lustre/lvfs/llog_lvfs.c

   1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
   2  * vim:expandtab:shiftwidth=8:tabstop=8:
   3  *
   4  *  Copyright (C) 2001-2003 Cluster File Systems, Inc.
   5  *   Author: Andreas Dilger <adilger@clusterfs.com>
   6  *
   7  *   This file is part of Lustre, http://www.lustre.org.
   8  *
   9  *   Lustre is free software; you can redistribute it and/or
  10  *   modify it under the terms of version 2 of the GNU General Public
  11  *   License as published by the Free Software Foundation.
  12  *
  13  *   Lustre is distributed in the hope that it will be useful,
  14  *   but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16  *   GNU General Public License for more details.
  17  *
  18  *   You should have received a copy of the GNU General Public License
  19  *   along with Lustre; if not, write to the Free Software
  20  *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  21  *
  22  * OST<->MDS recovery logging infrastructure.
  23  *
  24  * Invariants in implementation:
  25  * - we do not share logs among different OST<->MDS connections, so that
  26  *   if an OST or MDS fails it need only look at log(s) relevant to itself
  27  */
  28
  29 #define DEBUG_SUBSYSTEM S_LOG
  30
  31 #ifndef EXPORT_SYMTAB
  32 #define EXPORT_SYMTAB
  33 #endif
  34
  35 #ifdef __KERNEL__
  36 #include <linux/fs.h>
  37 #else
  38 #include <liblustre.h>
  39 #endif
  40
  41 #include <linux/lvfs.h>
  42 #include <linux/lustre_fsfilt.h>
  43 #include <linux/lustre_log.h>
  44
  45 #ifdef __KERNEL__
  46
  47 static int llog_lvfs_pad(struct llog_ctxt *ctxt, struct l_file *file,
  48                          int len, int index)
  49 {
  50         struct llog_rec_hdr rec;
  51         struct llog_rec_tail tail;
  52         int rc;
  53         ENTRY;
  54
  55         LASSERT(len >= LLOG_MIN_REC_SIZE && (len & 0x7) == 0);
  56
  57         tail.lrt_len = rec.lrh_len = cpu_to_le32(len);
  58         tail.lrt_index = rec.lrh_index = cpu_to_le32(index);
  59         rec.lrh_type = 0;
  60
  61         rc = llog_fsfilt_write_record(ctxt, file, &rec, sizeof(rec),
  62                                       &file->f_pos, 0);
  63         if (rc) {
  64                 CERROR("error writing padding record: rc %d\n", rc);
  65                 goto out;
  66         }
  67
  68         file->f_pos += len - sizeof(rec) - sizeof(tail);
  69         rc = llog_fsfilt_write_record(ctxt, file, &tail, sizeof(tail),
  70                                       &file->f_pos, 0);
  71         if (rc) {
  72                 CERROR("error writing padding record: rc %d\n", rc);
  73                 goto out;
  74         }
  75
  76  out:
  77         RETURN(rc);
  78 }
  79
  80 static int llog_lvfs_write_blob(struct llog_ctxt *ctxt, struct l_file *file,
  81                                 struct llog_rec_hdr *rec, void *buf, loff_t off)
  82 {
  83         int rc;
  84         struct llog_rec_tail end;
  85         loff_t saved_off = file->f_pos;
  86         int buflen = le32_to_cpu(rec->lrh_len);
  87
  88         ENTRY;
  89         file->f_pos = off;
  90
  91         if (!buf) {
  92                 rc = llog_fsfilt_write_record(ctxt, file, rec, buflen,
  93                                               &file->f_pos, 0);
  94                 if (rc) {
  95                         CERROR("error writing log record: rc %d\n", rc);
  96                         goto out;
  97                 }
  98                 GOTO(out, rc = 0);
  99         }
 100
 101         /* the buf case */
 102         rec->lrh_len = cpu_to_le32(sizeof(*rec) + buflen + sizeof(end));
 103         rc = llog_fsfilt_write_record(ctxt, file, rec, sizeof(*rec),
 104                                       &file->f_pos, 0);
 105         if (rc) {
 106                 CERROR("error writing log hdr: rc %d\n", rc);
 107                 goto out;
 108         }
 109
 110         rc = llog_fsfilt_write_record(ctxt, file, buf, buflen,
 111                                       &file->f_pos, 0);
 112         if (rc) {
 113                 CERROR("error writing log buffer: rc %d\n", rc);
 114                 goto out;
 115         }
 116
 117         end.lrt_len = rec->lrh_len;
 118         end.lrt_index = rec->lrh_index;
 119         rc = llog_fsfilt_write_record(ctxt, file, &end, sizeof(end),
 120                                       &file->f_pos, 0);
 121         if (rc) {
 122                 CERROR("error writing log tail: rc %d\n", rc);
