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, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.gnu.org/licenses/gpl-2.0.html
23 * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
24 * Use is subject to license terms.
26 * Copyright (c) 2011, 2017, Intel Corporation.
29 * This file is part of Lustre, http://www.lustre.org/
31 * lustre/mdt/mdt_reint.c
33 * Lustre Metadata Target (mdt) reintegration routines
35 * Author: Peter Braam <braam@clusterfs.com>
36 * Author: Andreas Dilger <adilger@clusterfs.com>
37 * Author: Phil Schwan <phil@clusterfs.com>
38 * Author: Huang Hua <huanghua@clusterfs.com>
39 * Author: Yury Umanets <umka@clusterfs.com>
42 #define DEBUG_SUBSYSTEM S_MDS
44 #include <lprocfs_status.h>
45 #include "mdt_internal.h"
46 #include <lustre_lmv.h>
47 #include <lustre_crypto.h>
49 static inline void mdt_reint_init_ma(struct mdt_thread_info *info,
52 ma->ma_need = MA_INODE;
57 * Get version of object by fid.
59 * Return real version or ENOENT_VERSION if object doesn't exist
61 static void mdt_obj_version_get(struct mdt_thread_info *info,
62 struct mdt_object *o, __u64 *version)
66 if (mdt_object_exists(o) && !mdt_object_remote(o) &&
67 !fid_is_obf(mdt_object_fid(o)))
68 *version = dt_version_get(info->mti_env, mdt_obj2dt(o));
70 *version = ENOENT_VERSION;
71 CDEBUG(D_INODE, "FID "DFID" version is %#llx\n",
72 PFID(mdt_object_fid(o)), *version);
76 * Check version is correct.
78 * Should be called only during replay.
80 static int mdt_version_check(struct ptlrpc_request *req,
81 __u64 version, int idx)
83 __u64 *pre_ver = lustre_msg_get_versions(req->rq_reqmsg);
86 if (!exp_connect_vbr(req->rq_export))
89 LASSERT(req_is_replay(req));
90 /** VBR: version is checked always because costs nothing */
91 LASSERT(idx < PTLRPC_NUM_VERSIONS);
92 /** Sanity check for malformed buffers */
93 if (pre_ver == NULL) {
94 CERROR("No versions in request buffer\n");
95 spin_lock(&req->rq_export->exp_lock);
96 req->rq_export->exp_vbr_failed = 1;
97 spin_unlock(&req->rq_export->exp_lock);
99 } else if (pre_ver[idx] != version) {
100 CDEBUG(D_INODE, "Version mismatch %#llx != %#llx\n",
101 pre_ver[idx], version);
102 spin_lock(&req->rq_export->exp_lock);
103 req->rq_export->exp_vbr_failed = 1;
104 spin_unlock(&req->rq_export->exp_lock);
111 * Save pre-versions in reply.
113 static void mdt_version_save(struct ptlrpc_request *req, __u64 version,
118 if (!exp_connect_vbr(req->rq_export))
121 LASSERT(!req_is_replay(req));
122 LASSERT(req->rq_repmsg != NULL);
123 reply_ver = lustre_msg_get_versions(req->rq_repmsg);
125 reply_ver[idx] = version;
129 * Save enoent version, it is needed when it is obvious that object doesn't
130 * exist, e.g. child during create.
132 static void mdt_enoent_version_save(struct mdt_thread_info *info, int idx)
134 /* save version of file name for replay, it must be ENOENT here */
135 if (!req_is_replay(mdt_info_req(info))) {
136 info->mti_ver[idx] = ENOENT_VERSION;
137 mdt_version_save(mdt_info_req(info), info->mti_ver[idx], idx);
142 * Get version from disk and save in reply buffer.
144 * Versions are saved in reply only during normal operations not replays.
146 void mdt_version_get_save(struct mdt_thread_info *info,
147 struct mdt_object *mto, int idx)
149 /* don't save versions during replay */
150 if (!req_is_replay(mdt_info_req(info))) {
151 mdt_obj_version_get(info, mto, &info->mti_ver[idx]);
152 mdt_version_save(mdt_info_req(info), info->mti_ver[idx], idx);
157 * Get version from disk and check it, no save in reply.
159 int mdt_version_get_check(struct mdt_thread_info *info,
160 struct mdt_object *mto, int idx)
162 /* only check versions during replay */
163 if (!req_is_replay(mdt_info_req(info)))
166 mdt_obj_version_get(info, mto, &info->mti_ver[idx]);
167 return mdt_version_check(mdt_info_req(info), info->mti_ver[idx], idx);
171 * Get version from disk and check if recovery or just save.
173 int mdt_version_get_check_save(struct mdt_thread_info *info,
174 struct mdt_object *mto, int idx)
178 mdt_obj_version_get(info, mto, &info->mti_ver[idx]);
179 if (req_is_replay(mdt_info_req(info)))
180 rc = mdt_version_check(mdt_info_req(info), info->mti_ver[idx],
183 mdt_version_save(mdt_info_req(info), info->mti_ver[idx], idx);
188 * Lookup with version checking.
190 * This checks version of 'name'. Many reint functions uses 'name' for child not
191 * FID, therefore we need to get object by name and check its version.
193 int mdt_lookup_version_check(struct mdt_thread_info *info,
194 struct mdt_object *p,
195 const struct lu_name *lname,
196 struct lu_fid *fid, int idx)
200 rc = mdo_lookup(info->mti_env, mdt_object_child(p), lname, fid,
202 /* Check version only during replay */
203 if (!req_is_replay(mdt_info_req(info)))
206 info->mti_ver[idx] = ENOENT_VERSION;
208 struct mdt_object *child;
210 child = mdt_object_find(info->mti_env, info->mti_mdt, fid);
211 if (likely(!IS_ERR(child))) {
212 mdt_obj_version_get(info, child, &info->mti_ver[idx]);
213 mdt_object_put(info->mti_env, child);
216 vbrc = mdt_version_check(mdt_info_req(info), info->mti_ver[idx], idx);
217 return vbrc ? vbrc : rc;
221 static int mdt_unlock_slaves(struct mdt_thread_info *mti,
222 struct mdt_object *obj,
223 struct ldlm_enqueue_info *einfo,
226 union ldlm_policy_data *policy = &mti->mti_policy;
227 struct mdt_lock_handle *lh = &mti->mti_lh[MDT_LH_LOCAL];
228 struct lustre_handle_array *slave_locks = einfo->ei_cbdata;
231 LASSERT(S_ISDIR(obj->mot_header.loh_attr));
232 LASSERT(slave_locks);
234 memset(policy, 0, sizeof(*policy));
235 policy->l_inodebits.bits = einfo->ei_inodebits;
236 mdt_lock_handle_init(lh);
237 mdt_lock_reg_init(lh, einfo->ei_mode);
238 for (i = 0; i < slave_locks->ha_count; i++) {
239 if (test_bit(i, (void *)slave_locks->ha_map))
240 lh->mlh_rreg_lh = slave_locks->ha_handles[i];
242 lh->mlh_reg_lh = slave_locks->ha_handles[i];
243 mdt_object_unlock(mti, NULL, lh, decref);
244 slave_locks->ha_handles[i].cookie = 0ull;
247 return mo_object_unlock(mti->mti_env, mdt_object_child(obj), einfo,
251 static inline int mdt_object_striped(struct mdt_thread_info *mti,
252 struct mdt_object *obj)
254 struct lu_device *bottom_dev;
255 struct lu_object *bottom_obj;
258 if (!S_ISDIR(obj->mot_header.loh_attr))
261 /* getxattr from bottom obj to avoid reading in shard FIDs */
262 bottom_dev = dt2lu_dev(mti->mti_mdt->mdt_bottom);
263 bottom_obj = lu_object_find_slice(mti->mti_env, bottom_dev,
264 mdt_object_fid(obj), NULL);
265 if (IS_ERR(bottom_obj))
266 return PTR_ERR(bottom_obj);
268 rc = dt_xattr_get(mti->mti_env, lu2dt(bottom_obj), &LU_BUF_NULL,
270 lu_object_put(mti->mti_env, bottom_obj);
272 return (rc > 0) ? 1 : (rc == -ENODATA) ? 0 : rc;
276 * Lock slave stripes if necessary, the lock handles of slave stripes
277 * will be stored in einfo->ei_cbdata.
279 static int mdt_lock_slaves(struct mdt_thread_info *mti, struct mdt_object *obj,
280 enum ldlm_mode mode, __u64 ibits,
281 struct ldlm_enqueue_info *einfo)
283 union ldlm_policy_data *policy = &mti->mti_policy;
285 LASSERT(S_ISDIR(obj->mot_header.loh_attr));
287 einfo->ei_type = LDLM_IBITS;
288 einfo->ei_mode = mode;
289 einfo->ei_cb_bl = mdt_remote_blocking_ast;
290 einfo->ei_cb_local_bl = mdt_blocking_ast;
291 einfo->ei_cb_cp = ldlm_completion_ast;
292 einfo->ei_enq_slave = 1;
293 einfo->ei_namespace = mti->mti_mdt->mdt_namespace;
294 einfo->ei_inodebits = ibits;
295 memset(policy, 0, sizeof(*policy));
296 policy->l_inodebits.bits = ibits;
298 return mo_object_lock(mti->mti_env, mdt_object_child(obj), NULL, einfo,
302 int mdt_reint_striped_lock(struct mdt_thread_info *info,
303 struct mdt_object *o,
304 struct mdt_lock_handle *lh,
306 struct ldlm_enqueue_info *einfo,
311 LASSERT(!mdt_object_remote(o));
313 memset(einfo, 0, sizeof(*einfo));
315 rc = mdt_reint_object_lock(info, o, lh, ibits, cos_incompat);
319 rc = mdt_object_striped(info, o);
322 mdt_object_unlock(info, o, lh, rc);
326 rc = mdt_lock_slaves(info, o, lh->mlh_reg_mode, ibits, einfo);
328 mdt_object_unlock(info, o, lh, rc);
329 if (rc == -EIO && OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_SLAVE_NAME))
336 void mdt_reint_striped_unlock(struct mdt_thread_info *info,
337 struct mdt_object *o,
338 struct mdt_lock_handle *lh,
339 struct ldlm_enqueue_info *einfo, int decref)
341 if (einfo->ei_cbdata)
342 mdt_unlock_slaves(info, o, einfo, decref);
343 mdt_object_unlock(info, o, lh, decref);
346 static int mdt_restripe(struct mdt_thread_info *info,
347 struct mdt_object *parent,
348 const struct lu_name *lname,
349 const struct lu_fid *tfid,
350 struct md_op_spec *spec,
353 struct mdt_device *mdt = info->mti_mdt;
354 struct lu_fid *fid = &info->mti_tmp_fid2;
355 struct ldlm_enqueue_info *einfo = &info->mti_einfo[0];
356 struct lmv_user_md *lum = spec->u.sp_ea.eadata;
357 struct lmv_mds_md_v1 *lmv;
358 struct mdt_object *child;
359 struct mdt_lock_handle *lhp;
360 struct mdt_lock_handle *lhc;
361 struct mdt_body *repbody;
365 if (!mdt->mdt_enable_dir_restripe)
369 lum->lum_hash_type |= cpu_to_le32(LMV_HASH_FLAG_FIXED);
371 rc = mdt_version_get_check_save(info, parent, 0);
375 lhp = &info->mti_lh[MDT_LH_PARENT];
376 mdt_lock_pdo_init(lhp, LCK_PW, lname);
377 rc = mdt_reint_object_lock(info, parent, lhp, MDS_INODELOCK_UPDATE,
382 rc = mdt_stripe_get(info, parent, ma, XATTR_NAME_LMV);
384 GOTO(unlock_parent, rc);
386 if (ma->ma_valid & MA_LMV) {
387 /* don't allow restripe if parent dir layout is changing */
388 lmv = &ma->ma_lmv->lmv_md_v1;
389 if (!lmv_is_sane2(lmv))
390 GOTO(unlock_parent, rc = -EBADF);
392 if (lmv_is_layout_changing(lmv))
393 GOTO(unlock_parent, rc = -EBUSY);
397 rc = mdt_lookup_version_check(info, parent, lname, fid, 1);
399 GOTO(unlock_parent, rc);
401 child = mdt_object_find(info->mti_env, mdt, fid);
403 GOTO(unlock_parent, rc = PTR_ERR(child));
405 if (!mdt_object_exists(child))
406 GOTO(out_child, rc = -ENOENT);
408 if (mdt_object_remote(child)) {
409 struct mdt_body *repbody;
411 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
413 GOTO(out_child, rc = -EPROTO);
415 repbody->mbo_fid1 = *fid;
416 repbody->mbo_valid |= (OBD_MD_FLID | OBD_MD_MDS);
417 GOTO(out_child, rc = -EREMOTE);
420 if (!S_ISDIR(lu_object_attr(&child->mot_obj)))
421 GOTO(out_child, rc = -ENOTDIR);
423 rc = mdt_stripe_get(info, child, ma, XATTR_NAME_LMV);
427 /* race with migrate? */
428 if ((ma->ma_valid & MA_LMV) &&
429 lmv_is_migrating(&ma->ma_lmv->lmv_md_v1))
430 GOTO(out_child, rc = -EBUSY);
433 lhc = &info->mti_lh[MDT_LH_CHILD];
434 mdt_lock_reg_init(lhc, LCK_EX);
436 /* enqueue object remote LOOKUP lock */
437 if (mdt_object_remote(parent)) {
438 rc = mdt_remote_object_lock(info, parent, fid,
441 MDS_INODELOCK_LOOKUP, false);
446 rc = mdt_reint_striped_lock(info, child, lhc, MDS_INODELOCK_FULL, einfo,
449 GOTO(unlock_child, rc);
451 tgt_vbr_obj_set(info->mti_env, mdt_obj2dt(child));
452 rc = mdt_version_get_check_save(info, child, 1);
454 GOTO(unlock_child, rc);
456 spin_lock(&mdt->mdt_restriper.mdr_lock);
457 if (child->mot_restriping) {
459 spin_unlock(&mdt->mdt_restriper.mdr_lock);
460 GOTO(unlock_child, rc = -EBUSY);
462 child->mot_restriping = 1;
463 spin_unlock(&mdt->mdt_restriper.mdr_lock);
466 rc = mdt_restripe_internal(info, parent, child, lname, fid, spec, ma);
468 GOTO(restriping_clear, rc);
470 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
472 GOTO(restriping_clear, rc = -EPROTO);
474 mdt_pack_attr2body(info, repbody, &ma->ma_attr, fid);
478 child->mot_restriping = 0;
480 mdt_reint_striped_unlock(info, child, lhc, einfo, rc);
482 mdt_object_put(info->mti_env, child);
484 mdt_object_unlock(info, parent, lhp, rc);
490 * VBR: we save three versions in reply:
491 * 0 - parent. Check that parent version is the same during replay.
