inet_timewait_sock.c 12.2 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
/*
 * INET		An implementation of the TCP/IP protocol suite for the LINUX
 *		operating system.  INET is implemented using the  BSD Socket
 *		interface as the means of communication with the user level.
 *
 *		Generic TIME_WAIT sockets functions
 *
 *		From code orinally in TCP
 */

11
#include <linux/kernel.h>
12 13
#include <net/inet_hashtables.h>
#include <net/inet_timewait_sock.h>
14
#include <net/ip.h>
15 16

/* Must be called with locally disabled BHs. */
17 18
static void __inet_twsk_kill(struct inet_timewait_sock *tw,
			     struct inet_hashinfo *hashinfo)
19 20 21 22
{
	struct inet_bind_hashbucket *bhead;
	struct inet_bind_bucket *tb;
	/* Unlink from established hashes. */
23
	rwlock_t *lock = inet_ehash_lockp(hashinfo, tw->tw_hash);
24

25
	write_lock(lock);
26
	if (hlist_unhashed(&tw->tw_node)) {
27
		write_unlock(lock);
28 29 30 31
		return;
	}
	__hlist_del(&tw->tw_node);
	sk_node_init(&tw->tw_node);
32
	write_unlock(lock);
33 34

	/* Disassociate with bind bucket. */
35 36
	bhead = &hashinfo->bhash[inet_bhashfn(twsk_net(tw), tw->tw_num,
			hashinfo->bhash_size)];
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
	spin_lock(&bhead->lock);
	tb = tw->tw_tb;
	__hlist_del(&tw->tw_bind_node);
	tw->tw_tb = NULL;
	inet_bind_bucket_destroy(hashinfo->bind_bucket_cachep, tb);
	spin_unlock(&bhead->lock);
#ifdef SOCK_REFCNT_DEBUG
	if (atomic_read(&tw->tw_refcnt) != 1) {
		printk(KERN_DEBUG "%s timewait_sock %p refcnt=%d\n",
		       tw->tw_prot->name, tw, atomic_read(&tw->tw_refcnt));
	}
#endif
	inet_twsk_put(tw);
}

52 53 54 55 56 57 58 59 60
void inet_twsk_put(struct inet_timewait_sock *tw)
{
	if (atomic_dec_and_test(&tw->tw_refcnt)) {
		struct module *owner = tw->tw_prot->owner;
		twsk_destructor((struct sock *)tw);
#ifdef SOCK_REFCNT_DEBUG
		printk(KERN_DEBUG "%s timewait_sock %p released\n",
		       tw->tw_prot->name, tw);
#endif
61
		release_net(twsk_net(tw));
62 63 64 65 66 67
		kmem_cache_free(tw->tw_prot->twsk_prot->twsk_slab, tw);
		module_put(owner);
	}
}
EXPORT_SYMBOL_GPL(inet_twsk_put);

68 69 70 71 72 73 74 75 76
/*
 * Enter the time wait state. This is called with locally disabled BH.
 * Essentially we whip up a timewait bucket, copy the relevant info into it
 * from the SK, and mess with hash chains and list linkage.
 */
void __inet_twsk_hashdance(struct inet_timewait_sock *tw, struct sock *sk,
			   struct inet_hashinfo *hashinfo)
{
	const struct inet_sock *inet = inet_sk(sk);
77
	const struct inet_connection_sock *icsk = inet_csk(sk);
78
	struct inet_ehash_bucket *ehead = inet_ehash_bucket(hashinfo, sk->sk_hash);
79
	rwlock_t *lock = inet_ehash_lockp(hashinfo, sk->sk_hash);
80 81 82 83 84
	struct inet_bind_hashbucket *bhead;
	/* Step 1: Put TW into bind hash. Original socket stays there too.
	   Note, that any socket with inet->num != 0 MUST be bound in
	   binding cache, even if it is closed.
	 */
85 86
	bhead = &hashinfo->bhash[inet_bhashfn(twsk_net(tw), inet->num,
			hashinfo->bhash_size)];
87
	spin_lock(&bhead->lock);
88
	tw->tw_tb = icsk->icsk_bind_hash;
89
	WARN_ON(!icsk->icsk_bind_hash);
90 91 92
	inet_twsk_add_bind_node(tw, &tw->tw_tb->owners);
	spin_unlock(&bhead->lock);

93
	write_lock(lock);
94 95 96

	/* Step 2: Remove SK from established hash. */
	if (__sk_del_node_init(sk))
97
		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
98