 123                 goto out;
 124         }
 125
 126         rc = 0;
 127  out:
 128         if (saved_off > file->f_pos)
 129                 file->f_pos = saved_off;
 130         LASSERT(rc <= 0);
 131         RETURN(rc);
 132 }
 133
 134 static int llog_lvfs_read_blob(struct llog_ctxt *ctxt, struct l_file *file,
 135                                void *buf, int size, loff_t off)
 136 {
 137         loff_t offset = off;
 138         int rc;
 139         ENTRY;
 140
 141         rc = llog_fsfilt_read_record(ctxt, file, buf, size, &offset);
 142         if (rc) {
 143                 CERROR("error reading log record: rc %d\n", rc);
 144                 RETURN(rc);
 145         }
 146         RETURN(0);
 147 }
 148
 149 static int llog_lvfs_read_header(struct llog_handle *handle)
 150 {
 151         struct llog_ctxt *ctxt = handle->lgh_ctxt;
 152         int rc;
 153         ENTRY;
 154
 155         LASSERT(sizeof(*handle->lgh_hdr) == LLOG_CHUNK_SIZE);
 156         LASSERT(ctxt != NULL);
 157
 158         if (handle->lgh_file->f_dentry->d_inode->i_size == 0) {
 159                 CDEBUG(D_HA, "not reading header from 0-byte log\n");
 160                 RETURN(LLOG_EEMPTY);
 161         }
 162
 163         rc = llog_lvfs_read_blob(ctxt, handle->lgh_file, handle->lgh_hdr,
 164                                  LLOG_CHUNK_SIZE, 0);
 165         if (rc)
 166                 CERROR("error reading log header\n");
 167
 168         handle->lgh_last_idx = le32_to_cpu(handle->lgh_hdr->llh_tail.lrt_index);
 169         handle->lgh_file->f_pos = handle->lgh_file->f_dentry->d_inode->i_size;
 170
 171         RETURN(rc);
 172 }
 173
 174 /* returns negative in on error; 0 if success && reccookie == 0; 1 otherwise */
 175 /* appends if idx == -1, otherwise overwrites record idx. */
 176 static int llog_lvfs_write_rec(struct llog_handle *loghandle,
 177                                struct llog_rec_hdr *rec,
 178                                struct llog_cookie *reccookie,
 179                                int cookiecount,
 180                                void *buf, int idx)
 181 {
 182         struct llog_log_hdr *llh;
 183         int reclen = le32_to_cpu(rec->lrh_len), index, rc;
 184         struct llog_rec_tail *lrt;
 185         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 186         struct file *file;
 187         loff_t offset;
 188         size_t left;
 189         ENTRY;
 190
 191         llh = loghandle->lgh_hdr;
 192         file = loghandle->lgh_file;
 193
 194         /* record length should not bigger than LLOG_CHUNK_SIZE */
 195         if (buf)
 196                 rc = (reclen > LLOG_CHUNK_SIZE - sizeof(struct llog_rec_hdr)
 197                       - sizeof(struct llog_rec_tail)) ? -E2BIG : 0;
 198         else
 199                 rc = (reclen > LLOG_CHUNK_SIZE) ? -E2BIG : 0;
 200         if (rc)
 201                 RETURN(rc);
 202
 203         if (idx != -1) {
 204                 loff_t saved_offset;
 205
 206                 /* no header: only allowed to insert record 1 */
 207                 if (idx != 1 && !file->f_dentry->d_inode->i_size) {
 208                         CERROR("idx != -1 in empty log\n");
 209                         LBUG();
 210                 }
 211
 212                 if (idx && llh->llh_size && llh->llh_size != reclen)
 213                         RETURN(-EINVAL);
 214
 215                 rc = llog_lvfs_write_blob(ctxt, file, &llh->llh_hdr, NULL, 0);
 216                 /* we are done if we only write the header or on error */
 217                 if (rc || idx == 0)
 218                         RETURN(rc);
 219
 220                 saved_offset = sizeof(*llh) + (idx-1)*le32_to_cpu(rec->lrh_len);
 221                 rc = llog_lvfs_write_blob(ctxt, file, rec, buf, saved_offset);
 222                 if (rc == 0 && reccookie) {
 223                         reccookie->lgc_lgl = loghandle->lgh_id;
 224                         reccookie->lgc_index = idx;
 225                         rc = 1;
 226                 }
 227                 RETURN(rc);
 228         }
 229
 230         /* Make sure that records don't cross a chunk boundary, so we can
 231          * process them page-at-a-time if needed.  If it will cross a chunk
 232          * boundary, write in a fake (but referenced) entry to pad the chunk.
 233          *
 234          * We know that llog_current_log() will return a loghandle that is
 235          * big enough to hold reclen, so all we care about is padding here.
