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 (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
30 * Use is subject to license terms.
33 * Copyright (c) 2011 Whamcloud, Inc.
36 * This file is part of Lustre, http://www.lustre.org/
37 * Lustre is a trademark of Sun Microsystems, Inc.
39 * lustre/mdd/mdd_object.c
41 * Lustre Metadata Server (mdd) routines
43 * Author: Wang Di <wangdi@clusterfs.com>
47 # define EXPORT_SYMTAB
49 #define DEBUG_SUBSYSTEM S_MDS
51 #include <linux/module.h>
53 #include <obd_class.h>
54 #include <obd_support.h>
55 #include <lprocfs_status.h>
56 /* fid_be_cpu(), fid_cpu_to_be(). */
57 #include <lustre_fid.h>
60 #include <lustre_param.h>
61 #include <lustre_mds.h>
62 #include <lustre/lustre_idl.h>
64 #include "mdd_internal.h"
66 static const struct lu_object_operations mdd_lu_obj_ops;
68 static int mdd_xattr_get(const struct lu_env *env,
69 struct md_object *obj, struct lu_buf *buf,
72 int mdd_data_get(const struct lu_env *env, struct mdd_object *obj,
75 LASSERTF(mdd_object_exists(obj), "FID is "DFID"\n",
76 PFID(mdd_object_fid(obj)));
77 mdo_data_get(env, obj, data);
81 int mdd_la_get(const struct lu_env *env, struct mdd_object *obj,
82 struct lu_attr *la, struct lustre_capa *capa)
84 LASSERTF(mdd_object_exists(obj), "FID is "DFID"\n",
85 PFID(mdd_object_fid(obj)));
86 return mdo_attr_get(env, obj, la, capa);
89 static void mdd_flags_xlate(struct mdd_object *obj, __u32 flags)
91 obj->mod_flags &= ~(APPEND_OBJ|IMMUTE_OBJ);
93 if (flags & LUSTRE_APPEND_FL)
94 obj->mod_flags |= APPEND_OBJ;
96 if (flags & LUSTRE_IMMUTABLE_FL)
97 obj->mod_flags |= IMMUTE_OBJ;
100 struct mdd_thread_info *mdd_env_info(const struct lu_env *env)
102 struct mdd_thread_info *info;
104 info = lu_context_key_get(&env->le_ctx, &mdd_thread_key);
105 LASSERT(info != NULL);
109 struct lu_buf *mdd_buf_get(const struct lu_env *env, void *area, ssize_t len)
113 buf = &mdd_env_info(env)->mti_buf;
119 void mdd_buf_put(struct lu_buf *buf)
121 if (buf == NULL || buf->lb_buf == NULL)
123 OBD_FREE_LARGE(buf->lb_buf, buf->lb_len);
128 const struct lu_buf *mdd_buf_get_const(const struct lu_env *env,
129 const void *area, ssize_t len)
133 buf = &mdd_env_info(env)->mti_buf;
134 buf->lb_buf = (void *)area;
139 struct lu_buf *mdd_buf_alloc(const struct lu_env *env, ssize_t len)
141 struct lu_buf *buf = &mdd_env_info(env)->mti_big_buf;
143 if ((len > buf->lb_len) && (buf->lb_buf != NULL)) {
144 OBD_FREE_LARGE(buf->lb_buf, buf->lb_len);
147 if (buf->lb_buf == NULL) {
149 OBD_ALLOC_LARGE(buf->lb_buf, buf->lb_len);
150 if (buf->lb_buf == NULL)
156 /** Increase the size of the \a mti_big_buf.
157 * preserves old data in buffer
158 * old buffer remains unchanged on error
159 * \retval 0 or -ENOMEM
161 int mdd_buf_grow(const struct lu_env *env, ssize_t len)
163 struct lu_buf *oldbuf = &mdd_env_info(env)->mti_big_buf;
166 LASSERT(len >= oldbuf->lb_len);
167 OBD_ALLOC_LARGE(buf.lb_buf, len);
169 if (buf.lb_buf == NULL)
173 memcpy(buf.lb_buf, oldbuf->lb_buf, oldbuf->lb_len);
175 OBD_FREE_LARGE(oldbuf->lb_buf, oldbuf->lb_len);
177 memcpy(oldbuf, &buf, sizeof(buf));
182 struct llog_cookie *mdd_max_cookie_get(const struct lu_env *env,
183 struct mdd_device *mdd)
185 struct mdd_thread_info *mti = mdd_env_info(env);
188 max_cookie_size = mdd_lov_cookiesize(env, mdd);
189 if (unlikely(mti->mti_max_cookie_size < max_cookie_size)) {
190 if (mti->mti_max_cookie)
191 OBD_FREE_LARGE(mti->mti_max_cookie,
192 mti->mti_max_cookie_size);
193 mti->mti_max_cookie = NULL;
194 mti->mti_max_cookie_size = 0;
196 if (unlikely(mti->mti_max_cookie == NULL)) {
197 OBD_ALLOC_LARGE(mti->mti_max_cookie, max_cookie_size);
198 if (likely(mti->mti_max_cookie != NULL))
199 mti->mti_max_cookie_size = max_cookie_size;
201 if (likely(mti->mti_max_cookie != NULL))
202 memset(mti->mti_max_cookie, 0, mti->mti_max_cookie_size);
203 return mti->mti_max_cookie;
206 struct lov_mds_md *mdd_max_lmm_get(const struct lu_env *env,
207 struct mdd_device *mdd)
209 struct mdd_thread_info *mti = mdd_env_info(env);
212 max_lmm_size = mdd_lov_mdsize(env, mdd);
213 if (unlikely(mti->mti_max_lmm_size < max_lmm_size)) {
214 if (mti->mti_max_lmm)
215 OBD_FREE_LARGE(mti->mti_max_lmm, mti->mti_max_lmm_size);
216 mti->mti_max_lmm = NULL;
217 mti->mti_max_lmm_size = 0;
219 if (unlikely(mti->mti_max_lmm == NULL)) {
220 OBD_ALLOC_LARGE(mti->mti_max_lmm, max_lmm_size);
221 if (likely(mti->mti_max_lmm != NULL))
222 mti->mti_max_lmm_size = max_lmm_size;
224 return mti->mti_max_lmm;
227 struct lu_object *mdd_object_alloc(const struct lu_env *env,
228 const struct lu_object_header *hdr,
231 struct mdd_object *mdd_obj;
233 OBD_ALLOC_PTR(mdd_obj);
234 if (mdd_obj != NULL) {
237 o = mdd2lu_obj(mdd_obj);
238 lu_object_init(o, NULL, d);
239 mdd_obj->mod_obj.mo_ops = &mdd_obj_ops;
240 mdd_obj->mod_obj.mo_dir_ops = &mdd_dir_ops;
241 mdd_obj->mod_count = 0;
242 o->lo_ops = &mdd_lu_obj_ops;
249 static int mdd_object_init(const struct lu_env *env, struct lu_object *o,
250 const struct lu_object_conf *unused)
252 struct mdd_device *d = lu2mdd_dev(o->lo_dev);
253 struct mdd_object *mdd_obj = lu2mdd_obj(o);
254 struct lu_object *below;
255 struct lu_device *under;
258 mdd_obj->mod_cltime = 0;
259 under = &d->mdd_child->dd_lu_dev;
260 below = under->ld_ops->ldo_object_alloc(env, o->lo_header, under);
261 mdd_pdlock_init(mdd_obj);
265 lu_object_add(o, below);
270 static int mdd_object_start(const struct lu_env *env, struct lu_object *o)
272 if (lu_object_exists(o))
273 return mdd_get_flags(env, lu2mdd_obj(o));
278 static void mdd_object_free(const struct lu_env *env, struct lu_object *o)
280 struct mdd_object *mdd = lu2mdd_obj(o);
286 static int mdd_object_print(const struct lu_env *env, void *cookie,
287 lu_printer_t p, const struct lu_object *o)
289 struct mdd_object *mdd = lu2mdd_obj((struct lu_object *)o);
290 return (*p)(env, cookie, LUSTRE_MDD_NAME"-object@%p(open_count=%d, "
291 "valid=%x, cltime="LPU64", flags=%lx)",
292 mdd, mdd->mod_count, mdd->mod_valid,
293 mdd->mod_cltime, mdd->mod_flags);
296 static const struct lu_object_operations mdd_lu_obj_ops = {
297 .loo_object_init = mdd_object_init,
298 .loo_object_start = mdd_object_start,
299 .loo_object_free = mdd_object_free,
300 .loo_object_print = mdd_object_print,
303 struct mdd_object *mdd_object_find(const struct lu_env *env,
304 struct mdd_device *d,
305 const struct lu_fid *f)
307 return md2mdd_obj(md_object_find_slice(env, &d->mdd_md_dev, f));
310 static int mdd_path2fid(const struct lu_env *env, struct mdd_device *mdd,
311 const char *path, struct lu_fid *fid)
314 struct lu_fid *f = &mdd_env_info(env)->mti_fid;
315 struct mdd_object *obj;
316 struct lu_name *lname = &mdd_env_info(env)->mti_name;
321 /* temp buffer for path element */
322 buf = mdd_buf_alloc(env, PATH_MAX);
323 if (buf->lb_buf == NULL)
326 lname->ln_name = name = buf->lb_buf;
327 lname->ln_namelen = 0;
328 *f = mdd->mdd_root_fid;
335 while (*path != '/' && *path != '\0') {
343 /* find obj corresponding to fid */
344 obj = mdd_object_find(env, mdd, f);
346 GOTO(out, rc = -EREMOTE);
348 GOTO(out, rc = PTR_ERR(obj));
349 /* get child fid from parent and name */
350 rc = mdd_lookup(env, &obj->mod_obj, lname, f, NULL);
351 mdd_object_put(env, obj);
356 lname->ln_namelen = 0;
365 /** The maximum depth that fid2path() will search.
366 * This is limited only because we want to store the fids for
367 * historical path lookup purposes.
