1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
2 * vim:expandtab:shiftwidth=8:tabstop=8:
6 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License version 2 only,
10 * as published by the Free Software Foundation.
12 * This program is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * General Public License version 2 for more details (a copy is included
16 * in the LICENSE file that accompanied this code).
18 * You should have received a copy of the GNU General Public License
19 * version 2 along with this program; If not, see
20 * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
22 * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
23 * CA 95054 USA or visit www.sun.com if you need additional information or
29 * Copyright 2008 Sun Microsystems, Inc. All rights reserved
30 * Use is subject to license terms.
33 * This file is part of Lustre, http://www.lustre.org/
34 * Lustre is a trademark of Sun Microsystems, Inc.
36 * lustre/mdd/mdd_dir.c
38 * Lustre Metadata Server (mdd) routines
40 * Author: Wang Di <wangdi@clusterfs.com>
44 # define EXPORT_SYMTAB
46 #define DEBUG_SUBSYSTEM S_MDS
48 #include <linux/module.h>
49 #include <linux/jbd.h>
51 #include <obd_class.h>
52 #include <lustre_ver.h>
53 #include <obd_support.h>
54 #include <lprocfs_status.h>
56 #include <linux/ldiskfs_fs.h>
57 #include <lustre_mds.h>
58 #include <lustre/lustre_idl.h>
59 #include <lustre_fid.h>
61 #include "mdd_internal.h"
63 static const char dot[] = ".";
64 static const char dotdot[] = "..";
66 static struct lu_name lname_dotdot = {
71 static int __mdd_lookup(const struct lu_env *env, struct md_object *pobj,
72 const struct lu_name *lname, struct lu_fid* fid,
74 static int mdd_links_add(const struct lu_env *env,
75 struct mdd_object *mdd_obj,
76 const struct lu_fid *pfid,
77 const struct lu_name *lname,
78 struct thandle *handle);
79 static int mdd_links_rename(const struct lu_env *env,
80 struct mdd_object *mdd_obj,
81 const struct lu_fid *oldpfid,
82 const struct lu_name *oldlname,
83 const struct lu_fid *newpfid,
84 const struct lu_name *newlname,
85 struct thandle *handle);
88 __mdd_lookup_locked(const struct lu_env *env, struct md_object *pobj,
89 const struct lu_name *lname, struct lu_fid* fid, int mask)
91 const char *name = lname->ln_name;
92 struct mdd_object *mdd_obj = md2mdd_obj(pobj);
93 struct dynlock_handle *dlh;
96 dlh = mdd_pdo_read_lock(env, mdd_obj, name, MOR_TGT_PARENT);
97 if (unlikely(dlh == NULL))
99 rc = __mdd_lookup(env, pobj, lname, fid, mask);
100 mdd_pdo_read_unlock(env, mdd_obj, dlh);
105 int mdd_lookup(const struct lu_env *env,
106 struct md_object *pobj, const struct lu_name *lname,
107 struct lu_fid* fid, struct md_op_spec *spec)
111 rc = __mdd_lookup_locked(env, pobj, lname, fid, MAY_EXEC);
115 static int mdd_parent_fid(const struct lu_env *env, struct mdd_object *obj,
118 return __mdd_lookup_locked(env, &obj->mod_obj, &lname_dotdot, fid, 0);
122 * For root fid use special function, which does not compare version component
123 * of fid. Version component is different for root fids on all MDTs.
125 int mdd_is_root(struct mdd_device *mdd, const struct lu_fid *fid)
127 return fid_seq(&mdd->mdd_root_fid) == fid_seq(fid) &&
128 fid_oid(&mdd->mdd_root_fid) == fid_oid(fid);
132 * return 1: if lf is the fid of the ancestor of p1;
135 * return -EREMOTE: if remote object is found, in this
136 * case fid of remote object is saved to @pf;
138 * otherwise: values < 0, errors.
140 static int mdd_is_parent(const struct lu_env *env,
141 struct mdd_device *mdd,
142 struct mdd_object *p1,
143 const struct lu_fid *lf,
146 struct mdd_object *parent = NULL;
151 LASSERT(!lu_fid_eq(mdo2fid(p1), lf));
152 pfid = &mdd_env_info(env)->mti_fid;
154 /* Check for root first. */
155 if (mdd_is_root(mdd, mdo2fid(p1)))
159 /* this is done recursively, bypass capa for each obj */
160 mdd_set_capainfo(env, 4, p1, BYPASS_CAPA);
161 rc = mdd_parent_fid(env, p1, pfid);
164 if (mdd_is_root(mdd, pfid))
166 if (lu_fid_eq(pfid, lf))
169 mdd_object_put(env, parent);
170 parent = mdd_object_find(env, mdd, pfid);
172 /* cross-ref parent */
173 if (parent == NULL) {
176 GOTO(out, rc = -EREMOTE);
177 } else if (IS_ERR(parent))
178 GOTO(out, rc = PTR_ERR(parent));
183 if (parent && !IS_ERR(parent))
184 mdd_object_put(env, parent);
189 * No permission check is needed.
191 * returns 1: if fid is ancestor of @mo;
192 * returns 0: if fid is not a ancestor of @mo;
194 * returns EREMOTE if remote object is found, fid of remote object is saved to
197 * returns < 0: if error
199 static int mdd_is_subdir(const struct lu_env *env,
200 struct md_object *mo, const struct lu_fid *fid,
203 struct mdd_device *mdd = mdo2mdd(mo);
207 if (!S_ISDIR(mdd_object_type(md2mdd_obj(mo))))
210 rc = mdd_is_parent(env, mdd, md2mdd_obj(mo), fid, sfid);
214 } else if (rc == 1) {
215 /* found @fid is parent */
223 * Check that @dir contains no entries except (possibly) dot and dotdot.
228 * -ENOTDIR not a directory object
229 * -ENOTEMPTY not empty
233 static int mdd_dir_is_empty(const struct lu_env *env,
234 struct mdd_object *dir)
237 struct dt_object *obj;
238 const struct dt_it_ops *iops;
242 obj = mdd_object_child(dir);
243 if (!dt_try_as_dir(env, obj))
246 iops = &obj->do_index_ops->dio_it;
247 it = iops->init(env, obj, BYPASS_CAPA);
249 result = iops->get(env, it, (const void *)"");
252 for (result = 0, i = 0; result == 0 && i < 3; ++i)
253 result = iops->next(env, it);
256 else if (result == +1)
258 } else if (result == 0)
260 * Huh? Index contains no zero key?
271 static int __mdd_may_link(const struct lu_env *env, struct mdd_object *obj)
273 struct mdd_device *m = mdd_obj2mdd_dev(obj);
274 struct lu_attr *la = &mdd_env_info(env)->mti_la;
278 rc = mdd_la_get(env, obj, la, BYPASS_CAPA);
283 * Subdir count limitation can be broken through.
285 if (la->la_nlink >= m->mdd_dt_conf.ddp_max_nlink &&
286 !S_ISDIR(la->la_mode))
293 * Check whether it may create the cobj under the pobj.
296 int mdd_may_create(const struct lu_env *env, struct mdd_object *pobj,
297 struct mdd_object *cobj, int check_perm, int check_nlink)
302 if (cobj && mdd_object_exists(cobj))
305 if (mdd_is_dead_obj(pobj))
309 rc = mdd_permission_internal_locked(env, pobj, NULL,
310 MAY_WRITE | MAY_EXEC,
313 if (!rc && check_nlink)
314 rc = __mdd_may_link(env, pobj);
320 * Check whether can unlink from the pobj in the case of "cobj == NULL".
322 int mdd_may_unlink(const struct lu_env *env, struct mdd_object *pobj,
323 const struct md_attr *ma)
328 if (mdd_is_dead_obj(pobj))
331 if ((ma->ma_attr.la_valid & LA_FLAGS) &&
332 (ma->ma_attr.la_flags & (LUSTRE_APPEND_FL | LUSTRE_IMMUTABLE_FL)))
335 rc = mdd_permission_internal_locked(env, pobj, NULL,
336 MAY_WRITE | MAY_EXEC,
341 if (mdd_is_append(pobj))
348 * pobj == NULL is remote ops case, under such case, pobj's
349 * VTX feature has been checked already, no need check again.
351 static inline int mdd_is_sticky(const struct lu_env *env,
352 struct mdd_object *pobj,
353 struct mdd_object *cobj)
355 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
356 struct md_ucred *uc = md_ucred(env);
360 rc = mdd_la_get(env, pobj, tmp_la, BYPASS_CAPA);
364 if (!(tmp_la->la_mode & S_ISVTX) ||
365 (tmp_la->la_uid == uc->mu_fsuid))
369 rc = mdd_la_get(env, cobj, tmp_la, BYPASS_CAPA);
373 if (tmp_la->la_uid == uc->mu_fsuid)
376 return !mdd_capable(uc, CFS_CAP_FOWNER);
380 * Check whether it may delete the cobj from the pobj.
