | return err; | ||
157 | *mrt = res.mrt; | ||
158 | return 0; | ||
159 | } | ||
160 | |||
161 | static int ipmr_rule_action(struct fib_rule *rule, struct flowi *flp, | ||
162 | int flags, struct fib_lookup_arg *arg) | ||
163 | { | ||
164 | struct ipmr_result *res = arg->result; | ||
165 | struct mr_table *mrt; | ||
166 | |||
167 | switch (rule->action) { | ||
168 | case FR_ACT_TO_TBL: | ||
169 | break; | ||
170 | case FR_ACT_UNREACHABLE: | ||
171 | return -ENETUNREACH; | ||
172 | case FR_ACT_PROHIBIT: | ||
173 | return -EACCES; | ||
174 | case FR_ACT_BLACKHOLE: | ||
175 | default: | ||
176 | return -EINVAL; | ||
177 | } | ||
178 | |||
179 | mrt = ipmr_get_table(rule->fr_net, rule->table); | ||
180 | if (mrt == NULL) | ||
181 | return -EAGAIN; | ||
182 | res->mrt = mrt; | ||
183 | return 0; | ||
184 | } | ||
185 | |||
186 | static int ipmr_rule_match(struct fib_rule *rule, struct flowi *fl, int flags) | ||
187 | { | ||
188 | return 1; | ||
189 | } | ||
190 | |||
191 | static const struct nla_policy ipmr_rule_policy[FRA_MAX + 1] = { | ||
192 | FRA_GENERIC_POLICY, | ||
193 | }; | ||
194 | |||
195 | static int ipmr_rule_configure(struct fib_rule *rule, struct sk_buff *skb, | ||
196 | struct fib_rule_hdr *frh, struct nlattr **tb) | ||
197 | { | ||
198 | return 0; | ||
199 | } | ||
200 | |||
201 | static int ipmr_rule_compare(struct fib_rule *rule, struct fib_rule_hdr *frh, | ||
202 | struct nlattr **tb) | ||
203 | { | ||
204 | return 1; | ||
205 | } | ||
206 | |||
207 | static int ipmr_rule_fill(struct fib_rule *rule, struct sk_buff *skb, | ||
208 | struct fib_rule_hdr *frh) | ||
209 | { | ||
210 | frh->dst_len = 0; | ||
211 | frh->src_len = 0; | ||
212 | frh->tos = 0; | ||
213 | return 0; | ||
214 | } | ||
215 | |||
216 | static struct fib_rules_ops ipmr_rules_ops_template = { | ||
217 | .family = FIB_RULES_IPMR, | ||
218 | .rule_size = sizeof(struct ipmr_rule), | ||
219 | .addr_size = sizeof(u32), | ||
220 | .action = ipmr_rule_action, | ||
221 | .match = ipmr_rule_match, | ||
222 | .configure = ipmr_rule_configure, | ||
223 | .compare = ipmr_rule_compare, | ||
224 | .default_pref = fib_default_rule_pref, | ||
225 | .fill = ipmr_rule_fill, | ||
226 | .nlgroup = RTNLGRP_IPV4_RULE, | ||
227 | .policy = ipmr_rule_policy, | ||
228 | .owner = THIS_MODULE, | ||
229 | }; | ||
230 | |||
231 | static int __net_init ipmr_rules_init(struct net *net) | ||
232 | { | ||
233 | struct fib_rules_ops *ops; | ||
234 | struct mr_table *mrt; | ||
235 | int err; | ||
236 | |||
237 | ops = fib_rules_register(&ipmr_rules_ops_template, net); | ||
238 | if (IS_ERR(ops)) | ||
239 | return PTR_ERR(ops); | ||
240 | |||
241 | INIT_LIST_HEAD(&net->ipv4.mr_tables); | ||
242 | |||
243 | mrt = ipmr_new_table(net, RT_TABLE_DEFAULT); | ||
244 | if (mrt == NULL) { | ||
245 | err = -ENOMEM; | ||
246 | goto err1; | ||
247 | } | ||
248 | |||
249 | err = fib_default_rule_add(ops, 0x7fff, RT_TABLE_DEFAULT, 0); | ||
250 | if (err < 0) | ||
251 | goto err2; | ||
252 | |||
253 | net->ipv4.mr_rules_ops = ops; | ||
254 | return 0; | ||
255 | |||
256 | err2: | ||
257 | kfree(mrt); | ||
258 | err1: | ||
259 | fib_rules_unregister(ops); | ||
260 | return err; | ||
261 | } | ||
262 | |||
263 | static void __net_exit ipmr_rules_exit(struct net *net) | ||
264 | { | ||
265 | struct mr_table *mrt, *next; | ||
266 | |||
267 | list_for_each_entry_safe(mrt, next, &net->ipv4.mr_tables, list) | ||
268 | kfree(mrt); | ||