 236          */
 237         left = LLOG_CHUNK_SIZE - (file->f_pos & (LLOG_CHUNK_SIZE - 1));
 238         if (buf)
 239                 reclen = sizeof(*rec) + le32_to_cpu(rec->lrh_len) +
 240                          sizeof(struct llog_rec_tail);
 241
 242         /* NOTE: padding is a record, but no bit is set */
 243         if (left != 0 && left != reclen &&
 244             left < (reclen + LLOG_MIN_REC_SIZE)) {
 245                 loghandle->lgh_last_idx++;
 246                 rc = llog_lvfs_pad(ctxt, file, left, loghandle->lgh_last_idx);
 247                 if (rc)
 248                         RETURN(rc);
 249         }
 250
 251         loghandle->lgh_last_idx++;
 252         index = loghandle->lgh_last_idx;
 253         rec->lrh_index = cpu_to_le32(index);
 254         if (buf == NULL) {
 255                 lrt = (void *)rec + le32_to_cpu(rec->lrh_len) - sizeof(*lrt);
 256                 lrt->lrt_len = rec->lrh_len;
 257                 lrt->lrt_index = rec->lrh_index;
 258         }
 259         if (ext2_set_bit(index, llh->llh_bitmap)) {
 260                 CERROR("argh, index %u already set in log bitmap?\n", index);
 261                 LBUG(); /* should never happen */
 262         }
 263         llh->llh_count = cpu_to_le32(le32_to_cpu(llh->llh_count) + 1);
 264         llh->llh_tail.lrt_index = cpu_to_le32(index);
 265
 266         offset = 0;
 267         rc = llog_lvfs_write_blob(ctxt, file, &llh->llh_hdr, NULL, 0);
 268         if (rc)
 269                 RETURN(rc);
 270
 271         rc = llog_lvfs_write_blob(ctxt, file, rec, buf, file->f_pos);
 272         if (rc)
 273                 RETURN(rc);
 274
 275         CDEBUG(D_HA, "added record "LPX64": idx: %u, %u bytes\n",
 276                loghandle->lgh_id.lgl_oid, index, le32_to_cpu(rec->lrh_len));
 277         if (rc == 0 && reccookie) {
 278                 reccookie->lgc_lgl = loghandle->lgh_id;
 279                 reccookie->lgc_index = index;
 280                 if (le32_to_cpu(rec->lrh_type) == MDS_UNLINK_REC)
 281                         reccookie->lgc_subsys = LLOG_UNLINK_ORIG_CTXT;
 282                 else if (le32_to_cpu(rec->lrh_type) == OST_SZ_REC)
 283                         reccookie->lgc_subsys = LLOG_SIZE_ORIG_CTXT;
 284                 else if (le32_to_cpu(rec->lrh_type) == OST_RAID1_REC)
 285                         reccookie->lgc_subsys = LLOG_RD1_ORIG_CTXT;
 286                 else
 287                         reccookie->lgc_subsys = -1;
 288                 rc = 1;
 289         }
 290         if (rc == 0 && (le32_to_cpu(rec->lrh_type) == LLOG_GEN_REC ||
 291             le32_to_cpu(rec->lrh_type) == SMFS_UPDATE_REC))
 292                 rc = 1;
 293
 294         RETURN(rc);
 295 }
 296
 297 /* We can skip reading at least as many log blocks as the number of
 298 * minimum sized log records we are skipping.  If it turns out
 299 * that we are not far enough along the log (because the
 300 * actual records are larger than minimum size) we just skip
 301 * some more records. */
 302
 303 static void llog_skip_over(__u64 *off, int curr, int goal)
 304 {
 305         if (goal <= curr)
 306                 return;
 307         *off = (*off + (goal-curr-1) * LLOG_MIN_REC_SIZE) &
 308                 ~(LLOG_CHUNK_SIZE - 1);
 309 }
 310
 311 /* sets:
 312  *  - curr_offset to the furthest point read in the log file
 313  *  - curr_idx to the log index preceeding curr_offset
 314  * returns -EIO/-EINVAL on error
 315  */
 316 static int llog_lvfs_next_block(struct llog_handle *loghandle, int *curr_idx,
 317                                 int next_idx, __u64 *curr_offset, void *buf,
 318                                 int len)
 319 {
 320         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 321         int rc;
 322         ENTRY;
 323
 324         if (len == 0 || len & (LLOG_CHUNK_SIZE - 1))
 325                 RETURN(-EINVAL);
 326
 327         CDEBUG(D_OTHER, "looking for log index %u (cur idx %u off "LPU64")\n",
 328                next_idx, *curr_idx, *curr_offset);
 329
 330         while (*curr_offset < loghandle->lgh_file->f_dentry->d_inode->i_size) {
 331                 struct llog_rec_hdr *rec;
 332                 struct llog_rec_tail *tail;
 333                 loff_t ppos;
 334
 335                 llog_skip_over(curr_offset, *curr_idx, next_idx);
 336
 337                 ppos = *curr_offset;
 338                 rc = llog_fsfilt_read_record(ctxt, loghandle->lgh_file,
 339                                              buf, len, &ppos);
 340
 341                 if (rc) {
 342                         CERROR("Cant read llog block at log id "LPU64
 343                                "/%u offset "LPU64"\n",
 344                                loghandle->lgh_id.lgl_oid,
 345                                loghandle->lgh_id.lgl_ogen,
 346                                *curr_offset);
 347                         RETURN(rc);
 348                 }
 349
 350                 /* put number of bytes read into rc to make code simpler */
 351                 rc = ppos - *curr_offset;
 352                 *curr_offset = ppos;
 353
 354                 if (rc == 0) /* end of file, nothing to do */
 355                         RETURN(0);
 356
 357                 if (rc < sizeof(*tail)) {
 358                         CERROR("Invalid llog block at log id "LPU64"/%u offset "