383 int mdd_may_delete(const struct lu_env *env, struct mdd_object *pobj,
384 struct mdd_object *cobj, struct md_attr *ma,
385 int check_perm, int check_empty)
391 if (!mdd_object_exists(cobj))
395 if (mdd_is_dead_obj(pobj))
399 rc = mdd_permission_internal_locked(env, pobj, NULL,
400 MAY_WRITE | MAY_EXEC,
406 if (mdd_is_append(pobj))
410 if (!(ma->ma_attr_flags & MDS_VTX_BYPASS) &&
411 mdd_is_sticky(env, pobj, cobj))
414 if (mdd_is_immutable(cobj) || mdd_is_append(cobj))
417 if ((ma->ma_attr.la_valid & LA_FLAGS) &&
418 (ma->ma_attr.la_flags & (LUSTRE_APPEND_FL | LUSTRE_IMMUTABLE_FL)))
421 if (S_ISDIR(ma->ma_attr.la_mode)) {
422 struct mdd_device *mdd = mdo2mdd(&cobj->mod_obj);
424 if (!S_ISDIR(mdd_object_type(cobj)))
427 if (lu_fid_eq(mdo2fid(cobj), &mdd->mdd_root_fid))
429 } else if (S_ISDIR(mdd_object_type(cobj)))
432 if (S_ISDIR(ma->ma_attr.la_mode) && check_empty)
433 rc = mdd_dir_is_empty(env, cobj);
440 * has mdd_write_lock on src already, but not on tgt yet
442 int mdd_link_sanity_check(const struct lu_env *env,
443 struct mdd_object *tgt_obj,
444 const struct lu_name *lname,
445 struct mdd_object *src_obj)
447 struct mdd_device *m = mdd_obj2mdd_dev(src_obj);
451 /* Local ops, no lookup before link, check filename length here. */
452 if (lname && (lname->ln_namelen > m->mdd_dt_conf.ddp_max_name_len))
453 RETURN(-ENAMETOOLONG);
455 if (mdd_is_immutable(src_obj) || mdd_is_append(src_obj))
458 if (S_ISDIR(mdd_object_type(src_obj)))
461 LASSERT(src_obj != tgt_obj);
463 rc = mdd_may_create(env, tgt_obj, NULL, 1, 0);
468 rc = __mdd_may_link(env, src_obj);
474 * If subdir count is up to ddp_max_nlink, then enable MNLINK_OBJ flag and
475 * assign i_nlink to 1 which means the i_nlink for subdir count is incredible
476 * (maybe too large to be represented). It is a trick to break through the
477 * "i_nlink" limitation for subdir count.
479 void __mdd_ref_add(const struct lu_env *env, struct mdd_object *obj,
480 struct thandle *handle)
482 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
483 struct mdd_device *m = mdd_obj2mdd_dev(obj);
485 if (!mdd_is_mnlink(obj)) {
486 if (S_ISDIR(mdd_object_type(obj))) {
487 if (mdd_la_get(env, obj, tmp_la, BYPASS_CAPA))
490 if (tmp_la->la_nlink >= m->mdd_dt_conf.ddp_max_nlink) {
491 obj->mod_flags |= MNLINK_OBJ;
492 tmp_la->la_nlink = 1;
493 tmp_la->la_valid = LA_NLINK;
494 mdd_attr_set_internal(env, obj, tmp_la, handle,
499 mdo_ref_add(env, obj, handle);
503 void __mdd_ref_del(const struct lu_env *env, struct mdd_object *obj,
504 struct thandle *handle, int is_dot)
506 if (!mdd_is_mnlink(obj) || is_dot)
507 mdo_ref_del(env, obj, handle);
510 /* insert named index, add reference if isdir */
511 static int __mdd_index_insert(const struct lu_env *env, struct mdd_object *pobj,
512 const struct lu_fid *lf, const char *name, int is_dir,
513 struct thandle *handle, struct lustre_capa *capa)
515 struct dt_object *next = mdd_object_child(pobj);
519 if (dt_try_as_dir(env, next)) {
520 struct md_ucred *uc = md_ucred(env);
522 rc = next->do_index_ops->dio_insert(env, next,
523 __mdd_fid_rec(env, lf),
524 (const struct dt_key *)name,
525 handle, capa, uc->mu_cap &
526 CFS_CAP_SYS_RESOURCE_MASK);
533 mdd_write_lock(env, pobj, MOR_TGT_PARENT);
534 __mdd_ref_add(env, pobj, handle);
535 mdd_write_unlock(env, pobj);
541 /* delete named index, drop reference if isdir */
542 static int __mdd_index_delete(const struct lu_env *env, struct mdd_object *pobj,
543 const char *name, int is_dir, struct thandle *handle,
544 struct lustre_capa *capa)
546 struct dt_object *next = mdd_object_child(pobj);
550 if (dt_try_as_dir(env, next)) {
551 rc = next->do_index_ops->dio_delete(env, next,
552 (struct dt_key *)name,
554 if (rc == 0 && is_dir) {
557 if (name != NULL && name[0] == '.' && name[1] == 0)
559 mdd_write_lock(env, pobj, MOR_TGT_PARENT);
560 __mdd_ref_del(env, pobj, handle, is_dot);
561 mdd_write_unlock(env, pobj);
570 __mdd_index_insert_only(const struct lu_env *env, struct mdd_object *pobj,
571 const struct lu_fid *lf, const char *name,
572 struct thandle *handle, struct lustre_capa *capa)
574 struct dt_object *next = mdd_object_child(pobj);
578 if (dt_try_as_dir(env, next)) {
579 struct md_ucred *uc = md_ucred(env);
581 rc = next->do_index_ops->dio_insert(env, next,
582 __mdd_fid_rec(env, lf),
583 (const struct dt_key *)name,
584 handle, capa, uc->mu_cap &
585 CFS_CAP_SYS_RESOURCE_MASK);
592 /** Store a namespace change changelog record
593 * If this fails, we must fail the whole transaction; we don't
594 * want the change to commit without the log entry.
595 * \param target - mdd_object of change
596 * \param parent - parent dir/object
597 * \param tf - target lu_fid, overrides fid of \a target if this is non-null
598 * \param tname - target name string
599 * \param handle - transacion handle
601 static int mdd_changelog_ns_store(const struct lu_env *env,
602 struct mdd_device *mdd,
603 enum changelog_rec_type type,
604 struct mdd_object *target,
605 struct mdd_object *parent,
606 const struct lu_fid *tf,
607 const struct lu_name *tname,
608 struct thandle *handle)
610 const struct lu_fid *tfid;
611 const struct lu_fid *tpfid = mdo2fid(parent);
612 struct llog_changelog_rec *rec;
618 if (!(mdd->mdd_cl.mc_flags & CLM_ON))
621 LASSERT(parent != NULL);
622 LASSERT(tname != NULL);
623 LASSERT(handle != NULL);
626 reclen = llog_data_len(sizeof(*rec) + tname->ln_namelen);
627 buf = mdd_buf_alloc(env, reclen);
628 if (buf->lb_buf == NULL)
630 rec = (struct llog_changelog_rec *)buf->lb_buf;
632 rec->cr_flags = CLF_VERSION;
633 rec->cr_type = (__u32)type;
634 tfid = tf ? tf : mdo2fid(target);
635 rec->cr_tfid = *tfid;
636 rec->cr_pfid = *tpfid;
637 rec->cr_namelen = tname->ln_namelen;
638 memcpy(rec->cr_name, tname->ln_name, rec->cr_namelen);
640 target->mod_cltime = cfs_time_current_64();
642 rc = mdd_changelog_llog_write(mdd, rec, handle);
644 CERROR("changelog failed: rc=%d, op%d %s c"DFID" p"DFID"\n",
645 rc, type, tname->ln_name, PFID(tfid), PFID(tpfid));
652 static int mdd_link(const struct lu_env *env, struct md_object *tgt_obj,
653 struct md_object *src_obj, const struct lu_name *lname,
656 const char *name = lname->ln_name;
657 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
658 struct mdd_object *mdd_tobj = md2mdd_obj(tgt_obj);
659 struct mdd_object *mdd_sobj = md2mdd_obj(src_obj);
660 struct mdd_device *mdd = mdo2mdd(src_obj);
661 struct dynlock_handle *dlh;
662 struct thandle *handle;
663 #ifdef HAVE_QUOTA_SUPPORT
664 struct obd_device *obd = mdd->mdd_obd_dev;
665 struct mds_obd *mds = &obd->u.mds;
666 unsigned int qids[MAXQUOTAS] = { 0, 0 };
667 int quota_opc = 0, rec_pending = 0;
672 #ifdef HAVE_QUOTA_SUPPORT
673 if (mds->mds_quota) {
674 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
676 rc = mdd_la_get(env, mdd_tobj, la_tmp, BYPASS_CAPA);
678 quota_opc = FSFILT_OP_LINK;
679 mdd_quota_wrapper(la_tmp, qids);
680 /* get block quota for parent */
681 lquota_chkquota(mds_quota_interface_ref, obd,
682 qids[USRQUOTA], qids[GRPQUOTA], 1,
683 &rec_pending, NULL, LQUOTA_FLAGS_BLK,
689 mdd_txn_param_build(env, mdd, MDD_TXN_LINK_OP);
690 handle = mdd_trans_start(env, mdd);
692 GOTO(out_pending, rc = PTR_ERR(handle));
694 dlh = mdd_pdo_write_lock(env, mdd_tobj, name, MOR_TGT_CHILD);
696 GOTO(out_trans, rc = -ENOMEM);
697 mdd_write_lock(env, mdd_sobj, MOR_TGT_CHILD);
699 rc = mdd_link_sanity_check(env, mdd_tobj, lname, mdd_sobj);
701 GOTO(out_unlock, rc);
703 rc = __mdd_index_insert_only(env, mdd_tobj, mdo2fid(mdd_sobj),
705 mdd_object_capa(env, mdd_tobj));
707 GOTO(out_unlock, rc);
709 __mdd_ref_add(env, mdd_sobj, handle);
711 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
