drivers/staging/lustre/lnet/lnet/net_fault.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /*
   3  * GPL HEADER START
   4  *
   5  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License version 2 only,
   9  * as published by the Free Software Foundation.
  10  *
  11  * This program is distributed in the hope that it will be useful, but
  12  * WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * General Public License version 2 for more details (a copy is included
  15  * in the LICENSE file that accompanied this code).
  16  *
  17  * You should have received a copy of the GNU General Public License
  18  * version 2 along with this program; If not, see
  19  * http://www.gnu.org/licenses/gpl-2.0.html
  20  *
  21  * GPL HEADER END
  22  */
  23 /*
  24  * Copyright (c) 2014, Intel Corporation.
  25  */
  26 /*
  27  * This file is part of Lustre, http://www.lustre.org/
  28  * Lustre is a trademark of Seagate, Inc.
  29  *
  30  * lnet/lnet/net_fault.c
  31  *
  32  * Lustre network fault simulation
  33  *
  34  * Author: liang.zhen@intel.com
  35  */
  36
  37 #define DEBUG_SUBSYSTEM S_LNET
  38
  39 #include <linux/lnet/lib-lnet.h>
  40 #include <uapi/linux/lnet/lnetctl.h>
  41
  42 #define LNET_MSG_MASK           (LNET_PUT_BIT | LNET_ACK_BIT | \
  43                                  LNET_GET_BIT | LNET_REPLY_BIT)
  44
  45 struct lnet_drop_rule {
  46         /** link chain on the_lnet.ln_drop_rules */
  47         struct list_head        dr_link;
  48         /** attributes of this rule */
  49         struct lnet_fault_attr  dr_attr;
  50         /** lock to protect \a dr_drop_at and \a dr_stat */
  51         spinlock_t              dr_lock;
  52         /**
  53          * the message sequence to drop, which means message is dropped when
  54          * dr_stat.drs_count == dr_drop_at
  55          */
  56         unsigned long           dr_drop_at;
  57         /**
  58          * seconds to drop the next message, it's exclusive with dr_drop_at
  59          */
  60         unsigned long           dr_drop_time;
  61         /** baseline to caculate dr_drop_time */
  62         unsigned long           dr_time_base;
  63         /** statistic of dropped messages */
  64         struct lnet_fault_stat  dr_stat;
  65 };
  66
  67 static bool
  68 lnet_fault_nid_match(lnet_nid_t nid, lnet_nid_t msg_nid)
  69 {
  70         if (nid == msg_nid || nid == LNET_NID_ANY)
  71                 return true;
  72
  73         if (LNET_NIDNET(nid) != LNET_NIDNET(msg_nid))
  74                 return false;
  75
  76         /* 255.255.255.255@net is wildcard for all addresses in a network */
  77         return LNET_NIDADDR(nid) == LNET_NIDADDR(LNET_NID_ANY);
  78 }
  79
  80 static bool
  81 lnet_fault_attr_match(struct lnet_fault_attr *attr, lnet_nid_t src,
  82                       lnet_nid_t dst, unsigned int type, unsigned int portal)
  83 {
  84         if (!lnet_fault_nid_match(attr->fa_src, src) ||
  85             !lnet_fault_nid_match(attr->fa_dst, dst))
  86                 return false;
  87
  88         if (!(attr->fa_msg_mask & (1 << type)))
  89                 return false;
  90
  91         /**
  92          * NB: ACK and REPLY have no portal, but they should have been
  93          * rejected by message mask
  94          */
  95         if (attr->fa_ptl_mask && /* has portal filter */
  96             !(attr->fa_ptl_mask & (1ULL << portal)))
  97                 return false;
  98
  99         return true;
 100 }
 101
 102 static int
 103 lnet_fault_attr_validate(struct lnet_fault_attr *attr)
 104 {
 105         if (!attr->fa_msg_mask)
 106                 attr->fa_msg_mask = LNET_MSG_MASK; /* all message types */
 107
 108         if (!attr->fa_ptl_mask) /* no portal filter */
 109                 return 0;
 110
 111         /* NB: only PUT and GET can be filtered if portal filter has been set */
 112         attr->fa_msg_mask &= LNET_GET_BIT | LNET_PUT_BIT;
 113         if (!attr->fa_msg_mask) {
 114                 CDEBUG(D_NET, "can't find valid message type bits %x\n",
 115                        attr->fa_msg_mask);
 116                 return -EINVAL;
 117         }
 118         return 0;
 119 }
 120
 121 static void
 122 lnet_fault_stat_inc(struct lnet_fault_stat *stat, unsigned int type)
 123 {
 124         /* NB: fs_counter is NOT updated by this function */
 125         switch (type) {
 126         case LNET_MSG_PUT:
 127                 stat->fs_put++;
 128                 return;
 129         case LNET_MSG_ACK:
 130                 stat->fs_ack++;
 131                 return;
 132         case LNET_MSG_GET:
 133                 stat->fs_get++;
 134                 return;
 135         case LNET_MSG_REPLY:
 136                 stat->fs_reply++;
 137                 return;
 138         }
 139 }
 140
 141 /**
 142  * LNet message drop simulation
 143  */
 144
 145 /**
 146  * Add a new drop rule to LNet
 147  * There is no check for duplicated drop rule, all rules will be checked for
 148  * incoming message.