 359                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 360                                loghandle->lgh_id.lgl_ogen, *curr_offset);
 361                         RETURN(-EINVAL);
 362                 }
 363
 364                 tail = buf + rc - sizeof(struct llog_rec_tail);
 365                 *curr_idx = le32_to_cpu(tail->lrt_index);
 366
 367                 /* this shouldn't happen */
 368                 if (tail->lrt_index == 0) {
 369                         CERROR("Invalid llog tail at log id "LPU64"/%u offset "
 370                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 371                                loghandle->lgh_id.lgl_ogen, *curr_offset);
 372                         RETURN(-EINVAL);
 373                 }
 374                 if (le32_to_cpu(tail->lrt_index) < next_idx)
 375                         continue;
 376
 377                 /* sanity check that the start of the new buffer is no farther
 378                  * than the record that we wanted.  This shouldn't happen. */
 379                 rec = buf;
 380                 if (le32_to_cpu(rec->lrh_index) > next_idx) {
 381                         CERROR("missed desired record? %u > %u\n",
 382                                le32_to_cpu(rec->lrh_index), next_idx);
 383                         RETURN(-ENOENT);
 384                 }
 385                 RETURN(0);
 386         }
 387         RETURN(-EIO);
 388 }
 389
 390 static int llog_lvfs_prev_block(struct llog_handle *loghandle,
 391                                 int prev_idx, void *buf, int len)
 392 {
 393         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 394         __u64 curr_offset;
 395         int rc;
 396         ENTRY;
 397
 398         if (len == 0 || len & (LLOG_CHUNK_SIZE - 1))
 399                 RETURN(-EINVAL);
 400
 401         CDEBUG(D_OTHER, "looking for log index %u n", prev_idx);
 402
 403         curr_offset = LLOG_CHUNK_SIZE;
 404         llog_skip_over(&curr_offset, 0, prev_idx);
 405
 406         while (curr_offset < loghandle->lgh_file->f_dentry->d_inode->i_size) {
 407                 struct llog_rec_hdr *rec;
 408                 struct llog_rec_tail *tail;
 409                 loff_t ppos;
 410
 411                 ppos = curr_offset;
 412                 rc = llog_fsfilt_read_record(ctxt, loghandle->lgh_file,
 413                                              buf, len, &ppos);
 414
 415                 if (rc) {
 416                         CERROR("Cant read llog block at log id "LPU64
 417                                "/%u offset "LPU64"\n",
 418                                loghandle->lgh_id.lgl_oid,
 419                                loghandle->lgh_id.lgl_ogen,
 420                                curr_offset);
 421                         RETURN(rc);
 422                 }
 423
 424                 /* put number of bytes read into rc to make code simpler */
 425                 rc = ppos - curr_offset;
 426                 curr_offset = ppos;
 427
 428                 if (rc == 0) /* end of file, nothing to do */
 429                         RETURN(0);
 430
 431                 if (rc < sizeof(*tail)) {
 432                         CERROR("Invalid llog block at log id "LPU64"/%u offset "
 433                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 434                                loghandle->lgh_id.lgl_ogen, curr_offset);
 435                         RETURN(-EINVAL);
 436                 }
 437
 438                 tail = buf + rc - sizeof(struct llog_rec_tail);
 439
 440                 /* this shouldn't happen */
 441                 if (tail->lrt_index == 0) {
 442                         CERROR("Invalid llog tail at log id "LPU64"/%u offset "
 443                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 444                                loghandle->lgh_id.lgl_ogen, curr_offset);
 445                         RETURN(-EINVAL);
 446                 }
 447                 if (le32_to_cpu(tail->lrt_index) < prev_idx)
 448                         continue;
 449
 450                 /* sanity check that the start of the new buffer is no farther
 451                  * than the record that we wanted.  This shouldn't happen. */
 452                 rec = buf;
 453                 if (le32_to_cpu(rec->lrh_index) > prev_idx) {
 454                         CERROR("missed desired record? %u > %u\n",
 455                                le32_to_cpu(rec->lrh_index), prev_idx);
 456                         RETURN(-ENOENT);
 457                 }
 458                 RETURN(0);
 459         }
 460         RETURN(-EIO);
 461 }
 462
 463 static struct file *llog_filp_open(char *name, int flags, int mode)
 464 {
 465         char *logname;
 466         struct file *filp;
 467         int len;
 468
 469         OBD_ALLOC(logname, PATH_MAX);
 470         if (logname == NULL)
 471                 return ERR_PTR(-ENOMEM);
 472
 473         len = snprintf(logname, PATH_MAX, "LOGS/%s", name);
 474         if (len >= PATH_MAX - 1) {
 475                 filp = ERR_PTR(-ENAMETOOLONG);
 476         } else {
 477                 filp = l_filp_open(logname, flags, mode);
 478                 if (IS_ERR(filp)) {
 479                         CERROR("logfile creation %s: %ld\n", logname,
 480                                PTR_ERR(filp));
 481                 }