712 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
714 la->la_valid = LA_CTIME | LA_MTIME;
715 rc = mdd_attr_check_set_internal_locked(env, mdd_tobj, la, handle, 0);
717 GOTO(out_unlock, rc);
719 la->la_valid = LA_CTIME;
720 rc = mdd_attr_check_set_internal(env, mdd_sobj, la, handle, 0);
722 mdd_links_add(env, mdd_sobj, mdo2fid(mdd_tobj), lname, handle);
726 mdd_write_unlock(env, mdd_sobj);
727 mdd_pdo_write_unlock(env, mdd_tobj, dlh);
730 rc = mdd_changelog_ns_store(env, mdd, CL_HARDLINK, mdd_sobj,
731 mdd_tobj, NULL, lname, handle);
732 mdd_trans_stop(env, mdd, rc, handle);
734 #ifdef HAVE_QUOTA_SUPPORT
737 lquota_pending_commit(mds_quota_interface_ref, obd,
738 qids[USRQUOTA], qids[GRPQUOTA],
740 /* Trigger dqacq for the parent owner. If failed,
741 * the next call for lquota_chkquota will process it. */
742 lquota_adjust(mds_quota_interface_ref, obd, 0, qids, rc,
749 /* caller should take a lock before calling */
750 int mdd_finish_unlink(const struct lu_env *env,
751 struct mdd_object *obj, struct md_attr *ma,
758 rc = mdd_iattr_get(env, obj, ma);
759 if (rc == 0 && ma->ma_attr.la_nlink == 0) {
760 /* add new orphan and the object
761 * will be deleted during mdd_close() */
762 if (obj->mod_count) {
763 rc = __mdd_orphan_add(env, obj, th);
765 obj->mod_flags |= ORPHAN_OBJ;
768 obj->mod_flags |= DEAD_OBJ;
769 if (!(obj->mod_flags & ORPHAN_OBJ)) {
770 rc = mdd_object_kill(env, obj, ma);
777 ma->ma_valid &= ~(MA_LOV | MA_COOKIE);
784 * has mdd_write_lock on cobj already, but not on pobj yet
786 int mdd_unlink_sanity_check(const struct lu_env *env, struct mdd_object *pobj,
787 struct mdd_object *cobj, struct md_attr *ma)
792 rc = mdd_may_delete(env, pobj, cobj, ma, 1, 1);
797 static int mdd_unlink(const struct lu_env *env, struct md_object *pobj,
798 struct md_object *cobj, const struct lu_name *lname,
801 const char *name = lname->ln_name;
802 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
803 struct mdd_object *mdd_pobj = md2mdd_obj(pobj);
804 struct mdd_object *mdd_cobj = md2mdd_obj(cobj);
805 struct mdd_device *mdd = mdo2mdd(pobj);
806 struct dynlock_handle *dlh;
807 struct thandle *handle;
808 #ifdef HAVE_QUOTA_SUPPORT
809 struct obd_device *obd = mdd->mdd_obd_dev;
810 struct mds_obd *mds = &obd->u.mds;
811 unsigned int qcids[MAXQUOTAS] = { 0, 0 };
812 unsigned int qpids[MAXQUOTAS] = { 0, 0 };
815 int is_dir = S_ISDIR(ma->ma_attr.la_mode);
819 LASSERTF(mdd_object_exists(mdd_cobj) > 0, "FID is "DFID"\n",
820 PFID(mdd_object_fid(mdd_cobj)));
822 rc = mdd_log_txn_param_build(env, cobj, ma, MDD_TXN_UNLINK_OP);
826 handle = mdd_trans_start(env, mdd);
828 RETURN(PTR_ERR(handle));
830 dlh = mdd_pdo_write_lock(env, mdd_pobj, name, MOR_TGT_PARENT);
832 GOTO(out_trans, rc = -ENOMEM);
833 mdd_write_lock(env, mdd_cobj, MOR_TGT_CHILD);
835 rc = mdd_unlink_sanity_check(env, mdd_pobj, mdd_cobj, ma);
839 rc = __mdd_index_delete(env, mdd_pobj, name, is_dir, handle,
840 mdd_object_capa(env, mdd_pobj));
844 __mdd_ref_del(env, mdd_cobj, handle, 0);
847 __mdd_ref_del(env, mdd_cobj, handle, 1);
849 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
850 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
852 la->la_valid = LA_CTIME | LA_MTIME;
853 rc = mdd_attr_check_set_internal_locked(env, mdd_pobj, la, handle, 0);
857 la->la_valid = LA_CTIME;
858 rc = mdd_attr_check_set_internal(env, mdd_cobj, la, handle, 0);
862 rc = mdd_finish_unlink(env, mdd_cobj, ma, handle);
863 #ifdef HAVE_QUOTA_SUPPORT
864 if (mds->mds_quota && ma->ma_valid & MA_INODE &&
865 ma->ma_attr.la_nlink == 0) {
866 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
868 rc = mdd_la_get(env, mdd_pobj, la_tmp, BYPASS_CAPA);
870 mdd_quota_wrapper(la_tmp, qpids);
871 if (mdd_cobj->mod_count == 0) {
872 quota_opc = FSFILT_OP_UNLINK;
873 mdd_quota_wrapper(&ma->ma_attr, qcids);
875 quota_opc = FSFILT_OP_UNLINK_PARTIAL_PARENT;
882 obd_set_info_async(mdd2obd_dev(mdd)->u.mds.mds_osc_exp,
883 sizeof(KEY_UNLINKED), KEY_UNLINKED, 0,
886 /* old files may not have link ea; ignore errors */
887 mdd_links_rename(env, mdd_cobj, mdo2fid(mdd_pobj),
888 lname, NULL, NULL, handle);
892 mdd_write_unlock(env, mdd_cobj);
893 mdd_pdo_write_unlock(env, mdd_pobj, dlh);
896 rc = mdd_changelog_ns_store(env, mdd,
897 is_dir ? CL_RMDIR : CL_UNLINK,
898 mdd_cobj, mdd_pobj, NULL, lname,
901 mdd_trans_stop(env, mdd, rc, handle);
902 #ifdef HAVE_QUOTA_SUPPORT
904 /* Trigger dqrel on the owner of child and parent. If failed,
905 * the next call for lquota_chkquota will process it. */
906 lquota_adjust(mds_quota_interface_ref, obd, qcids, qpids, rc,
912 /* has not lock on pobj yet */
913 static int mdd_ni_sanity_check(const struct lu_env *env,
914 struct md_object *pobj,
915 const struct md_attr *ma)
917 struct mdd_object *obj = md2mdd_obj(pobj);
921 if (ma->ma_attr_flags & MDS_PERM_BYPASS)
924 rc = mdd_may_create(env, obj, NULL, 1, S_ISDIR(ma->ma_attr.la_mode));
932 static int mdd_name_insert(const struct lu_env *env,
933 struct md_object *pobj,
934 const struct lu_name *lname,
935 const struct lu_fid *fid,
936 const struct md_attr *ma)
938 const char *name = lname->ln_name;
939 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
940 struct mdd_object *mdd_obj = md2mdd_obj(pobj);
941 struct mdd_device *mdd = mdo2mdd(pobj);
942 struct dynlock_handle *dlh;
943 struct thandle *handle;
944 int is_dir = S_ISDIR(ma->ma_attr.la_mode);
945 #ifdef HAVE_QUOTA_SUPPORT
946 struct md_ucred *uc = md_ucred(env);
947 struct obd_device *obd = mdd->mdd_obd_dev;
948 struct mds_obd *mds = &obd->u.mds;
949 unsigned int qids[MAXQUOTAS] = { 0, 0 };
950 int quota_opc = 0, rec_pending = 0;
951 cfs_cap_t save = uc->mu_cap;
956 #ifdef HAVE_QUOTA_SUPPORT
957 if (mds->mds_quota) {
958 if (!(ma->ma_attr_flags & MDS_QUOTA_IGNORE)) {
959 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
961 rc = mdd_la_get(env, mdd_obj, la_tmp, BYPASS_CAPA);
963 quota_opc = FSFILT_OP_LINK;
964 mdd_quota_wrapper(la_tmp, qids);
965 /* get block quota for parent */
966 lquota_chkquota(mds_quota_interface_ref, obd,
967 qids[USRQUOTA], qids[GRPQUOTA],
968 1, &rec_pending, NULL,
969 LQUOTA_FLAGS_BLK, NULL, 0);
972 uc->mu_cap |= CFS_CAP_SYS_RESOURCE_MASK;
976 mdd_txn_param_build(env, mdd, MDD_TXN_INDEX_INSERT_OP);
977 handle = mdd_trans_start(env, mdo2mdd(pobj));
979 GOTO(out_pending, rc = PTR_ERR(handle));
981 dlh = mdd_pdo_write_lock(env, mdd_obj, name, MOR_TGT_PARENT);
983 GOTO(out_trans, rc = -ENOMEM);
985 rc = mdd_ni_sanity_check(env, pobj, ma);
987 GOTO(out_unlock, rc);
989 rc = __mdd_index_insert(env, mdd_obj, fid, name, is_dir,
990 handle, BYPASS_CAPA);
992 GOTO(out_unlock, rc);
995 * For some case, no need update obj's ctime (LA_CTIME is not set),
997 * For other cases, update obj's ctime (LA_CTIME is set),
1000 if (ma->ma_attr.la_valid & LA_CTIME) {
1001 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
1002 la->la_valid = LA_CTIME | LA_MTIME;
1003 rc = mdd_attr_check_set_internal_locked(env, mdd_obj, la,
1008 mdd_pdo_write_unlock(env, mdd_obj, dlh);
1010 mdd_trans_stop(env, mdo2mdd(pobj), rc, handle);
1012 #ifdef HAVE_QUOTA_SUPPORT
1013 if (mds->mds_quota) {
1016 lquota_pending_commit(mds_quota_interface_ref,
1017 obd, qids[USRQUOTA],
1020 /* Trigger dqacq for the parent owner. If failed,
1021 * the next call for lquota_chkquota will process it*/
1022 lquota_adjust(mds_quota_interface_ref, obd, 0, qids,
1032 /* has not lock on pobj yet */
1033 static int mdd_nr_sanity_check(const struct lu_env *env,
1034 struct md_object *pobj,
1035 const struct md_attr *ma)
1037 struct mdd_object *obj = md2mdd_obj(pobj);
1041 if (ma->ma_attr_flags & MDS_PERM_BYPASS)
1044 rc = mdd_may_unlink(env, obj, ma);
1050 * Partial operation.