269 | fib_rules_unregister(net->ipv4.mr_rules_ops); | ||
270 | } | ||
271 | #else | ||
272 | #define ipmr_for_each_table(mrt, net) \ | ||
273 | for (mrt = net->ipv4.mrt; mrt; mrt = NULL) | ||
274 | |||
275 | static struct mr_table *ipmr_get_table(struct net *net, u32 id) | ||
276 | { | ||
277 | return net->ipv4.mrt; | ||
278 | } | ||
279 | |||
280 | static int ipmr_fib_lookup(struct net *net, struct flowi *flp, | ||
281 | struct mr_table **mrt) | ||
282 | { | ||
283 | *mrt = net->ipv4.mrt; | ||
284 | return 0; | ||
285 | } | ||
286 | |||
287 | static int __net_init ipmr_rules_init(struct net *net) | ||
288 | { | ||
289 | net->ipv4.mrt = ipmr_new_table(net, RT_TABLE_DEFAULT); | ||
290 | return net->ipv4.mrt ? 0 : -ENOMEM; | ||
291 | } | ||
292 | |||
293 | static void __net_exit ipmr_rules_exit(struct net *net) | ||
294 | { | ||
295 | kfree(net->ipv4.mrt); | ||
296 | } | ||
297 | #endif | ||
298 | |||
299 | static struct mr_table *ipmr_new_table(struct net *net, u32 id) | ||
300 | { | ||
301 | struct mr_table *mrt; | ||
302 | unsigned int i; | ||
303 | |||
304 | mrt = ipmr_get_table(net, id); | ||
305 | if (mrt != NULL) | ||
306 | return mrt; | ||
307 | |||
308 | mrt = kzalloc(sizeof(*mrt), GFP_KERNEL); | ||
309 | if (mrt == NULL) | ||
310 | return NULL; | ||
311 | mrt->id = id; | ||
312 | |||
313 | /* Forwarding cache */ | ||
314 | for (i = 0; i < MFC_LINES; i++) | ||
315 | INIT_LIST_HEAD(&mrt->mfc_cache_array[i]); | ||
316 | |||
317 | INIT_LIST_HEAD(&mrt->mfc_unres_queue); | ||
318 | |||
319 | setup_timer(&mrt->ipmr_expire_timer, ipmr_expire_process, | ||
320 | (unsigned long)mrt); | ||
321 | |||
322 | #ifdef CONFIG_IP_PIMSM | ||
323 | mrt->mroute_reg_vif_num = -1; | ||
324 | #endif | ||
325 | #ifdef CONFIG_IP_MROUTE_MULTIPLE_TABLES | ||
326 | list_add_tail_rcu(&mrt->list, &net->ipv4.mr_tables); | ||
327 | #endif | ||
328 | return mrt; | ||
329 | } | ||
118 | 330 | ||
119 | /* Service routines creating virtual interfaces: DVMRP tunnels and PIMREG */ | 331 | /* Service routines creating virtual interfaces: DVMRP tunnels and PIMREG */ |
120 | 332 | ||
@@ -215,7 +427,17 @@ failure: | |||
215 | static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, struct net_device *dev) | 427 | static netdev_tx_t reg_vif_xmit(struct sk_buff *skb, struct net_device *dev) |
216 | { | 428 | { |
217 | struct net *net = dev_net(dev); | 429 | struct net *net = dev_net(dev); |
218 | struct mr_table *mrt = net->ipv4.mrt; | 430 | struct mr_table *mrt; |
431 | struct flowi fl = { | ||
432 | .oif = dev->ifindex, | ||
433 | .iif = skb->skb_iif, | ||
434 | .mark = skb->mark, | ||
435 | }; | ||
436 | int err; | ||
437 | |||
438 | err = ipmr_fib_lookup(net, &fl, &mrt); | ||
439 | if (err < 0) | ||
440 | return err; | ||
219 | 441 | ||
220 | read_lock(&mrt_lock); | 442 | read_lock(&mrt_lock); |
221 | dev->stats.tx_bytes += skb->len; | 443 | dev->stats.tx_bytes += skb->len; |
@@ -240,12 +462,18 @@ static void reg_vif_setup(struct net_device *dev) | |||
240 | dev->features |= NETIF_F_NETNS_LOCAL; | 462 | dev->features |= NETIF_F_NETNS_LOCAL; |
241 | } | 463 | } |
242 | 464 | ||
243 | static struct net_device *ipmr_reg_vif(struct net *net) | 465 | static struct net_device *ipmr_reg_vif(struct net *net, struct mr_table *mrt) |
244 | { | 466 | { |
245 | struct net_device *dev; | 467 | struct net_device *dev; |
246 | struct in_device *in_dev; | 468 | struct in_device *in_dev; |
469 | char name[IFNAMSIZ]; | ||