369 #define MAX_PATH_DEPTH 100
371 /** mdd_path() lookup structure. */
372 struct path_lookup_info {
373 __u64 pli_recno; /**< history point */
374 __u64 pli_currec; /**< current record */
375 struct lu_fid pli_fid;
376 struct lu_fid pli_fids[MAX_PATH_DEPTH]; /**< path, in fids */
377 struct mdd_object *pli_mdd_obj;
378 char *pli_path; /**< full path */
380 int pli_linkno; /**< which hardlink to follow */
381 int pli_fidcount; /**< number of \a pli_fids */
384 static int mdd_path_current(const struct lu_env *env,
385 struct path_lookup_info *pli)
387 struct mdd_device *mdd = mdo2mdd(&pli->pli_mdd_obj->mod_obj);
388 struct mdd_object *mdd_obj;
389 struct lu_buf *buf = NULL;
390 struct link_ea_header *leh;
391 struct link_ea_entry *lee;
392 struct lu_name *tmpname = &mdd_env_info(env)->mti_name;
393 struct lu_fid *tmpfid = &mdd_env_info(env)->mti_fid;
399 ptr = pli->pli_path + pli->pli_pathlen - 1;
402 pli->pli_fidcount = 0;
403 pli->pli_fids[0] = *(struct lu_fid *)mdd_object_fid(pli->pli_mdd_obj);
405 while (!mdd_is_root(mdd, &pli->pli_fids[pli->pli_fidcount])) {
406 mdd_obj = mdd_object_find(env, mdd,
407 &pli->pli_fids[pli->pli_fidcount]);
409 GOTO(out, rc = -EREMOTE);
411 GOTO(out, rc = PTR_ERR(mdd_obj));
412 rc = lu_object_exists(&mdd_obj->mod_obj.mo_lu);
414 mdd_object_put(env, mdd_obj);
418 /* Do I need to error out here? */
423 /* Get parent fid and object name */
424 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
425 buf = mdd_links_get(env, mdd_obj);
426 mdd_read_unlock(env, mdd_obj);
427 mdd_object_put(env, mdd_obj);
429 GOTO(out, rc = PTR_ERR(buf));
432 lee = (struct link_ea_entry *)(leh + 1); /* link #0 */
433 mdd_lee_unpack(lee, &reclen, tmpname, tmpfid);
435 /* If set, use link #linkno for path lookup, otherwise use
436 link #0. Only do this for the final path element. */
437 if ((pli->pli_fidcount == 0) &&
438 (pli->pli_linkno < leh->leh_reccount)) {
440 for (count = 0; count < pli->pli_linkno; count++) {
441 lee = (struct link_ea_entry *)
442 ((char *)lee + reclen);
443 mdd_lee_unpack(lee, &reclen, tmpname, tmpfid);
445 if (pli->pli_linkno < leh->leh_reccount - 1)
446 /* indicate to user there are more links */
450 /* Pack the name in the end of the buffer */
451 ptr -= tmpname->ln_namelen;
452 if (ptr - 1 <= pli->pli_path)
453 GOTO(out, rc = -EOVERFLOW);
454 strncpy(ptr, tmpname->ln_name, tmpname->ln_namelen);
457 /* Store the parent fid for historic lookup */
458 if (++pli->pli_fidcount >= MAX_PATH_DEPTH)
459 GOTO(out, rc = -EOVERFLOW);
460 pli->pli_fids[pli->pli_fidcount] = *tmpfid;
463 /* Verify that our path hasn't changed since we started the lookup.
464 Record the current index, and verify the path resolves to the
465 same fid. If it does, then the path is correct as of this index. */
466 cfs_spin_lock(&mdd->mdd_cl.mc_lock);
467 pli->pli_currec = mdd->mdd_cl.mc_index;
468 cfs_spin_unlock(&mdd->mdd_cl.mc_lock);
469 rc = mdd_path2fid(env, mdd, ptr, &pli->pli_fid);
471 CDEBUG(D_INFO, "mdd_path2fid(%s) failed %d\n", ptr, rc);
472 GOTO (out, rc = -EAGAIN);
474 if (!lu_fid_eq(&pli->pli_fids[0], &pli->pli_fid)) {
475 CDEBUG(D_INFO, "mdd_path2fid(%s) found another FID o="DFID
476 " n="DFID"\n", ptr, PFID(&pli->pli_fids[0]),
477 PFID(&pli->pli_fid));
478 GOTO(out, rc = -EAGAIN);
480 ptr++; /* skip leading / */
481 memmove(pli->pli_path, ptr, pli->pli_path + pli->pli_pathlen - ptr);
485 if (buf && !IS_ERR(buf) && buf->lb_len > OBD_ALLOC_BIG)
486 /* if we vmalloced a large buffer drop it */
492 static int mdd_path_historic(const struct lu_env *env,
493 struct path_lookup_info *pli)
498 /* Returns the full path to this fid, as of changelog record recno. */
499 static int mdd_path(const struct lu_env *env, struct md_object *obj,
500 char *path, int pathlen, __u64 *recno, int *linkno)
502 struct path_lookup_info *pli;
510 if (mdd_is_root(mdo2mdd(obj), mdd_object_fid(md2mdd_obj(obj)))) {
519 pli->pli_mdd_obj = md2mdd_obj(obj);
520 pli->pli_recno = *recno;
521 pli->pli_path = path;
522 pli->pli_pathlen = pathlen;
523 pli->pli_linkno = *linkno;
525 /* Retry multiple times in case file is being moved */
526 while (tries-- && rc == -EAGAIN)
527 rc = mdd_path_current(env, pli);
529 /* For historical path lookup, the current links may not have existed
530 * at "recno" time. We must switch over to earlier links/parents
531 * by using the changelog records. If the earlier parent doesn't
532 * exist, we must search back through the changelog to reconstruct
533 * its parents, then check if it exists, etc.
534 * We may ignore this problem for the initial implementation and
535 * state that an "original" hardlink must still exist for us to find
536 * historic path name. */
537 if (pli->pli_recno != -1) {
538 rc = mdd_path_historic(env, pli);
540 *recno = pli->pli_currec;
541 /* Return next link index to caller */
542 *linkno = pli->pli_linkno;
550 int mdd_get_flags(const struct lu_env *env, struct mdd_object *obj)
552 struct lu_attr *la = &mdd_env_info(env)->mti_la;
556 rc = mdd_la_get(env, obj, la, BYPASS_CAPA);
558 mdd_flags_xlate(obj, la->la_flags);
559 if (S_ISDIR(la->la_mode) && la->la_nlink == 1)
560 obj->mod_flags |= MNLINK_OBJ;
565 /* get only inode attributes */
566 int mdd_iattr_get(const struct lu_env *env, struct mdd_object *mdd_obj,
572 if (ma->ma_valid & MA_INODE)
575 rc = mdd_la_get(env, mdd_obj, &ma->ma_attr,
576 mdd_object_capa(env, mdd_obj));
578 ma->ma_valid |= MA_INODE;
582 int mdd_get_default_md(struct mdd_object *mdd_obj, struct lov_mds_md *lmm)
584 struct lov_desc *ldesc;
585 struct mdd_device *mdd = mdo2mdd(&mdd_obj->mod_obj);
586 struct lov_user_md *lum = (struct lov_user_md*)lmm;
592 ldesc = &mdd->mdd_obd_dev->u.mds.mds_lov_desc;
593 LASSERT(ldesc != NULL);
595 lum->lmm_magic = LOV_MAGIC_V1;
596 lum->lmm_object_seq = FID_SEQ_LOV_DEFAULT;
597 lum->lmm_pattern = ldesc->ld_pattern;
598 lum->lmm_stripe_size = ldesc->ld_default_stripe_size;
599 lum->lmm_stripe_count = ldesc->ld_default_stripe_count;
600 lum->lmm_stripe_offset = ldesc->ld_default_stripe_offset;
602 RETURN(sizeof(*lum));
605 static int is_rootdir(struct mdd_object *mdd_obj)
607 const struct mdd_device *mdd_dev = mdd_obj2mdd_dev(mdd_obj);
608 const struct lu_fid *fid = mdo2fid(mdd_obj);
610 return lu_fid_eq(&mdd_dev->mdd_root_fid, fid);
613 /* get lov EA only */
614 static int __mdd_lmm_get(const struct lu_env *env,
615 struct mdd_object *mdd_obj, struct md_attr *ma)
620 if (ma->ma_valid & MA_LOV)
623 rc = mdd_get_md(env, mdd_obj, ma->ma_lmm, &ma->ma_lmm_size,
625 if (rc == 0 && (ma->ma_need & MA_LOV_DEF) && is_rootdir(mdd_obj))
626 rc = mdd_get_default_md(mdd_obj, ma->ma_lmm);
628 ma->ma_lmm_size = rc;
629 ma->ma_layout_gen = ma->ma_lmm->lmm_layout_gen;
630 ma->ma_valid |= MA_LOV | MA_LAY_GEN;
636 /* get the first parent fid from link EA */
637 static int mdd_pfid_get(const struct lu_env *env,
638 struct mdd_object *mdd_obj, struct md_attr *ma)
641 struct link_ea_header *leh;
642 struct link_ea_entry *lee;
643 struct lu_fid *pfid = &ma->ma_pfid;
646 if (ma->ma_valid & MA_PFID)
649 buf = mdd_links_get(env, mdd_obj);
651 RETURN(PTR_ERR(buf));
654 lee = (struct link_ea_entry *)(leh + 1);
655 memcpy(pfid, &lee->lee_parent_fid, sizeof(*pfid));
656 fid_be_to_cpu(pfid, pfid);
657 ma->ma_valid |= MA_PFID;
658 if (buf->lb_len > OBD_ALLOC_BIG)
659 /* if we vmalloced a large buffer drop it */
664 int mdd_lmm_get_locked(const struct lu_env *env, struct mdd_object *mdd_obj,
670 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
671 rc = __mdd_lmm_get(env, mdd_obj, ma);
672 mdd_read_unlock(env, mdd_obj);
677 static int __mdd_lmv_get(const struct lu_env *env,
678 struct mdd_object *mdd_obj, struct md_attr *ma)
683 if (ma->ma_valid & MA_LMV)
686 rc = mdd_get_md(env, mdd_obj, ma->ma_lmv, &ma->ma_lmv_size,
689 ma->ma_valid |= MA_LMV;
695 static int __mdd_lma_get(const struct lu_env *env, struct mdd_object *mdd_obj,
698 struct mdd_thread_info *info = mdd_env_info(env);
699 struct lustre_mdt_attrs *lma =
700 (struct lustre_mdt_attrs *)info->mti_xattr_buf;
705 /* If all needed data are already valid, nothing to do */
706 if ((ma->ma_valid & (MA_HSM | MA_SOM)) ==
707 (ma->ma_need & (MA_HSM | MA_SOM)))
710 /* Read LMA from disk EA */
711 lma_size = sizeof(info->mti_xattr_buf);
712 rc = mdd_get_md(env, mdd_obj, lma, &lma_size, XATTR_NAME_LMA);
716 /* Useless to check LMA incompatibility because this is already done in
717 * osd_ea_fid_get(), and this will fail long before this code is
719 * So, if we are here, LMA is compatible.