1052 static int mdd_name_remove(const struct lu_env *env,
1053 struct md_object *pobj,
1054 const struct lu_name *lname,
1055 const struct md_attr *ma)
1057 const char *name = lname->ln_name;
1058 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
1059 struct mdd_object *mdd_obj = md2mdd_obj(pobj);
1060 struct mdd_device *mdd = mdo2mdd(pobj);
1061 struct dynlock_handle *dlh;
1062 struct thandle *handle;
1063 int is_dir = S_ISDIR(ma->ma_attr.la_mode);
1064 #ifdef HAVE_QUOTA_SUPPORT
1065 struct obd_device *obd = mdd->mdd_obd_dev;
1066 struct mds_obd *mds = &obd->u.mds;
1067 unsigned int qids[MAXQUOTAS] = { 0, 0 };
1073 #ifdef HAVE_QUOTA_SUPPORT
1074 if (mds->mds_quota) {
1075 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
1077 rc = mdd_la_get(env, mdd_obj, la_tmp, BYPASS_CAPA);
1079 quota_opc = FSFILT_OP_UNLINK_PARTIAL_PARENT;
1080 mdd_quota_wrapper(la_tmp, qids);
1084 mdd_txn_param_build(env, mdd, MDD_TXN_INDEX_DELETE_OP);
1085 handle = mdd_trans_start(env, mdd);
1087 GOTO(out_pending, rc = PTR_ERR(handle));
1089 dlh = mdd_pdo_write_lock(env, mdd_obj, name, MOR_TGT_PARENT);
1091 GOTO(out_trans, rc = -ENOMEM);
1093 rc = mdd_nr_sanity_check(env, pobj, ma);
1095 GOTO(out_unlock, rc);
1097 rc = __mdd_index_delete(env, mdd_obj, name, is_dir,
1098 handle, BYPASS_CAPA);
1100 GOTO(out_unlock, rc);
1103 * For some case, no need update obj's ctime (LA_CTIME is not set),
1105 * For other cases, update obj's ctime (LA_CTIME is set),
1108 if (ma->ma_attr.la_valid & LA_CTIME) {
1109 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
1110 la->la_valid = LA_CTIME | LA_MTIME;
1111 rc = mdd_attr_check_set_internal_locked(env, mdd_obj, la,
1116 mdd_pdo_write_unlock(env, mdd_obj, dlh);
1118 mdd_trans_stop(env, mdd, rc, handle);
1120 #ifdef HAVE_QUOTA_SUPPORT
1121 /* Trigger dqrel for the parent owner.
1122 * If failed, the next call for lquota_chkquota will process it. */
1124 lquota_adjust(mds_quota_interface_ref, obd, 0, qids, rc,
1132 * has mdd_write_lock on tobj alreay, but not on tgt_pobj yet
1134 static int mdd_rt_sanity_check(const struct lu_env *env,
1135 struct mdd_object *tgt_pobj,
1136 struct mdd_object *tobj,
1142 if (unlikely(ma->ma_attr_flags & MDS_PERM_BYPASS))
1145 /* XXX: for mdd_rename_tgt, "tobj == NULL" does not mean tobj not
1146 * exist. In fact, tobj must exist, otherwise the call trace will be:
1147 * mdt_reint_rename_tgt -> mdo_name_insert -> ... -> mdd_name_insert.
1148 * When get here, tobj must be NOT NULL, the other case has been
1149 * processed in cmr_rename_tgt before mdd_rename_tgt and enable
1151 * So check may_delete, but not check nlink of tgt_pobj. */
1153 rc = mdd_may_delete(env, tgt_pobj, tobj, ma, 1, 1);
1158 /* Partial rename op on slave MDD */
1159 static int mdd_rename_tgt(const struct lu_env *env,
1160 struct md_object *pobj, struct md_object *tobj,
1161 const struct lu_fid *lf, const struct lu_name *lname,
1164 const char *name = lname->ln_name;
1165 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
1166 struct mdd_object *mdd_tpobj = md2mdd_obj(pobj);
1167 struct mdd_object *mdd_tobj = md2mdd_obj(tobj);
1168 struct mdd_device *mdd = mdo2mdd(pobj);
1169 struct dynlock_handle *dlh;
1170 struct thandle *handle;
1171 #ifdef HAVE_QUOTA_SUPPORT
1172 struct obd_device *obd = mdd->mdd_obd_dev;
1173 struct mds_obd *mds = &obd->u.mds;
1174 unsigned int qcids[MAXQUOTAS] = { 0, 0 };
1175 unsigned int qpids[MAXQUOTAS] = { 0, 0 };
1176 int quota_opc = 0, rec_pending = 0;
1181 #ifdef HAVE_QUOTA_SUPPORT
1182 if (mds->mds_quota && !tobj) {
1183 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
1185 rc = mdd_la_get(env, mdd_tpobj, la_tmp, BYPASS_CAPA);
1187 quota_opc = FSFILT_OP_LINK;
1188 mdd_quota_wrapper(la_tmp, qpids);
1189 /* get block quota for target parent */
1190 lquota_chkquota(mds_quota_interface_ref, obd,
1191 qpids[USRQUOTA], qpids[GRPQUOTA], 1,
1192 &rec_pending, NULL, LQUOTA_FLAGS_BLK,
1197 mdd_txn_param_build(env, mdd, MDD_TXN_RENAME_TGT_OP);
1198 handle = mdd_trans_start(env, mdd);
1200 GOTO(out_pending, rc = PTR_ERR(handle));
1202 dlh = mdd_pdo_write_lock(env, mdd_tpobj, name, MOR_TGT_PARENT);
1204 GOTO(out_trans, rc = -ENOMEM);
1206 mdd_write_lock(env, mdd_tobj, MOR_TGT_CHILD);
1208 rc = mdd_rt_sanity_check(env, mdd_tpobj, mdd_tobj, ma);
1213 * If rename_tgt is called then we should just re-insert name with
1214 * correct fid, no need to dec/inc parent nlink if obj is dir.
1216 rc = __mdd_index_delete(env, mdd_tpobj, name, 0, handle, BYPASS_CAPA);
1220 rc = __mdd_index_insert_only(env, mdd_tpobj, lf, name, handle,
1225 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
1226 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
1228 la->la_valid = LA_CTIME | LA_MTIME;
1229 rc = mdd_attr_check_set_internal_locked(env, mdd_tpobj, la, handle, 0);
1234 * For tobj is remote case cmm layer has processed
1235 * and pass NULL tobj to here. So when tobj is NOT NULL,
1236 * it must be local one.
1238 if (tobj && mdd_object_exists(mdd_tobj)) {
1239 __mdd_ref_del(env, mdd_tobj, handle, 0);
1241 /* Remove dot reference. */
1242 if (S_ISDIR(ma->ma_attr.la_mode))
1243 __mdd_ref_del(env, mdd_tobj, handle, 1);
1245 la->la_valid = LA_CTIME;
1246 rc = mdd_attr_check_set_internal(env, mdd_tobj, la, handle, 0);
1250 rc = mdd_finish_unlink(env, mdd_tobj, ma, handle);
1254 #ifdef HAVE_QUOTA_SUPPORT
1255 if (mds->mds_quota && ma->ma_valid & MA_INODE &&
1256 ma->ma_attr.la_nlink == 0 && mdd_tobj->mod_count == 0) {
1257 quota_opc = FSFILT_OP_UNLINK_PARTIAL_CHILD;
1258 mdd_quota_wrapper(&ma->ma_attr, qcids);
1265 mdd_write_unlock(env, mdd_tobj);
1266 mdd_pdo_write_unlock(env, mdd_tpobj, dlh);
1269 /* Bare EXT record with no RENAME in front of it signifies
1270 a partial slave op */
1271 rc = mdd_changelog_ns_store(env, mdd, CL_EXT, mdd_tobj,
1272 mdd_tpobj, NULL, lname, handle);
1274 mdd_trans_stop(env, mdd, rc, handle);
1276 #ifdef HAVE_QUOTA_SUPPORT
1277 if (mds->mds_quota) {
1279 lquota_pending_commit(mds_quota_interface_ref, obd,
1284 /* Trigger dqrel/dqacq on the target owner of child and
1285 * parent. If failed, the next call for lquota_chkquota
1286 * will process it. */
1287 lquota_adjust(mds_quota_interface_ref, obd, qcids,
1288 qpids, rc, quota_opc);
1295 * The permission has been checked when obj created, no need check again.
