4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
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.
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.
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
23 * Copyright (c) 2017, Intel Corporation.
26 * lustre/osd-zfs/osd_scrub.c
28 * Top-level entry points into osd module
30 * The OI scrub is used for rebuilding Object Index files when restores MDT from
33 * The otable based iterator scans ZFS objects to feed up layer LFSCK.
35 * Author: Fan Yong <fan.yong@intel.com>
38 #define DEBUG_SUBSYSTEM S_LFSCK
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>
51 #include <sys/zap_leaf.h>
53 #include "osd_internal.h"
55 #define OSD_OTABLE_MAX_HASH ((1ULL << 48) - 1)
56 #define OTABLE_PREFETCH 256
58 #define DTO_INDEX_INSERT 1
59 #define DTO_INDEX_DELETE 2
60 #define DTO_INDEX_UPDATE 3
62 static inline bool osd_scrub_has_window(struct osd_otable_it *it)
64 return it->ooi_prefetched < OTABLE_PREFETCH;
68 * update/insert/delete the specified OI mapping (@fid @id) according to the ops
70 * \retval 1, changed nothing
71 * \retval 0, changed successfully
72 * \retval -ve, on error
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)
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;
89 if (dev->od_scrub.os_file.sf_param & SP_DRYRUN && !force)
92 tx = dmu_tx_create(dev->od_os);
94 GOTO(log, rc = -ENOMEM);
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);
107 case DTO_INDEX_UPDATE:
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,
113 if (unlikely(rc == -ENOENT)) {
114 /* Some unlink thread may removed the OI mapping. */
118 case DTO_INDEX_INSERT:
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,
124 if (unlikely(rc == -EEXIST))
127 case DTO_INDEX_DELETE:
128 rc = osd_zap_remove(dev, zapid, dn, buf, tx);
130 /* It is normal that the unlink thread has removed the
131 * OI mapping already. */
136 LASSERTF(0, "Unexpected ops %d\n", ops);
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);
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)
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;
162 int ops = DTO_INDEX_UPDATE;
166 down_write(&scrub->os_rwsem);
167 scrub->os_new_checked++;
171 if (scrub->os_in_prior)
172 oii = list_entry(scrub->os_inconsistent_items.next,
173 struct osd_inconsistent_item, oii_list);
175 if (oid < sf->sf_pos_latest_start && !oii)
178 if (oii && oii->oii_insert) {
179 ops = DTO_INDEX_INSERT;
183 rc = osd_fid_lookup(env, dev, fid, &oid2);
188 ops = DTO_INDEX_INSERT;
191 rc = __osd_obj2dnode(dev->od_os, oid, &dn);
193 /* Someone removed the object by race. */
194 if (rc == -ENOENT || rc == -EEXIST)
199 scrub->os_full_speed = 1;
200 sf->sf_flags |= SF_INCONSISTENT;
201 } else if (oid == oid2) {
204 struct lustre_mdt_attrs *lma = NULL;
207 rc = __osd_xattr_load_by_oid(dev, oid2, &nvbuf);
208 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
213 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
214 (uchar_t **)&lma, &size);
215 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
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);
226 GOTO(out, rc = -EEXIST);
230 scrub->os_full_speed = 1;
231 sf->sf_flags |= SF_INCONSISTENT;
234 rc = osd_scrub_refresh_mapping(env, dev, fid, oid, ops, false, NULL);
236 if (scrub->os_in_prior)
237 sf->sf_items_updated_prior++;
239 sf->sf_items_updated++;
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;
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);
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);
275 RETURN(sf->sf_param & SP_FAILOUT ? rc : 0);
278 static int osd_scrub_prep(const struct lu_env *env, struct osd_device *dev)
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;
285 bool drop_dryrun = false;
288 CDEBUG(D_LFSCK, "%s: OI scrub prep, flags = 0x%x\n",
289 scrub->os_name, flags);
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;
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;
304 if (flags & SS_RESET)
305 scrub_file_reset(scrub, dev->od_uuid, 0);
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 |
313 scrub->os_full_speed = 1;
315 scrub->os_full_speed = 0;
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;
331 sf->sf_pos_latest_start = 1;
333 scrub->os_pos_current = sf->sf_pos_latest_start;
334 sf->sf_status = SS_SCANNING;
335 sf->sf_time_latest_start = cfs_time_current_sec();
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);
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);
345 up_write(&scrub->os_rwsem);
350 static int osd_scrub_post(const struct lu_env *env, struct osd_device *dev,
353 struct lustre_scrub *scrub = &dev->od_scrub;
354 struct scrub_file *sf = &scrub->os_file;
358 CDEBUG(D_LFSCK, "%s: OI scrub post with result = %d\n",
359 scrub->os_name, result);
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;
370 sf->sf_time_last_checkpoint = cfs_time_current_sec();
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);
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;
384 sf->sf_status = SS_STOPPED;
386 sf->sf_status = SS_FAILED;
388 sf->sf_run_time += cfs_duration_sec(cfs_time_current() + HALF_SEC -
389 scrub->os_time_last_checkpoint);
