lustre/osd-zfs/osd_scrub.c

   1 /*
   2  * GPL HEADER START
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 only,
   8  * as published by the Free Software Foundation.
   9
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  * GNU General Public License version 2 for more details.  A copy is
  14  * included in the COPYING file that accompanied this code.
  15
  16  * You should have received a copy of the GNU General Public License
  17  * along with this program; if not, write to the Free Software
  18  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
  19  *
  20  * GPL HEADER END
  21  */
  22 /*
  23  * Copyright (c) 2017, Intel Corporation.
  24  */
  25 /*
  26  * lustre/osd-zfs/osd_scrub.c
  27  *
  28  * Top-level entry points into osd module
  29  *
  30  * The OI scrub is used for rebuilding Object Index files when restores MDT from
  31  * file-level backup.
  32  *
  33  * The otable based iterator scans ZFS objects to feed up layer LFSCK.
  34  *
  35  * Author: Fan Yong <fan.yong@intel.com>
  36  */
  37
  38 #define DEBUG_SUBSYSTEM S_LFSCK
  39
  40 #include <linux/kthread.h>
  41 #include <uapi/linux/lustre/lustre_idl.h>
  42 #include <lustre_disk.h>
  43 #include <dt_object.h>
  44 #include <linux/xattr.h>
  45 #include <lustre_scrub.h>
  46 #include <obd_class.h>
  47 #include <lustre_nodemap.h>
  48 #include <sys/dsl_dataset.h>
  49 #include <sys/zap_impl.h>
  50 #include <sys/zap.h>
  51 #include <sys/zap_leaf.h>
  52
  53 #include "osd_internal.h"
  54
  55 #define OSD_OTABLE_MAX_HASH             ((1ULL << 48) - 1)
  56 #define OTABLE_PREFETCH                 256
  57
  58 #define DTO_INDEX_INSERT                1
  59 #define DTO_INDEX_DELETE                2
  60 #define DTO_INDEX_UPDATE                3
  61
  62 static inline bool osd_scrub_has_window(struct osd_otable_it *it)
  63 {
  64         return it->ooi_prefetched < OTABLE_PREFETCH;
  65 }
  66
  67 /**
  68  * update/insert/delete the specified OI mapping (@fid @id) according to the ops
  69  *
  70  * \retval   1, changed nothing
  71  * \retval   0, changed successfully
  72  * \retval -ve, on error
  73  */
  74 static int osd_scrub_refresh_mapping(const struct lu_env *env,
  75                                      struct osd_device *dev,
  76                                      const struct lu_fid *fid,
  77                                      uint64_t oid, int ops,
  78                                      bool force, const char *name)
  79 {
  80         struct osd_thread_info *info = osd_oti_get(env);
  81         struct zpl_direntry *zde = &info->oti_zde.lzd_reg;
  82         char *buf = info->oti_str;
  83         dmu_tx_t *tx = NULL;
  84         dnode_t *dn = NULL;
  85         uint64_t zapid;
  86         int rc;
  87         ENTRY;
  88
  89         if (dev->od_scrub.os_file.sf_param & SP_DRYRUN && !force)
  90                 GOTO(log, rc = 0);
  91
  92         tx = dmu_tx_create(dev->od_os);
  93         if (!tx)
  94                 GOTO(log, rc = -ENOMEM);
  95
  96         zapid = osd_get_name_n_idx(env, dev, fid, buf,
  97                                    sizeof(info->oti_str), &dn);
  98         osd_tx_hold_zap(tx, zapid, dn,
  99                         ops == DTO_INDEX_INSERT ? TRUE : FALSE, NULL);
 100         rc = -dmu_tx_assign(tx, TXG_WAIT);
 101         if (rc) {
 102                 dmu_tx_abort(tx);
 103                 GOTO(log, rc);
 104         }
 105
 106         switch (ops) {
 107         case DTO_INDEX_UPDATE:
 108                 zde->zde_pad = 0;
 109                 zde->zde_dnode = oid;
 110                 zde->zde_type = 0; /* The type in OI mapping is useless. */
 111                 rc = -zap_update(dev->od_os, zapid, buf, 8, sizeof(*zde) / 8,
 112                                  zde, tx);
 113                 if (unlikely(rc == -ENOENT)) {
 114                         /* Some unlink thread may removed the OI mapping. */
 115                         rc = 1;
 116                 }
 117                 break;
 118         case DTO_INDEX_INSERT:
 119                 zde->zde_pad = 0;
 120                 zde->zde_dnode = oid;
 121                 zde->zde_type = 0; /* The type in OI mapping is useless. */
 122                 rc = osd_zap_add(dev, zapid, dn, buf, 8, sizeof(*zde) / 8,
 123                                  zde, tx);
 124                 if (unlikely(rc == -EEXIST))
 125                         rc = 1;
 126                 break;
 127         case DTO_INDEX_DELETE:
 128                 rc = osd_zap_remove(dev, zapid, dn, buf, tx);
 129                 if (rc == -ENOENT) {
 130                         /* It is normal that the unlink thread has removed the
 131                          * OI mapping already. */
 132                         rc = 1;
 133                 }
 134                 break;
 135         default:
 136                 LASSERTF(0, "Unexpected ops %d\n", ops);
 137                 rc = -EINVAL;
 138                 break;
 139         }
 140
 141         dmu_tx_commit(tx);
 142         GOTO(log, rc);
 143
 144 log:
 145         CDEBUG(D_LFSCK, "%s: refresh OI map for scrub, op %d, force %s, "
 146                DFID" => %llu (%s): rc = %d\n", osd_name(dev), ops,
 147                force ? "yes" : "no", PFID(fid), oid, name ? name : "null", rc);
 148
 149         return rc;
 150 }
 151
 152 static int
 153 osd_scrub_check_update(const struct lu_env *env, struct osd_device *dev,
 154                        const struct lu_fid *fid, uint64_t oid, int val)
 155 {
 156         struct lustre_scrub *scrub = &dev->od_scrub;
 157         struct scrub_file *sf = &scrub->os_file;
 158         struct osd_inconsistent_item *oii = NULL;
 159         nvlist_t *nvbuf = NULL;
 160         dnode_t *dn = NULL;
 161         uint64_t oid2;
 162         int ops = DTO_INDEX_UPDATE;
 163         int rc;
 164         ENTRY;
 165
 166         down_write(&scrub->os_rwsem);
 167         scrub->os_new_checked++;
 168         if (val < 0)
 169                 GOTO(out, rc = val);
 170
 171         if (scrub->os_in_prior)
 172                 oii = list_entry(scrub->os_inconsistent_items.next,
 173                                  struct osd_inconsistent_item, oii_list);
 174
 175         if (oid < sf->sf_pos_latest_start && !oii)
 176                 GOTO(out, rc = 0);
 177
 178         if (oii && oii->oii_insert) {
 179                 ops = DTO_INDEX_INSERT;
 180                 goto zget;
 181         }
 182
 183         rc = osd_fid_lookup(env, dev, fid, &oid2);
 184         if (rc) {
 185                 if (rc != -ENOENT)
 186                         GOTO(out, rc);
 187
 188                 ops = DTO_INDEX_INSERT;
 189
 190 zget:
 191                 rc = __osd_obj2dnode(dev->od_os, oid, &dn);
 192                 if (rc) {
 193                         /* Someone removed the object by race. */
 194                         if (rc == -ENOENT || rc == -EEXIST)
 195                                 rc = 0;
 196                         GOTO(out, rc);
 197                 }
 198
 199                 scrub->os_full_speed = 1;
 200                 sf->sf_flags |= SF_INCONSISTENT;
 201         } else if (oid == oid2) {
 202                 GOTO(out, rc = 0);
 203         } else {
 204                 struct lustre_mdt_attrs *lma = NULL;
 205                 int size;
 206
 207                 rc = __osd_xattr_load_by_oid(dev, oid2, &nvbuf);
 208                 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
 209                         goto update;
 210                 if (rc)
 211                         GOTO(out, rc);
 212
 213                 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
 214                                                (uchar_t **)&lma, &size);
 215                 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
 216                         goto update;
 217                 if (rc)
 218                         GOTO(out, rc);
 219
 220                 lustre_lma_swab(lma);
 221                 if (unlikely(lu_fid_eq(&lma->lma_self_fid, fid))) {
 222                         CDEBUG(D_LFSCK, "%s: the FID "DFID" is used by "
 223                                "two objects: %llu and %llu (in OI)\n",
 224                                osd_name(dev), PFID(fid), oid, oid2);
 225
 226                         GOTO(out, rc = -EEXIST);
 227                 }
 228
 229 update:
 230                 scrub->os_full_speed = 1;
 231                 sf->sf_flags |= SF_INCONSISTENT;
 232         }
 233
 234         rc = osd_scrub_refresh_mapping(env, dev, fid, oid, ops, false, NULL);
 235         if (!rc) {
 236                 if (scrub->os_in_prior)
 237                         sf->sf_items_updated_prior++;
 238                 else
 239                         sf->sf_items_updated++;
 240         }
 241
 242         GOTO(out, rc);
 243
 244 out:
 245         if (nvbuf)
 246                 nvlist_free(nvbuf);
 247
 248         if (rc < 0) {
 249                 sf->sf_items_failed++;
 250                 if (sf->sf_pos_first_inconsistent == 0 ||
 251                     sf->sf_pos_first_inconsistent > oid)
 252                         sf->sf_pos_first_inconsistent = oid;
 253         } else {
 254                 rc = 0;
 255         }
 256
 257         /* There may be conflict unlink during the OI scrub,
 258          * if happend, then remove the new added OI mapping. */
 259         if (ops == DTO_INDEX_INSERT && dn && dn->dn_free_txg)
 260                 osd_scrub_refresh_mapping(env, dev, fid, oid,
 261                                           DTO_INDEX_DELETE, false, NULL);
