lustre/obdclass/llog_osd.c

   1 /*
   2  * GPL HEADER START
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 only,
   8  * as published by the Free Software Foundation.
   9  *
  10  * This program is distributed in the hope that it will be useful, but
  11  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * General Public License version 2 for more details (a copy is included
  14  * in the LICENSE file that accompanied this code).
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * version 2 along with this program; If not, see
  18  * http://www.gnu.org/licenses/gpl-2.0.html
  19  *
  20  * GPL HEADER END
  21  */
  22 /*
  23  * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
  24  * Use is subject to license terms.
  25  *
  26  * Copyright (c) 2012, 2017, Intel Corporation.
  27  */
  28 /*
  29  * This file is part of Lustre, http://www.lustre.org/
  30  */
  31 /*
  32  * lustre/obdclass/llog_osd.c
  33  *
  34  * Low level llog routines on top of OSD API
  35  *
  36  * This file provides set of methods for llog operations on top of
  37  * dt_device. It contains all supported llog_operations interfaces and
  38  * supplimental functions.
  39  *
  40  * Author: Alexey Zhuravlev <alexey.zhuravlev@intel.com>
  41  * Author: Mikhail Pershin <mike.pershin@intel.com>
  42  */
  43
  44 #define DEBUG_SUBSYSTEM S_LOG
  45
  46 #include <linux/delay.h>
  47
  48 #include <dt_object.h>
  49 #include <llog_swab.h>
  50 #include <lustre_fid.h>
  51 #include <obd.h>
  52 #include <obd_class.h>
  53
  54 #include "llog_internal.h"
  55 #include "local_storage.h"
  56
  57 /**
  58  * Implementation of the llog_operations::lop_declare_create
  59  *
  60  * This function is a wrapper over local_storage API function
  61  * local_object_declare_create().
  62  *
  63  * \param[in] env       execution environment
  64  * \param[in] los       local_storage for bottom storage device
  65  * \param[in] o         dt_object to create
  66  * \param[in] th        current transaction handle
  67  *
  68  * \retval              0 on successful declaration of the new object
  69  * \retval              negative error if declaration was failed
  70  */
  71 static int llog_osd_declare_new_object(const struct lu_env *env,
  72                                        struct local_oid_storage *los,
  73                                        struct dt_object *o,
  74                                        struct thandle *th)
  75 {
  76         struct llog_thread_info *lgi = llog_info(env);
  77
  78         lgi->lgi_attr.la_valid = LA_MODE;
  79         lgi->lgi_attr.la_mode = S_IFREG | S_IRUGO | S_IWUSR;
  80         lgi->lgi_dof.dof_type = dt_mode_to_dft(S_IFREG);
  81
  82         return local_object_declare_create(env, los, o, &lgi->lgi_attr,
  83                                            &lgi->lgi_dof, th);
  84 }
  85
  86 /**
  87  * Implementation of the llog_operations::lop_create
  88  *
  89  * This function is a wrapper over local_storage API function
  90  * local_object_create().
  91  *
  92  * \param[in] env       execution environment
  93  * \param[in] los       local_storage for bottom storage device
  94  * \param[in] o         dt_object to create
  95  * \param[in] th        current transaction handle
  96  *
  97  * \retval              0 on successful creation of the new object
  98  * \retval              negative error if creation was failed
  99  */
 100 static int llog_osd_create_new_object(const struct lu_env *env,
 101                                       struct local_oid_storage *los,
 102                                       struct dt_object *o,
 103                                       struct thandle *th)
 104 {
 105         struct llog_thread_info *lgi = llog_info(env);
 106
 107         lgi->lgi_attr.la_valid = LA_MODE;
 108         lgi->lgi_attr.la_mode = S_IFREG | S_IRUGO | S_IWUSR;
 109         lgi->lgi_dof.dof_type = dt_mode_to_dft(S_IFREG);
 110
 111         return local_object_create(env, los, o, &lgi->lgi_attr,
 112                                    &lgi->lgi_dof, th);
 113 }
 114
 115 /**
 116  * Implementation of the llog_operations::lop_exist
 117  *
 118  * This function checks that llog exists on storage.
 119  *
 120  * \param[in] handle    llog handle of the current llog
 121  *
 122  * \retval              true if llog object exists and is not just destroyed
 123  * \retval              false if llog doesn't exist or just destroyed
 124  */
 125 static int llog_osd_exist(struct llog_handle *handle)
 126 {
 127         LASSERT(handle->lgh_obj);
 128         return dt_object_exists(handle->lgh_obj) && !handle->lgh_destroyed;
 129 }
 130
 131 static void *rec_tail(struct llog_rec_hdr *rec)
 132 {
 133         return (void *)((char *)rec + rec->lrh_len -
 134                         sizeof(struct llog_rec_tail));
 135 }
 136
 137 /**
 138  * Write a padding record to the llog
 139  *
 140  * This function writes a padding record to the end of llog. That may
 141  * be needed if llog contains records of variable size, e.g. config logs
 142  * or changelogs.
 143  * The padding record just aligns llog to the llog chunk_size boundary if
 144  * the current record doesn't fit in the remaining space.
 145  *
 146  * It allocates full length to avoid two separate writes for header and tail.
 147  * Such 2-steps scheme needs extra protection and complex error handling.
 148  *
 149  * \param[in]     env   execution environment
 150  * \param[in]     o     dt_object to create
 151  * \param[in,out] off   pointer to the padding start offset
 152  * \param[in]     len   padding length
 153  * \param[in]     index index of the padding record in a llog
 154  * \param[in]     th    current transaction handle
 155  *
 156  * \retval              0 on successful padding write
 157  * \retval              negative error if write failed
 158  */
 159 static int llog_osd_pad(const struct lu_env *env, struct dt_object *o,
 160                         loff_t *off, int len, int index, struct thandle *th)
 161 {
 162         struct llog_thread_info *lgi = llog_info(env);
 163         struct llog_rec_hdr     *rec;
 164         struct llog_rec_tail    *tail;
 165         int                      rc;
 166
 167         ENTRY;
 168
 169         LASSERT(th);
 170         LASSERT(off);
 171         LASSERT(len >= LLOG_MIN_REC_SIZE && (len & 0x7) == 0);
 172
 173         OBD_ALLOC(rec, len);
 174         if (rec == NULL)
 175                 RETURN(-ENOMEM);
 176
 177         rec->lrh_len = len;
 178         rec->lrh_index = index;
 179         rec->lrh_type = LLOG_PAD_MAGIC;
 180
 181         tail = rec_tail(rec);
 182         tail->lrt_len = len;
 183         tail->lrt_index = index;
 184
 185         lgi->lgi_buf.lb_buf = rec;
 186         lgi->lgi_buf.lb_len = len;
 187         rc = dt_record_write(env, o, &lgi->lgi_buf, off, th);
 188         if (rc)
 189                 CERROR("%s: error writing padding record: rc = %d\n",
 190                        o->do_lu.lo_dev->ld_obd->obd_name, rc);
 191
 192         OBD_FREE(rec, len);
 193         RETURN(rc);
 194 }
 195
 196 /**
 197  * Implementation of the llog_operations::lop_read_header
 198  *
 199  * This function reads the current llog header from the bottom storage
 200  * device.
 201  *
 202  * \param[in] env       execution environment
 203  * \param[in] handle    llog handle of the current llog
 204  *
 205  * \retval              0 on successful header read
 206  * \retval              negative error if read failed
 207  */
 208 static int llog_osd_read_header(const struct lu_env *env,
 209                                 struct llog_handle *handle)
 210 {
 211         struct llog_rec_hdr *llh_hdr;
 212         struct dt_object *o;
 213         struct llog_thread_info *lgi;
 214         enum llog_flag flags;
 215         int rc;
 216
 217         ENTRY;
 218
 219         o = handle->lgh_obj;
 220         LASSERT(o);
 221
 222         lgi = llog_info(env);
 223
 224         dt_read_lock(env, o, 0);
 225
 226         rc = dt_attr_get(env, o, &lgi->lgi_attr);
 227         if (rc)
 228                 GOTO(unlock, rc = -EIO);
 229
 230         LASSERT(lgi->lgi_attr.la_valid & LA_SIZE);
 231
 232         if (lgi->lgi_attr.la_size == 0) {
 233                 CDEBUG(D_HA, "not reading header from 0-byte log\n");
 234                 GOTO(unlock, rc = LLOG_EEMPTY);
 235         }
 236
 237         flags = handle->lgh_hdr->llh_flags;
 238
 239         lgi->lgi_off = 0;
 240         lgi->lgi_buf.lb_buf = handle->lgh_hdr;
 241         lgi->lgi_buf.lb_len = handle->lgh_hdr_size;
 242         rc = dt_read(env, o, &lgi->lgi_buf, &lgi->lgi_off);
 243         llh_hdr = &handle->lgh_hdr->llh_hdr;
 244         if (rc < 0) {
 245                 CERROR("%s: can't read llog "DFID" header: rc = %d\n",
 246                        o->do_lu.lo_dev->ld_obd->obd_name,
 247                        PFID(lu_object_fid(&o->do_lu)), rc);
 248                 GOTO(unlock, rc = -EIO);
 249         }
 250         if (rc < sizeof(*llh_hdr) || rc < LLOG_MIN_CHUNK_SIZE) {
 251                 /* consider short header as non-initialized llog */
 252                 CERROR("%s: llog "DFID" header too small: rc = %d\n",
 253                        o->do_lu.lo_dev->ld_obd->obd_name,
 254                        PFID(lu_object_fid(&o->do_lu)), rc);
 255                 /* caller flags to be initialized */
 256                 handle->lgh_hdr->llh_flags = flags;
 257                 GOTO(unlock, rc = LLOG_EEMPTY);
 258         }
 259
 260         if (LLOG_REC_HDR_NEEDS_SWABBING(llh_hdr))
 261                 lustre_swab_llog_hdr(handle->lgh_hdr);
 262
 263         if (llh_hdr->lrh_type != LLOG_HDR_MAGIC) {
 264                 CERROR("%s: bad log %s "DFID" header magic: %#x "
 265                        "(expected %#x)\n", o->do_lu.lo_dev->ld_obd->obd_name,
 266                        handle->lgh_name ? handle->lgh_name : "",
 267                        PFID(lu_object_fid(&o->do_lu)),
 268                        llh_hdr->lrh_type, LLOG_HDR_MAGIC);
 269                 GOTO(unlock, rc = -EINVAL);
 270         } else if (llh_hdr->lrh_len < LLOG_MIN_CHUNK_SIZE ||
 271                    llh_hdr->lrh_len > handle->lgh_hdr_size) {
 272                 CERROR("%s: incorrectly sized log %s "DFID" header: "
 273                        "%#x (expected at least %#x)\n"
 274                        "you may need to re-run lconf --write_conf.\n",
 275                        o->do_lu.lo_dev->ld_obd->obd_name,
 276                        handle->lgh_name ? handle->lgh_name : "",
 277                        PFID(lu_object_fid(&o->do_lu)),
 278                        llh_hdr->lrh_len, LLOG_MIN_CHUNK_SIZE);
 279                 GOTO(unlock, rc = -EINVAL);
 280         } else if (LLOG_HDR_TAIL(handle->lgh_hdr)->lrt_index >
 281                    LLOG_HDR_BITMAP_SIZE(handle->lgh_hdr) ||
 282                    LLOG_HDR_TAIL(handle->lgh_hdr)->lrt_len !=
 283                         llh_hdr->lrh_len) {
 284                 CERROR("%s: incorrectly sized log %s "DFID" tailer: "
 285                        "%#x : rc = %d\n",
 286                        o->do_lu.lo_dev->ld_obd->obd_name,
 287                        handle->lgh_name ? handle->lgh_name : "",
 288                        PFID(lu_object_fid(&o->do_lu)),
 289                        LLOG_HDR_TAIL(handle->lgh_hdr)->lrt_len, -EIO);
 290                 GOTO(unlock, rc = -EINVAL);
 291         }
 292
 293         handle->lgh_hdr->llh_flags |= (flags & LLOG_F_EXT_MASK);
 294         handle->lgh_last_idx = LLOG_HDR_TAIL(handle->lgh_hdr)->lrt_index;
 295         rc = 0;
 296
 297 unlock:
 298         dt_read_unlock(env, o);
 299         RETURN(rc);
 300 }
 301
 302 /**
 303  * Implementation of the llog_operations::lop_declare_write
 304  *
 305  * This function declares the new record write.