390 rc = scrub_file_store(env, scrub);
391 up_write(&scrub->os_rwsem);
393 RETURN(rc < 0 ? rc : result);
396 /* iteration engine */
399 osd_scrub_wakeup(struct lustre_scrub *scrub, struct osd_otable_it *it)
401 spin_lock(&scrub->os_lock);
402 if (osd_scrub_has_window(it) ||
403 !list_empty(&scrub->os_inconsistent_items) ||
404 it->ooi_waiting || !thread_is_running(&scrub->os_thread))
405 scrub->os_waiting = 0;
407 scrub->os_waiting = 1;
408 spin_unlock(&scrub->os_lock);
410 return !scrub->os_waiting;
413 static int osd_scrub_next(const struct lu_env *env, struct osd_device *dev,
414 struct lu_fid *fid, uint64_t *oid)
416 struct l_wait_info lwi = { 0 };
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;
426 if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_DELAY) && cfs_fail_val > 0) {
427 lwi = LWI_TIMEOUT(cfs_time_seconds(cfs_fail_val), NULL, NULL);
428 if (likely(lwi.lwi_timeout > 0)) {
429 l_wait_event(thread->t_ctl_waitq,
430 !list_empty(&scrub->os_inconsistent_items) ||
431 !thread_is_running(thread),
433 if (unlikely(!thread_is_running(thread)))
434 RETURN(SCRUB_NEXT_EXIT);
438 if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_CRASH)) {
439 spin_lock(&scrub->os_lock);
440 thread_set_flags(thread, SVC_STOPPING);
441 spin_unlock(&scrub->os_lock);
442 RETURN(SCRUB_NEXT_CRASH);
445 if (OBD_FAIL_CHECK(OBD_FAIL_OSD_SCRUB_FATAL))
446 RETURN(SCRUB_NEXT_FATAL);
455 if (!list_empty(&scrub->os_inconsistent_items)) {
456 spin_lock(&scrub->os_lock);
457 if (likely(!list_empty(&scrub->os_inconsistent_items))) {
458 struct osd_inconsistent_item *oii;
460 oii = list_entry(scrub->os_inconsistent_items.next,
461 struct osd_inconsistent_item, oii_list);
462 *fid = oii->oii_cache.oic_fid;
463 *oid = oii->oii_cache.oic_dnode;
464 scrub->os_in_prior = 1;
465 spin_unlock(&scrub->os_lock);
469 spin_unlock(&scrub->os_lock);
472 if (!scrub->os_full_speed && !osd_scrub_has_window(it)) {
473 memset(&lwi, 0, sizeof(lwi));
474 l_wait_event(thread->t_ctl_waitq,
475 osd_scrub_wakeup(scrub, it),
479 if (unlikely(!thread_is_running(thread)))
480 GOTO(out, rc = SCRUB_NEXT_EXIT);
482 rc = -dmu_object_next(dev->od_os, &scrub->os_pos_current, B_FALSE, 0);
484 GOTO(out, rc = (rc == -ESRCH ? SCRUB_NEXT_BREAK : rc));
486 rc = __osd_xattr_load_by_oid(dev, scrub->os_pos_current, &nvbuf);
487 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
493 LASSERT(nvbuf != NULL);
494 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
495 (uchar_t **)&lma, &size);
497 lustre_lma_swab(lma);
498 if (likely(!(lma->lma_compat & LMAC_NOT_IN_OI) &&
499 !(lma->lma_incompat & LMAI_AGENT))) {
500 *fid = lma->lma_self_fid;
501 *oid = scrub->os_pos_current;
507 if (!scrub->os_full_speed) {
508 spin_lock(&scrub->os_lock);
509 it->ooi_prefetched++;
510 if (it->ooi_waiting) {
512 wake_up_all(&thread->t_ctl_waitq);
514 spin_unlock(&scrub->os_lock);
526 static int osd_scrub_exec(const struct lu_env *env, struct osd_device *dev,
527 const struct lu_fid *fid, uint64_t oid, int rc)
529 struct lustre_scrub *scrub = &dev->od_scrub;
530 struct ptlrpc_thread *thread = &scrub->os_thread;
531 struct osd_otable_it *it = dev->od_otable_it;
533 rc = osd_scrub_check_update(env, dev, fid, oid, rc);
534 if (!scrub->os_in_prior) {
535 if (!scrub->os_full_speed) {
536 spin_lock(&scrub->os_lock);
537 it->ooi_prefetched++;
538 if (it->ooi_waiting) {
540 wake_up_all(&thread->t_ctl_waitq);
542 spin_unlock(&scrub->os_lock);
545 scrub->os_in_prior = 0;
551 rc = scrub_checkpoint(env, scrub);
553 CDEBUG(D_LFSCK, "%s: fail to checkpoint, pos = %llu: "
554 "rc = %d\n", scrub->os_name, scrub->os_pos_current, rc);
555 /* Continue, as long as the scrub itself can go ahead. */
561 static int osd_scrub_main(void *args)
564 struct osd_device *dev = (struct osd_device *)args;
565 struct lustre_scrub *scrub = &dev->od_scrub;
566 struct ptlrpc_thread *thread = &scrub->os_thread;
572 rc = lu_env_init(&env, LCT_LOCAL | LCT_DT_THREAD);
574 CDEBUG(D_LFSCK, "%s: OI scrub fail to init env: rc = %d\n",
579 rc = osd_scrub_prep(&env, dev);
581 CDEBUG(D_LFSCK, "%s: OI scrub fail to scrub prep: rc = %d\n",
586 if (!scrub->os_full_speed) {
587 struct l_wait_info lwi = { 0 };
588 struct osd_otable_it *it = dev->od_otable_it;
590 l_wait_event(thread->t_ctl_waitq,
591 it->ooi_user_ready || !thread_is_running(thread),
593 if (unlikely(!thread_is_running(thread)))
596 scrub->os_pos_current = it->ooi_pos;
599 CDEBUG(D_LFSCK, "%s: OI scrub start, flags = 0x%x, pos = %llu\n",
600 scrub->os_name, scrub->os_start_flags,
601 scrub->os_pos_current);
603 fid = &osd_oti_get(&env)->oti_fid;
604 while (!rc && thread_is_running(thread)) {
605 rc = osd_scrub_next(&env, dev, fid, &oid);
607 case SCRUB_NEXT_EXIT:
609 case SCRUB_NEXT_CRASH:
610 spin_lock(&scrub->os_lock);
611 thread_set_flags(&scrub->os_thread, SVC_STOPPING);
612 spin_unlock(&scrub->os_lock);
613 GOTO(out, rc = -EINVAL);
614 case SCRUB_NEXT_FATAL:
615 GOTO(post, rc = -EINVAL);
616 case SCRUB_NEXT_BREAK:
620 rc = osd_scrub_exec(&env, dev, fid, oid, rc);
626 rc = osd_scrub_post(&env, dev, rc);
627 CDEBUG(D_LFSCK, "%s: OI scrub: stop, pos = %llu: rc = %d\n",