 262         up_write(&scrub->os_rwsem);
 263
 264         if (dn)
 265                 osd_dnode_rele(dn);
 266
 267         if (oii) {
 268                 spin_lock(&scrub->os_lock);
 269                 if (likely(!list_empty(&oii->oii_list)))
 270                         list_del(&oii->oii_list);
 271                 spin_unlock(&scrub->os_lock);
 272                 OBD_FREE_PTR(oii);
 273         }
 274
 275         RETURN(sf->sf_param & SP_FAILOUT ? rc : 0);
 276 }
 277
 278 static int osd_scrub_prep(const struct lu_env *env, struct osd_device *dev)
 279 {
 280         struct lustre_scrub *scrub = &dev->od_scrub;
 281         struct ptlrpc_thread *thread = &scrub->os_thread;
 282         struct scrub_file *sf = &scrub->os_file;
 283         __u32 flags = scrub->os_start_flags;
 284         int rc;
 285         bool drop_dryrun = false;
 286         ENTRY;
 287
 288         CDEBUG(D_LFSCK, "%s: OI scrub prep, flags = 0x%x\n",
 289                scrub->os_name, flags);
 290
 291         down_write(&scrub->os_rwsem);
 292         if (flags & SS_SET_FAILOUT)
 293                 sf->sf_param |= SP_FAILOUT;
 294         else if (flags & SS_CLEAR_FAILOUT)
 295                 sf->sf_param &= ~SP_FAILOUT;
 296
 297         if (flags & SS_SET_DRYRUN) {
 298                 sf->sf_param |= SP_DRYRUN;
 299         } else if (flags & SS_CLEAR_DRYRUN && sf->sf_param & SP_DRYRUN) {
 300                 sf->sf_param &= ~SP_DRYRUN;
 301                 drop_dryrun = true;
 302         }
 303
 304         if (flags & SS_RESET)
 305                 scrub_file_reset(scrub, dev->od_uuid, 0);
 306
 307         scrub->os_partial_scan = 0;
 308         if (flags & SS_AUTO_FULL) {
 309                 scrub->os_full_speed = 1;
 310                 sf->sf_flags |= SF_AUTO;
 311         } else if (sf->sf_flags & (SF_RECREATED | SF_INCONSISTENT |
 312                                    SF_UPGRADE)) {
 313                 scrub->os_full_speed = 1;
 314         } else {
 315                 scrub->os_full_speed = 0;
 316         }
 317
 318         spin_lock(&scrub->os_lock);
 319         scrub->os_in_prior = 0;
 320         scrub->os_waiting = 0;
 321         scrub->os_paused = 0;
 322         scrub->os_in_join = 0;
 323         scrub->os_full_scrub = 0;
 324         spin_unlock(&scrub->os_lock);
 325         scrub->os_new_checked = 0;
 326         if (drop_dryrun && sf->sf_pos_first_inconsistent != 0)
 327                 sf->sf_pos_latest_start = sf->sf_pos_first_inconsistent;
 328         else if (sf->sf_pos_last_checkpoint != 0)
 329                 sf->sf_pos_latest_start = sf->sf_pos_last_checkpoint + 1;
 330         else
 331                 sf->sf_pos_latest_start = 1;
 332
 333         scrub->os_pos_current = sf->sf_pos_latest_start;
 334         sf->sf_status = SS_SCANNING;
 335         sf->sf_time_latest_start = ktime_get_real_seconds();
 336         sf->sf_time_last_checkpoint = sf->sf_time_latest_start;
 337         sf->sf_pos_last_checkpoint = sf->sf_pos_latest_start - 1;
 338         rc = scrub_file_store(env, scrub);
 339         if (!rc) {
 340                 spin_lock(&scrub->os_lock);
 341                 thread_set_flags(thread, SVC_RUNNING);
 342                 spin_unlock(&scrub->os_lock);
 343                 wake_up_all(&thread->t_ctl_waitq);
 344         }
 345         up_write(&scrub->os_rwsem);
 346
 347         RETURN(rc);
 348 }
 349
 350 static int osd_scrub_post(const struct lu_env *env, struct osd_device *dev,
 351                           int result)
 352 {
 353         struct lustre_scrub *scrub = &dev->od_scrub;
 354         struct scrub_file *sf = &scrub->os_file;
 355         int rc;
 356         ENTRY;
 357
 358         CDEBUG(D_LFSCK, "%s: OI scrub post with result = %d\n",
 359                scrub->os_name, result);
 360
 361         down_write(&scrub->os_rwsem);
 362         spin_lock(&scrub->os_lock);
 363         thread_set_flags(&scrub->os_thread, SVC_STOPPING);
 364         spin_unlock(&scrub->os_lock);
 365         if (scrub->os_new_checked > 0) {
 366                 sf->sf_items_checked += scrub->os_new_checked;
 367                 scrub->os_new_checked = 0;
 368                 sf->sf_pos_last_checkpoint = scrub->os_pos_current;
 369         }
 370         sf->sf_time_last_checkpoint = ktime_get_real_seconds();
 371         if (result > 0) {
 372                 sf->sf_status = SS_COMPLETED;
 373                 if (!(sf->sf_param & SP_DRYRUN)) {
 374                         memset(sf->sf_oi_bitmap, 0, SCRUB_OI_BITMAP_SIZE);
 375                         sf->sf_flags &= ~(SF_RECREATED | SF_INCONSISTENT |
 376                                           SF_UPGRADE | SF_AUTO);
 377                 }
 378                 sf->sf_time_last_complete = sf->sf_time_last_checkpoint;
 379                 sf->sf_success_count++;
 380         } else if (result == 0) {
 381                 if (scrub->os_paused)
 382                         sf->sf_status = SS_PAUSED;
 383                 else
 384                         sf->sf_status = SS_STOPPED;
 385         } else {
 386                 sf->sf_status = SS_FAILED;
 387         }
 388         sf->sf_run_time += ktime_get_seconds() -
 389                            scrub->os_time_last_checkpoint;
 390
 391         rc = scrub_file_store(env, scrub);
 392         up_write(&scrub->os_rwsem);
 393
 394         RETURN(rc < 0 ? rc : result);
 395 }
 396
 397 /* iteration engine */
 398
 399 static inline int
 400 osd_scrub_wakeup(struct lustre_scrub *scrub, struct osd_otable_it *it)
 401 {
 402         spin_lock(&scrub->os_lock);
 403         if (osd_scrub_has_window(it) ||
 404             !list_empty(&scrub->os_inconsistent_items) ||
 405             it->ooi_waiting || !thread_is_running(&scrub->os_thread))
 406                 scrub->os_waiting = 0;
 407         else
 408                 scrub->os_waiting = 1;
 409         spin_unlock(&scrub->os_lock);
 410
 411         return !scrub->os_waiting;
 412 }
 413
 414 static int osd_scrub_next(const struct lu_env *env, struct osd_device *dev,
 415                           struct lu_fid *fid, uint64_t *oid)
 416 {
 417         struct lustre_scrub *scrub = &dev->od_scrub;
 418         struct ptlrpc_thread *thread = &scrub->os_thread;
 419         struct osd_otable_it *it = dev->od_otable_it;
 420         struct lustre_mdt_attrs *lma = NULL;
 421         nvlist_t *nvbuf = NULL;
 422         int size = 0;
 423         int rc = 0;
 424         ENTRY;
 425
 426         if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_DELAY) && cfs_fail_val > 0) {
 427                 wait_event_idle_timeout(
 428                         thread->t_ctl_waitq,
 429                         !list_empty(&scrub->os_inconsistent_items) ||
 430                         !thread_is_running(thread),
 431                         cfs_time_seconds(cfs_fail_val));
 432
 433                 if (unlikely(!thread_is_running(thread)))
 434                         RETURN(SCRUB_NEXT_EXIT);
 435         }
 436
 437         if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_CRASH)) {
 438                 spin_lock(&scrub->os_lock);
 439                 thread_set_flags(thread, SVC_STOPPING);
 440                 spin_unlock(&scrub->os_lock);
 441                 RETURN(SCRUB_NEXT_CRASH);
 442         }
 443
 444         if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_FATAL))
 445                 RETURN(SCRUB_NEXT_FATAL);
 446
 447 again:
 448         if (nvbuf) {
 449                 nvlist_free(nvbuf);
 450                 nvbuf = NULL;
 451                 lma = NULL;
 452         }
 453
 454         if (!list_empty(&scrub->os_inconsistent_items)) {
 455                 spin_lock(&scrub->os_lock);
 456                 if (likely(!list_empty(&scrub->os_inconsistent_items))) {
 457                         struct osd_inconsistent_item *oii;
 458
 459                         oii = list_entry(scrub->os_inconsistent_items.next,
 460                                 struct osd_inconsistent_item, oii_list);
 461                         *fid = oii->oii_cache.oic_fid;
 462                         *oid = oii->oii_cache.oic_dnode;
 463                         scrub->os_in_prior = 1;
 464                         spin_unlock(&scrub->os_lock);
 465
 466                         GOTO(out, rc = 0);
 467                 }
 468                 spin_unlock(&scrub->os_lock);
 469         }
 470
 471         if (!scrub->os_full_speed && !osd_scrub_has_window(it))
 472                 wait_event_idle(thread->t_ctl_waitq,
 473                                 osd_scrub_wakeup(scrub, it));
 474
 475         if (unlikely(!thread_is_running(thread)))
 476                 GOTO(out, rc = SCRUB_NEXT_EXIT);
 477
 478         rc = -dmu_object_next(dev->od_os, &scrub->os_pos_current, B_FALSE, 0);
 479         if (rc)
 480                 GOTO(out, rc = (rc == -ESRCH ? SCRUB_NEXT_BREAK : rc));
 481
 482         rc = __osd_xattr_load_by_oid(dev, scrub->os_pos_current, &nvbuf);
 483         if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
 484                 goto again;
 485
 486         if (rc)
 487                 GOTO(out, rc);
 488
 489         LASSERT(nvbuf != NULL);
 490         rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
 491                                        (uchar_t **)&lma, &size);
 492         if (!rc) {
 493                 lustre_lma_swab(lma);
 494                 if (likely(!(lma->lma_compat & LMAC_NOT_IN_OI) &&
 495                            !(lma->lma_incompat & LMAI_AGENT))) {
 496                         *fid = lma->lma_self_fid;
