1/* 2 * Copyright (c) 2004 Topspin Communications. All rights reserved. 3 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved. 4 * Copyright (c) 2004 Voltaire, Inc. All rights reserved. 5 * 6 * This software is available to you under a choice of one of two 7 * licenses. You may choose to be licensed under the terms of the GNU 8 * General Public License (GPL) Version 2, available from the file 9 * COPYING in the main directory of this source tree, or the 10 * OpenIB.org BSD license below: 11 * 12 * Redistribution and use in source and binary forms, with or 13 * without modification, are permitted provided that the following 14 * conditions are met: 15 * 16 * - Redistributions of source code must retain the above 17 * copyright notice, this list of conditions and the following 18 * disclaimer. 19 * 20 * - Redistributions in binary form must reproduce the above 21 * copyright notice, this list of conditions and the following 22 * disclaimer in the documentation and/or other materials 23 * provided with the distribution. 24 * 25 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 26 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 27 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND 28 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS 29 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN 30 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN 31 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 32 * SOFTWARE. 33 */ 34 35#include "ipoib.h" 36 37#include <linux/module.h> 38 39#include <linux/init.h> 40#include <linux/slab.h> 41#include <linux/kernel.h> 42#include <linux/vmalloc.h> 43 44#include <linux/if_arp.h> /* For ARPHRD_xxx */ 45 46#include <linux/ip.h> 47#include <linux/in.h> 48 49#include <linux/jhash.h> 50#include <net/arp.h> 51#include <net/addrconf.h> 52#include <linux/inetdevice.h> 53#include <rdma/ib_cache.h> 54 55MODULE_AUTHOR("Roland Dreier"); 56MODULE_DESCRIPTION("IP-over-InfiniBand net driver"); 57MODULE_LICENSE("Dual BSD/GPL"); 58 59int ipoib_sendq_size __read_mostly = IPOIB_TX_RING_SIZE; 60int ipoib_recvq_size __read_mostly = IPOIB_RX_RING_SIZE; 61 62module_param_named(send_queue_size, ipoib_sendq_size, int, 0444); 63MODULE_PARM_DESC(send_queue_size, "Number of descriptors in send queue"); 64module_param_named(recv_queue_size, ipoib_recvq_size, int, 0444); 65MODULE_PARM_DESC(recv_queue_size, "Number of descriptors in receive queue"); 66 67#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 68int ipoib_debug_level; 69 70module_param_named(debug_level, ipoib_debug_level, int, 0644); 71MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0"); 72#endif 73 74struct ipoib_path_iter { 75 struct net_device *dev; 76 struct ipoib_path path; 77}; 78 79static const u8 ipv4_bcast_addr[] = { 80 0x00, 0xff, 0xff, 0xff, 81 0xff, 0x12, 0x40, 0x1b, 0x00, 0x00, 0x00, 0x00, 82 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff 83}; 84 85struct workqueue_struct *ipoib_workqueue; 86 87struct ib_sa_client ipoib_sa_client; 88 89static int ipoib_add_one(struct ib_device *device); 90static void ipoib_remove_one(struct ib_device *device, void *client_data); 91static void ipoib_neigh_reclaim(struct rcu_head *rp); 92static struct net_device *ipoib_get_net_dev_by_params( 93 struct ib_device *dev, u8 port, u16 pkey, 94 const union ib_gid *gid, const struct sockaddr *addr, 95 void *client_data); 96static int ipoib_set_mac(struct net_device *dev, void *addr); 97static int ipoib_ioctl(struct net_device *dev, struct ifreq *ifr, 98 int cmd); 99 100static struct ib_client ipoib_client = { 101 .name = "ipoib", 102 .add = ipoib_add_one, 103 .remove = ipoib_remove_one, 104 .get_net_dev_by_params = ipoib_get_net_dev_by_params, 105}; 106 107#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 108static int ipoib_netdev_event(struct notifier_block *this, 109 unsigned long event, void *ptr) 110{ 111 struct netdev_notifier_info *ni = ptr; 112 struct net_device *dev = ni->dev; 113 114 if (dev->netdev_ops->ndo_open != ipoib_open) 115 return NOTIFY_DONE; 116 117 switch (event) { 118 case NETDEV_REGISTER: 119 ipoib_create_debug_files(dev); 120 break; 121 case NETDEV_CHANGENAME: 122 ipoib_delete_debug_files(dev); 123 ipoib_create_debug_files(dev); 124 break; 125 case NETDEV_UNREGISTER: 126 ipoib_delete_debug_files(dev); 127 break; 128 } 129 130 return NOTIFY_DONE; 131} 132#endif 133 134int ipoib_open(struct net_device *dev) 135{ 136 struct ipoib_dev_priv *priv = ipoib_priv(dev); 137 138 ipoib_dbg(priv, "bringing up interface\n"); 139 140 netif_carrier_off(dev); 141 142 set_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); 143 144 priv->sm_fullmember_sendonly_support = false; 145 146 if (ipoib_ib_dev_open(dev)) { 147 if (!test_bit(IPOIB_PKEY_ASSIGNED, &priv->flags)) 148 return 0; 149 goto err_disable; 150 } 151 152 ipoib_ib_dev_up(dev); 153 154 if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { 155 struct ipoib_dev_priv *cpriv; 156 157 /* Bring up any child interfaces too */ 158 down_read(&priv->vlan_rwsem); 159 list_for_each_entry(cpriv, &priv->child_intfs, list) { 160 int flags; 161 162 flags = cpriv->dev->flags; 163 if (flags & IFF_UP) 164 continue; 165 166 dev_change_flags(cpriv->dev, flags | IFF_UP, NULL); 167 } 168 up_read(&priv->vlan_rwsem); 169 } 170 171 netif_start_queue(dev); 172 173 return 0; 174 175err_disable: 176 clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); 177 178 return -EINVAL; 179} 180 181static int ipoib_stop(struct net_device *dev) 182{ 183 struct ipoib_dev_priv *priv = ipoib_priv(dev); 184 185 ipoib_dbg(priv, "stopping interface\n"); 186 187 clear_bit(IPOIB_FLAG_ADMIN_UP, &priv->flags); 188 189 netif_stop_queue(dev); 190 191 ipoib_ib_dev_down(dev); 192 ipoib_ib_dev_stop(dev); 193 194 if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { 195 struct ipoib_dev_priv *cpriv; 196 197 /* Bring down any child interfaces too */ 198 down_read(&priv->vlan_rwsem); 199 list_for_each_entry(cpriv, &priv->child_intfs, list) { 200 int flags; 201 202 flags = cpriv->dev->flags; 203 if (!(flags & IFF_UP)) 204 continue; 205 206 dev_change_flags(cpriv->dev, flags & ~IFF_UP, NULL); 207 } 208 up_read(&priv->vlan_rwsem); 209 } 210 211 return 0; 212} 213 214static netdev_features_t ipoib_fix_features(struct net_device *dev, netdev_features_t features) 215{ 216 struct ipoib_dev_priv *priv = ipoib_priv(dev); 217 218 if (test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags)) 219 features &= ~(NETIF_F_IP_CSUM | NETIF_F_TSO); 220 221 return features; 222} 223 224static int ipoib_change_mtu(struct net_device *dev, int new_mtu) 225{ 226 struct ipoib_dev_priv *priv = ipoib_priv(dev); 227 int ret = 0; 228 229 /* dev->mtu > 2K ==> connected mode */ 230 if (ipoib_cm_admin_enabled(dev)) { 231 if (new_mtu > ipoib_cm_max_mtu(dev)) 232 return -EINVAL; 233 234 if (new_mtu > priv->mcast_mtu) 235 ipoib_warn(priv, "mtu > %d will cause multicast packet drops.\n", 236 priv->mcast_mtu); 237 238 dev->mtu = new_mtu; 239 return 0; 240 } 241 242 if (new_mtu < (ETH_MIN_MTU + IPOIB_ENCAP_LEN) || 243 new_mtu > IPOIB_UD_MTU(priv->max_ib_mtu)) 244 return -EINVAL; 245 246 priv->admin_mtu = new_mtu; 247 248 if (priv->mcast_mtu < priv->admin_mtu) 249 ipoib_dbg(priv, "MTU must be smaller than the underlying " 250 "link layer MTU - 4 (%u)\n", priv->mcast_mtu); 251 252 new_mtu = min(priv->mcast_mtu, priv->admin_mtu); 253 254 if (priv->rn_ops->ndo_change_mtu) { 255 bool carrier_status = netif_carrier_ok(dev); 256 257 netif_carrier_off(dev); 258 259 /* notify lower level on the real mtu */ 260 ret = priv->rn_ops->ndo_change_mtu(dev, new_mtu); 261 262 if (carrier_status) 263 netif_carrier_on(dev); 264 } else { 265 dev->mtu = new_mtu; 266 } 267 268 return ret; 269} 270 271static void ipoib_get_stats(struct net_device *dev, 272 struct rtnl_link_stats64 *stats) 273{ 274 struct ipoib_dev_priv *priv = ipoib_priv(dev); 275 276 if (priv->rn_ops->ndo_get_stats64) 277 priv->rn_ops->ndo_get_stats64(dev, stats); 278 else 279 netdev_stats_to_stats64(stats, &dev->stats); 280} 281 282/* Called with an RCU read lock taken */ 283static bool ipoib_is_dev_match_addr_rcu(const struct sockaddr *addr, 284 struct net_device *dev) 285{ 286 struct net *net = dev_net(dev); 287 struct in_device *in_dev; 288 struct sockaddr_in *addr_in = (struct sockaddr_in *)addr; 289 struct sockaddr_in6 *addr_in6 = (struct sockaddr_in6 *)addr; 290 __be32 ret_addr; 291 292 switch (addr->sa_family) { 293 case AF_INET: 294 in_dev = in_dev_get(dev); 295 if (!in_dev) 296 return false; 297 298 ret_addr = inet_confirm_addr(net, in_dev, 0, 299 addr_in->sin_addr.s_addr, 300 RT_SCOPE_HOST); 301 in_dev_put(in_dev); 302 if (ret_addr) 303 return true; 304 305 break; 306 case AF_INET6: 307 if (IS_ENABLED(CONFIG_IPV6) && 308 ipv6_chk_addr(net, &addr_in6->sin6_addr, dev, 1)) 309 return true; 310 311 break; 312 } 313 return false; 314} 315 316/** 317 * Find the master net_device on top of the given net_device. 318 * @dev: base IPoIB net_device 319 * 320 * Returns the master net_device with a reference held, or the same net_device 321 * if no master exists. 