connection.c 16.2 KB
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577
/*
 * Copyright (c) 2006 Oracle.  All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 */
#include <linux/kernel.h>
#include <linux/list.h>
#include <linux/slab.h>
#include <linux/export.h>
#include <net/inet_hashtables.h>

#include "rds.h"
#include "loop.h"

#define RDS_CONNECTION_HASH_BITS 12
#define RDS_CONNECTION_HASH_ENTRIES (1 << RDS_CONNECTION_HASH_BITS)
#define RDS_CONNECTION_HASH_MASK (RDS_CONNECTION_HASH_ENTRIES - 1)

/* converting this to RCU is a chore for another day.. */
static DEFINE_SPINLOCK(rds_conn_lock);
static unsigned long rds_conn_count;
static struct hlist_head rds_conn_hash[RDS_CONNECTION_HASH_ENTRIES];
static struct kmem_cache *rds_conn_slab;

static struct hlist_head *rds_conn_bucket(__be32 laddr, __be32 faddr)
{
	static u32 rds_hash_secret __read_mostly;

	unsigned long hash;

	net_get_random_once(&rds_hash_secret, sizeof(rds_hash_secret));

	/* Pass NULL, don't need struct net for hash */
	hash = __inet_ehashfn(be32_to_cpu(laddr), 0,
			      be32_to_cpu(faddr), 0,
			      rds_hash_secret);
	return &rds_conn_hash[hash & RDS_CONNECTION_HASH_MASK];
}

#define rds_conn_info_set(var, test, suffix) do {		\
	if (test)						\
		var |= RDS_INFO_CONNECTION_FLAG_##suffix;	\
} while (0)

/* rcu read lock must be held or the connection spinlock */
static struct rds_connection *rds_conn_lookup(struct hlist_head *head,
					      __be32 laddr, __be32 faddr,
					      struct rds_transport *trans)
{
	struct rds_connection *conn, *ret = NULL;

	hlist_for_each_entry_rcu(conn, head, c_hash_node) {
		if (conn->c_faddr == faddr && conn->c_laddr == laddr &&
				conn->c_trans == trans) {
			ret = conn;
			break;
		}
	}
	rdsdebug("returning conn %p for %pI4 -> %pI4\n", ret,
		 &laddr, &faddr);
	return ret;
}

/*
 * This is called by transports as they're bringing down a connection.
 * It clears partial message state so that the transport can start sending
 * and receiving over this connection again in the future.  It is up to
 * the transport to have serialized this call with its send and recv.
 */
static void rds_conn_reset(struct rds_connection *conn)
{
	rdsdebug("connection %pI4 to %pI4 reset\n",
	  &conn->c_laddr, &conn->c_faddr);

	rds_stats_inc(s_conn_reset);
	rds_send_reset(conn);
	conn->c_flags = 0;

	/* Do not clear next_rx_seq here, else we cannot distinguish
	 * retransmitted packets from new packets, and will hand all
	 * of them to the application. That is not consistent with the
	 * reliability guarantees of RDS. */
}

/*
 * There is only every one 'conn' for a given pair of addresses in the
 * system at a time.  They contain messages to be retransmitted and so
 * span the lifetime of the actual underlying transport connections.
 *
 * For now they are not garbage collected once they're created.  They
 * are torn down as the module is removed, if ever.
 */
static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
				       struct rds_transport *trans, gfp_t gfp,
				       int is_outgoing)
{
	struct rds_connection *conn, *parent = NULL;
	struct hlist_head *head = rds_conn_bucket(laddr, faddr);
	struct rds_transport *loop_trans;
	unsigned long flags;
	int ret;

	rcu_read_lock();
	conn = rds_conn_lookup(head, laddr, faddr, trans);
	if (conn && conn->c_loopback && conn->c_trans != &rds_loop_transport &&
	    !is_outgoing) {
		/* This is a looped back IB connection, and we're
		 * called by the code handling the incoming connect.
		 * We need a second connection object into which we
		 * can stick the other QP. */
		parent = conn;
		conn = parent->c_passive;
	}
	rcu_read_unlock();
	if (conn)
		goto out;

	conn = kmem_cache_zalloc(rds_conn_slab, gfp);
	if (!conn) {
		conn = ERR_PTR(-ENOMEM);
		goto out;
	}

	INIT_HLIST_NODE(&conn->c_hash_node);
	conn->c_laddr = laddr;
	conn->c_faddr = faddr;
	spin_lock_init(&conn->c_lock);
	conn->c_next_tx_seq = 1;

	init_waitqueue_head(&conn->c_waitq);
	INIT_LIST_HEAD(&conn->c_send_queue);
	INIT_LIST_HEAD(&conn->c_retrans);

	ret = rds_cong_get_maps(conn);
	if (ret) {
		kmem_cache_free(rds_conn_slab, conn);
		conn = ERR_PTR(ret);
		goto out;
	}