 306  *
 307  * \param[in] env       execution environment
 308  * \param[in] loghandle llog handle of the current llog
 309  * \param[in] rec       llog record header. This is a real header of the full
 310  *                      llog record to write. This is the beginning of buffer
 311  *                      to write, the length of buffer is stored in
 312  *                      \a rec::lrh_len
 313  * \param[in] idx       index of the llog record. If \a idx == -1 then this is
 314  *                      append case, otherwise \a idx is the index of record
 315  *                      to modify
 316  * \param[in] th        current transaction handle
 317  *
 318  * \retval              0 on successful declaration
 319  * \retval              negative error if declaration failed
 320  */
 321 static int llog_osd_declare_write_rec(const struct lu_env *env,
 322                                       struct llog_handle *loghandle,
 323                                       struct llog_rec_hdr *rec,
 324                                       int idx, struct thandle *th)
 325 {
 326         struct llog_thread_info *lgi = llog_info(env);
 327         __u32                   chunk_size;
 328         struct dt_object        *o;
 329         int                      rc;
 330
 331         ENTRY;
 332
 333         LASSERT(env);
 334         LASSERT(th);
 335         LASSERT(loghandle);
 336         LASSERT(rec);
 337         LASSERT(rec->lrh_len <= loghandle->lgh_ctxt->loc_chunk_size);
 338
 339         o = loghandle->lgh_obj;
 340         LASSERT(o);
 341
 342         chunk_size = loghandle->lgh_ctxt->loc_chunk_size;
 343         lgi->lgi_buf.lb_len = chunk_size;
 344         lgi->lgi_buf.lb_buf = NULL;
 345         /* each time we update header */
 346         rc = dt_declare_record_write(env, o, &lgi->lgi_buf, 0, th);
 347         if (rc || idx == 0) /* if error or just header */
 348                 RETURN(rc);
 349
 350         /**
 351          * the pad record can be inserted so take into account double
 352          * record size: pad and the actual record into a new block
 353          */
 354         lgi->lgi_buf.lb_len = rec->lrh_len * 2;
 355         lgi->lgi_buf.lb_buf = NULL;
 356         /* XXX: implement declared window or multi-chunks approach */
 357         rc = dt_declare_record_write(env, o, &lgi->lgi_buf, -1, th);
 358
 359         RETURN(rc);
 360 }
 361
 362 /**
 363  * Implementation of the llog_operations::lop_write
 364  *
 365  * This function writes the new record in the llog or modify the existed one.
 366  *
 367  * \param[in]  env              execution environment
 368  * \param[in]  loghandle        llog handle of the current llog
 369  * \param[in]  rec              llog record header. This is a real header of
 370  *                              the full llog record to write. This is
 371  *                              the beginning of buffer to write, the length
 372  *                              of buffer is stored in \a rec::lrh_len
 373  * \param[in,out] reccookie     pointer to the cookie to return back if needed.
 374  *                              It is used for further cancel of this llog
 375  *                              record.
 376  * \param[in]  idx              index of the llog record. If \a idx == -1 then
 377  *                              this is append case, otherwise \a idx is
 378  *                              the index of record to modify
 379  * \param[in]  th               current transaction handle
 380  *
 381  * \retval                      0 on successful write && \a reccookie == NULL
 382  *                              1 on successful write && \a reccookie != NULL
 383  * \retval                      negative error if write failed
 384  */
 385 static int llog_osd_write_rec(const struct lu_env *env,
 386                               struct llog_handle *loghandle,
 387                               struct llog_rec_hdr *rec,
 388                               struct llog_cookie *reccookie,
 389                               int idx, struct thandle *th)
 390 {
 391         struct llog_thread_info *lgi = llog_info(env);
 392         struct llog_log_hdr *llh;
 393         int reclen = rec->lrh_len;
 394         int index, rc;
 395         struct llog_rec_tail *lrt;
 396         struct dt_object *o;
 397         __u32 chunk_size;
 398         size_t left;
 399         __u32 orig_last_idx;
 400         bool pad = false;
 401         ENTRY;
 402
 403         llh = loghandle->lgh_hdr;
 404         o = loghandle->lgh_obj;
 405
 406         chunk_size = llh->llh_hdr.lrh_len;
 407         CDEBUG(D_OTHER, "new record %x to "DFID"\n",
 408                rec->lrh_type, PFID(lu_object_fid(&o->do_lu)));
 409
 410         if (!llog_osd_exist(loghandle))
 411                 RETURN(-ENOENT);
 412
 413         /* record length should not bigger than  */
 414         if (reclen > loghandle->lgh_hdr->llh_hdr.lrh_len)
 415                 RETURN(-E2BIG);
 416
 417         /* sanity check for fixed-records llog */
 418         if (idx != LLOG_HEADER_IDX && (llh->llh_flags & LLOG_F_IS_FIXSIZE)) {
 419                 LASSERT(llh->llh_size != 0);
 420                 LASSERT(llh->llh_size == reclen);
 421         }
 422
 423         /* return error if osp object is stale */
 424         if (idx != LLOG_HEADER_IDX && dt_object_stale(o))
 425                 RETURN(-ESTALE);
 426         rc = dt_attr_get(env, o, &lgi->lgi_attr);
 427         if (rc)
 428                 RETURN(rc);
 429
 430         /**
 431          * The modification case.
 432          * If idx set then the record with that index must be modified.
 433          * There are three cases possible:
 434          * 1) the common case is the llog header update (idx == 0)
 435          * 2) the llog record modification during llog process.
 436          *    This is indicated by the \a loghandle::lgh_cur_idx > 0.
 437          *    In that case the \a loghandle::lgh_cur_offset
 438          * 3) otherwise this is assumed that llog consist of records of
 439          *    fixed size, i.e. catalog. The llog header must has llh_size
 440          *    field equal to record size. The record offset is calculated
 441          *    just by /a idx value
 442          *
 443          * During modification we don't need extra header update because
 444          * the bitmap and record count are not changed. The record header
 445          * and tail remains the same too.
 446          */
 447         if (idx != LLOG_NEXT_IDX) {
 448                 /* llog can be empty only when first record is being written */
 449                 LASSERT(ergo(idx > 0, lgi->lgi_attr.la_size > 0));
 450
 451                 if (!test_bit_le(idx, LLOG_HDR_BITMAP(llh))) {
 452                         CERROR("%s: modify unset record %u\n",
 453                                o->do_lu.lo_dev->ld_obd->obd_name, idx);
 454                         RETURN(-ENOENT);
 455                 }
 456
 457                 if (idx != rec->lrh_index) {
 458                         CERROR("%s: modify index mismatch %d %u\n",
 459                                o->do_lu.lo_dev->ld_obd->obd_name, idx,
 460                                rec->lrh_index);
 461                         RETURN(-EFAULT);
 462                 }
 463
 464                 if (idx == LLOG_HEADER_IDX) {
 465                         /* llog header update */
 466                         __u32   *bitmap = LLOG_HDR_BITMAP(llh);
 467
 468                         lgi->lgi_off = 0;
 469
 470                         /* If it does not indicate the bitmap index
 471                          * (reccookie == NULL), then it means update
 472                          * the whole update header. Otherwise only
 473                          * update header and bits needs to be updated,
 474                          * and in DNE cases, it will signaficantly
 475                          * shrink the RPC size.
 476                          * see distribute_txn_cancel_records()*/
 477                         if (reccookie == NULL) {
 478                                 lgi->lgi_buf.lb_len = reclen;
 479                                 lgi->lgi_buf.lb_buf = rec;
 480                                 rc = dt_record_write(env, o, &lgi->lgi_buf,
 481                                                      &lgi->lgi_off, th);
 482                                 RETURN(rc);
 483                         }
 484
 485                         /* update the header */
 486                         lgi->lgi_buf.lb_len = llh->llh_bitmap_offset;
 487                         lgi->lgi_buf.lb_buf = llh;
 488                         rc = dt_record_write(env, o, &lgi->lgi_buf,
 489                                              &lgi->lgi_off, th);
 490                         if (rc != 0)
 491                                 RETURN(rc);
 492
 493                         /* update the bitmap */
 494                         index = reccookie->lgc_index;
 495                         lgi->lgi_off = llh->llh_bitmap_offset +
 496                                       (index / (sizeof(*bitmap) * 8)) *
 497                                                         sizeof(*bitmap);
 498                         lgi->lgi_buf.lb_len = sizeof(*bitmap);
 499                         lgi->lgi_buf.lb_buf =
 500                                         &bitmap[index/(sizeof(*bitmap)*8)];
 501                         rc = dt_record_write(env, o, &lgi->lgi_buf,
 502                                              &lgi->lgi_off, th);
 503
 504                         RETURN(rc);
 505                 } else if (llh->llh_flags & LLOG_F_IS_FIXSIZE) {
 506                         lgi->lgi_off = llh->llh_hdr.lrh_len +
 507                                        (idx - 1) * reclen;
 508                 } else if (reccookie != NULL && reccookie->lgc_index > 0) {
 509                         /**
 510                          * The lgc_offset can be used only if index is
 511                          * the same.
 512                          */
 513                         if (idx != reccookie->lgc_index) {
 514                                 CERROR("%s: modify index mismatch %d %d\n",
 515                                        o->do_lu.lo_dev->ld_obd->obd_name, idx,
 516                                        reccookie->lgc_index);
 517                                 RETURN(-EFAULT);
 518                         }
 519
 520                         lgi->lgi_off = reccookie->lgc_offset;
 521                         CDEBUG(D_OTHER, "modify record "DFID": idx:%u, "
 522                                "len:%u offset %llu\n",
 523                                PLOGID(&loghandle->lgh_id), idx,
 524                                rec->lrh_len, (long long)lgi->lgi_off);
 525                 } else {
 526                         /* This can be result of lgh_cur_idx is not set during
 527                          * llog processing or llh_size is not set to proper
 528                          * record size for fixed records llog. Therefore it is
 529                          * impossible to get record offset. */
 530                         CERROR("%s: can't get record offset, idx:%d, "
 531                                "len:%u.\n", o->do_lu.lo_dev->ld_obd->obd_name,
 532                                idx, rec->lrh_len);
 533                         RETURN(-EFAULT);
 534                 }
 535
 536                 /* update only data, header and tail remain the same */
 537                 lgi->lgi_off += sizeof(struct llog_rec_hdr);
 538                 lgi->lgi_buf.lb_len = REC_DATA_LEN(rec);
 539                 lgi->lgi_buf.lb_buf = REC_DATA(rec);
 540                 rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 541                 if (rc == 0 && reccookie) {
 542                         reccookie->lgc_lgl = loghandle->lgh_id;
 543                         reccookie->lgc_index = idx;
 544                         rc = 1;
 545                 }
 546                 RETURN(rc);
 547         }
 548
 549         /**
 550          * The append case.
 551          * The most common case of using llog. The new index is assigned to
 552          * the new record, new bit is set in llog bitmap and llog count is
 553          * incremented.
 554          *
 555          * Make sure that records don't cross a chunk boundary, so we can
 556          * process them page-at-a-time if needed.  If it will cross a chunk
 557          * boundary, write in a fake (but referenced) entry to pad the chunk.
 558          */
 559
 560
 561         /* simulate ENOSPC when new plain llog is being added to the
 562          * catalog */
 563         if (CFS_FAIL_CHECK(OBD_FAIL_MDS_LLOG_CREATE_FAILED2) &&
 564             llh->llh_flags & LLOG_F_IS_CAT)
 565                 RETURN(-ENOSPC);
 566
 567         LASSERT(lgi->lgi_attr.la_valid & LA_SIZE);
 568         orig_last_idx = loghandle->lgh_last_idx;
 569         lgi->lgi_off = lgi->lgi_attr.la_size;
 570
 571         if (loghandle->lgh_max_size > 0 &&
 572             lgi->lgi_off >= loghandle->lgh_max_size) {
 573                 CDEBUG(D_OTHER, "llog is getting too large (%u > %u) at %u "
 574                        DFID"\n", (unsigned)lgi->lgi_off,
 575                        loghandle->lgh_max_size, (int)loghandle->lgh_last_idx,
 576                        PLOGID(&loghandle->lgh_id));
 577                 /* this is to signal that this llog is full */
 578                 loghandle->lgh_last_idx = llog_max_idx(llh);
 579                 RETURN(-ENOSPC);
 580         }
 581
 582         left = chunk_size - (lgi->lgi_off & (chunk_size - 1));
 583         /* NOTE: padding is a record, but no bit is set */
 584         if (left != 0 && left != reclen &&
 585             left < (reclen + LLOG_MIN_REC_SIZE)) {
 586                 index = loghandle->lgh_last_idx + 1;
 587                 rc = llog_osd_pad(env, o, &lgi->lgi_off, left, index, th);
 588                 if (rc)
 589                         RETURN(rc);
 590
 591                 loghandle->lgh_last_idx++; /* for pad rec */
 592                 pad = true;
 593         }
 594         /* if it's the last idx in log file, then return -ENOSPC
 595          * or wrap around if a catalog */
 596         if (llog_is_full(loghandle)) {
 597                 if (llh->llh_flags & LLOG_F_IS_CAT)
 598                         loghandle->lgh_last_idx = 0;
 599                 else
 600                         RETURN(-ENOSPC);
 601         }
 602
 603         down_write(&loghandle->lgh_last_sem);
 604         /* increment the last_idx along with llh_tail index, they should
 605          * be equal for a llog lifetime */
 606         if (CFS_FAIL_CHECK(OBD_FAIL_LLOG_ADD_GAP) && --cfs_fail_val == 0)
 607                 loghandle->lgh_last_idx++;
 608         loghandle->lgh_last_idx++;
 609         index = loghandle->lgh_last_idx;
 610         LLOG_HDR_TAIL(llh)->lrt_index = index;
 611         /**
 612          * NB: the caller should make sure only 1 process access
 613          * the lgh_last_idx, e.g. append should be exclusive.