492 * 1 - name. Version of 'name' if file exists with the same name or
493 * ENOENT_VERSION, it is needed because file may appear due to missed replays.
494 * 2 - child. Version of child by FID. Must be ENOENT. It is mostly sanity
497 static int mdt_create(struct mdt_thread_info *info)
499 struct mdt_device *mdt = info->mti_mdt;
500 struct mdt_object *parent;
501 struct mdt_object *child;
502 struct mdt_lock_handle *lh;
503 struct mdt_body *repbody;
504 struct md_attr *ma = &info->mti_attr;
505 struct mdt_reint_record *rr = &info->mti_rr;
506 struct md_op_spec *spec = &info->mti_spec;
507 bool restripe = false;
511 DEBUG_REQ(D_INODE, mdt_info_req(info),
512 "Create ("DNAME"->"DFID") in "DFID,
513 PNAME(&rr->rr_name), PFID(rr->rr_fid2), PFID(rr->rr_fid1));
515 if (!fid_is_md_operative(rr->rr_fid1))
518 if (S_ISDIR(ma->ma_attr.la_mode) &&
519 spec->u.sp_ea.eadata != NULL && spec->u.sp_ea.eadatalen != 0) {
520 const struct lmv_user_md *lum = spec->u.sp_ea.eadata;
521 struct lu_ucred *uc = mdt_ucred(info);
522 struct obd_export *exp = mdt_info_req(info)->rq_export;
524 /* Only new clients can create remote dir( >= 2.4) and
525 * striped dir(>= 2.6), old client will return -ENOTSUPP
527 if (!mdt_is_dne_client(exp))
530 if (le32_to_cpu(lum->lum_stripe_count) > 1) {
531 if (!mdt_is_striped_client(exp))
534 if (!mdt->mdt_enable_striped_dir)
536 } else if (!mdt->mdt_enable_remote_dir) {
540 if ((!(exp_connect_flags2(exp) & OBD_CONNECT2_CRUSH)) &&
541 (le32_to_cpu(lum->lum_hash_type) & LMV_HASH_TYPE_MASK) ==
545 if (!cap_raised(uc->uc_cap, CAP_SYS_ADMIN) &&
546 uc->uc_gid != mdt->mdt_enable_remote_dir_gid &&
547 mdt->mdt_enable_remote_dir_gid != -1)
550 /* restripe if later found dir exists, MDS_OPEN_CREAT means
551 * this is create only, don't try restripe.
553 if (mdt->mdt_enable_dir_restripe &&
554 le32_to_cpu(lum->lum_stripe_offset) == LMV_OFFSET_DEFAULT &&
555 !(spec->sp_cr_flags & MDS_OPEN_CREAT))
559 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
561 parent = mdt_object_find(info->mti_env, info->mti_mdt, rr->rr_fid1);
563 RETURN(PTR_ERR(parent));
565 if (!mdt_object_exists(parent))
566 GOTO(put_parent, rc = -ENOENT);
569 * LU-10235: check if name exists locklessly first to avoid massive
570 * lock recalls on existing directories.
572 rc = mdt_lookup_version_check(info, parent, &rr->rr_name,
573 &info->mti_tmp_fid1, 1);
576 GOTO(put_parent, rc = -EEXIST);
578 rc = mdt_restripe(info, parent, &rr->rr_name, rr->rr_fid2, spec,
582 /* -ENOENT is expected here */
584 GOTO(put_parent, rc);
586 /* save version of file name for replay, it must be ENOENT here */
587 mdt_enoent_version_save(info, 1);
589 OBD_RACE(OBD_FAIL_MDS_CREATE_RACE);
591 lh = &info->mti_lh[MDT_LH_PARENT];
592 mdt_lock_pdo_init(lh, LCK_PW, &rr->rr_name);
593 rc = mdt_object_lock(info, parent, lh, MDS_INODELOCK_UPDATE);
595 GOTO(put_parent, rc);
597 if (!mdt_object_remote(parent)) {
598 rc = mdt_version_get_check_save(info, parent, 0);
600 GOTO(unlock_parent, rc);
603 child = mdt_object_new(info->mti_env, mdt, rr->rr_fid2);
604 if (unlikely(IS_ERR(child)))
605 GOTO(unlock_parent, rc = PTR_ERR(child));
607 ma->ma_need = MA_INODE;
610 mdt_fail_write(info->mti_env, info->mti_mdt->mdt_bottom,
611 OBD_FAIL_MDS_REINT_CREATE_WRITE);
613 /* Version of child will be updated on disk. */
614 tgt_vbr_obj_set(info->mti_env, mdt_obj2dt(child));
615 rc = mdt_version_get_check_save(info, child, 2);
620 * Do not perform lookup sanity check. We know that name does
623 info->mti_spec.sp_cr_lookup = 0;
624 info->mti_spec.sp_feat = &dt_directory_features;
626 rc = mdo_create(info->mti_env, mdt_object_child(parent), &rr->rr_name,
627 mdt_object_child(child), &info->mti_spec, ma);
629 rc = mdt_attr_get_complex(info, child, ma);
635 * On DNE, we need to eliminate dependey between 'mkdir a' and
636 * 'mkdir a/b' if b is a striped directory, to achieve this, two
637 * things are done below:
638 * 1. save child and slaves lock.
639 * 2. if the child is a striped directory, relock parent so to
640 * compare against with COS locks to ensure parent was
643 if (mdt_slc_is_enabled(mdt) && S_ISDIR(ma->ma_attr.la_mode)) {
644 struct mdt_lock_handle *lhc;
645 struct ldlm_enqueue_info *einfo = &info->mti_einfo[0];
648 rc = mdt_object_striped(info, child);
654 if (!mdt_object_remote(parent)) {
655 mdt_object_unlock(info, parent, lh, 1);
656 mdt_lock_pdo_init(lh, LCK_PW, &rr->rr_name);
657 rc = mdt_reint_object_lock(info, parent, lh,
658 MDS_INODELOCK_UPDATE,
665 lhc = &info->mti_lh[MDT_LH_CHILD];
666 mdt_lock_handle_init(lhc);
667 mdt_lock_reg_init(lhc, LCK_PW);
668 rc = mdt_reint_striped_lock(info, child, lhc,
669 MDS_INODELOCK_UPDATE, einfo,
674 mdt_reint_striped_unlock(info, child, lhc, einfo, rc);
677 /* Return fid & attr to client. */
678 if (ma->ma_valid & MA_INODE)
679 mdt_pack_attr2body(info, repbody, &ma->ma_attr,
680 mdt_object_fid(child));
683 mdt_object_put(info->mti_env, child);
685 mdt_object_unlock(info, parent, lh, rc);
687 mdt_object_put(info->mti_env, parent);
691 static int mdt_attr_set(struct mdt_thread_info *info, struct mdt_object *mo,
694 struct mdt_lock_handle *lh;
695 int do_vbr = ma->ma_attr.la_valid &
696 (LA_MODE | LA_UID | LA_GID | LA_PROJID | LA_FLAGS);
697 __u64 lockpart = MDS_INODELOCK_UPDATE;
698 struct ldlm_enqueue_info *einfo = &info->mti_einfo[0];
703 rc = mdt_object_striped(info, mo);
709 lh = &info->mti_lh[MDT_LH_PARENT];
710 mdt_lock_reg_init(lh, LCK_PW);
712 /* Even though the new MDT will grant PERM lock to the old
713 * client, but the old client will almost ignore that during
714 * So it needs to revoke both LOOKUP and PERM lock here, so
715 * both new and old client can cancel the dcache
717 if (ma->ma_attr.la_valid & (LA_MODE|LA_UID|LA_GID))
718 lockpart |= MDS_INODELOCK_LOOKUP | MDS_INODELOCK_PERM;
720 rc = mdt_reint_striped_lock(info, mo, lh, lockpart, einfo,
725 /* all attrs are packed into mti_attr in unpack_setattr */
726 mdt_fail_write(info->mti_env, info->mti_mdt->mdt_bottom,
727 OBD_FAIL_MDS_REINT_SETATTR_WRITE);
729 /* VBR: update version if attr changed are important for recovery */
731 /* update on-disk version of changed object */
732 tgt_vbr_obj_set(info->mti_env, mdt_obj2dt(mo));
733 rc = mdt_version_get_check_save(info, mo, 0);
735 GOTO(out_unlock, rc);
738 /* Ensure constant striping during chown(). See LU-2789. */
739 if (ma->ma_attr.la_valid & (LA_UID|LA_GID|LA_PROJID))
740 mutex_lock(&mo->mot_lov_mutex);
742 /* all attrs are packed into mti_attr in unpack_setattr */
743 rc = mo_attr_set(info->mti_env, mdt_object_child(mo), ma);
745 if (ma->ma_attr.la_valid & (LA_UID|LA_GID|LA_PROJID))
746 mutex_unlock(&mo->mot_lov_mutex);
749 GOTO(out_unlock, rc);
750 mdt_dom_obj_lvb_update(info->mti_env, mo, NULL, false);
753 mdt_reint_striped_unlock(info, mo, lh, einfo, rc);
758 * Check HSM flags and add HS_DIRTY flag if relevant.
760 * A file could be set dirty only if it has a copy in the backend (HS_EXISTS)
761 * and is not RELEASED.
763 int mdt_add_dirty_flag(struct mdt_thread_info *info, struct mdt_object *mo,
766 struct lu_ucred *uc = mdt_ucred(info);
767 kernel_cap_t cap_saved;
771 /* If the file was modified, add the dirty flag */
772 ma->ma_need = MA_HSM;
773 rc = mdt_attr_get_complex(info, mo, ma);
775 CERROR("file attribute read error for "DFID": %d.\n",
776 PFID(mdt_object_fid(mo)), rc);
780 /* If an up2date copy exists in the backend, add dirty flag */
781 if ((ma->ma_valid & MA_HSM) && (ma->ma_hsm.mh_flags & HS_EXISTS)
782 && !(ma->ma_hsm.mh_flags & (HS_DIRTY|HS_RELEASED))) {
783 ma->ma_hsm.mh_flags |= HS_DIRTY;
785 /* Bump cap so that closes from non-owner writers can
786 * set the HSM state to dirty.