470 | |||
471 | if (mrt->id == RT_TABLE_DEFAULT) | ||
472 | sprintf(name, "pimreg"); | ||
473 | else | ||
474 | sprintf(name, "pimreg%u", mrt->id); | ||
247 | 475 | ||
248 | dev = alloc_netdev(0, "pimreg", reg_vif_setup); | 476 | dev = alloc_netdev(0, name, reg_vif_setup); |
249 | 477 | ||
250 | if (dev == NULL) | 478 | if (dev == NULL) |
251 | return NULL; | 479 | return NULL; |
@@ -461,7 +689,7 @@ static int vif_add(struct net *net, struct mr_table *mrt, | |||
461 | */ | 689 | */ |
462 | if (mrt->mroute_reg_vif_num >= 0) | 690 | if (mrt->mroute_reg_vif_num >= 0) |
463 | return -EADDRINUSE; | 691 | return -EADDRINUSE; |
464 | dev = ipmr_reg_vif(net); | 692 | dev = ipmr_reg_vif(net, mrt); |
465 | if (!dev) | 693 | if (!dev) |
466 | return -ENOBUFS; | 694 | return -ENOBUFS; |
467 | err = dev_set_allmulti(dev, 1); | 695 | err = dev_set_allmulti(dev, 1); |
@@ -928,17 +1156,19 @@ static void mroute_clean_tables(struct mr_table *mrt) | |||
928 | static void mrtsock_destruct(struct sock *sk) | 1156 | static void mrtsock_destruct(struct sock *sk) |
929 | { | 1157 | { |
930 | struct net *net = sock_net(sk); | 1158 | struct net *net = sock_net(sk); |
931 | struct mr_table *mrt = net->ipv4.mrt; | 1159 | struct mr_table *mrt; |
932 | 1160 | ||
933 | rtnl_lock(); | 1161 | rtnl_lock(); |
934 | if (sk == mrt->mroute_sk) { | 1162 | ipmr_for_each_table(mrt, net) { |
935 | IPV4_DEVCONF_ALL(net, MC_FORWARDING)--; | 1163 | if (sk == mrt->mroute_sk) { |
1164 | IPV4_DEVCONF_ALL(net, MC_FORWARDING)--; | ||
936 | 1165 | ||
937 | write_lock_bh(&mrt_lock); | 1166 | write_lock_bh(&mrt_lock); |
938 | mrt->mroute_sk = NULL; | 1167 | mrt->mroute_sk = NULL; |
939 | write_unlock_bh(&mrt_lock); | 1168 | write_unlock_bh(&mrt_lock); |
940 | 1169 | ||
941 | mroute_clean_tables(mrt); | 1170 | mroute_clean_tables(mrt); |
1171 | } | ||
942 | } | 1172 | } |
943 | rtnl_unlock(); | 1173 | rtnl_unlock(); |
944 | } | 1174 | } |
@@ -956,7 +1186,11 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi | |||
956 | struct vifctl vif; | 1186 | struct vifctl vif; |
957 | struct mfcctl mfc; | 1187 | struct mfcctl mfc; |
958 | struct net *net = sock_net(sk); | 1188 | struct net *net = sock_net(sk); |
959 | struct mr_table *mrt = net->ipv4.mrt; | 1189 | struct mr_table *mrt; |
1190 | |||
1191 | mrt = ipmr_get_table(net, raw_sk(sk)->ipmr_table ? : RT_TABLE_DEFAULT); | ||
1192 | if (mrt == NULL) | ||
1193 | return -ENOENT; | ||
960 | 1194 | ||
961 | if (optname != MRT_INIT) { | 1195 | if (optname != MRT_INIT) { |
962 | if (sk != mrt->mroute_sk && !capable(CAP_NET_ADMIN)) | 1196 | if (sk != mrt->mroute_sk && !capable(CAP_NET_ADMIN)) |
@@ -1055,6 +1289,27 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi | |||
1055 | return ret; | 1289 | return ret; |
1056 | } | 1290 | } |
1057 | #endif | 1291 | #endif |
1292 | #ifdef CONFIG_IP_MROUTE_MULTIPLE_TABLES | ||
1293 | case MRT_TABLE: | ||
1294 | { | ||
1295 | u32 v; | ||
1296 | |||
1297 | if (optlen != sizeof(u32)) | ||
1298 | return -EINVAL; | ||
1299 | if (get_user(v, (u32 __user *)optval)) | ||
1300 | return -EFAULT; | ||
1301 | if (sk == mrt->mroute_sk) | ||
1302 | return -EBUSY; | ||
1303 | |||
1304 | rtnl_lock(); | ||
1305 | ret = 0; | ||
1306 | if (!ipmr_new_table(net, v)) | ||
1307 | ret = -ENOMEM; | ||
1308 | raw_sk(sk)->ipmr_table = v; | ||
1309 | rtnl_unlock(); | ||
1310 | return ret; | ||
1311 | } | ||
1312 | #endif | ||
1058 | /* | 1313 | /* |