 614          * Otherwise it might hit the assert.
 615          */
 616         LASSERT(index < LLOG_HDR_BITMAP_SIZE(llh));
 617         rec->lrh_index = index;
 618         lrt = rec_tail(rec);
 619         lrt->lrt_len = rec->lrh_len;
 620         lrt->lrt_index = rec->lrh_index;
 621
 622         /* the lgh_hdr_mutex protects llog header data from concurrent
 623          * update/cancel, the llh_count and llh_bitmap are protected */
 624         mutex_lock(&loghandle->lgh_hdr_mutex);
 625         if (__test_and_set_bit_le(index, LLOG_HDR_BITMAP(llh))) {
 626                 CERROR("%s: index %u already set in llog bitmap "DFID"\n",
 627                        o->do_lu.lo_dev->ld_obd->obd_name, index,
 628                        PFID(lu_object_fid(&o->do_lu)));
 629                 mutex_unlock(&loghandle->lgh_hdr_mutex);
 630                 LBUG(); /* should never happen */
 631         }
 632         llh->llh_count++;
 633
 634         if (!(llh->llh_flags & LLOG_F_IS_FIXSIZE)) {
 635                 /* Update the minimum size of the llog record */
 636                 if (llh->llh_size == 0)
 637                         llh->llh_size = reclen;
 638                 else if (reclen < llh->llh_size)
 639                         llh->llh_size = reclen;
 640         }
 641
 642         /*
 643          * readers (e.g. llog_osd_read_header()) must not find
 644          * llog updated partially (bitmap/counter claims record,
 645          * but a record hasn't been added yet) as this results
 646          * in EIO.
 647          */
 648         dt_write_lock(env, o, 0);
 649
 650         if (lgi->lgi_attr.la_size == 0) {
 651                 lgi->lgi_off = 0;
 652                 lgi->lgi_buf.lb_len = llh->llh_hdr.lrh_len;
 653                 lgi->lgi_buf.lb_buf = &llh->llh_hdr;
 654                 rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 655                 if (rc != 0)
 656                         GOTO(out_unlock, rc);
 657         } else {
 658                 __u32   *bitmap = LLOG_HDR_BITMAP(llh);
 659
 660                 /* Note: If this is not initialization (size == 0), then do not
 661                  * write the whole header (8k bytes), only update header/tail
 662                  * and bits needs to be updated. Because this update might be
 663                  * part of cross-MDT operation, which needs to write these
 664                  * updates into the update log(32KB limit) and also pack inside
 665                  * the RPC (1MB limit), if we write 8K for each operation, which
 666                  * will cost a lot space, and keep us adding more updates to one
 667                  * update log.*/
 668                 lgi->lgi_off = 0;
 669                 lgi->lgi_buf.lb_len = llh->llh_bitmap_offset;
 670                 lgi->lgi_buf.lb_buf = &llh->llh_hdr;
 671                 rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 672                 if (rc != 0)
 673                         GOTO(out_unlock, rc);
 674
 675                 lgi->lgi_off = llh->llh_bitmap_offset +
 676                               (index / (sizeof(*bitmap) * 8)) * sizeof(*bitmap);
 677                 lgi->lgi_buf.lb_len = sizeof(*bitmap);
 678                 lgi->lgi_buf.lb_buf = &bitmap[index/(sizeof(*bitmap)*8)];
 679                 rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 680                 if (rc != 0)
 681                         GOTO(out_unlock, rc);
 682
 683                 lgi->lgi_off =  (unsigned long)LLOG_HDR_TAIL(llh) -
 684                                 (unsigned long)llh;
 685                 lgi->lgi_buf.lb_len = sizeof(llh->llh_tail);
 686                 lgi->lgi_buf.lb_buf = LLOG_HDR_TAIL(llh);
 687                 rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 688                 if (rc != 0)
 689                         GOTO(out_unlock, rc);
 690         }
 691         if (CFS_FAIL_PRECHECK(OBD_FAIL_LLOG_PAUSE_AFTER_PAD) && pad) {
 692                 /* a window for concurrent llog reader, see LU-12577 */
 693                 CFS_FAIL_TIMEOUT(OBD_FAIL_LLOG_PAUSE_AFTER_PAD,
 694                                  cfs_fail_val ?: 1);
 695         }
 696
 697 out_unlock:
 698         /* unlock here for remote object */
 699         mutex_unlock(&loghandle->lgh_hdr_mutex);
 700         if (rc) {
 701                 dt_write_unlock(env, o);
 702                 GOTO(out, rc);
 703         }
 704
 705         if (CFS_FAIL_PRECHECK(OBD_FAIL_LLOG_PROCESS_TIMEOUT) &&
 706            cfs_fail_val == (unsigned int)(loghandle->lgh_id.lgl_oi.oi.oi_id &
 707                                           0xFFFFFFFF)) {
 708                 CFS_RACE(OBD_FAIL_LLOG_PROCESS_TIMEOUT);
 709                 msleep(1 * MSEC_PER_SEC);
 710         }
 711         /* computed index can be used to determine offset for fixed-size
 712          * records. This also allows to handle Catalog wrap around case */
 713         if (llh->llh_flags & LLOG_F_IS_FIXSIZE) {
 714                 lgi->lgi_off = llh->llh_hdr.lrh_len + (index - 1) * reclen;
 715         } else {
 716                 rc = dt_attr_get(env, o, &lgi->lgi_attr);
 717                 if (rc) {
 718                         dt_write_unlock(env, o);
 719                         GOTO(out, rc);
 720                 }
 721
 722                 LASSERT(lgi->lgi_attr.la_valid & LA_SIZE);
 723                 lgi->lgi_off = max_t(__u64, lgi->lgi_attr.la_size,
 724                                      lgi->lgi_off);
 725         }
 726
 727         lgi->lgi_buf.lb_len = reclen;
 728         lgi->lgi_buf.lb_buf = rec;
 729         rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
 730
 731         dt_write_unlock(env, o);
 732         if (rc < 0)
 733                 GOTO(out, rc);
 734
 735         up_write(&loghandle->lgh_last_sem);
 736
 737         CDEBUG(D_HA, "added record "DFID".%u, %u off%llu\n",
 738                PFID(lu_object_fid(&o->do_lu)), index, rec->lrh_len,
 739                lgi->lgi_off);
 740         if (reccookie != NULL) {
 741                 reccookie->lgc_lgl = loghandle->lgh_id;
 742                 reccookie->lgc_index = index;
 743                 if ((rec->lrh_type == MDS_UNLINK_REC) ||
 744                     (rec->lrh_type == MDS_SETATTR64_REC))
 745                         reccookie->lgc_subsys = LLOG_MDS_OST_ORIG_CTXT;
 746                 else if (rec->lrh_type == OST_SZ_REC)
 747                         reccookie->lgc_subsys = LLOG_SIZE_ORIG_CTXT;
 748                 else
 749                         reccookie->lgc_subsys = -1;
 750                 rc = 1;
 751         }
 752         RETURN(rc);
 753 out:
 754         /* cleanup llog for error case */
 755         mutex_lock(&loghandle->lgh_hdr_mutex);
 756         clear_bit_le(index, LLOG_HDR_BITMAP(llh));
 757         llh->llh_count--;
 758         mutex_unlock(&loghandle->lgh_hdr_mutex);
 759
 760         /* restore llog last_idx */
 761         if (dt_object_remote(o)) {
 762                 loghandle->lgh_last_idx = orig_last_idx;
 763         } else if (--loghandle->lgh_last_idx == 0 &&
 764             (llh->llh_flags & LLOG_F_IS_CAT) && llh->llh_cat_idx != 0) {
 765                 /* catalog had just wrap-around case */
 766                 loghandle->lgh_last_idx = llog_max_idx(llh);
 767         }
 768
 769         LLOG_HDR_TAIL(llh)->lrt_index = loghandle->lgh_last_idx;
 770         up_write(&loghandle->lgh_last_sem);
 771
 772         RETURN(rc);
 773 }
 774
 775 /**
 776  * We can skip reading at least as many log blocks as the number of
 777  * minimum sized log records we are skipping.  If it turns out
 778  * that we are not far enough along the log (because the
 779  * actual records are larger than minimum size) we just skip
 780  * some more records.
 781  *
 782  * Note: in llog_process_thread, it will use bitmap offset as
 783  * the index to locate the record, which also includs some pad
 784  * records, whose record size is very small, and it also does not
 785  * consider pad record when recording minimum record size (otherwise
 786  * min_record size might be too small), so in some rare cases,
 787  * it might skip too much record for @goal, see llog_osd_next_block().
 788  *
 789  * When force_mini_rec is true, it means we have to use LLOG_MIN_REC_SIZE
 790  * as the min record size to skip over, usually because in the previous
 791  * try, it skip too much record, see loog_osd_next(prev)_block().
 792  */
 793 static inline void llog_skip_over(struct llog_handle *lgh, __u64 *off,
 794                                   int curr, int goal, __u32 chunk_size,
 795                                   bool force_mini_rec)
 796 {
 797         struct llog_log_hdr *llh = lgh->lgh_hdr;
 798
 799         /* Goal should not bigger than the record count */
 800         if (goal > lgh->lgh_last_idx)
 801                 goal = lgh->lgh_last_idx;
 802
 803         if (goal > curr) {
 804                 if (llh->llh_flags & LLOG_F_IS_FIXSIZE) {
 805                         *off = chunk_size + (goal - 1) * llh->llh_size;
 806                 } else {
 807                         __u64 min_rec_size = LLOG_MIN_REC_SIZE;
 808
 809                         if (llh->llh_size > 0 && !force_mini_rec)
 810                                 min_rec_size = llh->llh_size;
 811
 812                         *off = *off + (goal - curr - 1) * min_rec_size;
 813                 }
 814         }
 815         /* always align with lower chunk boundary*/
 816         *off &= ~(chunk_size - 1);
 817 }
 818
 819 /**
 820  * Remap a record to the desired format as specified by the crf flags.
 821  * The record must be big enough to contain the final remapped version.
 822  * Superfluous extension fields are removed and missing ones are added
 823  * and zeroed. The flags of the record are updated accordingly to what
 824  * the calling llog layer can support. Only influence user land has is
 825  * to store the NID in large NID format. The user land end user will
 826  * recieve all fields that supported by the kernel.
 827  *
 828  * The jobid and rename extensions will be added to a record, to match the
 829  * format an application expects, typically. In this case, the newly added
 830  * fields will be zeroed.
 831  * The Jobid field can be removed, to guarantee compatibility with older
 832  * clients that don't expect this field in the records they process.
 833  *
 834  * The following assumptions are being made:
 835  *   - CLF_RENAME will not be removed
 836  *   - CLF_JOBID will not be added without CLF_RENAME being added too
 837  *   - CLF_EXTRA_FLAGS will not be added without CLF_JOBID being added too
 838  *
 839  * @rec:        The record to remap.
 840  * @crf_wanted: Flags describing the desired extensions.
 841  * @cref_want:  Flags describing the desired extra extensions.
