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                 /* if it's the last idx in log file, then return -ENOSPC */
 250                 if (loghandle->lgh_last_idx == LLOG_BITMAP_SIZE(llh) - 1)
 251                         RETURN(-ENOSPC);
 252         }
 253
 254         loghandle->lgh_last_idx++;
 255         index = loghandle->lgh_last_idx;
 256         rec->lrh_index = cpu_to_le32(index);
 257         if (buf == NULL) {
 258                 lrt = (void *)rec + le32_to_cpu(rec->lrh_len) - sizeof(*lrt);
 259                 lrt->lrt_len = rec->lrh_len;
 260                 lrt->lrt_index = rec->lrh_index;
 261         }
 262         if (ext2_set_bit(index, llh->llh_bitmap)) {
 263                 CERROR("argh, index %u already set in log bitmap?\n", index);
 264                 LBUG(); /* should never happen */
 265         }
 266         llh->llh_count = cpu_to_le32(le32_to_cpu(llh->llh_count) + 1);
 267         llh->llh_tail.lrt_index = cpu_to_le32(index);
 268
 269         offset = 0;
 270         rc = llog_lvfs_write_blob(ctxt, file, &llh->llh_hdr, NULL, 0);
 271         if (rc)
 272                 RETURN(rc);
 273
 274         CDEBUG(D_HA, "adding record "LPX64": idx: %u, %u bytes off: %lld\n",
 275                loghandle->lgh_id.lgl_oid, index, le32_to_cpu(rec->lrh_len),
 276                file->f_pos);
 277
 278         rc = llog_lvfs_write_blob(ctxt, file, rec, buf, file->f_pos);
 279         if (rc)
 280                 RETURN(rc);
 281
 282         if (rc == 0 && reccookie) {
 283                 if (llog_cookie_get_flags(reccookie) & LLOG_COOKIE_REPLAY) {
 284                         LASSERT(EQ_LOGID(reccookie->lgc_lgl, loghandle->lgh_id));
 285                         LASSERT(reccookie->lgc_index == index);
 286                 } else {
 287                         reccookie->lgc_lgl = loghandle->lgh_id;
 288                         reccookie->lgc_index = index;
 289                         llog_cookie_add_flags(reccookie, LLOG_COOKIE_REPLAY);
 290                 }
 291
 292                 if (le32_to_cpu(rec->lrh_type) == MDS_UNLINK_REC)
 293                         reccookie->lgc_subsys = LLOG_UNLINK_ORIG_CTXT;
 294                 else if (le32_to_cpu(rec->lrh_type) == OST_SZ_REC)
 295                         reccookie->lgc_subsys = LLOG_SIZE_ORIG_CTXT;
 296                 else if (le32_to_cpu(rec->lrh_type) == OST_RAID1_REC)
 297                         reccookie->lgc_subsys = LLOG_RD1_ORIG_CTXT;
 298                 else
 299                         reccookie->lgc_subsys = -1;
 300                 rc = 1;
 301         }
 302         if (rc == 0 && (le32_to_cpu(rec->lrh_type) == LLOG_GEN_REC ||
 303             le32_to_cpu(rec->lrh_type) == SMFS_UPDATE_REC))
 304                 rc = 1;
 305
 306         RETURN(rc);
 307 }
 308
 309 /* We can skip reading at least as many log blocks as the number of
 310 * minimum sized log records we are skipping.  If it turns out
 311 * that we are not far enough along the log (because the
 312 * actual records are larger than minimum size) we just skip
 313 * some more records. */
 314
 315 static void llog_skip_over(__u64 *off, int curr, int goal)
 316 {
 317         if (goal <= curr)
 318                 return;
 319         *off = (*off + (goal-curr-1) * LLOG_MIN_REC_SIZE) &
 320                 ~(LLOG_CHUNK_SIZE - 1);
 321 }
 322
 323 /* sets:
 324  *  - curr_offset to the furthest point read in the log file
 325  *  - curr_idx to the log index preceeding curr_offset
 326  * returns -EIO/-EINVAL on error
 327  */
 328 static int llog_lvfs_next_block(struct llog_handle *loghandle, int *curr_idx,
 329                                 int next_idx, __u64 *curr_offset, void *buf,
 330                                 int len)
 331 {
 332         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 333         ENTRY;
 334
 335         if (len == 0 || len & (LLOG_CHUNK_SIZE - 1))
 336                 RETURN(-EINVAL);
 337
 338         CDEBUG(D_OTHER, "looking for log index %u (cur idx %u off "LPU64")\n",
 339                next_idx, *curr_idx, *curr_offset);
 340
 341         while (*curr_offset < loghandle->lgh_file->f_dentry->d_inode->i_size) {
 342                 struct llog_rec_hdr *rec;
 343                 struct llog_rec_tail *tail;
 344                 loff_t ppos;
 345                 int nbytes, rc;
 346
 347                 llog_skip_over(curr_offset, *curr_idx, next_idx);
 348
 349                 ppos = *curr_offset;
 350                 rc = llog_fsfilt_read_record(ctxt, loghandle->lgh_file,
 351                                              buf, len, &ppos);
 352
 353                 if (rc) {
 354                         CERROR("Cant read llog block at log id "LPU64
 355                                "/%u offset "LPU64"\n",
 356                                loghandle->lgh_id.lgl_oid,
 357                                loghandle->lgh_id.lgl_ogen,
 358                                *curr_offset);
 359                         RETURN(rc);
 360                 }
 361
 362                 nbytes = ppos - *curr_offset;
 363                 *curr_offset = ppos;
 364
 365                 if (nbytes == 0) /* end of file, nothing to do */
 366                         RETURN(0);
 367
 368                 if (nbytes < sizeof(*tail)) {
 369                         CERROR("Invalid llog block 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