 497                         *oid = scrub->os_pos_current;
 498
 499                         GOTO(out, rc = 0);
 500                 }
 501         }
 502
 503         if (!scrub->os_full_speed) {
 504                 spin_lock(&scrub->os_lock);
 505                 it->ooi_prefetched++;
 506                 if (it->ooi_waiting) {
 507                         it->ooi_waiting = 0;
 508                         wake_up_all(&thread->t_ctl_waitq);
 509                 }
 510                 spin_unlock(&scrub->os_lock);
 511         }
 512
 513         goto again;
 514
 515 out:
 516         if (nvbuf)
 517                 nvlist_free(nvbuf);
 518
 519         return rc;
 520 }
 521
 522 static int osd_scrub_exec(const struct lu_env *env, struct osd_device *dev,
 523                           const struct lu_fid *fid, uint64_t oid, int rc)
 524 {
 525         struct lustre_scrub *scrub = &dev->od_scrub;
 526         struct ptlrpc_thread *thread = &scrub->os_thread;
 527         struct osd_otable_it *it = dev->od_otable_it;
 528
 529         rc = osd_scrub_check_update(env, dev, fid, oid, rc);
 530         if (!scrub->os_in_prior) {
 531                 if (!scrub->os_full_speed) {
 532                         spin_lock(&scrub->os_lock);
 533                         it->ooi_prefetched++;
 534                         if (it->ooi_waiting) {
 535                                 it->ooi_waiting = 0;
 536                                 wake_up_all(&thread->t_ctl_waitq);
 537                         }
 538                         spin_unlock(&scrub->os_lock);
 539                 }
 540         } else {
 541                 scrub->os_in_prior = 0;
 542         }
 543
 544         if (rc)
 545                 return rc;
 546
 547         rc = scrub_checkpoint(env, scrub);
 548         if (rc) {
 549                 CDEBUG(D_LFSCK, "%s: fail to checkpoint, pos = %llu: "
 550                        "rc = %d\n", scrub->os_name, scrub->os_pos_current, rc);
 551                 /* Continue, as long as the scrub itself can go ahead. */
 552         }
 553
 554         return 0;
 555 }
 556
 557 static int osd_scrub_main(void *args)
 558 {
 559         struct lu_env env;
 560         struct osd_device *dev = (struct osd_device *)args;
 561         struct lustre_scrub *scrub = &dev->od_scrub;
 562         struct ptlrpc_thread *thread = &scrub->os_thread;
 563         struct lu_fid *fid;
 564         uint64_t oid;
 565         int rc = 0;
 566         ENTRY;
 567
 568         rc = lu_env_init(&env, LCT_LOCAL | LCT_DT_THREAD);
 569         if (rc) {
 570                 CDEBUG(D_LFSCK, "%s: OI scrub fail to init env: rc = %d\n",
 571                        scrub->os_name, rc);
 572                 GOTO(noenv, rc);
 573         }
 574
 575         rc = osd_scrub_prep(&env, dev);
 576         if (rc) {
 577                 CDEBUG(D_LFSCK, "%s: OI scrub fail to scrub prep: rc = %d\n",
 578                        scrub->os_name, rc);
 579                 GOTO(out, rc);
 580         }
 581
 582         if (!scrub->os_full_speed) {
 583                 struct osd_otable_it *it = dev->od_otable_it;
 584
 585                 wait_event_idle(thread->t_ctl_waitq,
 586                                 it->ooi_user_ready ||
 587                                 !thread_is_running(thread));
 588
 589                 if (unlikely(!thread_is_running(thread)))
 590                         GOTO(post, rc = 0);
 591
 592                 scrub->os_pos_current = it->ooi_pos;
 593         }
 594
 595         CDEBUG(D_LFSCK, "%s: OI scrub start, flags = 0x%x, pos = %llu\n",
 596                scrub->os_name, scrub->os_start_flags,
 597                scrub->os_pos_current);
 598
 599         fid = &osd_oti_get(&env)->oti_fid;
 600         while (!rc && thread_is_running(thread)) {
 601                 rc = osd_scrub_next(&env, dev, fid, &oid);
 602                 switch (rc) {
 603                 case SCRUB_NEXT_EXIT:
 604                         GOTO(post, rc = 0);
 605                 case SCRUB_NEXT_CRASH:
 606                         spin_lock(&scrub->os_lock);
 607                         thread_set_flags(&scrub->os_thread, SVC_STOPPING);
 608                         spin_unlock(&scrub->os_lock);
 609                         GOTO(out, rc = -EINVAL);
 610                 case SCRUB_NEXT_FATAL:
 611                         GOTO(post, rc = -EINVAL);
 612                 case SCRUB_NEXT_BREAK:
 613                         GOTO(post, rc = 1);
 614                 }
 615
 616                 rc = osd_scrub_exec(&env, dev, fid, oid, rc);
 617         }
 618
 619         GOTO(post, rc);
 620
 621 post:
 622         rc = osd_scrub_post(&env, dev, rc);
 623         CDEBUG(D_LFSCK, "%s: OI scrub: stop, pos = %llu: rc = %d\n",
 624                scrub->os_name, scrub->os_pos_current, rc);
 625
 626 out:
 627         while (!list_empty(&scrub->os_inconsistent_items)) {
 628                 struct osd_inconsistent_item *oii;
 629
 630                 oii = list_entry(scrub->os_inconsistent_items.next,
 631                                  struct osd_inconsistent_item, oii_list);
 632                 list_del_init(&oii->oii_list);
 633                 OBD_FREE_PTR(oii);
 634         }
 635
 636         lu_env_fini(&env);
 637
 638 noenv:
 639         spin_lock(&scrub->os_lock);
 640         thread_set_flags(thread, SVC_STOPPED);
 641         wake_up_all(&thread->t_ctl_waitq);
 642         spin_unlock(&scrub->os_lock);
 643         return rc;
 644 }
 645
 646 /* initial OI scrub */
 647
 648 struct osd_lf_map;
 649
 650 typedef int (*handle_dirent_t)(const struct lu_env *, struct osd_device *,
 651                                const char *, uint64_t, uint64_t,
 652                                enum osd_lf_flags, bool);
 653 static int osd_ios_varfid_hd(const struct lu_env *, struct osd_device *,
 654                              const char *, uint64_t, uint64_t,
 655                              enum osd_lf_flags, bool);
 656 static int osd_ios_uld_hd(const struct lu_env *, struct osd_device *,
 657                           const char *, uint64_t, uint64_t,
 658                           enum osd_lf_flags, bool);
 659
 660 typedef int (*scan_dir_t)(const struct lu_env *, struct osd_device *,
 661                           uint64_t, handle_dirent_t, enum osd_lf_flags);
 662 static int osd_ios_general_sd(const struct lu_env *, struct osd_device *,
 663                               uint64_t, handle_dirent_t, enum osd_lf_flags);
 664 static int osd_ios_ROOT_sd(const struct lu_env *, struct osd_device *,
 665                            uint64_t, handle_dirent_t, enum osd_lf_flags);
 666
 667 struct osd_lf_map {
 668         char                    *olm_name;
 669         struct lu_fid            olm_fid;
 670         enum osd_lf_flags        olm_flags;
 671         scan_dir_t               olm_scan_dir;
 672         handle_dirent_t          olm_handle_dirent;
 673 };
 674
 675 /* Add the new introduced local files in the list in the future. */
 676 static const struct osd_lf_map osd_lf_maps[] = {
 677         /* CONFIGS */
 678         {
 679                 .olm_name               = MOUNT_CONFIGS_DIR,
 680                 .olm_fid                = {
 681                         .f_seq  = FID_SEQ_LOCAL_FILE,
 682                         .f_oid  = MGS_CONFIGS_OID,
 683                 },
 684                 .olm_flags              = OLF_SCAN_SUBITEMS,
 685                 .olm_scan_dir           = osd_ios_general_sd,
 686                 .olm_handle_dirent      = osd_ios_varfid_hd,
 687         },
 688
 689         /* NIDTBL_VERSIONS */
 690         {
 691                 .olm_name               = MGS_NIDTBL_DIR,
 692                 .olm_flags              = OLF_SCAN_SUBITEMS,
 693                 .olm_scan_dir           = osd_ios_general_sd,
 694                 .olm_handle_dirent      = osd_ios_varfid_hd,
 695         },
 696
 697         /* PENDING */
 698         {
 699                 .olm_name               = MDT_ORPHAN_DIR,
 700         },
 701
 702         /* ROOT */
 703         {
 704                 .olm_name               = "ROOT",
 705                 .olm_fid                = {
 706                         .f_seq  = FID_SEQ_ROOT,
 707                         .f_oid  = FID_OID_ROOT,
 708                 },
 709                 .olm_flags              = OLF_SCAN_SUBITEMS,
 710                 .olm_scan_dir           = osd_ios_ROOT_sd,
 711         },
 712
 713         /* fld */
 714         {
 715                 .olm_name               = "fld",
 716                 .olm_fid                = {
 717                         .f_seq  = FID_SEQ_LOCAL_FILE,
 718                         .f_oid  = FLD_INDEX_OID,
 719                 },
 720         },
 721
 722         /* changelog_catalog */
 723         {
 724                 .olm_name               = CHANGELOG_CATALOG,
 725         },
 726
 727         /* changelog_users */
 728         {
 729                 .olm_name               = CHANGELOG_USERS,
 730         },
 731
 732         /* quota_master */
 733         {
 734                 .olm_name               = QMT_DIR,
 735                 .olm_flags              = OLF_SCAN_SUBITEMS,
 736                 .olm_scan_dir           = osd_ios_general_sd,
 737                 .olm_handle_dirent      = osd_ios_varfid_hd,
 738         },
 739
 740         /* quota_slave */
 741         {
 742                 .olm_name               = QSD_DIR,
 743                 .olm_flags              = OLF_SCAN_SUBITEMS,
 744                 .olm_scan_dir           = osd_ios_general_sd,