1297 static int mdd_cd_sanity_check(const struct lu_env *env,
1298 struct mdd_object *obj)
1303 if (!obj || mdd_is_dead_obj(obj))
1310 static int mdd_create_data(const struct lu_env *env, struct md_object *pobj,
1311 struct md_object *cobj, const struct md_op_spec *spec,
1314 struct mdd_device *mdd = mdo2mdd(cobj);
1315 struct mdd_object *mdd_pobj = md2mdd_obj(pobj);
1316 struct mdd_object *son = md2mdd_obj(cobj);
1317 struct lu_attr *attr = &ma->ma_attr;
1318 struct lov_mds_md *lmm = NULL;
1320 struct thandle *handle;
1324 rc = mdd_cd_sanity_check(env, son);
1328 if (!md_should_create(spec->sp_cr_flags))
1331 rc = mdd_lov_create(env, mdd, mdd_pobj, son, &lmm, &lmm_size,
1336 mdd_txn_param_build(env, mdd, MDD_TXN_CREATE_DATA_OP);
1337 handle = mdd_trans_start(env, mdd);
1339 GOTO(out_free, rc = PTR_ERR(handle));
1342 * XXX: Setting the lov ea is not locked but setting the attr is locked?
1343 * Should this be fixed?
1346 /* Replay creates has objects already */
1348 if (spec->no_create) {
1349 CDEBUG(D_INFO, "we already have lov ea\n");
1350 rc = mdd_lov_set_md(env, mdd_pobj, son,
1351 (struct lov_mds_md *)spec->u.sp_ea.eadata,
1352 spec->u.sp_ea.eadatalen, handle, 0);
1355 /* No need mdd_lsm_sanity_check here */
1356 rc = mdd_lov_set_md(env, mdd_pobj, son, lmm,
1357 lmm_size, handle, 0);
1360 rc = mdd_attr_get_internal_locked(env, son, ma);
1362 /* update lov_objid data, must be before transaction stop! */
1364 mdd_lov_objid_update(mdd, lmm);
1366 mdd_trans_stop(env, mdd, rc, handle);
1368 /* Finish mdd_lov_create() stuff. */
1369 mdd_lov_create_finish(env, mdd, lmm, lmm_size, spec);
1373 /* Get fid from name and parent */
1375 __mdd_lookup(const struct lu_env *env, struct md_object *pobj,
1376 const struct lu_name *lname, struct lu_fid* fid, int mask)
1378 const char *name = lname->ln_name;
1379 const struct dt_key *key = (const struct dt_key *)name;
1380 struct mdd_object *mdd_obj = md2mdd_obj(pobj);
1381 struct mdd_device *m = mdo2mdd(pobj);
1382 struct dt_object *dir = mdd_object_child(mdd_obj);
1383 struct lu_fid_pack *pack = &mdd_env_info(env)->mti_pack;
1387 if (unlikely(mdd_is_dead_obj(mdd_obj)))
1390 rc = mdd_object_exists(mdd_obj);
1391 if (unlikely(rc == 0))
1393 else if (unlikely(rc < 0)) {
1394 CERROR("Object "DFID" locates on remote server\n",
1395 PFID(mdo2fid(mdd_obj)));
1399 /* The common filename length check. */
1400 if (unlikely(lname->ln_namelen > m->mdd_dt_conf.ddp_max_name_len))
1401 RETURN(-ENAMETOOLONG);
1403 rc = mdd_permission_internal_locked(env, mdd_obj, NULL, mask,
1408 if (likely(S_ISDIR(mdd_object_type(mdd_obj)) &&
1409 dt_try_as_dir(env, dir))) {
1411 rc = dir->do_index_ops->dio_lookup(env, dir,
1412 (struct dt_rec *)pack, key,
1413 mdd_object_capa(env, mdd_obj));
1415 rc = fid_unpack(pack, fid);
1424 int mdd_object_initialize(const struct lu_env *env, const struct lu_fid *pfid,
1425 const struct lu_name *lname, struct mdd_object *child,
1426 struct md_attr *ma, struct thandle *handle,
1427 const struct md_op_spec *spec)
1433 * Update attributes for child.
1436 * (1) the valid bits should be converted between Lustre and Linux;
1437 * (2) maybe, the child attributes should be set in OSD when creation.
1440 rc = mdd_attr_set_internal(env, child, &ma->ma_attr, handle, 0);
1444 if (S_ISDIR(ma->ma_attr.la_mode)) {
1445 /* Add "." and ".." for newly created dir */
1446 __mdd_ref_add(env, child, handle);
1447 rc = __mdd_index_insert_only(env, child, mdo2fid(child),
1448 dot, handle, BYPASS_CAPA);
1450 rc = __mdd_index_insert_only(env, child, pfid,
1456 rc2 = __mdd_index_delete(env, child, dot, 1,
1457 handle, BYPASS_CAPA);
1459 CERROR("Failure to cleanup after dotdot"
1460 " creation: %d (%d)\n", rc2, rc);
1465 mdd_links_add(env, child, pfid, lname, handle);
1470 /* has not lock on pobj yet */
1471 static int mdd_create_sanity_check(const struct lu_env *env,
1472 struct md_object *pobj,
1473 const struct lu_name *lname,
1475 struct md_op_spec *spec)
1477 struct mdd_thread_info *info = mdd_env_info(env);
1478 struct lu_attr *la = &info->mti_la;
1479 struct lu_fid *fid = &info->mti_fid;
1480 struct mdd_object *obj = md2mdd_obj(pobj);
1481 struct mdd_device *m = mdo2mdd(pobj);
1482 int lookup = spec->sp_cr_lookup;
1487 if (mdd_is_dead_obj(obj))
1491 * In some cases this lookup is not needed - we know before if name
1492 * exists or not because MDT performs lookup for it.
1493 * name length check is done in lookup.
1497 * Check if the name already exist, though it will be checked in
1498 * _index_insert also, for avoiding rolling back if exists
1501 rc = __mdd_lookup_locked(env, pobj, lname, fid,
1502 MAY_WRITE | MAY_EXEC);
1504 RETURN(rc ? : -EEXIST);
1507 * Check WRITE permission for the parent.
1508 * EXEC permission have been checked
1509 * when lookup before create already.
1511 rc = mdd_permission_internal_locked(env, obj, NULL, MAY_WRITE,
1518 rc = mdd_la_get(env, obj, la, BYPASS_CAPA);
1522 if (la->la_mode & S_ISGID) {
1523 ma->ma_attr.la_gid = la->la_gid;
1524 if (S_ISDIR(ma->ma_attr.la_mode)) {
1525 ma->ma_attr.la_mode |= S_ISGID;
1526 ma->ma_attr.la_valid |= LA_MODE;
1530 switch (ma->ma_attr.la_mode & S_IFMT) {
1532 unsigned int symlen = strlen(spec->u.sp_symname) + 1;
1534 if (symlen > (1 << m->mdd_dt_conf.ddp_block_shift))
1535 RETURN(-ENAMETOOLONG);
1555 * Create object and insert it into namespace.
1557 static int mdd_create(const struct lu_env *env,
1558 struct md_object *pobj,
1559 const struct lu_name *lname,
1560 struct md_object *child,
1561 struct md_op_spec *spec,
1564 struct mdd_thread_info *info = mdd_env_info(env);
1565 struct lu_attr *la = &info->mti_la_for_fix;
1566 struct md_attr *ma_acl = &info->mti_ma;
1567 struct mdd_object *mdd_pobj = md2mdd_obj(pobj);
1568 struct mdd_object *son = md2mdd_obj(child);
1569 struct mdd_device *mdd = mdo2mdd(pobj);
1570 struct lu_attr *attr = &ma->ma_attr;
1571 struct lov_mds_md *lmm = NULL;
1572 struct thandle *handle;
1573 struct dynlock_handle *dlh;
1574 const char *name = lname->ln_name;
1575 int rc, created = 0, initialized = 0, inserted = 0, lmm_size = 0;
1576 int got_def_acl = 0;
1577 #ifdef HAVE_QUOTA_SUPPORT
1578 struct obd_device *obd = mdd->mdd_obd_dev;
1579 struct mds_obd *mds = &obd->u.mds;
1580 unsigned int qcids[MAXQUOTAS] = { 0, 0 };
1581 unsigned int qpids[MAXQUOTAS] = { 0, 0 };
1582 int quota_opc = 0, block_count = 0;
1583 int inode_pending = 0, block_pending = 0, parent_pending = 0;
1588 * Two operations have to be performed:
1590 * - an allocation of a new object (->do_create()), and
1592 * - an insertion into a parent index (->dio_insert()).
1594 * Due to locking, operation order is not important, when both are
1595 * successful, *but* error handling cases are quite different:
1597 * - if insertion is done first, and following object creation fails,
1598 * insertion has to be rolled back, but this operation might fail
1599 * also leaving us with dangling index entry.
1601 * - if creation is done first, is has to be undone if insertion
1602 * fails, leaving us with leaked space, which is neither good, nor
1605 * It seems that creation-first is simplest solution, but it is
1606 * sub-optimal in the frequent
1611 * case, because second mkdir is bound to create object, only to
1612 * destroy it immediately.