788 cap_saved = uc->uc_cap;
789 cap_raise(uc->uc_cap, CAP_FOWNER);
790 rc = mdt_hsm_attr_set(info, mo, &ma->ma_hsm);
791 uc->uc_cap = cap_saved;
793 CERROR("file attribute change error for "DFID": %d\n",
794 PFID(mdt_object_fid(mo)), rc);
800 static int mdt_reint_setattr(struct mdt_thread_info *info,
801 struct mdt_lock_handle *lhc)
803 struct mdt_device *mdt = info->mti_mdt;
804 struct md_attr *ma = &info->mti_attr;
805 struct mdt_reint_record *rr = &info->mti_rr;
806 struct ptlrpc_request *req = mdt_info_req(info);
807 struct mdt_object *mo;
808 struct mdt_body *repbody;
809 ktime_t kstart = ktime_get();
813 DEBUG_REQ(D_INODE, req, "setattr "DFID" %x", PFID(rr->rr_fid1),
814 (unsigned int)ma->ma_attr.la_valid);
816 if (info->mti_dlm_req)
817 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
819 OBD_RACE(OBD_FAIL_PTLRPC_RESEND_RACE);
821 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
822 mo = mdt_object_find(info->mti_env, mdt, rr->rr_fid1);
824 GOTO(out, rc = PTR_ERR(mo));
826 if (!mdt_object_exists(mo))
827 GOTO(out_put, rc = -ENOENT);
829 if (mdt_object_remote(mo))
830 GOTO(out_put, rc = -EREMOTE);
832 ma->ma_enable_chprojid_gid = mdt->mdt_enable_chprojid_gid;
833 /* revoke lease lock if size is going to be changed */
834 if (unlikely(ma->ma_attr.la_valid & LA_SIZE &&
835 !(ma->ma_attr_flags & MDS_TRUNC_KEEP_LEASE) &&
836 atomic_read(&mo->mot_lease_count) > 0)) {
837 down_read(&mo->mot_open_sem);
839 if (atomic_read(&mo->mot_lease_count) > 0) { /* lease exists */
840 lhc = &info->mti_lh[MDT_LH_LOCAL];
841 mdt_lock_reg_init(lhc, LCK_CW);
843 rc = mdt_object_lock(info, mo, lhc, MDS_INODELOCK_OPEN);
845 up_read(&mo->mot_open_sem);
849 /* revoke lease lock */
850 mdt_object_unlock(info, mo, lhc, 1);
852 up_read(&mo->mot_open_sem);
855 if (ma->ma_attr.la_valid & LA_SIZE || rr->rr_flags & MRF_OPEN_TRUNC) {
856 /* Check write access for the O_TRUNC case */
857 if (mdt_write_read(mo) < 0)
858 GOTO(out_put, rc = -ETXTBSY);
860 /* LU-10286: compatibility check for FLR.
861 * Please check the comment in mdt_finish_open() for details
863 if (!exp_connect_flr(info->mti_exp) ||
864 !exp_connect_overstriping(info->mti_exp)) {
865 rc = mdt_big_xattr_get(info, mo, XATTR_NAME_LOV);
866 if (rc < 0 && rc != -ENODATA)
869 if (!exp_connect_flr(info->mti_exp)) {
871 mdt_lmm_is_flr(info->mti_big_lmm))
872 GOTO(out_put, rc = -EOPNOTSUPP);
875 if (!exp_connect_overstriping(info->mti_exp)) {
877 mdt_lmm_is_overstriping(info->mti_big_lmm))
878 GOTO(out_put, rc = -EOPNOTSUPP);
882 /* For truncate, the file size sent from client
883 * is believable, but the blocks are incorrect,
884 * which makes the block size in LSOM attribute
885 * inconsisent with the real block size.
887 rc = mdt_lsom_update(info, mo, true);
892 if ((ma->ma_valid & MA_INODE) && ma->ma_attr.la_valid) {
893 if (ma->ma_valid & MA_LOV)
894 GOTO(out_put, rc = -EPROTO);
896 /* MDT supports FMD for regular files due to Data-on-MDT */
897 if (S_ISREG(lu_object_attr(&mo->mot_obj)) &&
898 ma->ma_attr.la_valid & (LA_ATIME | LA_MTIME | LA_CTIME)) {
899 tgt_fmd_update(info->mti_exp, mdt_object_fid(mo),
902 if (ma->ma_attr.la_valid & LA_MTIME) {
903 rc = mdt_attr_get_pfid(info, mo, &ma->ma_pfid);
905 ma->ma_valid |= MA_PFID;
909 rc = mdt_attr_set(info, mo, ma);
912 } else if ((ma->ma_valid & (MA_LOV | MA_LMV)) &&
913 (ma->ma_valid & MA_INODE)) {
914 struct lu_buf *buf = &info->mti_buf;
915 struct lu_ucred *uc = mdt_ucred(info);
916 struct mdt_lock_handle *lh;
918 __u64 lockpart = MDS_INODELOCK_XATTR;
920 /* reject if either remote or striped dir is disabled */
921 if (ma->ma_valid & MA_LMV) {
922 if (!mdt->mdt_enable_remote_dir ||
923 !mdt->mdt_enable_striped_dir)
924 GOTO(out_put, rc = -EPERM);
926 if (!cap_raised(uc->uc_cap, CAP_SYS_ADMIN) &&
927 uc->uc_gid != mdt->mdt_enable_remote_dir_gid &&
928 mdt->mdt_enable_remote_dir_gid != -1)
929 GOTO(out_put, rc = -EPERM);
932 if (!S_ISDIR(lu_object_attr(&mo->mot_obj)))
933 GOTO(out_put, rc = -ENOTDIR);
935 if (ma->ma_attr.la_valid != 0)
936 GOTO(out_put, rc = -EPROTO);
938 if (ma->ma_valid & MA_LOV) {
939 buf->lb_buf = ma->ma_lmm;
940 buf->lb_len = ma->ma_lmm_size;
941 name = XATTR_NAME_LOV;
943 struct lmv_user_md *lmu = &ma->ma_lmv->lmv_user_md;
946 buf->lb_len = ma->ma_lmv_size;
947 name = XATTR_NAME_DEFAULT_LMV;
948 /* force client to update dir default layout */
949 lockpart |= MDS_INODELOCK_LOOKUP;
952 lh = &info->mti_lh[MDT_LH_PARENT];
953 mdt_lock_reg_init(lh, LCK_PW);
955 rc = mdt_object_lock(info, mo, lh, lockpart);
959 rc = mo_xattr_set(info->mti_env, mdt_object_child(mo), buf,
962 mdt_object_unlock(info, mo, lh, rc);
966 GOTO(out_put, rc = -EPROTO);
969 /* If file data is modified, add the dirty flag */
970 if (ma->ma_attr_flags & MDS_DATA_MODIFIED)
971 rc = mdt_add_dirty_flag(info, mo, ma);
973 ma->ma_need = MA_INODE;
975 rc = mdt_attr_get_complex(info, mo, ma);
979 mdt_pack_attr2body(info, repbody, &ma->ma_attr, mdt_object_fid(mo));
983 mdt_object_put(info->mti_env, mo);
986 mdt_counter_incr(req, LPROC_MDT_SETATTR,
987 ktime_us_delta(ktime_get(), kstart));
989 mdt_client_compatibility(info);
990 rc2 = mdt_fix_reply(info);
996 static int mdt_reint_create(struct mdt_thread_info *info,
997 struct mdt_lock_handle *lhc)
999 struct ptlrpc_request *req = mdt_info_req(info);
1000 ktime_t kstart = ktime_get();
1004 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_REINT_CREATE))
1005 RETURN(err_serious(-ESTALE));
1007 if (info->mti_dlm_req)
1008 ldlm_request_cancel(mdt_info_req(info),
1009 info->mti_dlm_req, 0, LATF_SKIP);
1011 if (!lu_name_is_valid(&info->mti_rr.rr_name))
1014 switch (info->mti_attr.ma_attr.la_mode & S_IFMT) {
1024 CERROR("%s: Unsupported mode %o\n",
1025 mdt_obd_name(info->mti_mdt),
1026 info->mti_attr.ma_attr.la_mode);
1027 RETURN(err_serious(-EOPNOTSUPP));
1030 rc = mdt_create(info);
1032 if ((info->mti_attr.ma_attr.la_mode & S_IFMT) == S_IFDIR)
1033 mdt_counter_incr(req, LPROC_MDT_MKDIR,
1034 ktime_us_delta(ktime_get(), kstart));
1036 /* Special file should stay on the same node as parent*/
1037 mdt_counter_incr(req, LPROC_MDT_MKNOD,
1038 ktime_us_delta(ktime_get(), kstart));
1045 * VBR: save parent version in reply and child version getting by its name.
1046 * Version of child is getting and checking during its lookup. If
1048 static int mdt_reint_unlink(struct mdt_thread_info *info,
1049 struct mdt_lock_handle *lhc)
1051 struct mdt_reint_record *rr = &info->mti_rr;
1052 struct ptlrpc_request *req = mdt_info_req(info);
1053 struct md_attr *ma = &info->mti_attr;
1054 struct lu_fid *child_fid = &info->mti_tmp_fid1;
1055 struct mdt_object *mp;
1056 struct mdt_object *mc;
1057 struct mdt_lock_handle *parent_lh;
1058 struct mdt_lock_handle *child_lh;
1059 struct ldlm_enqueue_info *einfo = &info->mti_einfo[0];
1061 bool cos_incompat = false;
1063 ktime_t kstart = ktime_get();
1067 DEBUG_REQ(D_INODE, req, "unlink "DFID"/"DNAME"", PFID(rr->rr_fid1),
1068 PNAME(&rr->rr_name));
1070 if (info->mti_dlm_req)
1071 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
1073 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_REINT_UNLINK))
1074 RETURN(err_serious(-ENOENT));
1076 if (!fid_is_md_operative(rr->rr_fid1))
1079 mp = mdt_object_find(info->mti_env, info->mti_mdt, rr->rr_fid1);
1081 RETURN(PTR_ERR(mp));
1083 if (mdt_object_remote(mp)) {
1084 cos_incompat = true;
1086 rc = mdt_version_get_check_save(info, mp, 0);
1088 GOTO(put_parent, rc);
1091 OBD_RACE(OBD_FAIL_MDS_REINT_OPEN);
1092 OBD_RACE(OBD_FAIL_MDS_REINT_OPEN2);
1094 parent_lh = &info->mti_lh[MDT_LH_PARENT];
1095 mdt_lock_pdo_init(parent_lh, LCK_PW, &rr->rr_name);
1096 rc = mdt_reint_object_lock(info, mp, parent_lh, MDS_INODELOCK_UPDATE,
1099 GOTO(put_parent, rc);
1101 if (info->mti_spec.sp_cr_flags & MDS_OP_WITH_FID) {
1102 *child_fid = *rr->rr_fid2;
1104 /* lookup child object along with version checking */
1105 fid_zero(child_fid);
1106 rc = mdt_lookup_version_check(info, mp, &rr->rr_name, child_fid,
1109 /* Name might not be able to find during resend of
1110 * remote unlink, considering following case.
1111 * dir_A is a remote directory, the name entry of
1112 * dir_A is on MDT0, the directory is on MDT1,
1114 * 1. client sends unlink req to MDT1.
1115 * 2. MDT1 sends name delete update to MDT0.
1116 * 3. name entry is being deleted in MDT0 synchronously.
1117 * 4. MDT1 is restarted.
1118 * 5. client resends unlink req to MDT1. So it can not
1119 * find the name entry on MDT0 anymore.
1120 * In this case, MDT1 only needs to destory the local
1123 if (mdt_object_remote(mp) && rc == -ENOENT &&
1124 !fid_is_zero(rr->rr_fid2) &&
1125 lustre_msg_get_flags(req->rq_reqmsg) & MSG_RESENT) {
1127 *child_fid = *rr->rr_fid2;
1129 GOTO(unlock_parent, rc);
1134 if (!fid_is_md_operative(child_fid))
1135 GOTO(unlock_parent, rc = -EPERM);
1137 /* We will lock the child regardless it is local or remote. No harm. */
1138 mc = mdt_object_find(info->mti_env, info->mti_mdt, child_fid);
1140 GOTO(unlock_parent, rc = PTR_ERR(mc));
1142 if (info->mti_spec.sp_cr_flags & MDS_OP_WITH_FID) {
1143 /* In this case, child fid is embedded in the request, and we do
1144 * not have a proper name as rr_name contains an encoded
1145 * hash. So find name that matches provided hash.
1147 if (!find_name_matching_hash(info, &rr->rr_name,
1149 GOTO(put_child, rc = -ENOENT);
1152 if (!cos_incompat) {
1153 rc = mdt_object_striped(info, mc);
1155 GOTO(put_child, rc);
1159 mdt_object_put(info->mti_env, mc);
1160 mdt_object_unlock(info, mp, parent_lh, -EAGAIN);
1165 child_lh = &info->mti_lh[MDT_LH_CHILD];
1166 mdt_lock_reg_init(child_lh, LCK_EX);
1167 if (info->mti_spec.sp_rm_entry) {
1168 struct lu_ucred *uc = mdt_ucred(info);
1170 if (!mdt_is_dne_client(req->rq_export))
1171 /* Return -ENOTSUPP for old client */
1172 GOTO(put_child, rc = -ENOTSUPP);
1174 if (!cap_raised(uc->uc_cap, CAP_SYS_ADMIN))
1175 GOTO(put_child, rc = -EPERM);
1177 ma->ma_need = MA_INODE;
1179 rc = mdo_unlink(info->mti_env, mdt_object_child(mp),
1180 NULL, &rr->rr_name, ma, no_name);
1181 GOTO(put_child, rc);
1184 if (mdt_object_remote(mc)) {
1185 struct mdt_body *repbody;
1187 if (!fid_is_zero(rr->rr_fid2)) {
1188 CDEBUG(D_INFO, "%s: name "DNAME" cannot find "DFID"\n",
1189 mdt_obd_name(info->mti_mdt),
1190 PNAME(&rr->rr_name), PFID(mdt_object_fid(mc)));
1191 GOTO(put_child, rc = -ENOENT);
1193 CDEBUG(D_INFO, "%s: name "DNAME": "DFID" is on another MDT\n",
1194 mdt_obd_name(info->mti_mdt),
1195 PNAME(&rr->rr_name), PFID(mdt_object_fid(mc)));
1197 if (!mdt_is_dne_client(req->rq_export))
1198 /* Return -ENOTSUPP for old client */
1199 GOTO(put_child, rc = -ENOTSUPP);
1201 /* Revoke the LOOKUP lock of the remote object granted by
1202 * this MDT. Since the unlink will happen on another MDT,
1203 * it will release the LOOKUP lock right away. Then What
1204 * would happen if another client try to grab the LOOKUP
1205 * lock at the same time with unlink XXX
1207 mdt_object_lock(info, mc, child_lh, MDS_INODELOCK_LOOKUP);
1208 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
1209 LASSERT(repbody != NULL);
1210 repbody->mbo_fid1 = *mdt_object_fid(mc);
1211 repbody->mbo_valid |= (OBD_MD_FLID | OBD_MD_MDS);
1212 GOTO(unlock_child, rc = -EREMOTE);
1214 /* We used to acquire MDS_INODELOCK_FULL here but we can't do
1215 * this now because a running HSM restore on the child (unlink
1216 * victim) will hold the layout lock. See LU-4002.