 745                 .olm_handle_dirent      = osd_ios_varfid_hd,
 746         },
 747
 748         /* LFSCK */
 749         {
 750                 .olm_name               = LFSCK_DIR,
 751                 .olm_flags              = OLF_SCAN_SUBITEMS | OLF_NOT_BACKUP,
 752                 .olm_scan_dir           = osd_ios_general_sd,
 753                 .olm_handle_dirent      = osd_ios_varfid_hd,
 754         },
 755
 756         /* lfsck_bookmark */
 757         {
 758                 .olm_name               = LFSCK_BOOKMARK,
 759         },
 760
 761         /* lfsck_layout */
 762         {
 763                 .olm_name               = LFSCK_LAYOUT,
 764         },
 765
 766         /* lfsck_namespace */
 767         {
 768                 .olm_name               = LFSCK_NAMESPACE,
 769         },
 770
 771         /* OSP update logs update_log{_dir} use f_seq = FID_SEQ_UPDATE_LOG{_DIR}
 772          * and f_oid = index for their log files.  See lu_update_log{_dir}_fid()
 773          * for more details. */
 774
 775         /* update_log */
 776         {
 777                 .olm_name               = "update_log",
 778                 .olm_fid                = {
 779                         .f_seq  = FID_SEQ_UPDATE_LOG,
 780                 },
 781                 .olm_flags              = OLF_IDX_IN_FID,
 782         },
 783
 784         /* update_log_dir */
 785         {
 786                 .olm_name               = "update_log_dir",
 787                 .olm_fid        = {
 788                         .f_seq  = FID_SEQ_UPDATE_LOG_DIR,
 789                 },
 790                 .olm_flags              = OLF_SCAN_SUBITEMS | OLF_IDX_IN_FID,
 791                 .olm_scan_dir           = osd_ios_general_sd,
 792                 .olm_handle_dirent      = osd_ios_uld_hd,
 793         },
 794
 795         /* hsm_actions */
 796         {
 797                 .olm_name               = HSM_ACTIONS,
 798         },
 799
 800         /* nodemap */
 801         {
 802                 .olm_name               = LUSTRE_NODEMAP_NAME,
 803         },
 804
 805         /* index_backup */
 806         {
 807                 .olm_name               = INDEX_BACKUP_DIR,
 808                 .olm_fid                = {
 809                         .f_seq  = FID_SEQ_LOCAL_FILE,
 810                         .f_oid  = INDEX_BACKUP_OID,
 811                 },
 812                 .olm_flags              = OLF_SCAN_SUBITEMS | OLF_NOT_BACKUP,
 813                 .olm_scan_dir           = osd_ios_general_sd,
 814                 .olm_handle_dirent      = osd_ios_varfid_hd,
 815         },
 816
 817         {
 818                 .olm_name               = NULL
 819         }
 820 };
 821
 822 /* Add the new introduced files under .lustre/ in the list in the future. */
 823 static const struct osd_lf_map osd_dl_maps[] = {
 824         /* .lustre/fid */
 825         {
 826                 .olm_name               = "fid",
 827                 .olm_fid                = {
 828                         .f_seq  = FID_SEQ_DOT_LUSTRE,
 829                         .f_oid  = FID_OID_DOT_LUSTRE_OBF,
 830                 },
 831         },
 832
 833         /* .lustre/lost+found */
 834         {
 835                 .olm_name               = "lost+found",
 836                 .olm_fid                = {
 837                         .f_seq  = FID_SEQ_DOT_LUSTRE,
 838                         .f_oid  = FID_OID_DOT_LUSTRE_LPF,
 839                 },
 840         },
 841
 842         {
 843                 .olm_name               = NULL
 844         }
 845 };
 846
 847 struct osd_ios_item {
 848         struct list_head        oii_list;
 849         uint64_t                oii_parent;
 850         enum osd_lf_flags       oii_flags;
 851         scan_dir_t              oii_scan_dir;
 852         handle_dirent_t         oii_handle_dirent;
 853 };
 854
 855 static int osd_ios_new_item(struct osd_device *dev, uint64_t parent,
 856                             enum osd_lf_flags flags, scan_dir_t scan_dir,
 857                             handle_dirent_t handle_dirent)
 858 {
 859         struct osd_ios_item *item;
 860
 861         OBD_ALLOC_PTR(item);
 862         if (!item) {
 863                 CWARN("%s: initial OI scrub failed to add item for %llu\n",
 864                       osd_name(dev), parent);
 865                 return -ENOMEM;
 866         }
 867
 868         INIT_LIST_HEAD(&item->oii_list);
 869         item->oii_parent = parent;
 870         item->oii_flags = flags;
 871         item->oii_scan_dir = scan_dir;
 872         item->oii_handle_dirent = handle_dirent;
 873         list_add_tail(&item->oii_list, &dev->od_ios_list);
 874
 875         return 0;
 876 }
 877
 878 static bool osd_index_need_recreate(const struct lu_env *env,
 879                                     struct osd_device *dev, uint64_t oid)
 880 {
 881         struct osd_thread_info *info = osd_oti_get(env);
 882         zap_attribute_t *za = &info->oti_za2;
 883         zap_cursor_t *zc = &info->oti_zc2;
 884         int rc;
 885         ENTRY;
 886
 887         zap_cursor_init_serialized(zc, dev->od_os, oid, 0);
 888         rc = -zap_cursor_retrieve(zc, za);
 889         zap_cursor_fini(zc);
 890         if (rc && rc != -ENOENT)
 891                 RETURN(true);
 892
 893         RETURN(false);
 894 }
 895
 896 static void osd_ios_index_register(const struct lu_env *env,
 897                                    struct osd_device *osd,
 898                                    const struct lu_fid *fid, uint64_t oid)
 899 {
 900         struct osd_thread_info *info = osd_oti_get(env);
 901         zap_attribute_t *za = &info->oti_za2;
 902         zap_cursor_t *zc = &info->oti_zc2;
 903         struct zap_leaf_entry *le;
 904         dnode_t *dn = NULL;
 905         sa_handle_t *hdl;
 906         __u64 mode = 0;
 907         __u32 keysize = 0;
 908         __u32 recsize = 0;
 909         int rc;
 910         ENTRY;
 911
 912         rc = __osd_obj2dnode(osd->od_os, oid, &dn);
 913         if (rc == -EEXIST || rc == -ENOENT)
 914                 RETURN_EXIT;
 915
 916         if (rc < 0)
 917                 GOTO(log, rc);
 918
 919         if (!osd_object_is_zap(dn))
 920                 GOTO(log, rc = 1);
 921
 922         rc = -sa_handle_get(osd->od_os, oid, NULL, SA_HDL_PRIVATE, &hdl);
 923         if (rc)
 924                 GOTO(log, rc);
 925
 926         rc = -sa_lookup(hdl, SA_ZPL_MODE(osd), &mode, sizeof(mode));
 927         sa_handle_destroy(hdl);
 928         if (rc)
 929                 GOTO(log, rc);
 930
 931         if (!S_ISREG(mode))
 932                 GOTO(log, rc = 1);
 933
 934         zap_cursor_init_serialized(zc, osd->od_os, oid, 0);
 935         rc = -zap_cursor_retrieve(zc, za);
 936         if (rc)
 937                 /* Skip empty index object */
 938                 GOTO(fini, rc = (rc == -ENOENT ? 1 : rc));
 939
 940         if (zc->zc_zap->zap_ismicro ||
 941             !(zap_f_phys(zc->zc_zap)->zap_flags & ZAP_FLAG_UINT64_KEY))
 942                 GOTO(fini, rc = 1);
 943
 944         le = ZAP_LEAF_ENTRY(zc->zc_leaf, 0);
 945         keysize = le->le_name_numints * 8;
 946         recsize = za->za_integer_length * za->za_num_integers;
 947         if (likely(keysize && recsize))
 948                 rc = osd_index_register(osd, fid, keysize, recsize);
 949
 950         GOTO(fini, rc);
 951
 952 fini:
 953         zap_cursor_fini(zc);
 954
 955 log:
 956         if (dn)
 957                 osd_dnode_rele(dn);
 958         if (rc < 0)
 959                 CWARN("%s: failed to register index "DFID" (%u/%u): rc = %d\n",
 960                       osd_name(osd), PFID(fid), keysize, recsize, rc);
 961         else if (!rc)
 962                 CDEBUG(D_LFSCK, "%s: registered index "DFID" (%u/%u)\n",
 963                        osd_name(osd), PFID(fid), keysize, recsize);
 964 }
 965
 966 static void osd_index_restore(const struct lu_env *env, struct osd_device *dev,
 967                               struct lustre_index_restore_unit *liru, void *buf,
 968                               int bufsize)
 969 {
 970         struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
 971         struct lu_fid *tgt_fid = &liru->liru_cfid;
 972         struct lu_fid bak_fid;
 973         int rc;
 974         ENTRY;
 975
 976         lustre_fid2lbx(buf, tgt_fid, bufsize);
 977         rc = -zap_lookup(dev->od_os, dev->od_index_backup_id, buf, 8,
 978                          sizeof(*zde) / 8, (void *)zde);
 979         if (rc)
 980                 GOTO(log, rc);
 981
 982         rc = osd_get_fid_by_oid(env, dev, zde->lzd_reg.zde_dnode, &bak_fid);
 983         if (rc)
 984                 GOTO(log, rc);
 985
 986         /* The OI mapping for index may be invalid, since it will be
 987          * re-created, not update the OI mapping, just cache it in RAM. */
 988         rc = osd_idc_find_and_init_with_oid(env, dev, tgt_fid,
 989                                             liru->liru_clid);
 990         if (!rc)
 991                 rc = lustre_index_restore(env, &dev->od_dt_dev,
 992                                 &liru->liru_pfid, tgt_fid, &bak_fid,
 993                                 liru->liru_name, &dev->od_index_backup_list,
 994                                 &dev->od_lock, buf, bufsize);
 995         GOTO(log, rc);
 996
 997 log:
 998         CDEBUG(D_WARNING, "%s: restore index '%s' with "DFID": rc = %d\n",
 999                osd_name(dev), liru->liru_name, PFID(tgt_fid), rc);
1000 }
1001
1002 /**
1003  * verify FID-in-LMA and OI entry for one object
1004  *
1005  * ios: Initial OI Scrub.