722 lustre_lma_swab(lma);
724 /* Swab and copy LMA */
725 if (ma->ma_need & MA_HSM) {
726 if (lma->lma_compat & LMAC_HSM)
727 ma->ma_hsm.mh_flags = lma->lma_flags & HSM_FLAGS_MASK;
729 ma->ma_hsm.mh_flags = 0;
730 ma->ma_valid |= MA_HSM;
734 if (ma->ma_need & MA_SOM && lma->lma_compat & LMAC_SOM) {
735 LASSERT(ma->ma_som != NULL);
736 ma->ma_som->msd_ioepoch = lma->lma_ioepoch;
737 ma->ma_som->msd_size = lma->lma_som_size;
738 ma->ma_som->msd_blocks = lma->lma_som_blocks;
739 ma->ma_som->msd_mountid = lma->lma_som_mountid;
740 ma->ma_valid |= MA_SOM;
746 int mdd_attr_get_internal(const struct lu_env *env, struct mdd_object *mdd_obj,
752 if (ma->ma_need & MA_INODE)
753 rc = mdd_iattr_get(env, mdd_obj, ma);
755 if (rc == 0 && ma->ma_need & MA_LOV) {
756 if (S_ISREG(mdd_object_type(mdd_obj)) ||
757 S_ISDIR(mdd_object_type(mdd_obj)))
758 rc = __mdd_lmm_get(env, mdd_obj, ma);
760 if (rc == 0 && ma->ma_need & MA_PFID && !(ma->ma_valid & MA_LOV)) {
761 if (S_ISREG(mdd_object_type(mdd_obj)))
762 rc = mdd_pfid_get(env, mdd_obj, ma);
764 if (rc == 0 && ma->ma_need & MA_LMV) {
765 if (S_ISDIR(mdd_object_type(mdd_obj)))
766 rc = __mdd_lmv_get(env, mdd_obj, ma);
768 if (rc == 0 && ma->ma_need & (MA_HSM | MA_SOM)) {
769 if (S_ISREG(mdd_object_type(mdd_obj)))
770 rc = __mdd_lma_get(env, mdd_obj, ma);
772 #ifdef CONFIG_FS_POSIX_ACL
773 if (rc == 0 && ma->ma_need & MA_ACL_DEF) {
774 if (S_ISDIR(mdd_object_type(mdd_obj)))
775 rc = mdd_def_acl_get(env, mdd_obj, ma);
778 CDEBUG(D_INODE, "after getattr rc = %d, ma_valid = "LPX64" ma_lmm=%p\n",
779 rc, ma->ma_valid, ma->ma_lmm);
783 int mdd_attr_get_internal_locked(const struct lu_env *env,
784 struct mdd_object *mdd_obj, struct md_attr *ma)
787 int needlock = ma->ma_need &
788 (MA_LOV | MA_LMV | MA_ACL_DEF | MA_HSM | MA_SOM | MA_PFID);
791 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
792 rc = mdd_attr_get_internal(env, mdd_obj, ma);
794 mdd_read_unlock(env, mdd_obj);
799 * No permission check is needed.
801 static int mdd_attr_get(const struct lu_env *env, struct md_object *obj,
804 struct mdd_object *mdd_obj = md2mdd_obj(obj);
808 rc = mdd_attr_get_internal_locked(env, mdd_obj, ma);
813 * No permission check is needed.
815 static int mdd_xattr_get(const struct lu_env *env,
816 struct md_object *obj, struct lu_buf *buf,
819 struct mdd_object *mdd_obj = md2mdd_obj(obj);
824 LASSERT(mdd_object_exists(mdd_obj));
826 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
827 rc = mdo_xattr_get(env, mdd_obj, buf, name,
828 mdd_object_capa(env, mdd_obj));
829 mdd_read_unlock(env, mdd_obj);
835 * Permission check is done when open,
836 * no need check again.
838 static int mdd_readlink(const struct lu_env *env, struct md_object *obj,
841 struct mdd_object *mdd_obj = md2mdd_obj(obj);
842 struct dt_object *next;
847 LASSERT(mdd_object_exists(mdd_obj));
849 next = mdd_object_child(mdd_obj);
850 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
851 rc = next->do_body_ops->dbo_read(env, next, buf, &pos,
852 mdd_object_capa(env, mdd_obj));
853 mdd_read_unlock(env, mdd_obj);
858 * No permission check is needed.
860 static int mdd_xattr_list(const struct lu_env *env, struct md_object *obj,
863 struct mdd_object *mdd_obj = md2mdd_obj(obj);
868 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
869 rc = mdo_xattr_list(env, mdd_obj, buf, mdd_object_capa(env, mdd_obj));
870 mdd_read_unlock(env, mdd_obj);
875 int mdd_declare_object_create_internal(const struct lu_env *env,
876 struct mdd_object *p,
877 struct mdd_object *c,
879 struct thandle *handle,
880 const struct md_op_spec *spec)
882 struct dt_object_format *dof = &mdd_env_info(env)->mti_dof;
883 const struct dt_index_features *feat = spec->sp_feat;
887 if (feat != &dt_directory_features && feat != NULL)
888 dof->dof_type = DFT_INDEX;
890 dof->dof_type = dt_mode_to_dft(ma->ma_attr.la_mode);
892 dof->u.dof_idx.di_feat = feat;
894 rc = mdo_declare_create_obj(env, c, &ma->ma_attr, NULL, dof, handle);
899 int mdd_object_create_internal(const struct lu_env *env, struct mdd_object *p,
900 struct mdd_object *c, struct md_attr *ma,
901 struct thandle *handle,
902 const struct md_op_spec *spec)
904 struct lu_attr *attr = &ma->ma_attr;
905 struct dt_allocation_hint *hint = &mdd_env_info(env)->mti_hint;
906 struct dt_object_format *dof = &mdd_env_info(env)->mti_dof;
907 const struct dt_index_features *feat = spec->sp_feat;
911 if (!mdd_object_exists(c)) {
912 struct dt_object *next = mdd_object_child(c);
915 if (feat != &dt_directory_features && feat != NULL)
916 dof->dof_type = DFT_INDEX;
918 dof->dof_type = dt_mode_to_dft(attr->la_mode);
920 dof->u.dof_idx.di_feat = feat;
922 /* @hint will be initialized by underlying device. */
923 next->do_ops->do_ah_init(env, hint,
924 p ? mdd_object_child(p) : NULL,
925 attr->la_mode & S_IFMT);
927 rc = mdo_create_obj(env, c, attr, hint, dof, handle);
928 LASSERT(ergo(rc == 0, mdd_object_exists(c)));
936 * Make sure the ctime is increased only.
938 static inline int mdd_attr_check(const struct lu_env *env,
939 struct mdd_object *obj,
940 struct lu_attr *attr)
942 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
946 if (attr->la_valid & LA_CTIME) {
947 rc = mdd_la_get(env, obj, tmp_la, BYPASS_CAPA);
951 if (attr->la_ctime < tmp_la->la_ctime)
952 attr->la_valid &= ~(LA_MTIME | LA_CTIME);
953 else if (attr->la_valid == LA_CTIME &&
954 attr->la_ctime == tmp_la->la_ctime)
955 attr->la_valid &= ~LA_CTIME;
960 int mdd_attr_set_internal(const struct lu_env *env,
961 struct mdd_object *obj,
962 struct lu_attr *attr,
963 struct thandle *handle,
969 rc = mdo_attr_set(env, obj, attr, handle, mdd_object_capa(env, obj));
970 #ifdef CONFIG_FS_POSIX_ACL
971 if (!rc && (attr->la_valid & LA_MODE) && needacl)
972 rc = mdd_acl_chmod(env, obj, attr->la_mode, handle);
977 int mdd_attr_check_set_internal(const struct lu_env *env,
978 struct mdd_object *obj,
979 struct lu_attr *attr,
980 struct thandle *handle,
986 rc = mdd_attr_check(env, obj, attr);
991 rc = mdd_attr_set_internal(env, obj, attr, handle, needacl);
995 static int mdd_attr_set_internal_locked(const struct lu_env *env,
996 struct mdd_object *obj,
997 struct lu_attr *attr,
998 struct thandle *handle,
1004 needacl = needacl && (attr->la_valid & LA_MODE);
1006 mdd_write_lock(env, obj, MOR_TGT_CHILD);
1007 rc = mdd_attr_set_internal(env, obj, attr, handle, needacl);
1009 mdd_write_unlock(env, obj);
1013 int mdd_attr_check_set_internal_locked(const struct lu_env *env,
1014 struct mdd_object *obj,
1015 struct lu_attr *attr,
1016 struct thandle *handle,
1022 needacl = needacl && (attr->la_valid & LA_MODE);
1024 mdd_write_lock(env, obj, MOR_TGT_CHILD);
1025 rc = mdd_attr_check_set_internal(env, obj, attr, handle, needacl);
1027 mdd_write_unlock(env, obj);
1031 int __mdd_xattr_set(const struct lu_env *env, struct mdd_object *obj,
1032 const struct lu_buf *buf, const char *name,
1033 int fl, struct thandle *handle)
1035 struct lustre_capa *capa = mdd_object_capa(env, obj);
1039 if (buf->lb_buf && buf->lb_len > 0)
1040 rc = mdo_xattr_set(env, obj, buf, name, 0, handle, capa);
1041 else if (buf->lb_buf == NULL && buf->lb_len == 0)
1042 rc = mdo_xattr_del(env, obj, name, handle, capa);
1048 * This gives the same functionality as the code between
1049 * sys_chmod and inode_setattr
1050 * chown_common and inode_setattr
1051 * utimes and inode_setattr
1052 * This API is ported from mds_fix_attr but remove some unnecesssary stuff.