628 scrub->os_name, scrub->os_pos_current, rc);
631 while (!list_empty(&scrub->os_inconsistent_items)) {
632 struct osd_inconsistent_item *oii;
634 oii = list_entry(scrub->os_inconsistent_items.next,
635 struct osd_inconsistent_item, oii_list);
636 list_del_init(&oii->oii_list);
643 spin_lock(&scrub->os_lock);
644 thread_set_flags(thread, SVC_STOPPED);
645 wake_up_all(&thread->t_ctl_waitq);
646 spin_unlock(&scrub->os_lock);
650 /* initial OI scrub */
654 typedef int (*handle_dirent_t)(const struct lu_env *, struct osd_device *,
655 const char *, uint64_t, uint64_t,
656 enum osd_lf_flags, bool);
657 static int osd_ios_varfid_hd(const struct lu_env *, struct osd_device *,
658 const char *, uint64_t, uint64_t,
659 enum osd_lf_flags, bool);
660 static int osd_ios_uld_hd(const struct lu_env *, struct osd_device *,
661 const char *, uint64_t, uint64_t,
662 enum osd_lf_flags, bool);
664 typedef int (*scan_dir_t)(const struct lu_env *, struct osd_device *,
665 uint64_t, handle_dirent_t, enum osd_lf_flags);
666 static int osd_ios_general_sd(const struct lu_env *, struct osd_device *,
667 uint64_t, handle_dirent_t, enum osd_lf_flags);
668 static int osd_ios_ROOT_sd(const struct lu_env *, struct osd_device *,
669 uint64_t, handle_dirent_t, enum osd_lf_flags);
673 struct lu_fid olm_fid;
674 enum osd_lf_flags olm_flags;
675 scan_dir_t olm_scan_dir;
676 handle_dirent_t olm_handle_dirent;
679 /* Add the new introduced local files in the list in the future. */
680 static const struct osd_lf_map osd_lf_maps[] = {
683 .olm_name = MOUNT_CONFIGS_DIR,
685 .f_seq = FID_SEQ_LOCAL_FILE,
686 .f_oid = MGS_CONFIGS_OID,
688 .olm_flags = OLF_SCAN_SUBITEMS,
689 .olm_scan_dir = osd_ios_general_sd,
690 .olm_handle_dirent = osd_ios_varfid_hd,
693 /* NIDTBL_VERSIONS */
695 .olm_name = MGS_NIDTBL_DIR,
696 .olm_flags = OLF_SCAN_SUBITEMS,
697 .olm_scan_dir = osd_ios_general_sd,
698 .olm_handle_dirent = osd_ios_varfid_hd,
703 .olm_name = "PENDING",
710 .f_seq = FID_SEQ_ROOT,
711 .f_oid = FID_OID_ROOT,
713 .olm_flags = OLF_SCAN_SUBITEMS,
714 .olm_scan_dir = osd_ios_ROOT_sd,
721 .f_seq = FID_SEQ_LOCAL_FILE,
722 .f_oid = FLD_INDEX_OID,
726 /* changelog_catalog */
728 .olm_name = CHANGELOG_CATALOG,
731 /* changelog_users */
733 .olm_name = CHANGELOG_USERS,
739 .olm_flags = OLF_SCAN_SUBITEMS,
740 .olm_scan_dir = osd_ios_general_sd,
741 .olm_handle_dirent = osd_ios_varfid_hd,
747 .olm_flags = OLF_SCAN_SUBITEMS,
748 .olm_scan_dir = osd_ios_general_sd,
749 .olm_handle_dirent = osd_ios_varfid_hd,
754 .olm_name = LFSCK_DIR,
755 .olm_flags = OLF_SCAN_SUBITEMS | OLF_NOT_BACKUP,
756 .olm_scan_dir = osd_ios_general_sd,
757 .olm_handle_dirent = osd_ios_varfid_hd,
762 .olm_name = LFSCK_BOOKMARK,
767 .olm_name = LFSCK_LAYOUT,
770 /* lfsck_namespace */
772 .olm_name = LFSCK_NAMESPACE,
775 /* OSP update logs update_log{_dir} use f_seq = FID_SEQ_UPDATE_LOG{_DIR}
776 * and f_oid = index for their log files. See lu_update_log{_dir}_fid()
777 * for more details. */
781 .olm_name = "update_log",
783 .f_seq = FID_SEQ_UPDATE_LOG,
785 .olm_flags = OLF_IDX_IN_FID,
790 .olm_name = "update_log_dir",
792 .f_seq = FID_SEQ_UPDATE_LOG_DIR,
794 .olm_flags = OLF_SCAN_SUBITEMS | OLF_IDX_IN_FID,
795 .olm_scan_dir = osd_ios_general_sd,
796 .olm_handle_dirent = osd_ios_uld_hd,
801 .olm_name = HSM_ACTIONS,
806 .olm_name = LUSTRE_NODEMAP_NAME,
811 .olm_name = INDEX_BACKUP_DIR,
813 .f_seq = FID_SEQ_LOCAL_FILE,
814 .f_oid = INDEX_BACKUP_OID,
816 .olm_flags = OLF_SCAN_SUBITEMS | OLF_NOT_BACKUP,
817 .olm_scan_dir = osd_ios_general_sd,
818 .olm_handle_dirent = osd_ios_varfid_hd,
826 /* Add the new introduced files under .lustre/ in the list in the future. */
827 static const struct osd_lf_map osd_dl_maps[] = {
832 .f_seq = FID_SEQ_DOT_LUSTRE,
833 .f_oid = FID_OID_DOT_LUSTRE_OBF,
837 /* .lustre/lost+found */
839 .olm_name = "lost+found",
841 .f_seq = FID_SEQ_DOT_LUSTRE,
842 .f_oid = FID_OID_DOT_LUSTRE_LPF,
851 struct osd_ios_item {
852 struct list_head oii_list;
854 enum osd_lf_flags oii_flags;
855 scan_dir_t oii_scan_dir;
856 handle_dirent_t oii_handle_dirent;
859 static int osd_ios_new_item(struct osd_device *dev, uint64_t parent,
860 enum osd_lf_flags flags, scan_dir_t scan_dir,
861 handle_dirent_t handle_dirent)
863 struct osd_ios_item *item;
867 CWARN("%s: initial OI scrub failed to add item for %llu\n",
868 osd_name(dev), parent);
872 INIT_LIST_HEAD(&item->oii_list);
873 item->oii_parent = parent;
874 item->oii_flags = flags;
875 item->oii_scan_dir = scan_dir;
876 item->oii_handle_dirent = handle_dirent;
877 list_add_tail(&item->oii_list, &dev->od_ios_list);
882 static bool osd_index_need_recreate(const struct lu_env *env,
883 struct osd_device *dev, uint64_t oid)
885 struct osd_thread_info *info = osd_oti_get(env);
886 zap_attribute_t *za = &info->oti_za2;
887 zap_cursor_t *zc = &info->oti_zc2;
891 zap_cursor_init_serialized(zc, dev->od_os, oid, 0);
892 rc = -zap_cursor_retrieve(zc, za);
894 if (rc && rc != -ENOENT)
900 static void osd_ios_index_register(const struct lu_env *env,
901 struct osd_device *osd,
902 const struct lu_fid *fid, uint64_t oid)
904 struct osd_thread_info *info = osd_oti_get(env);
905 zap_attribute_t *za = &info->oti_za2;