1059 | * Spurious command, or MRT_VERSION which you cannot | 1314 | * Spurious command, or MRT_VERSION which you cannot |
1060 | * set. | 1315 | * set. |
@@ -1073,7 +1328,11 @@ int ip_mroute_getsockopt(struct sock *sk, int optname, char __user *optval, int | |||
1073 | int olr; | 1328 | int olr; |
1074 | int val; | 1329 | int val; |
1075 | struct net *net = sock_net(sk); | 1330 | struct net *net = sock_net(sk); |
1076 | struct mr_table *mrt = net->ipv4.mrt; | 1331 | struct mr_table *mrt; |
1332 | |||
1333 | mrt = ipmr_get_table(net, raw_sk(sk)->ipmr_table ? : RT_TABLE_DEFAULT); | ||
1334 | if (mrt == NULL) | ||
1335 | return -ENOENT; | ||
1077 | 1336 | ||
1078 | if (optname != MRT_VERSION && | 1337 | if (optname != MRT_VERSION && |
1079 | #ifdef CONFIG_IP_PIMSM | 1338 | #ifdef CONFIG_IP_PIMSM |
@@ -1115,7 +1374,11 @@ int ipmr_ioctl(struct sock *sk, int cmd, void __user *arg) | |||
1115 | struct vif_device *vif; | 1374 | struct vif_device *vif; |
1116 | struct mfc_cache *c; | 1375 | struct mfc_cache *c; |
1117 | struct net *net = sock_net(sk); | 1376 | struct net *net = sock_net(sk); |
1118 | struct mr_table *mrt = net->ipv4.mrt; | 1377 | struct mr_table *mrt; |
1378 | |||
1379 | mrt = ipmr_get_table(net, raw_sk(sk)->ipmr_table ? : RT_TABLE_DEFAULT); | ||
1380 | if (mrt == NULL) | ||
1381 | return -ENOENT; | ||
1119 | 1382 | ||
1120 | switch (cmd) { | 1383 | switch (cmd) { |
1121 | case SIOCGETVIFCNT: | 1384 | case SIOCGETVIFCNT: |
@@ -1166,17 +1429,20 @@ static int ipmr_device_event(struct notifier_block *this, unsigned long event, v | |||
1166 | { | 1429 | { |
1167 | struct net_device *dev = ptr; | 1430 | struct net_device *dev = ptr; |
1168 | struct net *net = dev_net(dev); | 1431 | struct net *net = dev_net(dev); |
1169 | struct mr_table *mrt = net->ipv4.mrt; | 1432 | struct mr_table *mrt; |
1170 | struct vif_device *v; | 1433 | struct vif_device *v; |
1171 | int ct; | 1434 | int ct; |
1172 | LIST_HEAD(list); | 1435 | LIST_HEAD(list); |
1173 | 1436 | ||
1174 | if (event != NETDEV_UNREGISTER) | 1437 | if (event != NETDEV_UNREGISTER) |
1175 | return NOTIFY_DONE; | 1438 | return NOTIFY_DONE; |
1176 | v = &mrt->vif_table[0]; | 1439 | |
1177 | for (ct = 0; ct < mrt->maxvif; ct++, v++) { | 1440 | ipmr_for_each_table(mrt, net) { |
1178 | if (v->dev == dev) | 1441 | v = &mrt->vif_table[0]; |
1179 | vif_delete(mrt, ct, 1, &list); | 1442 | for (ct = 0; ct < mrt->maxvif; ct++, v++) { |
1443 | if (v->dev == dev) | ||
1444 | vif_delete(mrt, ct, 1, &list); | ||
1445 | } | ||
1180 | } | 1446 | } |
1181 | unregister_netdevice_many(&list); | 1447 | unregister_netdevice_many(&list); |
1182 | return NOTIFY_DONE; | 1448 | return NOTIFY_DONE; |
@@ -1443,8 +1709,9 @@ int ip_mr_input(struct sk_buff *skb) | |||
1443 | { | 1709 | { |
1444 | struct mfc_cache *cache; | 1710 | struct mfc_cache *cache; |
1445 | struct net *net = dev_net(skb->dev); | 1711 | struct net *net = dev_net(skb->dev); |
1446 | struct mr_table *mrt = net->ipv4.mrt; | ||
1447 | int local = skb_rtable(skb)->rt_flags & RTCF_LOCAL; | 1712 | int local = skb_rtable(skb)->rt_flags & RTCF_LOCAL; |
1713 | struct mr_table *mrt; | ||
1714 | int err; | ||
1448 | 1715 | ||
1449 | /* Packet is looped back after forward, it should not be | 1716 | /* Packet is looped back after forward, it should not be |
1450 | forwarded second time, but still can be delivered locally. | 1717 | forwarded second time, but still can be delivered locally. |