 482         }
 483
 484         OBD_FREE(logname, PATH_MAX);
 485         return filp;
 486 }
 487
 488 /* creates object for the case when we have no obd (smfs). */
 489 static struct file *
 490 llog_object_create_alone(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 491 {
 492         unsigned int tmpname = ll_insecure_random_int();
 493         char fidname[LL_FID_NAMELEN];
 494         struct file *filp;
 495         struct dentry *new_child, *parent;
 496         void *handle;
 497         int rc = 0, err, namelen;
 498         ENTRY;
 499
 500         sprintf(fidname, "OBJECTS/%u", tmpname);
 501         filp = filp_open(fidname, O_CREAT | O_EXCL, 0644);
 502         if (IS_ERR(filp)) {
 503                 rc = PTR_ERR(filp);
 504                 if (rc == -EEXIST) {
 505                         CERROR("impossible object name collision %u\n",
 506                                tmpname);
 507                         LBUG();
 508                 }
 509                 CERROR("error creating tmp object %u: rc %d\n", tmpname, rc);
 510                 RETURN(filp);
 511         }
 512
 513         namelen = ll_fid2str(fidname, filp->f_dentry->d_inode->i_ino,
 514                              filp->f_dentry->d_inode->i_generation);
 515         parent = filp->f_dentry->d_parent;
 516         down(&parent->d_inode->i_sem);
 517         new_child = lookup_one_len(fidname, parent, namelen);
 518         if (IS_ERR(new_child)) {
 519                 CERROR("getting neg dentry for obj rename: %d\n", rc);
 520                 GOTO(out_close, rc = PTR_ERR(new_child));
 521         }
 522         if (new_child->d_inode != NULL) {
 523                 CERROR("impossible non-negative obj dentry %lu:%u!\n",
 524                        filp->f_dentry->d_inode->i_ino,
 525                        filp->f_dentry->d_inode->i_generation);
 526                 LBUG();
 527         }
 528
 529         handle = llog_fsfilt_start(ctxt, parent->d_inode, FSFILT_OP_RENAME, NULL);
 530         if (IS_ERR(handle))
 531                 GOTO(out_dput, rc = PTR_ERR(handle));
 532
 533         lock_kernel();
 534         rc = vfs_rename(parent->d_inode, filp->f_dentry,
 535                         parent->d_inode, new_child);
 536         unlock_kernel();
 537         if (rc)
 538                 CERROR("error renaming new object %lu:%u: rc %d\n",
 539                        filp->f_dentry->d_inode->i_ino,
 540                        filp->f_dentry->d_inode->i_generation, rc);
 541
 542         err = llog_fsfilt_commit(ctxt, parent->d_inode, handle, 0);
 543         if (!rc)
 544                 rc = err;
 545
 546 out_dput:
 547         dput(new_child);
 548 out_close:
 549         up(&parent->d_inode->i_sem);
 550         if (rc) {
 551                 filp_close(filp, 0);
 552                 filp = (struct file *)rc;
 553         } else {
 554                 /* FIXME: is this group 1 is correct? */
 555                 lgh_id->lgl_ogr = 1;
 556                 lgh_id->lgl_oid = filp->f_dentry->d_inode->i_ino;
 557                 lgh_id->lgl_ogen = filp->f_dentry->d_inode->i_generation;
 558         }
 559
 560         RETURN(filp);
 561 }
 562
 563 /* creates object for generic case (obd exists) */
 564 static struct file *
 565 llog_object_create_generic(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 566 {
 567         int rc = 0;
 568         struct file *filp;
 569         struct dentry *dchild;
 570         struct obd_device *obd;
 571         struct obdo *oa = NULL;
 572         int open_flags = O_RDWR | O_CREAT | O_LARGEFILE;
 573         ENTRY;
 574
 575         obd = ctxt->loc_exp->exp_obd;
 576
 577         /* this is important to work here over obd_create() as it manages
 578            groups and we need it. Yet another reason is that mds_obd_create()
 579            is fully the same as old version of this function and this helps
 580            us to avoid code duplicating and layering violating. */
 581         OBD_ALLOC(oa, sizeof(*oa));
 582         if (!oa)
 583                 RETURN(ERR_PTR(-ENOMEM));
 584
 585         oa->o_gr = FILTER_GROUP_LLOG;
 586         oa->o_valid = OBD_MD_FLGENER | OBD_MD_FLGROUP;
 587         rc = obd_create(ctxt->loc_exp, oa, NULL, NULL);
 588         if (rc)
 589                 GOTO(out_free_oa, rc);
 590
 591         dchild = obd_lvfs_fid2dentry(ctxt->loc_exp, oa->o_id,
 592                                      oa->o_generation, oa->o_gr);
 593         if (IS_ERR(dchild))
 594                 GOTO(out_free_oa, rc = PTR_ERR(dchild));
 595
 596         filp = l_dentry_open(&obd->obd_lvfs_ctxt, dchild,
 597                              open_flags);
 598         if (IS_ERR(filp)) {
 599                 l_dput(dchild);
 600                 GOTO(out_free_oa, rc = PTR_ERR(filp));
 601         }
 602
 603         /* group 1 is not longer valid, we use the group which is set
 604            by obd_create()->mds_obd_create(). */
 605         lgh_id->lgl_ogr = oa->o_gr;
 606         lgh_id->lgl_oid = oa->o_id;
 607         lgh_id->lgl_ogen = oa->o_generation;
 608         OBD_FREE(oa, sizeof(*oa));
 609         RETURN(filp);
 610
 611 out_free_oa:
 612         OBD_FREE(oa, sizeof(*oa));
 613         RETURN(ERR_PTR(rc));
 614 }
 615
 616 static struct file *
 617 llog_object_create(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 618 {