 149  */
 150 static int
 151 lnet_drop_rule_add(struct lnet_fault_attr *attr)
 152 {
 153         struct lnet_drop_rule *rule;
 154
 155         if (attr->u.drop.da_rate & attr->u.drop.da_interval) {
 156                 CDEBUG(D_NET, "please provide either drop rate or drop interval, but not both at the same time %d/%d\n",
 157                        attr->u.drop.da_rate, attr->u.drop.da_interval);
 158                 return -EINVAL;
 159         }
 160
 161         if (lnet_fault_attr_validate(attr))
 162                 return -EINVAL;
 163
 164         CFS_ALLOC_PTR(rule);
 165         if (!rule)
 166                 return -ENOMEM;
 167
 168         spin_lock_init(&rule->dr_lock);
 169
 170         rule->dr_attr = *attr;
 171         if (attr->u.drop.da_interval) {
 172                 rule->dr_time_base = cfs_time_shift(attr->u.drop.da_interval);
 173                 rule->dr_drop_time = cfs_time_shift(cfs_rand() %
 174                                                     attr->u.drop.da_interval);
 175         } else {
 176                 rule->dr_drop_at = cfs_rand() % attr->u.drop.da_rate;
 177         }
 178
 179         lnet_net_lock(LNET_LOCK_EX);
 180         list_add(&rule->dr_link, &the_lnet.ln_drop_rules);
 181         lnet_net_unlock(LNET_LOCK_EX);
 182
 183         CDEBUG(D_NET, "Added drop rule: src %s, dst %s, rate %d, interval %d\n",
 184                libcfs_nid2str(attr->fa_src), libcfs_nid2str(attr->fa_src),
 185                attr->u.drop.da_rate, attr->u.drop.da_interval);
 186         return 0;
 187 }
 188
 189 /**
 190  * Remove matched drop rules from lnet, all rules that can match \a src and
 191  * \a dst will be removed.
 192  * If \a src is zero, then all rules have \a dst as destination will be remove
 193  * If \a dst is zero, then all rules have \a src as source will be removed
 194  * If both of them are zero, all rules will be removed
 195  */
 196 static int
 197 lnet_drop_rule_del(lnet_nid_t src, lnet_nid_t dst)
 198 {
 199         struct lnet_drop_rule *rule;
 200         struct lnet_drop_rule *tmp;
 201         struct list_head zombies;
 202         int n = 0;
 203
 204         INIT_LIST_HEAD(&zombies);
 205
 206         lnet_net_lock(LNET_LOCK_EX);
 207         list_for_each_entry_safe(rule, tmp, &the_lnet.ln_drop_rules, dr_link) {
 208                 if (rule->dr_attr.fa_src != src && src)
 209                         continue;
 210
 211                 if (rule->dr_attr.fa_dst != dst && dst)
 212                         continue;
 213
 214                 list_move(&rule->dr_link, &zombies);
 215         }
 216         lnet_net_unlock(LNET_LOCK_EX);
 217
 218         list_for_each_entry_safe(rule, tmp, &zombies, dr_link) {
 219                 CDEBUG(D_NET, "Remove drop rule: src %s->dst: %s (1/%d, %d)\n",
 220                        libcfs_nid2str(rule->dr_attr.fa_src),
 221                        libcfs_nid2str(rule->dr_attr.fa_dst),
 222                        rule->dr_attr.u.drop.da_rate,
 223                        rule->dr_attr.u.drop.da_interval);
 224
 225                 list_del(&rule->dr_link);
 226                 CFS_FREE_PTR(rule);
 227                 n++;
 228         }
 229
 230         return n;
 231 }
 232
 233 /**
 234  * List drop rule at position of \a pos
 235  */
 236 static int
 237 lnet_drop_rule_list(int pos, struct lnet_fault_attr *attr,
 238                     struct lnet_fault_stat *stat)
 239 {
 240         struct lnet_drop_rule *rule;
 241         int cpt;
 242         int i = 0;
 243         int rc = -ENOENT;
 244
 245         cpt = lnet_net_lock_current();
 246         list_for_each_entry(rule, &the_lnet.ln_drop_rules, dr_link) {
 247                 if (i++ < pos)
 248                         continue;
 249
 250                 spin_lock(&rule->dr_lock);
 251                 *attr = rule->dr_attr;
 252                 *stat = rule->dr_stat;
 253                 spin_unlock(&rule->dr_lock);
 254                 rc = 0;
 255                 break;
 256         }
 257
 258         lnet_net_unlock(cpt);
 259         return rc;
 260 }
 261
 262 /**
 263  * reset counters for all drop rules
 264  */
 265 static void
 266 lnet_drop_rule_reset(void)
 267 {
 268         struct lnet_drop_rule *rule;
 269         int cpt;
 270
 271         cpt = lnet_net_lock_current();
 272
 273         list_for_each_entry(rule, &the_lnet.ln_drop_rules, dr_link) {
 274                 struct lnet_fault_attr *attr = &rule->dr_attr;
 275
 276                 spin_lock(&rule->dr_lock);
 277
 278                 memset(&rule->dr_stat, 0, sizeof(rule->dr_stat));
 279                 if (attr->u.drop.da_rate) {
 280                         rule->dr_drop_at = cfs_rand() % attr->u.drop.da_rate;
 281                 } else {
 282                         rule->dr_drop_time = cfs_time_shift(cfs_rand() %
 283                                                 attr->u.drop.da_interval);
 284                         rule->dr_time_base = cfs_time_shift(attr->u.drop.da_interval);
 285                 }
 286                 spin_unlock(&rule->dr_lock);