1054 static int mdd_fix_attr(const struct lu_env *env, struct mdd_object *obj,
1055 struct lu_attr *la, const struct md_attr *ma)
1057 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
1058 struct md_ucred *uc;
1065 /* Do not permit change file type */
1066 if (la->la_valid & LA_TYPE)
1069 /* They should not be processed by setattr */
1070 if (la->la_valid & (LA_NLINK | LA_RDEV | LA_BLKSIZE))
1073 /* export destroy does not have ->le_ses, but we may want
1074 * to drop LUSTRE_SOM_FL. */
1080 rc = mdd_la_get(env, obj, tmp_la, BYPASS_CAPA);
1084 if (la->la_valid == LA_CTIME) {
1085 if (!(ma->ma_attr_flags & MDS_PERM_BYPASS))
1086 /* This is only for set ctime when rename's source is
1088 rc = mdd_may_delete(env, NULL, obj,
1089 (struct md_attr *)ma, 1, 0);
1090 if (rc == 0 && la->la_ctime <= tmp_la->la_ctime)
1091 la->la_valid &= ~LA_CTIME;
1095 if (la->la_valid == LA_ATIME) {
1096 /* This is atime only set for read atime update on close. */
1097 if (la->la_atime >= tmp_la->la_atime &&
1098 la->la_atime < (tmp_la->la_atime +
1099 mdd_obj2mdd_dev(obj)->mdd_atime_diff))
1100 la->la_valid &= ~LA_ATIME;
1104 /* Check if flags change. */
1105 if (la->la_valid & LA_FLAGS) {
1106 unsigned int oldflags = 0;
1107 unsigned int newflags = la->la_flags &
1108 (LUSTRE_IMMUTABLE_FL | LUSTRE_APPEND_FL);
1110 if ((uc->mu_fsuid != tmp_la->la_uid) &&
1111 !mdd_capable(uc, CFS_CAP_FOWNER))
1114 /* XXX: the IMMUTABLE and APPEND_ONLY flags can
1115 * only be changed by the relevant capability. */
1116 if (mdd_is_immutable(obj))
1117 oldflags |= LUSTRE_IMMUTABLE_FL;
1118 if (mdd_is_append(obj))
1119 oldflags |= LUSTRE_APPEND_FL;
1120 if ((oldflags ^ newflags) &&
1121 !mdd_capable(uc, CFS_CAP_LINUX_IMMUTABLE))
1124 if (!S_ISDIR(tmp_la->la_mode))
1125 la->la_flags &= ~LUSTRE_DIRSYNC_FL;
1128 if ((mdd_is_immutable(obj) || mdd_is_append(obj)) &&
1129 (la->la_valid & ~LA_FLAGS) &&
1130 !(ma->ma_attr_flags & MDS_PERM_BYPASS))
1133 /* Check for setting the obj time. */
1134 if ((la->la_valid & (LA_MTIME | LA_ATIME | LA_CTIME)) &&
1135 !(la->la_valid & ~(LA_MTIME | LA_ATIME | LA_CTIME))) {
1136 if ((uc->mu_fsuid != tmp_la->la_uid) &&
1137 !mdd_capable(uc, CFS_CAP_FOWNER)) {
1138 rc = mdd_permission_internal_locked(env, obj, tmp_la,
1146 if (la->la_valid & LA_KILL_SUID) {
1147 la->la_valid &= ~LA_KILL_SUID;
1148 if ((tmp_la->la_mode & S_ISUID) &&
1149 !(la->la_valid & LA_MODE)) {
1150 la->la_mode = tmp_la->la_mode;
1151 la->la_valid |= LA_MODE;
1153 la->la_mode &= ~S_ISUID;
1156 if (la->la_valid & LA_KILL_SGID) {
1157 la->la_valid &= ~LA_KILL_SGID;
1158 if (((tmp_la->la_mode & (S_ISGID | S_IXGRP)) ==
1159 (S_ISGID | S_IXGRP)) &&
1160 !(la->la_valid & LA_MODE)) {
1161 la->la_mode = tmp_la->la_mode;
1162 la->la_valid |= LA_MODE;
1164 la->la_mode &= ~S_ISGID;
1167 /* Make sure a caller can chmod. */
1168 if (la->la_valid & LA_MODE) {
1169 if (!(ma->ma_attr_flags & MDS_PERM_BYPASS) &&
1170 (uc->mu_fsuid != tmp_la->la_uid) &&
1171 !mdd_capable(uc, CFS_CAP_FOWNER))
1174 if (la->la_mode == (cfs_umode_t) -1)
1175 la->la_mode = tmp_la->la_mode;
1177 la->la_mode = (la->la_mode & S_IALLUGO) |
1178 (tmp_la->la_mode & ~S_IALLUGO);
1180 /* Also check the setgid bit! */
1181 if (!lustre_in_group_p(uc, (la->la_valid & LA_GID) ?
1182 la->la_gid : tmp_la->la_gid) &&
1183 !mdd_capable(uc, CFS_CAP_FSETID))
1184 la->la_mode &= ~S_ISGID;
1186 la->la_mode = tmp_la->la_mode;
1189 /* Make sure a caller can chown. */
1190 if (la->la_valid & LA_UID) {
1191 if (la->la_uid == (uid_t) -1)
1192 la->la_uid = tmp_la->la_uid;
1193 if (((uc->mu_fsuid != tmp_la->la_uid) ||
1194 (la->la_uid != tmp_la->la_uid)) &&
1195 !mdd_capable(uc, CFS_CAP_CHOWN))
1198 /* If the user or group of a non-directory has been
1199 * changed by a non-root user, remove the setuid bit.
1200 * 19981026 David C Niemi <niemi@tux.org>
1202 * Changed this to apply to all users, including root,
1203 * to avoid some races. This is the behavior we had in
1204 * 2.0. The check for non-root was definitely wrong
1205 * for 2.2 anyway, as it should have been using
1206 * CAP_FSETID rather than fsuid -- 19990830 SD. */
1207 if (((tmp_la->la_mode & S_ISUID) == S_ISUID) &&
1208 !S_ISDIR(tmp_la->la_mode)) {
1209 la->la_mode &= ~S_ISUID;
1210 la->la_valid |= LA_MODE;
1214 /* Make sure caller can chgrp. */
1215 if (la->la_valid & LA_GID) {
1216 if (la->la_gid == (gid_t) -1)
1217 la->la_gid = tmp_la->la_gid;
1218 if (((uc->mu_fsuid != tmp_la->la_uid) ||
1219 ((la->la_gid != tmp_la->la_gid) &&
1220 !lustre_in_group_p(uc, la->la_gid))) &&
1221 !mdd_capable(uc, CFS_CAP_CHOWN))
1224 /* Likewise, if the user or group of a non-directory
1225 * has been changed by a non-root user, remove the
1226 * setgid bit UNLESS there is no group execute bit
1227 * (this would be a file marked for mandatory
1228 * locking). 19981026 David C Niemi <niemi@tux.org>
1230 * Removed the fsuid check (see the comment above) --
1232 if (((tmp_la->la_mode & (S_ISGID | S_IXGRP)) ==
1233 (S_ISGID | S_IXGRP)) && !S_ISDIR(tmp_la->la_mode)) {
1234 la->la_mode &= ~S_ISGID;
1235 la->la_valid |= LA_MODE;
1239 /* For both Size-on-MDS case and truncate case,
1240 * "la->la_valid & (LA_SIZE | LA_BLOCKS)" are ture.
1241 * We distinguish them by "ma->ma_attr_flags & MDS_SOM".
1242 * For SOM case, it is true, the MAY_WRITE perm has been checked
1243 * when open, no need check again. For truncate case, it is false,
1244 * the MAY_WRITE perm should be checked here. */
1245 if (ma->ma_attr_flags & MDS_SOM) {
1246 /* For the "Size-on-MDS" setattr update, merge coming
1247 * attributes with the set in the inode. BUG 10641 */
1248 if ((la->la_valid & LA_ATIME) &&
1249 (la->la_atime <= tmp_la->la_atime))
1250 la->la_valid &= ~LA_ATIME;
1252 /* OST attributes do not have a priority over MDS attributes,
1253 * so drop times if ctime is equal. */
1254 if ((la->la_valid & LA_CTIME) &&
1255 (la->la_ctime <= tmp_la->la_ctime))
1256 la->la_valid &= ~(LA_MTIME | LA_CTIME);
1258 if (la->la_valid & (LA_SIZE | LA_BLOCKS)) {
1259 if (!((ma->ma_attr_flags & MDS_OPEN_OWNEROVERRIDE) &&
1260 (uc->mu_fsuid == tmp_la->la_uid)) &&
1261 !(ma->ma_attr_flags & MDS_PERM_BYPASS)) {
1262 rc = mdd_permission_internal_locked(env, obj,
1269 if (la->la_valid & LA_CTIME) {
1270 /* The pure setattr, it has the priority over what is
1271 * already set, do not drop it if ctime is equal. */
1272 if (la->la_ctime < tmp_la->la_ctime)
1273 la->la_valid &= ~(LA_ATIME | LA_MTIME |
1281 /** Store a data change changelog record
1282 * If this fails, we must fail the whole transaction; we don't
1283 * want the change to commit without the log entry.
1284 * \param mdd_obj - mdd_object of change
1285 * \param handle - transacion handle
1287 static int mdd_changelog_data_store(const struct lu_env *env,
1288 struct mdd_device *mdd,
1289 enum changelog_rec_type type,
1291 struct mdd_object *mdd_obj,
1292 struct thandle *handle)
1294 const struct lu_fid *tfid = mdo2fid(mdd_obj);
1295 struct llog_changelog_rec *rec;
1296 struct thandle *th = NULL;
1302 if (!(mdd->mdd_cl.mc_flags & CLM_ON))
1304 if ((mdd->mdd_cl.mc_mask & (1 << type)) == 0)
1307 LASSERT(mdd_obj != NULL);
1308 LASSERT(handle != NULL);
1310 if ((type >= CL_MTIME) && (type <= CL_ATIME) &&
1311 cfs_time_before_64(mdd->mdd_cl.mc_starttime, mdd_obj->mod_cltime)) {
1312 /* Don't need multiple updates in this log */
1313 /* Don't check under lock - no big deal if we get an extra
1318 reclen = llog_data_len(sizeof(*rec));
1319 buf = mdd_buf_alloc(env, reclen);
1320 if (buf->lb_buf == NULL)
1322 rec = (struct llog_changelog_rec *)buf->lb_buf;
1324 rec->cr.cr_flags = CLF_VERSION | (CLF_FLAGMASK & flags);
1325 rec->cr.cr_type = (__u32)type;
1326 rec->cr.cr_tfid = *tfid;
1327 rec->cr.cr_namelen = 0;
1328 mdd_obj->mod_cltime = cfs_time_current_64();
1330 rc = mdd_changelog_llog_write(mdd, rec, handle ? : th);
1333 mdd_trans_stop(env, mdd, rc, th);
1336 CERROR("changelog failed: rc=%d op%d t"DFID"\n",
1337 rc, type, PFID(tfid));
1344 int mdd_changelog(const struct lu_env *env, enum changelog_rec_type type,
1345 int flags, struct md_object *obj)
1347 struct thandle *handle;
1348 struct mdd_object *mdd_obj = md2mdd_obj(obj);
1349 struct mdd_device *mdd = mdo2mdd(obj);
1353 handle = mdd_trans_create(env, mdd);
1355 return(PTR_ERR(handle));
1357 rc = mdd_declare_changelog_store(env, mdd, NULL, handle);
1361 rc = mdd_trans_start(env, mdd, handle);
1365 rc = mdd_changelog_data_store(env, mdd, type, flags, mdd_obj,
1369 mdd_trans_stop(env, mdd, rc, handle);
1375 * Should be called with write lock held.
1377 * \see mdd_lma_set_locked().