1218 lock_ibits = MDS_INODELOCK_LOOKUP | MDS_INODELOCK_UPDATE;
1219 if (mdt_object_remote(mp)) {
1220 /* Enqueue lookup lock from parent MDT */
1221 rc = mdt_remote_object_lock(info, mp, mdt_object_fid(mc),
1222 &child_lh->mlh_rreg_lh,
1223 child_lh->mlh_rreg_mode,
1224 MDS_INODELOCK_LOOKUP, false);
1226 GOTO(put_child, rc);
1228 lock_ibits &= ~MDS_INODELOCK_LOOKUP;
1231 rc = mdt_reint_striped_lock(info, mc, child_lh, lock_ibits, einfo,
1234 GOTO(put_child, rc);
1237 * Now we can only make sure we need MA_INODE, in mdd layer, will check
1238 * whether need MA_LOV and MA_COOKIE.
1240 ma->ma_need = MA_INODE;
1243 mdt_fail_write(info->mti_env, info->mti_mdt->mdt_bottom,
1244 OBD_FAIL_MDS_REINT_UNLINK_WRITE);
1245 /* save version when object is locked */
1246 mdt_version_get_save(info, mc, 1);
1248 mutex_lock(&mc->mot_lov_mutex);
1250 rc = mdo_unlink(info->mti_env, mdt_object_child(mp),
1251 mdt_object_child(mc), &rr->rr_name, ma, no_name);
1253 mutex_unlock(&mc->mot_lov_mutex);
1255 GOTO(unlock_child, rc);
1257 if (!lu_object_is_dying(&mc->mot_header)) {
1258 rc = mdt_attr_get_complex(info, mc, ma);
1261 } else if (mdt_dom_check_for_discard(info, mc)) {
1262 mdt_dom_discard_data(info, mc);
1264 mdt_handle_last_unlink(info, mc, ma);
1267 if (ma->ma_valid & MA_INODE) {
1268 switch (ma->ma_attr.la_mode & S_IFMT) {
1270 mdt_counter_incr(req, LPROC_MDT_RMDIR,
1271 ktime_us_delta(ktime_get(), kstart));
1279 mdt_counter_incr(req, LPROC_MDT_UNLINK,
1280 ktime_us_delta(ktime_get(), kstart));
1283 LASSERTF(0, "bad file type %o unlinking\n",
1284 ma->ma_attr.la_mode);
1291 mdt_reint_striped_unlock(info, mc, child_lh, einfo, rc);
1293 if (info->mti_spec.sp_cr_flags & MDS_OP_WITH_FID &&
1294 info->mti_big_buf.lb_buf)
1295 lu_buf_free(&info->mti_big_buf);
1296 mdt_object_put(info->mti_env, mc);
1298 mdt_object_unlock(info, mp, parent_lh, rc);
1300 mdt_object_put(info->mti_env, mp);
1301 CFS_RACE_WAKEUP(OBD_FAIL_OBD_ZERO_NLINK_RACE);
1306 * VBR: save versions in reply: 0 - parent; 1 - child by fid; 2 - target by
1309 static int mdt_reint_link(struct mdt_thread_info *info,
1310 struct mdt_lock_handle *lhc)
1312 struct mdt_reint_record *rr = &info->mti_rr;
1313 struct ptlrpc_request *req = mdt_info_req(info);
1314 struct md_attr *ma = &info->mti_attr;
1315 struct mdt_object *ms;
1316 struct mdt_object *mp;
1317 struct mdt_lock_handle *lhs;
1318 struct mdt_lock_handle *lhp;
1319 ktime_t kstart = ktime_get();
1324 DEBUG_REQ(D_INODE, req, "link "DFID" to "DFID"/"DNAME,
1325 PFID(rr->rr_fid1), PFID(rr->rr_fid2), PNAME(&rr->rr_name));
1327 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_REINT_LINK))
1328 RETURN(err_serious(-ENOENT));
1330 if (OBD_FAIL_PRECHECK(OBD_FAIL_PTLRPC_RESEND_RACE) ||
1331 OBD_FAIL_PRECHECK(OBD_FAIL_PTLRPC_ENQ_RESEND)) {
1332 req->rq_no_reply = 1;
1333 RETURN(err_serious(-ENOENT));
1336 if (info->mti_dlm_req)
1337 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
1339 /* Invalid case so return error immediately instead of
1342 if (lu_fid_eq(rr->rr_fid1, rr->rr_fid2))
1345 if (!fid_is_md_operative(rr->rr_fid1) ||
1346 !fid_is_md_operative(rr->rr_fid2))
1349 /* step 1: find target parent dir */
1350 mp = mdt_object_find(info->mti_env, info->mti_mdt, rr->rr_fid2);
1352 RETURN(PTR_ERR(mp));
1354 rc = mdt_version_get_check_save(info, mp, 0);
1356 GOTO(put_parent, rc);
1358 /* step 2: find source */
1359 ms = mdt_object_find(info->mti_env, info->mti_mdt, rr->rr_fid1);
1361 GOTO(put_parent, rc = PTR_ERR(ms));
1363 if (!mdt_object_exists(ms)) {
1364 CDEBUG(D_INFO, "%s: "DFID" does not exist.\n",
1365 mdt_obd_name(info->mti_mdt), PFID(rr->rr_fid1));
1366 GOTO(put_source, rc = -ENOENT);
1369 cos_incompat = (mdt_object_remote(mp) || mdt_object_remote(ms));
1371 OBD_RACE(OBD_FAIL_MDS_LINK_RENAME_RACE);
1373 lhp = &info->mti_lh[MDT_LH_PARENT];
1374 mdt_lock_pdo_init(lhp, LCK_PW, &rr->rr_name);
1375 rc = mdt_reint_object_lock(info, mp, lhp, MDS_INODELOCK_UPDATE,
1378 GOTO(put_source, rc);
1380 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME3, 5);
1382 lhs = &info->mti_lh[MDT_LH_CHILD];
1383 mdt_lock_reg_init(lhs, LCK_EX);
1384 rc = mdt_reint_object_lock(info, ms, lhs,
1385 MDS_INODELOCK_UPDATE | MDS_INODELOCK_XATTR,
1388 GOTO(unlock_parent, rc);
1390 /* step 3: link it */
1391 mdt_fail_write(info->mti_env, info->mti_mdt->mdt_bottom,
1392 OBD_FAIL_MDS_REINT_LINK_WRITE);
1394 tgt_vbr_obj_set(info->mti_env, mdt_obj2dt(ms));
1395 rc = mdt_version_get_check_save(info, ms, 1);
1397 GOTO(unlock_source, rc);
1399 /** check target version by name during replay */
1400 rc = mdt_lookup_version_check(info, mp, &rr->rr_name,
1401 &info->mti_tmp_fid1, 2);
1402 if (rc != 0 && rc != -ENOENT)
1403 GOTO(unlock_source, rc);
1404 /* save version of file name for replay, it must be ENOENT here */
1405 if (!req_is_replay(mdt_info_req(info))) {
1406 if (rc != -ENOENT) {
1407 CDEBUG(D_INFO, "link target "DNAME" existed!\n",
1408 PNAME(&rr->rr_name));
1409 GOTO(unlock_source, rc = -EEXIST);
1411 info->mti_ver[2] = ENOENT_VERSION;
1412 mdt_version_save(mdt_info_req(info), info->mti_ver[2], 2);
1415 rc = mdo_link(info->mti_env, mdt_object_child(mp),
1416 mdt_object_child(ms), &rr->rr_name, ma);
1419 mdt_counter_incr(req, LPROC_MDT_LINK,
1420 ktime_us_delta(ktime_get(), kstart));
1424 mdt_object_unlock(info, ms, lhs, rc);
1426 mdt_object_unlock(info, mp, lhp, rc);
1428 mdt_object_put(info->mti_env, ms);
1430 mdt_object_put(info->mti_env, mp);
1434 * lock the part of the directory according to the hash of the name
1435 * (lh->mlh_pdo_hash) in parallel directory lock.
1437 static int mdt_pdir_hash_lock(struct mdt_thread_info *info,
1438 struct mdt_lock_handle *lh,
1439 struct mdt_object *obj, __u64 ibits,
1442 struct ldlm_res_id *res = &info->mti_res_id;
1443 struct ldlm_namespace *ns = info->mti_mdt->mdt_namespace;
1444 union ldlm_policy_data *policy = &info->mti_policy;
1445 __u64 dlmflags = LDLM_FL_LOCAL_ONLY | LDLM_FL_ATOMIC_CB;
1449 * Finish res_id initializing by name hash marking part of
1450 * directory which is taking modification.
1452 LASSERT(lh->mlh_pdo_hash != 0);
1453 fid_build_pdo_res_name(mdt_object_fid(obj), lh->mlh_pdo_hash, res);
1454 memset(policy, 0, sizeof(*policy));
1455 policy->l_inodebits.bits = ibits;
1457 (lh->mlh_reg_mode == LCK_PW || lh->mlh_reg_mode == LCK_EX))
1458 dlmflags |= LDLM_FL_COS_INCOMPAT;
1460 * Use LDLM_FL_LOCAL_ONLY for this lock. We do not know yet if it is
1461 * going to be sent to client. If it is - mdt_intent_policy() path will
1462 * fix it up and turn FL_LOCAL flag off.
1464 rc = mdt_fid_lock(info->mti_env, ns, &lh->mlh_reg_lh, lh->mlh_reg_mode,
1465 policy, res, dlmflags,
1466 &info->mti_exp->exp_handle.h_cookie);
1471 * Get BFL lock for rename or migrate process.
1473 static int mdt_rename_lock(struct mdt_thread_info *info,
1474 struct lustre_handle *lh)
1479 if (mdt_seq_site(info->mti_mdt)->ss_node_id != 0) {
1480 struct lu_fid *fid = &info->mti_tmp_fid1;
1481 struct mdt_object *obj;
1483 /* XXX, right now, it has to use object API to
1484 * enqueue lock cross MDT, so it will enqueue
1485 * rename lock(with LUSTRE_BFL_FID) by root object
1488 obj = mdt_object_find(info->mti_env, info->mti_mdt, fid);
1490 RETURN(PTR_ERR(obj));
1492 rc = mdt_remote_object_lock(info, obj,
1493 &LUSTRE_BFL_FID, lh,
1495 MDS_INODELOCK_UPDATE, false);
1496 mdt_object_put(info->mti_env, obj);
1498 struct ldlm_namespace *ns = info->mti_mdt->mdt_namespace;
1499 union ldlm_policy_data *policy = &info->mti_policy;
1500 struct ldlm_res_id *res_id = &info->mti_res_id;
1503 fid_build_reg_res_name(&LUSTRE_BFL_FID, res_id);
1504 memset(policy, 0, sizeof(*policy));
1505 policy->l_inodebits.bits = MDS_INODELOCK_UPDATE;
1506 flags = LDLM_FL_LOCAL_ONLY | LDLM_FL_ATOMIC_CB;
1507 rc = ldlm_cli_enqueue_local(info->mti_env, ns, res_id,
1508 LDLM_IBITS, policy, LCK_EX, &flags,
1510 ldlm_completion_ast, NULL, NULL, 0,
1512 &info->mti_exp->exp_handle.h_cookie,
1519 static void mdt_rename_unlock(struct lustre_handle *lh)
1522 LASSERT(lustre_handle_is_used(lh));
1523 /* Cancel the single rename lock right away */
1524 ldlm_lock_decref_and_cancel(lh, LCK_EX);
1528 static struct mdt_object *mdt_parent_find_check(struct mdt_thread_info *info,
1529 const struct lu_fid *fid,
1532 struct mdt_object *dir;
1536 dir = mdt_object_find(info->mti_env, info->mti_mdt, fid);
1540 /* check early, the real version will be saved after locking */
1541 rc = mdt_version_get_check(info, dir, idx);
1545 if (!mdt_object_exists(dir))
1546 GOTO(out_put, rc = -ENOENT);
1548 if (!S_ISDIR(lu_object_attr(&dir->mot_obj)))
1549 GOTO(out_put, rc = -ENOTDIR);
1553 mdt_object_put(info->mti_env, dir);
1558 * in case obj is remote obj on its parent, revoke LOOKUP lock,
1559 * herein we don't really check it, just do revoke.