 287         }
 288
 289         lnet_net_unlock(cpt);
 290 }
 291
 292 /**
 293  * check source/destination NID, portal, message type and drop rate,
 294  * decide whether should drop this message or not
 295  */
 296 static bool
 297 drop_rule_match(struct lnet_drop_rule *rule, lnet_nid_t src,
 298                 lnet_nid_t dst, unsigned int type, unsigned int portal)
 299 {
 300         struct lnet_fault_attr *attr = &rule->dr_attr;
 301         bool drop;
 302
 303         if (!lnet_fault_attr_match(attr, src, dst, type, portal))
 304                 return false;
 305
 306         /* match this rule, check drop rate now */
 307         spin_lock(&rule->dr_lock);
 308         if (rule->dr_drop_time) { /* time based drop */
 309                 unsigned long now = cfs_time_current();
 310
 311                 rule->dr_stat.fs_count++;
 312                 drop = cfs_time_aftereq(now, rule->dr_drop_time);
 313                 if (drop) {
 314                         if (cfs_time_after(now, rule->dr_time_base))
 315                                 rule->dr_time_base = now;
 316
 317                         rule->dr_drop_time = rule->dr_time_base +
 318                                              cfs_time_seconds(cfs_rand() %
 319                                                 attr->u.drop.da_interval);
 320                         rule->dr_time_base += cfs_time_seconds(attr->u.drop.da_interval);
 321
 322                         CDEBUG(D_NET, "Drop Rule %s->%s: next drop : %lu\n",
 323                                libcfs_nid2str(attr->fa_src),
 324                                libcfs_nid2str(attr->fa_dst),
 325                                rule->dr_drop_time);
 326                 }
 327
 328         } else { /* rate based drop */
 329                 drop = rule->dr_stat.fs_count++ == rule->dr_drop_at;
 330
 331                 if (!do_div(rule->dr_stat.fs_count, attr->u.drop.da_rate)) {
 332                         rule->dr_drop_at = rule->dr_stat.fs_count +
 333                                            cfs_rand() % attr->u.drop.da_rate;
 334                         CDEBUG(D_NET, "Drop Rule %s->%s: next drop: %lu\n",
 335                                libcfs_nid2str(attr->fa_src),
 336                                libcfs_nid2str(attr->fa_dst), rule->dr_drop_at);
 337                 }
 338         }
 339
 340         if (drop) { /* drop this message, update counters */
 341                 lnet_fault_stat_inc(&rule->dr_stat, type);
 342                 rule->dr_stat.u.drop.ds_dropped++;
 343         }
 344
 345         spin_unlock(&rule->dr_lock);
 346         return drop;
 347 }
 348
 349 /**
 350  * Check if message from \a src to \a dst can match any existed drop rule
 351  */
 352 bool
 353 lnet_drop_rule_match(struct lnet_hdr *hdr)
 354 {
 355         struct lnet_drop_rule *rule;
 356         lnet_nid_t src = le64_to_cpu(hdr->src_nid);
 357         lnet_nid_t dst = le64_to_cpu(hdr->dest_nid);
 358         unsigned int typ = le32_to_cpu(hdr->type);
 359         unsigned int ptl = -1;
 360         bool drop = false;
 361         int cpt;
 362
 363         /**
 364          * NB: if Portal is specified, then only PUT and GET will be
 365          * filtered by drop rule
 366          */
 367         if (typ == LNET_MSG_PUT)
 368                 ptl = le32_to_cpu(hdr->msg.put.ptl_index);
 369         else if (typ == LNET_MSG_GET)
 370                 ptl = le32_to_cpu(hdr->msg.get.ptl_index);
 371
 372         cpt = lnet_net_lock_current();
 373         list_for_each_entry(rule, &the_lnet.ln_drop_rules, dr_link) {
 374                 drop = drop_rule_match(rule, src, dst, typ, ptl);
 375                 if (drop)
 376                         break;
 377         }
 378
 379         lnet_net_unlock(cpt);
 380         return drop;
 381 }
 382
 383 /**
 384  * LNet Delay Simulation
 385  */
 386 /** timestamp (second) to send delayed message */
 387 #define msg_delay_send           msg_ev.hdr_data
 388
 389 struct lnet_delay_rule {
 390         /** link chain on the_lnet.ln_delay_rules */
 391         struct list_head        dl_link;
 392         /** link chain on delay_dd.dd_sched_rules */
 393         struct list_head        dl_sched_link;
 394         /** attributes of this rule */
 395         struct lnet_fault_attr  dl_attr;
 396         /** lock to protect \a below members */
 397         spinlock_t              dl_lock;
 398         /** refcount of delay rule */
 399         atomic_t                dl_refcount;
 400         /**
 401          * the message sequence to delay, which means message is delayed when
 402          * dl_stat.fs_count == dl_delay_at
 403          */
 404         unsigned long           dl_delay_at;
 405         /**
 406          * seconds to delay the next message, it's exclusive with dl_delay_at
 407          */
 408         unsigned long           dl_delay_time;
 409         /** baseline to caculate dl_delay_time */
 410         unsigned long           dl_time_base;
 411         /** jiffies to send the next delayed message */
