Branch data Line data Source code
1 : : /*
2 : : * INETPEER - A storage for permanent information about peers
3 : : *
4 : : * This source is covered by the GNU GPL, the same as all kernel sources.
5 : : *
6 : : * Authors: Andrey V. Savochkin <saw@msu.ru>
7 : : */
8 : :
9 : : #include <linux/cache.h>
10 : : #include <linux/module.h>
11 : : #include <linux/types.h>
12 : : #include <linux/slab.h>
13 : : #include <linux/interrupt.h>
14 : : #include <linux/spinlock.h>
15 : : #include <linux/random.h>
16 : : #include <linux/timer.h>
17 : : #include <linux/time.h>
18 : : #include <linux/kernel.h>
19 : : #include <linux/mm.h>
20 : : #include <linux/net.h>
21 : : #include <linux/workqueue.h>
22 : : #include <net/ip.h>
23 : : #include <net/inetpeer.h>
24 : : #include <net/secure_seq.h>
25 : :
26 : : /*
27 : : * Theory of operations.
28 : : * We keep one entry for each peer IP address. The nodes contains long-living
29 : : * information about the peer which doesn't depend on routes.
30 : : *
31 : : * Nodes are removed only when reference counter goes to 0.
32 : : * When it's happened the node may be removed when a sufficient amount of
33 : : * time has been passed since its last use. The less-recently-used entry can
34 : : * also be removed if the pool is overloaded i.e. if the total amount of
35 : : * entries is greater-or-equal than the threshold.
36 : : *
37 : : * Node pool is organised as an RB tree.
38 : : * Such an implementation has been chosen not just for fun. It's a way to
39 : : * prevent easy and efficient DoS attacks by creating hash collisions. A huge
40 : : * amount of long living nodes in a single hash slot would significantly delay
41 : : * lookups performed with disabled BHs.
42 : : *
43 : : * Serialisation issues.
44 : : * 1. Nodes may appear in the tree only with the pool lock held.
45 : : * 2. Nodes may disappear from the tree only with the pool lock held
46 : : * AND reference count being 0.
47 : : * 3. Global variable peer_total is modified under the pool lock.
48 : : * 4. struct inet_peer fields modification:
49 : : * rb_node: pool lock
50 : : * refcnt: atomically against modifications on other CPU;
51 : : * usually under some other lock to prevent node disappearing
52 : : * daddr: unchangeable
53 : : */
54 : :
55 : : static struct kmem_cache *peer_cachep __ro_after_init;
56 : :
57 : 39 : void inet_peer_base_init(struct inet_peer_base *bp)
58 : : {
59 : 39 : bp->rb_root = RB_ROOT;
60 : 39 : seqlock_init(&bp->lock);
61 : 39 : bp->total = 0;
62 : 39 : }
63 : : EXPORT_SYMBOL_GPL(inet_peer_base_init);
64 : :
65 : : #define PEER_MAX_GC 32
66 : :
67 : : /* Exported for sysctl_net_ipv4. */
68 : : int inet_peer_threshold __read_mostly = 65536 + 128; /* start to throw entries more
69 : : * aggressively at this stage */
70 : : int inet_peer_minttl __read_mostly = 120 * HZ; /* TTL under high load: 120 sec */
71 : : int inet_peer_maxttl __read_mostly = 10 * 60 * HZ; /* usual time to live: 10 min */
72 : :
73 : : /* Called from ip_output.c:ip_init */
74 : 13 : void __init inet_initpeers(void)
75 : : {
76 : 13 : struct sysinfo si;
77 : :
78 : : /* Use the straight interface to information about memory. */
79 : 13 : si_meminfo(&si);
80 : : /* The values below were suggested by Alexey Kuznetsov
81 : : * <kuznet@ms2.inr.ac.ru>. I don't have any opinion about the values
82 : : * myself. --SAW
83 : : */
84 [ - + ]: 13 : if (si.totalram <= (32768*1024)/PAGE_SIZE)
85 : 0 : inet_peer_threshold >>= 1; /* max pool size about 1MB on IA32 */
86 [ - + ]: 13 : if (si.totalram <= (16384*1024)/PAGE_SIZE)
87 : 0 : inet_peer_threshold >>= 1; /* about 512KB */
88 [ - + ]: 13 : if (si.totalram <= (8192*1024)/PAGE_SIZE)
89 : 0 : inet_peer_threshold >>= 2; /* about 128KB */
90 : :