@@ -1452,6 +1719,10 @@ int ip_mr_input(struct sk_buff *skb) | |||
1452 | if (IPCB(skb)->flags&IPSKB_FORWARDED) | 1719 | if (IPCB(skb)->flags&IPSKB_FORWARDED) |
1453 | goto dont_forward; | 1720 | goto dont_forward; |
1454 | 1721 | ||
1722 | err = ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt); | ||
1723 | if (err < 0) | ||
1724 | return err; | ||
1725 | |||
1455 | if (!local) { | 1726 | if (!local) { |
1456 | if (IPCB(skb)->opt.router_alert) { | 1727 | if (IPCB(skb)->opt.router_alert) { |
1457 | if (ip_call_ra_chain(skb)) | 1728 | if (ip_call_ra_chain(skb)) |
@@ -1522,12 +1793,11 @@ dont_forward: | |||
1522 | } | 1793 | } |
1523 | 1794 | ||
1524 | #ifdef CONFIG_IP_PIMSM | 1795 | #ifdef CONFIG_IP_PIMSM |
1525 | static int __pim_rcv(struct sk_buff *skb, unsigned int pimlen) | 1796 | static int __pim_rcv(struct mr_table *mrt, struct sk_buff *skb, |
1797 | unsigned int pimlen) | ||
1526 | { | 1798 | { |
1527 | struct net_device *reg_dev = NULL; | 1799 | struct net_device *reg_dev = NULL; |
1528 | struct iphdr *encap; | 1800 | struct iphdr *encap; |
1529 | struct net *net = dev_net(skb->dev); | ||
1530 | struct mr_table *mrt = net->ipv4.mrt; | ||
1531 | 1801 | ||
1532 | encap = (struct iphdr *)(skb_transport_header(skb) + pimlen); | 1802 | encap = (struct iphdr *)(skb_transport_header(skb) + pimlen); |
1533 | /* | 1803 | /* |
@@ -1578,18 +1848,21 @@ int pim_rcv_v1(struct sk_buff * skb) | |||
1578 | { | 1848 | { |
1579 | struct igmphdr *pim; | 1849 | struct igmphdr *pim; |
1580 | struct net *net = dev_net(skb->dev); | 1850 | struct net *net = dev_net(skb->dev); |
1581 | struct mr_table *mrt = net->ipv4.mrt; | 1851 | struct mr_table *mrt; |
1582 | 1852 | ||
1583 | if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr))) | 1853 | if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr))) |
1584 | goto drop; | 1854 | goto drop; |
1585 | 1855 | ||
1586 | pim = igmp_hdr(skb); | 1856 | pim = igmp_hdr(skb); |
1587 | 1857 | ||
1858 | if (ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt) < 0) | ||
1859 | goto drop; | ||
1860 | |||
1588 | if (!mrt->mroute_do_pim || | 1861 | if (!mrt->mroute_do_pim || |
1589 | pim->group != PIM_V1_VERSION || pim->code != PIM_V1_REGISTER) | 1862 | pim->group != PIM_V1_VERSION || pim->code != PIM_V1_REGISTER) |
1590 | goto drop; | 1863 | goto drop; |
1591 | 1864 | ||
1592 | if (__pim_rcv(skb, sizeof(*pim))) { | 1865 | if (__pim_rcv(mrt, skb, sizeof(*pim))) { |
1593 | drop: | 1866 | drop: |
1594 | kfree_skb(skb); | 1867 | kfree_skb(skb); |
1595 | } | 1868 | } |
@@ -1601,6 +1874,8 @@ drop: | |||
1601 | static int pim_rcv(struct sk_buff * skb) | 1874 | static int pim_rcv(struct sk_buff * skb) |
1602 | { | 1875 | { |
1603 | struct pimreghdr *pim; | 1876 | struct pimreghdr *pim; |
1877 | struct net *net = dev_net(skb->dev); | ||
1878 | struct mr_table *mrt; | ||
1604 | 1879 | ||
1605 | if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr))) | 1880 | if (!pskb_may_pull(skb, sizeof(*pim) + sizeof(struct iphdr))) |
1606 | goto drop; | 1881 | goto drop; |
@@ -1612,7 +1887,10 @@ static int pim_rcv(struct sk_buff * skb) | |||
1612 | csum_fold(skb_checksum(skb, 0, skb->len, 0)))) | 1887 | csum_fold(skb_checksum(skb, 0, skb->len, 0)))) |
1613 | goto drop; | 1888 | goto drop; |
1614 | 1889 | ||
1615 | if (__pim_rcv(skb, sizeof(*pim))) { | 1890 | if (ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt) < 0) |
1891 | goto drop; | ||
1892 | |||