1561 int mdt_revoke_remote_lookup_lock(struct mdt_thread_info *info,
1562 struct mdt_object *pobj,
1563 struct mdt_object *obj)
1565 struct mdt_lock_handle *lh = &info->mti_lh[MDT_LH_LOCAL];
1568 mdt_lock_handle_init(lh);
1569 mdt_lock_reg_init(lh, LCK_EX);
1571 if (mdt_object_remote(pobj)) {
1572 /* don't bother to check if pobj and obj are on the same MDT. */
1573 rc = mdt_remote_object_lock(info, pobj, mdt_object_fid(obj),
1574 &lh->mlh_rreg_lh, LCK_EX,
1575 MDS_INODELOCK_LOOKUP, false);
1576 } else if (mdt_object_remote(obj)) {
1577 struct ldlm_res_id *res = &info->mti_res_id;
1578 union ldlm_policy_data *policy = &info->mti_policy;
1579 __u64 dlmflags = LDLM_FL_LOCAL_ONLY | LDLM_FL_ATOMIC_CB |
1580 LDLM_FL_COS_INCOMPAT;
1582 fid_build_reg_res_name(mdt_object_fid(obj), res);
1583 memset(policy, 0, sizeof(*policy));
1584 policy->l_inodebits.bits = MDS_INODELOCK_LOOKUP;
1585 rc = mdt_fid_lock(info->mti_env, info->mti_mdt->mdt_namespace,
1586 &lh->mlh_reg_lh, LCK_EX, policy, res,
1589 /* do nothing if both are local */
1597 * TODO, currently we don't save this lock because there is no place to
1598 * hold this lock handle, but to avoid race we need to save this lock.
1600 mdt_object_unlock(info, NULL, lh, 1);
1606 * operation may takes locks of linkea, or directory stripes, group them in
1609 struct mdt_sub_lock {
1610 struct mdt_object *msl_obj;
1611 struct mdt_lock_handle msl_lh;
1612 struct list_head msl_linkage;
1615 static void mdt_unlock_list(struct mdt_thread_info *info,
1616 struct list_head *list, int decref)
1618 struct mdt_sub_lock *msl;
1619 struct mdt_sub_lock *tmp;
1621 list_for_each_entry_safe(msl, tmp, list, msl_linkage) {
1622 mdt_object_unlock_put(info, msl->msl_obj, &msl->msl_lh, decref);
1623 list_del(&msl->msl_linkage);
1628 static inline void mdt_migrate_object_unlock(struct mdt_thread_info *info,
1629 struct mdt_object *obj,
1630 struct mdt_lock_handle *lh,
1631 struct ldlm_enqueue_info *einfo,
1632 struct list_head *slave_locks,
1635 if (mdt_object_remote(obj)) {
1636 mdt_unlock_list(info, slave_locks, decref);
1637 mdt_object_unlock(info, obj, lh, decref);
1639 mdt_reint_striped_unlock(info, obj, lh, einfo, decref);
1644 * lock parents of links, and also check whether total locks don't exceed
1647 * \retval 0 on success, and locks can be saved in ptlrpc_reply_stat
1648 * \retval 1 on success, but total lock count may exceed RS_MAX_LOCKS
1649 * \retval -ev negative errno upon error
1651 static int mdt_link_parents_lock(struct mdt_thread_info *info,
1652 struct mdt_object *pobj,
1653 const struct md_attr *ma,
1654 struct mdt_object *obj,
1655 struct mdt_lock_handle *lhp,
1656 struct ldlm_enqueue_info *peinfo,
1657 struct list_head *parent_slave_locks,
1658 struct list_head *link_locks)
1660 struct mdt_device *mdt = info->mti_mdt;
1661 struct lu_buf *buf = &info->mti_big_buf;
1662 struct lu_name *lname = &info->mti_name;
1663 struct linkea_data ldata = { NULL };
1664 bool blocked = false;
1665 int local_lnkp_cnt = 0;
1669 if (S_ISDIR(lu_object_attr(&obj->mot_obj)))
1672 buf = lu_buf_check_and_alloc(buf, MAX_LINKEA_SIZE);
1673 if (buf->lb_buf == NULL)
1677 rc = mdt_links_read(info, obj, &ldata);
1679 if (rc == -ENOENT || rc == -ENODATA)
1684 for (linkea_first_entry(&ldata); ldata.ld_lee && !rc;
1685 linkea_next_entry(&ldata)) {
1686 struct mdt_object *lnkp;
1687 struct mdt_sub_lock *msl;
1691 linkea_entry_unpack(ldata.ld_lee, &ldata.ld_reclen, lname,
1694 /* check if it's also linked to parent */
1695 if (lu_fid_eq(mdt_object_fid(pobj), &fid)) {
1696 CDEBUG(D_INFO, "skip parent "DFID", reovke "DNAME"\n",
1697 PFID(&fid), PNAME(lname));
1698 /* in case link is remote object, revoke LOOKUP lock */
1699 rc = mdt_revoke_remote_lookup_lock(info, pobj, obj);
1705 /* check if it's linked to a stripe of parent */
1706 if (ma->ma_valid & MA_LMV) {
1707 struct lmv_mds_md_v1 *lmv = &ma->ma_lmv->lmv_md_v1;
1708 struct lu_fid *stripe_fid = &info->mti_tmp_fid1;
1711 for (; j < le32_to_cpu(lmv->lmv_stripe_count); j++) {
1712 fid_le_to_cpu(stripe_fid,
1713 &lmv->lmv_stripe_fids[j]);
1714 if (lu_fid_eq(stripe_fid, &fid)) {
1715 CDEBUG(D_INFO, "skip stripe "DFID
1716 ", reovke "DNAME"\n",
1717 PFID(&fid), PNAME(lname));
1718 lnkp = mdt_object_find(info->mti_env,
1721 GOTO(out, rc = PTR_ERR(lnkp));
1727 rc = mdt_revoke_remote_lookup_lock(info, lnkp,
1729 mdt_object_put(info->mti_env, lnkp);
1734 /* Check if it's already locked */
1735 list_for_each_entry(msl, link_locks, msl_linkage) {
1736 if (lu_fid_eq(mdt_object_fid(msl->msl_obj), &fid)) {
1738 DFID" was locked, revoke "DNAME"\n",
1739 PFID(&fid), PNAME(lname));
1740 lnkp = msl->msl_obj;
1746 rc = mdt_revoke_remote_lookup_lock(info, lnkp, obj);
1750 CDEBUG(D_INFO, "lock "DFID":"DNAME"\n",
1751 PFID(&fid), PNAME(lname));
1753 lnkp = mdt_object_find(info->mti_env, mdt, &fid);
1755 CWARN("%s: cannot find obj "DFID": %ld\n",
1756 mdt_obd_name(mdt), PFID(&fid), PTR_ERR(lnkp));
1760 if (!mdt_object_exists(lnkp)) {
1761 CDEBUG(D_INFO, DFID" doesn't exist, skip "DNAME"\n",
1762 PFID(&fid), PNAME(lname));
1763 mdt_object_put(info->mti_env, lnkp);
1767 if (!mdt_object_remote(lnkp))
1772 GOTO(out, rc = -ENOMEM);
1775 * we can't follow parent-child lock order like other MD
1776 * operations, use lock_try here to avoid deadlock, if the lock
1777 * cannot be taken, drop all locks taken, revoke the blocked
1778 * one, and continue processing the remaining entries, and in
1779 * the end of the loop restart from beginning.
1781 mdt_lock_pdo_init(&msl->msl_lh, LCK_PW, lname);
1783 rc = mdt_object_lock_try(info, lnkp, &msl->msl_lh, &ibits,
1784 MDS_INODELOCK_UPDATE, true);
1785 if (!(ibits & MDS_INODELOCK_UPDATE)) {
1787 CDEBUG(D_INFO, "busy lock on "DFID" "DNAME"\n",
1788 PFID(&fid), PNAME(lname));
1790 mdt_unlock_list(info, link_locks, 1);
1791 /* also unlock parent locks to avoid deadlock */
1793 mdt_migrate_object_unlock(info, pobj, lhp,
1800 mdt_lock_pdo_init(&msl->msl_lh, LCK_PW, lname);
1801 rc = mdt_object_lock(info, lnkp, &msl->msl_lh,
1802 MDS_INODELOCK_UPDATE);
1804 mdt_object_put(info->mti_env, lnkp);
1809 if (mdt_object_remote(lnkp)) {
1810 struct ldlm_lock *lock;
1813 * for remote object, set lock cb_atomic,
1814 * so lock can be released in blocking_ast()
1815 * immediately, then the next lock_try will
1816 * have better chance of success.
1818 lock = ldlm_handle2lock(
1819 &msl->msl_lh.mlh_rreg_lh);
1820 LASSERT(lock != NULL);
1821 lock_res_and_lock(lock);
1822 ldlm_set_atomic_cb(lock);
1823 unlock_res_and_lock(lock);
1824 LDLM_LOCK_PUT(lock);
1827 mdt_object_unlock_put(info, lnkp, &msl->msl_lh, 1);
1832 INIT_LIST_HEAD(&msl->msl_linkage);
1833 msl->msl_obj = lnkp;
1834 list_add_tail(&msl->msl_linkage, link_locks);
1836 rc = mdt_revoke_remote_lookup_lock(info, lnkp, obj);
1840 GOTO(out, rc = -EBUSY);
1845 mdt_unlock_list(info, link_locks, rc);
1846 } else if (local_lnkp_cnt > RS_MAX_LOCKS - 5) {
1847 CDEBUG(D_INFO, "Too many links (%d), sync operations\n",
1850 * parent may have 3 local objects: master object and 2 stripes
1851 * (if it's being migrated too); source may have 1 local objects
1852 * as regular file; target has 1 local object.
1853 * Note, source may have 2 local locks if it is directory but it
1854 * can't have hardlinks, so it is not considered here.
1861 static int mdt_lock_remote_slaves(struct mdt_thread_info *info,
1862 struct mdt_object *obj,
1863 const struct md_attr *ma,
1864 struct list_head *slave_locks)
1866 struct mdt_device *mdt = info->mti_mdt;
1867 const struct lmv_mds_md_v1 *lmv = &ma->ma_lmv->lmv_md_v1;
1868 struct lu_fid *fid = &info->mti_tmp_fid1;
1869 struct mdt_object *slave;
1870 struct mdt_sub_lock *msl;
1875 LASSERT(mdt_object_remote(obj));
1876 LASSERT(ma->ma_valid & MA_LMV);
1879 if (!lmv_is_sane(lmv))
1882 for (i = 0; i < le32_to_cpu(lmv->lmv_stripe_count); i++) {
1883 fid_le_to_cpu(fid, &lmv->lmv_stripe_fids[i]);
1885 if (!fid_is_sane(fid))
1888 slave = mdt_object_find(info->mti_env, mdt, fid);
1890 GOTO(out, rc = PTR_ERR(slave));
1894 mdt_object_put(info->mti_env, slave);
1895 GOTO(out, rc = -ENOMEM);
1898 mdt_lock_reg_init(&msl->msl_lh, LCK_EX);
1899 rc = mdt_reint_object_lock(info, slave, &msl->msl_lh,
1900 MDS_INODELOCK_UPDATE, true);
1903 mdt_object_put(info->mti_env, slave);
1907 INIT_LIST_HEAD(&msl->msl_linkage);
1908 msl->msl_obj = slave;
1909 list_add_tail(&msl->msl_linkage, slave_locks);
1915 mdt_unlock_list(info, slave_locks, rc);
1919 /* lock parent and its stripes */
1920 static int mdt_migrate_parent_lock(struct mdt_thread_info *info,
1921 struct mdt_object *obj,
1922 const struct md_attr *ma,
1923 struct mdt_lock_handle *lh,
1924 struct ldlm_enqueue_info *einfo,
1925 struct list_head *slave_locks)
1929 if (mdt_object_remote(obj)) {
1930 rc = mdt_remote_object_lock(info, obj, mdt_object_fid(obj),
1931 &lh->mlh_rreg_lh, LCK_PW,
1932 MDS_INODELOCK_UPDATE, false);
1937 * if obj is remote and striped, lock its stripes explicitly
1938 * because it's not striped in LOD layer on this MDT.
1940 if (ma->ma_valid & MA_LMV) {
1941 rc = mdt_lock_remote_slaves(info, obj, ma, slave_locks);
1943 mdt_object_unlock(info, obj, lh, rc);
1946 rc = mdt_reint_striped_lock(info, obj, lh, MDS_INODELOCK_UPDATE,
1954 * in migration, object may be remote, and we need take full lock of it and its
1955 * stripes if it's directory, besides, object may be a remote object on its
1956 * parent, revoke its LOOKUP lock on where its parent is located.