322 */ 323static struct net_device *ipoib_get_master_net_dev(struct net_device *dev) 324{ 325 struct net_device *master; 326 327 rcu_read_lock(); 328 master = netdev_master_upper_dev_get_rcu(dev); 329 if (master) 330 dev_hold(master); 331 rcu_read_unlock(); 332 333 if (master) 334 return master; 335 336 dev_hold(dev); 337 return dev; 338} 339 340struct ipoib_walk_data { 341 const struct sockaddr *addr; 342 struct net_device *result; 343}; 344 345static int ipoib_upper_walk(struct net_device *upper, 346 struct netdev_nested_priv *priv) 347{ 348 struct ipoib_walk_data *data = (struct ipoib_walk_data *)priv->data; 349 int ret = 0; 350 351 if (ipoib_is_dev_match_addr_rcu(data->addr, upper)) { 352 dev_hold(upper); 353 data->result = upper; 354 ret = 1; 355 } 356 357 return ret; 358} 359 360/** 361 * Find a net_device matching the given address, which is an upper device of 362 * the given net_device. 363 * @addr: IP address to look for. 364 * @dev: base IPoIB net_device 365 * 366 * If found, returns the net_device with a reference held. Otherwise return 367 * NULL. 368 */ 369static struct net_device *ipoib_get_net_dev_match_addr( 370 const struct sockaddr *addr, struct net_device *dev) 371{ 372 struct netdev_nested_priv priv; 373 struct ipoib_walk_data data = { 374 .addr = addr, 375 }; 376 377 priv.data = (void *)&data; 378 rcu_read_lock(); 379 if (ipoib_is_dev_match_addr_rcu(addr, dev)) { 380 dev_hold(dev); 381 data.result = dev; 382 goto out; 383 } 384 385 netdev_walk_all_upper_dev_rcu(dev, ipoib_upper_walk, &priv); 386out: 387 rcu_read_unlock(); 388 return data.result; 389} 390 391/* returns the number of IPoIB netdevs on top a given ipoib device matching a 392 * pkey_index and address, if one exists. 393 * 394 * @found_net_dev: contains a matching net_device if the return value >= 1, 395 * with a reference held. */ 396static int ipoib_match_gid_pkey_addr(struct ipoib_dev_priv *priv, 397 const union ib_gid *gid, 398 u16 pkey_index, 399 const struct sockaddr *addr, 400 int nesting, 401 struct net_device **found_net_dev) 402{ 403 struct ipoib_dev_priv *child_priv; 404 struct net_device *net_dev = NULL; 405 int matches = 0; 406 407 if (priv->pkey_index == pkey_index && 408 (!gid || !memcmp(gid, &priv->local_gid, sizeof(*gid)))) { 409 if (!addr) { 410 net_dev = ipoib_get_master_net_dev(priv->dev); 411 } else { 412 /* Verify the net_device matches the IP address, as 413 * IPoIB child devices currently share a GID. */ 414 net_dev = ipoib_get_net_dev_match_addr(addr, priv->dev); 415 } 416 if (net_dev) { 417 if (!*found_net_dev) 418 *found_net_dev = net_dev; 419 else 420 dev_put(net_dev); 421 ++matches; 422 } 423 } 424 425 /* Check child interfaces */ 426 down_read_nested(&priv->vlan_rwsem, nesting); 427 list_for_each_entry(child_priv, &priv->child_intfs, list) { 428 matches += ipoib_match_gid_pkey_addr(child_priv, gid, 429 pkey_index, addr, 430 nesting + 1, 431 found_net_dev); 432 if (matches > 1) 433 break; 434 } 435 up_read(&priv->vlan_rwsem); 436 437 return matches; 438} 439 440/* Returns the number of matching net_devs found (between 0 and 2). Also 441 * return the matching net_device in the @net_dev parameter, holding a 442 * reference to the net_device, if the number of matches >= 1 */ 443static int __ipoib_get_net_dev_by_params(struct list_head *dev_list, u8 port, 444 u16 pkey_index, 445 const union ib_gid *gid, 446 const struct sockaddr *addr, 447 struct net_device **net_dev) 448{ 449 struct ipoib_dev_priv *priv; 450 int matches = 0; 451 452 *net_dev = NULL; 453 454 list_for_each_entry(priv, dev_list, list) { 455 if (priv->port != port) 456 continue; 457 458 matches += ipoib_match_gid_pkey_addr(priv, gid, pkey_index, 459 addr, 0, net_dev); 460 if (matches > 1) 461 break; 462 } 463 464 return matches; 465} 466 467static struct net_device *ipoib_get_net_dev_by_params( 468 struct ib_device *dev, u8 port, u16 pkey, 469 const union ib_gid *gid, const struct sockaddr *addr, 470 void *client_data) 471{ 472 struct net_device *net_dev; 473 struct list_head *dev_list = client_data; 474 u16 pkey_index; 475 int matches; 476 int ret; 477 478 if (!rdma_protocol_ib(dev, port)) 479 return NULL; 480 481 ret = ib_find_cached_pkey(dev, port, pkey, &pkey_index); 482 if (ret) 483 return NULL; 484 485 /* See if we can find a unique device matching the L2 parameters */ 486 matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, 487 gid, NULL, &net_dev); 488 489 switch (matches) { 490 case 0: 491 return NULL; 492 case 1: 493 return net_dev; 494 } 495 496 dev_put(net_dev); 497 498 /* Couldn't find a unique device with L2 parameters only. Use L3 499 * address to uniquely match the net device */ 500 matches = __ipoib_get_net_dev_by_params(dev_list, port, pkey_index, 501 gid, addr, &net_dev); 502 switch (matches) { 503 case 0: 504 return NULL; 505 default: 506 dev_warn_ratelimited(&dev->dev, 507 "duplicate IP address detected\n"); 508 fallthrough; 509 case 1: 510 return net_dev; 511 } 512} 513 514int ipoib_set_mode(struct net_device *dev, const char *buf) 515{ 516 struct ipoib_dev_priv *priv = ipoib_priv(dev); 517 518 if ((test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags) && 519 !strcmp(buf, "connected\n")) || 520 (!test_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags) && 521 !strcmp(buf, "datagram\n"))) { 522 return 0; 523 } 524 525 /* flush paths if we switch modes so that connections are restarted */ 526 if (IPOIB_CM_SUPPORTED(dev->dev_addr) && !strcmp(buf, "connected\n")) { 527 set_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); 528 ipoib_warn(priv, "enabling connected mode " 529 "will cause multicast packet drops\n"); 530 netdev_update_features(dev); 531 dev_set_mtu(dev, ipoib_cm_max_mtu(dev)); 532 netif_set_real_num_tx_queues(dev, 1); 533 rtnl_unlock(); 534 priv->tx_wr.wr.send_flags &= ~IB_SEND_IP_CSUM; 535 536 ipoib_flush_paths(dev); 537 return (!rtnl_trylock()) ? -EBUSY : 0; 538 } 539 540 if (!strcmp(buf, "datagram\n")) { 541 clear_bit(IPOIB_FLAG_ADMIN_CM, &priv->flags); 542 netdev_update_features(dev); 543 dev_set_mtu(dev, min(priv->mcast_mtu, dev->mtu)); 544 netif_set_real_num_tx_queues(dev, dev->num_tx_queues); 545 rtnl_unlock(); 546 ipoib_flush_paths(dev); 547 return (!rtnl_trylock()) ? -EBUSY : 0; 548 } 549 550 return -EINVAL; 551} 552 553struct ipoib_path *__path_find(struct net_device *dev, void *gid) 554{ 555 struct ipoib_dev_priv *priv = ipoib_priv(dev); 556 struct rb_node *n = priv->path_tree.rb_node; 557 struct ipoib_path *path; 558 int ret; 559 560 while (n) { 561 path = rb_entry(n, struct ipoib_path, rb_node); 562 563 ret = memcmp(gid, path->pathrec.dgid.raw, 564 sizeof (union ib_gid)); 565 566 if (ret < 0) 567 n = n->rb_left; 568 else if (ret > 0) 569 n = n->rb_right; 570 else 571 return path; 572 } 573 574 return NULL; 575} 576 577static int __path_add(struct net_device *dev, struct ipoib_path *path) 578{ 579 struct ipoib_dev_priv *priv = ipoib_priv(dev); 580 struct rb_node **n = &priv->path_tree.rb_node; 581 struct rb_node *pn = NULL; 582 struct ipoib_path *tpath; 583 int ret; 584 585 while (*n) { 586 pn = *n; 587 tpath = rb_entry(pn, struct ipoib_path, rb_node); 588 589 ret = memcmp(path->pathrec.dgid.raw, tpath->pathrec.dgid.raw, 590 sizeof (union ib_gid)); 591 if (ret < 0) 592 n = &pn->rb_left; 593 else if (ret > 0) 594 n = &pn->rb_right; 595 else 596 return -EEXIST; 597 } 598 599 rb_link_node(&path->rb_node, pn, n); 600 rb_insert_color(&path->rb_node, &priv->path_tree); 601 602 list_add_tail(&path->list, &priv->path_list); 603 604 return 0; 605} 606 607static void path_free(struct net_device *dev, struct ipoib_path *path) 608{ 609 struct sk_buff *skb; 610 611 while ((skb = __skb_dequeue(&path->queue))) 612 dev_kfree_skb_irq(skb); 613 614 ipoib_dbg(ipoib_priv(dev), "%s\n", __func__); 615 616 /* remove all neigh connected to this path */ 617 ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw); 618 619 if (path->ah) 620 ipoib_put_ah(path->ah); 621 622 kfree(path); 623} 624 625#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 626 627struct ipoib_path_iter *ipoib_path_iter_init(struct net_device *dev) 628{ 629 struct ipoib_path_iter *iter; 630 631 iter = kmalloc(sizeof(*iter), GFP_KERNEL); 632 if (!iter) 633 return NULL; 634 635 iter->dev = dev; 636 memset(iter->path.pathrec.dgid.raw, 0, 16); 637 638 if (ipoib_path_iter_next(iter)) { 639 kfree(iter); 640 return NULL; 641 } 642 643 return iter; 644} 645 646int ipoib_path_iter_next(struct ipoib_path_iter *iter) 647{ 648 struct ipoib_dev_priv *priv = ipoib_priv(iter->dev); 649 struct rb_node *n; 650 struct ipoib_path *path; 651 int ret = 1; 652 653 spin_lock_irq(&priv->lock); 654 655 n = rb_first(&priv->path_tree); 656 657 while (n) { 658 path = rb_entry(n, struct ipoib_path, rb_node); 659 660 if (memcmp(iter->path.pathrec.dgid.raw, path->pathrec.dgid.raw, 661 sizeof (union ib_gid)) < 0) { 662 iter->path = *path; 663 ret = 0; 664 break; 665 } 666 667 n = rb_next(n); 668 } 669 670 spin_unlock_irq(&priv->lock); 671 672 return ret; 673} 674 675void ipoib_path_iter_read(struct ipoib_path_iter *iter, 676 struct ipoib_path *path) 677{ 678 *path = iter->path; 679} 680 681#endif /* CONFIG_INFINIBAND_IPOIB_DEBUG */ 682 683void ipoib_mark_paths_invalid(struct net_device *dev) 684{ 685 struct ipoib_dev_priv *priv = ipoib_priv(dev); 686 struct ipoib_path *path, *tp; 687 688 spin_lock_irq(&priv->lock); 689 690 list_for_each_entry_safe(path, tp, &priv->path_list, list) { 691 ipoib_dbg(priv, "mark path LID 0x%08x GID %pI6 invalid\n", 692 be32_to_cpu(sa_path_get_dlid(&path->pathrec)), 693 path->pathrec.dgid.raw); 694 if (path->ah) 695 path->ah->valid = 0; 696 } 697 698 spin_unlock_irq(&priv->lock); 699} 700 701static void push_pseudo_header(struct sk_buff *skb, const char *daddr) 702{ 703 struct ipoib_pseudo_header *phdr; 704 705 phdr = skb_push(skb, sizeof(*phdr)); 706 memcpy(phdr->hwaddr, daddr, INFINIBAND_ALEN); 707} 708 709void ipoib_flush_paths(struct net_device *dev) 710{ 711 struct