1614 * To avoid this follow local file systems that do double lookup:
1616 * 0. lookup -> -EEXIST (mdd_create_sanity_check())
1618 * 1. create (mdd_object_create_internal())
1620 * 2. insert (__mdd_index_insert(), lookup again)
1623 /* Sanity checks before big job. */
1624 rc = mdd_create_sanity_check(env, pobj, lname, ma, spec);
1628 #ifdef HAVE_QUOTA_SUPPORT
1629 if (mds->mds_quota) {
1630 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
1632 rc = mdd_la_get(env, mdd_pobj, la_tmp, BYPASS_CAPA);
1636 quota_opc = FSFILT_OP_CREATE;
1637 mdd_quota_wrapper(&ma->ma_attr, qcids);
1638 mdd_quota_wrapper(la_tmp, qpids);
1639 /* get file quota for child */
1640 lquota_chkquota(mds_quota_interface_ref, obd,
1641 qcids[USRQUOTA], qcids[GRPQUOTA], 1,
1642 &inode_pending, NULL, 0, NULL, 0);
1643 switch (ma->ma_attr.la_mode & S_IFMT) {
1652 if (qcids[USRQUOTA] == qpids[USRQUOTA] &&
1653 qcids[GRPQUOTA] == qpids[GRPQUOTA]) {
1657 /* get block quota for child and parent */
1659 lquota_chkquota(mds_quota_interface_ref, obd,
1660 qcids[USRQUOTA], qcids[GRPQUOTA],
1662 &block_pending, NULL,
1663 LQUOTA_FLAGS_BLK, NULL, 0);
1665 lquota_chkquota(mds_quota_interface_ref, obd,
1666 qpids[USRQUOTA], qpids[GRPQUOTA], 1,
1667 &parent_pending, NULL,
1668 LQUOTA_FLAGS_BLK, NULL, 0);
1674 * No RPC inside the transaction, so OST objects should be created at
1677 if (S_ISREG(attr->la_mode)) {
1678 rc = mdd_lov_create(env, mdd, mdd_pobj, son, &lmm, &lmm_size,
1681 GOTO(out_pending, rc);
1684 if (!S_ISLNK(attr->la_mode)) {
1685 ma_acl->ma_acl_size = sizeof info->mti_xattr_buf;
1686 ma_acl->ma_acl = info->mti_xattr_buf;
1687 ma_acl->ma_need = MA_ACL_DEF;
1688 ma_acl->ma_valid = 0;
1690 mdd_read_lock(env, mdd_pobj, MOR_TGT_PARENT);
1691 rc = mdd_def_acl_get(env, mdd_pobj, ma_acl);
1692 mdd_read_unlock(env, mdd_pobj);
1695 else if (ma_acl->ma_valid & MA_ACL_DEF)
1699 mdd_txn_param_build(env, mdd, MDD_TXN_MKDIR_OP);
1700 handle = mdd_trans_start(env, mdd);
1702 GOTO(out_free, rc = PTR_ERR(handle));
1704 dlh = mdd_pdo_write_lock(env, mdd_pobj, name, MOR_TGT_PARENT);
1706 GOTO(out_trans, rc = -ENOMEM);
1708 mdd_write_lock(env, son, MOR_TGT_CHILD);
1709 rc = mdd_object_create_internal(env, mdd_pobj, son, ma, handle, spec);
1711 mdd_write_unlock(env, son);
1717 #ifdef CONFIG_FS_POSIX_ACL
1719 struct lu_buf *acl_buf = &info->mti_buf;
1720 acl_buf->lb_buf = ma_acl->ma_acl;
1721 acl_buf->lb_len = ma_acl->ma_acl_size;
1723 rc = __mdd_acl_init(env, son, acl_buf, &attr->la_mode, handle);
1725 mdd_write_unlock(env, son);
1728 ma->ma_attr.la_valid |= LA_MODE;
1733 rc = mdd_object_initialize(env, mdo2fid(mdd_pobj), lname,
1734 son, ma, handle, spec);
1735 mdd_write_unlock(env, son);
1738 * Object has no links, so it will be destroyed when last
1739 * reference is released. (XXX not now.)
1745 rc = __mdd_index_insert(env, mdd_pobj, mdo2fid(son),
1746 name, S_ISDIR(attr->la_mode), handle,
1747 mdd_object_capa(env, mdd_pobj));
1754 /* No need mdd_lsm_sanity_check here */
1755 rc = mdd_lov_set_md(env, mdd_pobj, son, lmm, lmm_size, handle, 0);
1757 CERROR("error on stripe info copy %d \n", rc);
1760 if (lmm && lmm_size > 0) {
1761 /* Set Lov here, do not get lmm again later */
1762 memcpy(ma->ma_lmm, lmm, lmm_size);
1763 ma->ma_lmm_size = lmm_size;
1764 ma->ma_valid |= MA_LOV;
1767 if (S_ISLNK(attr->la_mode)) {
1768 struct md_ucred *uc = md_ucred(env);
1769 struct dt_object *dt = mdd_object_child(son);
1770 const char *target_name = spec->u.sp_symname;
1771 int sym_len = strlen(target_name);
1772 const struct lu_buf *buf;
1775 buf = mdd_buf_get_const(env, target_name, sym_len);
1776 rc = dt->do_body_ops->dbo_write(env, dt, buf, &pos, handle,
1777 mdd_object_capa(env, son),
1779 CFS_CAP_SYS_RESOURCE_MASK);
1784 GOTO(cleanup, rc = -EFAULT);
1788 la->la_valid = LA_CTIME | LA_MTIME;
1789 rc = mdd_attr_check_set_internal_locked(env, mdd_pobj, la, handle, 0);
1793 /* Return attr back. */
1794 rc = mdd_attr_get_internal_locked(env, son, ma);
1797 if (rc && created) {
1801 rc2 = __mdd_index_delete(env, mdd_pobj, name,
1802 S_ISDIR(attr->la_mode),
1803 handle, BYPASS_CAPA);
1805 CERROR("error can not cleanup destroy %d\n",
1810 mdd_write_lock(env, son, MOR_TGT_CHILD);
1811 __mdd_ref_del(env, son, handle, 0);
1812 if (initialized && S_ISDIR(attr->la_mode))
1813 __mdd_ref_del(env, son, handle, 1);
1814 mdd_write_unlock(env, son);
1818 /* update lov_objid data, must be before transaction stop! */
1820 mdd_lov_objid_update(mdd, lmm);
1822 mdd_pdo_write_unlock(env, mdd_pobj, dlh);
1825 rc = mdd_changelog_ns_store(env, mdd,
1826 S_ISDIR(attr->la_mode) ? CL_MKDIR :
1827 S_ISREG(attr->la_mode) ? CL_CREATE :
1828 S_ISLNK(attr->la_mode) ? CL_SOFTLINK : CL_MKNOD,
1829 son, mdd_pobj, NULL, lname, handle);
1830 mdd_trans_stop(env, mdd, rc, handle);
1832 /* finis lov_create stuff, free all temporary data */
1833 mdd_lov_create_finish(env, mdd, lmm, lmm_size, spec);
1835 #ifdef HAVE_QUOTA_SUPPORT
1838 lquota_pending_commit(mds_quota_interface_ref, obd,
1839 qcids[USRQUOTA], qcids[GRPQUOTA],
1842 lquota_pending_commit(mds_quota_interface_ref, obd,
1843 qcids[USRQUOTA], qcids[GRPQUOTA],
1846 lquota_pending_commit(mds_quota_interface_ref, obd,
1847 qpids[USRQUOTA], qpids[GRPQUOTA],
1849 /* Trigger dqacq on the owner of child and parent. If failed,
1850 * the next call for lquota_chkquota will process it. */
1851 lquota_adjust(mds_quota_interface_ref, obd, qcids, qpids, rc,
1859 * Get locks on parents in proper order
1860 * RETURN: < 0 - error, rename_order if successful
1868 static int mdd_rename_order(const struct lu_env *env,
1869 struct mdd_device *mdd,
1870 struct mdd_object *src_pobj,
1871 struct mdd_object *tgt_pobj)
1873 /* order of locking, 1 - tgt-src, 0 - src-tgt*/
1877 if (src_pobj == tgt_pobj)
1878 RETURN(MDD_RN_SAME);
1880 /* compared the parent child relationship of src_p&tgt_p */
1881 if (lu_fid_eq(&mdd->mdd_root_fid, mdo2fid(src_pobj))){
1883 } else if (lu_fid_eq(&mdd->mdd_root_fid, mdo2fid(tgt_pobj))) {
1886 rc = mdd_is_parent(env, mdd, src_pobj, mdo2fid(tgt_pobj), NULL);
1899 /* has not mdd_write{read}_lock on any obj yet. */
1900 static int mdd_rename_sanity_check(const struct lu_env *env,
1901 struct mdd_object *src_pobj,
1902 struct mdd_object *tgt_pobj,
1903 struct mdd_object *sobj,
1904 struct mdd_object *tobj,
1910 if (unlikely(ma->ma_attr_flags & MDS_PERM_BYPASS))
1913 /* XXX: when get here, sobj must NOT be NULL,
1914 * the other case has been processed in cml_rename
1915 * before mdd_rename and enable MDS_PERM_BYPASS. */
1917 rc = mdd_may_delete(env, src_pobj, sobj, ma, 1, 0);
1921 /* XXX: when get here, "tobj == NULL" means tobj must
1922 * NOT exist (neither on remote MDS, such case has been