 842  */
 843 static void changelog_remap_rec(struct changelog_rec *rec,
 844                                 enum changelog_rec_flags crf_wanted,
 845                                 enum changelog_rec_extra_flags cref_want)
 846 {
 847         char *xattr_mov = NULL;
 848         char *omd_mov = NULL;
 849         char *nid_mov = NULL;
 850         char *uidgid_mov = NULL;
 851         char *ef_mov;
 852         char *jid_mov;
 853         char *rnm_mov;
 854         enum changelog_rec_extra_flags cref = CLFE_INVALID;
 855
 856         crf_wanted = (enum changelog_rec_flags)
 857             (crf_wanted & CLF_SUPPORTED);
 858         cref_want = (enum changelog_rec_extra_flags)
 859             (cref_want & CLFE_SUPPORTED);
 860
 861         if ((rec->cr_flags & CLF_SUPPORTED) == crf_wanted) {
 862                 if (!(rec->cr_flags & CLF_EXTRA_FLAGS) ||
 863                     (rec->cr_flags & CLF_EXTRA_FLAGS &&
 864                     (changelog_rec_extra_flags(rec)->cr_extra_flags &
 865                                                         CLFE_SUPPORTED) ==
 866                                                                      cref_want))
 867                         return;
 868         }
 869
 870         /* First move the variable-length name field */
 871         memmove((char *)rec + changelog_rec_offset(crf_wanted, cref_want),
 872                 changelog_rec_name(rec), rec->cr_namelen);
 873
 874         /* Locations of extensions in the remapped record */
 875         if (rec->cr_flags & CLF_EXTRA_FLAGS) {
 876                 xattr_mov = (char *)rec +
 877                         changelog_rec_offset(
 878                             (enum changelog_rec_flags)
 879                                     (crf_wanted & CLF_SUPPORTED),
 880                             (enum changelog_rec_extra_flags)
 881                                     (cref_want & ~CLFE_XATTR));
 882                 omd_mov = (char *)rec +
 883                         changelog_rec_offset(
 884                             (enum changelog_rec_flags)
 885                                     (crf_wanted & CLF_SUPPORTED),
 886                             (enum changelog_rec_extra_flags)
 887                                     (cref_want & ~(CLFE_OPEN | CLFE_XATTR)));
 888                 nid_mov = (char *)rec +
 889                         changelog_rec_offset(
 890                             (enum changelog_rec_flags)
 891                                 (crf_wanted & CLF_SUPPORTED),
 892                             (enum changelog_rec_extra_flags)
 893                                 (cref_want &
 894                                  ~(CLFE_NID | CLFE_OPEN | CLFE_XATTR)));
 895                 uidgid_mov = (char *)rec +
 896                         changelog_rec_offset(
 897                                 (enum changelog_rec_flags)
 898                                     (crf_wanted & CLF_SUPPORTED),
 899                                 (enum changelog_rec_extra_flags)
 900                                     (cref_want & ~(CLFE_UIDGID |
 901                                                            CLFE_NID |
 902                                                            CLFE_OPEN |
 903                                                            CLFE_XATTR)));
 904                 cref = (enum changelog_rec_extra_flags)
 905                         changelog_rec_extra_flags(rec)->cr_extra_flags;
 906         }
 907
 908         ef_mov  = (char *)rec +
 909                   changelog_rec_offset(
 910                                 (enum changelog_rec_flags)
 911                                  (crf_wanted & ~CLF_EXTRA_FLAGS), CLFE_INVALID);
 912         jid_mov = (char *)rec +
 913                   changelog_rec_offset((enum changelog_rec_flags)(crf_wanted &
 914                                        ~(CLF_EXTRA_FLAGS | CLF_JOBID)),
 915                                        CLFE_INVALID);
 916         rnm_mov = (char *)rec +
 917                   changelog_rec_offset((enum changelog_rec_flags)(crf_wanted &
 918                                        ~(CLF_EXTRA_FLAGS |
 919                                          CLF_JOBID |
 920                                          CLF_RENAME)),
 921                                        CLFE_INVALID);
 922
 923         /* Move the extension fields to the desired positions */
 924         if ((crf_wanted & CLF_EXTRA_FLAGS) &&
 925             (rec->cr_flags & CLF_EXTRA_FLAGS)) {
 926                 if ((cref_want & CLFE_XATTR) && (cref & CLFE_XATTR))
 927                         memmove(xattr_mov, changelog_rec_xattr(rec),
 928                                 sizeof(struct changelog_ext_xattr));
 929
 930                 if ((cref_want & CLFE_OPEN) && (cref & CLFE_OPEN))
 931                         memmove(omd_mov, changelog_rec_openmode(rec),
 932                                 sizeof(struct changelog_ext_openmode));
 933
 934                 if ((cref_want & CLFE_NID) && (cref & CLFE_NID)) {
 935                         struct changelog_ext_nid *cen = changelog_rec_nid(rec);
 936
 937                         if ((cref_want & CLFE_NID_BE) != (cref & CLFE_NID_BE)) {
 938                                 struct lnet_nid *nid;
 939
 940                                 if (!(cref_want & CLFE_NID_BE)) {
 941                                         nid = (struct lnet_nid *)cen;
 942                                         if (nid_is_nid4(nid)) {
 943                                                 struct changelog_ext_nid *mov;
 944
 945                                                 mov = (struct changelog_ext_nid *)nid_mov;
 946                                                 mov->cr_nid = lnet_nid_to_nid4(nid);
 947                                                 cref &= ~CLFE_NID_BE;
 948                                         } else {
 949                                                 cref &= ~(CLFE_NID |
 950                                                           CLFE_NID_BE);
 951                                         }
 952                                 } else {
 953                                         nid = (struct lnet_nid *)nid_mov;
 954                                         lnet_nid4_to_nid(cen->cr_nid, nid);
 955                                 }
 956                                 changelog_rec_extra_flags(rec)->cr_extra_flags =
 957                                         cref;
 958                         } else {
 959                                 memmove(nid_mov, cen, sizeof(*cen));
 960                         }
 961                 }
 962
 963                 if ((cref_want & CLFE_UIDGID) && (cref & CLFE_UIDGID))
 964                         memmove(uidgid_mov, changelog_rec_uidgid(rec),
 965                                 sizeof(struct changelog_ext_uidgid));
 966
 967                 memmove(ef_mov, changelog_rec_extra_flags(rec),
 968                         sizeof(struct changelog_ext_extra_flags));
 969         }
 970
 971         if ((crf_wanted & CLF_JOBID) && (rec->cr_flags & CLF_JOBID))
 972                 memmove(jid_mov, changelog_rec_jobid(rec),
 973                         sizeof(struct changelog_ext_jobid));
 974
 975         if ((crf_wanted & CLF_RENAME) && (rec->cr_flags & CLF_RENAME))
 976                 memmove(rnm_mov, changelog_rec_rename(rec),
 977                         sizeof(struct changelog_ext_rename));
 978
 979         /* Clear newly added fields */
 980         if (xattr_mov && (cref_want & CLFE_XATTR) &&
 981             !(cref & CLFE_XATTR))
 982                 memset(xattr_mov, 0, sizeof(struct changelog_ext_xattr));
 983
 984         if (omd_mov && (cref_want & CLFE_OPEN) &&
 985             !(cref & CLFE_OPEN))
 986                 memset(omd_mov, 0, sizeof(struct changelog_ext_openmode));
 987
 988         if (nid_mov && (cref_want & CLFE_NID) &&
 989             !(cref & CLFE_NID))
 990                 memset(nid_mov, 0, sizeof(struct changelog_ext_nid));
 991
 992         if (uidgid_mov && (cref_want & CLFE_UIDGID) &&
 993             !(cref & CLFE_UIDGID))
 994                 memset(uidgid_mov, 0, sizeof(struct changelog_ext_uidgid));
 995
 996         if ((crf_wanted & CLF_EXTRA_FLAGS) &&
 997             !(rec->cr_flags & CLF_EXTRA_FLAGS))
 998                 memset(ef_mov, 0, sizeof(struct changelog_ext_extra_flags));
 999
1000         if ((crf_wanted & CLF_JOBID) && !(rec->cr_flags & CLF_JOBID))
1001                 memset(jid_mov, 0, sizeof(struct changelog_ext_jobid));
1002
1003         if ((crf_wanted & CLF_RENAME) && !(rec->cr_flags & CLF_RENAME))
1004                 memset(rnm_mov, 0, sizeof(struct changelog_ext_rename));
1005
1006         /* Update the record's flags accordingly */
1007         rec->cr_flags = (rec->cr_flags & CLF_FLAGMASK) | crf_wanted;
1008         if (rec->cr_flags & CLF_EXTRA_FLAGS)
1009                 changelog_rec_extra_flags(rec)->cr_extra_flags =
1010                         changelog_rec_extra_flags(rec)->cr_extra_flags |
1011                         cref_want;
1012 }
1013
1014 /**
1015  * Remove optional fields that the client doesn't expect.
1016  * This is typically in order to ensure compatibility with older clients.
1017  * It is assumed that since we exclusively remove fields, the block will be
1018  * big enough to handle the remapped records. It is also assumed that records
1019  * of a block have the same format (i.e.: the same features enabled).
1020  *
1021  * \param[in,out]    hdr           Header of the block of records to remap.
1022  * \param[in,out]    last_hdr      Last header, don't read past this point.
1023  * \param[in]        flags         Flags describing the fields to keep.
1024  * \param[in]        extra_flags   Flags describing the extra fields to keep.
1025  */
1026 static void changelog_block_trim_ext(struct llog_rec_hdr *hdr,
1027                                      struct llog_rec_hdr *last_hdr,
1028                                      struct llog_handle *loghandle)
1029 {
1030         enum changelog_rec_flags flags = CLF_SUPPORTED;
1031         enum changelog_rec_extra_flags extra_flags = CLFE_SUPPORTED;
1032
1033         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_X_XATTR))
1034                 extra_flags &= ~CLFE_XATTR;
1035         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_X_OMODE))
1036                 extra_flags &= ~CLFE_OPEN;
1037         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_X_NID))
1038                 extra_flags &= ~(CLFE_NID | CLFE_NID_BE);
1039         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_X_NID_BE)) {
1040                 if (extra_flags & CLFE_NID_BE) {
1041                         /* The large nid won't be understood */
1042                         extra_flags &= ~CLFE_NID_BE;
1043                 }
1044         }
1045         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_X_UIDGID))
1046                 extra_flags &= ~CLFE_UIDGID;
1047         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_EXTRA_FLAGS))
1048                 flags &= ~CLF_EXTRA_FLAGS;
1049         if (!(loghandle->lgh_hdr->llh_flags & LLOG_F_EXT_JOBID))
1050                 flags &= ~CLF_JOBID;
1051
1052         if (flags == CLF_SUPPORTED && extra_flags == CLFE_SUPPORTED)
1053                 return;
1054
1055         if (hdr->lrh_type != CHANGELOG_REC)
1056                 return;
1057
1058         do {
1059                 struct changelog_rec *rec = (struct changelog_rec *)(hdr + 1);
1060                 enum changelog_rec_extra_flags xflag = CLFE_INVALID;
1061
1062                 if (flags & CLF_EXTRA_FLAGS &&
1063                     rec->cr_flags & CLF_EXTRA_FLAGS) {
1064                         xflag = changelog_rec_extra_flags(rec)->cr_extra_flags &
1065                                 extra_flags;
1066                 }
1067
1068                 if (unlikely(hdr->lrh_len == 0)) {
1069                         /* It is corruption case, we cannot know the next rec,
1070                          * jump to the last one directly to avoid dead loop. */
1071                         LCONSOLE(D_WARNING, "Hit invalid llog record: "
1072                                  "idx %u, type %u, id %u\n",
1073                                  hdr->lrh_index, hdr->lrh_type, hdr->lrh_id);
1074                         hdr = llog_rec_hdr_next(last_hdr);
1075                         if (unlikely(hdr == last_hdr))
1076                                 LCONSOLE(D_WARNING, "The last record crashed: "
1077                                          "idx %u, type %u, id %u\n",
1078                                          hdr->lrh_index, hdr->lrh_type,
1079                                          hdr->lrh_id);
1080                         break;
1081                 }
1082
1083                 /* Fill up the changelog record with everything the kernel
1084                  * version supports.
1085                  */
1086                 changelog_remap_rec(rec, rec->cr_flags & flags, xflag);
1087                 hdr = llog_rec_hdr_next(hdr);
1088                 /* Yield CPU to avoid soft-lockup if there are too many records
1089                  * to be handled. */
1090                 cond_resched();
1091         } while ((char *)hdr <= (char *)last_hdr);
1092 }
1093
1094 /**
1095  * Implementation of the llog_operations::lop_next_block
1096  *
1097  * This function finds the the next llog block to return which contains
1098  * record with required index. It is main part of llog processing.
1099  *
1100  * \param[in]     env           execution environment
1101  * \param[in]     loghandle     llog handle of the current llog
1102  * \param[in,out] cur_idx       index preceeding cur_offset
1103  * \param[in]     next_idx      target index to find
1104  * \param[in,out] cur_offset    furtherst point read in the file
1105  * \param[in]     buf           pointer to data buffer to fill
1106  * \param[in]     len           required len to read, it is
1107  *                              usually llog chunk_size.