ipoib_dev_priv *priv = ipoib_priv(dev); 712 struct ipoib_path *path, *tp; 713 LIST_HEAD(remove_list); 714 unsigned long flags; 715 716 netif_tx_lock_bh(dev); 717 spin_lock_irqsave(&priv->lock, flags); 718 719 list_splice_init(&priv->path_list, &remove_list); 720 721 list_for_each_entry(path, &remove_list, list) 722 rb_erase(&path->rb_node, &priv->path_tree); 723 724 list_for_each_entry_safe(path, tp, &remove_list, list) { 725 if (path->query) 726 ib_sa_cancel_query(path->query_id, path->query); 727 spin_unlock_irqrestore(&priv->lock, flags); 728 netif_tx_unlock_bh(dev); 729 wait_for_completion(&path->done); 730 path_free(dev, path); 731 netif_tx_lock_bh(dev); 732 spin_lock_irqsave(&priv->lock, flags); 733 } 734 735 spin_unlock_irqrestore(&priv->lock, flags); 736 netif_tx_unlock_bh(dev); 737} 738 739static void path_rec_completion(int status, 740 struct sa_path_rec *pathrec, 741 void *path_ptr) 742{ 743 struct ipoib_path *path = path_ptr; 744 struct net_device *dev = path->dev; 745 struct ipoib_dev_priv *priv = ipoib_priv(dev); 746 struct ipoib_ah *ah = NULL; 747 struct ipoib_ah *old_ah = NULL; 748 struct ipoib_neigh *neigh, *tn; 749 struct sk_buff_head skqueue; 750 struct sk_buff *skb; 751 unsigned long flags; 752 753 if (!status) 754 ipoib_dbg(priv, "PathRec LID 0x%04x for GID %pI6\n", 755 be32_to_cpu(sa_path_get_dlid(pathrec)), 756 pathrec->dgid.raw); 757 else 758 ipoib_dbg(priv, "PathRec status %d for GID %pI6\n", 759 status, path->pathrec.dgid.raw); 760 761 skb_queue_head_init(&skqueue); 762 763 if (!status) { 764 struct rdma_ah_attr av; 765 766 if (!ib_init_ah_attr_from_path(priv->ca, priv->port, 767 pathrec, &av, NULL)) { 768 ah = ipoib_create_ah(dev, priv->pd, &av); 769 rdma_destroy_ah_attr(&av); 770 } 771 } 772 773 spin_lock_irqsave(&priv->lock, flags); 774 775 if (!IS_ERR_OR_NULL(ah)) { 776 /* 777 * pathrec.dgid is used as the database key from the LLADDR, 778 * it must remain unchanged even if the SA returns a different 779 * GID to use in the AH. 780 */ 781 if (memcmp(pathrec->dgid.raw, path->pathrec.dgid.raw, 782 sizeof(union ib_gid))) { 783 ipoib_dbg( 784 priv, 785 "%s got PathRec for gid %pI6 while asked for %pI6\n", 786 dev->name, pathrec->dgid.raw, 787 path->pathrec.dgid.raw); 788 memcpy(pathrec->dgid.raw, path->pathrec.dgid.raw, 789 sizeof(union ib_gid)); 790 } 791 792 path->pathrec = *pathrec; 793 794 old_ah = path->ah; 795 path->ah = ah; 796 797 ipoib_dbg(priv, "created address handle %p for LID 0x%04x, SL %d\n", 798 ah, be32_to_cpu(sa_path_get_dlid(pathrec)), 799 pathrec->sl); 800 801 while ((skb = __skb_dequeue(&path->queue))) 802 __skb_queue_tail(&skqueue, skb); 803 804 list_for_each_entry_safe(neigh, tn, &path->neigh_list, list) { 805 if (neigh->ah) { 806 WARN_ON(neigh->ah != old_ah); 807 /* 808 * Dropping the ah reference inside 809 * priv->lock is safe here, because we 810 * will hold one more reference from 811 * the original value of path->ah (ie 812 * old_ah). 813 */ 814 ipoib_put_ah(neigh->ah); 815 } 816 kref_get(&path->ah->ref); 817 neigh->ah = path->ah; 818 819 if (ipoib_cm_enabled(dev, neigh->daddr)) { 820 if (!ipoib_cm_get(neigh)) 821 ipoib_cm_set(neigh, ipoib_cm_create_tx(dev, 822 path, 823 neigh)); 824 if (!ipoib_cm_get(neigh)) { 825 ipoib_neigh_free(neigh); 826 continue; 827 } 828 } 829 830 while ((skb = __skb_dequeue(&neigh->queue))) 831 __skb_queue_tail(&skqueue, skb); 832 } 833 path->ah->valid = 1; 834 } 835 836 path->query = NULL; 837 complete(&path->done); 838 839 spin_unlock_irqrestore(&priv->lock, flags); 840 841 if (IS_ERR_OR_NULL(ah)) 842 ipoib_del_neighs_by_gid(dev, path->pathrec.dgid.raw); 843 844 if (old_ah) 845 ipoib_put_ah(old_ah); 846 847 while ((skb = __skb_dequeue(&skqueue))) { 848 int ret; 849 skb->dev = dev; 850 ret = dev_queue_xmit(skb); 851 if (ret) 852 ipoib_warn(priv, "%s: dev_queue_xmit failed to re-queue packet, ret:%d\n", 853 __func__, ret); 854 } 855} 856 857static void init_path_rec(struct ipoib_dev_priv *priv, struct ipoib_path *path, 858 void *gid) 859{ 860 path->dev = priv->dev; 861 862 if (rdma_cap_opa_ah(priv->ca, priv->port)) 863 path->pathrec.rec_type = SA_PATH_REC_TYPE_OPA; 864 else 865 path->pathrec.rec_type = SA_PATH_REC_TYPE_IB; 866 867 memcpy(path->pathrec.dgid.raw, gid, sizeof(union ib_gid)); 868 path->pathrec.sgid = priv->local_gid; 869 path->pathrec.pkey = cpu_to_be16(priv->pkey); 870 path->pathrec.numb_path = 1; 871 path->pathrec.traffic_class = priv->broadcast->mcmember.traffic_class; 872} 873 874static struct ipoib_path *path_rec_create(struct net_device *dev, void *gid) 875{ 876 struct ipoib_dev_priv *priv = ipoib_priv(dev); 877 struct ipoib_path *path; 878 879 if (!priv->broadcast) 880 return NULL; 881 882 path = kzalloc(sizeof(*path), GFP_ATOMIC); 883 if (!path) 884 return NULL; 885 886 skb_queue_head_init(&path->queue); 887 888 INIT_LIST_HEAD(&path->neigh_list); 889 890 init_path_rec(priv, path, gid); 891 892 return path; 893} 894 895static int path_rec_start(struct net_device *dev, 896 struct ipoib_path *path) 897{ 898 struct ipoib_dev_priv *priv = ipoib_priv(dev); 899 900 ipoib_dbg(priv, "Start path record lookup for %pI6\n", 901 path->pathrec.dgid.raw); 902 903 init_completion(&path->done); 904 905 path->query_id = 906 ib_sa_path_rec_get(&ipoib_sa_client, priv->ca, priv->port, 907 &path->pathrec, 908 IB_SA_PATH_REC_DGID | 909 IB_SA_PATH_REC_SGID | 910 IB_SA_PATH_REC_NUMB_PATH | 911 IB_SA_PATH_REC_TRAFFIC_CLASS | 912 IB_SA_PATH_REC_PKEY, 913 1000, GFP_ATOMIC, 914 path_rec_completion, 915 path, &path->query); 916 if (path->query_id < 0) { 917 ipoib_warn(priv, "ib_sa_path_rec_get failed: %d\n", path->query_id); 918 path->query = NULL; 919 complete(&path->done); 920 return path->query_id; 921 } 922 923 return 0; 924} 925 926static void neigh_refresh_path(struct ipoib_neigh *neigh, u8 *daddr, 927 struct net_device *dev) 928{ 929 struct ipoib_dev_priv *priv = ipoib_priv(dev); 930 struct ipoib_path *path; 931 unsigned long flags; 932 933 spin_lock_irqsave(&priv->lock, flags); 934 935 path = __path_find(dev, daddr + 4); 936 if (!path) 937 goto out; 938 if (!path->query) 939 path_rec_start(dev, path); 940out: 941 spin_unlock_irqrestore(&priv->lock, flags); 942} 943 944static struct ipoib_neigh *neigh_add_path(struct sk_buff *skb, u8 *daddr, 945 struct net_device *dev) 946{ 947 struct ipoib_dev_priv *priv = ipoib_priv(dev); 948 struct rdma_netdev *rn = netdev_priv(dev); 949 struct ipoib_path *path; 950 struct ipoib_neigh *neigh; 951 unsigned long flags; 952 953 spin_lock_irqsave(&priv->lock, flags); 954 neigh = ipoib_neigh_alloc(daddr, dev); 955 if (!neigh) { 956 spin_unlock_irqrestore(&priv->lock, flags); 957 ++dev->stats.tx_dropped; 958 dev_kfree_skb_any(skb); 959 return NULL; 960 } 961 962 /* To avoid race condition, make sure that the 963 * neigh will be added only once. 964 */ 965 if (unlikely(!list_empty(&neigh->list))) { 966 spin_unlock_irqrestore(&priv->lock, flags); 967 return neigh; 968 } 969 970 path = __path_find(dev, daddr + 4); 971 if (!path) { 972 path = path_rec_create(dev, daddr + 4); 973 if (!path) 974 goto err_path; 975 976 __path_add(dev, path); 977 } 978 979 list_add_tail(&neigh->list, &path->neigh_list); 980 981 if (path->ah && path->ah->valid) { 982 kref_get(&path->ah->ref); 983 neigh->ah = path->ah; 984 985 if (ipoib_cm_enabled(dev, neigh->daddr)) { 986 if (!ipoib_cm_get(neigh)) 987 ipoib_cm_set(neigh, ipoib_cm_create_tx(dev, path, neigh)); 988 if (!ipoib_cm_get(neigh)) { 989 ipoib_neigh_free(neigh); 990 goto err_drop; 991 } 992 if (skb_queue_len(&neigh->queue) < 993 IPOIB_MAX_PATH_REC_QUEUE) { 994 push_pseudo_header(skb, neigh->daddr); 995 __skb_queue_tail(&neigh->queue, skb); 996 } else { 997 ipoib_warn(priv, "queue length limit %d. Packet drop.\n", 998 skb_queue_len(&neigh->queue)); 999 goto err_drop; 1000 } 1001 } else { 1002 spin_unlock_irqrestore(&priv->lock, flags); 1003 path->ah->last_send = rn->send(dev, skb, path->ah->ah, 1004 IPOIB_QPN(daddr)); 1005 ipoib_neigh_put(neigh); 1006 return NULL; 1007 } 1008 } else { 1009 neigh->ah = NULL; 1010 1011 if (!path->query && path_rec_start(dev, path)) 1012 goto err_path; 1013 if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) { 1014 push_pseudo_header(skb, neigh->daddr); 1015 __skb_queue_tail(&neigh->queue, skb); 1016 } else { 1017 goto err_drop; 1018 } 1019 } 1020 1021 spin_unlock_irqrestore(&priv->lock, flags); 1022 ipoib_neigh_put(neigh); 1023 return NULL; 1024 1025err_path: 1026 ipoib_neigh_free(neigh); 1027err_drop: 1028 ++dev->stats.tx_dropped; 1029 dev_kfree_skb_any(skb); 1030 1031 spin_unlock_irqrestore(&priv->lock, flags); 1032 ipoib_neigh_put(neigh); 1033 1034 return NULL; 1035} 1036 1037static void unicast_arp_send(struct sk_buff *skb, struct net_device *dev, 1038 struct ipoib_pseudo_header *phdr) 1039{ 1040 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1041 struct rdma_netdev *rn = netdev_priv(dev); 1042 struct ipoib_path *path; 1043 unsigned long flags; 1044 1045 spin_lock_irqsave(&priv->lock, flags); 1046 1047 /* no broadcast means that all paths are (going to be) not valid */ 1048 if (!priv->broadcast) 1049 goto drop_and_unlock; 1050 1051 path = __path_find(dev, phdr->hwaddr + 4); 1052 if (!path || !path->ah || !path->ah->valid) { 1053 if (!path) { 1054 path = path_rec_create(dev, phdr->hwaddr + 4); 1055 if (!path) 1056 goto drop_and_unlock; 1057 __path_add(dev, path); 1058 } else { 1059 /* 1060 * make sure there are no changes in the existing 1061 * path record 1062 */ 1063 init_path_rec(priv, path, phdr->hwaddr + 4); 1064 } 1065 if (!path->query && path_rec_start(dev, path)) { 1066 goto drop_and_unlock; 1067 } 1068 1069 if (skb_queue_len(&path->queue) < IPOIB_MAX_PATH_REC_QUEUE) { 1070 push_pseudo_header(skb, phdr->hwaddr); 1071 __skb_queue_tail(&path->queue, skb); 1072 goto unlock; 1073 } else { 1074 goto drop_and_unlock; 1075 } 1076 } 1077 1078 spin_unlock_irqrestore(&priv->lock, flags); 1079 ipoib_dbg(priv, "Send unicast ARP to %08x\n", 1080 be32_to_cpu(sa_path_get_dlid(&path->pathrec))); 1081 path->ah->last_send = rn->send(dev, skb, path->ah->ah, 1082 IPOIB_QPN(phdr->hwaddr)); 1083 return; 1084 1085drop_and_unlock: 1086 ++dev->stats.tx_dropped; 1087 dev_kfree_skb_any(skb); 1088unlock: 1089 spin_unlock_irqrestore(&priv->lock, flags); 1090} 1091 1092static netdev_tx_t ipoib_start_xmit(struct sk_buff *skb, struct net_device *dev) 1093{ 1094 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1095 struct rdma_netdev *rn = netdev_priv(dev); 1096 struct ipoib_neigh *neigh; 1097 struct ipoib_pseudo_header *phdr; 1098 struct ipoib_header *header; 1099 unsigned long flags; 1100 1101 phdr = (struct ipoib_pseudo_header *) skb->data; 1102 skb_pull(skb, sizeof(*phdr)); 1103 header = (struct ipoib_header *) skb->data; 1104 1105 if (unlikely(phdr->hwaddr[4] == 0xff)) { 1106 /* multicast, arrange "if" according to probability */ 1107 if ((header->proto != htons(ETH_P_IP)) && 1108 (header->proto != htons(ETH_P_IPV6)) && 1109 (header->proto != htons(ETH_P_ARP)) && 1110 (header->proto != htons(ETH_P_RARP)) && 1111 (header->proto != htons(ETH_P_TIPC))) { 1112 /* ethertype not supported by IPoIB */ 1113 ++dev->stats.tx_dropped; 1114 dev_kfree_skb_any(skb); 1115 return NETDEV_TX_OK; 1116 } 1117 /* Add in the P_Key for multicast*/ 1118 phdr->hwaddr[8] = (priv->pkey >> 8) & 0xff; 1119 phdr->hwaddr[9] = priv->pkey & 0xff; 1120 1121 neigh = ipoib_neigh_get(dev, phdr->hwaddr); 1122 if (likely(neigh)) 1123 goto send_using_neigh; 1124 ipoib_mcast_send(dev, phdr->hwaddr, skb); 1125 return NETDEV_TX_OK; 1126 } 1127 1128 /* unicast, arrange "switch" according to probability */ 1129 switch (header->proto) { 1130 case htons(ETH_P_IP): 1131 case htons(ETH_P_IPV6): 1132 case htons(ETH_P_TIPC): 1133 neigh = ipoib_neigh_get(dev, phdr->hwaddr); 1134 if (unlikely(!neigh)) { 1135 neigh = neigh_add_path(skb, phdr->hwaddr, dev); 1136 if (likely(!neigh)) 1137 return NETDEV_TX_OK; 1138 } 1139 break; 1140 case htons(ETH_P_ARP): 1141 case htons(ETH_P_RARP): 1142 /* for unicast ARP and RARP should always perform path find */ 1143 unicast_arp_send(skb, dev, phdr); 1144 return NETDEV_TX_OK; 1145 default: 1146 /* ethertype not supported by IPoIB */ 1147 ++dev->stats.tx_dropped; 1148 dev_kfree_skb_any(skb); 1149 return NETDEV_TX_OK; 1150 } 1151 1152send_using_neigh: 1153 /* note we now hold a ref to neigh */ 1154 if (ipoib_cm_get(neigh)) { 1155 if (ipoib_cm_up(neigh)) { 1156 ipoib_cm_send(dev, skb, ipoib_cm_get(neigh)); 1157 goto unref; 1158 } 1159 } else if (neigh->ah && neigh->ah->valid) { 1160 neigh->ah->last_send = rn->send(dev, skb, neigh->ah->ah, 1161 IPOIB_QPN(phdr->hwaddr)); 1162 goto unref; 1163 } else if (neigh->ah) { 1164 neigh_refresh_path(neigh, phdr->hwaddr, dev); 1165 } 1166 1167 if (skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE) { 1168 push_pseudo_header(skb, phdr->hwaddr); 1169 spin_lock_irqsave(&priv->lock, flags); 1170 __skb_queue_tail(&neigh->queue, skb); 1171 spin_unlock_irqrestore(&priv->lock, flags); 1172 } else { 1173 ++dev->stats.tx_dropped; 1174 dev_kfree_skb_any(skb); 1175 } 1176 1177unref: 1178 ipoib_neigh_put(neigh); 1179 1180 return NETDEV_TX_OK; 1181} 1182 1183static void ipoib_timeout(struct net_device *dev, unsigned int txqueue) 1184{ 1185 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1186 1187 ipoib_warn(priv, "transmit timeout: latency %d msecs\n", 1188 jiffies_to_msecs(jiffies - dev_trans_start(dev))); 1189 ipoib_warn(priv, 1190 "queue stopped %d, tx_head %u, tx_tail %u, global_tx_head %u, global_tx_tail %u\n", 1191 netif_queue_stopped(dev), priv->tx_head, priv->tx_tail, 1192 priv->global_tx_head, priv->global_tx_tail); 1193 1194 /* XXX reset QP, etc. */ 1195} 1196 1197static int ipoib_hard_header(struct sk_buff *skb, 1198 struct net_device *dev, 1199 unsigned short type, 1200 const void *daddr, 1201 const void *saddr, 1202 unsigned int len) 1203{ 1204 struct ipoib_header *header; 1205 1206 header = skb_push(skb, sizeof(*header)); 1207 1208 header->proto = htons(type); 1209 header->reserved = 0; 1210 1211 /* 1212 * we don't rely on dst_entry structure, always stuff the 1213 * destination address into skb hard header so we can figure out where 1214 * to send the packet later. 1215 */ 1216 push_pseudo_header(skb, daddr); 1217 1218 return IPOIB_HARD_LEN; 1219} 1220 1221static void ipoib_set_mcast_list(struct net_device *dev) 1222{ 1223 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1224 1225 if (!test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)) { 1226 ipoib_dbg(priv, "IPOIB_FLAG_OPER_UP not set"); 1227 return; 1228 } 1229 1230 queue_work(priv->wq, &priv->restart_task); 1231} 1232 1233static int ipoib_get_iflink(const struct net_device *dev) 1234{ 1235 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1236 1237 /* parent interface */ 1238 if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) 1239 return dev->ifindex; 1240 1241 /* child/vlan interface */ 1242 return priv->parent->ifindex; 1243} 1244 1245static u32 ipoib_addr_hash(struct ipoib_neigh_hash *htbl, u8 *daddr) 1246{ 1247 /* 1248 * Use only the address parts that contributes to spreading 1249 * The subnet prefix is not used as one can not connect to 1250 * same remote port (GUID) using the same remote QPN via two 1251 * different subnets. 1252 */ 1253 /* qpn octets[1:4) & port GUID octets[12:20) */ 1254 u32 *d32 = (u32 *) daddr; 1255 u32 hv; 1256 1257 hv = jhash_3words(d32[3], d32[4], IPOIB_QPN_MASK & d32[0], 0); 1258 return hv & htbl->mask; 1259} 1260 1261struct ipoib_neigh *ipoib_neigh_get(struct net_device *dev, u8 *daddr) 1262{ 1263 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1264 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1265 struct ipoib_neigh_hash *htbl; 1266 struct ipoib_neigh *neigh = NULL; 1267 u32 hash_val; 1268 1269 rcu_read_lock_bh(); 1270 1271 htbl = rcu_dereference_bh(ntbl->htbl); 1272 1273 if (!htbl) 1274 goto out_unlock; 1275 1276 hash_val = ipoib_addr_hash(htbl, daddr); 1277 for (neigh = rcu_dereference_bh(htbl->buckets[hash_val]); 1278 neigh != NULL; 1279 neigh = rcu_dereference_bh(neigh->hnext)) { 1280 if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) { 1281 /* found, take one ref on behalf of the caller */ 1282 if (!atomic_inc_not_zero(&neigh->refcnt)) { 1283 /* deleted */ 1284 neigh = NULL; 1285 goto out_unlock; 1286 } 1287 1288 if (likely(skb_queue_len(&neigh->queue) < IPOIB_MAX_PATH_REC_QUEUE)) 1289 neigh->alive = jiffies; 1290 goto out_unlock; 1291 } 1292 } 1293 1294out_unlock: 1295 rcu_read_unlock_bh(); 1296 return neigh; 1297} 1298 1299static void __ipoib_reap_neigh(struct ipoib_dev_priv *priv) 1300{ 1301 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1302 struct ipoib_neigh_hash *htbl; 1303 unsigned long neigh_obsolete; 1304 unsigned long dt; 1305 unsigned long flags; 1306 int i; 1307 LIST_HEAD(remove_list); 1308 1309 spin_lock_irqsave(&priv->lock, flags); 1310 1311 htbl = rcu_dereference_protected(ntbl->htbl, 1312 lockdep_is_held(&priv->lock)); 1313 1314 if (!htbl) 1315 goto out_unlock; 1316 1317 /* neigh is obsolete if it was idle for two GC periods */ 1318 dt = 2 * arp_tbl.gc_interval; 1319 neigh_obsolete = jiffies - dt; 1320 1321 for (i = 0; i < htbl->size; i++) { 1322 struct ipoib_neigh *neigh; 1323 struct ipoib_neigh __rcu **np = &htbl->buckets[i]; 1324 1325 while ((neigh = rcu_dereference_protected(*np, 1326 lockdep_is_held(&priv->lock))) != NULL) { 1327 /* was the neigh idle for two GC periods */ 1328 if (time_after(neigh_obsolete, neigh->alive)) { 1329 1330 ipoib_check_and_add_mcast_sendonly(priv, neigh->daddr + 4, &remove_list); 1331 1332 rcu_assign_pointer(*np, 1333 rcu_dereference_protected(neigh->hnext, 1334 lockdep_is_held(&priv->lock))); 1335 /* remove from path/mc list */ 1336 list_del_init(&neigh->list); 1337 call_rcu(&neigh->rcu, ipoib_neigh_reclaim); 1338 } else { 1339 np = &neigh->hnext; 1340 } 1341 1342 } 1343 } 1344 1345out_unlock: 1346 spin_unlock_irqrestore(&priv->lock, flags); 1347 ipoib_mcast_remove_list(&remove_list); 1348} 1349 1350static void ipoib_reap_neigh(struct work_struct *work) 1351{ 1352 struct ipoib_dev_priv *priv = 1353 container_of(work, struct ipoib_dev_priv, neigh_reap_task.work); 1354 1355 __ipoib_reap_neigh(priv); 1356 1357 queue_delayed_work(priv->wq, &priv->neigh_reap_task, 1358 arp_tbl.gc_interval); 1359} 1360 1361 1362static struct ipoib_neigh *ipoib_neigh_ctor(u8 *daddr, 1363 struct net_device *dev) 1364{ 1365 struct ipoib_neigh *neigh; 1366 1367 neigh = kzalloc(sizeof(*neigh), GFP_ATOMIC); 1368 if (!neigh) 1369 return NULL; 1370 1371 neigh->dev = dev; 1372 memcpy(&neigh->daddr, daddr, sizeof(neigh->daddr)); 1373 skb_queue_head_init(&neigh->queue); 1374 INIT_LIST_HEAD(&neigh->list); 1375 ipoib_cm_set(neigh, NULL); 1376 /* one ref on behalf of the caller */ 1377 atomic_set(&neigh->refcnt, 1); 1378 1379 return neigh; 1380} 1381 1382struct ipoib_neigh *ipoib_neigh_alloc(u8 *daddr, 1383 struct net_device *dev) 1384{ 1385 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1386 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1387 struct ipoib_neigh_hash *htbl; 1388 struct ipoib_neigh *neigh; 1389 u32 hash_val; 1390 1391 htbl = rcu_dereference_protected(ntbl->htbl, 1392 lockdep_is_held(&priv->lock)); 1393 if (!htbl) { 1394 neigh = NULL; 1395 goto out_unlock; 1396 } 1397 1398 /* need to add a new neigh, but maybe some other thread succeeded? 