1379 static int __mdd_lma_set(const struct lu_env *env, struct mdd_object *mdd_obj,
1380 const struct md_attr *ma, struct thandle *handle)
1382 struct mdd_thread_info *info = mdd_env_info(env);
1384 struct lustre_mdt_attrs *lma =
1385 (struct lustre_mdt_attrs *) info->mti_xattr_buf;
1386 int lmasize = sizeof(struct lustre_mdt_attrs);
1391 /* Either HSM or SOM part is not valid, we need to read it before */
1392 if ((!ma->ma_valid) & (MA_HSM | MA_SOM)) {
1393 rc = mdd_get_md(env, mdd_obj, lma, &lmasize, XATTR_NAME_LMA);
1397 lustre_lma_swab(lma);
1399 memset(lma, 0, lmasize);
1403 if (ma->ma_valid & MA_HSM) {
1404 lma->lma_flags |= ma->ma_hsm.mh_flags & HSM_FLAGS_MASK;
1405 lma->lma_compat |= LMAC_HSM;
1409 if (ma->ma_valid & MA_SOM) {
1410 LASSERT(ma->ma_som != NULL);
1411 if (ma->ma_som->msd_ioepoch == IOEPOCH_INVAL) {
1412 lma->lma_compat &= ~LMAC_SOM;
1414 lma->lma_compat |= LMAC_SOM;
1415 lma->lma_ioepoch = ma->ma_som->msd_ioepoch;
1416 lma->lma_som_size = ma->ma_som->msd_size;
1417 lma->lma_som_blocks = ma->ma_som->msd_blocks;
1418 lma->lma_som_mountid = ma->ma_som->msd_mountid;
1423 memcpy(&lma->lma_self_fid, mdo2fid(mdd_obj), sizeof(lma->lma_self_fid));
1425 lustre_lma_swab(lma);
1426 buf = mdd_buf_get(env, lma, lmasize);
1427 rc = __mdd_xattr_set(env, mdd_obj, buf, XATTR_NAME_LMA, 0, handle);
1433 * Save LMA extended attributes with data from \a ma.
1435 * HSM and Size-On-MDS data will be extracted from \ma if they are valid, if
1436 * not, LMA EA will be first read from disk, modified and write back.
1439 static int mdd_lma_set_locked(const struct lu_env *env,
1440 struct mdd_object *mdd_obj,
1441 const struct md_attr *ma, struct thandle *handle)
1445 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
1446 rc = __mdd_lma_set(env, mdd_obj, ma, handle);
1447 mdd_write_unlock(env, mdd_obj);
1451 /* Precedence for choosing record type when multiple
1452 * attributes change: setattr > mtime > ctime > atime
1453 * (ctime changes when mtime does, plus chmod/chown.
1454 * atime and ctime are independent.) */
1455 static int mdd_attr_set_changelog(const struct lu_env *env,
1456 struct md_object *obj, struct thandle *handle,
1459 struct mdd_device *mdd = mdo2mdd(obj);
1462 bits = (valid & ~(LA_CTIME|LA_MTIME|LA_ATIME)) ? 1 << CL_SETATTR : 0;
1463 bits |= (valid & LA_MTIME) ? 1 << CL_MTIME : 0;
1464 bits |= (valid & LA_CTIME) ? 1 << CL_CTIME : 0;
1465 bits |= (valid & LA_ATIME) ? 1 << CL_ATIME : 0;
1466 bits = bits & mdd->mdd_cl.mc_mask;
1470 /* The record type is the lowest non-masked set bit */
1471 while (bits && ((bits & 1) == 0)) {
1476 /* FYI we only store the first CLF_FLAGMASK bits of la_valid */
1477 return mdd_changelog_data_store(env, mdd, type, (int)valid,
1478 md2mdd_obj(obj), handle);
1481 static int mdd_declare_attr_set(const struct lu_env *env,
1482 struct mdd_device *mdd,
1483 struct mdd_object *obj,
1484 const struct md_attr *ma,
1485 struct lov_mds_md *lmm,
1486 struct thandle *handle)
1488 struct lu_buf *buf = &mdd_env_info(env)->mti_buf;
1491 rc = mdo_declare_attr_set(env, obj, &ma->ma_attr, handle);
1495 rc = mdd_declare_changelog_store(env, mdd, NULL, handle);
1499 if (ma->ma_valid & MA_LOV) {
1501 buf->lb_len = ma->ma_lmm_size;
1502 rc = mdo_declare_xattr_set(env, obj, buf, XATTR_NAME_LOV,
1508 if (ma->ma_valid & (MA_HSM | MA_SOM)) {
1510 buf->lb_len = sizeof(struct lustre_mdt_attrs);
1511 rc = mdo_declare_xattr_set(env, obj, buf, XATTR_NAME_LMA,
1517 /* basically the log is the same as in unlink case */
1521 if (le32_to_cpu(lmm->lmm_magic) != LOV_MAGIC_V1 &&
1522 le32_to_cpu(lmm->lmm_magic) != LOV_MAGIC_V3) {
1523 CERROR("%s: invalid LOV_MAGIC %08x on object "DFID"\n",
1524 mdd->mdd_obd_dev->obd_name,
1525 le32_to_cpu(lmm->lmm_magic),
1526 PFID(lu_object_fid(&obj->mod_obj.mo_lu)));
1530 stripe = le16_to_cpu(lmm->lmm_stripe_count);
1531 if (stripe == LOV_ALL_STRIPES) {
1532 struct lov_desc *ldesc;
1534 ldesc = &mdd->mdd_obd_dev->u.mds.mds_lov_desc;
1535 LASSERT(ldesc != NULL);
1536 stripe = ldesc->ld_tgt_count;
1539 for (i = 0; i < stripe; i++) {
1540 rc = mdd_declare_llog_record(env, mdd,
1541 sizeof(struct llog_unlink_rec),
1551 /* set attr and LOV EA at once, return updated attr */
1552 static int mdd_attr_set(const struct lu_env *env, struct md_object *obj,
1553 const struct md_attr *ma)
1555 struct mdd_object *mdd_obj = md2mdd_obj(obj);
1556 struct mdd_device *mdd = mdo2mdd(obj);
1557 struct thandle *handle;
1558 struct lov_mds_md *lmm = NULL;
1559 struct llog_cookie *logcookies = NULL;
1560 int rc, lmm_size = 0, cookie_size = 0;
1561 struct lu_attr *la_copy = &mdd_env_info(env)->mti_la_for_fix;
1562 struct obd_device *obd = mdd->mdd_obd_dev;
1563 struct mds_obd *mds = &obd->u.mds;
1564 #ifdef HAVE_QUOTA_SUPPORT
1565 unsigned int qnids[MAXQUOTAS] = { 0, 0 };
1566 unsigned int qoids[MAXQUOTAS] = { 0, 0 };
1567 int quota_opc = 0, block_count = 0;
1568 int inode_pending[MAXQUOTAS] = { 0, 0 };
1569 int block_pending[MAXQUOTAS] = { 0, 0 };
1573 *la_copy = ma->ma_attr;
1574 rc = mdd_fix_attr(env, mdd_obj, la_copy, ma);
1578 /* setattr on "close" only change atime, or do nothing */
1579 if (ma->ma_valid == MA_INODE &&
1580 ma->ma_attr.la_valid == LA_ATIME && la_copy->la_valid == 0)
1583 if (S_ISREG(mdd_object_type(mdd_obj)) &&
1584 ma->ma_attr.la_valid & (LA_UID | LA_GID)) {
1585 lmm_size = mdd_lov_mdsize(env, mdd);
1586 lmm = mdd_max_lmm_get(env, mdd);
1590 rc = mdd_get_md_locked(env, mdd_obj, lmm, &lmm_size,
1597 handle = mdd_trans_create(env, mdd);
1599 RETURN(PTR_ERR(handle));
1601 rc = mdd_declare_attr_set(env, mdd, mdd_obj, ma,
1602 lmm_size > 0 ? lmm : NULL, handle);
1606 /* permission changes may require sync operation */
1607 if (ma->ma_attr.la_valid & (LA_MODE|LA_UID|LA_GID))
1608 handle->th_sync = !!mdd->mdd_sync_permission;
1610 rc = mdd_trans_start(env, mdd, handle);
1614 /* permission changes may require sync operation */
1615 if (ma->ma_attr.la_valid & (LA_MODE|LA_UID|LA_GID))
1616 handle->th_sync |= mdd->mdd_sync_permission;
1618 if (ma->ma_attr.la_valid & (LA_MTIME | LA_CTIME))
1619 CDEBUG(D_INODE, "setting mtime "LPU64", ctime "LPU64"\n",
1620 ma->ma_attr.la_mtime, ma->ma_attr.la_ctime);
1622 #ifdef HAVE_QUOTA_SUPPORT
1623 if (mds->mds_quota && la_copy->la_valid & (LA_UID | LA_GID)) {
1624 struct obd_export *exp = md_quota(env)->mq_exp;
1625 struct lu_attr *la_tmp = &mdd_env_info(env)->mti_la;
1627 rc = mdd_la_get(env, mdd_obj, la_tmp, BYPASS_CAPA);
1629 quota_opc = FSFILT_OP_SETATTR;
1630 mdd_quota_wrapper(la_copy, qnids);
1631 mdd_quota_wrapper(la_tmp, qoids);
1632 /* get file quota for new owner */
1633 lquota_chkquota(mds_quota_interface_ref, obd, exp,
1634 qnids, inode_pending, 1, NULL, 0,
1636 block_count = (la_tmp->la_blocks + 7) >> 3;
1639 mdd_data_get(env, mdd_obj, &data);
1640 /* get block quota for new owner */
1641 lquota_chkquota(mds_quota_interface_ref, obd,
1642 exp, qnids, block_pending,
1644 LQUOTA_FLAGS_BLK, data, 1);
1650 if (la_copy->la_valid & LA_FLAGS) {
1651 rc = mdd_attr_set_internal_locked(env, mdd_obj, la_copy,
1654 mdd_flags_xlate(mdd_obj, la_copy->la_flags);
1655 } else if (la_copy->la_valid) { /* setattr */
1656 rc = mdd_attr_set_internal_locked(env, mdd_obj, la_copy,
1658 /* journal chown/chgrp in llog, just like unlink */
1659 if (rc == 0 && lmm_size){
1660 cookie_size = mdd_lov_cookiesize(env, mdd);
1661 logcookies = mdd_max_cookie_get(env, mdd);
1662 if (logcookies == NULL)
1663 GOTO(cleanup, rc = -ENOMEM);
1665 if (mdd_setattr_log(env, mdd, ma, lmm, lmm_size,
1666 logcookies, cookie_size) <= 0)
1671 if (rc == 0 && ma->ma_valid & MA_LOV) {
1674 mode = mdd_object_type(mdd_obj);
1675 if (S_ISREG(mode) || S_ISDIR(mode)) {
1676 rc = mdd_lsm_sanity_check(env, mdd_obj);
1680 rc = mdd_lov_set_md(env, NULL, mdd_obj, ma->ma_lmm,
1681 ma->ma_lmm_size, handle, 1);
1685 if (rc == 0 && ma->ma_valid & (MA_HSM | MA_SOM)) {
1688 mode = mdd_object_type(mdd_obj);
1690 rc = mdd_lma_set_locked(env, mdd_obj, ma, handle);
1695 rc = mdd_attr_set_changelog(env, obj, handle,
1696 ma->ma_attr.la_valid);
1698 mdd_trans_stop(env, mdd, rc, handle);
1699 if (rc == 0 && (lmm != NULL && lmm_size > 0 )) {
1700 /*set obd attr, if needed*/
1701 rc = mdd_lov_setattr_async(env, mdd_obj, lmm, lmm_size,
1704 #ifdef HAVE_QUOTA_SUPPORT
1706 lquota_pending_commit(mds_quota_interface_ref, obd, qnids,
1708 lquota_pending_commit(mds_quota_interface_ref, obd, qnids,
1710 /* Trigger dqrel/dqacq for original owner and new owner.