906 zap_cursor_t *zc = &info->oti_zc2;
907 struct zap_leaf_entry *le;
916 rc = __osd_obj2dnode(osd->od_os, oid, &dn);
917 if (rc == -EEXIST || rc == -ENOENT)
923 if (!osd_object_is_zap(dn))
926 rc = -sa_handle_get(osd->od_os, oid, NULL, SA_HDL_PRIVATE, &hdl);
930 rc = -sa_lookup(hdl, SA_ZPL_MODE(osd), &mode, sizeof(mode));
931 sa_handle_destroy(hdl);
938 zap_cursor_init_serialized(zc, osd->od_os, oid, 0);
939 rc = -zap_cursor_retrieve(zc, za);
941 /* Skip empty index object */
942 GOTO(fini, rc = (rc == -ENOENT ? 1 : rc));
944 if (zc->zc_zap->zap_ismicro ||
945 !(zap_f_phys(zc->zc_zap)->zap_flags & ZAP_FLAG_UINT64_KEY))
948 le = ZAP_LEAF_ENTRY(zc->zc_leaf, 0);
949 keysize = le->le_name_numints * 8;
950 recsize = za->za_integer_length * za->za_num_integers;
951 if (likely(keysize && recsize))
952 rc = osd_index_register(osd, fid, keysize, recsize);
963 CWARN("%s: failed to register index "DFID" (%u/%u): rc = %d\n",
964 osd_name(osd), PFID(fid), keysize, recsize, rc);
966 CDEBUG(D_LFSCK, "%s: registered index "DFID" (%u/%u)\n",
967 osd_name(osd), PFID(fid), keysize, recsize);
970 static void osd_index_restore(const struct lu_env *env, struct osd_device *dev,
971 struct lustre_index_restore_unit *liru, void *buf,
974 struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
975 struct lu_fid *tgt_fid = &liru->liru_cfid;
976 struct lu_fid bak_fid;
980 lustre_fid2lbx(buf, tgt_fid, bufsize);
981 rc = -zap_lookup(dev->od_os, dev->od_index_backup_id, buf, 8,
982 sizeof(*zde) / 8, (void *)zde);
986 rc = osd_get_fid_by_oid(env, dev, zde->lzd_reg.zde_dnode, &bak_fid);
990 /* The OI mapping for index may be invalid, since it will be
991 * re-created, not update the OI mapping, just cache it in RAM. */
992 rc = osd_idc_find_and_init_with_oid(env, dev, tgt_fid,
995 rc = lustre_index_restore(env, &dev->od_dt_dev,
996 &liru->liru_pfid, tgt_fid, &bak_fid,
997 liru->liru_name, &dev->od_index_backup_list,
998 &dev->od_lock, buf, bufsize);
1002 CDEBUG(D_WARNING, "%s: restore index '%s' with "DFID": rc = %d\n",
1003 osd_name(dev), liru->liru_name, PFID(tgt_fid), rc);
1007 * verify FID-in-LMA and OI entry for one object
1009 * ios: Initial OI Scrub.
1011 static int osd_ios_scan_one(const struct lu_env *env, struct osd_device *dev,
1012 const struct lu_fid *fid, uint64_t parent,
1013 uint64_t oid, const char *name,
1014 enum osd_lf_flags flags)
1016 struct lustre_scrub *scrub = &dev->od_scrub;
1017 struct scrub_file *sf = &scrub->os_file;
1018 struct lustre_mdt_attrs *lma = NULL;
1019 nvlist_t *nvbuf = NULL;
1028 rc = __osd_xattr_load_by_oid(dev, oid, &nvbuf);
1029 if (unlikely(rc == -ENOENT || rc == -EEXIST))
1032 if (rc && rc != -ENODATA) {
1033 CWARN("%s: initial OI scrub failed to get lma for %llu: "
1034 "rc = %d\n", osd_name(dev), oid, rc);
1040 LASSERT(nvbuf != NULL);
1041 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
1042 (uchar_t **)&lma, &size);
1043 if (rc || size == 0) {
1044 LASSERT(lma == NULL);
1047 LASSERTF(lma != NULL, "corrupted LMA, size %d\n", size);
1048 lustre_lma_swab(lma);
1049 if (lma->lma_compat & LMAC_NOT_IN_OI) {
1054 if (lma->lma_compat & LMAC_IDX_BACKUP &&
1055 osd_index_need_recreate(env, dev, oid)) {
1056 if (parent == dev->od_root) {
1057 lu_local_obj_fid(&tfid,
1060 rc = osd_get_fid_by_oid(env, dev,
1068 rc = lustre_liru_new(
1069 &dev->od_index_restore_list,
1070 &tfid, &lma->lma_self_fid, oid,
1071 name, strlen(name));
1076 tfid = lma->lma_self_fid;
1077 if (!(flags & OLF_NOT_BACKUP))
1078 osd_ios_index_register(env, dev, &tfid, oid);
1083 if (rc == -ENODATA) {
1085 /* Skip the object without FID-in-LMA */
1086 CDEBUG(D_LFSCK, "%s: %llu has no FID-in-LMA, skip it\n",
1087 osd_name(dev), oid);
1092 LASSERT(!fid_is_zero(fid));
1095 if (flags & OLF_IDX_IN_FID) {
1096 LASSERT(dev->od_index >= 0);
1098 tfid.f_oid = dev->od_index;
1102 rc = osd_fid_lookup(env, dev, &tfid, &oid2);
1104 if (rc != -ENOENT) {
1105 CWARN("%s: initial OI scrub failed to lookup fid for "
1106 DFID"=>%llu: rc = %d\n",
1107 osd_name(dev), PFID(&tfid), oid, rc);
1112 flag = SF_RECREATED;
1113 op = DTO_INDEX_INSERT;
1118 flag = SF_INCONSISTENT;
1119 op = DTO_INDEX_UPDATE;
1122 if (!(sf->sf_flags & flag)) {
1123 scrub_file_reset(scrub, dev->od_uuid, flag);
1124 rc = scrub_file_store(env, scrub);
1129 rc = osd_scrub_refresh_mapping(env, dev, &tfid, oid, op, true, name);
1131 RETURN(rc > 0 ? 0 : rc);
1134 static int osd_ios_varfid_hd(const struct lu_env *env, struct osd_device *dev,
1135 const char *name, uint64_t parent, uint64_t oid,
1136 enum osd_lf_flags flags, bool is_dir)
1141 rc = osd_ios_scan_one(env, dev, NULL, parent, oid, name, 0);
1143 rc = osd_ios_new_item(dev, oid, flags, osd_ios_general_sd,
1149 static int osd_ios_uld_hd(const struct lu_env *env, struct osd_device *dev,
1150 const char *name, uint64_t parent, uint64_t oid,
1151 enum osd_lf_flags flags, bool is_dir)
1157 /* skip any non-DFID format name */
1161 /* skip the start '[' */
1162 sscanf(&name[1], SFID, RFID(&tfid));
1163 if (fid_is_sane(&tfid))
1164 rc = osd_ios_scan_one(env, dev, &tfid, parent, oid, name, 0);
1172 * General scanner for the directories execpt /ROOT during initial OI scrub.