1108  *
1109  * \retval                      0 on successful buffer read
1110  * \retval                      negative value on error
1111  */
1112 static int llog_osd_next_block(const struct lu_env *env,
1113                                struct llog_handle *loghandle, int *cur_idx,
1114                                int next_idx, __u64 *cur_offset, void *buf,
1115                                int len)
1116 {
1117         struct llog_thread_info *lgi = llog_info(env);
1118         struct dt_object        *o;
1119         struct dt_device        *dt;
1120         int                      rc;
1121         __u32                   chunk_size;
1122         int last_idx = *cur_idx;
1123         __u64 last_offset = *cur_offset;
1124         bool force_mini_rec = !next_idx;
1125
1126         ENTRY;
1127
1128         LASSERT(env);
1129         LASSERT(lgi);
1130
1131         chunk_size = loghandle->lgh_hdr->llh_hdr.lrh_len;
1132         if (len == 0 || len & (chunk_size - 1))
1133                 RETURN(-EINVAL);
1134
1135         LASSERT(loghandle);
1136         LASSERT(loghandle->lgh_ctxt);
1137
1138         if (CFS_FAIL_PRECHECK(OBD_FAIL_MDS_CHANGELOG_DEL) &&
1139             cfs_fail_val == ((unsigned long)loghandle & 0xFFFFFFFF)) {
1140                 CFS_RACE(OBD_FAIL_MDS_CHANGELOG_DEL);
1141                 msleep(MSEC_PER_SEC >> 2);
1142         }
1143
1144         o = loghandle->lgh_obj;
1145         LASSERT(o);
1146         dt_read_lock(env, o, 0);
1147         if (!llog_osd_exist(loghandle))
1148                 GOTO(out, rc = -ESTALE); //object was destroyed
1149
1150         dt = lu2dt_dev(o->do_lu.lo_dev);
1151         LASSERT(dt);
1152
1153         rc = dt_attr_get(env, o, &lgi->lgi_attr);
1154         if (rc)
1155                 GOTO(out, rc);
1156
1157         CDEBUG(D_OTHER,
1158                "looking for log index %u (cur idx %u off %llu), size %llu\n",
1159                next_idx, *cur_idx,
1160                *cur_offset, lgi->lgi_attr.la_size);
1161
1162         while (*cur_offset < lgi->lgi_attr.la_size) {
1163                 struct llog_rec_hdr     *rec, *last_rec;
1164                 struct llog_rec_tail    *tail;
1165
1166                 llog_skip_over(loghandle, cur_offset, *cur_idx,
1167                                next_idx, chunk_size, force_mini_rec);
1168
1169                 /* read up to next llog chunk_size block */
1170                 lgi->lgi_buf.lb_len = chunk_size -
1171                                       (*cur_offset & (chunk_size - 1));
1172                 lgi->lgi_buf.lb_buf = buf;
1173
1174                 rc = dt_read(env, o, &lgi->lgi_buf, cur_offset);
1175                 if (rc < 0) {
1176                         if (rc == -EBADR) {
1177                                 /* no goal is valid case */
1178                                 if (!next_idx)
1179                                         GOTO(out, rc);
1180                                 if (!force_mini_rec)
1181                                         goto retry;
1182                         }
1183                         CERROR("%s: can't read llog block from log "DFID
1184                                " offset %llu: rc = %d\n",
1185                                o->do_lu.lo_dev->ld_obd->obd_name,
1186                                PFID(lu_object_fid(&o->do_lu)), *cur_offset,
1187                                rc);
1188                         GOTO(out, rc);
1189                 }
1190
1191                 if (rc < len) {
1192                         /* signal the end of the valid buffer to
1193                          * llog_process */
1194                         memset(buf + rc, 0, len - rc);
1195                 }
1196
1197                 if (rc == 0) { /* end of file, nothing to do */
1198                         if (!force_mini_rec)
1199                                 goto retry;
1200                         GOTO(out, rc);
1201                 }
1202
1203                 if (rc < sizeof(*tail)) {
1204                         if (!force_mini_rec)
1205                                 goto retry;
1206
1207                         CERROR("%s: invalid llog block at log id "DFID" offset %llu\n",
1208                                o->do_lu.lo_dev->ld_obd->obd_name,
1209                                PLOGID(&loghandle->lgh_id), *cur_offset);
1210                         GOTO(out, rc = -EINVAL);
1211                 }
1212
1213                 rec = buf;
1214                 if (LLOG_REC_HDR_NEEDS_SWABBING(rec))
1215                         lustre_swab_llog_rec(rec);
1216                 tail = (struct llog_rec_tail *)((char *)buf + rc -
1217                                                 sizeof(struct llog_rec_tail));
1218
1219                 /* caller handles bad records if any */
1220                 if (llog_verify_record(loghandle, rec))
1221                         GOTO(out, rc = 0);
1222
1223                 /* get the last record in block */
1224                 last_rec = (struct llog_rec_hdr *)((char *)buf + rc -
1225                                                    tail->lrt_len);
1226
1227                 if (LLOG_REC_HDR_NEEDS_SWABBING(last_rec))
1228                         lustre_swab_llog_rec(last_rec);
1229
1230                 if (last_rec->lrh_index != tail->lrt_index) {
1231                         CERROR("%s: invalid llog tail at log id "DFID" offset %llu last_rec idx %u tail idx %u lrt len %u read_size %d\n",
1232                                o->do_lu.lo_dev->ld_obd->obd_name,
1233                                PLOGID(&loghandle->lgh_id), *cur_offset,
1234                                last_rec->lrh_index, tail->lrt_index,
1235                                tail->lrt_len, rc);
1236                         GOTO(out, rc = -EINVAL);
1237                 }
1238
1239                 *cur_idx = tail->lrt_index;
1240
1241                 /* this shouldn't happen */
1242                 if (tail->lrt_index == 0) {
1243                         CERROR("%s: invalid llog tail at log id "DFID"offset %llu bytes %d\n",
1244                                o->do_lu.lo_dev->ld_obd->obd_name,
1245                                PLOGID(&loghandle->lgh_id), *cur_offset, rc);
1246                         GOTO(out, rc = -EINVAL);
1247                 }
1248                 if (tail->lrt_index < next_idx) {
1249                         last_idx = *cur_idx;
1250                         last_offset = *cur_offset;
1251                         continue;
1252                 }
1253
1254                 /* sanity check that the start of the new buffer is no farther
1255                  * than the record that we wanted.  This shouldn't happen. */
1256                 if (next_idx && rec->lrh_index > next_idx) {
1257                         if (!force_mini_rec && next_idx > last_idx)
1258                                 goto retry;
1259
1260                         CERROR("%s: missed desired record? %u > %u\n",
1261                                o->do_lu.lo_dev->ld_obd->obd_name,
1262                                rec->lrh_index, next_idx);
1263                         GOTO(out, rc = -ENOENT);
1264                 }
1265
1266                 /* Trim unsupported extensions for compat w/ older clients */
1267                 changelog_block_trim_ext(rec, last_rec, loghandle);
1268
1269                 GOTO(out, rc = 0);
1270
1271 retry:
1272                 /* Note: because there are some pad records in the
1273                  * llog, so llog_skip_over() might skip too much
1274                  * records, let's try skip again with minimum record */
1275                 force_mini_rec = true;
1276                 *cur_offset = last_offset;
1277                 *cur_idx = last_idx;
1278         }
1279         /* being here means we reach end of llog but didn't find needed idx
1280          * normally could happen while processing remote llog, return -EBADR
1281          * to indicate access beyond end of file like dt_read() does and to
1282          * distunguish this situation from real IO or network issues.
1283          */
1284         GOTO(out, rc = -EBADR);
1285 out:
1286         dt_read_unlock(env, o);
1287         return rc;
1288 }
1289
1290 /**
1291  * Implementation of the llog_operations::lop_prev_block
1292  *
1293  * This function finds the llog block to return which contains
1294  * record with required index but in reverse order - from end of llog
1295  * to the beginning.
1296  * It is main part of reverse llog processing.
1297  *
1298  * \param[in] env       execution environment
1299  * \param[in] loghandle llog handle of the current llog
1300  * \param[in] prev_idx  target index to find
1301  * \param[in] buf       pointer to data buffer to fill
1302  * \param[in] len       required len to read, it is llog_chunk_size usually.
1303  *
1304  * \retval              0 on successful buffer read
1305  * \retval              negative value on error
1306  */
1307 static int llog_osd_prev_block(const struct lu_env *env,
1308                                struct llog_handle *loghandle,
1309                                int prev_idx, void *buf, int len)
1310 {
1311         struct llog_thread_info *lgi = llog_info(env);
1312         struct dt_object        *o;
1313         struct dt_device        *dt;
1314         loff_t                   cur_offset;
1315         __u32                   chunk_size;
1316         int                      rc;
1317
1318         ENTRY;
1319
1320         chunk_size = loghandle->lgh_hdr->llh_hdr.lrh_len;
1321         if (len == 0 || len & (chunk_size - 1))
1322                 RETURN(-EINVAL);
1323
1324         CDEBUG(D_OTHER, "looking for log index %u\n", prev_idx);
1325
1326         LASSERT(loghandle);
1327         LASSERT(loghandle->lgh_ctxt);
1328
1329         o = loghandle->lgh_obj;
1330         LASSERT(o);
1331         dt_read_lock(env, o, 0);
1332         if (!llog_osd_exist(loghandle))
1333                 GOTO(out, rc = -ESTALE);
1334
1335         dt = lu2dt_dev(o->do_lu.lo_dev);
1336         LASSERT(dt);
1337
1338         /* Let's only use mini record size for previous block read
1339          * for now XXX */
1340         cur_offset = chunk_size;
1341         llog_skip_over(loghandle, &cur_offset, 0, prev_idx,
1342                        chunk_size, true);
1343
1344         rc = dt_attr_get(env, o, &lgi->lgi_attr);
1345         if (rc)
1346                 GOTO(out, rc);
1347
1348         while (cur_offset < lgi->lgi_attr.la_size) {
1349                 struct llog_rec_hdr     *rec, *last_rec;
1350                 struct llog_rec_tail    *tail;
1351
1352                 lgi->lgi_buf.lb_len = len;
1353                 lgi->lgi_buf.lb_buf = buf;
1354                 rc = dt_read(env, o, &lgi->lgi_buf, &cur_offset);
1355                 if (rc < 0) {
1356                         CERROR("%s: can't read llog block from log "DFID
1357                                " offset %llu: rc = %d\n",
1358                                o->do_lu.lo_dev->ld_obd->obd_name,
1359                                PFID(lu_object_fid(&o->do_lu)), cur_offset, rc);
1360                         GOTO(out, rc);
1361                 }
1362
1363                 if (rc == 0) /* end of file, nothing to do */
1364                         GOTO(out, rc);
1365
1366                 if (rc < sizeof(*tail)) {
1367                         CERROR("%s: invalid llog block at log id "DFID" offset %llu\n",
1368                                o->do_lu.lo_dev->ld_obd->obd_name,
1369                                PLOGID(&loghandle->lgh_id), cur_offset);
1370                         GOTO(out, rc = -EINVAL);
1371                 }
1372
1373                 rec = buf;
1374                 if (LLOG_REC_HDR_NEEDS_SWABBING(rec))
1375                         lustre_swab_llog_rec(rec);
1376
1377                 tail = (struct llog_rec_tail *)((char *)buf + rc -
1378                                                 sizeof(struct llog_rec_tail));
1379                 /* get the last record in block */
1380                 last_rec = (struct llog_rec_hdr *)((char *)buf + rc -
1381                                                    le32_to_cpu(tail->lrt_len));
1382
1383                 if (LLOG_REC_HDR_NEEDS_SWABBING(last_rec))
1384                         lustre_swab_llog_rec(last_rec);
1385                 LASSERT(last_rec->lrh_index == tail->lrt_index);
1386
1387                 /* this shouldn't happen */
1388                 if (tail->lrt_index == 0) {
1389                         CERROR("%s: invalid llog tail at log id "DFID" offset %llu\n",
1390                                o->do_lu.lo_dev->ld_obd->obd_name,
1391                                PLOGID(&loghandle->lgh_id), cur_offset);
1392                         GOTO(out, rc = -EINVAL);
1393                 }
1394                 if (tail->lrt_index < prev_idx)
1395                         continue;
1396
1397                 /* sanity check that the start of the new buffer is no farther
1398                  * than the record that we wanted.  This shouldn't happen. */
1399                 if (rec->lrh_index > prev_idx) {
1400                         CERROR("%s: missed desired record? %u > %u\n",
1401                                o->do_lu.lo_dev->ld_obd->obd_name,
1402                                rec->lrh_index, prev_idx);
1403                         GOTO(out, rc = -ENOENT);
1404                 }
1405
1406                 /* Trim unsupported extensions for compat w/ older clients */
1407                 changelog_block_trim_ext(rec, last_rec, loghandle);
1408
1409                 GOTO(out, rc = 0);
1410         }
1411         GOTO(out, rc = -EIO);
1412 out:
1413         dt_read_unlock(env, o);
1414         return rc;
1415 }
1416
1417 /**
1418  * This is helper function to get llog directory object. It is used by named
1419  * llog operations to find/insert/delete llog entry from llog directory.
1420  *
1421  * \param[in] env       execution environment
1422  * \param[in] ctxt      llog context
1423  *
1424  * \retval              dt_object of llog directory
1425  * \retval              ERR_PTR of negative value on error
1426  */
1427 static struct dt_object *llog_osd_dir_get(const struct lu_env *env,
1428                                           struct llog_ctxt *ctxt)
1429 {
1430         struct dt_device        *dt;
1431         struct dt_thread_info   *dti = dt_info(env);
1432         struct dt_object        *dir;
1433         int                      rc;
1434
1435         dt = ctxt->loc_exp->exp_obd->obd_lvfs_ctxt.dt;
1436         if (ctxt->loc_dir == NULL) {
1437                 rc = dt_root_get(env, dt, &dti->dti_fid);
1438                 if (rc)
1439                         return ERR_PTR(rc);
1440                 dir = dt_locate(env, dt, &dti->dti_fid);
1441
1442                 if (!IS_ERR(dir) && !dt_try_as_dir(env, dir, false)) {
1443                         dt_object_put(env, dir);
1444                         return ERR_PTR(-ENOTDIR);
1445                 }
1446         } else {
1447                 lu_object_get(&ctxt->loc_dir->do_lu);
1448                 dir = ctxt->loc_dir;
1449         }
1450
1451         return dir;
1452 }
1453
1454 /**
1455  * Implementation of the llog_operations::lop_open
1456  *
1457  * This function opens the llog by its logid or by name, it may open also
1458  * non existent llog and assing then new id to it.
1459  * The llog_open/llog_close pair works similar to lu_object_find/put,
1460  * the object may not exist prior open. The result of open is just dt_object
1461  * in the llog header.