 412         unsigned long           dl_msg_send;
 413         /** delayed message list */
 414         struct list_head        dl_msg_list;
 415         /** statistic of delayed messages */
 416         struct lnet_fault_stat  dl_stat;
 417         /** timer to wakeup delay_daemon */
 418         struct timer_list       dl_timer;
 419 };
 420
 421 struct delay_daemon_data {
 422         /** serialise rule add/remove */
 423         struct mutex            dd_mutex;
 424         /** protect rules on \a dd_sched_rules */
 425         spinlock_t              dd_lock;
 426         /** scheduled delay rules (by timer) */
 427         struct list_head        dd_sched_rules;
 428         /** daemon thread sleeps at here */
 429         wait_queue_head_t       dd_waitq;
 430         /** controller (lctl command) wait at here */
 431         wait_queue_head_t       dd_ctl_waitq;
 432         /** daemon is running */
 433         unsigned int            dd_running;
 434         /** daemon stopped */
 435         unsigned int            dd_stopped;
 436 };
 437
 438 static struct delay_daemon_data delay_dd;
 439
 440 static unsigned long
 441 round_timeout(unsigned long timeout)
 442 {
 443         return cfs_time_seconds((unsigned int)
 444                         cfs_duration_sec(cfs_time_sub(timeout, 0)) + 1);
 445 }
 446
 447 static void
 448 delay_rule_decref(struct lnet_delay_rule *rule)
 449 {
 450         if (atomic_dec_and_test(&rule->dl_refcount)) {
 451                 LASSERT(list_empty(&rule->dl_sched_link));
 452                 LASSERT(list_empty(&rule->dl_msg_list));
 453                 LASSERT(list_empty(&rule->dl_link));
 454
 455                 CFS_FREE_PTR(rule);
 456         }
 457 }
 458
 459 /**
 460  * check source/destination NID, portal, message type and delay rate,
 461  * decide whether should delay this message or not
 462  */
 463 static bool
 464 delay_rule_match(struct lnet_delay_rule *rule, lnet_nid_t src,
 465                  lnet_nid_t dst, unsigned int type, unsigned int portal,
 466                  struct lnet_msg *msg)
 467 {
 468         struct lnet_fault_attr *attr = &rule->dl_attr;
 469         bool delay;
 470
 471         if (!lnet_fault_attr_match(attr, src, dst, type, portal))
 472                 return false;
 473
 474         /* match this rule, check delay rate now */
 475         spin_lock(&rule->dl_lock);
 476         if (rule->dl_delay_time) { /* time based delay */
 477                 unsigned long now = cfs_time_current();
 478
 479                 rule->dl_stat.fs_count++;
 480                 delay = cfs_time_aftereq(now, rule->dl_delay_time);
 481                 if (delay) {
 482                         if (cfs_time_after(now, rule->dl_time_base))
 483                                 rule->dl_time_base = now;
 484
 485                         rule->dl_delay_time = rule->dl_time_base +
 486                                              cfs_time_seconds(cfs_rand() %
 487                                                 attr->u.delay.la_interval);
 488                         rule->dl_time_base += cfs_time_seconds(attr->u.delay.la_interval);
 489
 490                         CDEBUG(D_NET, "Delay Rule %s->%s: next delay : %lu\n",
 491                                libcfs_nid2str(attr->fa_src),
 492                                libcfs_nid2str(attr->fa_dst),
 493                                rule->dl_delay_time);
 494                 }
 495
 496         } else { /* rate based delay */
 497                 delay = rule->dl_stat.fs_count++ == rule->dl_delay_at;
 498                 /* generate the next random rate sequence */
 499                 if (!do_div(rule->dl_stat.fs_count, attr->u.delay.la_rate)) {
 500                         rule->dl_delay_at = rule->dl_stat.fs_count +
 501                                             cfs_rand() % attr->u.delay.la_rate;
 502                         CDEBUG(D_NET, "Delay Rule %s->%s: next delay: %lu\n",
 503                                libcfs_nid2str(attr->fa_src),
 504                                libcfs_nid2str(attr->fa_dst), rule->dl_delay_at);
 505                 }
 506         }
 507
 508         if (!delay) {
 509                 spin_unlock(&rule->dl_lock);
 510                 return false;
 511         }
 512
 513         /* delay this message, update counters */
 514         lnet_fault_stat_inc(&rule->dl_stat, type);
 515         rule->dl_stat.u.delay.ls_delayed++;
 516
 517         list_add_tail(&msg->msg_list, &rule->dl_msg_list);
 518         msg->msg_delay_send = round_timeout(
 519                         cfs_time_shift(attr->u.delay.la_latency));
 520         if (rule->dl_msg_send == -1) {
 521                 rule->dl_msg_send = msg->msg_delay_send;
 522                 mod_timer(&rule->dl_timer, rule->dl_msg_send);
 523         }
 524
 525         spin_unlock(&rule->dl_lock);
 526         return true;
 527 }
 528
 529 /**
 530  * check if \a msg can match any Delay Rule, receiving of this message
 531  * will be delayed if there is a match.