99 100
	/* Step 3: Hash TW into TIMEWAIT chain. */
	inet_twsk_add_node(tw, &ehead->twchain);
101 102
	atomic_inc(&tw->tw_refcnt);

103
	write_unlock(lock);
104
}
105

106 107
EXPORT_SYMBOL_GPL(__inet_twsk_hashdance);

108 109
struct inet_timewait_sock *inet_twsk_alloc(const struct sock *sk, const int state)
{
110 111
	struct inet_timewait_sock *tw =
		kmem_cache_alloc(sk->sk_prot_creator->twsk_prot->twsk_slab,
112
				 GFP_ATOMIC);
113 114 115 116 117 118 119 120 121 122 123 124 125 126
	if (tw != NULL) {
		const struct inet_sock *inet = inet_sk(sk);

		/* Give us an identity. */
		tw->tw_daddr	    = inet->daddr;
		tw->tw_rcv_saddr    = inet->rcv_saddr;
		tw->tw_bound_dev_if = sk->sk_bound_dev_if;
		tw->tw_num	    = inet->num;
		tw->tw_state	    = TCP_TIME_WAIT;
		tw->tw_substate	    = state;
		tw->tw_sport	    = inet->sport;
		tw->tw_dport	    = inet->dport;
		tw->tw_family	    = sk->sk_family;
		tw->tw_reuse	    = sk->sk_reuse;
127
		tw->tw_hash	    = sk->sk_hash;
128 129
		tw->tw_ipv6only	    = 0;
		tw->tw_prot	    = sk->sk_prot_creator;
130
		twsk_net_set(tw, hold_net(sock_net(sk)));
131 132
		atomic_set(&tw->tw_refcnt, 1);
		inet_twsk_dead_node_init(tw);
133
		__module_get(tw->tw_prot->owner);
134 135 136 137
	}

	return tw;
}
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

EXPORT_SYMBOL_GPL(inet_twsk_alloc);

/* Returns non-zero if quota exceeded.  */
static int inet_twdr_do_twkill_work(struct inet_timewait_death_row *twdr,
				    const int slot)
{
	struct inet_timewait_sock *tw;
	struct hlist_node *node;
	unsigned int killed;
	int ret;

	/* NOTE: compare this to previous version where lock
	 * was released after detaching chain. It was racy,
	 * because tw buckets are scheduled in not serialized context
	 * in 2.3 (with netfilter), and with softnet it is common, because
	 * soft irqs are not sequenced.
	 */
	killed = 0;
	ret = 0;
rescan:
	inet_twsk_for_each_inmate(tw, node, &twdr->cells[slot]) {
		__inet_twsk_del_dead_node(tw);
		spin_unlock(&twdr->death_lock);
		__inet_twsk_kill(tw, twdr->hashinfo);
163 164 165
#ifdef CONFIG_NET_NS
		NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITED);
#endif
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183
		inet_twsk_put(tw);
		killed++;
		spin_lock(&twdr->death_lock);
		if (killed > INET_TWDR_TWKILL_QUOTA) {
			ret = 1;
			break;
		}

		/* While we dropped twdr->death_lock, another cpu may have
		 * killed off the next TW bucket in the list, therefore
		 * do a fresh re-read of the hlist head node with the
		 * lock reacquired.  We still use the hlist traversal
		 * macro in order to get the prefetches.
		 */
		goto rescan;
	}

	twdr->tw_count -= killed;
184 185 186
#ifndef CONFIG_NET_NS
	NET_ADD_STATS_BH(&init_net, LINUX_MIB_TIMEWAITED, killed);
#endif
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219
	return ret;
}

void inet_twdr_hangman(unsigned long data)
{
	struct inet_timewait_death_row *twdr;
	int unsigned need_timer;

	twdr = (struct inet_timewait_death_row *)data;
	spin_lock(&twdr->death_lock);

	if (twdr->tw_count == 0)
		goto out;

	need_timer = 0;
	if (inet_twdr_do_twkill_work(twdr, twdr->slot)) {
		twdr->thread_slots |= (1 << twdr->slot);
		schedule_work(&twdr->twkill_work);
		need_timer = 1;
	} else {
		/* We purged the entire slot, anything left?  */
		if (twdr->tw_count)
			need_timer = 1;
	}
	twdr->slot = ((twdr->slot + 1) & (INET_TWDR_TWKILL_SLOTS - 1));
	if (need_timer)
		mod_timer(&twdr->tw_timer, jiffies + twdr->period);
out:
	spin_unlock(&twdr->death_lock);
}