1173 * It scans the name entries under the given directory one by one. For each
1174 * entry, verifies its OI mapping via the given @handle_dirent.
1176 static int osd_ios_general_sd(const struct lu_env *env, struct osd_device *dev,
1177 uint64_t parent, handle_dirent_t handle_dirent,
1178 enum osd_lf_flags flags)
1180 struct osd_thread_info *info = osd_oti_get(env);
1181 struct luz_direntry *zde = &info->oti_zde;
1182 zap_attribute_t *za = &info->oti_za;
1183 zap_cursor_t *zc = &info->oti_zc;
1187 zap_cursor_init_serialized(zc, dev->od_os, parent, 0);
1188 rc = -zap_cursor_retrieve(zc, za);
1190 zap_cursor_advance(zc);
1195 rc = -zap_cursor_retrieve(zc, za);
1197 GOTO(log, rc = (rc == -ENOENT ? 0 : rc));
1199 /* skip the entry started with '.' */
1200 if (likely(za->za_name[0] != '.')) {
1201 rc = osd_zap_lookup(dev, parent, NULL, za->za_name,
1202 za->za_integer_length,
1203 sizeof(*zde) / za->za_integer_length,
1206 CWARN("%s: initial OI scrub failed to lookup "
1207 "%s under %llu: rc = %d\n",
1208 osd_name(dev), za->za_name, parent, rc);
1212 rc = handle_dirent(env, dev, za->za_name, parent,
1213 zde->lzd_reg.zde_dnode, flags,
1214 S_ISDIR(DTTOIF(zde->lzd_reg.zde_type)) ?
1216 CDEBUG(D_LFSCK, "%s: initial OI scrub handled %s under "
1218 osd_name(dev), za->za_name, parent, rc);
1221 zap_cursor_advance(zc);
1226 CWARN("%s: initial OI scrub failed to scan the directory %llu: "
1227 "rc = %d\n", osd_name(dev), parent, rc);
1228 zap_cursor_fini(zc);
1234 * The scanner for /ROOT directory. It is not all the items under /ROOT will
1235 * be scanned during the initial OI scrub, instead, only the .lustre and the
1236 * sub-items under .lustre will be handled.
1238 static int osd_ios_ROOT_sd(const struct lu_env *env, struct osd_device *dev,
1239 uint64_t parent, handle_dirent_t handle_dirent,
1240 enum osd_lf_flags flags)
1242 struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
1243 const struct osd_lf_map *map;
1249 rc = osd_zap_lookup(dev, parent, NULL, dot_lustre_name, 8,
1250 sizeof(*zde) / 8, (void *)zde);
1251 if (rc == -ENOENT) {
1252 /* The .lustre directory is lost. That is not fatal. It can
1253 * be re-created in the subsequent MDT start processing. */
1258 CWARN("%s: initial OI scrub failed to find .lustre: "
1259 "rc = %d\n", osd_name(dev), rc);
1264 oid = zde->lzd_reg.zde_dnode;
1265 rc = osd_ios_scan_one(env, dev, &LU_DOT_LUSTRE_FID, parent, oid,
1266 dot_lustre_name, 0);
1270 for (map = osd_dl_maps; map->olm_name; map++) {
1271 rc = osd_zap_lookup(dev, oid, NULL, map->olm_name, 8,
1272 sizeof(*zde) / 8, (void *)zde);
1275 CWARN("%s: initial OI scrub failed to find"
1276 "the entry %s under .lustre: rc = %d\n",
1277 osd_name(dev), map->olm_name, rc);
1278 else if (!fid_is_zero(&map->olm_fid))
1279 /* Try to remove the stale OI mapping. */
1280 osd_scrub_refresh_mapping(env, dev,
1282 DTO_INDEX_DELETE, true,
1287 rc = osd_ios_scan_one(env, dev, &map->olm_fid, oid,
1288 zde->lzd_reg.zde_dnode, map->olm_name,
1297 static void osd_initial_OI_scrub(const struct lu_env *env,
1298 struct osd_device *dev)
1300 struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
1301 const struct osd_lf_map *map;
1305 for (map = osd_lf_maps; map->olm_name; map++) {
1306 rc = osd_zap_lookup(dev, dev->od_root, NULL, map->olm_name, 8,
1307 sizeof(*zde) / 8, (void *)zde);
1310 CWARN("%s: initial OI scrub failed "
1311 "to find the entry %s: rc = %d\n",
1312 osd_name(dev), map->olm_name, rc);
1313 else if (!fid_is_zero(&map->olm_fid))
1314 /* Try to remove the stale OI mapping. */
1315 osd_scrub_refresh_mapping(env, dev,
1317 DTO_INDEX_DELETE, true,
1322 rc = osd_ios_scan_one(env, dev, &map->olm_fid, dev->od_root,
1323 zde->lzd_reg.zde_dnode, map->olm_name,
1325 if (!rc && map->olm_flags & OLF_SCAN_SUBITEMS)
1326 osd_ios_new_item(dev, zde->lzd_reg.zde_dnode,
1327 map->olm_flags, map->olm_scan_dir,
1328 map->olm_handle_dirent);
1331 while (!list_empty(&dev->od_ios_list)) {
1332 struct osd_ios_item *item;
1334 item = list_entry(dev->od_ios_list.next,
1335 struct osd_ios_item, oii_list);
1336 list_del_init(&item->oii_list);
1337 item->oii_scan_dir(env, dev, item->oii_parent,
1338 item->oii_handle_dirent, item->oii_flags);
1342 if (!list_empty(&dev->od_index_restore_list)) {
1345 OBD_ALLOC_LARGE(buf, INDEX_BACKUP_BUFSIZE);
1347 CERROR("%s: not enough RAM for rebuild index\n",