1711 * If failed, the next call for lquota_chkquota will
1713 lquota_adjust(mds_quota_interface_ref, obd, qnids, qoids, rc,
1720 int mdd_xattr_set_txn(const struct lu_env *env, struct mdd_object *obj,
1721 const struct lu_buf *buf, const char *name, int fl,
1722 struct thandle *handle)
1727 mdd_write_lock(env, obj, MOR_TGT_CHILD);
1728 rc = __mdd_xattr_set(env, obj, buf, name, fl, handle);
1729 mdd_write_unlock(env, obj);
1734 static int mdd_xattr_sanity_check(const struct lu_env *env,
1735 struct mdd_object *obj)
1737 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
1738 struct md_ucred *uc = md_ucred(env);
1742 if (mdd_is_immutable(obj) || mdd_is_append(obj))
1745 rc = mdd_la_get(env, obj, tmp_la, BYPASS_CAPA);
1749 if ((uc->mu_fsuid != tmp_la->la_uid) &&
1750 !mdd_capable(uc, CFS_CAP_FOWNER))
1756 static int mdd_declare_xattr_set(const struct lu_env *env,
1757 struct mdd_device *mdd,
1758 struct mdd_object *obj,
1759 const struct lu_buf *buf,
1761 struct thandle *handle)
1766 rc = mdo_declare_xattr_set(env, obj, buf, name, 0, handle);
1770 /* Only record user xattr changes */
1771 if ((strncmp("user.", name, 5) == 0))
1772 rc = mdd_declare_changelog_store(env, mdd, NULL, handle);
1778 * The caller should guarantee to update the object ctime
1779 * after xattr_set if needed.
1781 static int mdd_xattr_set(const struct lu_env *env, struct md_object *obj,
1782 const struct lu_buf *buf, const char *name,
1785 struct mdd_object *mdd_obj = md2mdd_obj(obj);
1786 struct mdd_device *mdd = mdo2mdd(obj);
1787 struct thandle *handle;
1791 rc = mdd_xattr_sanity_check(env, mdd_obj);
1795 handle = mdd_trans_create(env, mdd);
1797 RETURN(PTR_ERR(handle));
1799 /* security-replated changes may require sync */
1800 if (!strcmp(name, XATTR_NAME_ACL_ACCESS) &&
1801 mdd->mdd_sync_permission == 1)
1802 handle->th_sync = 1;
1804 rc = mdd_declare_xattr_set(env, mdd, mdd_obj, buf, name, handle);
1808 rc = mdd_trans_start(env, mdd, handle);
1812 /* security-replated changes may require sync */
1813 if (!strcmp(name, XATTR_NAME_ACL_ACCESS))
1814 handle->th_sync |= mdd->mdd_sync_permission;
1816 rc = mdd_xattr_set_txn(env, mdd_obj, buf, name, fl, handle);
1818 /* Only record system & user xattr changes */
1819 if ((rc == 0) && (strncmp(XATTR_USER_PREFIX, name,
1820 sizeof(XATTR_USER_PREFIX) - 1) == 0 ||
1821 strncmp(POSIX_ACL_XATTR_ACCESS, name,
1822 sizeof(POSIX_ACL_XATTR_ACCESS) - 1) == 0 ||
1823 strncmp(POSIX_ACL_XATTR_DEFAULT, name,
1824 sizeof(POSIX_ACL_XATTR_DEFAULT) - 1) == 0))
1825 rc = mdd_changelog_data_store(env, mdd, CL_XATTR, 0, mdd_obj,
1829 mdd_trans_stop(env, mdd, rc, handle);
1834 static int mdd_declare_xattr_del(const struct lu_env *env,
1835 struct mdd_device *mdd,
1836 struct mdd_object *obj,
1838 struct thandle *handle)
1842 rc = mdo_declare_xattr_del(env, obj, name, handle);
1846 /* Only record user xattr changes */
1847 if ((strncmp("user.", name, 5) == 0))
1848 rc = mdd_declare_changelog_store(env, mdd, NULL, handle);
1854 * The caller should guarantee to update the object ctime
1855 * after xattr_set if needed.
1857 int mdd_xattr_del(const struct lu_env *env, struct md_object *obj,
1860 struct mdd_object *mdd_obj = md2mdd_obj(obj);
1861 struct mdd_device *mdd = mdo2mdd(obj);
1862 struct thandle *handle;
1866 rc = mdd_xattr_sanity_check(env, mdd_obj);
1870 handle = mdd_trans_create(env, mdd);
1872 RETURN(PTR_ERR(handle));
1874 rc = mdd_declare_xattr_del(env, mdd, mdd_obj, name, handle);
1878 rc = mdd_trans_start(env, mdd, handle);
1882 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
1883 rc = mdo_xattr_del(env, mdd_obj, name, handle,
1884 mdd_object_capa(env, mdd_obj));
1885 mdd_write_unlock(env, mdd_obj);
1887 /* Only record system & user xattr changes */
1888 if ((rc == 0) && (strncmp(XATTR_USER_PREFIX, name,
1889 sizeof(XATTR_USER_PREFIX) - 1) == 0 ||
1890 strncmp(POSIX_ACL_XATTR_ACCESS, name,
1891 sizeof(POSIX_ACL_XATTR_ACCESS) - 1) == 0 ||
1892 strncmp(POSIX_ACL_XATTR_DEFAULT, name,
1893 sizeof(POSIX_ACL_XATTR_DEFAULT) - 1) == 0))
1894 rc = mdd_changelog_data_store(env, mdd, CL_XATTR, 0, mdd_obj,
1898 mdd_trans_stop(env, mdd, rc, handle);
1903 /* partial unlink */
1904 static int mdd_ref_del(const struct lu_env *env, struct md_object *obj,
1907 struct lu_attr *la_copy = &mdd_env_info(env)->mti_la_for_fix;
1908 struct mdd_object *mdd_obj = md2mdd_obj(obj);
1909 struct mdd_device *mdd = mdo2mdd(obj);
1910 struct thandle *handle;
1911 #ifdef HAVE_QUOTA_SUPPORT
1912 struct obd_device *obd = mdd->mdd_obd_dev;
1913 struct mds_obd *mds = &obd->u.mds;
1914 unsigned int qids[MAXQUOTAS] = { 0, 0 };
1920 /* XXX: this code won't be used ever:
1921 * DNE uses slightly different approach */
1925 * Check -ENOENT early here because we need to get object type
1926 * to calculate credits before transaction start
1928 if (!mdd_object_exists(mdd_obj))
1931 LASSERT(mdd_object_exists(mdd_obj) > 0);
1933 handle = mdd_trans_create(env, mdd);
1937 rc = mdd_trans_start(env, mdd, handle);
1939 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
1941 rc = mdd_unlink_sanity_check(env, NULL, mdd_obj, ma);
1945 __mdd_ref_del(env, mdd_obj, handle, 0);
1947 if (S_ISDIR(lu_object_attr(&obj->mo_lu))) {
1949 __mdd_ref_del(env, mdd_obj, handle, 1);
1952 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
1953 la_copy->la_ctime = ma->ma_attr.la_ctime;
1955 la_copy->la_valid = LA_CTIME;
1956 rc = mdd_attr_check_set_internal(env, mdd_obj, la_copy, handle, 0);
1960 rc = mdd_finish_unlink(env, mdd_obj, ma, handle);
1961 #ifdef HAVE_QUOTA_SUPPORT
1962 if (mds->mds_quota && ma->ma_valid & MA_INODE &&
1963 ma->ma_attr.la_nlink == 0 && mdd_obj->mod_count == 0) {
1964 quota_opc = FSFILT_OP_UNLINK_PARTIAL_CHILD;
1965 mdd_quota_wrapper(&ma->ma_attr, qids);
1972 mdd_write_unlock(env, mdd_obj);
1973 mdd_trans_stop(env, mdd, rc, handle);
1974 #ifdef HAVE_QUOTA_SUPPORT
1976 /* Trigger dqrel on the owner of child. If failed,
1977 * the next call for lquota_chkquota will process it */
1978 lquota_adjust(mds_quota_interface_ref, obd, qids, 0, rc,
1984 /* partial operation */
1985 static int mdd_oc_sanity_check(const struct lu_env *env,
1986 struct mdd_object *obj,
1992 switch (ma->ma_attr.la_mode & S_IFMT) {
2009 static int mdd_object_create(const struct lu_env *env,
2010 struct md_object *obj,
2011 const struct md_op_spec *spec,
2015 struct mdd_device *mdd = mdo2mdd(obj);
2016 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2017 const struct lu_fid *pfid = spec->u.sp_pfid;
2018 struct thandle *handle;
2019 #ifdef HAVE_QUOTA_SUPPORT
2020 struct obd_device *obd = mdd->mdd_obd_dev;
2021 struct obd_export *exp = md_quota(env)->mq_exp;
2022 struct mds_obd *mds = &obd->u.mds;