 532  */
 533 bool
 534 lnet_delay_rule_match_locked(struct lnet_hdr *hdr, struct lnet_msg *msg)
 535 {
 536         struct lnet_delay_rule *rule;
 537         lnet_nid_t src = le64_to_cpu(hdr->src_nid);
 538         lnet_nid_t dst = le64_to_cpu(hdr->dest_nid);
 539         unsigned int typ = le32_to_cpu(hdr->type);
 540         unsigned int ptl = -1;
 541
 542         /* NB: called with hold of lnet_net_lock */
 543
 544         /**
 545          * NB: if Portal is specified, then only PUT and GET will be
 546          * filtered by delay rule
 547          */
 548         if (typ == LNET_MSG_PUT)
 549                 ptl = le32_to_cpu(hdr->msg.put.ptl_index);
 550         else if (typ == LNET_MSG_GET)
 551                 ptl = le32_to_cpu(hdr->msg.get.ptl_index);
 552
 553         list_for_each_entry(rule, &the_lnet.ln_delay_rules, dl_link) {
 554                 if (delay_rule_match(rule, src, dst, typ, ptl, msg))
 555                         return true;
 556         }
 557
 558         return false;
 559 }
 560
 561 /** check out delayed messages for send */
 562 static void
 563 delayed_msg_check(struct lnet_delay_rule *rule, bool all,
 564                   struct list_head *msg_list)
 565 {
 566         struct lnet_msg *msg;
 567         struct lnet_msg *tmp;
 568         unsigned long now = cfs_time_current();
 569
 570         if (!all && rule->dl_msg_send > now)
 571                 return;
 572
 573         spin_lock(&rule->dl_lock);
 574         list_for_each_entry_safe(msg, tmp, &rule->dl_msg_list, msg_list) {
 575                 if (!all && msg->msg_delay_send > now)
 576                         break;
 577
 578                 msg->msg_delay_send = 0;
 579                 list_move_tail(&msg->msg_list, msg_list);
 580         }
 581
 582         if (list_empty(&rule->dl_msg_list)) {
 583                 del_timer(&rule->dl_timer);
 584                 rule->dl_msg_send = -1;
 585
 586         } else if (!list_empty(msg_list)) {
 587                 /*
 588                  * dequeued some timedout messages, update timer for the
 589                  * next delayed message on rule
 590                  */
 591                 msg = list_entry(rule->dl_msg_list.next,
 592                                  struct lnet_msg, msg_list);
 593                 rule->dl_msg_send = msg->msg_delay_send;
 594                 mod_timer(&rule->dl_timer, rule->dl_msg_send);
 595         }
 596         spin_unlock(&rule->dl_lock);
 597 }
 598
 599 static void
 600 delayed_msg_process(struct list_head *msg_list, bool drop)
 601 {
 602         struct lnet_msg *msg;
 603
 604         while (!list_empty(msg_list)) {
 605                 struct lnet_ni *ni;
 606                 int cpt;
 607                 int rc;
 608
 609                 msg = list_entry(msg_list->next, struct lnet_msg, msg_list);
 610                 LASSERT(msg->msg_rxpeer);
 611
 612                 ni = msg->msg_rxpeer->lp_ni;
 613                 cpt = msg->msg_rx_cpt;
 614
 615                 list_del_init(&msg->msg_list);
 616                 if (drop) {
 617                         rc = -ECANCELED;
 618
 619                 } else if (!msg->msg_routing) {
 620                         rc = lnet_parse_local(ni, msg);
 621                         if (!rc)
 622                                 continue;
 623
 624                 } else {
 625                         lnet_net_lock(cpt);
 626                         rc = lnet_parse_forward_locked(ni, msg);
 627                         lnet_net_unlock(cpt);
 628
 629                         switch (rc) {
 630                         case LNET_CREDIT_OK:
 631                                 lnet_ni_recv(ni, msg->msg_private, msg, 0,
 632                                              0, msg->msg_len, msg->msg_len);
 633                                 /* fall through */
 634                         case LNET_CREDIT_WAIT:
 635                                 continue;
 636                         default: /* failures */
 637                                 break;
 638                         }
 639                 }
 640
 641                 lnet_drop_message(ni, cpt, msg->msg_private, msg->msg_len);