1958 static int mdt_migrate_object_lock(struct mdt_thread_info *info,
1959 struct mdt_object *pobj,
1960 struct mdt_object *obj,
1961 struct mdt_lock_handle *lh,
1962 struct ldlm_enqueue_info *einfo,
1963 struct list_head *slave_locks)
1967 if (mdt_object_remote(obj)) {
1968 rc = mdt_revoke_remote_lookup_lock(info, pobj, obj);
1972 rc = mdt_remote_object_lock(info, obj, mdt_object_fid(obj),
1973 &lh->mlh_rreg_lh, LCK_EX,
1974 MDS_INODELOCK_FULL, false);
1979 * if obj is remote and striped, lock its stripes explicitly
1980 * because it's not striped in LOD layer on this MDT.
1982 if (S_ISDIR(lu_object_attr(&obj->mot_obj))) {
1983 struct md_attr *ma = &info->mti_attr;
1985 rc = mdt_stripe_get(info, obj, ma, XATTR_NAME_LMV);
1987 mdt_object_unlock(info, obj, lh, rc);
1991 if (ma->ma_valid & MA_LMV) {
1992 rc = mdt_lock_remote_slaves(info, obj, ma,
1995 mdt_object_unlock(info, obj, lh, rc);
1999 if (mdt_object_remote(pobj)) {
2000 rc = mdt_revoke_remote_lookup_lock(info, pobj, obj);
2005 rc = mdt_reint_striped_lock(info, obj, lh, MDS_INODELOCK_FULL,
2013 * lookup source by name, if parent is striped directory, we need to find the
2014 * corresponding stripe where source is located, and then lookup there.
2016 * besides, if parent is migrating too, and file is already in target stripe,
2017 * this should be a redo of 'lfs migrate' on client side.
2019 static int mdt_migrate_lookup(struct mdt_thread_info *info,
2020 struct mdt_object *pobj,
2021 const struct md_attr *ma,
2022 const struct lu_name *lname,
2023 struct mdt_object **spobj,
2024 struct mdt_object **sobj)
2026 const struct lu_env *env = info->mti_env;
2027 struct lu_fid *fid = &info->mti_tmp_fid1;
2028 struct mdt_object *stripe;
2031 if (ma->ma_valid & MA_LMV) {
2032 /* if parent is striped, lookup on corresponding stripe */
2033 struct lmv_mds_md_v1 *lmv = &ma->ma_lmv->lmv_md_v1;
2035 if (!lmv_is_sane(lmv))
2038 rc = lmv_name_to_stripe_index_old(lmv, lname->ln_name,
2043 fid_le_to_cpu(fid, &lmv->lmv_stripe_fids[rc]);
2045 stripe = mdt_object_find(env, info->mti_mdt, fid);
2047 return PTR_ERR(stripe);
2050 rc = mdo_lookup(env, mdt_object_child(stripe), lname, fid,
2052 if (rc == -ENOENT && lmv_is_layout_changing(lmv)) {
2054 * if parent layout is changeing, and lookup child
2055 * failed on source stripe, lookup again on target
2056 * stripe, if it exists, it means previous migration
2057 * was interrupted, and current file was migrated
2060 mdt_object_put(env, stripe);
2062 rc = lmv_name_to_stripe_index(lmv, lname->ln_name,
2067 fid_le_to_cpu(fid, &lmv->lmv_stripe_fids[rc]);
2069 stripe = mdt_object_find(env, info->mti_mdt, fid);
2071 return PTR_ERR(stripe);
2074 rc = mdo_lookup(env, mdt_object_child(stripe), lname,
2075 fid, &info->mti_spec);
2076 mdt_object_put(env, stripe);
2077 return rc ?: -EALREADY;
2079 mdt_object_put(env, stripe);
2084 rc = mdo_lookup(env, mdt_object_child(pobj), lname, fid,
2090 mdt_object_get(env, stripe);
2095 *sobj = mdt_object_find(env, info->mti_mdt, fid);
2096 if (IS_ERR(*sobj)) {
2097 mdt_object_put(env, stripe);
2098 rc = PTR_ERR(*sobj);
2106 /* end lease and close file for regular file */
2107 static int mdd_migrate_close(struct mdt_thread_info *info,
2108 struct mdt_object *obj)
2110 struct close_data *data;
2111 struct mdt_body *repbody;
2112 struct ldlm_lock *lease;
2117 if (!req_capsule_field_present(info->mti_pill, &RMF_MDT_EPOCH,
2119 !req_capsule_field_present(info->mti_pill, &RMF_CLOSE_DATA,
2123 data = req_capsule_client_get(info->mti_pill, &RMF_CLOSE_DATA);
2128 lease = ldlm_handle2lock(&data->cd_handle);
2132 /* check if the lease was already canceled */
2133 lock_res_and_lock(lease);
2134 rc = ldlm_is_cancel(lease);
2135 unlock_res_and_lock(lease);
2139 LDLM_DEBUG(lease, DFID" lease broken",
2140 PFID(mdt_object_fid(obj)));
2144 * cancel server side lease, client side counterpart should have been
2145 * cancelled, it's okay to cancel it now as we've held mot_open_sem.
2147 ldlm_lock_cancel(lease);
2148 ldlm_reprocess_all(lease->l_resource, lease);
2149 LDLM_LOCK_PUT(lease);
2152 rc2 = mdt_close_internal(info, mdt_info_req(info), NULL);
2153 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
2154 repbody->mbo_valid |= OBD_MD_CLOSE_INTENT_EXECED;
2160 * migrate file in below steps:
2161 * 1. lock parent and its stripes
2162 * 2. lookup source by name
2163 * 3. lock parents of source links if source is not directory
2164 * 4. reject if source is in HSM
2165 * 5. take source open_sem and close file if source is regular file
2166 * 6. lock source and its stripes if it's directory
2167 * 7. lock target so subsequent change to it can trigger COS
2169 * 9. unlock above locks
2170 * 10. sync device if source has links
2172 int mdt_reint_migrate(struct mdt_thread_info *info,
2173 struct mdt_lock_handle *unused)
2175 const struct lu_env *env = info->mti_env;
2176 struct mdt_device *mdt = info->mti_mdt;
2177 struct ptlrpc_request *req = mdt_info_req(info);
2178 struct mdt_reint_record *rr = &info->mti_rr;
2179 struct lu_ucred *uc = mdt_ucred(info);
2180 struct md_attr *ma = &info->mti_attr;
2181 struct ldlm_enqueue_info *peinfo = &info->mti_einfo[0];
2182 struct ldlm_enqueue_info *seinfo = &info->mti_einfo[1];
2183 struct mdt_object *pobj;
2184 struct mdt_object *spobj = NULL;
2185 struct mdt_object *sobj = NULL;
2186 struct mdt_object *tobj;
2187 struct lustre_handle rename_lh = { 0 };
2188 struct mdt_lock_handle *lhp;
2189 struct mdt_lock_handle *lhs;
2190 struct mdt_lock_handle *lht;
2191 LIST_HEAD(parent_slave_locks);
2192 LIST_HEAD(child_slave_locks);
2193 LIST_HEAD(link_locks);
2194 int lock_retries = 5;
2195 bool open_sem_locked = false;
2196 bool do_sync = false;
2200 CDEBUG(D_INODE, "migrate "DFID"/"DNAME" to "DFID"\n", PFID(rr->rr_fid1),
2201 PNAME(&rr->rr_name), PFID(rr->rr_fid2));
2203 if (info->mti_dlm_req)
2204 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
2206 if (!fid_is_md_operative(rr->rr_fid1) ||
2207 !fid_is_md_operative(rr->rr_fid2))
2210 /* don't allow migrate . or .. */
2211 if (lu_name_is_dot_or_dotdot(&rr->rr_name))
2214 if (!mdt->mdt_enable_remote_dir || !mdt->mdt_enable_dir_migration)
2217 if (uc && !cap_raised(uc->uc_cap, CAP_SYS_ADMIN) &&
2218 uc->uc_gid != mdt->mdt_enable_remote_dir_gid &&
2219 mdt->mdt_enable_remote_dir_gid != -1)
2223 * Note: do not enqueue rename lock for replay request, because
2224 * if other MDT holds rename lock, but being blocked to wait for
2225 * this MDT to finish its recovery, and the failover MDT can not
2226 * get rename lock, which will cause deadlock.
2228 * req is NULL if this is called by directory auto-split.
2230 if (req && !req_is_replay(req)) {
2231 rc = mdt_rename_lock(info, &rename_lh);
2233 CERROR("%s: can't lock FS for rename: rc = %d\n",
2234 mdt_obd_name(info->mti_mdt), rc);
2239 /* pobj is master object of parent */
2240 pobj = mdt_object_find(env, mdt, rr->rr_fid1);
2242 GOTO(unlock_rename, rc = PTR_ERR(pobj));
2245 rc = mdt_version_get_check(info, pobj, 0);
2247 GOTO(put_parent, rc);
2250 if (!mdt_object_exists(pobj))
2251 GOTO(put_parent, rc = -ENOENT);
2253 if (!S_ISDIR(lu_object_attr(&pobj->mot_obj)))
2254 GOTO(put_parent, rc = -ENOTDIR);
2256 rc = mdt_stripe_get(info, pobj, ma, XATTR_NAME_LMV);
2258 GOTO(put_parent, rc);
2261 /* lock parent object */
2262 lhp = &info->mti_lh[MDT_LH_PARENT];
2263 mdt_lock_reg_init(lhp, LCK_PW);
2264 rc = mdt_migrate_parent_lock(info, pobj, ma, lhp, peinfo,
2265 &parent_slave_locks);
2267 GOTO(put_parent, rc);
2270 * spobj is the corresponding stripe against name if pobj is striped
2271 * directory, which is the real parent, and no need to lock, because
2272 * we've taken full lock of pobj.
2274 rc = mdt_migrate_lookup(info, pobj, ma, &rr->rr_name, &spobj, &sobj);
2276 GOTO(unlock_parent, rc);
2278 /* lock parents of source links, and revoke LOOKUP lock of links */
2279 rc = mdt_link_parents_lock(info, pobj, ma, sobj, lhp, peinfo,
2280 &parent_slave_locks, &link_locks);
2281 if (rc == -EBUSY && lock_retries-- > 0) {
2282 mdt_object_put(env, sobj);
2283 mdt_object_put(env, spobj);
2288 GOTO(put_source, rc);
2291 * RS_MAX_LOCKS is the limit of number of locks that can be saved along
2292 * with one request, if total lock count exceeds this limit, we will
2293 * drop all locks after migration, and synchronous device in the end.