2023 unsigned int qids[MAXQUOTAS] = { 0, 0 };
2024 int quota_opc = 0, block_count = 0;
2025 int inode_pending[MAXQUOTAS] = { 0, 0 };
2026 int block_pending[MAXQUOTAS] = { 0, 0 };
2031 /* XXX: this code won't be used ever:
2032 * DNE uses slightly different approach */
2035 #ifdef HAVE_QUOTA_SUPPORT
2036 if (mds->mds_quota) {
2037 quota_opc = FSFILT_OP_CREATE_PARTIAL_CHILD;
2038 mdd_quota_wrapper(&ma->ma_attr, qids);
2039 /* get file quota for child */
2040 lquota_chkquota(mds_quota_interface_ref, obd, exp,
2041 qids, inode_pending, 1, NULL, 0,
2043 switch (ma->ma_attr.la_mode & S_IFMT) {
2052 /* get block quota for child */
2054 lquota_chkquota(mds_quota_interface_ref, obd, exp,
2055 qids, block_pending, block_count,
2056 NULL, LQUOTA_FLAGS_BLK, NULL, 0);
2060 handle = mdd_trans_create(env, mdd);
2062 GOTO(out_pending, rc = PTR_ERR(handle));
2064 rc = mdd_trans_start(env, mdd, handle);
2066 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
2067 rc = mdd_oc_sanity_check(env, mdd_obj, ma);
2071 rc = mdd_object_create_internal(env, NULL, mdd_obj, ma, handle, spec);
2075 if (spec->sp_cr_flags & MDS_CREATE_SLAVE_OBJ) {
2076 /* If creating the slave object, set slave EA here. */
2077 int lmv_size = spec->u.sp_ea.eadatalen;
2078 struct lmv_stripe_md *lmv;
2080 lmv = (struct lmv_stripe_md *)spec->u.sp_ea.eadata;
2081 LASSERT(lmv != NULL && lmv_size > 0);
2083 rc = __mdd_xattr_set(env, mdd_obj,
2084 mdd_buf_get_const(env, lmv, lmv_size),
2085 XATTR_NAME_LMV, 0, handle);
2089 rc = mdd_attr_set_internal(env, mdd_obj, &ma->ma_attr,
2092 #ifdef CONFIG_FS_POSIX_ACL
2093 if (spec->sp_cr_flags & MDS_CREATE_RMT_ACL) {
2094 struct lu_buf *buf = &mdd_env_info(env)->mti_buf;
2096 buf->lb_buf = (void *)spec->u.sp_ea.eadata;
2097 buf->lb_len = spec->u.sp_ea.eadatalen;
2098 if ((buf->lb_len > 0) && (buf->lb_buf != NULL)) {
2099 rc = __mdd_acl_init(env, mdd_obj, buf,
2100 &ma->ma_attr.la_mode,
2105 ma->ma_attr.la_valid |= LA_MODE;
2108 pfid = spec->u.sp_ea.fid;
2111 rc = mdd_object_initialize(env, pfid, NULL, mdd_obj, ma, handle,
2117 rc = mdd_attr_get_internal(env, mdd_obj, ma);
2118 mdd_write_unlock(env, mdd_obj);
2120 mdd_trans_stop(env, mdd, rc, handle);
2122 #ifdef HAVE_QUOTA_SUPPORT
2124 lquota_pending_commit(mds_quota_interface_ref, obd, qids,
2126 lquota_pending_commit(mds_quota_interface_ref, obd, qids,
2128 /* Trigger dqacq on the owner of child. If failed,
2129 * the next call for lquota_chkquota will process it. */
2130 lquota_adjust(mds_quota_interface_ref, obd, qids, 0, rc,
2138 static int mdd_ref_add(const struct lu_env *env, struct md_object *obj,
2139 const struct md_attr *ma)
2141 struct lu_attr *la_copy = &mdd_env_info(env)->mti_la_for_fix;
2142 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2143 struct mdd_device *mdd = mdo2mdd(obj);
2144 struct thandle *handle;
2148 /* XXX: this code won't be used ever:
2149 * DNE uses slightly different approach */
2152 handle = mdd_trans_create(env, mdd);
2156 rc = mdd_trans_start(env, mdd, handle);
2158 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
2159 rc = mdd_link_sanity_check(env, NULL, NULL, mdd_obj);
2161 __mdd_ref_add(env, mdd_obj, handle);
2162 mdd_write_unlock(env, mdd_obj);
2164 LASSERT(ma->ma_attr.la_valid & LA_CTIME);
2165 la_copy->la_ctime = ma->ma_attr.la_ctime;
2167 la_copy->la_valid = LA_CTIME;
2168 rc = mdd_attr_check_set_internal_locked(env, mdd_obj, la_copy,
2171 mdd_trans_stop(env, mdd, 0, handle);
2177 * do NOT or the MAY_*'s, you'll get the weakest
2179 int accmode(const struct lu_env *env, struct lu_attr *la, int flags)
2183 /* Sadly, NFSD reopens a file repeatedly during operation, so the
2184 * "acc_mode = 0" allowance for newly-created files isn't honoured.
2185 * NFSD uses the MDS_OPEN_OWNEROVERRIDE flag to say that a file
2186 * owner can write to a file even if it is marked readonly to hide
2187 * its brokenness. (bug 5781) */
2188 if (flags & MDS_OPEN_OWNEROVERRIDE) {
2189 struct md_ucred *uc = md_ucred(env);
2191 if ((uc == NULL) || (uc->mu_valid == UCRED_INIT) ||
2192 (la->la_uid == uc->mu_fsuid))
2196 if (flags & FMODE_READ)
2198 if (flags & (FMODE_WRITE | MDS_OPEN_TRUNC | MDS_OPEN_APPEND))
2200 if (flags & MDS_FMODE_EXEC)
2205 static int mdd_open_sanity_check(const struct lu_env *env,
2206 struct mdd_object *obj, int flag)
2208 struct lu_attr *tmp_la = &mdd_env_info(env)->mti_la;
2213 if (mdd_is_dead_obj(obj))
2216 rc = mdd_la_get(env, obj, tmp_la, BYPASS_CAPA);
2220 if (S_ISLNK(tmp_la->la_mode))
2223 mode = accmode(env, tmp_la, flag);
2225 if (S_ISDIR(tmp_la->la_mode) && (mode & MAY_WRITE))
2228 if (!(flag & MDS_OPEN_CREATED)) {
2229 rc = mdd_permission_internal(env, obj, tmp_la, mode);
2234 if (S_ISFIFO(tmp_la->la_mode) || S_ISSOCK(tmp_la->la_mode) ||
2235 S_ISBLK(tmp_la->la_mode) || S_ISCHR(tmp_la->la_mode))
2236 flag &= ~MDS_OPEN_TRUNC;
2238 /* For writing append-only file must open it with append mode. */
2239 if (mdd_is_append(obj)) {
2240 if ((flag & FMODE_WRITE) && !(flag & MDS_OPEN_APPEND))
2242 if (flag & MDS_OPEN_TRUNC)
2248 * Now, flag -- O_NOATIME does not be packed by client.
2250 if (flag & O_NOATIME) {
2251 struct md_ucred *uc = md_ucred(env);
2253 if (uc && ((uc->mu_valid == UCRED_OLD) ||
2254 (uc->mu_valid == UCRED_NEW)) &&
2255 (uc->mu_fsuid != tmp_la->la_uid) &&
2256 !mdd_capable(uc, CFS_CAP_FOWNER))
2264 static int mdd_open(const struct lu_env *env, struct md_object *obj,
2267 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2270 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
2272 rc = mdd_open_sanity_check(env, mdd_obj, flags);
2274 mdd_obj->mod_count++;
2276 mdd_write_unlock(env, mdd_obj);
2280 int mdd_declare_object_kill(const struct lu_env *env, struct mdd_object *obj,
2281 struct md_attr *ma, struct thandle *handle)
2285 rc = mdd_declare_unlink_log(env, obj, ma, handle);
2289 return mdo_declare_destroy(env, obj, handle);
2292 /* return md_attr back,
2293 * if it is last unlink then return lov ea + llog cookie*/
2294 int mdd_object_kill(const struct lu_env *env, struct mdd_object *obj,
2295 struct md_attr *ma, struct thandle *handle)
2300 if (S_ISREG(mdd_object_type(obj))) {
2301 /* Return LOV & COOKIES unconditionally here. We clean evth up.
2302 * Caller must be ready for that. */
2304 rc = __mdd_lmm_get(env, obj, ma);
2305 if ((ma->ma_valid & MA_LOV))
2306 rc = mdd_unlink_log(env, mdo2mdd(&obj->mod_obj),
2311 rc = mdo_destroy(env, obj, handle);
2316 static int mdd_declare_close(const struct lu_env *env,
2317 struct mdd_object *obj,
2319 struct thandle *handle)
2323 rc = orph_declare_index_delete(env, obj, handle);
2327 return mdd_declare_object_kill(env, obj, ma, handle);
2331 * No permission check is needed.