1350 while (!list_empty(&dev->od_index_restore_list)) {
1351 struct lustre_index_restore_unit *liru;
1353 liru = list_entry(dev->od_index_restore_list.next,
1354 struct lustre_index_restore_unit,
1356 list_del(&liru->liru_link);
1358 osd_index_restore(env, dev, liru, buf,
1359 INDEX_BACKUP_BUFSIZE);
1360 OBD_FREE(liru, liru->liru_len);
1364 OBD_FREE_LARGE(buf, INDEX_BACKUP_BUFSIZE);
1370 /* OI scrub start/stop */
1372 int osd_scrub_start(const struct lu_env *env, struct osd_device *dev,
1378 if (dev->od_dt_dev.dd_rdonly)
1381 /* od_otable_sem: prevent concurrent start/stop */
1382 down(&dev->od_otable_sem);
1383 rc = scrub_start(osd_scrub_main, &dev->od_scrub, dev, flags);
1384 up(&dev->od_otable_sem);
1386 RETURN(rc == -EALREADY ? 0 : rc);
1389 static void osd_scrub_stop(struct osd_device *dev)
1391 struct lustre_scrub *scrub = &dev->od_scrub;
1394 /* od_otable_sem: prevent concurrent start/stop */
1395 down(&dev->od_otable_sem);
1396 scrub->os_paused = 1;
1398 up(&dev->od_otable_sem);
1403 /* OI scrub setup/cleanup */
1405 static const char osd_scrub_name[] = "OI_scrub";
1407 int osd_scrub_setup(const struct lu_env *env, struct osd_device *dev)
1409 struct osd_thread_info *info = osd_oti_get(env);
1410 struct lustre_scrub *scrub = &dev->od_scrub;
1411 struct scrub_file *sf = &scrub->os_file;
1412 struct lu_fid *fid = &info->oti_fid;
1413 struct dt_object *obj;
1419 memcpy(dev->od_uuid,
1420 &dsl_dataset_phys(dev->od_os->os_dsl_dataset)->ds_guid,
1421 sizeof(dsl_dataset_phys(dev->od_os->os_dsl_dataset)->ds_guid));
1422 memset(&dev->od_scrub, 0, sizeof(struct lustre_scrub));
1423 init_waitqueue_head(&scrub->os_thread.t_ctl_waitq);
1424 init_rwsem(&scrub->os_rwsem);
1425 spin_lock_init(&scrub->os_lock);
1426 INIT_LIST_HEAD(&scrub->os_inconsistent_items);
1427 scrub->os_name = osd_name(dev);
1429 /* 'What the @fid is' is not imporatant, because the object
1430 * has no OI mapping, and only is visible inside the OSD.*/
1431 fid->f_seq = FID_SEQ_IGIF_MAX;
1433 fid->f_oid = ((1 << 31) | dev->od_index) + 1;
1435 fid->f_oid = dev->od_index + 1;
1437 rc = osd_obj_find_or_create(env, dev, dev->od_root,
1438 osd_scrub_name, &oid, fid, false);
1442 rc = osd_idc_find_and_init_with_oid(env, dev, fid, oid);
1446 obj = lu2dt(lu_object_find_slice(env, osd2lu_dev(dev), fid, NULL));
1447 if (IS_ERR_OR_NULL(obj))
1448 RETURN(obj ? PTR_ERR(obj) : -ENOENT);
1450 scrub->os_obj = obj;
1451 rc = scrub_file_load(env, scrub);
1452 if (rc == -ENOENT || rc == -EFAULT) {
1453 scrub_file_init(scrub, dev->od_uuid);
1455 } else if (rc < 0) {
1456 GOTO(cleanup_obj, rc);
1458 if (memcmp(sf->sf_uuid, dev->od_uuid, 16) != 0) {
1459 struct obd_uuid *old_uuid;
1460 struct obd_uuid *new_uuid;
1462 OBD_ALLOC_PTR(old_uuid);
1463 OBD_ALLOC_PTR(new_uuid);
1464 if (!old_uuid || !new_uuid) {
1465 CERROR("%s: UUID has been changed, but"
1466 "failed to allocate RAM for report\n",
1469 class_uuid_unparse(sf->sf_uuid, old_uuid);
1470 class_uuid_unparse(dev->od_uuid, new_uuid);
1471 CDEBUG(D_LFSCK, "%s: UUID has been changed "
1472 "from %s to %s\n", osd_name(dev),
1473 old_uuid->uuid, new_uuid->uuid);
1475 scrub_file_reset(scrub, dev->od_uuid, SF_INCONSISTENT);
1478 OBD_FREE_PTR(old_uuid);
1480 OBD_FREE_PTR(new_uuid);
1481 } else if (sf->sf_status == SS_SCANNING) {
1482 sf->sf_status = SS_CRASHED;
1486 if ((sf->sf_oi_count & (sf->sf_oi_count - 1)) != 0) {
1487 LCONSOLE_WARN("%s: invalid oi count %d, set it to %d\n",
1488 osd_name(dev), sf->sf_oi_count,
1490 sf->sf_oi_count = osd_oi_count;
1495 if (sf->sf_pos_last_checkpoint != 0)
1496 scrub->os_pos_current = sf->sf_pos_last_checkpoint + 1;
1498 scrub->os_pos_current = 1;
1501 rc = scrub_file_store(env, scrub);
1503 GOTO(cleanup_obj, rc);
1506 /* Initialize OI files. */
1507 rc = osd_oi_init(env, dev);
1509 GOTO(cleanup_obj, rc);
1511 if (!dev->od_dt_dev.dd_rdonly)
1512 osd_initial_OI_scrub(env, dev);
1514 if (!dev->od_dt_dev.dd_rdonly &&
1515 dev->od_auto_scrub_interval != AS_NEVER &&
1516 ((sf->sf_status == SS_PAUSED) ||
1517 (sf->sf_status == SS_CRASHED &&
1518 sf->sf_flags & (SF_RECREATED | SF_INCONSISTENT |
1519 SF_UPGRADE | SF_AUTO)) ||
1520 (sf->sf_status == SS_INIT &&
1521 sf->sf_flags & (SF_RECREATED | SF_INCONSISTENT |
1523 rc = osd_scrub_start(env, dev, SS_AUTO_FULL);
1526 GOTO(cleanup_oi, rc);
1531 osd_oi_fini(env, dev);