2297 /* TODO: DoM migration is not supported, migrate dirent only */
2298 if (S_ISREG(lu_object_attr(&sobj->mot_obj))) {
2299 rc = mdt_stripe_get(info, sobj, ma, XATTR_NAME_LOV);
2301 GOTO(unlock_links, rc);
2303 if (ma->ma_valid & MA_LOV && mdt_lmm_dom_stripesize(ma->ma_lmm))
2304 info->mti_spec.sp_migrate_nsonly = 1;
2305 } else if (S_ISDIR(lu_object_attr(&sobj->mot_obj))) {
2306 rc = mdt_stripe_get(info, sobj, ma, XATTR_NAME_LMV);
2308 GOTO(unlock_links, rc);
2310 /* race with restripe/auto-split? */
2311 if ((ma->ma_valid & MA_LMV) &&
2312 lmv_is_restriping(&ma->ma_lmv->lmv_md_v1))
2313 GOTO(unlock_links, rc = -EBUSY);
2316 /* if migration HSM is allowed */
2317 if (!mdt->mdt_opts.mo_migrate_hsm_allowed) {
2318 ma->ma_need = MA_HSM;
2320 rc = mdt_attr_get_complex(info, sobj, ma);
2322 GOTO(unlock_links, rc);
2324 if ((ma->ma_valid & MA_HSM) && ma->ma_hsm.mh_flags != 0)
2325 GOTO(unlock_links, rc = -EOPNOTSUPP);
2328 /* end lease and close file for regular file */
2329 if (info->mti_spec.sp_migrate_close) {
2330 /* try to hold open_sem so that nobody else can open the file */
2331 if (!down_write_trylock(&sobj->mot_open_sem)) {
2333 mdd_migrate_close(info, sobj);
2334 GOTO(unlock_links, rc = -EBUSY);
2336 open_sem_locked = true;
2337 rc = mdd_migrate_close(info, sobj);
2339 GOTO(unlock_open_sem, rc);
2344 lhs = &info->mti_lh[MDT_LH_OLD];
2345 mdt_lock_reg_init(lhs, LCK_EX);
2346 rc = mdt_migrate_object_lock(info, spobj, sobj, lhs, seinfo,
2347 &child_slave_locks);
2349 GOTO(unlock_open_sem, rc);
2352 tobj = mdt_object_find(env, mdt, rr->rr_fid2);
2354 GOTO(unlock_source, rc = PTR_ERR(tobj));
2356 lht = &info->mti_lh[MDT_LH_NEW];
2357 mdt_lock_reg_init(lht, LCK_EX);
2358 rc = mdt_reint_object_lock(info, tobj, lht, MDS_INODELOCK_FULL, true);
2360 GOTO(put_target, rc);
2362 /* Don't do lookup sanity check. We know name doesn't exist. */
2363 info->mti_spec.sp_cr_lookup = 0;
2364 info->mti_spec.sp_feat = &dt_directory_features;
2366 rc = mdo_migrate(env, mdt_object_child(pobj),
2367 mdt_object_child(sobj), &rr->rr_name,
2368 mdt_object_child(tobj),
2369 &info->mti_spec, ma);
2371 lprocfs_counter_incr(mdt->mdt_lu_dev.ld_obd->obd_md_stats,
2372 LPROC_MDT_MIGRATE + LPROC_MD_LAST_OPC);
2375 mdt_object_unlock(info, tobj, lht, rc);
2377 mdt_object_put(env, tobj);
2379 mdt_migrate_object_unlock(info, sobj, lhs, seinfo,
2380 &child_slave_locks, rc);
2382 if (open_sem_locked)
2383 up_write(&sobj->mot_open_sem);
2385 /* if we've got too many locks to save into RPC,
2386 * then just commit before the locks are released
2389 mdt_device_sync(env, mdt);
2390 mdt_unlock_list(info, &link_locks, do_sync ? 1 : rc);
2392 mdt_object_put(env, sobj);
2393 mdt_object_put(env, spobj);
2395 mdt_migrate_object_unlock(info, pobj, lhp, peinfo,
2396 &parent_slave_locks, rc);
2398 mdt_object_put(env, pobj);
2400 if (lustre_handle_is_used(&rename_lh))
2401 mdt_rename_unlock(&rename_lh);
2406 static int mdt_object_lock_save(struct mdt_thread_info *info,
2407 struct mdt_object *dir,
2408 struct mdt_lock_handle *lh,
2409 int idx, bool cos_incompat)
2413 /* we lock the target dir if it is local */
2414 rc = mdt_reint_object_lock(info, dir, lh, MDS_INODELOCK_UPDATE,
2419 /* get and save correct version after locking */
2420 mdt_version_get_save(info, dir, idx);
2425 * determine lock order of sobj and tobj
2427 * there are two situations we need to lock tobj before sobj:
2428 * 1. sobj is child of tobj
2429 * 2. sobj and tobj are stripes of a directory, and stripe index of sobj is
2430 * larger than that of tobj
2432 * \retval 1 lock tobj before sobj
2433 * \retval 0 lock sobj before tobj
2434 * \retval -ev negative errno upon error
2436 static int mdt_rename_determine_lock_order(struct mdt_thread_info *info,
2437 struct mdt_object *sobj,
2438 struct mdt_object *tobj)
2440 struct md_attr *ma = &info->mti_attr;
2441 struct lu_fid *spfid = &info->mti_tmp_fid1;
2442 struct lu_fid *tpfid = &info->mti_tmp_fid2;
2443 struct lmv_mds_md_v1 *lmv;
2448 /* sobj and tobj are the same */
2452 if (fid_is_root(mdt_object_fid(sobj)))
2455 if (fid_is_root(mdt_object_fid(tobj)))
2458 /* check whether sobj is child of tobj */
2459 rc = mdo_is_subdir(info->mti_env, mdt_object_child(sobj),
2460 mdt_object_fid(tobj));
2467 /* check whether sobj and tobj are children of the same parent */
2468 rc = mdt_attr_get_pfid(info, sobj, spfid);
2472 rc = mdt_attr_get_pfid(info, tobj, tpfid);
2476 if (!lu_fid_eq(spfid, tpfid))
2479 /* check whether sobj and tobj are sibling stripes */
2480 rc = mdt_stripe_get(info, sobj, ma, XATTR_NAME_LMV);
2484 if (!(ma->ma_valid & MA_LMV))
2487 lmv = &ma->ma_lmv->lmv_md_v1;
2488 if (!(le32_to_cpu(lmv->lmv_magic) & LMV_MAGIC_STRIPE))
2490 sindex = le32_to_cpu(lmv->lmv_master_mdt_index);
2493 rc = mdt_stripe_get(info, tobj, ma, XATTR_NAME_LMV);
2497 if (!(ma->ma_valid & MA_LMV))
2500 lmv = &ma->ma_lmv->lmv_md_v1;
2501 if (!(le32_to_cpu(lmv->lmv_magic) & LMV_MAGIC_STRIPE))
2503 tindex = le32_to_cpu(lmv->lmv_master_mdt_index);
2505 /* check stripe index of sobj and tobj */
2506 if (sindex == tindex)
2509 return sindex < tindex ? 0 : 1;
2513 * lock rename source object.
2515 * Both source and source parent may be remote, and source may be a remote
2516 * object on source parent, to avoid overriding lock handle, store remote
2517 * LOOKUP lock separately in @lhr.
2519 * \retval 0 on success
2520 * \retval -ev negative errno upon error
2522 static int mdt_rename_source_lock(struct mdt_thread_info *info,
2523 struct mdt_object *parent,
2524 struct mdt_object *child,
2525 struct mdt_lock_handle *lhc,
2526 struct mdt_lock_handle *lhr,
2532 rc = mdt_is_remote_object(info, parent, child);
2537 /* enqueue remote LOOKUP lock from the parent MDT */
2538 __u64 rmt_ibits = MDS_INODELOCK_LOOKUP;
2540 if (mdt_object_remote(parent)) {
2541 rc = mdt_remote_object_lock(info, parent,
2542 mdt_object_fid(child),
2549 LASSERT(mdt_object_remote(child));
2550 rc = mdt_object_local_lock(info, child, lhr,
2551 &rmt_ibits, 0, true);
2556 ibits &= ~MDS_INODELOCK_LOOKUP;
2559 if (mdt_object_remote(child)) {
2560 rc = mdt_remote_object_lock(info, child, mdt_object_fid(child),
2567 rc = mdt_reint_object_lock(info, child, lhc, ibits,
2572 mdt_object_unlock(info, child, lhr, rc);
2578 * VBR: rename versions in reply: 0 - srcdir parent; 1 - tgtdir parent;
2579 * 2 - srcdir child; 3 - tgtdir child.
2580 * Update on disk version of srcdir child.
2582 static int mdt_reint_rename(struct mdt_thread_info *info,
2583 struct mdt_lock_handle *unused)
2585 struct mdt_device *mdt = info->mti_mdt;
2586 struct mdt_reint_record *rr = &info->mti_rr;
2587 struct md_attr *ma = &info->mti_attr;
2588 struct ptlrpc_request *req = mdt_info_req(info);
2589 struct mdt_object *msrcdir = NULL;
2590 struct mdt_object *mtgtdir = NULL;
2591 struct mdt_object *mold;
2592 struct mdt_object *mnew = NULL;
2593 struct lustre_handle rename_lh = { 0 };
2594 struct mdt_lock_handle *lh_srcdirp;
2595 struct mdt_lock_handle *lh_tgtdirp;
2596 struct mdt_lock_handle *lh_oldp = NULL;
2597 struct mdt_lock_handle *lh_rmt = NULL;
2598 struct mdt_lock_handle *lh_newp = NULL;
2599 struct lu_fid *old_fid = &info->mti_tmp_fid1;
2600 struct lu_fid *new_fid = &info->mti_tmp_fid2;
2602 bool reverse = false, discard = false;
2604 ktime_t kstart = ktime_get();
2608 DEBUG_REQ(D_INODE, req, "rename "DFID"/"DNAME" to "DFID"/"DNAME,
2609 PFID(rr->rr_fid1), PNAME(&rr->rr_name),
2610 PFID(rr->rr_fid2), PNAME(&rr->rr_tgt_name));
2612 if (info->mti_dlm_req)
2613 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
2615 if (!fid_is_md_operative(rr->rr_fid1) ||
2616 !fid_is_md_operative(rr->rr_fid2))
2619 /* find both parents. */
2620 msrcdir = mdt_parent_find_check(info, rr->rr_fid1, 0);
2621 if (IS_ERR(msrcdir))
2622 RETURN(PTR_ERR(msrcdir));
2624 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME3, 5);
2626 if (lu_fid_eq(rr->rr_fid1, rr->rr_fid2)) {
2628 mdt_object_get(info->mti_env, mtgtdir);
2630 mtgtdir = mdt_parent_find_check(info, rr->rr_fid2, 1);
2631 if (IS_ERR(mtgtdir))
2632 GOTO(out_put_srcdir, rc = PTR_ERR(mtgtdir));
2636 * Note: do not enqueue rename lock for replay request, because
2637 * if other MDT holds rename lock, but being blocked to wait for
2638 * this MDT to finish its recovery, and the failover MDT can not
2639 * get rename lock, which will cause deadlock.
2641 if (!req_is_replay(req)) {
2643 * Normally rename RPC is handled on the MDT with the target
2644 * directory (if target exists, it's on the MDT with the
2645 * target), if the source directory is remote, it's a hint that
2646 * source is remote too (this may not be true, but it won't
2647 * cause any issue), return -EXDEV early to avoid taking
2650 if (!mdt->mdt_enable_remote_rename &&
2651 mdt_object_remote(msrcdir))
2652 GOTO(out_put_tgtdir, rc = -EXDEV);
2654 /* This might be further relaxed in the future for regular file
2655 * renames in different source and target parents. Start with
2656 * only same-directory renames for simplicity and because this
2657 * is by far the most the common use case.
2659 if (msrcdir != mtgtdir) {
2660 rc = mdt_rename_lock(info, &rename_lh);
2662 CERROR("%s: cannot lock for rename: rc = %d\n",
2663 mdt_obd_name(mdt), rc);
2664 GOTO(out_put_tgtdir, rc);
2667 CDEBUG(D_INFO, "%s: samedir rename "DFID"/"DNAME"\n",
2668 mdt_obd_name(mdt), PFID(rr->rr_fid1),
2669 PNAME(&rr->rr_name));
2673 rc = mdt_rename_determine_lock_order(info, msrcdir, mtgtdir);
2675 GOTO(out_unlock_rename, rc);
2679 /* source needs to be looked up after locking source parent, otherwise
2680 * this rename may race with unlink source, and cause rename hang, see
2681 * sanityn.sh 55b, so check parents first, if later we found source is
2682 * remote, relock parents.
2684 cos_incompat = (mdt_object_remote(msrcdir) ||
2685 mdt_object_remote(mtgtdir));
2687 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME4, 5);
2689 /* lock parents in the proper order. */
2690 lh_srcdirp = &info->mti_lh[MDT_LH_PARENT];
2691 lh_tgtdirp = &info->mti_lh[MDT_LH_CHILD];
2693 OBD_RACE(OBD_FAIL_MDS_REINT_OPEN);
2694 OBD_RACE(OBD_FAIL_MDS_REINT_OPEN2);
2696 mdt_lock_pdo_init(lh_srcdirp, LCK_PW, &rr->rr_name);
2697 mdt_lock_pdo_init(lh_tgtdirp, LCK_PW, &rr->rr_tgt_name);
2700 rc = mdt_object_lock_save(info, mtgtdir, lh_tgtdirp, 1,
2703 GOTO(out_unlock_rename, rc);
2705 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME, 5);
2707 rc = mdt_object_lock_save(info, msrcdir, lh_srcdirp, 0,
2710 mdt_object_unlock(info, mtgtdir, lh_tgtdirp, rc);
2711 GOTO(out_unlock_rename, rc);
2714 rc = mdt_object_lock_save(info, msrcdir, lh_srcdirp, 0,
2717 GOTO(out_unlock_rename, rc);
2719 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME, 5);
2721 if (mtgtdir != msrcdir) {
2722 rc = mdt_object_lock_save(info, mtgtdir, lh_tgtdirp, 1,
2724 } else if (!mdt_object_remote(mtgtdir) &&
2725 lh_srcdirp->mlh_pdo_hash !=
2726 lh_tgtdirp->mlh_pdo_hash) {
2727 rc = mdt_pdir_hash_lock(info, lh_tgtdirp, mtgtdir,
2728 MDS_INODELOCK_UPDATE,
2730 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_PDO_LOCK2, 10);
2733 mdt_object_unlock(info, msrcdir, lh_srcdirp, rc);
2734 GOTO(out_unlock_rename, rc);
2738 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME4, 5);
2739 OBD_FAIL_TIMEOUT(OBD_FAIL_MDS_RENAME2, 5);
2741 /* find mold object. */
2743 rc = mdt_lookup_version_check(info, msrcdir, &rr->rr_name, old_fid, 2);
2745 GOTO(out_unlock_parents, rc);
2747 if (lu_fid_eq(old_fid, rr->rr_fid1) || lu_fid_eq(old_fid, rr->rr_fid2))
2748 GOTO(out_unlock_parents, rc = -EINVAL);
2750 if (!fid_is_md_operative(old_fid))
2751 GOTO(out_unlock_parents, rc = -EPERM);
2753 mold = mdt_object_find(info->mti_env, info->mti_mdt, old_fid);
2755 GOTO(out_unlock_parents, rc = PTR_ERR(mold));
2757 if (!mdt_object_exists(mold)) {
2758 LU_OBJECT_DEBUG(D_INODE, info->mti_env,
2760 "object does not exist");
2761 GOTO(out_put_old, rc = -ENOENT);
2764 if (mdt_object_remote(mold) && !mdt->mdt_enable_remote_rename)
2765 GOTO(out_put_old, rc = -EXDEV);
2767 /* Check if @mtgtdir is subdir of @mold, before locking child
2768 * to avoid reverse locking.