2333 static int mdd_close(const struct lu_env *env, struct md_object *obj,
2334 struct md_attr *ma, int mode)
2336 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2337 struct mdd_device *mdd = mdo2mdd(obj);
2338 struct thandle *handle = NULL;
2342 #ifdef HAVE_QUOTA_SUPPORT
2343 struct obd_device *obd = mdo2mdd(obj)->mdd_obd_dev;
2344 struct mds_obd *mds = &obd->u.mds;
2345 unsigned int qids[MAXQUOTAS] = { 0, 0 };
2350 if (ma->ma_valid & MA_FLAGS && ma->ma_attr_flags & MDS_KEEP_ORPHAN) {
2351 mdd_obj->mod_count--;
2353 if (mdd_obj->mod_flags & ORPHAN_OBJ && !mdd_obj->mod_count)
2354 CDEBUG(D_HA, "Object "DFID" is retained in orphan "
2355 "list\n", PFID(mdd_object_fid(mdd_obj)));
2359 /* check without any lock */
2360 if (mdd_obj->mod_count == 1 &&
2361 (mdd_obj->mod_flags & (ORPHAN_OBJ | DEAD_OBJ)) != 0) {
2363 handle = mdd_trans_create(env, mdo2mdd(obj));
2365 RETURN(PTR_ERR(handle));
2367 rc = mdd_declare_close(env, mdd_obj, ma, handle);
2371 rc = mdd_declare_changelog_store(env, mdd, NULL, handle);
2375 rc = mdd_trans_start(env, mdo2mdd(obj), handle);
2380 mdd_write_lock(env, mdd_obj, MOR_TGT_CHILD);
2381 if (handle == NULL && mdd_obj->mod_count == 1 &&
2382 (mdd_obj->mod_flags & ORPHAN_OBJ) != 0) {
2383 mdd_write_unlock(env, mdd_obj);
2387 /* release open count */
2388 mdd_obj->mod_count --;
2390 if (mdd_obj->mod_count == 0 && mdd_obj->mod_flags & ORPHAN_OBJ) {
2391 /* remove link to object from orphan index */
2392 LASSERT(handle != NULL);
2393 rc = __mdd_orphan_del(env, mdd_obj, handle);
2395 CDEBUG(D_HA, "Object "DFID" is deleted from orphan "
2396 "list, OSS objects to be destroyed.\n",
2397 PFID(mdd_object_fid(mdd_obj)));
2399 CERROR("Object "DFID" can not be deleted from orphan "
2400 "list, maybe cause OST objects can not be "
2401 "destroyed (err: %d).\n",
2402 PFID(mdd_object_fid(mdd_obj)), rc);
2403 /* If object was not deleted from orphan list, do not
2404 * destroy OSS objects, which will be done when next
2410 rc = mdd_iattr_get(env, mdd_obj, ma);
2411 /* Object maybe not in orphan list originally, it is rare case for
2412 * mdd_finish_unlink() failure. */
2413 if (rc == 0 && ma->ma_attr.la_nlink == 0) {
2414 #ifdef HAVE_QUOTA_SUPPORT
2415 if (mds->mds_quota) {
2416 quota_opc = FSFILT_OP_UNLINK_PARTIAL_CHILD;
2417 mdd_quota_wrapper(&ma->ma_attr, qids);
2420 /* MDS_CLOSE_CLEANUP means destroy OSS objects by MDS. */
2421 if (ma->ma_valid & MA_FLAGS &&
2422 ma->ma_attr_flags & MDS_CLOSE_CLEANUP) {
2423 rc = mdd_lov_destroy(env, mdd, mdd_obj, &ma->ma_attr);
2425 if (handle == NULL) {
2426 handle = mdd_trans_create(env, mdo2mdd(obj));
2428 GOTO(out, rc = PTR_ERR(handle));
2430 rc = mdd_declare_object_kill(env, mdd_obj, ma,
2435 rc = mdd_declare_changelog_store(env, mdd,
2440 rc = mdd_trans_start(env, mdo2mdd(obj), handle);
2445 rc = mdd_object_kill(env, mdd_obj, ma, handle);
2451 CERROR("Error when prepare to delete Object "DFID" , "
2452 "which will cause OST objects can not be "
2453 "destroyed.\n", PFID(mdd_object_fid(mdd_obj)));
2459 ma->ma_valid &= ~(MA_LOV | MA_COOKIE);
2461 mdd_write_unlock(env, mdd_obj);
2464 (mode & (FMODE_WRITE | MDS_OPEN_APPEND | MDS_OPEN_TRUNC)) &&
2465 !(ma->ma_valid & MA_FLAGS && ma->ma_attr_flags & MDS_RECOV_OPEN)) {
2466 if (handle == NULL) {
2467 handle = mdd_trans_create(env, mdo2mdd(obj));
2469 GOTO(stop, rc = IS_ERR(handle));
2471 rc = mdd_declare_changelog_store(env, mdd, NULL,
2476 rc = mdd_trans_start(env, mdo2mdd(obj), handle);
2481 mdd_changelog_data_store(env, mdd, CL_CLOSE, mode,
2487 mdd_trans_stop(env, mdd, rc, handle);
2488 #ifdef HAVE_QUOTA_SUPPORT
2490 /* Trigger dqrel on the owner of child. If failed,
2491 * the next call for lquota_chkquota will process it */
2492 lquota_adjust(mds_quota_interface_ref, obd, qids, 0, rc,
2499 * Permission check is done when open,
2500 * no need check again.
2502 static int mdd_readpage_sanity_check(const struct lu_env *env,
2503 struct mdd_object *obj)
2505 struct dt_object *next = mdd_object_child(obj);
2509 if (S_ISDIR(mdd_object_type(obj)) && dt_try_as_dir(env, next))
2517 static int mdd_dir_page_build(const struct lu_env *env, struct mdd_device *mdd,
2518 struct lu_dirpage *dp, int nob,
2519 const struct dt_it_ops *iops, struct dt_it *it,
2525 struct lu_dirent *ent;
2526 struct lu_dirent *last = NULL;
2529 memset(area, 0, sizeof (*dp));
2530 area += sizeof (*dp);
2531 nob -= sizeof (*dp);
2538 len = iops->key_size(env, it);
2540 /* IAM iterator can return record with zero len. */
2544 hash = iops->store(env, it);
2545 if (unlikely(first)) {
2547 dp->ldp_hash_start = cpu_to_le64(hash);
2550 /* calculate max space required for lu_dirent */
2551 recsize = lu_dirent_calc_size(len, attr);
2553 if (nob >= recsize) {
2554 result = iops->rec(env, it, (struct dt_rec *)ent, attr);
2555 if (result == -ESTALE)
2560 /* osd might not able to pack all attributes,
2561 * so recheck rec length */
2562 recsize = le16_to_cpu(ent->lde_reclen);
2564 result = (last != NULL) ? 0 :-EINVAL;
2568 ent = (void *)ent + recsize;
2572 result = iops->next(env, it);
2573 if (result == -ESTALE)
2575 } while (result == 0);
2578 dp->ldp_hash_end = cpu_to_le64(hash);
2580 if (last->lde_hash == dp->ldp_hash_end)
2581 dp->ldp_flags |= cpu_to_le32(LDF_COLLIDE);
2582 last->lde_reclen = 0; /* end mark */
2587 static int __mdd_readpage(const struct lu_env *env, struct mdd_object *obj,
2588 const struct lu_rdpg *rdpg)
2591 struct dt_object *next = mdd_object_child(obj);
2592 const struct dt_it_ops *iops;
2594 struct mdd_device *mdd = mdo2mdd(&obj->mod_obj);
2600 LASSERT(rdpg->rp_pages != NULL);
2601 LASSERT(next->do_index_ops != NULL);
2603 if (rdpg->rp_count <= 0)
2607 * iterate through directory and fill pages from @rdpg
2609 iops = &next->do_index_ops->dio_it;
2610 it = iops->init(env, next, rdpg->rp_attrs, mdd_object_capa(env, obj));
2614 rc = iops->load(env, it, rdpg->rp_hash);
2618 * Iterator didn't find record with exactly the key requested.
2620 * It is currently either
2622 * - positioned above record with key less than
2623 * requested---skip it.
2625 * - or not positioned at all (is in IAM_IT_SKEWED
2626 * state)---position it on the next item.
2628 rc = iops->next(env, it);
2633 * At this point and across for-loop:
2635 * rc == 0 -> ok, proceed.
2636 * rc > 0 -> end of directory.
2639 for (i = 0, nob = rdpg->rp_count; rc == 0 && nob > 0;
2640 i++, nob -= CFS_PAGE_SIZE) {
2641 struct lu_dirpage *dp;
2643 LASSERT(i < rdpg->rp_npages);
2644 pg = rdpg->rp_pages[i];
2646 #if CFS_PAGE_SIZE > LU_PAGE_SIZE
2649 rc = mdd_dir_page_build(env, mdd, dp,
2650 min_t(int, nob, LU_PAGE_SIZE),
2651 iops, it, rdpg->rp_attrs);
2656 dp->ldp_hash_end = cpu_to_le64(MDS_DIR_END_OFF);
2658 } else if (rc < 0) {
2659 CWARN("build page failed: %d!\n", rc);
2662 #if CFS_PAGE_SIZE > LU_PAGE_SIZE
2663 dp = (struct lu_dirpage *)((char *)dp + LU_PAGE_SIZE);
2664 if ((unsigned long)dp & ~CFS_PAGE_MASK)
2671 struct lu_dirpage *dp;
2673 dp = cfs_kmap(rdpg->rp_pages[0]);
2674 dp->ldp_hash_start = cpu_to_le64(rdpg->rp_hash);
2677 * No pages were processed, mark this for first page
2680 dp->ldp_flags = cpu_to_le32(LDF_EMPTY);
2683 cfs_kunmap(rdpg->rp_pages[0]);
2685 rc = min_t(unsigned int, nlupgs * LU_PAGE_SIZE, rdpg->rp_count);
2688 iops->fini(env, it);
2693 int mdd_readpage(const struct lu_env *env, struct md_object *obj,
2694 const struct lu_rdpg *rdpg)
2696 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2700 LASSERT(mdd_object_exists(mdd_obj));
2702 mdd_read_lock(env, mdd_obj, MOR_TGT_CHILD);
2703 rc = mdd_readpage_sanity_check(env, mdd_obj);
2705 GOTO(out_unlock, rc);
2707 if (mdd_is_dead_obj(mdd_obj)) {
2709 struct lu_dirpage *dp;
2712 * According to POSIX, please do not return any entry to client:
2713 * even dot and dotdot should not be returned.
2715 CWARN("readdir from dead object: "DFID"\n",
2716 PFID(mdd_object_fid(mdd_obj)));
2718 if (rdpg->rp_count <= 0)
2719 GOTO(out_unlock, rc = -EFAULT);
2720 LASSERT(rdpg->rp_pages != NULL);
2722 pg = rdpg->rp_pages[0];
2723 dp = (struct lu_dirpage*)cfs_kmap(pg);
2724 memset(dp, 0 , sizeof(struct lu_dirpage));
2725 dp->ldp_hash_start = cpu_to_le64(rdpg->rp_hash);
2726 dp->ldp_hash_end = cpu_to_le64(MDS_DIR_END_OFF);
2727 dp->ldp_flags = cpu_to_le32(LDF_EMPTY);
2729 GOTO(out_unlock, rc = LU_PAGE_SIZE);
2732 rc = __mdd_readpage(env, mdd_obj, rdpg);
2736 mdd_read_unlock(env, mdd_obj);
2740 static int mdd_object_sync(const struct lu_env *env, struct md_object *obj)
2742 struct mdd_object *mdd_obj = md2mdd_obj(obj);
2743 struct dt_object *next;
2745 LASSERT(mdd_object_exists(mdd_obj));
2746 next = mdd_object_child(mdd_obj);
2747 return next->do_ops->do_object_sync(env, next);
2750 const struct md_object_operations mdd_obj_ops = {
2751 .moo_permission = mdd_permission,
2752 .moo_attr_get = mdd_attr_get,
2753 .moo_attr_set = mdd_attr_set,
2754 .moo_xattr_get = mdd_xattr_get,
2755 .moo_xattr_set = mdd_xattr_set,
2756 .moo_xattr_list = mdd_xattr_list,
2757 .moo_xattr_del = mdd_xattr_del,
2758 .moo_object_create = mdd_object_create,
2759 .moo_ref_add = mdd_ref_add,
2760 .moo_ref_del = mdd_ref_del,
2761 .moo_open = mdd_open,
2762 .moo_close = mdd_close,
2763 .moo_readpage = mdd_readpage,
2764 .moo_readlink = mdd_readlink,
2765 .moo_changelog = mdd_changelog,
2766 .moo_capa_get = mdd_capa_get,
2767 .moo_object_sync = mdd_object_sync,
2768 .moo_path = mdd_path,
2769 .moo_file_lock = mdd_file_lock,
2770 .moo_file_unlock = mdd_file_unlock,