1462  *
1463  * \param[in] env               execution environment
1464  * \param[in] handle            llog handle of the current llog
1465  * \param[in] logid             logid of llog to open (nameless llog)
1466  * \param[in] name              name of llog to open (named llog)
1467  * \param[in] open_param
1468  *                              LLOG_OPEN_NEW - new llog, may not exist
1469  *                              LLOG_OPEN_EXIST - old llog, must exist
1470  *
1471  * \retval                      0 on successful open, llog_handle::lgh_obj
1472  *                              contains the dt_object of the llog.
1473  * \retval                      negative value on error
1474  */
1475 static int llog_osd_open(const struct lu_env *env, struct llog_handle *handle,
1476                          struct llog_logid *logid, char *name,
1477                          enum llog_open_param open_param)
1478 {
1479         struct llog_thread_info         *lgi = llog_info(env);
1480         struct llog_ctxt                *ctxt = handle->lgh_ctxt;
1481         struct dt_object                *o;
1482         struct dt_device                *dt;
1483         struct ls_device                *ls;
1484         struct local_oid_storage        *los = NULL;
1485         int                              rc = 0;
1486         bool new_id = false;
1487
1488         ENTRY;
1489
1490         LASSERT(env);
1491         LASSERT(ctxt);
1492         LASSERT(ctxt->loc_exp);
1493         LASSERT(ctxt->loc_exp->exp_obd);
1494         dt = ctxt->loc_exp->exp_obd->obd_lvfs_ctxt.dt;
1495         LASSERT(dt);
1496         if (ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
1497                 struct lu_object_conf conf = { 0 };
1498                 if (logid != NULL) {
1499                         logid_to_fid(logid, &lgi->lgi_fid);
1500                 } else {
1501                         /* If logid == NULL, then it means the caller needs
1502                          * to allocate new FID (llog_cat_declare_add_rec()). */
1503                         rc = dt_fid_alloc(env, dt, &lgi->lgi_fid, NULL, NULL);
1504                         if (rc < 0)
1505                                 RETURN(rc);
1506                         rc = 0;
1507                         conf.loc_flags = LOC_F_NEW;
1508                 }
1509
1510                 o = dt_locate_at(env, dt, &lgi->lgi_fid,
1511                                  dt->dd_lu_dev.ld_site->ls_top_dev, &conf);
1512                 if (IS_ERR(o))
1513                         RETURN(PTR_ERR(o));
1514
1515                 goto after_open;
1516         }
1517
1518         ls = ls_device_find_or_init(dt);
1519         if (IS_ERR(ls))
1520                 RETURN(PTR_ERR(ls));
1521
1522         mutex_lock(&ls->ls_los_mutex);
1523         los = dt_los_find(ls, name != NULL ? FID_SEQ_LLOG_NAME : FID_SEQ_LLOG);
1524         mutex_unlock(&ls->ls_los_mutex);
1525         LASSERT(los);
1526         ls_device_put(env, ls);
1527
1528         LASSERT(handle);
1529
1530         if (logid != NULL) {
1531                 logid_to_fid(logid, &lgi->lgi_fid);
1532         } else if (name) {
1533                 struct dt_object *llog_dir;
1534
1535                 llog_dir = llog_osd_dir_get(env, ctxt);
1536                 if (IS_ERR(llog_dir))
1537                         GOTO(out, rc = PTR_ERR(llog_dir));
1538                 dt_read_lock(env, llog_dir, 0);
1539                 rc = dt_lookup_dir(env, llog_dir, name, &lgi->lgi_fid);
1540                 dt_read_unlock(env, llog_dir);
1541                 dt_object_put(env, llog_dir);
1542                 if (rc == -ENOENT && open_param == LLOG_OPEN_NEW) {
1543                         /* generate fid for new llog */
1544                         rc = local_object_fid_generate(env, los,
1545                                                        &lgi->lgi_fid);
1546                         new_id = true;
1547                 }
1548                 if (rc < 0)
1549                         GOTO(out, rc);
1550                 OBD_ALLOC(handle->lgh_name, strlen(name) + 1);
1551                 if (handle->lgh_name)
1552                         strcpy(handle->lgh_name, name);
1553                 else
1554                         GOTO(out, rc = -ENOMEM);
1555         } else {
1556                 LASSERTF(open_param & LLOG_OPEN_NEW, "%#x\n", open_param);
1557                 /* generate fid for new llog */
1558 generate:
1559                 rc = local_object_fid_generate(env, los, &lgi->lgi_fid);
1560                 if (rc < 0)
1561                         GOTO(out, rc);
1562                 new_id = true;
1563         }
1564         if (CFS_FAIL_PRECHECK(OBD_FAIL_MDS_LLOG_UMOUNT_RACE) &&
1565             cfs_fail_val == 1) {
1566                 cfs_fail_val = 2;
1567                 CFS_RACE(OBD_FAIL_MDS_LLOG_UMOUNT_RACE);
1568                 msleep(MSEC_PER_SEC);
1569         }
1570         o = ls_locate(env, ls, &lgi->lgi_fid, NULL);
1571         if (IS_ERR(o))
1572                 GOTO(out_name, rc = PTR_ERR(o));
1573
1574         if (dt_object_exists(o) && new_id) {
1575                 /* llog exists with just generated ID, e.g. some old llog file
1576                  * still is in use or is orphan, drop a warn and skip it. */
1577                 CDEBUG(D_INFO, "%s: llog exists with the same FID: "DFID
1578                        ", skipping\n",
1579                        o->do_lu.lo_dev->ld_obd->obd_name,
1580                        PFID(lu_object_fid(&o->do_lu)));
1581                 dt_object_put(env, o);
1582                 /* just skip this llog ID, we shouldn't delete it because we
1583                  * don't know exactly what is its purpose and state. */
1584                 goto generate;
1585         }
1586
1587 after_open:
1588         /* No new llog is expected but doesn't exist */
1589         if (open_param != LLOG_OPEN_NEW && !dt_object_exists(o)) {
1590                 CDEBUG(D_INFO, "%s: llog FID: "DFID" obj %p doesn`t exist\n",
1591                        o->do_lu.lo_dev->ld_obd->obd_name,
1592                        PFID(lu_object_fid(&o->do_lu)), o);
1593                 GOTO(out_put, rc = -ENOENT);
1594         }
1595         fid_to_logid(&lgi->lgi_fid, &handle->lgh_id);
1596         handle->lgh_obj = o;
1597         handle->private_data = los;
1598         LASSERT(handle->lgh_ctxt);
1599
1600         RETURN(rc);
1601
1602 out_put:
1603         dt_object_put(env, o);
1604 out_name:
1605         OBD_FREE(handle->lgh_name, strlen(name) + 1);
1606 out:
1607         if (los != NULL)
1608                 dt_los_put(los);
1609         RETURN(rc);
1610 }
1611
1612 /**
1613  * Get dir for regular fid log object
1614  *
1615  * Get directory for regular fid log object, and these regular fid log
1616  * object will be inserted under this directory, to satisfy the FS
1617  * consistency check, e2fsck etc.
1618  *
1619  * \param [in] env      execution environment
1620  * \param [in] dto      llog object
1621  *
1622  * \retval              pointer to the directory if it is found.
1623  * \retval              ERR_PTR(negative errno) if it fails.
1624  */
1625 static struct dt_object *llog_osd_get_regular_fid_dir(const struct lu_env *env,
1626                                                       struct dt_object *dto)
1627 {
1628         struct llog_thread_info *lgi = llog_info(env);
1629         struct seq_server_site *ss = dto->do_lu.lo_dev->ld_site->ld_seq_site;
1630         struct lu_seq_range     *range = &lgi->lgi_range;
1631         struct lu_fid           *dir_fid = &lgi->lgi_fid;
1632         struct dt_object        *dir;
1633         int                     rc;
1634         ENTRY;
1635
1636         fld_range_set_any(range);
1637         LASSERT(ss != NULL);
1638         rc = ss->ss_server_fld->lsf_seq_lookup(env, ss->ss_server_fld,
1639                                    fid_seq(lu_object_fid(&dto->do_lu)), range);
1640         if (rc < 0)
1641                 RETURN(ERR_PTR(rc));
1642
1643         lu_update_log_dir_fid(dir_fid, range->lsr_index);
1644         dir = dt_locate(env, lu2dt_dev(dto->do_lu.lo_dev), dir_fid);
1645         if (IS_ERR(dir))
1646                 RETURN(dir);
1647
1648         if (!dt_try_as_dir(env, dir, false)) {
1649                 dt_object_put(env, dir);
1650                 RETURN(ERR_PTR(-ENOTDIR));
1651         }
1652
1653         RETURN(dir);
1654 }
1655
1656 /**
1657  * Add llog object with regular FID to name entry
1658  *
1659  * Add llog object with regular FID to name space, and each llog
1660  * object on each MDT will be /update_log_dir/[seq:oid:ver],
1661  * so to satisfy the namespace consistency check, e2fsck etc.
1662  *
1663  * \param [in] env      execution environment
1664  * \param [in] dto      llog object
1665  * \param [in] th       thandle
1666  * \param [in] declare  if it is declare or execution
1667  *
1668  * \retval              0 if insertion succeeds.
1669  * \retval              negative errno if insertion fails.
1670  */
1671 static int
1672 llog_osd_regular_fid_add_name_entry(const struct lu_env *env,
1673                                     struct dt_object *dto,
1674                                     struct thandle *th, bool declare)
1675 {
1676         struct llog_thread_info *lgi = llog_info(env);
1677         const struct lu_fid     *fid = lu_object_fid(&dto->do_lu);
1678         struct dt_insert_rec    *rec = &lgi->lgi_dt_rec;
1679         struct dt_object        *dir;
1680         char                    *name = lgi->lgi_name;
1681         int                     rc;
1682         ENTRY;
1683
1684         if (!fid_is_norm(fid))
1685                 RETURN(0);
1686
1687         dir = llog_osd_get_regular_fid_dir(env, dto);
1688         if (IS_ERR(dir))
1689                 RETURN(PTR_ERR(dir));
1690
1691         rec->rec_fid = fid;
1692         rec->rec_type = S_IFREG;
1693         snprintf(name, sizeof(lgi->lgi_name), DFID, PFID(fid));
1694         dt_write_lock(env, dir, 0);
1695         if (declare) {
1696                 rc = dt_declare_insert(env, dir, (struct dt_rec *)rec,
1697                                (struct dt_key *)name, th);
1698         } else {
1699                 rc = dt_insert(env, dir, (struct dt_rec *)rec,
1700                                (struct dt_key *)name, th);
1701         }
1702         dt_write_unlock(env, dir);
1703
1704         dt_object_put(env, dir);
1705         RETURN(rc);
1706 }
1707
1708
1709 /**
1710  * Implementation of the llog_operations::lop_declare_create
1711  *
1712  * This function declares the llog create. It declares also name insert
1713  * into llog directory in case of named llog.
1714  *
1715  * \param[in] env       execution environment
1716  * \param[in] res       llog handle of the current llog
1717  * \param[in] th        current transaction handle
1718  *
1719  * \retval              0 on successful create declaration
1720  * \retval              negative value on error
1721  */
1722 static int llog_osd_declare_create(const struct lu_env *env,
1723                                    struct llog_handle *res, struct thandle *th)
1724 {
1725         struct llog_thread_info         *lgi = llog_info(env);
1726         struct dt_insert_rec            *rec = &lgi->lgi_dt_rec;
1727         struct local_oid_storage        *los;
1728         struct dt_object                *o;
1729         int                              rc;
1730
1731         ENTRY;
1732
1733         LASSERT(res->lgh_obj);
1734         LASSERT(th);
1735
1736         /* object can be created by another thread */
1737         o = res->lgh_obj;
1738         if (dt_object_exists(o))
1739                 RETURN(0);
1740
1741         if (res->lgh_ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
1742                 struct llog_thread_info *lgi = llog_info(env);
1743
1744                 lgi->lgi_attr.la_valid = LA_MODE | LA_SIZE;
1745                 lgi->lgi_attr.la_size = 0;
1746                 lgi->lgi_attr.la_mode = S_IFREG | S_IRUGO | S_IWUSR;
1747                 lgi->lgi_dof.dof_type = dt_mode_to_dft(S_IFREG);
1748
1749                 rc = dt_declare_create(env, o, &lgi->lgi_attr, NULL,
1750                                        &lgi->lgi_dof, th);
1751                 if (rc < 0)
1752                         RETURN(rc);
1753
1754
1755                 rc = llog_osd_regular_fid_add_name_entry(env, o, th, true);
1756
1757                 RETURN(rc);
1758         }
1759         los = res->private_data;
1760         LASSERT(los);
1761
1762         rc = llog_osd_declare_new_object(env, los, o, th);
1763         if (rc)
1764                 RETURN(rc);
1765
1766         /* do not declare header initialization here as it's declared
1767          * in llog_osd_declare_write_rec() which is always called */
1768
1769         if (res->lgh_name) {
1770                 struct dt_object *llog_dir;
1771
1772                 llog_dir = llog_osd_dir_get(env, res->lgh_ctxt);
1773                 if (IS_ERR(llog_dir))
1774                         RETURN(PTR_ERR(llog_dir));
1775                 logid_to_fid(&res->lgh_id, &lgi->lgi_fid);
1776                 rec->rec_fid = &lgi->lgi_fid;
1777                 rec->rec_type = S_IFREG;
1778                 rc = dt_declare_insert(env, llog_dir,
1779                                        (struct dt_rec *)rec,
1780                                        (struct dt_key *)res->lgh_name, th);
1781                 dt_object_put(env, llog_dir);
1782                 if (rc)
1783                         CERROR("%s: can't declare named llog %s: rc = %d\n",
1784                                o->do_lu.lo_dev->ld_obd->obd_name,
1785                                res->lgh_name, rc);
1786         }
1787         RETURN(rc);
1788 }
1789
1790 /**
1791  * Implementation of the llog_operations::lop_create
1792  *
1793  * This function creates the llog according with llog_handle::lgh_obj
1794  * and llog_handle::lgh_name.