1923 * processed in cml_rename before mdd_rename and enable
1925 * So check may_create, but not check may_unlink. */
1927 rc = mdd_may_create(env, tgt_pobj, NULL,
1928 (src_pobj != tgt_pobj), 0);
1930 rc = mdd_may_delete(env, tgt_pobj, tobj, ma,
1931 (src_pobj != tgt_pobj), 1);
1933 if (!rc && !tobj && (src_pobj != tgt_pobj) &&
1934 S_ISDIR(ma->ma_attr.la_mode))
1935 rc = __mdd_may_link(env, tgt_pobj);
1940 /* src object can be remote that is why we use only fid and type of object */
1941 static int mdd_rename(const struct lu_env *env,
1942 struct md_object *src_pobj, struct md_object *tgt_pobj,
1943 const struct lu_fid *lf, const struct lu_name *lsname,
1944 struct md_object *tobj, const struct lu_name *ltname,
1947 const char *sname = lsname->ln_name;
1948 const char *tname = ltname->ln_name;
1949 struct lu_attr *la = &mdd_env_info(env)->mti_la_for_fix;
1950 struct mdd_object *mdd_spobj = md2mdd_obj(src_pobj); /* source parent */
1951 struct mdd_object *mdd_tpobj = md2mdd_obj(tgt_pobj);
1952 struct mdd_device *mdd = mdo2mdd(src_pobj);
1953 struct mdd_object *mdd_sobj = NULL; /* source object */
1954 struct mdd_object *mdd_tobj = NULL;
1955 struct dynlock_handle *sdlh, *tdlh;
1956 struct thandle *handle;
1957 const struct lu_fid *tpobj_fid = mdo2fid(mdd_tpobj);
1961 #ifdef HAVE_QUOTA_SUPPORT
1962 struct obd_device *obd = mdd->mdd_obd_dev;
1963 struct mds_obd *mds = &obd->u.mds;
1964 unsigned int qspids[MAXQUOTAS] = { 0, 0 };
1965 unsigned int qtcids[MAXQUOTAS] = { 0, 0 };
1966 unsigned int qtpids[MAXQUOTAS] = { 0, 0 };
1967 int quota_opc = 0, rec_pending = 0;
1971 LASSERT(ma->ma_attr.la_mode & S_IFMT);
1972 is_dir = S_ISDIR(ma->ma_attr.la_mode);
1975 mdd_tobj = md2mdd_obj(tobj);
1977 #ifdef HAVE_QUOTA_SUPPORT
1978 if (mds->mds_quota) {
1979 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
1981 rc = mdd_la_get(env, mdd_spobj, la_tmp, BYPASS_CAPA);
1983 mdd_quota_wrapper(la_tmp, qspids);
1985 rc = mdd_la_get(env, mdd_tpobj, la_tmp,
1988 quota_opc = FSFILT_OP_LINK;
1989 mdd_quota_wrapper(la_tmp, qtpids);
1990 /* get block quota for target parent */
1991 lquota_chkquota(mds_quota_interface_ref,
1992 obd, qtpids[USRQUOTA],
1993 qtpids[GRPQUOTA], 1,
2002 mdd_txn_param_build(env, mdd, MDD_TXN_RENAME_OP);
2003 handle = mdd_trans_start(env, mdd);
2005 GOTO(out_pending, rc = PTR_ERR(handle));
2007 /* FIXME: Should consider tobj and sobj too in rename_lock. */
2008 rc = mdd_rename_order(env, mdd, mdd_spobj, mdd_tpobj);
2010 GOTO(cleanup_unlocked, rc);
2012 /* Get locks in determined order */
2013 if (rc == MDD_RN_SAME) {
2014 sdlh = mdd_pdo_write_lock(env, mdd_spobj,
2015 sname, MOR_SRC_PARENT);
2016 /* check hashes to determine do we need one lock or two */
2017 if (mdd_name2hash(sname) != mdd_name2hash(tname))
2018 tdlh = mdd_pdo_write_lock(env, mdd_tpobj, tname,
2022 } else if (rc == MDD_RN_SRCTGT) {
2023 sdlh = mdd_pdo_write_lock(env, mdd_spobj, sname,MOR_SRC_PARENT);
2024 tdlh = mdd_pdo_write_lock(env, mdd_tpobj, tname,MOR_TGT_PARENT);
2026 tdlh = mdd_pdo_write_lock(env, mdd_tpobj, tname,MOR_SRC_PARENT);
2027 sdlh = mdd_pdo_write_lock(env, mdd_spobj, sname,MOR_TGT_PARENT);
2029 if (sdlh == NULL || tdlh == NULL)
2030 GOTO(cleanup, rc = -ENOMEM);
2032 mdd_sobj = mdd_object_find(env, mdd, lf);
2033 rc = mdd_rename_sanity_check(env, mdd_spobj, mdd_tpobj,
2034 mdd_sobj, mdd_tobj, ma);
2038 /* Remove source name from source directory */
2039 rc = __mdd_index_delete(env, mdd_spobj, sname, is_dir, handle,
2040 mdd_object_capa(env, mdd_spobj));
2044 /* "mv dir1 dir2" needs "dir1/.." link update */
2045 if (is_dir && mdd_sobj) {
2046 rc = __mdd_index_delete(env, mdd_sobj, dotdot, is_dir, handle,
2047 mdd_object_capa(env, mdd_spobj));
2051 rc = __mdd_index_insert(env, mdd_sobj, tpobj_fid, dotdot,
2053 mdd_object_capa(env, mdd_tpobj));
2058 /* Remove target name from target directory
2059 * Here tobj can be remote one, so we do index_delete unconditionally
2060 * and -ENOENT is allowed.
2062 rc = __mdd_index_delete(env, mdd_tpobj, tname, is_dir, handle,
2063 mdd_object_capa(env, mdd_tpobj));
2064 if (rc != 0 && rc != -ENOENT)
2067 /* Insert new fid with target name into target dir */
2068 rc = __mdd_index_insert(env, mdd_tpobj, lf, tname, is_dir, handle,
2069 mdd_object_capa(env, mdd_tpobj));
2073 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
2074 la->la_ctime = la->la_mtime = ma->ma_attr.la_ctime;
2076 /* XXX: mdd_sobj must be local one if it is NOT NULL. */
2078 la->la_valid = LA_CTIME;
2079 rc = mdd_attr_check_set_internal_locked(env, mdd_sobj, la,
2085 /* Remove old target object
2086 * For tobj is remote case cmm layer has processed
2087 * and set tobj to NULL then. So when tobj is NOT NULL,
2088 * it must be local one.
2090 if (tobj && mdd_object_exists(mdd_tobj)) {
2091 mdd_write_lock(env, mdd_tobj, MOR_TGT_CHILD);
2092 __mdd_ref_del(env, mdd_tobj, handle, 0);
2094 /* Remove dot reference. */
2096 __mdd_ref_del(env, mdd_tobj, handle, 1);
2098 la->la_valid = LA_CTIME;
2099 rc = mdd_attr_check_set_internal(env, mdd_tobj, la, handle, 0);
2103 rc = mdd_finish_unlink(env, mdd_tobj, ma, handle);
2104 mdd_write_unlock(env, mdd_tobj);
2108 #ifdef HAVE_QUOTA_SUPPORT
2109 if (mds->mds_quota && ma->ma_valid & MA_INODE &&
2110 ma->ma_attr.la_nlink == 0 && mdd_tobj->mod_count == 0) {
2111 quota_opc = FSFILT_OP_UNLINK_PARTIAL_CHILD;
2112 mdd_quota_wrapper(&ma->ma_attr, qtcids);
2117 la->la_valid = LA_CTIME | LA_MTIME;
2118 rc = mdd_attr_check_set_internal_locked(env, mdd_spobj, la, handle, 0);
2122 if (mdd_spobj != mdd_tpobj) {
2123 la->la_valid = LA_CTIME | LA_MTIME;
2124 rc = mdd_attr_check_set_internal_locked(env, mdd_tpobj, la,
2128 if (rc == 0 && mdd_sobj) {
2129 mdd_write_lock(env, mdd_sobj, MOR_SRC_CHILD);
2130 rc = mdd_links_rename(env, mdd_sobj, mdo2fid(mdd_spobj), lsname,
2131 mdo2fid(mdd_tpobj), ltname, handle);
2133 /* Old files might not have EA entry */
2134 mdd_links_add(env, mdd_sobj, mdo2fid(mdd_spobj),
2136 mdd_write_unlock(env, mdd_sobj);
2137 /* We don't fail the transaction if the link ea can't be
2138 updated -- fid2path will use alternate lookup method. */
2144 if (likely(tdlh) && sdlh != tdlh)
2145 mdd_pdo_write_unlock(env, mdd_tpobj, tdlh);
2147 mdd_pdo_write_unlock(env, mdd_spobj, sdlh);
2150 rc = mdd_changelog_ns_store(env, mdd, CL_RENAME, mdd_tobj,
2151 mdd_spobj, lf, lsname, handle);
2153 rc = mdd_changelog_ns_store(env, mdd, CL_EXT, mdd_tobj,
2154 mdd_tpobj, lf, ltname, handle);
2156 mdd_trans_stop(env, mdd, rc, handle);
2158 mdd_object_put(env, mdd_sobj);
2160 #ifdef HAVE_QUOTA_SUPPORT
2161 if (mds->mds_quota) {
2163 lquota_pending_commit(mds_quota_interface_ref, obd,
2167 /* Trigger dqrel on the source owner of parent.