	/*
	 * This is where a connection becomes loopback.  If *any* RDS sockets
	 * can bind to the destination address then we'd rather the messages
	 * flow through loopback rather than either transport.
	 */
	loop_trans = rds_trans_get_preferred(faddr);
	if (loop_trans) {
		rds_trans_put(loop_trans);
		conn->c_loopback = 1;
		if (is_outgoing && trans->t_prefer_loopback) {
			/* "outgoing" connection - and the transport
			 * says it wants the connection handled by the
			 * loopback transport. This is what TCP does.
			 */
			trans = &rds_loop_transport;
		}
	}

	conn->c_trans = trans;

	ret = trans->conn_alloc(conn, gfp);
	if (ret) {
		kmem_cache_free(rds_conn_slab, conn);
		conn = ERR_PTR(ret);
		goto out;
	}

	atomic_set(&conn->c_state, RDS_CONN_DOWN);
	conn->c_reconnect_jiffies = 0;
	INIT_DELAYED_WORK(&conn->c_send_w, rds_send_worker);
	INIT_DELAYED_WORK(&conn->c_recv_w, rds_recv_worker);
	INIT_DELAYED_WORK(&conn->c_conn_w, rds_connect_worker);
	INIT_WORK(&conn->c_down_w, rds_shutdown_worker);
	mutex_init(&conn->c_cm_lock);
	conn->c_flags = 0;

	rdsdebug("allocated conn %p for %pI4 -> %pI4 over %s %s\n",
	  conn, &laddr, &faddr,
	  trans->t_name ? trans->t_name : "[unknown]",
	  is_outgoing ? "(outgoing)" : "");

	/*
	 * Since we ran without holding the conn lock, someone could
	 * have created the same conn (either normal or passive) in the
	 * interim. We check while holding the lock. If we won, we complete
	 * init and return our conn. If we lost, we rollback and return the
	 * other one.
	 */
	spin_lock_irqsave(&rds_conn_lock, flags);
	if (parent) {
		/* Creating passive conn */
		if (parent->c_passive) {
			trans->conn_free(conn->c_transport_data);
			kmem_cache_free(rds_conn_slab, conn);
			conn = parent->c_passive;
		} else {
			parent->c_passive = conn;
			rds_cong_add_conn(conn);
			rds_conn_count++;
		}
	} else {
		/* Creating normal conn */
		struct rds_connection *found;

		found = rds_conn_lookup(head, laddr, faddr, trans);
		if (found) {
			trans->conn_free(conn->c_transport_data);
			kmem_cache_free(rds_conn_slab, conn);
			conn = found;
		} else {
			hlist_add_head_rcu(&conn->c_hash_node, head);
			rds_cong_add_conn(conn);
			rds_conn_count++;
		}
	}
	spin_unlock_irqrestore(&rds_conn_lock, flags);

out:
	return conn;
}

struct rds_connection *rds_conn_create(__be32 laddr, __be32 faddr,
				       struct rds_transport *trans, gfp_t gfp)
{
	return __rds_conn_create(laddr, faddr, trans, gfp, 0);
}
EXPORT_SYMBOL_GPL(rds_conn_create);

struct rds_connection *rds_conn_create_outgoing(__be32 laddr, __be32 faddr,
				       struct rds_transport *trans, gfp_t gfp)
{
	return __rds_conn_create(laddr, faddr, trans, gfp, 1);
}
EXPORT_SYMBOL_GPL(rds_conn_create_outgoing);

void rds_conn_shutdown(struct rds_connection *conn)
{
	/* shut it down unless it's down already */
	if (!rds_conn_transition(conn, RDS_CONN_DOWN, RDS_CONN_DOWN)) {
		/*
		 * Quiesce the connection mgmt handlers before we start tearing
		 * things down. We don't hold the mutex for the entire
		 * duration of the shutdown operation, else we may be
		 * deadlocking with the CM handler. Instead, the CM event
		 * handler is supposed to check for state DISCONNECTING
		 */
		mutex_lock(&conn->c_cm_lock);
		if (!rds_conn_transition(conn, RDS_CONN_UP, RDS_CONN_DISCONNECTING)
		 && !rds_conn_transition(conn, RDS_CONN_ERROR, RDS_CONN_DISCONNECTING)) {
			rds_conn_error(conn, "shutdown called in state %d\n",
					atomic_read(&conn->c_state));
			mutex_unlock(&conn->c_cm_lock);
			return;
		}
		mutex_unlock(&conn->c_cm_lock);

		wait_event(conn->c_waitq,
			   !test_bit(RDS_IN_XMIT, &conn->c_flags));

		conn->c_trans->conn_shutdown(conn);
		rds_conn_reset(conn);

		if (!rds_conn_transition(conn, RDS_CONN_DISCONNECTING, RDS_CONN_DOWN)) {
			/* This can happen - eg when we're in the middle of tearing
			 * down the connection, and someone unloads the rds module.
			 * Quite reproduceable with loopback connections.
			 * Mostly harmless.
			 */
			rds_conn_error(conn,
				"%s: failed to transition to state DOWN, "
				"current state is %d\n",
				__func__,
				atomic_read(&conn->c_state));
			return;
		}
	}