2770 if (mtgtdir != msrcdir) {
2771 rc = mdo_is_subdir(info->mti_env, mdt_object_child(mtgtdir),
2776 GOTO(out_put_old, rc);
2780 tgt_vbr_obj_set(info->mti_env, mdt_obj2dt(mold));
2781 /* save version after locking */
2782 mdt_version_get_save(info, mold, 2);
2784 if (!cos_incompat && mdt_object_remote(mold)) {
2785 cos_incompat = true;
2786 mdt_object_put(info->mti_env, mold);
2787 mdt_object_unlock(info, mtgtdir, lh_tgtdirp, -EAGAIN);
2788 mdt_object_unlock(info, msrcdir, lh_srcdirp, -EAGAIN);
2792 /* find mnew object:
2793 * mnew target object may not exist now
2794 * lookup with version checking
2797 rc = mdt_lookup_version_check(info, mtgtdir, &rr->rr_tgt_name, new_fid,
2800 /* the new_fid should have been filled at this moment */
2801 if (lu_fid_eq(old_fid, new_fid))
2802 GOTO(out_put_old, rc);
2804 if (lu_fid_eq(new_fid, rr->rr_fid1) ||
2805 lu_fid_eq(new_fid, rr->rr_fid2))
2806 GOTO(out_put_old, rc = -EINVAL);
2808 if (!fid_is_md_operative(new_fid))
2809 GOTO(out_put_old, rc = -EPERM);
2811 mnew = mdt_object_find(info->mti_env, info->mti_mdt, new_fid);
2813 GOTO(out_put_old, rc = PTR_ERR(mnew));
2815 if (!mdt_object_exists(mnew)) {
2816 LU_OBJECT_DEBUG(D_INODE, info->mti_env,
2818 "object does not exist");
2819 GOTO(out_put_new, rc = -ENOENT);
2822 if (mdt_object_remote(mnew)) {
2823 struct mdt_body *repbody;
2825 /* Always send rename req to the target child MDT */
2826 repbody = req_capsule_server_get(info->mti_pill,
2828 LASSERT(repbody != NULL);
2829 repbody->mbo_fid1 = *new_fid;
2830 repbody->mbo_valid |= (OBD_MD_FLID | OBD_MD_MDS);
2831 GOTO(out_put_new, rc = -EXDEV);
2833 /* Before locking the target dir, check we do not replace
2834 * a dir with a non-dir, otherwise it may deadlock with
2835 * link op which tries to create a link in this dir
2836 * back to this non-dir.
2838 if (S_ISDIR(lu_object_attr(&mnew->mot_obj)) &&
2839 !S_ISDIR(lu_object_attr(&mold->mot_obj)))
2840 GOTO(out_put_new, rc = -EISDIR);
2842 lh_oldp = &info->mti_lh[MDT_LH_OLD];
2843 lh_rmt = &info->mti_lh[MDT_LH_RMT];
2844 mdt_lock_reg_init(lh_oldp, LCK_EX);
2845 mdt_lock_reg_init(lh_rmt, LCK_EX);
2846 lock_ibits = MDS_INODELOCK_LOOKUP | MDS_INODELOCK_XATTR;
2847 rc = mdt_rename_source_lock(info, msrcdir, mold, lh_oldp,
2848 lh_rmt, lock_ibits, cos_incompat);
2850 GOTO(out_put_new, rc);
2852 /* Check if @msrcdir is subdir of @mnew, before locking child
2853 * to avoid reverse locking.
2855 if (mtgtdir != msrcdir) {
2856 rc = mdo_is_subdir(info->mti_env,
2857 mdt_object_child(msrcdir), new_fid);
2861 GOTO(out_unlock_old, rc);
2865 /* We used to acquire MDS_INODELOCK_FULL here but we
2866 * can't do this now because a running HSM restore on
2867 * the rename onto victim will hold the layout
2868 * lock. See LU-4002.
2871 lh_newp = &info->mti_lh[MDT_LH_NEW];
2872 mdt_lock_reg_init(lh_newp, LCK_EX);
2873 lock_ibits = MDS_INODELOCK_LOOKUP | MDS_INODELOCK_UPDATE;
2874 if (mdt_object_remote(mtgtdir)) {
2875 rc = mdt_remote_object_lock(info, mtgtdir,
2876 mdt_object_fid(mnew),
2877 &lh_newp->mlh_rreg_lh,
2878 lh_newp->mlh_rreg_mode,
2879 MDS_INODELOCK_LOOKUP,
2882 GOTO(out_unlock_old, rc);
2884 lock_ibits &= ~MDS_INODELOCK_LOOKUP;
2886 rc = mdt_reint_object_lock(info, mnew, lh_newp, lock_ibits,
2889 GOTO(out_unlock_new, rc);
2891 /* get and save version after locking */
2892 mdt_version_get_save(info, mnew, 3);
2893 } else if (rc != -ENOENT) {
2894 GOTO(out_put_old, rc);
2896 lh_oldp = &info->mti_lh[MDT_LH_OLD];
2897 lh_rmt = &info->mti_lh[MDT_LH_RMT];
2898 mdt_lock_reg_init(lh_oldp, LCK_EX);
2899 mdt_lock_reg_init(lh_rmt, LCK_EX);
2900 lock_ibits = MDS_INODELOCK_LOOKUP | MDS_INODELOCK_XATTR;
2901 rc = mdt_rename_source_lock(info, msrcdir, mold, lh_oldp,
2902 lh_rmt, lock_ibits, cos_incompat);
2904 GOTO(out_put_old, rc);
2906 mdt_enoent_version_save(info, 3);
2909 /* step 5: rename it */
2910 mdt_reint_init_ma(info, ma);
2912 mdt_fail_write(info->mti_env, info->mti_mdt->mdt_bottom,
2913 OBD_FAIL_MDS_REINT_RENAME_WRITE);
2916 mutex_lock(&mnew->mot_lov_mutex);
2918 rc = mdo_rename(info->mti_env, mdt_object_child(msrcdir),
2919 mdt_object_child(mtgtdir), old_fid, &rr->rr_name,
2920 mnew != NULL ? mdt_object_child(mnew) : NULL,
2921 &rr->rr_tgt_name, ma);
2924 mutex_unlock(&mnew->mot_lov_mutex);
2926 /* handle last link of tgt object */
2928 mdt_counter_incr(req, LPROC_MDT_RENAME,
2929 ktime_us_delta(ktime_get(), kstart));
2931 mdt_handle_last_unlink(info, mnew, ma);
2932 discard = mdt_dom_check_for_discard(info, mnew);
2934 mdt_rename_counter_tally(info, info->mti_mdt, req,
2936 ktime_us_delta(ktime_get(), kstart));
2942 mdt_object_unlock(info, mnew, lh_newp, rc);
2944 mdt_object_unlock(info, NULL, lh_rmt, rc);
2945 mdt_object_unlock(info, mold, lh_oldp, rc);
2947 if (mnew && !discard)
2948 mdt_object_put(info->mti_env, mnew);
2950 mdt_object_put(info->mti_env, mold);
2952 mdt_object_unlock(info, mtgtdir, lh_tgtdirp, rc);
2953 mdt_object_unlock(info, msrcdir, lh_srcdirp, rc);
2955 if (lustre_handle_is_used(&rename_lh))
2956 mdt_rename_unlock(&rename_lh);
2958 mdt_object_put(info->mti_env, mtgtdir);
2960 mdt_object_put(info->mti_env, msrcdir);
2962 /* The DoM discard can be done right in the place above where it is
2963 * assigned, meanwhile it is done here after rename unlock due to
2964 * compatibility with old clients, for them the discard blocks
2965 * the main thread until completion. Check LU-11359 for details.
2968 mdt_dom_discard_data(info, mnew);
2969 mdt_object_put(info->mti_env, mnew);
2971 OBD_RACE(OBD_FAIL_MDS_LINK_RENAME_RACE);
2975 static int mdt_reint_resync(struct mdt_thread_info *info,
2976 struct mdt_lock_handle *lhc)
2978 struct mdt_reint_record *rr = &info->mti_rr;
2979 struct ptlrpc_request *req = mdt_info_req(info);
2980 struct md_attr *ma = &info->mti_attr;
2981 struct mdt_object *mo;
2982 struct ldlm_lock *lease;
2983 struct mdt_body *repbody;
2984 struct md_layout_change layout = { .mlc_mirror_id = rr->rr_mirror_id };
2989 DEBUG_REQ(D_INODE, req, DFID", FLR file resync", PFID(rr->rr_fid1));
2991 if (info->mti_dlm_req)
2992 ldlm_request_cancel(req, info->mti_dlm_req, 0, LATF_SKIP);
2994 mo = mdt_object_find(info->mti_env, info->mti_mdt, rr->rr_fid1);
2996 GOTO(out, rc = PTR_ERR(mo));
2998 if (!mdt_object_exists(mo))
2999 GOTO(out_obj, rc = -ENOENT);
3001 if (!S_ISREG(lu_object_attr(&mo->mot_obj)))
3002 GOTO(out_obj, rc = -EINVAL);
3004 if (mdt_object_remote(mo))
3005 GOTO(out_obj, rc = -EREMOTE);
3007 lease = ldlm_handle2lock(rr->rr_lease_handle);
3009 GOTO(out_obj, rc = -ESTALE);
3011 /* It's really necessary to grab open_sem and check if the lease lock
3012 * has been lost. There would exist a concurrent writer coming in and
3013 * generating some dirty data in memory cache, the writeback would fail
3014 * after the layout version is increased by MDS_REINT_RESYNC RPC.
3016 if (!down_write_trylock(&mo->mot_open_sem))
3017 GOTO(out_put_lease, rc = -EBUSY);
3019 lock_res_and_lock(lease);
3020 lease_broken = ldlm_is_cancel(lease);
3021 unlock_res_and_lock(lease);
3023 GOTO(out_unlock, rc = -EBUSY);
3025 /* the file has yet opened by anyone else after we took the lease. */
3026 layout.mlc_opc = MD_LAYOUT_RESYNC;
3027 lhc = &info->mti_lh[MDT_LH_LOCAL];
3028 rc = mdt_layout_change(info, mo, lhc, &layout);
3030 GOTO(out_unlock, rc);
3032 mdt_object_unlock(info, mo, lhc, 0);
3034 ma->ma_need = MA_INODE;
3036 rc = mdt_attr_get_complex(info, mo, ma);
3038 GOTO(out_unlock, rc);
3040 repbody = req_capsule_server_get(info->mti_pill, &RMF_MDT_BODY);
3041 mdt_pack_attr2body(info, repbody, &ma->ma_attr, mdt_object_fid(mo));
3045 up_write(&mo->mot_open_sem);
3047 LDLM_LOCK_PUT(lease);
3049 mdt_object_put(info->mti_env, mo);
3051 mdt_client_compatibility(info);
3052 rc2 = mdt_fix_reply(info);
3058 struct mdt_reinter {
3059 int (*mr_handler)(struct mdt_thread_info *, struct mdt_lock_handle *);
3060 enum lprocfs_extra_opc mr_extra_opc;
3063 static const struct mdt_reinter mdt_reinters[] = {
3065 .mr_handler = &mdt_reint_setattr,
3066 .mr_extra_opc = MDS_REINT_SETATTR,
3069 .mr_handler = &mdt_reint_create,
3070 .mr_extra_opc = MDS_REINT_CREATE,
3073 .mr_handler = &mdt_reint_link,
3074 .mr_extra_opc = MDS_REINT_LINK,
3077 .mr_handler = &mdt_reint_unlink,
3078 .mr_extra_opc = MDS_REINT_UNLINK,
3081 .mr_handler = &mdt_reint_rename,
3082 .mr_extra_opc = MDS_REINT_RENAME,
3085 .mr_handler = &mdt_reint_open,
3086 .mr_extra_opc = MDS_REINT_OPEN,
3088 [REINT_SETXATTR] = {
3089 .mr_handler = &mdt_reint_setxattr,
3090 .mr_extra_opc = MDS_REINT_SETXATTR,
3093 .mr_handler = &mdt_reint_unlink,
3094 .mr_extra_opc = MDS_REINT_UNLINK,
3097 .mr_handler = &mdt_reint_migrate,
3098 .mr_extra_opc = MDS_REINT_RENAME,
3101 .mr_handler = &mdt_reint_resync,
3102 .mr_extra_opc = MDS_REINT_RESYNC,
3106 int mdt_reint_rec(struct mdt_thread_info *info,
3107 struct mdt_lock_handle *lhc)
3109 const struct mdt_reinter *mr;
3113 if (!(info->mti_rr.rr_opcode < ARRAY_SIZE(mdt_reinters)))
3116 mr = &mdt_reinters[info->mti_rr.rr_opcode];
3117 if (mr->mr_handler == NULL)
3120 rc = (*mr->mr_handler)(info, lhc);
3122 lprocfs_counter_incr(ptlrpc_req2svc(mdt_info_req(info))->srv_stats,
3123 PTLRPC_LAST_CNTR + mr->mr_extra_opc);