2168 * If failed, the next call for lquota_chkquota will
2170 lquota_adjust(mds_quota_interface_ref, obd, 0, qspids, rc,
2171 FSFILT_OP_UNLINK_PARTIAL_PARENT);
2173 /* Trigger dqrel/dqacq on the target owner of child and
2174 * parent. If failed, the next call for lquota_chkquota
2175 * will process it. */
2176 lquota_adjust(mds_quota_interface_ref, obd, qtcids,
2177 qtpids, rc, quota_opc);
2183 /** enable/disable storing of hardlink info */
2184 int mdd_linkea_enable = 1;
2185 CFS_MODULE_PARM(mdd_linkea_enable, "d", int, 0644,
2186 "record hardlink info in EAs");
2188 /** Read the link EA into a temp buffer.
2189 * Uses the name_buf since it is generally large.
2190 * \retval IS_ERR err
2191 * \retval ptr to \a lu_buf (always \a mti_big_buf)
2193 struct lu_buf *mdd_links_get(const struct lu_env *env,
2194 struct mdd_object *mdd_obj)
2197 struct lustre_capa *capa;
2198 struct link_ea_header *leh;
2201 /* First try a small buf */
2202 buf = mdd_buf_alloc(env, CFS_PAGE_SIZE);
2203 if (buf->lb_buf == NULL)
2204 return ERR_PTR(-ENOMEM);
2206 capa = mdd_object_capa(env, mdd_obj);
2207 rc = mdo_xattr_get(env, mdd_obj, buf, XATTR_NAME_LINK, capa);
2208 if (rc == -ERANGE) {
2209 /* Buf was too small, figure out what we need. */
2212 rc = mdo_xattr_get(env, mdd_obj, buf, XATTR_NAME_LINK, capa);
2215 buf = mdd_buf_alloc(env, rc);
2216 if (buf->lb_buf == NULL)
2217 return ERR_PTR(-ENOMEM);
2218 rc = mdo_xattr_get(env, mdd_obj, buf, XATTR_NAME_LINK, capa);
2224 if (leh->leh_magic == __swab32(LINK_EA_MAGIC)) {
2225 leh->leh_magic = LINK_EA_MAGIC;
2226 leh->leh_reccount = __swab32(leh->leh_reccount);
2227 leh->leh_len = __swab64(leh->leh_len);
2228 /* entries are swabbed by mdd_lee_unpack */
2230 if (leh->leh_magic != LINK_EA_MAGIC)
2231 return ERR_PTR(-EINVAL);
2236 /** Pack a link_ea_entry.
2237 * All elements are stored as chars to avoid alignment issues.
2238 * Numbers are always big-endian
2239 * \param packbuf is a temp fid buffer
2240 * \retval record length
2242 static int mdd_lee_pack(struct link_ea_entry *lee, const struct lu_name *lname,
2243 const struct lu_fid *pfid, struct lu_fid* packbuf)
2248 fid_pack(&lee->lee_parent_fid, pfid, packbuf);
2249 ptr = (char *)&lee->lee_parent_fid + lee->lee_parent_fid.fp_len;
2250 strncpy(ptr, lname->ln_name, lname->ln_namelen);
2251 reclen = lee->lee_parent_fid.fp_len + lname->ln_namelen +
2252 sizeof(lee->lee_reclen);
2253 lee->lee_reclen[0] = (reclen >> 8) & 0xff;
2254 lee->lee_reclen[1] = reclen & 0xff;
2258 void mdd_lee_unpack(const struct link_ea_entry *lee, int *reclen,
2259 struct lu_name *lname, struct lu_fid *pfid)
2261 *reclen = (lee->lee_reclen[0] << 8) | lee->lee_reclen[1];
2262 fid_unpack(&lee->lee_parent_fid, pfid);
2263 lname->ln_name = (char *)&lee->lee_parent_fid +
2264 lee->lee_parent_fid.fp_len;
2265 lname->ln_namelen = *reclen - lee->lee_parent_fid.fp_len -
2266 sizeof(lee->lee_reclen);
2269 /** Add a record to the end of link ea buf */
2270 static int __mdd_links_add(const struct lu_env *env, struct lu_buf *buf,
2271 const struct lu_fid *pfid,
2272 const struct lu_name *lname)
2274 struct link_ea_header *leh;
2275 struct link_ea_entry *lee;
2278 if (lname == NULL || pfid == NULL)
2281 /* Make sure our buf is big enough for the new one */
2283 reclen = lname->ln_namelen + sizeof(struct link_ea_entry);
2284 if (leh->leh_len + reclen > buf->lb_len) {
2285 mdd_buf_grow(env, leh->leh_len + reclen);
2286 if (buf->lb_buf == NULL)
2291 lee = buf->lb_buf + leh->leh_len;
2292 reclen = mdd_lee_pack(lee, lname, pfid, &mdd_env_info(env)->mti_fid2);
2293 leh->leh_len += reclen;
2294 leh->leh_reccount++;
2298 /* For pathologic linkers, we don't want to spend lots of time scanning the
2299 * link ea. Limit ourseleves to something reasonable; links not in the EA
2300 * can be looked up via (slower) parent lookup.
2302 #define LINKEA_MAX_COUNT 128
2304 static int mdd_links_add(const struct lu_env *env,
2305 struct mdd_object *mdd_obj,
2306 const struct lu_fid *pfid,
2307 const struct lu_name *lname,
2308 struct thandle *handle)
2311 struct link_ea_header *leh;
2315 if (!mdd_linkea_enable)
2318 buf = mdd_links_get(env, mdd_obj);
2321 if (rc != -ENODATA) {
2322 CERROR("link_ea read failed %d "DFID"\n", rc,
2323 PFID(mdd_object_fid(mdd_obj)));
2326 /* empty EA; start one */
2327 buf = mdd_buf_alloc(env, CFS_PAGE_SIZE);
2328 if (buf->lb_buf == NULL)
2331 leh->leh_magic = LINK_EA_MAGIC;
2332 leh->leh_len = sizeof(struct link_ea_header);
2333 leh->leh_reccount = 0;
2337 if (leh->leh_reccount > LINKEA_MAX_COUNT)
2340 rc = __mdd_links_add(env, buf, pfid, lname);
2345 rc = __mdd_xattr_set(env, mdd_obj,
2346 mdd_buf_get_const(env, buf->lb_buf, leh->leh_len),
2347 XATTR_NAME_LINK, 0, handle);
2349 CERROR("link_ea add failed %d "DFID"\n", rc,
2350 PFID(mdd_object_fid(mdd_obj)));
2352 if (buf->lb_vmalloc)
2353 /* if we vmalloced a large buffer drop it */
2359 static int mdd_links_rename(const struct lu_env *env,
2360 struct mdd_object *mdd_obj,
2361 const struct lu_fid *oldpfid,
2362 const struct lu_name *oldlname,
2363 const struct lu_fid *newpfid,
2364 const struct lu_name *newlname,
2365 struct thandle *handle)
2368 struct link_ea_header *leh;
2369 struct link_ea_entry *lee;
2370 struct lu_name *tmpname = &mdd_env_info(env)->mti_name;
2371 struct lu_fid *tmpfid = &mdd_env_info(env)->mti_fid;
2376 if (!mdd_linkea_enable)
2379 if (mdd_obj->mod_flags & DEAD_OBJ)
2380 /* No more links, don't bother */
2383 buf = mdd_links_get(env, mdd_obj);
2386 CERROR("link_ea read failed %d "DFID"\n",
2387 rc, PFID(mdd_object_fid(mdd_obj)));
2391 lee = (struct link_ea_entry *)(leh + 1); /* link #0 */
2393 /* Find the old record */
2394 for(count = 0; count <= leh->leh_reccount; count++) {
2395 mdd_lee_unpack(lee, &reclen, tmpname, tmpfid);
2396 if (tmpname->ln_namelen == oldlname->ln_namelen &&
2397 lu_fid_eq(tmpfid, oldpfid) &&
2398 (strncmp(tmpname->ln_name, oldlname->ln_name,
2399 tmpname->ln_namelen) == 0))
2401 lee = (struct link_ea_entry *)((char *)lee + reclen);
2403 if (count > leh->leh_reccount) {
2404 CDEBUG(D_INODE, "Old link_ea name '%.*s' not found\n",
2405 oldlname->ln_namelen, oldlname->ln_name);
2406 GOTO(out, rc = -ENOENT);
2409 /* Remove the old record */
2410 leh->leh_reccount--;
2411 leh->leh_len -= reclen;
2412 memmove(lee, (char *)lee + reclen, (char *)leh + leh->leh_len -
2415 /* If renaming, add the new record */
2416 if (newpfid != NULL) {
2417 rc = __mdd_links_add(env, buf, newpfid, newlname);
2423 rc = __mdd_xattr_set(env, mdd_obj,
2424 mdd_buf_get_const(env, buf->lb_buf, leh->leh_len),
2425 XATTR_NAME_LINK, 0, handle);
2429 CDEBUG(D_INODE, "link_ea mv/unlink '%.*s' failed %d "DFID"\n",
2430 oldlname->ln_namelen, oldlname->ln_name, rc,
2431 PFID(mdd_object_fid(mdd_obj)));
2433 if (buf->lb_vmalloc)
2434 /* if we vmalloced a large buffer drop it */
2440 const struct md_dir_operations mdd_dir_ops = {
2441 .mdo_is_subdir = mdd_is_subdir,
2442 .mdo_lookup = mdd_lookup,
2443 .mdo_create = mdd_create,
2444 .mdo_rename = mdd_rename,
2445 .mdo_link = mdd_link,
2446 .mdo_unlink = mdd_unlink,
2447 .mdo_name_insert = mdd_name_insert,
2448 .mdo_name_remove = mdd_name_remove,
2449 .mdo_rename_tgt = mdd_rename_tgt,
2450 .mdo_create_data = mdd_create_data