1893 | if (__pim_rcv(mrt, skb, sizeof(*pim))) { | ||
1616 | drop: | 1894 | drop: |
1617 | kfree_skb(skb); | 1895 | kfree_skb(skb); |
1618 | } | 1896 | } |
@@ -1663,10 +1941,14 @@ int ipmr_get_route(struct net *net, | |||
1663 | struct sk_buff *skb, struct rtmsg *rtm, int nowait) | 1941 | struct sk_buff *skb, struct rtmsg *rtm, int nowait) |
1664 | { | 1942 | { |
1665 | int err; | 1943 | int err; |
1666 | struct mr_table *mrt = net->ipv4.mrt; | 1944 | struct mr_table *mrt; |
1667 | struct mfc_cache *cache; | 1945 | struct mfc_cache *cache; |
1668 | struct rtable *rt = skb_rtable(skb); | 1946 | struct rtable *rt = skb_rtable(skb); |
1669 | 1947 | ||
1948 | mrt = ipmr_get_table(net, RT_TABLE_DEFAULT); | ||
1949 | if (mrt == NULL) | ||
1950 | return -ENOENT; | ||
1951 | |||
1670 | read_lock(&mrt_lock); | 1952 | read_lock(&mrt_lock); |
1671 | cache = ipmr_cache_find(mrt, rt->rt_src, rt->rt_dst); | 1953 | cache = ipmr_cache_find(mrt, rt->rt_src, rt->rt_dst); |
1672 | 1954 | ||
@@ -1717,6 +1999,7 @@ int ipmr_get_route(struct net *net, | |||
1717 | */ | 1999 | */ |
1718 | struct ipmr_vif_iter { | 2000 | struct ipmr_vif_iter { |
1719 | struct seq_net_private p; | 2001 | struct seq_net_private p; |
2002 | struct mr_table *mrt; | ||
1720 | int ct; | 2003 | int ct; |
1721 | }; | 2004 | }; |
1722 | 2005 | ||
@@ -1724,7 +2007,7 @@ static struct vif_device *ipmr_vif_seq_idx(struct net *net, | |||
1724 | struct ipmr_vif_iter *iter, | 2007 | struct ipmr_vif_iter *iter, |
1725 | loff_t pos) | 2008 | loff_t pos) |
1726 | { | 2009 | { |
1727 | struct mr_table *mrt = net->ipv4.mrt; | 2010 | struct mr_table *mrt = iter->mrt; |
1728 | 2011 | ||
1729 | for (iter->ct = 0; iter->ct < mrt->maxvif; ++iter->ct) { | 2012 | for (iter->ct = 0; iter->ct < mrt->maxvif; ++iter->ct) { |
1730 | if (!VIF_EXISTS(mrt, iter->ct)) | 2013 | if (!VIF_EXISTS(mrt, iter->ct)) |
@@ -1738,7 +2021,15 @@ static struct vif_device *ipmr_vif_seq_idx(struct net *net, | |||
1738 | static void *ipmr_vif_seq_start(struct seq_file *seq, loff_t *pos) | 2021 | static void *ipmr_vif_seq_start(struct seq_file *seq, loff_t *pos) |
1739 | __acquires(mrt_lock) | 2022 | __acquires(mrt_lock) |
1740 | { | 2023 | { |
2024 | struct ipmr_vif_iter *iter = seq->private; | ||
1741 | struct net *net = seq_file_net(seq); | 2025 | struct net *net = seq_file_net(seq); |
2026 | struct mr_table *mrt; | ||
2027 | |||
2028 | mrt = ipmr_get_table(net, RT_TABLE_DEFAULT); | ||
2029 | if (mrt == NULL) | ||
2030 | return ERR_PTR(-ENOENT); | ||
2031 | |||
2032 | iter->mrt = mrt; | ||
1742 | 2033 | ||
1743 | read_lock(&mrt_lock); | 2034 | read_lock(&mrt_lock); |
1744 | return *pos ? ipmr_vif_seq_idx(net, seq->private, *pos - 1) | 2035 | return *pos ? ipmr_vif_seq_idx(net, seq->private, *pos - 1) |
@@ -1749,7 +2040,7 @@ static void *ipmr_vif_seq_next(struct seq_file *seq, void *v, loff_t *pos) | |||
1749 | { | 2040 | { |
1750 | struct ipmr_vif_iter *iter = seq->private; | 2041 | struct ipmr_vif_iter *iter = seq->private; |
1751 | struct net *net = seq_file_net(seq); | 2042 | struct net *net = seq_file_net(seq); |
1752 | struct mr_table *mrt = net->ipv4.mrt; | 2043 | struct mr_table *mrt = iter->mrt; |
1753 | 2044 | ||
1754 | ++*pos; | 2045 | ++*pos; |
1755 | if (v == SEQ_START_TOKEN) | 2046 | if (v == SEQ_START_TOKEN) |
@@ -1771,8 +2062,8 @@ static void ipmr_vif_seq_stop(struct seq_file *seq, void *v) | |||
1771 | 2062 | ||