EXPORT_SYMBOL_GPL(inet_twdr_hangman);

220
void inet_twdr_twkill_work(struct work_struct *work)
221
{
222 223
	struct inet_timewait_death_row *twdr =
		container_of(work, struct inet_timewait_death_row, twkill_work);
224 225
	int i;

226 227
	BUILD_BUG_ON((INET_TWDR_TWKILL_SLOTS - 1) >
			(sizeof(twdr->thread_slots) * 8));
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316

	while (twdr->thread_slots) {
		spin_lock_bh(&twdr->death_lock);
		for (i = 0; i < INET_TWDR_TWKILL_SLOTS; i++) {
			if (!(twdr->thread_slots & (1 << i)))
				continue;

			while (inet_twdr_do_twkill_work(twdr, i) != 0) {
				if (need_resched()) {
					spin_unlock_bh(&twdr->death_lock);
					schedule();
					spin_lock_bh(&twdr->death_lock);
				}
			}

			twdr->thread_slots &= ~(1 << i);
		}
		spin_unlock_bh(&twdr->death_lock);
	}
}

EXPORT_SYMBOL_GPL(inet_twdr_twkill_work);

/* These are always called from BH context.  See callers in
 * tcp_input.c to verify this.
 */

/* This is for handling early-kills of TIME_WAIT sockets. */
void inet_twsk_deschedule(struct inet_timewait_sock *tw,
			  struct inet_timewait_death_row *twdr)
{
	spin_lock(&twdr->death_lock);
	if (inet_twsk_del_dead_node(tw)) {
		inet_twsk_put(tw);
		if (--twdr->tw_count == 0)
			del_timer(&twdr->tw_timer);
	}
	spin_unlock(&twdr->death_lock);
	__inet_twsk_kill(tw, twdr->hashinfo);
}

EXPORT_SYMBOL(inet_twsk_deschedule);

void inet_twsk_schedule(struct inet_timewait_sock *tw,
		       struct inet_timewait_death_row *twdr,
		       const int timeo, const int timewait_len)
{
	struct hlist_head *list;
	int slot;

	/* timeout := RTO * 3.5
	 *
	 * 3.5 = 1+2+0.5 to wait for two retransmits.
	 *
	 * RATIONALE: if FIN arrived and we entered TIME-WAIT state,
	 * our ACK acking that FIN can be lost. If N subsequent retransmitted
	 * FINs (or previous seqments) are lost (probability of such event
	 * is p^(N+1), where p is probability to lose single packet and
	 * time to detect the loss is about RTO*(2^N - 1) with exponential
	 * backoff). Normal timewait length is calculated so, that we
	 * waited at least for one retransmitted FIN (maximal RTO is 120sec).
	 * [ BTW Linux. following BSD, violates this requirement waiting
	 *   only for 60sec, we should wait at least for 240 secs.
	 *   Well, 240 consumes too much of resources 8)
	 * ]
	 * This interval is not reduced to catch old duplicate and
	 * responces to our wandering segments living for two MSLs.
	 * However, if we use PAWS to detect
	 * old duplicates, we can reduce the interval to bounds required
	 * by RTO, rather than MSL. So, if peer understands PAWS, we
	 * kill tw bucket after 3.5*RTO (it is important that this number
	 * is greater than TS tick!) and detect old duplicates with help
	 * of PAWS.
	 */
	slot = (timeo + (1 << INET_TWDR_RECYCLE_TICK) - 1) >> INET_TWDR_RECYCLE_TICK;

	spin_lock(&twdr->death_lock);

	/* Unlink it, if it was scheduled */
	if (inet_twsk_del_dead_node(tw))
		twdr->tw_count--;
	else
		atomic_inc(&tw->tw_refcnt);

	if (slot >= INET_TWDR_RECYCLE_SLOTS) {
		/* Schedule to slow timer */
		if (timeo >= timewait_len) {
			slot = INET_TWDR_TWKILL_SLOTS - 1;
		} else {
317
			slot = DIV_ROUND_UP(timeo, twdr->period);
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
			if (slot >= INET_TWDR_TWKILL_SLOTS)
				slot = INET_TWDR_TWKILL_SLOTS - 1;
		}
		tw->tw_ttd = jiffies + timeo;
		slot = (twdr->slot + slot) & (INET_TWDR_TWKILL_SLOTS - 1);
		list = &twdr->cells[slot];
	} else {
		tw->tw_ttd = jiffies + (slot << INET_TWDR_RECYCLE_TICK);