	/* Then reconnect if it's still live.
	 * The passive side of an IB loopback connection is never added
	 * to the conn hash, so we never trigger a reconnect on this
	 * conn - the reconnect is always triggered by the active peer. */
	cancel_delayed_work_sync(&conn->c_conn_w);
	rcu_read_lock();
	if (!hlist_unhashed(&conn->c_hash_node)) {
		rcu_read_unlock();
		rds_queue_reconnect(conn);
	} else {
		rcu_read_unlock();
	}
}

/*
 * Stop and free a connection.
 *
 * This can only be used in very limited circumstances.  It assumes that once
 * the conn has been shutdown that no one else is referencing the connection.
 * We can only ensure this in the rmmod path in the current code.
 */
void rds_conn_destroy(struct rds_connection *conn)
{
	struct rds_message *rm, *rtmp;
	unsigned long flags;

	rdsdebug("freeing conn %p for %pI4 -> "
		 "%pI4\n", conn, &conn->c_laddr,
		 &conn->c_faddr);

	/* Ensure conn will not be scheduled for reconnect */
	spin_lock_irq(&rds_conn_lock);
	hlist_del_init_rcu(&conn->c_hash_node);
	spin_unlock_irq(&rds_conn_lock);
	synchronize_rcu();

	/* shut the connection down */
	rds_conn_drop(conn);
	flush_work(&conn->c_down_w);

	/* make sure lingering queued work won't try to ref the conn */
	cancel_delayed_work_sync(&conn->c_send_w);
	cancel_delayed_work_sync(&conn->c_recv_w);

	/* tear down queued messages */
	list_for_each_entry_safe(rm, rtmp,
				 &conn->c_send_queue,
				 m_conn_item) {
		list_del_init(&rm->m_conn_item);
		BUG_ON(!list_empty(&rm->m_sock_item));
		rds_message_put(rm);
	}
	if (conn->c_xmit_rm)
		rds_message_put(conn->c_xmit_rm);

	conn->c_trans->conn_free(conn->c_transport_data);

	/*
	 * The congestion maps aren't freed up here.  They're
	 * freed by rds_cong_exit() after all the connections
	 * have been freed.
	 */
	rds_cong_remove_conn(conn);

	BUG_ON(!list_empty(&conn->c_retrans));
	kmem_cache_free(rds_conn_slab, conn);

	spin_lock_irqsave(&rds_conn_lock, flags);
	rds_conn_count--;
	spin_unlock_irqrestore(&rds_conn_lock, flags);
}
EXPORT_SYMBOL_GPL(rds_conn_destroy);

static void rds_conn_message_info(struct socket *sock, unsigned int len,
				  struct rds_info_iterator *iter,
				  struct rds_info_lengths *lens,
				  int want_send)
{
	struct hlist_head *head;
	struct list_head *list;
	struct rds_connection *conn;
	struct rds_message *rm;
	unsigned int total = 0;
	unsigned long flags;
	size_t i;

	len /= sizeof(struct rds_info_message);

	rcu_read_lock();

	for (i = 0, head = rds_conn_hash; i < ARRAY_SIZE(rds_conn_hash);
	     i++, head++) {
		hlist_for_each_entry_rcu(conn, head, c_hash_node) {
			if (want_send)
				list = &conn->c_send_queue;
			else
				list = &conn->c_retrans;

			spin_lock_irqsave(&conn->c_lock, flags);

			/* XXX too lazy to maintain counts.. */
			list_for_each_entry(rm, list, m_conn_item) {
				total++;
				if (total <= len)
					rds_inc_info_copy(&rm->m_inc, iter,
							  conn->c_laddr,
							  conn->c_faddr, 0);
			}

			spin_unlock_irqrestore(&conn->c_lock, flags);
		}
	}
	rcu_read_unlock();

	lens->nr = total;
	lens->each = sizeof(struct rds_info_message);
}

static void rds_conn_message_info_send(struct socket *sock, unsigned int len,
				       struct rds_info_iterator *iter,
				       struct rds_info_lengths *lens)
{
	rds_conn_message_info(sock, len, iter, lens, 1);
}

static void rds_conn_message_info_retrans(struct socket *sock,
					  unsigned int len,
					  struct rds_info_iterator *iter,
					  struct rds_info_lengths *lens)
{
	rds_conn_message_info(sock, len, iter, lens, 0);
}