 642                 lnet_finalize(ni, msg, rc);
 643         }
 644 }
 645
 646 /**
 647  * Process delayed messages for scheduled rules
 648  * This function can either be called by delay_rule_daemon, or by lnet_finalise
 649  */
 650 void
 651 lnet_delay_rule_check(void)
 652 {
 653         struct lnet_delay_rule *rule;
 654         struct list_head msgs;
 655
 656         INIT_LIST_HEAD(&msgs);
 657         while (1) {
 658                 if (list_empty(&delay_dd.dd_sched_rules))
 659                         break;
 660
 661                 spin_lock_bh(&delay_dd.dd_lock);
 662                 if (list_empty(&delay_dd.dd_sched_rules)) {
 663                         spin_unlock_bh(&delay_dd.dd_lock);
 664                         break;
 665                 }
 666
 667                 rule = list_entry(delay_dd.dd_sched_rules.next,
 668                                   struct lnet_delay_rule, dl_sched_link);
 669                 list_del_init(&rule->dl_sched_link);
 670                 spin_unlock_bh(&delay_dd.dd_lock);
 671
 672                 delayed_msg_check(rule, false, &msgs);
 673                 delay_rule_decref(rule); /* -1 for delay_dd.dd_sched_rules */
 674         }
 675
 676         if (!list_empty(&msgs))
 677                 delayed_msg_process(&msgs, false);
 678 }
 679
 680 /** daemon thread to handle delayed messages */
 681 static int
 682 lnet_delay_rule_daemon(void *arg)
 683 {
 684         delay_dd.dd_running = 1;
 685         wake_up(&delay_dd.dd_ctl_waitq);
 686
 687         while (delay_dd.dd_running) {
 688                 wait_event_interruptible(delay_dd.dd_waitq,
 689                                          !delay_dd.dd_running ||
 690                                          !list_empty(&delay_dd.dd_sched_rules));
 691                 lnet_delay_rule_check();
 692         }
 693
 694         /* in case more rules have been enqueued after my last check */
 695         lnet_delay_rule_check();
 696         delay_dd.dd_stopped = 1;
 697         wake_up(&delay_dd.dd_ctl_waitq);
 698
 699         return 0;
 700 }
 701
 702 static void
 703 delay_timer_cb(struct timer_list *t)
 704 {
 705         struct lnet_delay_rule *rule = from_timer(rule, t, dl_timer);
 706
 707         spin_lock_bh(&delay_dd.dd_lock);
 708         if (list_empty(&rule->dl_sched_link) && delay_dd.dd_running) {
 709                 atomic_inc(&rule->dl_refcount);
 710                 list_add_tail(&rule->dl_sched_link, &delay_dd.dd_sched_rules);
 711                 wake_up(&delay_dd.dd_waitq);
 712         }
 713         spin_unlock_bh(&delay_dd.dd_lock);
 714 }
 715
 716 /**
 717  * Add a new delay rule to LNet
 718  * There is no check for duplicated delay rule, all rules will be checked for
 719  * incoming message.
 720  */
 721 int
 722 lnet_delay_rule_add(struct lnet_fault_attr *attr)
 723 {
 724         struct lnet_delay_rule *rule;
 725         int rc = 0;
 726
 727         if (attr->u.delay.la_rate & attr->u.delay.la_interval) {
 728                 CDEBUG(D_NET, "please provide either delay rate or delay interval, but not both at the same time %d/%d\n",
 729                        attr->u.delay.la_rate, attr->u.delay.la_interval);
 730                 return -EINVAL;
 731         }
 732
 733         if (!attr->u.delay.la_latency) {
 734                 CDEBUG(D_NET, "delay latency cannot be zero\n");
 735                 return -EINVAL;
 736         }
 737
 738         if (lnet_fault_attr_validate(attr))
 739                 return -EINVAL;
 740
 741         CFS_ALLOC_PTR(rule);
 742         if (!rule)
 743                 return -ENOMEM;
 744
 745         mutex_lock(&delay_dd.dd_mutex);
 746         if (!delay_dd.dd_running) {
 747                 struct task_struct *task;
 748
 749                 /**
 750                  *  NB: although LND threads will process delayed message
 751                  * in lnet_finalize, but there is no guarantee that LND
 752                  * threads will be waken up if no other message needs to
 753                  * be handled.
 754                  * Only one daemon thread, performance is not the concern
 755                  * of this simualation module.