1006  */
1007 static int osd_ios_scan_one(const struct lu_env *env, struct osd_device *dev,
1008                             const struct lu_fid *fid, uint64_t parent,
1009                             uint64_t oid, const char *name,
1010                             enum osd_lf_flags flags)
1011 {
1012         struct lustre_scrub *scrub = &dev->od_scrub;
1013         struct scrub_file *sf = &scrub->os_file;
1014         struct lustre_mdt_attrs *lma = NULL;
1015         nvlist_t *nvbuf = NULL;
1016         struct lu_fid tfid;
1017         uint64_t oid2 = 0;
1018         __u64 flag = 0;
1019         int size = 0;
1020         int op = 0;
1021         int rc;
1022         ENTRY;
1023
1024         rc = __osd_xattr_load_by_oid(dev, oid, &nvbuf);
1025         if (unlikely(rc == -ENOENT || rc == -EEXIST))
1026                 RETURN(0);
1027
1028         if (rc && rc != -ENODATA) {
1029                 CWARN("%s: initial OI scrub failed to get lma for %llu: "
1030                       "rc = %d\n", osd_name(dev), oid, rc);
1031
1032                 RETURN(rc);
1033         }
1034
1035         if (!rc) {
1036                 LASSERT(nvbuf != NULL);
1037                 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
1038                                                (uchar_t **)&lma, &size);
1039                 if (rc || size == 0) {
1040                         LASSERT(lma == NULL);
1041                         rc = -ENODATA;
1042                 } else {
1043                         LASSERTF(lma != NULL, "corrupted LMA, size %d\n", size);
1044                         lustre_lma_swab(lma);
1045                         if (lma->lma_compat & LMAC_NOT_IN_OI) {
1046                                 nvlist_free(nvbuf);
1047                                 RETURN(0);
1048                         }
1049
1050                         if (lma->lma_compat & LMAC_IDX_BACKUP &&
1051                             osd_index_need_recreate(env, dev, oid)) {
1052                                 if (parent == dev->od_root) {
1053                                         lu_local_obj_fid(&tfid,
1054                                                          OSD_FS_ROOT_OID);
1055                                 } else {
1056                                         rc = osd_get_fid_by_oid(env, dev,
1057                                                                 parent, &tfid);
1058                                         if (rc) {
1059                                                 nvlist_free(nvbuf);
1060                                                 RETURN(rc);
1061                                         }
1062                                 }
1063
1064                                 rc = lustre_liru_new(
1065                                                 &dev->od_index_restore_list,
1066                                                 &tfid, &lma->lma_self_fid, oid,
1067                                                 name, strlen(name));
1068                                 nvlist_free(nvbuf);
1069                                 RETURN(rc);
1070                         }
1071
1072                         tfid = lma->lma_self_fid;
1073                         if (!(flags & OLF_NOT_BACKUP))
1074                                 osd_ios_index_register(env, dev, &tfid, oid);
1075                 }
1076                 nvlist_free(nvbuf);
1077         }
1078
1079         if (rc == -ENODATA) {
1080                 if (!fid) {
1081                         /* Skip the object without FID-in-LMA */
1082                         CDEBUG(D_LFSCK, "%s: %llu has no FID-in-LMA, skip it\n",
1083                                osd_name(dev), oid);
1084
1085                         RETURN(0);
1086                 }
1087
1088                 LASSERT(!fid_is_zero(fid));
1089
1090                 tfid = *fid;
1091                 if (flags & OLF_IDX_IN_FID) {
1092                         LASSERT(dev->od_index >= 0);
1093
1094                         tfid.f_oid = dev->od_index;
1095                 }
1096         }
1097
1098         rc = osd_fid_lookup(env, dev, &tfid, &oid2);
1099         if (rc) {
1100                 if (rc != -ENOENT) {
1101                         CWARN("%s: initial OI scrub failed to lookup fid for "
1102                               DFID"=>%llu: rc = %d\n",
1103                               osd_name(dev), PFID(&tfid), oid, rc);
1104
1105                         RETURN(rc);
1106                 }
1107
1108                 flag = SF_RECREATED;
1109                 op = DTO_INDEX_INSERT;
1110         } else {
1111                 if (oid == oid2)
1112                         RETURN(0);
1113
1114                 flag = SF_INCONSISTENT;
1115                 op = DTO_INDEX_UPDATE;
1116         }
1117
1118         if (!(sf->sf_flags & flag)) {
1119                 scrub_file_reset(scrub, dev->od_uuid, flag);
1120                 rc = scrub_file_store(env, scrub);
1121                 if (rc)
1122                         RETURN(rc);
1123         }
1124
1125         rc = osd_scrub_refresh_mapping(env, dev, &tfid, oid, op, true, name);
1126
1127         RETURN(rc > 0 ? 0 : rc);
1128 }
1129
1130 static int osd_ios_varfid_hd(const struct lu_env *env, struct osd_device *dev,
1131                              const char *name, uint64_t parent, uint64_t oid,
1132                              enum osd_lf_flags flags, bool is_dir)
1133 {
1134         int rc;
1135         ENTRY;
1136
1137         rc = osd_ios_scan_one(env, dev, NULL, parent, oid, name, 0);
1138         if (!rc && is_dir)
1139                 rc = osd_ios_new_item(dev, oid, flags, osd_ios_general_sd,
1140                                       osd_ios_varfid_hd);
1141
1142         RETURN(rc);
1143 }
1144
1145 static int osd_ios_uld_hd(const struct lu_env *env, struct osd_device *dev,
1146                           const char *name, uint64_t parent, uint64_t oid,
1147                           enum osd_lf_flags flags, bool is_dir)
1148 {
1149         struct lu_fid tfid;
1150         int rc;
1151         ENTRY;
1152
1153         /* skip any non-DFID format name */
1154         if (name[0] != '[')
1155                 RETURN(0);
1156
1157         /* skip the start '[' */
1158         sscanf(&name[1], SFID, RFID(&tfid));
1159         if (fid_is_sane(&tfid))
1160                 rc = osd_ios_scan_one(env, dev, &tfid, parent, oid, name, 0);
1161         else
1162                 rc = -EIO;
1163
1164         RETURN(rc);
1165 }
1166
1167 /*
1168  * General scanner for the directories execpt /ROOT during initial OI scrub.
1169  * It scans the name entries under the given directory one by one. For each
1170  * entry, verifies its OI mapping via the given @handle_dirent.