1533 dt_object_put_nocache(env, scrub->os_obj);
1534 scrub->os_obj = NULL;
1539 void osd_scrub_cleanup(const struct lu_env *env, struct osd_device *dev)
1541 struct lustre_scrub *scrub = &dev->od_scrub;
1543 LASSERT(!dev->od_otable_it);
1545 if (scrub->os_obj) {
1546 osd_scrub_stop(dev);
1547 dt_object_put_nocache(env, scrub->os_obj);
1548 scrub->os_obj = NULL;
1551 if (dev->od_oi_table)
1552 osd_oi_fini(env, dev);
1555 /* object table based iteration APIs */
1557 static struct dt_it *osd_otable_it_init(const struct lu_env *env,
1558 struct dt_object *dt, __u32 attr)
1560 enum dt_otable_it_flags flags = attr >> DT_OTABLE_IT_FLAGS_SHIFT;
1561 enum dt_otable_it_valid valid = attr & ~DT_OTABLE_IT_FLAGS_MASK;
1562 struct osd_device *dev = osd_dev(dt->do_lu.lo_dev);
1563 struct lustre_scrub *scrub = &dev->od_scrub;
1564 struct osd_otable_it *it;
1569 if (dev->od_dt_dev.dd_rdonly)
1570 RETURN(ERR_PTR(-EROFS));
1572 /* od_otable_sem: prevent concurrent init/fini */
1573 down(&dev->od_otable_sem);
1574 if (dev->od_otable_it)
1575 GOTO(out, it = ERR_PTR(-EALREADY));
1579 GOTO(out, it = ERR_PTR(-ENOMEM));
1581 if (flags & DOIF_OUTUSED)
1582 it->ooi_used_outside = 1;
1584 if (flags & DOIF_RESET)
1587 if (valid & DOIV_ERROR_HANDLE) {
1588 if (flags & DOIF_FAILOUT)
1589 start |= SS_SET_FAILOUT;
1591 start |= SS_CLEAR_FAILOUT;
1594 if (valid & DOIV_DRYRUN) {
1595 if (flags & DOIF_DRYRUN)
1596 start |= SS_SET_DRYRUN;
1598 start |= SS_CLEAR_DRYRUN;
1601 /* XXX: dmu_object_next() does NOT find dnodes allocated
1602 * in the current non-committed txg, so we force txg
1603 * commit to find all existing dnodes ... */
1604 txg_wait_synced(dmu_objset_pool(dev->od_os), 0ULL);
1606 dev->od_otable_it = it;
1608 rc = scrub_start(osd_scrub_main, scrub, dev, start & ~SS_AUTO_PARTIAL);
1609 if (rc == -EALREADY) {
1611 } else if (rc < 0) {
1612 dev->od_otable_it = NULL;
1616 it->ooi_pos = scrub->os_pos_current;
1622 up(&dev->od_otable_sem);
1623 return (struct dt_it *)it;
1626 static void osd_otable_it_fini(const struct lu_env *env, struct dt_it *di)
1628 struct osd_otable_it *it = (struct osd_otable_it *)di;
1629 struct osd_device *dev = it->ooi_dev;
1631 /* od_otable_sem: prevent concurrent init/fini */
1632 down(&dev->od_otable_sem);
1633 scrub_stop(&dev->od_scrub);
1634 LASSERT(dev->od_otable_it == it);
1636 dev->od_otable_it = NULL;
1637 up(&dev->od_otable_sem);
1641 static int osd_otable_it_get(const struct lu_env *env,
1642 struct dt_it *di, const struct dt_key *key)
1647 static void osd_otable_it_put(const struct lu_env *env, struct dt_it *di)
1651 static void osd_otable_it_preload(const struct lu_env *env,
1652 struct osd_otable_it *it)
1654 struct osd_device *dev = it->ooi_dev;
1657 /* can go negative on the very first access to the iterator
1658 * or if some non-Lustre objects were found */
1659 if (unlikely(it->ooi_prefetched < 0))
1660 it->ooi_prefetched = 0;
1662 if (it->ooi_prefetched >= (OTABLE_PREFETCH >> 1))
1665 if (it->ooi_prefetched_dnode == 0)
1666 it->ooi_prefetched_dnode = it->ooi_pos;
1668 while (it->ooi_prefetched < OTABLE_PREFETCH) {
1669 rc = -dmu_object_next(dev->od_os, &it->ooi_prefetched_dnode,
1674 osd_dmu_prefetch(dev->od_os, it->ooi_prefetched_dnode,
1675 0, 0, 0, ZIO_PRIORITY_ASYNC_READ);
1676 it->ooi_prefetched++;
1681 osd_otable_it_wakeup(struct lustre_scrub *scrub, struct osd_otable_it *it)
1683 spin_lock(&scrub->os_lock);
1684 if (it->ooi_pos < scrub->os_pos_current || scrub->os_waiting ||
1685 !thread_is_running(&scrub->os_thread))
1686 it->ooi_waiting = 0;
1688 it->ooi_waiting = 1;
1689 spin_unlock(&scrub->os_lock);
1691 return !it->ooi_waiting;
1694 static int osd_otable_it_next(const struct lu_env *env, struct dt_it *di)
1696 struct osd_otable_it *it = (struct osd_otable_it *)di;
1697 struct osd_device *dev = it->ooi_dev;
1698 struct lustre_scrub *scrub = &dev->od_scrub;
1699 struct ptlrpc_thread *thread = &scrub->os_thread;
1700 struct l_wait_info lwi = { 0 };
1701 struct lustre_mdt_attrs *lma = NULL;
1702 nvlist_t *nvbuf = NULL;
1707 LASSERT(it->ooi_user_ready);
1708 fid_zero(&it->ooi_fid);
1710 if (unlikely(it->ooi_all_cached))
1721 if (it->ooi_pos >= scrub->os_pos_current)
1722 l_wait_event(thread->t_ctl_waitq,
1723 osd_otable_it_wakeup(scrub, it),
1726 if (!thread_is_running(thread) && !it->ooi_used_outside)
1729 rc = -dmu_object_next(dev->od_os, &it->ooi_pos, B_FALSE, 0);