 619         if (ctxt->loc_alone)
 620                 return llog_object_create_alone(ctxt, lgh_id);
 621         else
 622                 return llog_object_create_generic(ctxt, lgh_id);
 623 }
 624
 625 static int llog_add_link_object(struct llog_ctxt *ctxt, struct llog_logid logid,
 626                                 struct dentry *dentry)
 627 {
 628         struct dentry *new_child;
 629         char fidname[LL_FID_NAMELEN];
 630         void *handle;
 631         int namelen, rc = 0, err;
 632         ENTRY;
 633
 634         namelen = ll_fid2str(fidname, logid.lgl_oid, logid.lgl_ogen);
 635         down(&ctxt->loc_objects_dir->d_inode->i_sem);
 636         new_child = lookup_one_len(fidname, ctxt->loc_objects_dir, namelen);
 637         if (IS_ERR(new_child)) {
 638                 CERROR("getting neg dentry for obj rename: %d\n", rc);
 639                 GOTO(out, rc = PTR_ERR(new_child));
 640         }
 641         if (new_child->d_inode == dentry->d_inode)
 642                 GOTO(out_dput, rc);
 643         if (new_child->d_inode != NULL) {
 644                 CERROR("impossible non-negative obj dentry "LPX64":%u!\n",
 645                        logid.lgl_oid, logid.lgl_ogen);
 646                 LBUG();
 647         }
 648         handle = llog_fsfilt_start(ctxt, ctxt->loc_objects_dir->d_inode,
 649                                    FSFILT_OP_LINK, NULL);
 650         if (IS_ERR(handle))
 651                 GOTO(out_dput, rc = PTR_ERR(handle));
 652
 653         lock_kernel();
 654         rc = vfs_link(dentry, ctxt->loc_objects_dir->d_inode, new_child);
 655         unlock_kernel();
 656         if (rc)
 657                 CERROR("error link new object "LPX64":%u: rc %d\n",
 658                        logid.lgl_oid, logid.lgl_ogen, rc);
 659         err = llog_fsfilt_commit(ctxt, ctxt->loc_objects_dir->d_inode, handle, 0);
 660 out_dput:
 661         l_dput(new_child);
 662 out:
 663         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 664         RETURN(rc);
 665 }
 666
 667 static int llog_lvfs_create(struct llog_ctxt *ctxt, struct llog_handle **res,
 668                             struct llog_logid *logid, char *name)
 669 {
 670         struct llog_handle *handle;
 671         struct lvfs_run_ctxt saved;
 672         int rc = 0;
 673         int open_flags = O_RDWR | O_CREAT | O_LARGEFILE;
 674         ENTRY;
 675
 676         handle = llog_alloc_handle();
 677         if (handle == NULL)
 678                 RETURN(-ENOMEM);
 679         *res = handle;
 680
 681         LASSERT(ctxt);
 682         if (ctxt->loc_lvfs_ctxt)
 683                 push_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 684
 685         if (logid != NULL) {
 686                 char logname[LL_FID_NAMELEN + 10] = "OBJECTS/";
 687                 char fidname[LL_FID_NAMELEN];
 688                 ll_fid2str(fidname, logid->lgl_oid, logid->lgl_ogen);
 689                 strcat(logname, fidname);
 690
 691                 handle->lgh_file = filp_open(logname, O_RDWR | O_LARGEFILE, 0644);
 692                 if (IS_ERR(handle->lgh_file)) {
 693                         CERROR("cannot open %s file, error = %ld\n",
 694                                logname, PTR_ERR(handle->lgh_file));
 695                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 696                 }
 697                 if (!S_ISREG(handle->lgh_file->f_dentry->d_inode->i_mode)) {
 698                         CERROR("%s is not a regular file!: mode = %o\n", logname,
 699                                handle->lgh_file->f_dentry->d_inode->i_mode);
 700                         GOTO(cleanup, rc = -ENOENT);
 701                 }
 702                 LASSERT(handle->lgh_file->f_dentry->d_parent == ctxt->loc_objects_dir);
 703                 handle->lgh_id = *logid;
 704         } else if (name) {
 705                 handle->lgh_file = llog_filp_open(name, open_flags, 0644);
 706                 if (IS_ERR(handle->lgh_file)) {
 707                         CERROR("cannot open %s file, error = %ld\n",
 708                                name, PTR_ERR(handle->lgh_file));
 709                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 710                 }
 711                 LASSERT(handle->lgh_file->f_dentry->d_parent == ctxt->loc_logs_dir);
 712
 713                 handle->lgh_id.lgl_ogr = 1;
 714                 handle->lgh_id.lgl_oid = handle->lgh_file->f_dentry->d_inode->i_ino;
 715                 handle->lgh_id.lgl_ogen = handle->lgh_file->f_dentry->d_inode->i_generation;
 716                 rc = llog_add_link_object(ctxt, handle->lgh_id, handle->lgh_file->f_dentry);
 717                 if (rc)
 718                         GOTO(cleanup, rc);
 719         } else {
 720                 handle->lgh_file = llog_object_create(ctxt, &handle->lgh_id);
 721                 if (IS_ERR(handle->lgh_file)) {
 722                         CERROR("cannot create llog object, error = %ld\n",
 723                                PTR_ERR(handle->lgh_file));
 724                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 725                 }
 726         }
 727
 728         handle->lgh_ctxt = ctxt;
 729 finish:
 730         if (ctxt->loc_lvfs_ctxt)
 731                 pop_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 732         RETURN(rc);
 733 cleanup:
 734         llog_free_handle(handle);
 735         goto finish;
 736 }
 737
 738 static int llog_lvfs_close(struct llog_handle *handle)
 739 {
 740         int rc;
 741         ENTRY;
 742
 743         rc = filp_close(handle->lgh_file, 0);
 744         if (rc)
 745                 CERROR("error closing log: rc %d\n", rc);
 746         RETURN(rc);
 747 }
 748
 749 static int llog_lvfs_destroy(struct llog_handle *loghandle)
 750 {
 751         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 752         struct lvfs_run_ctxt saved;
 753         struct dentry *fdentry;
 754         struct inode *parent_inode;
 755         char fidname[LL_FID_NAMELEN];
 756         void *handle;
 757         int rc = -EINVAL, err, namelen;
 758         ENTRY;
 759
 760         if (ctxt->loc_lvfs_ctxt)
 761                 push_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 762
 763         fdentry = loghandle->lgh_file->f_dentry;
 764         parent_inode = fdentry->d_parent->d_inode;
 765
 766         if (!strcmp(fdentry->d_parent->d_name.name, "LOGS")) {
 767                 LASSERT(parent_inode == ctxt->loc_logs_dir->d_inode);
 768
 769                 namelen = ll_fid2str(fidname, fdentry->d_inode->i_ino,
 770                                      fdentry->d_inode->i_generation);
 771                 dget(fdentry);
 772                 rc = llog_lvfs_close(loghandle);
 773                 if (rc) {
 774                         dput(fdentry);
 775                         GOTO(out, rc);
 776                 }
 777
 778                 handle = llog_fsfilt_start(ctxt, parent_inode,
 779                                            FSFILT_OP_UNLINK, NULL);
 780                 if (IS_ERR(handle)) {
 781                         dput(fdentry);
 782                         GOTO(out, rc = PTR_ERR(handle));
 783                 }
 784
 785                 down(&parent_inode->i_sem);
 786                 rc = vfs_unlink(parent_inode, fdentry);
 787                 up(&parent_inode->i_sem);
 788                 dput(fdentry);
 789
 790                 if (!rc) {
 791                         down(&ctxt->loc_objects_dir->d_inode->i_sem);
 792                         fdentry = lookup_one_len(fidname, ctxt->loc_objects_dir,
 793                                                  namelen);
 794                         if (fdentry == NULL || fdentry->d_inode == NULL) {
 795                                 CERROR("destroy non_existent object %s\n", fidname);
 796                                 GOTO(out_err, rc = IS_ERR(fdentry) ?
 797                                      PTR_ERR(fdentry) : -ENOENT);
 798                         }
 799                         rc = vfs_unlink(ctxt->loc_objects_dir->d_inode, fdentry);
 800                         l_dput(fdentry);
 801 out_err:
 802                         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 803                 }
 804                 err = llog_fsfilt_commit(ctxt, parent_inode, handle, 0);
 805                 if (err && !rc)
 806                         err = rc;
 807
 808                 GOTO(out, rc);
 809         }
 810         if (ctxt->loc_alone) {
 811                 if (!strcmp(fdentry->d_parent->d_name.name, "OBJECTS")) {
 812                         LASSERT(parent_inode == ctxt->loc_objects_dir->d_inode);
 813
 814                         dget(fdentry);
 815                         rc = llog_lvfs_close(loghandle);
 816                         if (rc == 0) {
 817                                 down(&parent_inode->i_sem);
 818                                 rc = vfs_unlink(parent_inode, fdentry);
 819                                 up(&parent_inode->i_sem);
 820                         }
 821                         dput(fdentry);
 822                 }
 823         } else {
 824                 struct obdo *oa = NULL;
 825
 826                 OBD_ALLOC(oa, sizeof(*oa));
 827                 if (!oa)
 828                         GOTO(out, rc = -ENOMEM);
 829
 830                 oa->o_id = loghandle->lgh_id.lgl_oid;
 831                 oa->o_gr = loghandle->lgh_id.lgl_ogr;
 832                 oa->o_generation = loghandle->lgh_id.lgl_ogen;
 833                 oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLGENER;
 834
 835                 rc = llog_lvfs_close(loghandle);
 836                 if (rc)
 837                         GOTO(out_free_oa, rc);
 838
 839                 rc = obd_destroy(loghandle->lgh_ctxt->loc_exp, oa, NULL, NULL);
 840 out_free_oa:
 841                 OBD_FREE(oa, sizeof(*oa));
 842         }
 843 out:
 844         if (ctxt->loc_lvfs_ctxt)
 845                 pop_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 846         RETURN(rc);
 847 }
 848
 849 /* reads the catalog list */
 850 int llog_get_cat_list(struct lvfs_run_ctxt *ctxt,
 851                       struct fsfilt_operations *fsops, char *name,
 852                       int count, struct llog_catid *idarray)
 853 {
 854         struct lvfs_run_ctxt saved;
 855         struct l_file *file;
 856         int size = sizeof(*idarray) * count;
 857         loff_t off = 0;
 858         int rc;
 859
 860         LASSERT(count);
 861
 862         if (ctxt)
 863                 push_ctxt(&saved, ctxt, NULL);
 864         file = l_filp_open(name, O_RDWR | O_CREAT | O_LARGEFILE, 0700);
 865         if (!file || IS_ERR(file)) {