1772 | static int ipmr_vif_seq_show(struct seq_file *seq, void *v) | 2063 | static int ipmr_vif_seq_show(struct seq_file *seq, void *v) |
1773 | { | 2064 | { |
1774 | struct net *net = seq_file_net(seq); | 2065 | struct ipmr_vif_iter *iter = seq->private; |
1775 | struct mr_table *mrt = net->ipv4.mrt; | 2066 | struct mr_table *mrt = iter->mrt; |
1776 | 2067 | ||
1777 | if (v == SEQ_START_TOKEN) { | 2068 | if (v == SEQ_START_TOKEN) { |
1778 | seq_puts(seq, | 2069 | seq_puts(seq, |
@@ -1814,6 +2105,7 @@ static const struct file_operations ipmr_vif_fops = { | |||
1814 | 2105 | ||
1815 | struct ipmr_mfc_iter { | 2106 | struct ipmr_mfc_iter { |
1816 | struct seq_net_private p; | 2107 | struct seq_net_private p; |
2108 | struct mr_table *mrt; | ||
1817 | struct list_head *cache; | 2109 | struct list_head *cache; |
1818 | int ct; | 2110 | int ct; |
1819 | }; | 2111 | }; |
@@ -1822,7 +2114,7 @@ struct ipmr_mfc_iter { | |||
1822 | static struct mfc_cache *ipmr_mfc_seq_idx(struct net *net, | 2114 | static struct mfc_cache *ipmr_mfc_seq_idx(struct net *net, |
1823 | struct ipmr_mfc_iter *it, loff_t pos) | 2115 | struct ipmr_mfc_iter *it, loff_t pos) |
1824 | { | 2116 | { |
1825 | struct mr_table *mrt = net->ipv4.mrt; | 2117 | struct mr_table *mrt = it->mrt; |
1826 | struct mfc_cache *mfc; | 2118 | struct mfc_cache *mfc; |
1827 | 2119 | ||
1828 | read_lock(&mrt_lock); | 2120 | read_lock(&mrt_lock); |
@@ -1850,7 +2142,13 @@ static void *ipmr_mfc_seq_start(struct seq_file *seq, loff_t *pos) | |||
1850 | { | 2142 | { |
1851 | struct ipmr_mfc_iter *it = seq->private; | 2143 | struct ipmr_mfc_iter *it = seq->private; |
1852 | struct net *net = seq_file_net(seq); | 2144 | struct net *net = seq_file_net(seq); |
2145 | struct mr_table *mrt; | ||
2146 | |||
2147 | mrt = ipmr_get_table(net, RT_TABLE_DEFAULT); | ||
2148 | if (mrt == NULL) | ||
2149 | return ERR_PTR(-ENOENT); | ||
1853 | 2150 | ||
2151 | it->mrt = mrt; | ||
1854 | it->cache = NULL; | 2152 | it->cache = NULL; |
1855 | it->ct = 0; | 2153 | it->ct = 0; |
1856 | return *pos ? ipmr_mfc_seq_idx(net, seq->private, *pos - 1) | 2154 | return *pos ? ipmr_mfc_seq_idx(net, seq->private, *pos - 1) |
@@ -1862,7 +2160,7 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos) | |||
1862 | struct mfc_cache *mfc = v; | 2160 | struct mfc_cache *mfc = v; |
1863 | struct ipmr_mfc_iter *it = seq->private; | 2161 | struct ipmr_mfc_iter *it = seq->private; |
1864 | struct net *net = seq_file_net(seq); | 2162 | struct net *net = seq_file_net(seq); |
1865 | struct mr_table *mrt = net->ipv4.mrt; | 2163 | struct mr_table *mrt = it->mrt; |
1866 | 2164 | ||
1867 | ++*pos; | 2165 | ++*pos; |
1868 | 2166 | ||
@@ -1903,8 +2201,7 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos) | |||
1903 | static void ipmr_mfc_seq_stop(struct seq_file *seq, void *v) | 2201 | static void ipmr_mfc_seq_stop(struct seq_file *seq, void *v) |
1904 | { | 2202 | { |
1905 | struct ipmr_mfc_iter *it = seq->private; | 2203 | struct ipmr_mfc_iter *it = seq->private; |
1906 | struct net *net = seq_file_net(seq); | 2204 | struct mr_table *mrt = it->mrt; |
1907 | struct mr_table *mrt = net->ipv4.mrt; | ||
1908 | 2205 | ||
1909 | if (it->cache == &mrt->mfc_unres_queue) | 2206 | if (it->cache == &mrt->mfc_unres_queue) |
1910 | spin_unlock_bh(&mfc_unres_lock); | 2207 | spin_unlock_bh(&mfc_unres_lock); |
@@ -1915,8 +2212,6 @@ static void ipmr_mfc_seq_stop(struct seq_file *seq, void *v) | |||