1399 * recalc hash, maybe hash resize took place so we do a search 1400 */ 1401 hash_val = ipoib_addr_hash(htbl, daddr); 1402 for (neigh = rcu_dereference_protected(htbl->buckets[hash_val], 1403 lockdep_is_held(&priv->lock)); 1404 neigh != NULL; 1405 neigh = rcu_dereference_protected(neigh->hnext, 1406 lockdep_is_held(&priv->lock))) { 1407 if (memcmp(daddr, neigh->daddr, INFINIBAND_ALEN) == 0) { 1408 /* found, take one ref on behalf of the caller */ 1409 if (!atomic_inc_not_zero(&neigh->refcnt)) { 1410 /* deleted */ 1411 neigh = NULL; 1412 break; 1413 } 1414 neigh->alive = jiffies; 1415 goto out_unlock; 1416 } 1417 } 1418 1419 neigh = ipoib_neigh_ctor(daddr, dev); 1420 if (!neigh) 1421 goto out_unlock; 1422 1423 /* one ref on behalf of the hash table */ 1424 atomic_inc(&neigh->refcnt); 1425 neigh->alive = jiffies; 1426 /* put in hash */ 1427 rcu_assign_pointer(neigh->hnext, 1428 rcu_dereference_protected(htbl->buckets[hash_val], 1429 lockdep_is_held(&priv->lock))); 1430 rcu_assign_pointer(htbl->buckets[hash_val], neigh); 1431 atomic_inc(&ntbl->entries); 1432 1433out_unlock: 1434 1435 return neigh; 1436} 1437 1438void ipoib_neigh_dtor(struct ipoib_neigh *neigh) 1439{ 1440 /* neigh reference count was dropprd to zero */ 1441 struct net_device *dev = neigh->dev; 1442 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1443 struct sk_buff *skb; 1444 if (neigh->ah) 1445 ipoib_put_ah(neigh->ah); 1446 while ((skb = __skb_dequeue(&neigh->queue))) { 1447 ++dev->stats.tx_dropped; 1448 dev_kfree_skb_any(skb); 1449 } 1450 if (ipoib_cm_get(neigh)) 1451 ipoib_cm_destroy_tx(ipoib_cm_get(neigh)); 1452 ipoib_dbg(ipoib_priv(dev), 1453 "neigh free for %06x %pI6\n", 1454 IPOIB_QPN(neigh->daddr), 1455 neigh->daddr + 4); 1456 kfree(neigh); 1457 if (atomic_dec_and_test(&priv->ntbl.entries)) { 1458 if (test_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags)) 1459 complete(&priv->ntbl.flushed); 1460 } 1461} 1462 1463static void ipoib_neigh_reclaim(struct rcu_head *rp) 1464{ 1465 /* Called as a result of removal from hash table */ 1466 struct ipoib_neigh *neigh = container_of(rp, struct ipoib_neigh, rcu); 1467 /* note TX context may hold another ref */ 1468 ipoib_neigh_put(neigh); 1469} 1470 1471void ipoib_neigh_free(struct ipoib_neigh *neigh) 1472{ 1473 struct net_device *dev = neigh->dev; 1474 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1475 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1476 struct ipoib_neigh_hash *htbl; 1477 struct ipoib_neigh __rcu **np; 1478 struct ipoib_neigh *n; 1479 u32 hash_val; 1480 1481 htbl = rcu_dereference_protected(ntbl->htbl, 1482 lockdep_is_held(&priv->lock)); 1483 if (!htbl) 1484 return; 1485 1486 hash_val = ipoib_addr_hash(htbl, neigh->daddr); 1487 np = &htbl->buckets[hash_val]; 1488 for (n = rcu_dereference_protected(*np, 1489 lockdep_is_held(&priv->lock)); 1490 n != NULL; 1491 n = rcu_dereference_protected(*np, 1492 lockdep_is_held(&priv->lock))) { 1493 if (n == neigh) { 1494 /* found */ 1495 rcu_assign_pointer(*np, 1496 rcu_dereference_protected(neigh->hnext, 1497 lockdep_is_held(&priv->lock))); 1498 /* remove from parent list */ 1499 list_del_init(&neigh->list); 1500 call_rcu(&neigh->rcu, ipoib_neigh_reclaim); 1501 return; 1502 } else { 1503 np = &n->hnext; 1504 } 1505 } 1506} 1507 1508static int ipoib_neigh_hash_init(struct ipoib_dev_priv *priv) 1509{ 1510 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1511 struct ipoib_neigh_hash *htbl; 1512 struct ipoib_neigh __rcu **buckets; 1513 u32 size; 1514 1515 clear_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags); 1516 ntbl->htbl = NULL; 1517 htbl = kzalloc(sizeof(*htbl), GFP_KERNEL); 1518 if (!htbl) 1519 return -ENOMEM; 1520 size = roundup_pow_of_two(arp_tbl.gc_thresh3); 1521 buckets = kvcalloc(size, sizeof(*buckets), GFP_KERNEL); 1522 if (!buckets) { 1523 kfree(htbl); 1524 return -ENOMEM; 1525 } 1526 htbl->size = size; 1527 htbl->mask = (size - 1); 1528 htbl->buckets = buckets; 1529 RCU_INIT_POINTER(ntbl->htbl, htbl); 1530 htbl->ntbl = ntbl; 1531 atomic_set(&ntbl->entries, 0); 1532 1533 /* start garbage collection */ 1534 queue_delayed_work(priv->wq, &priv->neigh_reap_task, 1535 arp_tbl.gc_interval); 1536 1537 return 0; 1538} 1539 1540static void neigh_hash_free_rcu(struct rcu_head *head) 1541{ 1542 struct ipoib_neigh_hash *htbl = container_of(head, 1543 struct ipoib_neigh_hash, 1544 rcu); 1545 struct ipoib_neigh __rcu **buckets = htbl->buckets; 1546 struct ipoib_neigh_table *ntbl = htbl->ntbl; 1547 1548 kvfree(buckets); 1549 kfree(htbl); 1550 complete(&ntbl->deleted); 1551} 1552 1553void ipoib_del_neighs_by_gid(struct net_device *dev, u8 *gid) 1554{ 1555 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1556 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1557 struct ipoib_neigh_hash *htbl; 1558 unsigned long flags; 1559 int i; 1560 1561 /* remove all neigh connected to a given path or mcast */ 1562 spin_lock_irqsave(&priv->lock, flags); 1563 1564 htbl = rcu_dereference_protected(ntbl->htbl, 1565 lockdep_is_held(&priv->lock)); 1566 1567 if (!htbl) 1568 goto out_unlock; 1569 1570 for (i = 0; i < htbl->size; i++) { 1571 struct ipoib_neigh *neigh; 1572 struct ipoib_neigh __rcu **np = &htbl->buckets[i]; 1573 1574 while ((neigh = rcu_dereference_protected(*np, 1575 lockdep_is_held(&priv->lock))) != NULL) { 1576 /* delete neighs belong to this parent */ 1577 if (!memcmp(gid, neigh->daddr + 4, sizeof (union ib_gid))) { 1578 rcu_assign_pointer(*np, 1579 rcu_dereference_protected(neigh->hnext, 1580 lockdep_is_held(&priv->lock))); 1581 /* remove from parent list */ 1582 list_del_init(&neigh->list); 1583 call_rcu(&neigh->rcu, ipoib_neigh_reclaim); 1584 } else { 1585 np = &neigh->hnext; 1586 } 1587 1588 } 1589 } 1590out_unlock: 1591 spin_unlock_irqrestore(&priv->lock, flags); 1592} 1593 1594static void ipoib_flush_neighs(struct ipoib_dev_priv *priv) 1595{ 1596 struct ipoib_neigh_table *ntbl = &priv->ntbl; 1597 struct ipoib_neigh_hash *htbl; 1598 unsigned long flags; 1599 int i, wait_flushed = 0; 1600 1601 init_completion(&priv->ntbl.flushed); 1602 set_bit(IPOIB_NEIGH_TBL_FLUSH, &priv->flags); 1603 1604 spin_lock_irqsave(&priv->lock, flags); 1605 1606 htbl = rcu_dereference_protected(ntbl->htbl, 1607 lockdep_is_held(&priv->lock)); 1608 if (!htbl) 1609 goto out_unlock; 1610 1611 wait_flushed = atomic_read(&priv->ntbl.entries); 1612 if (!wait_flushed) 1613 goto free_htbl; 1614 1615 for (i = 0; i < htbl->size; i++) { 1616 struct ipoib_neigh *neigh; 1617 struct ipoib_neigh __rcu **np = &htbl->buckets[i]; 1618 1619 while ((neigh = rcu_dereference_protected(*np, 1620 lockdep_is_held(&priv->lock))) != NULL) { 1621 rcu_assign_pointer(*np, 1622 rcu_dereference_protected(neigh->hnext, 1623 lockdep_is_held(&priv->lock))); 1624 /* remove from path/mc list */ 1625 list_del_init(&neigh->list); 1626 call_rcu(&neigh->rcu, ipoib_neigh_reclaim); 1627 } 1628 } 1629 1630free_htbl: 1631 rcu_assign_pointer(ntbl->htbl, NULL); 1632 call_rcu(&htbl->rcu, neigh_hash_free_rcu); 1633 1634out_unlock: 1635 spin_unlock_irqrestore(&priv->lock, flags); 1636 if (wait_flushed) 1637 wait_for_completion(&priv->ntbl.flushed); 1638} 1639 1640static void ipoib_neigh_hash_uninit(struct net_device *dev) 1641{ 1642 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1643 1644 ipoib_dbg(priv, "%s\n", __func__); 1645 init_completion(&priv->ntbl.deleted); 1646 1647 cancel_delayed_work_sync(&priv->neigh_reap_task); 1648 1649 ipoib_flush_neighs(priv); 1650 1651 wait_for_completion(&priv->ntbl.deleted); 1652} 1653 1654static void ipoib_napi_add(struct net_device *dev) 1655{ 1656 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1657 1658 netif_napi_add(dev, &priv->recv_napi, ipoib_rx_poll, IPOIB_NUM_WC); 1659 netif_napi_add(dev, &priv->send_napi, ipoib_tx_poll, MAX_SEND_CQE); 1660} 1661 1662static void ipoib_napi_del(struct net_device *dev) 1663{ 1664 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1665 1666 netif_napi_del(&priv->recv_napi); 1667 netif_napi_del(&priv->send_napi); 1668} 1669 1670static void ipoib_dev_uninit_default(struct net_device *dev) 1671{ 1672 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1673 1674 ipoib_transport_dev_cleanup(dev); 1675 1676 ipoib_napi_del(dev); 1677 1678 ipoib_cm_dev_cleanup(dev); 1679 1680 kfree(priv->rx_ring); 1681 vfree(priv->tx_ring); 1682 1683 priv->rx_ring = NULL; 1684 priv->tx_ring = NULL; 1685} 1686 1687static int ipoib_dev_init_default(struct net_device *dev) 1688{ 1689 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1690 1691 ipoib_napi_add(dev); 1692 1693 /* Allocate RX/TX "rings" to hold queued skbs */ 1694 priv->rx_ring = kcalloc(ipoib_recvq_size, 1695 sizeof(*priv->rx_ring), 1696 GFP_KERNEL); 1697 if (!priv->rx_ring) 1698 goto out; 1699 1700 priv->tx_ring = vzalloc(array_size(ipoib_sendq_size, 1701 sizeof(*priv->tx_ring))); 1702 if (!priv->tx_ring) { 1703 pr_warn("%s: failed to allocate TX ring (%d entries)\n", 1704 priv->ca->name, ipoib_sendq_size); 1705 goto out_rx_ring_cleanup; 1706 } 1707 1708 /* priv->tx_head, tx_tail and global_tx_tail/head are already 0 */ 1709 1710 if (ipoib_transport_dev_init(dev, priv->ca)) { 1711 pr_warn("%s: ipoib_transport_dev_init failed\n", 1712 priv->ca->name); 1713 goto out_tx_ring_cleanup; 1714 } 1715 1716 /* after qp created set dev address */ 1717 priv->dev->dev_addr[1] = (priv->qp->qp_num >> 16) & 0xff; 1718 priv->dev->dev_addr[2] = (priv->qp->qp_num >> 8) & 0xff; 1719 priv->dev->dev_addr[3] = (priv->qp->qp_num) & 0xff; 1720 1721 return 0; 1722 1723out_tx_ring_cleanup: 1724 vfree(priv->tx_ring); 1725 1726out_rx_ring_cleanup: 1727 kfree(priv->rx_ring); 1728 1729out: 1730 ipoib_napi_del(dev); 1731 return -ENOMEM; 1732} 1733 1734static int ipoib_ioctl(struct net_device *dev, struct ifreq *ifr, 1735 int cmd) 1736{ 1737 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1738 1739 if (!priv->rn_ops->ndo_do_ioctl) 1740 return -EOPNOTSUPP; 1741 1742 return priv->rn_ops->ndo_do_ioctl(dev, ifr, cmd); 1743} 1744 1745static int ipoib_dev_init(struct net_device *dev) 1746{ 1747 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1748 int ret = -ENOMEM; 1749 1750 priv->qp = NULL; 1751 1752 /* 1753 * the various IPoIB tasks assume they will never race against 1754 * themselves, so always use a single thread workqueue 1755 */ 1756 priv->wq = alloc_ordered_workqueue("ipoib_wq", WQ_MEM_RECLAIM); 1757 if (!priv->wq) { 1758 pr_warn("%s: failed to allocate device WQ\n", dev->name); 1759 goto out; 1760 } 1761 1762 /* create pd, which used both for control and datapath*/ 1763 priv->pd = ib_alloc_pd(priv->ca, 0); 1764 if (IS_ERR(priv->pd)) { 1765 pr_warn("%s: failed to allocate PD\n", priv->ca->name); 1766 goto clean_wq; 1767 } 1768 1769 ret = priv->rn_ops->ndo_init(dev); 1770 if (ret) { 1771 pr_warn("%s failed to init HW resource\n", dev->name); 1772 goto out_free_pd; 1773 } 1774 1775 ret = ipoib_neigh_hash_init(priv); 1776 if (ret) { 1777 pr_warn("%s failed to init neigh hash\n", dev->name); 1778 goto out_dev_uninit; 1779 } 1780 1781 if (dev->flags & IFF_UP) { 1782 if (ipoib_ib_dev_open(dev)) { 1783 pr_warn("%s failed to open device\n", dev->name); 1784 ret = -ENODEV; 1785 goto out_hash_uninit; 1786 } 1787 } 1788 1789 return 0; 1790 1791out_hash_uninit: 1792 ipoib_neigh_hash_uninit(dev); 1793 1794out_dev_uninit: 1795 ipoib_ib_dev_cleanup(dev); 1796 1797out_free_pd: 1798 if (priv->pd) { 1799 ib_dealloc_pd(priv->pd); 1800 priv->pd = NULL; 1801 } 1802 1803clean_wq: 1804 if (priv->wq) { 1805 destroy_workqueue(priv->wq); 1806 priv->wq = NULL; 1807 } 1808 1809out: 1810 return ret; 1811} 1812 1813/* 1814 * This must be called before doing an unregister_netdev on a parent device to 1815 * shutdown the IB event handler. 1816 */ 1817static void ipoib_parent_unregister_pre(struct net_device *ndev) 1818{ 1819 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 1820 1821 /* 1822 * ipoib_set_mac checks netif_running before pushing work, clearing 1823 * running ensures the it will not add more work. 1824 */ 1825 rtnl_lock(); 1826 dev_change_flags(priv->dev, priv->dev->flags & ~IFF_UP, NULL); 1827 rtnl_unlock(); 1828 1829 /* ipoib_event() cannot be running once this returns */ 1830 ib_unregister_event_handler(&priv->event_handler); 1831 1832 /* 1833 * Work on the queue grabs the rtnl lock, so this cannot be done while 1834 * also holding it. 1835 */ 1836 flush_workqueue(ipoib_workqueue); 1837} 1838 1839static void ipoib_set_dev_features(struct ipoib_dev_priv *priv) 1840{ 1841 priv->hca_caps = priv->ca->attrs.device_cap_flags; 1842 1843 if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) { 1844 priv->dev->hw_features |= NETIF_F_IP_CSUM | NETIF_F_RXCSUM; 1845 1846 if (priv->hca_caps & IB_DEVICE_UD_TSO) 1847 priv->dev->hw_features |= NETIF_F_TSO; 1848 1849 priv->dev->features |= priv->dev->hw_features; 1850 } 1851} 1852 1853static int ipoib_parent_init(struct net_device *ndev) 1854{ 1855 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 1856 struct ib_port_attr attr; 1857 int result; 1858 1859 result = ib_query_port(priv->ca, priv->port, &attr); 1860 if (result) { 1861 pr_warn("%s: ib_query_port %d failed\n", priv->ca->name, 1862 priv->port); 1863 return result; 1864 } 1865 priv->max_ib_mtu = rdma_mtu_from_attr(priv->ca, priv->port, &attr); 1866 1867 result = ib_query_pkey(priv->ca, priv->port, 0, &priv->pkey); 1868 if (result) { 1869 pr_warn("%s: ib_query_pkey port %d failed (ret = %d)\n", 1870 priv->ca->name, priv->port, result); 1871 return result; 1872 } 1873 1874 result = rdma_query_gid(priv->ca, priv->port, 0, &priv->local_gid); 1875 if (result) { 1876 pr_warn("%s: rdma_query_gid port %d failed (ret = %d)\n", 1877 priv->ca->name, priv->port, result); 1878 return result; 1879 } 1880 memcpy(priv->dev->dev_addr + 4, priv->local_gid.raw, 1881 sizeof(union ib_gid)); 1882 1883 SET_NETDEV_DEV(priv->dev, priv->ca->dev.parent); 1884 priv->dev->dev_port = priv->port - 1; 1885 /* Let's set this one too for backwards compatibility. */ 1886 priv->dev->dev_id = priv->port - 1; 1887 1888 return 0; 1889} 1890 1891static void ipoib_child_init(struct net_device *ndev) 1892{ 1893 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 1894 struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent); 1895 1896 priv->max_ib_mtu = ppriv->max_ib_mtu; 1897 set_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags); 1898 if (memchr_inv(priv->dev->dev_addr, 0, INFINIBAND_ALEN)) 1899 memcpy(&priv->local_gid, priv->dev->dev_addr + 4, 1900 sizeof(priv->local_gid)); 1901 else { 1902 memcpy(priv->dev->dev_addr, ppriv->dev->dev_addr, 1903 INFINIBAND_ALEN); 1904 memcpy(&priv->local_gid, &ppriv->local_gid, 1905 sizeof(priv->local_gid)); 1906 } 1907} 1908 1909static int ipoib_ndo_init(struct net_device *ndev) 1910{ 1911 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 1912 int rc; 1913 struct rdma_netdev *rn = netdev_priv(ndev); 1914 1915 if (priv->parent) { 1916 ipoib_child_init(ndev); 1917 } else { 1918 rc = ipoib_parent_init(ndev); 1919 if (rc) 1920 return rc; 1921 } 1922 1923 /* MTU will be reset when mcast join happens */ 1924 ndev->mtu = IPOIB_UD_MTU(priv->max_ib_mtu); 1925 priv->mcast_mtu = priv->admin_mtu = ndev->mtu; 1926 rn->mtu = priv->mcast_mtu; 1927 ndev->max_mtu = IPOIB_CM_MTU; 1928 1929 ndev->neigh_priv_len = sizeof(struct ipoib_neigh); 1930 1931 /* 1932 * Set the full membership bit, so that we join the right 1933 * broadcast group, etc. 1934 */ 1935 priv->pkey |= 0x8000; 1936 1937 ndev->broadcast[8] = priv->pkey >> 8; 1938 ndev->broadcast[9] = priv->pkey & 0xff; 1939 set_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags); 1940 1941 ipoib_set_dev_features(priv); 1942 1943 rc = ipoib_dev_init(ndev); 1944 if (rc) { 1945 pr_warn("%s: failed to initialize device: %s port %d (ret = %d)\n", 1946 priv->ca->name, priv->dev->name, priv->port, rc); 1947 return rc; 1948 } 1949 1950 if (priv->parent) { 1951 struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent); 1952 1953 dev_hold(priv->parent); 1954 1955 down_write(&ppriv->vlan_rwsem); 1956 list_add_tail(&priv->list, &ppriv->child_intfs); 1957 up_write(&ppriv->vlan_rwsem); 1958 } 1959 1960 return 0; 1961} 1962 1963static void ipoib_ndo_uninit(struct net_device *dev) 1964{ 1965 struct ipoib_dev_priv *priv = ipoib_priv(dev); 1966 1967 ASSERT_RTNL(); 1968 1969 /* 1970 * ipoib_remove_one guarantees the children are removed before the 1971 * parent, and that is the only place where a parent can be removed. 1972 */ 1973 WARN_ON(!list_empty(&priv->child_intfs)); 1974 1975 if (priv->parent) { 1976 struct ipoib_dev_priv *ppriv = ipoib_priv(priv->parent); 1977 1978 down_write(&ppriv->vlan_rwsem); 1979 list_del(&priv->list); 1980 up_write(&ppriv->vlan_rwsem); 1981 } 1982 1983 ipoib_neigh_hash_uninit(dev); 1984 1985 ipoib_ib_dev_cleanup(dev); 1986 1987 /* no more works over the priv->wq */ 1988 if (priv->wq) { 1989 /* See ipoib_mcast_carrier_on_task() */ 1990 WARN_ON(test_bit(IPOIB_FLAG_OPER_UP, &priv->flags)); 1991 flush_workqueue(priv->wq); 1992 destroy_workqueue(priv->wq); 1993 priv->wq = NULL; 1994 } 1995 1996 if (priv->parent) 1997 dev_put(priv->parent); 1998} 1999 2000static int ipoib_set_vf_link_state(struct net_device *dev, int vf, int link_state) 2001{ 2002 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2003 2004 return ib_set_vf_link_state(priv->ca, vf, priv->port, link_state); 2005} 2006 2007static int ipoib_get_vf_config(struct net_device *dev, int vf, 2008 struct ifla_vf_info *ivf) 2009{ 2010 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2011 int err; 2012 2013 err = ib_get_vf_config(priv->ca, vf, priv->port, ivf); 2014 if (err) 2015 return err; 2016 2017 ivf->vf = vf; 2018 memcpy(ivf->mac, dev->dev_addr, dev->addr_len); 2019 2020 return 0; 2021} 2022 2023static int ipoib_set_vf_guid(struct net_device *dev, int vf, u64 guid, int type) 2024{ 2025 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2026 2027 if (type != IFLA_VF_IB_NODE_GUID && type != IFLA_VF_IB_PORT_GUID) 2028 return -EINVAL; 2029 2030 return ib_set_vf_guid(priv->ca, vf, priv->port, guid, type); 2031} 2032 2033static int ipoib_get_vf_guid(struct net_device *dev, int vf, 2034 struct ifla_vf_guid *node_guid, 2035 struct ifla_vf_guid *port_guid) 2036{ 2037 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2038 2039 return ib_get_vf_guid(priv->ca, vf, priv->port, node_guid, port_guid); 2040} 2041 2042static int ipoib_get_vf_stats(struct net_device *dev, int vf, 2043 struct ifla_vf_stats *vf_stats) 2044{ 2045 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2046 2047 return ib_get_vf_stats(priv->ca, vf, priv->port, vf_stats); 2048} 2049 2050static const struct header_ops ipoib_header_ops = { 2051 .create = ipoib_hard_header, 2052}; 2053 2054static const struct net_device_ops ipoib_netdev_ops_pf = { 2055 .ndo_init = ipoib_ndo_init, 2056 .ndo_uninit = ipoib_ndo_uninit, 2057 .ndo_open = ipoib_open, 2058 .ndo_stop = ipoib_stop, 2059 .ndo_change_mtu = ipoib_change_mtu, 2060 .ndo_fix_features = ipoib_fix_features, 2061 .ndo_start_xmit = ipoib_start_xmit, 2062 .ndo_tx_timeout = ipoib_timeout, 2063 .ndo_set_rx_mode = ipoib_set_mcast_list, 2064 .ndo_get_iflink = ipoib_get_iflink, 2065 .ndo_set_vf_link_state = ipoib_set_vf_link_state, 2066 .ndo_get_vf_config = ipoib_get_vf_config, 2067 .ndo_get_vf_stats = ipoib_get_vf_stats, 2068 .ndo_get_vf_guid = ipoib_get_vf_guid, 2069 .ndo_set_vf_guid = ipoib_set_vf_guid, 2070 .ndo_set_mac_address = ipoib_set_mac, 2071 .ndo_get_stats64 = ipoib_get_stats, 2072 .ndo_do_ioctl = ipoib_ioctl, 2073}; 2074 2075static const struct net_device_ops ipoib_netdev_ops_vf = { 2076 .ndo_init = ipoib_ndo_init, 2077 .ndo_uninit = ipoib_ndo_uninit, 2078 .ndo_open = ipoib_open, 2079 .ndo_stop = ipoib_stop, 2080 .ndo_change_mtu = ipoib_change_mtu, 2081 .ndo_fix_features = ipoib_fix_features, 2082 .ndo_start_xmit = ipoib_start_xmit, 2083 .ndo_tx_timeout = ipoib_timeout, 2084 .ndo_set_rx_mode = ipoib_set_mcast_list, 2085 .ndo_get_iflink = ipoib_get_iflink, 2086 .ndo_get_stats64 = ipoib_get_stats, 2087 .ndo_do_ioctl = ipoib_ioctl, 2088}; 2089 2090static const struct net_device_ops ipoib_netdev_default_pf = { 2091 .ndo_init = ipoib_dev_init_default, 2092 .ndo_uninit = ipoib_dev_uninit_default, 2093 .ndo_open = ipoib_ib_dev_open_default, 2094 .ndo_stop = ipoib_ib_dev_stop_default, 2095}; 2096 2097void ipoib_setup_common(struct net_device *dev) 2098{ 2099 dev->header_ops = &ipoib_header_ops; 2100 dev->netdev_ops = &ipoib_netdev_default_pf; 2101 2102 ipoib_set_ethtool_ops(dev); 2103 2104 dev->watchdog_timeo = HZ; 2105 2106 dev->flags |= IFF_BROADCAST | IFF_MULTICAST; 2107 2108 dev->hard_header_len = IPOIB_HARD_LEN; 2109 dev->addr_len = INFINIBAND_ALEN; 2110 dev->type = ARPHRD_INFINIBAND; 2111 dev->tx_queue_len = ipoib_sendq_size * 2; 2112 dev->features = (NETIF_F_VLAN_CHALLENGED | 2113 NETIF_F_HIGHDMA); 2114 netif_keep_dst(dev); 2115 2116 memcpy(dev->broadcast, ipv4_bcast_addr, INFINIBAND_ALEN); 2117 2118 /* 2119 * unregister_netdev always frees the netdev, we use this mode 2120 * consistently to unify all the various unregister paths, including 2121 * those connected to rtnl_link_ops which require it. 2122 */ 2123 dev->needs_free_netdev = true; 2124} 2125 2126static void ipoib_build_priv(struct net_device *dev) 2127{ 2128 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2129 2130 priv->dev = dev; 2131 spin_lock_init(&priv->lock); 2132 init_rwsem(&priv->vlan_rwsem); 2133 mutex_init(&priv->mcast_mutex); 2134 2135 INIT_LIST_HEAD(&priv->path_list); 2136 INIT_LIST_HEAD(&priv->child_intfs); 2137 INIT_LIST_HEAD(&priv->dead_ahs); 2138 INIT_LIST_HEAD(&priv->multicast_list); 2139 2140 INIT_DELAYED_WORK(&priv->mcast_task, ipoib_mcast_join_task); 2141 INIT_WORK(&priv->carrier_on_task, ipoib_mcast_carrier_on_task); 2142 INIT_WORK(&priv->flush_light, ipoib_ib_dev_flush_light); 2143 INIT_WORK(&priv->flush_normal, ipoib_ib_dev_flush_normal); 2144 INIT_WORK(&priv->flush_heavy, ipoib_ib_dev_flush_heavy); 2145 INIT_WORK(&priv->restart_task, ipoib_mcast_restart_task); 2146 INIT_DELAYED_WORK(&priv->ah_reap_task, ipoib_reap_ah); 2147 INIT_DELAYED_WORK(&priv->neigh_reap_task, ipoib_reap_neigh); 2148} 2149 2150static struct net_device *ipoib_alloc_netdev(struct ib_device *hca, u8 port, 2151 const char *name) 2152{ 2153 struct net_device *dev; 2154 2155 dev = rdma_alloc_netdev(hca, port, RDMA_NETDEV_IPOIB, name, 2156 NET_NAME_UNKNOWN, ipoib_setup_common); 2157 if (!IS_ERR(dev) || PTR_ERR(dev) != -EOPNOTSUPP) 2158 return dev; 2159 2160 dev = alloc_netdev(sizeof(struct rdma_netdev), name, NET_NAME_UNKNOWN, 2161 ipoib_setup_common); 2162 if (!dev) 2163 return ERR_PTR(-ENOMEM); 2164 return dev; 2165} 2166 2167int ipoib_intf_init(struct ib_device *hca, u8 port, const char *name, 2168 struct net_device *dev) 2169{ 2170 struct rdma_netdev *rn = netdev_priv(dev); 2171 struct ipoib_dev_priv *priv; 2172 int rc; 2173 2174 priv = kzalloc(sizeof(*priv), GFP_KERNEL); 2175 if (!priv) 2176 return -ENOMEM; 2177 2178 priv->ca = hca; 2179 priv->port = port; 2180 2181 rc = rdma_init_netdev(hca, port, RDMA_NETDEV_IPOIB, name, 2182 NET_NAME_UNKNOWN, ipoib_setup_common, dev); 2183 if (rc) { 2184 if (rc != -EOPNOTSUPP) 2185 goto out; 2186 2187 rn->send = ipoib_send; 2188 rn->attach_mcast = ipoib_mcast_attach; 2189 rn->detach_mcast = ipoib_mcast_detach; 2190 rn->hca = hca; 2191 2192 rc = netif_set_real_num_tx_queues(dev, 1); 2193 if (rc) 2194 goto out; 2195 2196 rc = netif_set_real_num_rx_queues(dev, 1); 2197 if (rc) 2198 goto out; 2199 } 2200 2201 priv->rn_ops = dev->netdev_ops; 2202 2203 if (hca->attrs.device_cap_flags & IB_DEVICE_VIRTUAL_FUNCTION) 2204 dev->netdev_ops = &ipoib_netdev_ops_vf; 2205 else 2206 dev->netdev_ops = &ipoib_netdev_ops_pf; 2207 2208 rn->clnt_priv = priv; 2209 /* 2210 * Only the child register_netdev flows can handle priv_destructor 2211 * being set, so we force it to NULL here and handle manually until it 2212 * is safe to turn on. 2213 */ 2214 priv->next_priv_destructor = dev->priv_destructor; 2215 dev->priv_destructor = NULL; 2216 2217 ipoib_build_priv(dev); 2218 2219 return 0; 2220 2221out: 2222 kfree(priv); 2223 return rc; 2224} 2225 2226struct net_device *ipoib_intf_alloc(struct ib_device *hca, u8 port, 2227 const char *name) 2228{ 2229 struct net_device *dev; 2230 int rc; 2231 2232 dev = ipoib_alloc_netdev(hca, port, name); 2233 if (IS_ERR(dev)) 2234 return dev; 2235 2236 rc = ipoib_intf_init(hca, port, name, dev); 2237 if (rc) { 2238 free_netdev(dev); 2239 return ERR_PTR(rc); 2240 } 2241 2242 /* 2243 * Upon success the caller must ensure ipoib_intf_free is called or 2244 * register_netdevice succeed'd and priv_destructor is set to 2245 * ipoib_intf_free. 2246 */ 2247 return dev; 2248} 2249 2250void ipoib_intf_free(struct net_device *dev) 2251{ 2252 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2253 struct rdma_netdev *rn = netdev_priv(dev); 2254 2255 dev->priv_destructor = priv->next_priv_destructor; 2256 if (dev->priv_destructor) 2257 dev->priv_destructor(dev); 2258 2259 /* 2260 * There are some error flows around register_netdev failing that may 2261 * attempt to call priv_destructor twice, prevent that from happening. 2262 */ 2263 dev->priv_destructor = NULL; 2264 2265 /* unregister/destroy is very complicated. Make bugs more obvious. */ 2266 rn->clnt_priv = NULL; 2267 2268 kfree(priv); 2269} 2270 2271static ssize_t show_pkey(struct device *dev, 2272 struct device_attribute *attr, char *buf) 2273{ 2274 struct net_device *ndev = to_net_dev(dev); 2275 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 2276 2277 return sprintf(buf, "0x%04x\n", priv->pkey); 2278} 2279static DEVICE_ATTR(pkey, S_IRUGO, show_pkey, NULL); 2280 2281static ssize_t show_umcast(struct device *dev, 2282 struct device_attribute *attr, char *buf) 2283{ 2284 struct net_device *ndev = to_net_dev(dev); 2285 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 2286 2287 return sprintf(buf, "%d\n", test_bit(IPOIB_FLAG_UMCAST, &priv->flags)); 2288} 2289 2290void ipoib_set_umcast(struct net_device *ndev, int umcast_val) 2291{ 2292 struct ipoib_dev_priv *priv = ipoib_priv(ndev); 2293 2294 if (umcast_val > 0) { 2295 set_bit(IPOIB_FLAG_UMCAST, &priv->flags); 2296 ipoib_warn(priv, "ignoring multicast groups joined directly " 2297 "by userspace\n"); 2298 } else 2299 clear_bit(IPOIB_FLAG_UMCAST, &priv->flags); 2300} 2301 2302static ssize_t set_umcast(struct device *dev, 2303 struct device_attribute *attr, 2304 const char *buf, size_t count) 2305{ 2306 unsigned long umcast_val = simple_strtoul(buf, NULL, 0); 2307 2308 ipoib_set_umcast(to_net_dev(dev), umcast_val); 2309 2310 return count; 2311} 2312static DEVICE_ATTR(umcast, S_IWUSR | S_IRUGO, show_umcast, set_umcast); 2313 2314int ipoib_add_umcast_attr(struct net_device *dev) 2315{ 2316 return device_create_file(&dev->dev, &dev_attr_umcast); 2317} 2318 2319static void set_base_guid(struct ipoib_dev_priv *priv, union ib_gid *gid) 2320{ 2321 struct ipoib_dev_priv *child_priv; 2322 struct net_device *netdev = priv->dev; 2323 2324 netif_addr_lock_bh(netdev); 2325 2326 memcpy(&priv->local_gid.global.interface_id, 2327 &gid->global.interface_id, 2328 sizeof(gid->global.interface_id)); 2329 memcpy(netdev->dev_addr + 4, &priv->local_gid, sizeof(priv->local_gid)); 2330 clear_bit(IPOIB_FLAG_DEV_ADDR_SET, &priv->flags); 2331 2332 netif_addr_unlock_bh(netdev); 2333 2334 if (!test_bit(IPOIB_FLAG_SUBINTERFACE, &priv->flags)) { 2335 down_read(&priv->vlan_rwsem); 2336 list_for_each_entry(child_priv, &priv->child_intfs, list) 2337 set_base_guid(child_priv, gid); 2338 up_read(&priv->vlan_rwsem); 2339 } 2340} 2341 2342static int ipoib_check_lladdr(struct net_device *dev, 2343 struct sockaddr_storage *ss) 2344{ 2345 union ib_gid *gid = (union ib_gid *)(ss->__data + 4); 2346 int ret = 0; 2347 2348 netif_addr_lock_bh(dev); 2349 2350 /* Make sure the QPN, reserved and subnet prefix match the current 2351 * lladdr, it also makes sure the lladdr is unicast. 2352 */ 2353 if (memcmp(dev->dev_addr, ss->__data, 2354 4 + sizeof(gid->global.subnet_prefix)) || 2355 gid->global.interface_id == 0) 2356 ret = -EINVAL; 2357 2358 netif_addr_unlock_bh(dev); 2359 2360 return ret; 2361} 2362 2363static int ipoib_set_mac(struct net_device *dev, void *addr) 2364{ 2365 struct ipoib_dev_priv *priv = ipoib_priv(dev); 2366 struct sockaddr_storage *ss = addr; 2367 int ret; 2368 2369 if (!