 756                  */
 757                 task = kthread_run(lnet_delay_rule_daemon, NULL, "lnet_dd");
 758                 if (IS_ERR(task)) {
 759                         rc = PTR_ERR(task);
 760                         goto failed;
 761                 }
 762                 wait_event(delay_dd.dd_ctl_waitq, delay_dd.dd_running);
 763         }
 764
 765         timer_setup(&rule->dl_timer, delay_timer_cb, 0);
 766
 767         spin_lock_init(&rule->dl_lock);
 768         INIT_LIST_HEAD(&rule->dl_msg_list);
 769         INIT_LIST_HEAD(&rule->dl_sched_link);
 770
 771         rule->dl_attr = *attr;
 772         if (attr->u.delay.la_interval) {
 773                 rule->dl_time_base = cfs_time_shift(attr->u.delay.la_interval);
 774                 rule->dl_delay_time = cfs_time_shift(cfs_rand() %
 775                                                      attr->u.delay.la_interval);
 776         } else {
 777                 rule->dl_delay_at = cfs_rand() % attr->u.delay.la_rate;
 778         }
 779
 780         rule->dl_msg_send = -1;
 781
 782         lnet_net_lock(LNET_LOCK_EX);
 783         atomic_set(&rule->dl_refcount, 1);
 784         list_add(&rule->dl_link, &the_lnet.ln_delay_rules);
 785         lnet_net_unlock(LNET_LOCK_EX);
 786
 787         CDEBUG(D_NET, "Added delay rule: src %s, dst %s, rate %d\n",
 788                libcfs_nid2str(attr->fa_src), libcfs_nid2str(attr->fa_src),
 789                attr->u.delay.la_rate);
 790
 791         mutex_unlock(&delay_dd.dd_mutex);
 792         return 0;
 793 failed:
 794         mutex_unlock(&delay_dd.dd_mutex);
 795         CFS_FREE_PTR(rule);
 796         return rc;
 797 }
 798
 799 /**
 800  * Remove matched Delay Rules from lnet, if \a shutdown is true or both \a src
 801  * and \a dst are zero, all rules will be removed, otherwise only matched rules
 802  * will be removed.
 803  * If \a src is zero, then all rules have \a dst as destination will be remove
 804  * If \a dst is zero, then all rules have \a src as source will be removed
 805  *
 806  * When a delay rule is removed, all delayed messages of this rule will be
 807  * processed immediately.
 808  */
 809 int
 810 lnet_delay_rule_del(lnet_nid_t src, lnet_nid_t dst, bool shutdown)
 811 {
 812         struct lnet_delay_rule *rule;
 813         struct lnet_delay_rule *tmp;
 814         struct list_head rule_list;
 815         struct list_head msg_list;
 816         int n = 0;
 817         bool cleanup;
 818
 819         INIT_LIST_HEAD(&rule_list);
 820         INIT_LIST_HEAD(&msg_list);
 821
 822         if (shutdown) {
 823                 src = 0;
 824                 dst = 0;
 825         }
 826
 827         mutex_lock(&delay_dd.dd_mutex);
 828         lnet_net_lock(LNET_LOCK_EX);
 829
 830         list_for_each_entry_safe(rule, tmp, &the_lnet.ln_delay_rules, dl_link) {
 831                 if (rule->dl_attr.fa_src != src && src)
 832                         continue;
 833
 834                 if (rule->dl_attr.fa_dst != dst && dst)
 835                         continue;
 836
 837                 CDEBUG(D_NET, "Remove delay rule: src %s->dst: %s (1/%d, %d)\n",
 838                        libcfs_nid2str(rule->dl_attr.fa_src),
 839                        libcfs_nid2str(rule->dl_attr.fa_dst),
 840                        rule->dl_attr.u.delay.la_rate,
 841                        rule->dl_attr.u.delay.la_interval);
 842                 /* refcount is taken over by rule_list */
 843                 list_move(&rule->dl_link, &rule_list);
 844         }
 845
 846         /* check if we need to shutdown delay_daemon */
 847         cleanup = list_empty(&the_lnet.ln_delay_rules) &&
 848                   !list_empty(&rule_list);
 849         lnet_net_unlock(LNET_LOCK_EX);
 850
 851         list_for_each_entry_safe(rule, tmp, &rule_list, dl_link) {
 852                 list_del_init(&rule->dl_link);
 853
 854                 del_timer_sync(&rule->dl_timer);
 855                 delayed_msg_check(rule, true, &msg_list);
 856                 delay_rule_decref(rule); /* -1 for the_lnet.ln_delay_rules */
 857                 n++;
 858         }
 859
 860         if (cleanup) { /* no more delay rule, shutdown delay_daemon */
 861                 LASSERT(delay_dd.dd_running);
 862                 delay_dd.dd_running = 0;
 863                 wake_up(&delay_dd.dd_waitq);
 864
 865                 while (!delay_dd.dd_stopped)
 866                         wait_event(delay_dd.dd_ctl_waitq, delay_dd.dd_stopped);
 867         }
 868         mutex_unlock(&delay_dd.dd_mutex);
 869
 870         if (!list_empty(&msg_list))
 871                 delayed_msg_process(&msg_list, shutdown);
 872
 873         return n;
 874 }
 875
 876 /**
 877  * List Delay Rule at position of \a pos
 878  */
 879 int
 880 lnet_delay_rule_list(int pos, struct lnet_fault_attr *attr,