1171  */
1172 static int osd_ios_general_sd(const struct lu_env *env, struct osd_device *dev,
1173                               uint64_t parent, handle_dirent_t handle_dirent,
1174                               enum osd_lf_flags flags)
1175 {
1176         struct osd_thread_info *info = osd_oti_get(env);
1177         struct luz_direntry *zde = &info->oti_zde;
1178         zap_attribute_t *za = &info->oti_za;
1179         zap_cursor_t *zc = &info->oti_zc;
1180         int rc;
1181         ENTRY;
1182
1183         zap_cursor_init_serialized(zc, dev->od_os, parent, 0);
1184         rc = -zap_cursor_retrieve(zc, za);
1185         if (rc == -ENOENT)
1186                 zap_cursor_advance(zc);
1187         else if (rc)
1188                 GOTO(log, rc);
1189
1190         while (1) {
1191                 rc = -zap_cursor_retrieve(zc, za);
1192                 if (rc)
1193                         GOTO(log, rc = (rc == -ENOENT ? 0 : rc));
1194
1195                 /* skip the entry started with '.' */
1196                 if (likely(za->za_name[0] != '.')) {
1197                         rc = osd_zap_lookup(dev, parent, NULL, za->za_name,
1198                                         za->za_integer_length,
1199                                         sizeof(*zde) / za->za_integer_length,
1200                                         (void *)zde);
1201                         if (rc) {
1202                                 CWARN("%s: initial OI scrub failed to lookup "
1203                                       "%s under %llu: rc = %d\n",
1204                                       osd_name(dev), za->za_name, parent, rc);
1205                                 continue;
1206                         }
1207
1208                         rc = handle_dirent(env, dev, za->za_name, parent,
1209                                         zde->lzd_reg.zde_dnode, flags,
1210                                         S_ISDIR(DTTOIF(zde->lzd_reg.zde_type)) ?
1211                                         true : false);
1212                         CDEBUG(D_LFSCK, "%s: initial OI scrub handled %s under "
1213                                "%llu: rc = %d\n",
1214                                osd_name(dev), za->za_name, parent, rc);
1215                 }
1216
1217                 zap_cursor_advance(zc);
1218         }
1219
1220 log:
1221         if (rc)
1222                 CWARN("%s: initial OI scrub failed to scan the directory %llu: "
1223                       "rc = %d\n", osd_name(dev), parent, rc);
1224         zap_cursor_fini(zc);
1225
1226         return rc;
1227 }
1228
1229 /*
1230  * The scanner for /ROOT directory. It is not all the items under /ROOT will
1231  * be scanned during the initial OI scrub, instead, only the .lustre and the
1232  * sub-items under .lustre will be handled.
1233  */
1234 static int osd_ios_ROOT_sd(const struct lu_env *env, struct osd_device *dev,
1235                            uint64_t parent, handle_dirent_t handle_dirent,
1236                            enum osd_lf_flags flags)
1237 {
1238         struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
1239         const struct osd_lf_map *map;
1240         uint64_t oid;
1241         int rc;
1242         int rc1 = 0;
1243         ENTRY;
1244
1245         rc = osd_zap_lookup(dev, parent, NULL, dot_lustre_name, 8,
1246                             sizeof(*zde) / 8, (void *)zde);
1247         if (rc == -ENOENT) {
1248                 /* The .lustre directory is lost. That is not fatal. It can
1249                  * be re-created in the subsequent MDT start processing. */
1250                 RETURN(0);
1251         }
1252
1253         if (rc) {
1254                 CWARN("%s: initial OI scrub failed to find .lustre: "
1255                       "rc = %d\n", osd_name(dev), rc);
1256
1257                 RETURN(rc);
1258         }
1259
1260         oid = zde->lzd_reg.zde_dnode;
1261         rc = osd_ios_scan_one(env, dev, &LU_DOT_LUSTRE_FID, parent, oid,
1262                               dot_lustre_name, 0);
1263         if (rc)
1264                 RETURN(rc);
1265
1266         for (map = osd_dl_maps; map->olm_name; map++) {
1267                 rc = osd_zap_lookup(dev, oid, NULL, map->olm_name, 8,
1268                                     sizeof(*zde) / 8, (void *)zde);
1269                 if (rc) {
1270                         if (rc != -ENOENT)
1271                                 CWARN("%s: initial OI scrub failed to find the entry %s under .lustre: rc = %d\n",
1272                                       osd_name(dev), map->olm_name, rc);
1273                         else if (!fid_is_zero(&map->olm_fid))
1274                                 /* Try to remove the stale OI mapping. */
1275                                 osd_scrub_refresh_mapping(env, dev,
1276                                                 &map->olm_fid, 0,
1277                                                 DTO_INDEX_DELETE, true,
1278                                                 map->olm_name);
1279                         continue;
1280                 }
1281
1282                 rc = osd_ios_scan_one(env, dev, &map->olm_fid, oid,
1283                                       zde->lzd_reg.zde_dnode, map->olm_name,
1284                                       map->olm_flags);
1285                 if (rc)
1286                         rc1 = rc;
1287         }
1288
1289         RETURN(rc1);
1290 }
1291
1292 static void osd_initial_OI_scrub(const struct lu_env *env,
1293                                  struct osd_device *dev)
1294 {
1295         struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
1296         const struct osd_lf_map *map;
1297         int rc;
1298         ENTRY;
1299
1300         for (map = osd_lf_maps; map->olm_name; map++) {
1301                 rc = osd_zap_lookup(dev, dev->od_root, NULL, map->olm_name, 8,
1302                                     sizeof(*zde) / 8, (void *)zde);
1303                 if (rc) {
1304                         if (rc != -ENOENT)
1305                                 CWARN("%s: initial OI scrub failed "
1306                                       "to find the entry %s: rc = %d\n",
1307                                       osd_name(dev), map->olm_name, rc);
1308                         else if (!fid_is_zero(&map->olm_fid))
1309                                 /* Try to remove the stale OI mapping. */
1310                                 osd_scrub_refresh_mapping(env, dev,
1311                                                 &map->olm_fid, 0,
1312                                                 DTO_INDEX_DELETE, true,
1313                                                 map->olm_name);
1314                         continue;
1315                 }
1316
1317                 rc = osd_ios_scan_one(env, dev, &map->olm_fid, dev->od_root,
1318                                       zde->lzd_reg.zde_dnode, map->olm_name,
1319                                       map->olm_flags);
1320                 if (!rc && map->olm_flags & OLF_SCAN_SUBITEMS)
1321                         osd_ios_new_item(dev, zde->lzd_reg.zde_dnode,
1322                                          map->olm_flags, map->olm_scan_dir,
1323                                          map->olm_handle_dirent);
1324         }
1325
1326         while (!list_empty(&dev->od_ios_list)) {
1327                 struct osd_ios_item *item;
1328
1329                 item = list_entry(dev->od_ios_list.next,
1330                                   struct osd_ios_item, oii_list);
1331                 list_del_init(&item->oii_list);
1332                 item->oii_scan_dir(env, dev, item->oii_parent,
1333                                    item->oii_handle_dirent, item->oii_flags);
1334                 OBD_FREE_PTR(item);
1335         }
1336
1337         if (!list_empty(&dev->od_index_restore_list)) {
1338                 char *buf;
1339
1340                 OBD_ALLOC_LARGE(buf, INDEX_BACKUP_BUFSIZE);
1341                 if (!buf)
1342                         CERROR("%s: not enough RAM for rebuild index\n",
1343                                osd_name(dev));
1344
1345                 while (!list_empty(&dev->od_index_restore_list)) {
1346                         struct lustre_index_restore_unit *liru;
1347
1348                         liru = list_entry(dev->od_index_restore_list.next,
1349                                           struct lustre_index_restore_unit,
1350                                           liru_link);
1351                         list_del(&liru->liru_link);
1352                         if (buf)
1353                                 osd_index_restore(env, dev, liru, buf,
1354                                                   INDEX_BACKUP_BUFSIZE);
1355                         OBD_FREE(liru, liru->liru_len);
1356                 }
1357
1358                 if (buf)
1359                         OBD_FREE_LARGE(buf, INDEX_BACKUP_BUFSIZE);
1360         }
1361
1362         EXIT;
1363 }
1364
1365 /* OI scrub start/stop */
1366
1367 int osd_scrub_start(const struct lu_env *env, struct osd_device *dev,
1368                     __u32 flags)
1369 {
1370         int rc;
1371         ENTRY;
1372
1373         if (dev->od_dt_dev.dd_rdonly)
1374                 RETURN(-EROFS);
1375
1376         /* od_otable_sem: prevent concurrent start/stop */
1377         down(&dev->od_otable_sem);
1378         rc = scrub_start(osd_scrub_main, &dev->od_scrub, dev, flags);
1379         up(&dev->od_otable_sem);
1380
1381         RETURN(rc == -EALREADY ? 0 : rc);
1382 }
1383
1384 void osd_scrub_stop(struct osd_device *dev)
1385 {