91 : 13 : peer_cachep = kmem_cache_create("inet_peer_cache",
92 : : sizeof(struct inet_peer),
93 : : 0, SLAB_HWCACHE_ALIGN | SLAB_PANIC,
94 : : NULL);
95 : 13 : }
96 : :
97 : : /* Called with rcu_read_lock() or base->lock held */
98 : 0 : static struct inet_peer *lookup(const struct inetpeer_addr *daddr,
99 : : struct inet_peer_base *base,
100 : : unsigned int seq,
101 : : struct inet_peer *gc_stack[],
102 : : unsigned int *gc_cnt,
103 : : struct rb_node **parent_p,
104 : : struct rb_node ***pp_p)
105 : : {
106 : 0 : struct rb_node **pp, *parent, *next;
107 : 0 : struct inet_peer *p;
108 : :
109 : 0 : pp = &base->rb_root.rb_node;
110 : 0 : parent = NULL;
111 : 0 : while (1) {
112 : 0 : int cmp;
113 : :
114 [ # # ]: 0 : next = rcu_dereference_raw(*pp);
115 [ # # ]: 0 : if (!next)
116 : : break;
117 : 0 : parent = next;
118 : 0 : p = rb_entry(parent, struct inet_peer, rb_node);
119 [ # # ]: 0 : cmp = inetpeer_addr_cmp(daddr, &p->daddr);
120 : 0 : if (cmp == 0) {
121 [ # # ]: 0 : if (!refcount_inc_not_zero(&p->refcnt))
122 : : break;
123 : : return p;
124 : : }
125 [ # # ]: 0 : if (gc_stack) {
126 [ # # ]: 0 : if (*gc_cnt < PEER_MAX_GC)
127 : 0 : gc_stack[(*gc_cnt)++] = p;
128 [ # # ]: 0 : } else if (unlikely(read_seqretry(&base->lock, seq))) {
129 : : break;
130 : : }
131 [ # # ]: 0 : if (cmp == -1)
132 : 0 : pp = &next->rb_left;
133 : : else
134 : 0 : pp = &next->rb_right;
135 : : }
136 : 0 : *parent_p = parent;
137 : 0 : *pp_p = pp;
138 : 0 : return NULL;
139 : : }
140 : :
141 : 0 : static void inetpeer_free_rcu(struct rcu_head *head)
142 : : {
143 : 0 : kmem_cache_free(peer_cachep, container_of(head, struct inet_peer, rcu));
144 : 0 : }
145 : :
146 : : /* perform garbage collect on all items stacked during a lookup */
147 : 0 : static void inet_peer_gc(struct inet_peer_base *base,
148 : : struct inet_peer *gc_stack[],
149 : : unsigned int gc_cnt)
150 : : {
151 : 0 : struct inet_peer *p;
152 : 0 : __u32 delta, ttl;
153 : 0 : int i;
154 : :
155 [ # # ]: 0 : if (base->total >= inet_peer_threshold)
156 : : ttl = 0; /* be aggressive */
157 : : else
158 : 0 : ttl = inet_peer_maxttl
159 : 0 : - (inet_peer_maxttl - inet_peer_minttl) / HZ *
160 : 0 : base->total / inet_peer_threshold * HZ;
161 [ # # ]: 0 : for (i = 0; i < gc_cnt; i++) {
162 : 0 : p = gc_stack[i];
163 : :
164 : : /* The READ_ONCE() pairs with the WRITE_ONCE()
165 : : * in inet_putpeer()
166 : : */
167 [ # # ]: 0 : delta = (__u32)jiffies - READ_ONCE(p->dtime);
168 : :
169 [ # # # # ]: 0 : if (delta < ttl || !refcount_dec_if_one(&p->refcnt))
170 : 0 : gc_stack[i] = NULL;
171 : : }
172 [ # # ]: 0 : for (i = 0; i < gc_cnt; i++) {
173 : 0 : p = gc_stack[i];
174 [ # # ]: 0 : if (p) {
175 : 0 : rb_erase(&p->rb_node, &base->rb_root);
176 : 0 : base->total--;
177 : 0 : call_rcu(&p->rcu, inetpeer_free_rcu);
178 : : }
179 : : }
180 : 0 : }
181 : :
182 : 0 : struct inet_peer *inet_getpeer(struct inet_peer_base *base,
183 : : const struct inetpeer_addr *daddr,
184 : : int create)
185 : : {
186 : 0 : struct inet_peer *p, *gc_stack[PEER_MAX_GC];
187 : 0 : struct rb_node **pp, *parent;
188 : 0 : unsigned int gc_cnt, seq;
189 : 0 : int invalidated;
190 : :
191 : : /* Attempt a lockless lookup first.
192 : : * Because of a concurrent writer, we might not find an existing entry.
193 : : */
194 : 0 : rcu_read_lock();
195 : 0 : seq = read_seqbegin(&base->lock);
196 : 0 : p = lookup(daddr, base, seq, NULL, &gc_cnt, &parent, &pp);
197 : 0 : invalidated = read_seqretry(&base->lock, seq);
198 : 0 : rcu_read_unlock();
199 : :
200 [ # # ]: 0 : if (p)
201 : : return p;
202 : :
203 : : /* If no writer did a change during our lookup, we can return early. */
204 [ # # ]: 0 : if (!create && !invalidated)
205 : : return NULL;
206 : :
207 : : /* retry an exact lookup, taking the lock before.