 881                      struct lnet_fault_stat *stat)
 882 {
 883         struct lnet_delay_rule *rule;
 884         int cpt;
 885         int i = 0;
 886         int rc = -ENOENT;
 887
 888         cpt = lnet_net_lock_current();
 889         list_for_each_entry(rule, &the_lnet.ln_delay_rules, dl_link) {
 890                 if (i++ < pos)
 891                         continue;
 892
 893                 spin_lock(&rule->dl_lock);
 894                 *attr = rule->dl_attr;
 895                 *stat = rule->dl_stat;
 896                 spin_unlock(&rule->dl_lock);
 897                 rc = 0;
 898                 break;
 899         }
 900
 901         lnet_net_unlock(cpt);
 902         return rc;
 903 }
 904
 905 /**
 906  * reset counters for all Delay Rules
 907  */
 908 void
 909 lnet_delay_rule_reset(void)
 910 {
 911         struct lnet_delay_rule *rule;
 912         int cpt;
 913
 914         cpt = lnet_net_lock_current();
 915
 916         list_for_each_entry(rule, &the_lnet.ln_delay_rules, dl_link) {
 917                 struct lnet_fault_attr *attr = &rule->dl_attr;
 918
 919                 spin_lock(&rule->dl_lock);
 920
 921                 memset(&rule->dl_stat, 0, sizeof(rule->dl_stat));
 922                 if (attr->u.delay.la_rate) {
 923                         rule->dl_delay_at = cfs_rand() % attr->u.delay.la_rate;
 924                 } else {
 925                         rule->dl_delay_time = cfs_time_shift(cfs_rand() %
 926                                                 attr->u.delay.la_interval);
 927                         rule->dl_time_base = cfs_time_shift(attr->u.delay.la_interval);
 928                 }
 929                 spin_unlock(&rule->dl_lock);
 930         }
 931
 932         lnet_net_unlock(cpt);
 933 }
 934
 935 int
 936 lnet_fault_ctl(int opc, struct libcfs_ioctl_data *data)
 937 {
 938         struct lnet_fault_attr *attr;
 939         struct lnet_fault_stat *stat;
 940
 941         attr = (struct lnet_fault_attr *)data->ioc_inlbuf1;
 942
 943         switch (opc) {
 944         default:
 945                 return -EINVAL;
 946
 947         case LNET_CTL_DROP_ADD:
 948                 if (!attr)
 949                         return -EINVAL;
 950
 951                 return lnet_drop_rule_add(attr);
 952
 953         case LNET_CTL_DROP_DEL:
 954                 if (!attr)
 955                         return -EINVAL;
 956
 957                 data->ioc_count = lnet_drop_rule_del(attr->fa_src,
 958                                                      attr->fa_dst);
 959                 return 0;
 960
 961         case LNET_CTL_DROP_RESET:
 962                 lnet_drop_rule_reset();
 963                 return 0;
 964
 965         case LNET_CTL_DROP_LIST:
 966                 stat = (struct lnet_fault_stat *)data->ioc_inlbuf2;
 967                 if (!attr || !stat)
 968                         return -EINVAL;
 969
 970                 return lnet_drop_rule_list(data->ioc_count, attr, stat);
 971
 972         case LNET_CTL_DELAY_ADD:
 973                 if (!attr)
 974                         return -EINVAL;
 975
 976                 return lnet_delay_rule_add(attr);
 977
 978         case LNET_CTL_DELAY_DEL:
 979                 if (!attr)
 980                         return -EINVAL;
 981
 982                 data->ioc_count = lnet_delay_rule_del(attr->fa_src,
 983                                                       attr->fa_dst, false);
 984                 return 0;
 985
 986         case LNET_CTL_DELAY_RESET:
 987                 lnet_delay_rule_reset();
 988                 return 0;
 989
 990         case LNET_CTL_DELAY_LIST:
 991                 stat = (struct lnet_fault_stat *)data->ioc_inlbuf2;
 992                 if (!attr || !stat)
 993                         return -EINVAL;
 994
 995                 return lnet_delay_rule_list(data->ioc_count, attr, stat);
 996         }
 997 }
 998
 999 int
1000 lnet_fault_init(void)
1001 {
1002         BUILD_BUG_ON(LNET_PUT_BIT != 1 << LNET_MSG_PUT);
1003         BUILD_BUG_ON(LNET_ACK_BIT != 1 << LNET_MSG_ACK);
1004         BUILD_BUG_ON(LNET_GET_BIT != 1 << LNET_MSG_GET);
1005         BUILD_BUG_ON(LNET_REPLY_BIT != 1 << LNET_MSG_REPLY);
1006
1007         mutex_init(&delay_dd.dd_mutex);
1008         spin_lock_init(&delay_dd.dd_lock);
1009         init_waitqueue_head(&delay_dd.dd_waitq);
1010         init_waitqueue_head(&delay_dd.dd_ctl_waitq);
1011         INIT_LIST_HEAD(&delay_dd.dd_sched_rules);
1012
1013         return 0;
1014 }
1015
1016 void
1017 lnet_fault_fini(void)
1018 {
1019         lnet_drop_rule_del(0, 0);
1020         lnet_delay_rule_del(0, 0, true);
1021
1022         LASSERT(list_empty(&the_lnet.ln_drop_rules));
1023         LASSERT(list_empty(&the_lnet.ln_delay_rules));
1024         LASSERT(list_empty(&delay_dd.dd_sched_rules));
1025 }