 866                 rc = PTR_ERR(file);
 867                 CERROR("OBD filter: cannot open/create %s: rc = %d\n",
 868                        name, rc);
 869                 GOTO(out, rc);
 870         }
 871
 872         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 873                 CERROR("%s is not a regular file!: mode = %o\n", name,
 874                        file->f_dentry->d_inode->i_mode);
 875                 GOTO(out, rc = -ENOENT);
 876         }
 877
 878         rc = fsops->fs_read_record(file, idarray, size, &off);
 879         if (rc) {
 880                 CDEBUG(D_INODE,"OBD filter: error reading %s: rc %d\n",
 881                        name, rc);
 882                 GOTO(out, rc);
 883         }
 884
 885  out:
 886         if (file && !IS_ERR(file))
 887                 rc = filp_close(file, 0);
 888         if (ctxt)
 889                 pop_ctxt(&saved, ctxt, NULL);
 890         RETURN(rc);
 891 }
 892 EXPORT_SYMBOL(llog_get_cat_list);
 893
 894 /* writes the cat list */
 895 int llog_put_cat_list(struct lvfs_run_ctxt *ctxt,
 896                       struct fsfilt_operations *fsops, char *name,
 897                       int count, struct llog_catid *idarray)
 898 {
 899         struct lvfs_run_ctxt saved;
 900         struct l_file *file;
 901         int size = sizeof(*idarray) * count;
 902         loff_t off = 0;
 903         int rc;
 904
 905         LASSERT(count);
 906
 907         if (ctxt)
 908                 push_ctxt(&saved, ctxt, NULL);
 909         file = filp_open(name, O_RDWR | O_CREAT | O_LARGEFILE, 0700);
 910         if (!file || IS_ERR(file)) {
 911                 rc = PTR_ERR(file);
 912                 CERROR("OBD filter: cannot open/create %s: rc = %d\n",
 913                        name, rc);
 914                 GOTO(out, rc);
 915         }
 916
 917         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 918                 CERROR("%s is not a regular file!: mode = %o\n", name,
 919                        file->f_dentry->d_inode->i_mode);
 920                 GOTO(out, rc = -ENOENT);
 921         }
 922
 923         rc = fsops->fs_write_record(file, idarray, size, &off, 1);
 924         if (rc) {
 925                 CDEBUG(D_INODE,"OBD filter: error reading %s: rc %d\n",
 926                        name, rc);
 927                 GOTO(out, rc);
 928         }
 929
 930  out:
 931         if (file && !IS_ERR(file))
 932                 rc = filp_close(file, 0);
 933         if (ctxt)
 934                 pop_ctxt(&saved, ctxt, NULL);
 935         RETURN(rc);
 936 }
 937 EXPORT_SYMBOL(llog_put_cat_list);
 938
 939 struct llog_operations llog_lvfs_ops = {
 940         lop_create:      llog_lvfs_create,
 941         lop_destroy:     llog_lvfs_destroy,
 942         lop_close:       llog_lvfs_close,
 943         lop_read_header: llog_lvfs_read_header,
 944         lop_write_rec:   llog_lvfs_write_rec,
 945         lop_next_block:  llog_lvfs_next_block,
 946         lop_prev_block:  llog_lvfs_prev_block,
 947 };
 948 EXPORT_SYMBOL(llog_lvfs_ops);
 949
 950 #else /* !__KERNEL__ */
 951
 952 static int llog_lvfs_read_header(struct llog_handle *handle)
 953 {
 954         LBUG();
 955         return 0;
 956 }
 957
 958 static int llog_lvfs_write_rec(struct llog_handle *loghandle,
 959                                struct llog_rec_hdr *rec,
 960                                struct llog_cookie *reccookie, int cookiecount,
 961                                void *buf, int idx)
 962 {
 963         LBUG();
 964         return 0;
 965 }
 966
 967 static int llog_lvfs_create(struct llog_ctxt *ctxt, struct llog_handle **res,
 968                             struct llog_logid *logid, char *name)
 969 {
 970         LBUG();
 971         return 0;
 972 }
 973
 974 static int llog_lvfs_close(struct llog_handle *handle)
 975 {
 976         LBUG();
 977         return 0;
 978 }
 979
 980 static int llog_lvfs_destroy(struct llog_handle *handle)
 981 {
 982         LBUG();
 983         return 0;
 984 }
 985
 986 int llog_get_cat_list(struct lvfs_run_ctxt *ctxt,
 987                       struct fsfilt_operations *fsops, char *name,
 988                       int count, struct llog_catid *idarray)
 989 {
 990         LBUG();
 991         return 0;
 992 }
 993
 994 int llog_put_cat_list(struct lvfs_run_ctxt *ctxt,
 995                       struct fsfilt_operations *fsops, char *name,
 996                       int count, struct llog_catid *idarray)
 997 {
 998         LBUG();
 999         return 0;
1000 }
1001
1002 int llog_lvfs_prev_block(struct llog_handle *loghandle,
1003                          int prev_idx, void *buf, int len)
1004 {
1005         LBUG();
1006         return 0;
1007 }
1008
1009 int llog_lvfs_next_block(struct llog_handle *loghandle, int *curr_idx,
1010                          int next_idx, __u64 *offset, void *buf, int len)
1011 {
1012         LBUG();
1013         return 0;
1014 }
1015
1016 struct llog_operations llog_lvfs_ops = {
1017         lop_create:      llog_lvfs_create,
1018         lop_destroy:     llog_lvfs_destroy,
1019         lop_close:       llog_lvfs_close,
1020         lop_read_header: llog_lvfs_read_header,
1021         lop_write_rec:   llog_lvfs_write_rec,
1022         lop_next_block:  llog_lvfs_next_block,
1023         lop_prev_block:  llog_lvfs_prev_block,
1024 };
1025 #endif