(dev->priv_flags & IFF_LIVE_ADDR_CHANGE) && netif_running(dev)) 2370 return -EBUSY; 2371 2372 ret = ipoib_check_lladdr(dev, ss); 2373 if (ret) 2374 return ret; 2375 2376 set_base_guid(priv, (union ib_gid *)(ss->__data + 4)); 2377 2378 queue_work(ipoib_workqueue, &priv->flush_light); 2379 2380 return 0; 2381} 2382 2383static ssize_t create_child(struct device *dev, 2384 struct device_attribute *attr, 2385 const char *buf, size_t count) 2386{ 2387 int pkey; 2388 int ret; 2389 2390 if (sscanf(buf, "%i", &pkey) != 1) 2391 return -EINVAL; 2392 2393 if (pkey <= 0 || pkey > 0xffff || pkey == 0x8000) 2394 return -EINVAL; 2395 2396 ret = ipoib_vlan_add(to_net_dev(dev), pkey); 2397 2398 return ret ? ret : count; 2399} 2400static DEVICE_ATTR(create_child, S_IWUSR, NULL, create_child); 2401 2402static ssize_t delete_child(struct device *dev, 2403 struct device_attribute *attr, 2404 const char *buf, size_t count) 2405{ 2406 int pkey; 2407 int ret; 2408 2409 if (sscanf(buf, "%i", &pkey) != 1) 2410 return -EINVAL; 2411 2412 if (pkey < 0 || pkey > 0xffff) 2413 return -EINVAL; 2414 2415 ret = ipoib_vlan_delete(to_net_dev(dev), pkey); 2416 2417 return ret ? ret : count; 2418 2419} 2420static DEVICE_ATTR(delete_child, S_IWUSR, NULL, delete_child); 2421 2422int ipoib_add_pkey_attr(struct net_device *dev) 2423{ 2424 return device_create_file(&dev->dev, &dev_attr_pkey); 2425} 2426 2427/* 2428 * We erroneously exposed the iface's port number in the dev_id 2429 * sysfs field long after dev_port was introduced for that purpose[1], 2430 * and we need to stop everyone from relying on that. 2431 * Let's overload the shower routine for the dev_id file here 2432 * to gently bring the issue up. 2433 * 2434 * [1] https://www.spinics.net/lists/netdev/msg272123.html 2435 */ 2436static ssize_t dev_id_show(struct device *dev, 2437 struct device_attribute *attr, char *buf) 2438{ 2439 struct net_device *ndev = to_net_dev(dev); 2440 2441 /* 2442 * ndev->dev_port will be equal to 0 in old kernel prior to commit 2443 * 9b8b2a323008 ("IB/ipoib: Use dev_port to expose network interface 2444 * port numbers") Zero was chosen as special case for user space 2445 * applications to fallback and query dev_id to check if it has 2446 * different value or not. 2447 * 2448 * Don't print warning in such scenario. 2449 * 2450 * https://github.com/systemd/systemd/blob/master/src/udev/udev-builtin-net_id.c#L358 2451 */ 2452 if (ndev->dev_port && ndev->dev_id == ndev->dev_port) 2453 netdev_info_once(ndev, 2454 "\"%s\" wants to know my dev_id. Should it look at dev_port instead? See Documentation/ABI/testing/sysfs-class-net for more info.\n", 2455 current->comm); 2456 2457 return sprintf(buf, "%#x\n", ndev->dev_id); 2458} 2459static DEVICE_ATTR_RO(dev_id); 2460 2461static int ipoib_intercept_dev_id_attr(struct net_device *dev) 2462{ 2463 device_remove_file(&dev->dev, &dev_attr_dev_id); 2464 return device_create_file(&dev->dev, &dev_attr_dev_id); 2465} 2466 2467static struct net_device *ipoib_add_port(const char *format, 2468 struct ib_device *hca, u8 port) 2469{ 2470 struct rtnl_link_ops *ops = ipoib_get_link_ops(); 2471 struct rdma_netdev_alloc_params params; 2472 struct ipoib_dev_priv *priv; 2473 struct net_device *ndev; 2474 int result; 2475 2476 ndev = ipoib_intf_alloc(hca, port, format); 2477 if (IS_ERR(ndev)) { 2478 pr_warn("%s, %d: ipoib_intf_alloc failed %ld\n", hca->name, port, 2479 PTR_ERR(ndev)); 2480 return ndev; 2481 } 2482 priv = ipoib_priv(ndev); 2483 2484 INIT_IB_EVENT_HANDLER(&priv->event_handler, 2485 priv->ca, ipoib_event); 2486 ib_register_event_handler(&priv->event_handler); 2487 2488 /* call event handler to ensure pkey in sync */ 2489 queue_work(ipoib_workqueue, &priv->flush_heavy); 2490 2491 ndev->rtnl_link_ops = ipoib_get_link_ops(); 2492 2493 result = register_netdev(ndev); 2494 if (result) { 2495 pr_warn("%s: couldn't register ipoib port %d; error %d\n", 2496 hca->name, port, result); 2497 2498 ipoib_parent_unregister_pre(ndev); 2499 ipoib_intf_free(ndev); 2500 free_netdev(ndev); 2501 2502 return ERR_PTR(result); 2503 } 2504 2505 if (hca->ops.rdma_netdev_get_params) { 2506 int rc = hca->ops.rdma_netdev_get_params(hca, port, 2507 RDMA_NETDEV_IPOIB, 2508 ¶ms); 2509 2510 if (!rc && ops->priv_size < params.sizeof_priv) 2511 ops->priv_size = params.sizeof_priv; 2512 } 2513 /* 2514 * We cannot set priv_destructor before register_netdev because we 2515 * need priv to be always valid during the error flow to execute 2516 * ipoib_parent_unregister_pre(). Instead handle it manually and only 2517 * enter priv_destructor mode once we are completely registered. 2518 */ 2519 ndev->priv_destructor = ipoib_intf_free; 2520 2521 if (ipoib_intercept_dev_id_attr(ndev)) 2522 goto sysfs_failed; 2523 if (ipoib_cm_add_mode_attr(ndev)) 2524 goto sysfs_failed; 2525 if (ipoib_add_pkey_attr(ndev)) 2526 goto sysfs_failed; 2527 if (ipoib_add_umcast_attr(ndev)) 2528 goto sysfs_failed; 2529 if (device_create_file(&ndev->dev, &dev_attr_create_child)) 2530 goto sysfs_failed; 2531 if (device_create_file(&ndev->dev, &dev_attr_delete_child)) 2532 goto sysfs_failed; 2533 2534 return ndev; 2535 2536sysfs_failed: 2537 ipoib_parent_unregister_pre(ndev); 2538 unregister_netdev(ndev); 2539 return ERR_PTR(-ENOMEM); 2540} 2541 2542static int ipoib_add_one(struct ib_device *device) 2543{ 2544 struct list_head *dev_list; 2545 struct net_device *dev; 2546 struct ipoib_dev_priv *priv; 2547 unsigned int p; 2548 int count = 0; 2549 2550 dev_list = kmalloc(sizeof(*dev_list), GFP_KERNEL); 2551 if (!dev_list) 2552 return -ENOMEM; 2553 2554 INIT_LIST_HEAD(dev_list); 2555 2556 rdma_for_each_port (device, p) { 2557 if (!rdma_protocol_ib(device, p)) 2558 continue; 2559 dev = ipoib_add_port("ib%d", device, p); 2560 if (!IS_ERR(dev)) { 2561 priv = ipoib_priv(dev); 2562 list_add_tail(&priv->list, dev_list); 2563 count++; 2564 } 2565 } 2566 2567 if (!count) { 2568 kfree(dev_list); 2569 return -EOPNOTSUPP; 2570 } 2571 2572 ib_set_client_data(device, &ipoib_client, dev_list); 2573 return 0; 2574} 2575 2576static void ipoib_remove_one(struct ib_device *device, void *client_data) 2577{ 2578 struct ipoib_dev_priv *priv, *tmp, *cpriv, *tcpriv; 2579 struct list_head *dev_list = client_data; 2580 2581 list_for_each_entry_safe(priv, tmp, dev_list, list) { 2582 LIST_HEAD(head); 2583 ipoib_parent_unregister_pre(priv->dev); 2584 2585 rtnl_lock(); 2586 2587 list_for_each_entry_safe(cpriv, tcpriv, &priv->child_intfs, 2588 list) 2589 unregister_netdevice_queue(cpriv->dev, &head); 2590 unregister_netdevice_queue(priv->dev, &head); 2591 unregister_netdevice_many(&head); 2592 2593 rtnl_unlock(); 2594 } 2595 2596 kfree(dev_list); 2597} 2598 2599#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 2600static struct notifier_block ipoib_netdev_notifier = { 2601 .notifier_call = ipoib_netdev_event, 2602}; 2603#endif 2604 2605static int __init ipoib_init_module(void) 2606{ 2607 int ret; 2608 2609 ipoib_recvq_size = roundup_pow_of_two(ipoib_recvq_size); 2610 ipoib_recvq_size = min(ipoib_recvq_size, IPOIB_MAX_QUEUE_SIZE); 2611 ipoib_recvq_size = max(ipoib_recvq_size, IPOIB_MIN_QUEUE_SIZE); 2612 2613 ipoib_sendq_size = roundup_pow_of_two(ipoib_sendq_size); 2614 ipoib_sendq_size = min(ipoib_sendq_size, IPOIB_MAX_QUEUE_SIZE); 2615 ipoib_sendq_size = max3(ipoib_sendq_size, 2 * MAX_SEND_CQE, IPOIB_MIN_QUEUE_SIZE); 2616#ifdef CONFIG_INFINIBAND_IPOIB_CM 2617 ipoib_max_conn_qp = min(ipoib_max_conn_qp, IPOIB_CM_MAX_CONN_QP); 2618 ipoib_max_conn_qp = max(ipoib_max_conn_qp, 0); 2619#endif 2620 2621 /* 2622 * When copying small received packets, we only copy from the 2623 * linear data part of the SKB, so we rely on this condition. 2624 */ 2625 BUILD_BUG_ON(IPOIB_CM_COPYBREAK > IPOIB_CM_HEAD_SIZE); 2626 2627 ipoib_register_debugfs(); 2628 2629 /* 2630 * We create a global workqueue here that is used for all flush 2631 * operations. However, if you attempt to flush a workqueue 2632 * from a task on that same workqueue, it deadlocks the system. 2633 * We want to be able to flush the tasks associated with a 2634 * specific net device, so we also create a workqueue for each 2635 * netdevice. We queue up the tasks for that device only on 2636 * its private workqueue, and we only queue up flush events 2637 * on our global flush workqueue. This avoids the deadlocks. 2638 */ 2639 ipoib_workqueue = alloc_ordered_workqueue("ipoib_flush", 0); 2640 if (!ipoib_workqueue) { 2641 ret = -ENOMEM; 2642 goto err_fs; 2643 } 2644 2645 ib_sa_register_client(&ipoib_sa_client); 2646 2647 ret = ib_register_client(&ipoib_client); 2648 if (ret) 2649 goto err_sa; 2650 2651 ret = ipoib_netlink_init(); 2652 if (ret) 2653 goto err_client; 2654 2655#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 2656 register_netdevice_notifier(&ipoib_netdev_notifier); 2657#endif 2658 return 0; 2659 2660err_client: 2661 ib_unregister_client(&ipoib_client); 2662 2663err_sa: 2664 ib_sa_unregister_client(&ipoib_sa_client); 2665 destroy_workqueue(ipoib_workqueue); 2666 2667err_fs: 2668 ipoib_unregister_debugfs(); 2669 2670 return ret; 2671} 2672 2673static void __exit ipoib_cleanup_module(void) 2674{ 2675#ifdef CONFIG_INFINIBAND_IPOIB_DEBUG 2676 unregister_netdevice_notifier(&ipoib_netdev_notifier); 2677#endif 2678 ipoib_netlink_fini(); 2679 ib_unregister_client(&ipoib_client); 2680 ib_sa_unregister_client(&ipoib_sa_client); 2681 ipoib_unregister_debugfs(); 2682 destroy_workqueue(ipoib_workqueue); 2683} 2684 2685module_init(ipoib_init_module); 2686module_exit(ipoib_cleanup_module); 2687