208 : : * At least, nodes should be hot in our cache.
209 : : */
210 : 0 : parent = NULL;
211 : 0 : write_seqlock_bh(&base->lock);
212 : :
213 : 0 : gc_cnt = 0;
214 : 0 : p = lookup(daddr, base, seq, gc_stack, &gc_cnt, &parent, &pp);
215 [ # # ]: 0 : if (!p && create) {
216 : 0 : p = kmem_cache_alloc(peer_cachep, GFP_ATOMIC);
217 [ # # ]: 0 : if (p) {
218 : 0 : p->daddr = *daddr;
219 : 0 : p->dtime = (__u32)jiffies;
220 : 0 : refcount_set(&p->refcnt, 2);
221 : 0 : atomic_set(&p->rid, 0);
222 : 0 : p->metrics[RTAX_LOCK-1] = INETPEER_METRICS_NEW;
223 : 0 : p->rate_tokens = 0;
224 : 0 : p->n_redirects = 0;
225 : : /* 60*HZ is arbitrary, but chosen enough high so that the first
226 : : * calculation of tokens is at its maximum.
227 : : */
228 : 0 : p->rate_last = jiffies - 60*HZ;
229 : :
230 : 0 : rb_link_node(&p->rb_node, parent, pp);
231 : 0 : rb_insert_color(&p->rb_node, &base->rb_root);
232 : 0 : base->total++;
233 : : }
234 : : }
235 [ # # ]: 0 : if (gc_cnt)
236 : 0 : inet_peer_gc(base, gc_stack, gc_cnt);
237 : 0 : write_sequnlock_bh(&base->lock);
238 : :
239 : 0 : return p;
240 : : }
241 : : EXPORT_SYMBOL_GPL(inet_getpeer);
242 : :
243 : 0 : void inet_putpeer(struct inet_peer *p)
244 : : {
245 : : /* The WRITE_ONCE() pairs with itself (we run lockless)
246 : : * and the READ_ONCE() in inet_peer_gc()
247 : : */
248 : 0 : WRITE_ONCE(p->dtime, (__u32)jiffies);
249 : :
250 [ # # ]: 0 : if (refcount_dec_and_test(&p->refcnt))
251 : 0 : call_rcu(&p->rcu, inetpeer_free_rcu);
252 : 0 : }
253 : : EXPORT_SYMBOL_GPL(inet_putpeer);
254 : :
255 : : /*
256 : : * Check transmit rate limitation for given message.
257 : : * The rate information is held in the inet_peer entries now.
258 : : * This function is generic and could be used for other purposes
259 : : * too. It uses a Token bucket filter as suggested by Alexey Kuznetsov.
260 : : *
261 : : * Note that the same inet_peer fields are modified by functions in
262 : : * route.c too, but these work for packet destinations while xrlim_allow
263 : : * works for icmp destinations. This means the rate limiting information
264 : : * for one "ip object" is shared - and these ICMPs are twice limited:
265 : : * by source and by destination.
266 : : *
267 : : * RFC 1812: 4.3.2.8 SHOULD be able to limit error message rate
268 : : * SHOULD allow setting of rate limits
269 : : *
270 : : * Shared between ICMPv4 and ICMPv6.
271 : : */
272 : : #define XRLIM_BURST_FACTOR 6
273 : 0 : bool inet_peer_xrlim_allow(struct inet_peer *peer, int timeout)
274 : : {
275 : 0 : unsigned long now, token;
276 : 0 : bool rc = false;
277 : :
278 [ # # ]: 0 : if (!peer)
279 : : return true;
280 : :
281 : 0 : token = peer->rate_tokens;
282 : 0 : now = jiffies;
283 : 0 : token += now - peer->rate_last;
284 : 0 : peer->rate_last = now;
285 : 0 : if (token > XRLIM_BURST_FACTOR * timeout)
286 : : token = XRLIM_BURST_FACTOR * timeout;
287 [ # # ]: 0 : if (token >= timeout) {
288 : 0 : token -= timeout;
289 : 0 : rc = true;
290 : : }
291 : 0 : peer->rate_tokens = token;
292 : 0 : return rc;
293 : : }
294 : : EXPORT_SYMBOL(inet_peer_xrlim_allow);
295 : :
296 : 0 : void inetpeer_invalidate_tree(struct inet_peer_base *base)
297 : : {
298 : 0 : struct rb_node *p = rb_first(&base->rb_root);
299 : :
300 [ # # ]: 0 : while (p) {
301 : 0 : struct inet_peer *peer = rb_entry(p, struct inet_peer, rb_node);
302 : :
303 : 0 : p = rb_next(p);
304 : 0 : rb_erase(&peer->rb_node, &base->rb_root);
305 : 0 : inet_putpeer(peer);
306 : 0 : cond_resched();
307 : : }
308 : :
309 : 0 : base->total = 0;
310 : 0 : }
311 : : EXPORT_SYMBOL(inetpeer_invalidate_tree);
|