1795  *
1796  * \param[in] env       execution environment
1797  * \param[in] res       llog handle of the current llog
1798  * \param[in] th        current transaction handle
1799  *
1800  * \retval              0 on successful create
1801  * \retval              negative value on error
1802  */
1803 static int llog_osd_create(const struct lu_env *env, struct llog_handle *res,
1804                            struct thandle *th)
1805 {
1806         struct llog_thread_info *lgi = llog_info(env);
1807         struct dt_insert_rec    *rec = &lgi->lgi_dt_rec;
1808         struct local_oid_storage *los;
1809         struct dt_object        *o;
1810         int                      rc = 0;
1811
1812         ENTRY;
1813
1814         LASSERT(env);
1815         o = res->lgh_obj;
1816         LASSERT(o);
1817
1818         /* llog can be already created */
1819         if (dt_object_exists(o))
1820                 RETURN(-EEXIST);
1821
1822         if (res->lgh_ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
1823                 struct llog_thread_info *lgi = llog_info(env);
1824
1825                 lgi->lgi_attr.la_valid = LA_MODE | LA_SIZE | LA_TYPE;
1826                 lgi->lgi_attr.la_size = 0;
1827                 lgi->lgi_attr.la_mode = S_IFREG | S_IRUGO | S_IWUSR;
1828                 lgi->lgi_dof.dof_type = dt_mode_to_dft(S_IFREG);
1829
1830                 dt_write_lock(env, o, 0);
1831                 rc = dt_create(env, o, &lgi->lgi_attr, NULL,
1832                                &lgi->lgi_dof, th);
1833                 dt_write_unlock(env, o);
1834                 if (rc < 0)
1835                         RETURN(rc);
1836
1837                 rc = llog_osd_regular_fid_add_name_entry(env, o, th, false);
1838
1839                 RETURN(rc);
1840         }
1841
1842         los = res->private_data;
1843         LASSERT(los);
1844
1845         dt_write_lock(env, o, 0);
1846         if (!dt_object_exists(o))
1847                 rc = llog_osd_create_new_object(env, los, o, th);
1848         else
1849                 rc = -EEXIST;
1850
1851         dt_write_unlock(env, o);
1852         if (rc)
1853                 RETURN(rc);
1854
1855         if (res->lgh_name) {
1856                 struct dt_object *llog_dir;
1857
1858                 llog_dir = llog_osd_dir_get(env, res->lgh_ctxt);
1859                 if (IS_ERR(llog_dir))
1860                         RETURN(PTR_ERR(llog_dir));
1861
1862                 logid_to_fid(&res->lgh_id, &lgi->lgi_fid);
1863                 rec->rec_fid = &lgi->lgi_fid;
1864                 rec->rec_type = S_IFREG;
1865                 dt_read_lock(env, llog_dir, 0);
1866                 rc = dt_insert(env, llog_dir, (struct dt_rec *)rec,
1867                                (struct dt_key *)res->lgh_name, th);
1868                 dt_read_unlock(env, llog_dir);
1869                 dt_object_put(env, llog_dir);
1870                 if (rc)
1871                         CERROR("%s: can't create named llog %s: rc = %d\n",
1872                                o->do_lu.lo_dev->ld_obd->obd_name,
1873                                res->lgh_name, rc);
1874         }
1875         RETURN(rc);
1876 }
1877
1878 /**
1879  * Implementation of the llog_operations::lop_close
1880  *
1881  * This function closes the llog. It just put llog object and referenced
1882  * local storage.
1883  *
1884  * \param[in] env       execution environment
1885  * \param[in] handle    llog handle of the current llog
1886  *
1887  * \retval              0 on successful llog close
1888  * \retval              negative value on error
1889  */
1890 static int llog_osd_close(const struct lu_env *env, struct llog_handle *handle)
1891 {
1892         struct local_oid_storage        *los;
1893         int                              rc = 0;
1894
1895         ENTRY;
1896
1897         LASSERT(handle->lgh_obj);
1898
1899         if (handle->lgh_ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
1900                 /* Remove the object from the cache, otherwise it may
1901                  * hold LOD being released during cleanup process */
1902                 dt_object_put_nocache(env, handle->lgh_obj);
1903                 LASSERT(handle->private_data == NULL);
1904                 RETURN(rc);
1905         } else {
1906                 dt_object_put(env, handle->lgh_obj);
1907         }
1908         los = handle->private_data;
1909         LASSERT(los);
1910         dt_los_put(los);
1911
1912         OBD_FREE(handle->lgh_name, strlen(handle->lgh_name) + 1);
1913
1914         RETURN(rc);
1915 }
1916
1917 /**
1918  * delete llog object name entry
1919  *
1920  * Delete llog object (with regular FID) from name space (under
1921  * update_log_dir).
1922  *
1923  * \param [in] env      execution environment
1924  * \param [in] dto      llog object
1925  * \param [in] th       thandle
1926  * \param [in] declare  if it is declare or execution
1927  *
1928  * \retval              0 if deletion succeeds.
1929  * \retval              negative errno if deletion fails.
1930  */
1931 static int
1932 llog_osd_regular_fid_del_name_entry(const struct lu_env *env,
1933                                     struct dt_object *dto,
1934                                     struct thandle *th, bool declare)
1935 {
1936         struct llog_thread_info *lgi = llog_info(env);
1937         const struct lu_fid     *fid = lu_object_fid(&dto->do_lu);
1938         struct dt_object        *dir;
1939         char                    *name = lgi->lgi_name;
1940         int                     rc;
1941         ENTRY;
1942
1943         if (!fid_is_norm(fid))
1944                 RETURN(0);
1945
1946         dir = llog_osd_get_regular_fid_dir(env, dto);
1947         if (IS_ERR(dir))
1948                 RETURN(PTR_ERR(dir));
1949
1950         snprintf(name, sizeof(lgi->lgi_name), DFID, PFID(fid));
1951         dt_write_lock(env, dir, 0);
1952         if (declare) {
1953                 rc = dt_declare_delete(env, dir, (struct dt_key *)name,
1954                                        th);
1955         } else {
1956                 rc = dt_delete(env, dir, (struct dt_key *)name, th);
1957         }
1958         dt_write_unlock(env, dir);
1959
1960         dt_object_put(env, dir);
1961         RETURN(rc);
1962 }
1963
1964 /**
1965  * Implementation of the llog_operations::lop_declare_destroy
1966  *
1967  * This function declare destroys the llog and deletes also entry in the
1968  * llog directory in case of named llog. Llog should be opened prior that.
1969  *
1970  * \param[in] env               execution environment
1971  * \param[in] loghandle llog handle of the current llog
1972  *
1973  * \retval              0 on successful destroy
1974  * \retval              negative value on error
1975  */
1976 static int llog_osd_declare_destroy(const struct lu_env *env,
1977                                     struct llog_handle *loghandle,
1978                                     struct thandle *th)
1979 {
1980         struct llog_ctxt        *ctxt;
1981         struct dt_object        *o, *llog_dir = NULL;
1982         int                      rc;
1983
1984         ENTRY;
1985
1986         ctxt = loghandle->lgh_ctxt;
1987         LASSERT(ctxt);
1988
1989         o = loghandle->lgh_obj;
1990         LASSERT(o);
1991
1992         if (loghandle->lgh_name) {
1993                 llog_dir = llog_osd_dir_get(env, ctxt);
1994                 if (IS_ERR(llog_dir))
1995                         RETURN(PTR_ERR(llog_dir));
1996
1997                 rc = dt_declare_delete(env, llog_dir,
1998                                        (struct dt_key *)loghandle->lgh_name,
1999                                        th);
2000                 if (rc < 0)
2001                         GOTO(out_put, rc);
2002         }
2003
2004         rc = dt_declare_ref_del(env, o, th);
2005         if (rc < 0)
2006                 GOTO(out_put, rc);
2007
2008         rc = dt_declare_destroy(env, o, th);
2009         if (rc < 0)
2010                 GOTO(out_put, rc);
2011
2012         if (loghandle->lgh_ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
2013                 rc = llog_osd_regular_fid_del_name_entry(env, o, th, true);
2014                 if (rc < 0)
2015                         GOTO(out_put, rc);
2016         }
2017
2018 out_put:
2019         if (!(IS_ERR_OR_NULL(llog_dir)))
2020                 dt_object_put(env, llog_dir);
2021
2022         RETURN(rc);
2023 }
2024
2025
2026 /**
2027  * Implementation of the llog_operations::lop_destroy
2028  *
2029  * This function destroys the llog and deletes also entry in the
2030  * llog directory in case of named llog. Llog should be opened prior that.
2031  * Destroy method is not part of external transaction and does everything
2032  * inside.
2033  *
2034  * \param[in] env               execution environment
2035  * \param[in] loghandle llog handle of the current llog
2036  *
2037  * \retval              0 on successful destroy
2038  * \retval              negative value on error
2039  */
2040 static int llog_osd_destroy(const struct lu_env *env,
2041                             struct llog_handle *loghandle, struct thandle *th)
2042 {
2043         struct llog_ctxt        *ctxt;
2044         struct dt_object        *o, *llog_dir = NULL;
2045         int                      rc;
2046
2047         ENTRY;
2048
2049         ctxt = loghandle->lgh_ctxt;
2050         LASSERT(ctxt != NULL);
2051
2052         o = loghandle->lgh_obj;
2053         LASSERT(o != NULL);
2054
2055         dt_write_lock(env, o, 0);
2056         if (!llog_osd_exist(loghandle))
2057                 GOTO(out_unlock, rc = 0);
2058
2059         if (loghandle->lgh_name) {
2060                 llog_dir = llog_osd_dir_get(env, ctxt);
2061                 if (IS_ERR(llog_dir))
2062                         GOTO(out_unlock, rc = PTR_ERR(llog_dir));
2063
2064                 dt_read_lock(env, llog_dir, 0);
2065                 rc = dt_delete(env, llog_dir,
2066                                (struct dt_key *)loghandle->lgh_name,
2067                                th);
2068                 dt_read_unlock(env, llog_dir);
2069                 if (rc) {
2070                         CERROR("%s: can't remove llog %s: rc = %d\n",
2071                                o->do_lu.lo_dev->ld_obd->obd_name,
2072                                loghandle->lgh_name, rc);
2073                         GOTO(out_unlock, rc);
2074                 }
2075         }
2076
2077         dt_ref_del(env, o, th);
2078         rc = dt_destroy(env, o, th);
2079         if (rc < 0)
2080                 GOTO(out_unlock, rc);
2081
2082         loghandle->lgh_destroyed = true;
2083         if (loghandle->lgh_ctxt->loc_flags & LLOG_CTXT_FLAG_NORMAL_FID) {
2084                 rc = llog_osd_regular_fid_del_name_entry(env, o, th, false);
2085                 if (rc < 0)
2086                         GOTO(out_unlock, rc);
2087         }
2088
2089 out_unlock:
2090         dt_write_unlock(env, o);
2091         if (!(IS_ERR_OR_NULL(llog_dir)))
2092                 dt_object_put(env, llog_dir);
2093         RETURN(rc);
2094 }
2095
2096 /**
2097  * Implementation of the llog_operations::lop_setup
2098  *
2099  * This function setup the llog on local storage.
2100  *
2101  * \param[in] env       execution environment
2102  * \param[in] obd       obd device the llog belongs to
2103  * \param[in] olg       the llog group, it is always zero group now.
2104  * \param[in] ctxt_idx  the llog index, it defines the purpose of this llog.
2105  *                      Every new llog type have to use own index.
2106  * \param[in] disk_obd  the storage obd, where llog is stored.