void rds_for_each_conn_info(struct socket *sock, unsigned int len,
			  struct rds_info_iterator *iter,
			  struct rds_info_lengths *lens,
			  int (*visitor)(struct rds_connection *, void *),
			  size_t item_len)
{
	uint64_t buffer[(item_len + 7) / 8];
	struct hlist_head *head;
	struct rds_connection *conn;
	size_t i;

	rcu_read_lock();

	lens->nr = 0;
	lens->each = item_len;

	for (i = 0, head = rds_conn_hash; i < ARRAY_SIZE(rds_conn_hash);
	     i++, head++) {
		hlist_for_each_entry_rcu(conn, head, c_hash_node) {

			/* XXX no c_lock usage.. */
			if (!visitor(conn, buffer))
				continue;

			/* We copy as much as we can fit in the buffer,
			 * but we count all items so that the caller
			 * can resize the buffer. */
			if (len >= item_len) {
				rds_info_copy(iter, buffer, item_len);
				len -= item_len;
			}
			lens->nr++;
		}
	}
	rcu_read_unlock();
}
EXPORT_SYMBOL_GPL(rds_for_each_conn_info);

static int rds_conn_info_visitor(struct rds_connection *conn,
				  void *buffer)
{
	struct rds_info_connection *cinfo = buffer;

	cinfo->next_tx_seq = conn->c_next_tx_seq;
	cinfo->next_rx_seq = conn->c_next_rx_seq;
	cinfo->laddr = conn->c_laddr;
	cinfo->faddr = conn->c_faddr;
	strncpy(cinfo->transport, conn->c_trans->t_name,
		sizeof(cinfo->transport));
	cinfo->flags = 0;

	rds_conn_info_set(cinfo->flags, test_bit(RDS_IN_XMIT, &conn->c_flags),
			  SENDING);
	/* XXX Future: return the state rather than these funky bits */
	rds_conn_info_set(cinfo->flags,
			  atomic_read(&conn->c_state) == RDS_CONN_CONNECTING,
			  CONNECTING);
	rds_conn_info_set(cinfo->flags,
			  atomic_read(&conn->c_state) == RDS_CONN_UP,
			  CONNECTED);
	return 1;
}

static void rds_conn_info(struct socket *sock, unsigned int len,
			  struct rds_info_iterator *iter,
			  struct rds_info_lengths *lens)
{
	rds_for_each_conn_info(sock, len, iter, lens,
				rds_conn_info_visitor,
				sizeof(struct rds_info_connection));
}

int rds_conn_init(void)
{
	rds_conn_slab = kmem_cache_create("rds_connection",
					  sizeof(struct rds_connection),
					  0, 0, NULL);
	if (!rds_conn_slab)
		return -ENOMEM;

	rds_info_register_func(RDS_INFO_CONNECTIONS, rds_conn_info);
	rds_info_register_func(RDS_INFO_SEND_MESSAGES,
			       rds_conn_message_info_send);
	rds_info_register_func(RDS_INFO_RETRANS_MESSAGES,
			       rds_conn_message_info_retrans);

	return 0;
}

void rds_conn_exit(void)
{
	rds_loop_exit();

	WARN_ON(!hlist_empty(rds_conn_hash));

	kmem_cache_destroy(rds_conn_slab);

	rds_info_deregister_func(RDS_INFO_CONNECTIONS, rds_conn_info);
	rds_info_deregister_func(RDS_INFO_SEND_MESSAGES,
				 rds_conn_message_info_send);
	rds_info_deregister_func(RDS_INFO_RETRANS_MESSAGES,
				 rds_conn_message_info_retrans);
}

/*
 * Force a disconnect
 */
void rds_conn_drop(struct rds_connection *conn)
{
	atomic_set(&conn->c_state, RDS_CONN_ERROR);
	queue_work(rds_wq, &conn->c_down_w);
}
EXPORT_SYMBOL_GPL(rds_conn_drop);

/*
 * If the connection is down, trigger a connect. We may have scheduled a
 * delayed reconnect however - in this case we should not interfere.
 */
void rds_conn_connect_if_down(struct rds_connection *conn)
{
	if (rds_conn_state(conn) == RDS_CONN_DOWN &&
	    !test_and_set_bit(RDS_RECONNECT_PENDING, &conn->c_flags))
		queue_delayed_work(rds_wq, &conn->c_conn_w, 0);
}
EXPORT_SYMBOL_GPL(rds_conn_connect_if_down);

/*
 * An error occurred on the connection
 */
void
__rds_conn_error(struct rds_connection *conn, const char *fmt, ...)
{
	va_list ap;

	va_start(ap, fmt);
	vprintk(fmt, ap);
	va_end(ap);

	rds_conn_drop(conn);
}