		if (twdr->twcal_hand < 0) {
			twdr->twcal_hand = 0;
			twdr->twcal_jiffie = jiffies;
			twdr->twcal_timer.expires = twdr->twcal_jiffie +
					      (slot << INET_TWDR_RECYCLE_TICK);
			add_timer(&twdr->twcal_timer);
		} else {
			if (time_after(twdr->twcal_timer.expires,
				       jiffies + (slot << INET_TWDR_RECYCLE_TICK)))
				mod_timer(&twdr->twcal_timer,
					  jiffies + (slot << INET_TWDR_RECYCLE_TICK));
			slot = (twdr->twcal_hand + slot) & (INET_TWDR_RECYCLE_SLOTS - 1);
		}
		list = &twdr->twcal_row[slot];
	}

	hlist_add_head(&tw->tw_death_node, list);

	if (twdr->tw_count++ == 0)
		mod_timer(&twdr->tw_timer, jiffies + twdr->period);
	spin_unlock(&twdr->death_lock);
}

EXPORT_SYMBOL_GPL(inet_twsk_schedule);

void inet_twdr_twcal_tick(unsigned long data)
{
	struct inet_timewait_death_row *twdr;
	int n, slot;
	unsigned long j;
	unsigned long now = jiffies;
	int killed = 0;
	int adv = 0;

	twdr = (struct inet_timewait_death_row *)data;

	spin_lock(&twdr->death_lock);
	if (twdr->twcal_hand < 0)
		goto out;

	slot = twdr->twcal_hand;
	j = twdr->twcal_jiffie;

	for (n = 0; n < INET_TWDR_RECYCLE_SLOTS; n++) {
		if (time_before_eq(j, now)) {
			struct hlist_node *node, *safe;
			struct inet_timewait_sock *tw;

			inet_twsk_for_each_inmate_safe(tw, node, safe,
						       &twdr->twcal_row[slot]) {
				__inet_twsk_del_dead_node(tw);
				__inet_twsk_kill(tw, twdr->hashinfo);
379 380 381
#ifdef CONFIG_NET_NS
				NET_INC_STATS_BH(twsk_net(tw), LINUX_MIB_TIMEWAITKILLED);
#endif
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
				inet_twsk_put(tw);
				killed++;
			}
		} else {
			if (!adv) {
				adv = 1;
				twdr->twcal_jiffie = j;
				twdr->twcal_hand = slot;
			}

			if (!hlist_empty(&twdr->twcal_row[slot])) {
				mod_timer(&twdr->twcal_timer, j);
				goto out;
			}
		}
		j += 1 << INET_TWDR_RECYCLE_TICK;
		slot = (slot + 1) & (INET_TWDR_RECYCLE_SLOTS - 1);
	}
	twdr->twcal_hand = -1;

out:
	if ((twdr->tw_count -= killed) == 0)
		del_timer(&twdr->tw_timer);
405 406 407
#ifndef CONFIG_NET_NS
	NET_ADD_STATS_BH(&init_net, LINUX_MIB_TIMEWAITKILLED, killed);
#endif
408 409 410 411
	spin_unlock(&twdr->death_lock);
}

EXPORT_SYMBOL_GPL(inet_twdr_twcal_tick);
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446

void inet_twsk_purge(struct net *net, struct inet_hashinfo *hashinfo,
		     struct inet_timewait_death_row *twdr, int family)
{
	struct inet_timewait_sock *tw;
	struct sock *sk;
	struct hlist_node *node;
	int h;

	local_bh_disable();
	for (h = 0; h < (hashinfo->ehash_size); h++) {
		struct inet_ehash_bucket *head =
			inet_ehash_bucket(hashinfo, h);
		rwlock_t *lock = inet_ehash_lockp(hashinfo, h);
restart:
		write_lock(lock);
		sk_for_each(sk, node, &head->twchain) {

			tw = inet_twsk(sk);
			if (!net_eq(twsk_net(tw), net) ||
			    tw->tw_family != family)
				continue;

			atomic_inc(&tw->tw_refcnt);
			write_unlock(lock);
			inet_twsk_deschedule(tw, twdr);
			inet_twsk_put(tw);

			goto restart;
		}
		write_unlock(lock);
	}
	local_bh_enable();
}
EXPORT_SYMBOL_GPL(inet_twsk_purge);