1731 if (unlikely(rc == -ESRCH)) {
1732 it->ooi_all_cached = 1;
1739 rc = __osd_xattr_load_by_oid(dev, it->ooi_pos, &nvbuf);
1741 if (!scrub->os_full_speed)
1742 spin_lock(&scrub->os_lock);
1743 it->ooi_prefetched--;
1744 if (!scrub->os_full_speed) {
1745 if (scrub->os_waiting) {
1746 scrub->os_waiting = 0;
1747 wake_up_all(&thread->t_ctl_waitq);
1749 spin_unlock(&scrub->os_lock);
1752 if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
1758 LASSERT(nvbuf != NULL);
1759 rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA,
1760 (uchar_t **)&lma, &size);
1761 if (rc || size == 0)
1762 /* It is either non-Lustre object or OSD internal object,
1763 * ignore it, go ahead */
1766 LASSERTF(lma != NULL, "corrupted LMA, size %d\n", size);
1767 lustre_lma_swab(lma);
1768 if (unlikely(lma->lma_compat & LMAC_NOT_IN_OI ||
1769 lma->lma_incompat & LMAI_AGENT))
1772 it->ooi_fid = lma->lma_self_fid;
1780 if (!rc && scrub->os_full_speed)
1781 osd_otable_it_preload(env, it);
1786 static struct dt_key *osd_otable_it_key(const struct lu_env *env,
1787 const struct dt_it *di)
1792 static int osd_otable_it_key_size(const struct lu_env *env,
1793 const struct dt_it *di)
1795 return sizeof(__u64);
1798 static int osd_otable_it_rec(const struct lu_env *env, const struct dt_it *di,
1799 struct dt_rec *rec, __u32 attr)
1801 struct osd_otable_it *it = (struct osd_otable_it *)di;
1802 struct lu_fid *fid = (struct lu_fid *)rec;
1808 static __u64 osd_otable_it_store(const struct lu_env *env,
1809 const struct dt_it *di)
1811 struct osd_otable_it *it = (struct osd_otable_it *)di;
1817 * Set the OSD layer iteration start position as the specified hash.
1819 static int osd_otable_it_load(const struct lu_env *env,
1820 const struct dt_it *di, __u64 hash)
1822 struct osd_otable_it *it = (struct osd_otable_it *)di;
1823 struct osd_device *dev = it->ooi_dev;
1824 struct lustre_scrub *scrub = &dev->od_scrub;
1828 /* Forbid to set iteration position after iteration started. */
1829 if (it->ooi_user_ready)
1832 if (hash > OSD_OTABLE_MAX_HASH)
1833 hash = OSD_OTABLE_MAX_HASH;
1835 /* The hash is the last checkpoint position,
1836 * we will start from the next one. */
1837 it->ooi_pos = hash + 1;
1838 it->ooi_prefetched = 0;
1839 it->ooi_prefetched_dnode = 0;
1840 it->ooi_user_ready = 1;
1841 if (!scrub->os_full_speed)
1842 wake_up_all(&scrub->os_thread.t_ctl_waitq);
1844 /* Unplug OSD layer iteration by the first next() call. */
1845 rc = osd_otable_it_next(env, (struct dt_it *)it);
1850 static int osd_otable_it_key_rec(const struct lu_env *env,
1851 const struct dt_it *di, void *key_rec)
1856 const struct dt_index_operations osd_otable_ops = {
1858 .init = osd_otable_it_init,
1859 .fini = osd_otable_it_fini,
1860 .get = osd_otable_it_get,
1861 .put = osd_otable_it_put,
1862 .next = osd_otable_it_next,
1863 .key = osd_otable_it_key,
1864 .key_size = osd_otable_it_key_size,
1865 .rec = osd_otable_it_rec,
1866 .store = osd_otable_it_store,
1867 .load = osd_otable_it_load,
1868 .key_rec = osd_otable_it_key_rec,
1872 /* high priority inconsistent items list APIs */
1874 int osd_oii_insert(const struct lu_env *env, struct osd_device *dev,
1875 const struct lu_fid *fid, uint64_t oid, bool insert)
1877 struct lustre_scrub *scrub = &dev->od_scrub;
1878 struct ptlrpc_thread *thread = &scrub->os_thread;
1879 struct osd_inconsistent_item *oii;
1880 bool wakeup = false;
1883 osd_idc_find_and_init_with_oid(env, dev, fid, oid);
1888 INIT_LIST_HEAD(&oii->oii_list);
1889 oii->oii_cache.oic_dev = dev;
1890 oii->oii_cache.oic_fid = *fid;
1891 oii->oii_cache.oic_dnode = oid;
1892 oii->oii_insert = insert;
1894 spin_lock(&scrub->os_lock);
1895 if (unlikely(!thread_is_running(thread))) {
1896 spin_unlock(&scrub->os_lock);
1901 if (list_empty(&scrub->os_inconsistent_items))
1903 list_add_tail(&oii->oii_list, &scrub->os_inconsistent_items);
1904 spin_unlock(&scrub->os_lock);
1907 wake_up_all(&thread->t_ctl_waitq);
1912 int osd_oii_lookup(struct osd_device *dev, const struct lu_fid *fid,
1915 struct lustre_scrub *scrub = &dev->od_scrub;
1916 struct osd_inconsistent_item *oii;
1920 spin_lock(&scrub->os_lock);
1921 list_for_each_entry(oii, &scrub->os_inconsistent_items, oii_list) {
1922 if (lu_fid_eq(fid, &oii->oii_cache.oic_fid)) {
1923 *oid = oii->oii_cache.oic_dnode;
1928 spin_unlock(&scrub->os_lock);