1386         struct lustre_scrub *scrub = &dev->od_scrub;
1387         ENTRY;
1388
1389         /* od_otable_sem: prevent concurrent start/stop */
1390         down(&dev->od_otable_sem);
1391         scrub->os_paused = 1;
1392         scrub_stop(scrub);
1393         up(&dev->od_otable_sem);
1394
1395         EXIT;
1396 }
1397
1398 /* OI scrub setup/cleanup */
1399
1400 static const char osd_scrub_name[] = "OI_scrub";
1401
1402 int osd_scrub_setup(const struct lu_env *env, struct osd_device *dev)
1403 {
1404         struct osd_thread_info *info = osd_oti_get(env);
1405         struct lustre_scrub *scrub = &dev->od_scrub;
1406         struct scrub_file *sf = &scrub->os_file;
1407         struct lu_fid *fid = &info->oti_fid;
1408         struct dt_object *obj;
1409         uint64_t oid;
1410         int rc = 0;
1411         bool dirty = false;
1412         ENTRY;
1413
1414         memcpy(dev->od_uuid.b,
1415                &dsl_dataset_phys(dev->od_os->os_dsl_dataset)->ds_guid,
1416                sizeof(dsl_dataset_phys(dev->od_os->os_dsl_dataset)->ds_guid));
1417         memset(&dev->od_scrub, 0, sizeof(struct lustre_scrub));
1418         init_waitqueue_head(&scrub->os_thread.t_ctl_waitq);
1419         init_rwsem(&scrub->os_rwsem);
1420         spin_lock_init(&scrub->os_lock);
1421         INIT_LIST_HEAD(&scrub->os_inconsistent_items);
1422         scrub->os_name = osd_name(dev);
1423
1424         /* 'What the @fid is' is not imporatant, because the object
1425          * has no OI mapping, and only is visible inside the OSD.*/
1426         fid->f_seq = FID_SEQ_IGIF_MAX;
1427         if (dev->od_is_ost)
1428                 fid->f_oid = ((1 << 31) | dev->od_index) + 1;
1429         else
1430                 fid->f_oid = dev->od_index + 1;
1431         fid->f_ver = 0;
1432         rc = osd_obj_find_or_create(env, dev, dev->od_root,
1433                                     osd_scrub_name, &oid, fid, false);
1434         if (rc)
1435                 RETURN(rc);
1436
1437         rc = osd_idc_find_and_init_with_oid(env, dev, fid, oid);
1438         if (rc)
1439                 RETURN(rc);
1440
1441         obj = lu2dt(lu_object_find_slice(env, osd2lu_dev(dev), fid, NULL));
1442         if (IS_ERR_OR_NULL(obj))
1443                 RETURN(obj ? PTR_ERR(obj) : -ENOENT);
1444
1445         obj->do_body_ops = &osd_body_scrub_ops;
1446         scrub->os_obj = obj;
1447         rc = scrub_file_load(env, scrub);
1448         if (rc == -ENOENT || rc == -EFAULT) {
1449                 scrub_file_init(scrub, dev->od_uuid);
1450                 dirty = true;
1451         } else if (rc < 0) {
1452                 GOTO(cleanup_obj, rc);
1453         } else {
1454                 if (!uuid_equal(&sf->sf_uuid, &dev->od_uuid)) {
1455                         CDEBUG(D_LFSCK,
1456                                "%s: UUID has been changed from %pU to %pU\n",
1457                                osd_name(dev), &sf->sf_uuid, &dev->od_uuid);
1458                         scrub_file_reset(scrub, dev->od_uuid, SF_INCONSISTENT);
1459                         dirty = true;
1460                 } else if (sf->sf_status == SS_SCANNING) {
1461                         sf->sf_status = SS_CRASHED;
1462                         dirty = true;
1463                 }
1464
1465                 if ((sf->sf_oi_count & (sf->sf_oi_count - 1)) != 0) {
1466                         LCONSOLE_WARN("%s: invalid oi count %d, set it to %d\n",
1467                                       osd_name(dev), sf->sf_oi_count,
1468                                       osd_oi_count);
1469                         sf->sf_oi_count = osd_oi_count;
1470                         dirty = true;
1471                 }
1472         }
1473
1474         if (sf->sf_pos_last_checkpoint != 0)
1475                 scrub->os_pos_current = sf->sf_pos_last_checkpoint + 1;
1476         else
1477                 scrub->os_pos_current = 1;
1478
1479         if (dirty) {
1480                 rc = scrub_file_store(env, scrub);
1481                 if (rc)
1482                         GOTO(cleanup_obj, rc);
1483         }
1484
1485         /* Initialize OI files. */
1486         rc = osd_oi_init(env, dev);
1487         if (rc < 0)
1488                 GOTO(cleanup_obj, rc);
1489
1490         if (!dev->od_dt_dev.dd_rdonly)
1491                 osd_initial_OI_scrub(env, dev);
1492
1493         if (!dev->od_dt_dev.dd_rdonly &&
1494             dev->od_auto_scrub_interval != AS_NEVER &&
1495             ((sf->sf_status == SS_PAUSED) ||
1496              (sf->sf_status == SS_CRASHED &&
1497               sf->sf_flags & (SF_RECREATED | SF_INCONSISTENT |
1498                               SF_UPGRADE | SF_AUTO)) ||
1499              (sf->sf_status == SS_INIT &&
1500               sf->sf_flags & (SF_RECREATED | SF_INCONSISTENT |
1501                               SF_UPGRADE))))
1502                 rc = osd_scrub_start(env, dev, SS_AUTO_FULL);
1503
1504         if (rc)
1505                 GOTO(cleanup_oi, rc);
1506
1507         RETURN(0);
1508
1509 cleanup_oi:
1510         osd_oi_fini(env, dev);
1511 cleanup_obj:
1512         dt_object_put_nocache(env, scrub->os_obj);
1513         scrub->os_obj = NULL;
1514
1515         return rc;
1516 }
1517
1518 void osd_scrub_cleanup(const struct lu_env *env, struct osd_device *dev)
1519 {
1520         struct lustre_scrub *scrub = &dev->od_scrub;
1521
1522         LASSERT(!dev->od_otable_it);
1523
1524         if (scrub->os_obj) {
1525                 osd_scrub_stop(dev);
1526                 dt_object_put_nocache(env, scrub->os_obj);
1527                 scrub->os_obj = NULL;
1528         }
1529
1530         if (dev->od_oi_table)
1531                 osd_oi_fini(env, dev);
1532 }
1533
1534 /* object table based iteration APIs */
1535
1536 static struct dt_it *osd_otable_it_init(const struct lu_env *env,
1537                                        struct dt_object *dt, __u32 attr)
1538 {
1539         enum dt_otable_it_flags flags = attr >> DT_OTABLE_IT_FLAGS_SHIFT;
1540         enum dt_otable_it_valid valid = attr & ~DT_OTABLE_IT_FLAGS_MASK;
1541         struct osd_device *dev = osd_dev(dt->do_lu.lo_dev);
1542         struct lustre_scrub *scrub = &dev->od_scrub;
1543         struct osd_otable_it *it;
1544         __u32 start = 0;
1545         int rc;
1546         ENTRY;
1547
1548         if (dev->od_dt_dev.dd_rdonly)
1549                 RETURN(ERR_PTR(-EROFS));
1550
1551         /* od_otable_sem: prevent concurrent init/fini */
1552         down(&dev->od_otable_sem);
1553         if (dev->od_otable_it)
1554                 GOTO(out, it = ERR_PTR(-EALREADY));
1555
1556         OBD_ALLOC_PTR(it);
1557         if (!it)
1558                 GOTO(out, it = ERR_PTR(-ENOMEM));
1559
1560         if (flags & DOIF_OUTUSED)
1561                 it->ooi_used_outside = 1;
1562
1563         if (flags & DOIF_RESET)
1564                 start |= SS_RESET;
1565
1566         if (valid & DOIV_ERROR_HANDLE) {
1567                 if (flags & DOIF_FAILOUT)
1568                         start |= SS_SET_FAILOUT;
1569                 else
1570                         start |= SS_CLEAR_FAILOUT;
1571         }
1572
1573         if (valid & DOIV_DRYRUN) {
1574                 if (flags & DOIF_DRYRUN)
1575                         start |= SS_SET_DRYRUN;
1576                 else
1577                         start |= SS_CLEAR_DRYRUN;
1578         }
1579
1580         /* XXX: dmu_object_next() does NOT find dnodes allocated
1581          *      in the current non-committed txg, so we force txg
1582          *      commit to find all existing dnodes ... */
1583         txg_wait_synced(dmu_objset_pool(dev->od_os), 0ULL);
1584
1585         dev->od_otable_it = it;
1586         it->ooi_dev = dev;
1587         rc = scrub_start(osd_scrub_main, scrub, dev, start & ~SS_AUTO_PARTIAL);
1588         if (rc == -EALREADY) {
1589                 it->ooi_pos = 1;
1590         } else if (rc < 0) {
1591                 dev->od_otable_it = NULL;
1592                 OBD_FREE_PTR(it);
1593                 it = ERR_PTR(rc);
1594         } else {
1595                 it->ooi_pos = scrub->os_pos_current;
1596         }
1597
1598         GOTO(out, it);
1599
1600 out:
1601         up(&dev->od_otable_sem);
1602         return (struct dt_it *)it;
1603 }
1604
1605 static void osd_otable_it_fini(const struct lu_env *env, struct dt_it *di)
1606 {
1607         struct osd_otable_it *it = (struct osd_otable_it *)di;
1608         struct osd_device *dev = it->ooi_dev;
1609
1610         /* od_otable_sem: prevent concurrent init/fini */
1611         down(&dev->od_otable_sem);
1612         scrub_stop(&dev->od_scrub);
1613         LASSERT(dev->od_otable_it == it);
1614
1615         dev->od_otable_it = NULL;
1616         up(&dev->od_otable_sem);
1617         OBD_FREE_PTR(it);
1618 }
1619
1620 static int osd_otable_it_get(const struct lu_env *env,
1621                              struct dt_it *di, const struct dt_key *key)
1622 {
1623         return 0;
1624 }
1625
1626 static void osd_otable_it_put(const struct lu_env *env, struct dt_it *di)
1627 {
1628 }
1629
1630 static void osd_otable_it_preload(const struct lu_env *env,
1631                                   struct osd_otable_it *it)
1632 {
1633         struct osd_device *dev = it->ooi_dev;
1634         int rc;
1635
1636         /* can go negative on the very first access to the iterator
1637          * or if some non-Lustre objects were found */
1638         if (unlikely(it->ooi_prefetched < 0))
1639                 it->ooi_prefetched = 0;
1640
1641         if (it->ooi_prefetched >= (OTABLE_PREFETCH >> 1))