1915 | static int ipmr_mfc_seq_show(struct seq_file *seq, void *v) | 2212 | static int ipmr_mfc_seq_show(struct seq_file *seq, void *v) |
1916 | { | 2213 | { |
1917 | int n; | 2214 | int n; |
1918 | struct net *net = seq_file_net(seq); | ||
1919 | struct mr_table *mrt = net->ipv4.mrt; | ||
1920 | 2215 | ||
1921 | if (v == SEQ_START_TOKEN) { | 2216 | if (v == SEQ_START_TOKEN) { |
1922 | seq_puts(seq, | 2217 | seq_puts(seq, |
@@ -1924,6 +2219,7 @@ static int ipmr_mfc_seq_show(struct seq_file *seq, void *v) | |||
1924 | } else { | 2219 | } else { |
1925 | const struct mfc_cache *mfc = v; | 2220 | const struct mfc_cache *mfc = v; |
1926 | const struct ipmr_mfc_iter *it = seq->private; | 2221 | const struct ipmr_mfc_iter *it = seq->private; |
2222 | const struct mr_table *mrt = it->mrt; | ||
1927 | 2223 | ||
1928 | seq_printf(seq, "%08lX %08lX %-3hd", | 2224 | seq_printf(seq, "%08lX %08lX %-3hd", |
1929 | (unsigned long) mfc->mfc_mcastgrp, | 2225 | (unsigned long) mfc->mfc_mcastgrp, |
@@ -1989,28 +2285,11 @@ static const struct net_protocol pim_protocol = { | |||
1989 | */ | 2285 | */ |
1990 | static int __net_init ipmr_net_init(struct net *net) | 2286 | static int __net_init ipmr_net_init(struct net *net) |
1991 | { | 2287 | { |
1992 | struct mr_table *mrt; | 2288 | int err; |
1993 | unsigned int i; | ||
1994 | int err = 0; | ||
1995 | 2289 | ||
1996 | mrt = kzalloc(sizeof(*mrt), GFP_KERNEL); | 2290 | err = ipmr_rules_init(net); |
1997 | if (mrt == NULL) { | 2291 | if (err < 0) |
1998 | err = -ENOMEM; | ||
1999 | goto fail; | 2292 | goto fail; |
2000 | } | ||
2001 | |||
2002 | /* Forwarding cache */ | ||
2003 | for (i = 0; i < MFC_LINES; i++) | ||
2004 | INIT_LIST_HEAD(&mrt->mfc_cache_array[i]); | ||
2005 | |||
2006 | INIT_LIST_HEAD(&mrt->mfc_unres_queue); | ||
2007 | |||
2008 | setup_timer(&mrt->ipmr_expire_timer, ipmr_expire_process, | ||
2009 | (unsigned long)net); | ||
2010 | |||
2011 | #ifdef CONFIG_IP_PIMSM | ||
2012 | mrt->mroute_reg_vif_num = -1; | ||
2013 | #endif | ||
2014 | 2293 | ||
2015 | #ifdef CONFIG_PROC_FS | 2294 | #ifdef CONFIG_PROC_FS |
2016 | err = -ENOMEM; | 2295 | err = -ENOMEM; |
@@ -2019,15 +2298,13 @@ static int __net_init ipmr_net_init(struct net *net) | |||
2019 | if (!proc_net_fops_create(net, "ip_mr_cache", 0, &ipmr_mfc_fops)) | 2298 | if (!proc_net_fops_create(net, "ip_mr_cache", 0, &ipmr_mfc_fops)) |
2020 | goto proc_cache_fail; | 2299 | goto proc_cache_fail; |
2021 | #endif | 2300 | #endif |
2022 | |||
2023 | net->ipv4.mrt = mrt; | ||
2024 | return 0; | 2301 | return 0; |
2025 | 2302 | ||
2026 | #ifdef CONFIG_PROC_FS | 2303 | #ifdef CONFIG_PROC_FS |
2027 | proc_cache_fail: | 2304 | proc_cache_fail: |
2028 | proc_net_remove(net, "ip_mr_vif"); | 2305 | proc_net_remove(net, "ip_mr_vif"); |
2029 | proc_vif_fail: | 2306 | proc_vif_fail: |
2030 | kfree(mrt); | 2307 | ipmr_rules_exit(net); |
2031 | #endif | 2308 | #endif |
2032 | fail: | 2309 | fail: |
2033 | return err; | 2310 | return err; |
@@ -2039,7 +2316,7 @@ static void __net_exit ipmr_net_exit(struct net *net) | |||
2039 | proc_net_remove(net, "ip_mr_cache"); | 2316 | proc_net_remove(net, "ip_mr_cache"); |
2040 | proc_net_remove(net, "ip_mr_vif"); | 2317 | proc_net_remove(net, "ip_mr_vif"); |
2041 | #endif | 2318 | #endif |
2042 | kfree(net->ipv4.mrt); | 2319 | ipmr_rules_exit(net); |
2043 | } | 2320 | } |
2044 | 2321 | ||
2045 | static struct pernet_operations ipmr_net_ops = { | 2322 | static struct pernet_operations ipmr_net_ops = { |