2107  *
2108  * \retval              0 on successful llog setup
2109  * \retval              negative value on error
2110  */
2111 static int llog_osd_setup(const struct lu_env *env, struct obd_device *obd,
2112                           struct obd_llog_group *olg, int ctxt_idx,
2113                           struct obd_device *disk_obd)
2114 {
2115         struct llog_thread_info         *lgi = llog_info(env);
2116         struct llog_ctxt                *ctxt;
2117         int                              rc = 0;
2118         ENTRY;
2119
2120         LASSERT(obd);
2121         LASSERT(olg->olg_ctxts[ctxt_idx]);
2122
2123         ctxt = llog_ctxt_get(olg->olg_ctxts[ctxt_idx]);
2124         LASSERT(ctxt);
2125
2126         if (disk_obd == NULL)
2127                 GOTO(out, rc = 0);
2128
2129         /* initialize data allowing to generate new fids,
2130          * literally we need a sequece */
2131         lgi->lgi_fid.f_seq = FID_SEQ_LLOG;
2132         lgi->lgi_fid.f_oid = 1;
2133         lgi->lgi_fid.f_ver = 0;
2134         rc = local_oid_storage_init(env, disk_obd->obd_lvfs_ctxt.dt,
2135                                     &lgi->lgi_fid,
2136                                     &ctxt->loc_los_nameless);
2137         if (rc != 0)
2138                 GOTO(out, rc);
2139
2140         lgi->lgi_fid.f_seq = FID_SEQ_LLOG_NAME;
2141         lgi->lgi_fid.f_oid = 1;
2142         lgi->lgi_fid.f_ver = 0;
2143         rc = local_oid_storage_init(env, disk_obd->obd_lvfs_ctxt.dt,
2144                                     &lgi->lgi_fid,
2145                                     &ctxt->loc_los_named);
2146         if (rc != 0) {
2147                 local_oid_storage_fini(env, ctxt->loc_los_nameless);
2148                 ctxt->loc_los_nameless = NULL;
2149         }
2150
2151         GOTO(out, rc);
2152
2153 out:
2154         llog_ctxt_put(ctxt);
2155         return rc;
2156 }
2157
2158 /**
2159  * Implementation of the llog_operations::lop_cleanup
2160  *
2161  * This function cleanups the llog on local storage.
2162  *
2163  * \param[in] env       execution environment
2164  * \param[in] ctxt      the llog context
2165  *
2166  * \retval              0
2167  */
2168 static int llog_osd_cleanup(const struct lu_env *env, struct llog_ctxt *ctxt)
2169 {
2170         if (ctxt->loc_los_nameless != NULL) {
2171                 local_oid_storage_fini(env, ctxt->loc_los_nameless);
2172                 ctxt->loc_los_nameless = NULL;
2173         }
2174
2175         if (ctxt->loc_los_named != NULL) {
2176                 local_oid_storage_fini(env, ctxt->loc_los_named);
2177                 ctxt->loc_los_named = NULL;
2178         }
2179
2180         return 0;
2181 }
2182
2183 const struct llog_operations llog_osd_ops = {
2184         .lop_next_block         = llog_osd_next_block,
2185         .lop_prev_block         = llog_osd_prev_block,
2186         .lop_read_header        = llog_osd_read_header,
2187         .lop_declare_destroy    = llog_osd_declare_destroy,
2188         .lop_destroy            = llog_osd_destroy,
2189         .lop_setup              = llog_osd_setup,
2190         .lop_cleanup            = llog_osd_cleanup,
2191         .lop_open               = llog_osd_open,
2192         .lop_exist              = llog_osd_exist,
2193         .lop_declare_create     = llog_osd_declare_create,
2194         .lop_create             = llog_osd_create,
2195         .lop_declare_write_rec  = llog_osd_declare_write_rec,
2196         .lop_write_rec          = llog_osd_write_rec,
2197         .lop_close              = llog_osd_close,
2198 };
2199 EXPORT_SYMBOL(llog_osd_ops);
2200
2201 const struct llog_operations llog_common_cat_ops = {
2202         .lop_next_block         = llog_osd_next_block,
2203         .lop_prev_block         = llog_osd_prev_block,
2204         .lop_read_header        = llog_osd_read_header,
2205         .lop_declare_destroy    = llog_osd_declare_destroy,
2206         .lop_destroy            = llog_osd_destroy,
2207         .lop_setup              = llog_osd_setup,
2208         .lop_cleanup            = llog_osd_cleanup,
2209         .lop_open               = llog_osd_open,
2210         .lop_exist              = llog_osd_exist,
2211         .lop_declare_create     = llog_osd_declare_create,
2212         .lop_create             = llog_osd_create,
2213         .lop_declare_write_rec  = llog_osd_declare_write_rec,
2214         .lop_write_rec          = llog_osd_write_rec,
2215         .lop_close              = llog_osd_close,
2216         .lop_add                = llog_cat_add_rec,
2217         .lop_declare_add        = llog_cat_declare_add_rec,
2218 };
2219 EXPORT_SYMBOL(llog_common_cat_ops);
2220
2221 /**
2222  * Read the special file which contains the list of llog catalogs IDs
2223  *
2224  * This function reads the CATALOGS file which contains the array of llog
2225  * catalogs IDs. The main purpose of this file is to store OSP llogs indexed
2226  * by OST/MDT number.
2227  *
2228  * \param[in]  env              execution environment
2229  * \param[in]  d                corresponding storage device
2230  * \param[in]  idx              position to start from, usually OST/MDT index
2231  * \param[in]  count            how many catalog IDs to read
2232  * \param[out] idarray          the buffer for the data. If it is NULL then
2233  *                              function returns just number of catalog IDs
2234  *                              in the file.
2235  * \param[in]  fid              LLOG_CATALOGS_OID for CATALOG object
2236  *
2237  * \retval                      0 on successful read of catalog IDs
2238  * \retval                      negative value on error
2239  * \retval                      positive value which is number of records in
2240  *                              the file if \a idarray is NULL
2241  */
2242 int llog_osd_get_cat_list(const struct lu_env *env, struct dt_device *d,
2243                           int idx, int count, struct llog_catid *idarray,
2244                           const struct lu_fid *fid)
2245 {
2246         struct llog_thread_info *lgi = llog_info(env);
2247         struct dt_object        *o = NULL;
2248         struct thandle          *th;
2249         int                      rc, size;
2250
2251         ENTRY;
2252
2253         LASSERT(d);
2254
2255         size = sizeof(*idarray) * count;
2256         lgi->lgi_off = idx *  sizeof(*idarray);
2257
2258         lgi->lgi_fid = *fid;
2259         o = dt_locate(env, d, &lgi->lgi_fid);
2260         if (IS_ERR(o))
2261                 RETURN(PTR_ERR(o));
2262
2263         if (!dt_object_exists(o)) {
2264                 th = dt_trans_create(env, d);
2265                 if (IS_ERR(th))
2266                         GOTO(out, rc = PTR_ERR(th));
2267
2268                 lgi->lgi_attr.la_valid = LA_MODE | LA_TYPE;
2269                 lgi->lgi_attr.la_mode = S_IFREG | S_IRUGO | S_IWUSR;
2270                 lgi->lgi_dof.dof_type = dt_mode_to_dft(S_IFREG);
2271
2272                 th->th_wait_submit = 1;
2273                 /* Make the llog object creation synchronization, so
2274                  * it will be reliable to the reference, especially
2275                  * for remote reference */
2276                 th->th_sync = 1;
2277
2278                 rc = dt_declare_create(env, o, &lgi->lgi_attr, NULL,
2279                                        &lgi->lgi_dof, th);
2280                 if (rc)
2281                         GOTO(out_trans, rc);
2282
2283                 rc = dt_trans_start_local(env, d, th);
2284                 if (rc)
2285                         GOTO(out_trans, rc);
2286
2287                 dt_write_lock(env, o, 0);
2288                 if (!dt_object_exists(o))
2289                         rc = dt_create(env, o, &lgi->lgi_attr, NULL,
2290                                        &lgi->lgi_dof, th);
2291                 dt_write_unlock(env, o);
2292 out_trans:
2293                 dt_trans_stop(env, d, th);
2294                 if (rc)
2295                         GOTO(out, rc);
2296         }
2297
2298         rc = dt_attr_get(env, o, &lgi->lgi_attr);
2299         if (rc)
2300                 GOTO(out, rc);
2301
2302         if (!S_ISREG(lgi->lgi_attr.la_mode)) {
2303                 CERROR("%s: CATALOGS is not a regular file!: mode = %o\n",
2304                        o->do_lu.lo_dev->ld_obd->obd_name,
2305                        lgi->lgi_attr.la_mode);
2306                 GOTO(out, rc = -ENOENT);
2307         }
2308
2309         CDEBUG(D_CONFIG, "cat list: disk size=%d, read=%d\n",
2310                (int)lgi->lgi_attr.la_size, size);
2311
2312         /* return just number of llogs */
2313         if (idarray == NULL) {
2314                 rc = lgi->lgi_attr.la_size / sizeof(*idarray);
2315                 GOTO(out, rc);
2316         }
2317
2318         /* read for new ost index or for empty file */
2319         memset(idarray, 0, size);
2320         if (lgi->lgi_attr.la_size <= lgi->lgi_off)
2321                 GOTO(out, rc = 0);
2322         if (lgi->lgi_attr.la_size < lgi->lgi_off + size)
2323                 size = lgi->lgi_attr.la_size - lgi->lgi_off;
2324
2325         lgi->lgi_buf.lb_buf = idarray;
2326         lgi->lgi_buf.lb_len = size;
2327         rc = dt_record_read(env, o, &lgi->lgi_buf, &lgi->lgi_off);
2328         /* -EFAULT means the llog is a sparse file. This is not an error
2329          * after arbitrary OST index is supported. */
2330         if (rc < 0 && rc != -EFAULT) {
2331                 CERROR("%s: error reading CATALOGS: rc = %d\n",
2332                        o->do_lu.lo_dev->ld_obd->obd_name,  rc);
2333                 GOTO(out, rc);
2334         }
2335
2336         EXIT;
2337 out:
2338         dt_object_put(env, o);
2339         RETURN(rc);
2340 }
2341 EXPORT_SYMBOL(llog_osd_get_cat_list);
2342
2343 /**
2344  * Write the special file which contains the list of llog catalogs IDs
2345  *
2346  * This function writes the CATALOG file which contains the array of llog
2347  * catalogs IDs. It is used mostly to store OSP llogs indexed by OST/MDT
2348  * number.
2349  *
2350  * \param[in]  env      execution environment
2351  * \param[in]  d        corresponding storage device
2352  * \param[in]  idx      position to start from, usually OST/MDT index
2353  * \param[in]  count    how many catalog IDs to write
2354  * \param[out] idarray  the buffer with the data to write.
2355  * \param[in]  fid      LLOG_CATALOGS_OID for CATALOG object
2356  *
2357  * \retval              0 on successful write of catalog IDs
2358  * \retval              negative value on error
2359  */
2360 int llog_osd_put_cat_list(const struct lu_env *env, struct dt_device *d,
2361                           int idx, int count, struct llog_catid *idarray,
2362                           const struct lu_fid *fid)
2363 {
2364         struct llog_thread_info *lgi = llog_info(env);
2365         struct dt_object        *o = NULL;
2366         struct thandle          *th;
2367         int                      rc, size;
2368
2369         if (count == 0)
2370                 RETURN(0);
2371
2372         LASSERT(d);
2373
2374         size = sizeof(*idarray) * count;
2375         lgi->lgi_off = idx * sizeof(*idarray);
2376         lgi->lgi_fid = *fid;
2377
2378         o = dt_locate(env, d, &lgi->lgi_fid);
2379         if (IS_ERR(o))
2380                 RETURN(PTR_ERR(o));
2381
2382         if (!dt_object_exists(o))
2383                 GOTO(out, rc = -ENOENT);
2384
2385         rc = dt_attr_get(env, o, &lgi->lgi_attr);
2386         if (rc)
2387                 GOTO(out, rc);
2388
2389         if (!S_ISREG(lgi->lgi_attr.la_mode)) {
2390                 CERROR("%s: CATALOGS is not a regular file!: mode = %o\n",
2391                        o->do_lu.lo_dev->ld_obd->obd_name,
2392                        lgi->lgi_attr.la_mode);
2393                 GOTO(out, rc = -ENOENT);
2394         }
2395
2396         th = dt_trans_create(env, d);
2397         if (IS_ERR(th))
2398                 GOTO(out, rc = PTR_ERR(th));
2399
2400         lgi->lgi_buf.lb_len = size;
2401         lgi->lgi_buf.lb_buf = idarray;
2402         rc = dt_declare_record_write(env, o, &lgi->lgi_buf, lgi->lgi_off, th);
2403         if (rc)
2404                 GOTO(out_trans, rc);
2405
2406         /* For update log, this happens during initialization,
2407          * see lod_sub_prep_llog(), and we need make sure catlog
2408          * file ID is written to catlist file(committed) before
2409          * cross-MDT operation write update records to catlog FILE,
2410          * otherwise, during failover these update records might
2411          * missing */
2412         if (fid_is_update_log(fid))
2413                 th->th_sync = 1;
2414
2415         rc = dt_trans_start_local(env, d, th);
2416         if (rc)
2417                 GOTO(out_trans, rc);
2418
2419         th->th_wait_submit = 1;
2420
2421         rc = dt_record_write(env, o, &lgi->lgi_buf, &lgi->lgi_off, th);
2422         if (rc)
2423                 CDEBUG(D_INODE, "can't write CATALOGS at index %d: rc = %d\n",
2424                        idx, rc);
2425 out_trans:
2426         dt_trans_stop(env, d, th);
2427 out:
2428         dt_object_put(env, o);
2429         RETURN(rc);
2430 }
2431 EXPORT_SYMBOL(llog_osd_put_cat_list);