1642                 return;
1643
1644         if (it->ooi_prefetched_dnode == 0)
1645                 it->ooi_prefetched_dnode = it->ooi_pos;
1646
1647         while (it->ooi_prefetched < OTABLE_PREFETCH) {
1648                 rc = -dmu_object_next(dev->od_os, &it->ooi_prefetched_dnode,
1649                                       B_FALSE, 0);
1650                 if (rc)
1651                         break;
1652
1653                 osd_dmu_prefetch(dev->od_os, it->ooi_prefetched_dnode,
1654                                  0, 0, 0, ZIO_PRIORITY_ASYNC_READ);
1655                 it->ooi_prefetched++;
1656         }
1657 }
1658
1659 static inline int
1660 osd_otable_it_wakeup(struct lustre_scrub *scrub, struct osd_otable_it *it)
1661 {
1662         spin_lock(&scrub->os_lock);
1663         if (it->ooi_pos < scrub->os_pos_current || scrub->os_waiting ||
1664             !thread_is_running(&scrub->os_thread))
1665                 it->ooi_waiting = 0;
1666         else
1667                 it->ooi_waiting = 1;
1668         spin_unlock(&scrub->os_lock);
1669
1670         return !it->ooi_waiting;
1671 }
1672
1673 static int osd_otable_it_next(const struct lu_env *env, struct dt_it *di)
1674 {
1675         struct osd_otable_it *it = (struct osd_otable_it *)di;
1676         struct osd_device *dev = it->ooi_dev;
1677         struct lustre_scrub *scrub = &dev->od_scrub;
1678         struct ptlrpc_thread *thread = &scrub->os_thread;
1679         struct lustre_mdt_attrs *lma = NULL;
1680         nvlist_t *nvbuf = NULL;
1681         int rc, size = 0;
1682         bool locked;
1683         ENTRY;
1684
1685         LASSERT(it->ooi_user_ready);
1686         fid_zero(&it->ooi_fid);
1687
1688         if (unlikely(it->ooi_all_cached))
1689                 RETURN(1);
1690
1691 again:
1692         if (nvbuf) {
1693                 nvlist_free(nvbuf);
1694                 nvbuf = NULL;
1695                 lma = NULL;
1696                 size = 0;
1697         }
1698
1699         if (it->ooi_pos >= scrub->os_pos_current)
1700                 wait_event_idle(thread->t_ctl_waitq,
1701                                 osd_otable_it_wakeup(scrub, it));
1702
1703         if (!thread_is_running(thread) && !it->ooi_used_outside)
1704                 GOTO(out, rc = 1);
1705
1706         rc = -dmu_object_next(dev->od_os, &it->ooi_pos, B_FALSE, 0);
1707         if (rc) {
1708                 if (unlikely(rc == -ESRCH)) {
1709                         it->ooi_all_cached = 1;
1710                         rc = 1;
1711                 }
1712
1713                 GOTO(out, rc);
1714         }
1715
1716         rc = __osd_xattr_load_by_oid(dev, it->ooi_pos, &nvbuf);
1717
1718         locked = false;
1719         if (!scrub->os_full_speed) {
1720                 spin_lock(&scrub->os_lock);
1721                 locked = true;
1722         }
1723         it->ooi_prefetched--;
1724         if (!scrub->os_full_speed) {
1725                 if (scrub->os_waiting) {
1726                         scrub->os_waiting = 0;
1727                         wake_up_all(&thread->t_ctl_waitq);
1728                 }
1729         }
1730         if (locked)
1731                 spin_unlock(&scrub->os_lock);
1732
1733         if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
1734                 goto again;
1735
1736         if (rc)
1737                 GOTO(out, rc);
1738
1739         LASSERT(nvbuf != NULL);
1740         rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
1741                                        (uchar_t **)&lma, &size);
1742         if (rc || size == 0)
1743                 /* It is either non-Lustre object or OSD internal object,
1744                  * ignore it, go ahead */
1745                 goto again;
1746
1747         LASSERTF(lma != NULL, "corrupted LMA, size %d\n", size);
1748         lustre_lma_swab(lma);
1749         if (unlikely(lma->lma_compat & LMAC_NOT_IN_OI ||
1750                      lma->lma_incompat & LMAI_AGENT))
1751                 goto again;
1752
1753         it->ooi_fid = lma->lma_self_fid;
1754
1755         GOTO(out, rc = 0);
1756
1757 out:
1758         if (nvbuf)
1759                 nvlist_free(nvbuf);
1760
1761         if (!rc && scrub->os_full_speed)
1762                 osd_otable_it_preload(env, it);
1763
1764         return rc;
1765 }
1766
1767 static struct dt_key *osd_otable_it_key(const struct lu_env *env,
1768                                         const struct dt_it *di)
1769 {
1770         return NULL;
1771 }
1772
1773 static int osd_otable_it_key_size(const struct lu_env *env,
1774                                   const struct dt_it *di)
1775 {
1776         return sizeof(__u64);
1777 }
1778
1779 static int osd_otable_it_rec(const struct lu_env *env, const struct dt_it *di,
1780                              struct dt_rec *rec, __u32 attr)
1781 {
1782         struct osd_otable_it *it  = (struct osd_otable_it *)di;
1783         struct lu_fid *fid = (struct lu_fid *)rec;
1784
1785         *fid = it->ooi_fid;
1786         return 0;
1787 }
1788
1789 static __u64 osd_otable_it_store(const struct lu_env *env,
1790                                  const struct dt_it *di)
1791 {
1792         struct osd_otable_it *it = (struct osd_otable_it *)di;
1793
1794         return it->ooi_pos;
1795 }
1796
1797 /**
1798  * Set the OSD layer iteration start position as the specified hash.
1799  */
1800 static int osd_otable_it_load(const struct lu_env *env,
1801                               const struct dt_it *di, __u64 hash)
1802 {
1803         struct osd_otable_it *it = (struct osd_otable_it *)di;
1804         struct osd_device *dev = it->ooi_dev;
1805         struct lustre_scrub *scrub = &dev->od_scrub;
1806         int rc;
1807         ENTRY;
1808
1809         /* Forbid to set iteration position after iteration started. */
1810         if (it->ooi_user_ready)
1811                 RETURN(-EPERM);
1812
1813         if (hash > OSD_OTABLE_MAX_HASH)
1814                 hash = OSD_OTABLE_MAX_HASH;
1815
1816         /* The hash is the last checkpoint position,
1817          * we will start from the next one. */
1818         it->ooi_pos = hash + 1;
1819         it->ooi_prefetched = 0;
1820         it->ooi_prefetched_dnode = 0;
1821         it->ooi_user_ready = 1;
1822         if (!scrub->os_full_speed)
1823                 wake_up_all(&scrub->os_thread.t_ctl_waitq);
1824
1825         /* Unplug OSD layer iteration by the first next() call. */
1826         rc = osd_otable_it_next(env, (struct dt_it *)it);
1827
1828         RETURN(rc);
1829 }
1830
1831 static int osd_otable_it_key_rec(const struct lu_env *env,
1832                                  const struct dt_it *di, void *key_rec)
1833 {
1834         return 0;
1835 }
1836
1837 const struct dt_index_operations osd_otable_ops = {
1838         .dio_it = {
1839                 .init     = osd_otable_it_init,
1840                 .fini     = osd_otable_it_fini,
1841                 .get      = osd_otable_it_get,
1842                 .put      = osd_otable_it_put,
1843                 .next     = osd_otable_it_next,
1844                 .key      = osd_otable_it_key,
1845                 .key_size = osd_otable_it_key_size,
1846                 .rec      = osd_otable_it_rec,
1847                 .store    = osd_otable_it_store,
1848                 .load     = osd_otable_it_load,
1849                 .key_rec  = osd_otable_it_key_rec,
1850         }
1851 };
1852
1853 /* high priority inconsistent items list APIs */
1854
1855 int osd_oii_insert(const struct lu_env *env, struct osd_device *dev,
1856                    const struct lu_fid *fid, uint64_t oid, bool insert)
1857 {
1858         struct lustre_scrub *scrub = &dev->od_scrub;
1859         struct ptlrpc_thread *thread = &scrub->os_thread;
1860         struct osd_inconsistent_item *oii;
1861         bool wakeup = false;
1862         ENTRY;
1863
1864         osd_idc_find_and_init_with_oid(env, dev, fid, oid);
1865         OBD_ALLOC_PTR(oii);
1866         if (unlikely(!oii))
1867                 RETURN(-ENOMEM);
1868
1869         INIT_LIST_HEAD(&oii->oii_list);
1870         oii->oii_cache.oic_dev = dev;
1871         oii->oii_cache.oic_fid = *fid;
1872         oii->oii_cache.oic_dnode = oid;
1873         oii->oii_insert = insert;
1874
1875         spin_lock(&scrub->os_lock);
1876         if (unlikely(!thread_is_running(thread))) {
1877                 spin_unlock(&scrub->os_lock);
1878                 OBD_FREE_PTR(oii);
1879                 RETURN(-EAGAIN);
1880         }
1881
1882         if (list_empty(&scrub->os_inconsistent_items))
1883                 wakeup = true;
1884         list_add_tail(&oii->oii_list, &scrub->os_inconsistent_items);
1885         spin_unlock(&scrub->os_lock);
1886
1887         if (wakeup)
1888                 wake_up_all(&thread->t_ctl_waitq);
1889
1890         RETURN(0);
1891 }
1892
1893 int osd_oii_lookup(struct osd_device *dev, const struct lu_fid *fid,
1894                    uint64_t *oid)
1895 {
1896         struct lustre_scrub *scrub = &dev->od_scrub;
1897         struct osd_inconsistent_item *oii;
1898         int ret = -ENOENT;
1899         ENTRY;
1900
1901         spin_lock(&scrub->os_lock);
1902         list_for_each_entry(oii, &scrub->os_inconsistent_items, oii_list) {
1903                 if (lu_fid_eq(fid, &oii->oii_cache.oic_fid)) {
1904                         *oid = oii->oii_cache.oic_dnode;
1905                         ret = 0;
1906                         break;
1907                 }
1908         }
1909         spin_unlock(&scrub->os_lock);
1910
1911         RETURN(ret);
1912 }