 375                 tail = buf + nbytes - sizeof(struct llog_rec_tail);
 376                 *curr_idx = le32_to_cpu(tail->lrt_index);
 377
 378                 /* this shouldn't happen */
 379                 if (tail->lrt_index == 0) {
 380                         CERROR("Invalid llog tail at log id "LPU64"/%u offset "
 381                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 382                                loghandle->lgh_id.lgl_ogen, *curr_offset);
 383                         RETURN(-EINVAL);
 384                 }
 385                 if (le32_to_cpu(tail->lrt_index) < next_idx)
 386                         continue;
 387
 388                 /* sanity check that the start of the new buffer is no farther
 389                  * than the record that we wanted.  This shouldn't happen. */
 390                 rec = buf;
 391                 if (le32_to_cpu(rec->lrh_index) > next_idx) {
 392                         CERROR("missed desired record? %u > %u\n",
 393                                le32_to_cpu(rec->lrh_index), next_idx);
 394                         RETURN(-ENOENT);
 395                 }
 396                 RETURN(0);
 397         }
 398         RETURN(-EIO);
 399 }
 400
 401 static int llog_lvfs_prev_block(struct llog_handle *loghandle,
 402                                 int prev_idx, void *buf, int len)
 403 {
 404         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 405         __u64 curr_offset;
 406         int rc;
 407         ENTRY;
 408
 409         if (len == 0 || len & (LLOG_CHUNK_SIZE - 1))
 410                 RETURN(-EINVAL);
 411
 412         CDEBUG(D_OTHER, "looking for log index %u n", prev_idx);
 413
 414         curr_offset = LLOG_CHUNK_SIZE;
 415         llog_skip_over(&curr_offset, 0, prev_idx);
 416
 417         while (curr_offset < loghandle->lgh_file->f_dentry->d_inode->i_size) {
 418                 struct llog_rec_hdr *rec;
 419                 struct llog_rec_tail *tail;
 420                 loff_t ppos;
 421
 422                 ppos = curr_offset;
 423                 rc = llog_fsfilt_read_record(ctxt, loghandle->lgh_file,
 424                                              buf, len, &ppos);
 425
 426                 if (rc) {
 427                         CERROR("Cant read llog block at log id "LPU64
 428                                "/%u offset "LPU64"\n",
 429                                loghandle->lgh_id.lgl_oid,
 430                                loghandle->lgh_id.lgl_ogen,
 431                                curr_offset);
 432                         RETURN(rc);
 433                 }
 434
 435                 /* put number of bytes read into rc to make code simpler */
 436                 rc = ppos - curr_offset;
 437                 curr_offset = ppos;
 438
 439                 if (rc == 0) /* end of file, nothing to do */
 440                         RETURN(0);
 441
 442                 if (rc < sizeof(*tail)) {
 443                         CERROR("Invalid llog block at log id "LPU64"/%u offset "
 444                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 445                                loghandle->lgh_id.lgl_ogen, curr_offset);
 446                         RETURN(-EINVAL);
 447                 }
 448
 449                 tail = buf + rc - sizeof(struct llog_rec_tail);
 450
 451                 /* this shouldn't happen */
 452                 if (tail->lrt_index == 0) {
 453                         CERROR("Invalid llog tail at log id "LPU64"/%u offset "
 454                                LPU64"\n", loghandle->lgh_id.lgl_oid,
 455                                loghandle->lgh_id.lgl_ogen, curr_offset);
 456                         RETURN(-EINVAL);
 457                 }
 458                 if (le32_to_cpu(tail->lrt_index) < prev_idx)
 459                         continue;
 460
 461                 /* sanity check that the start of the new buffer is no farther
 462                  * than the record that we wanted.  This shouldn't happen. */
 463                 rec = buf;
 464                 if (le32_to_cpu(rec->lrh_index) > prev_idx) {
 465                         CERROR("missed desired record? %u > %u\n",
 466                                le32_to_cpu(rec->lrh_index), prev_idx);
 467                         RETURN(-ENOENT);
 468                 }
 469                 RETURN(0);
 470         }
 471         RETURN(-EIO);
 472 }
 473
 474 static struct file *llog_filp_open(char *name, int flags, int mode)
 475 {
 476         char *logname;
 477         struct file *filp;
 478         int len;
 479
 480         OBD_ALLOC(logname, PATH_MAX);
 481         if (logname == NULL)
 482                 return ERR_PTR(-ENOMEM);
 483
 484         len = snprintf(logname, PATH_MAX, "LOGS/%s", name);
 485         if (len >= PATH_MAX - 1) {
 486                 filp = ERR_PTR(-ENAMETOOLONG);
 487         } else {
 488                 filp = l_filp_open(logname, flags, mode);
 489                 if (IS_ERR(filp)) {
 490                         CERROR("logfile %s(%s): %ld\n",
 491                                flags & O_CREAT ? "create" : "open", logname,
 492                                PTR_ERR(filp));
 493                 }
 494         }
 495
 496         OBD_FREE(logname, PATH_MAX);
 497         return filp;
 498 }
 499
 500 /* creates object for the case when we have no obd (smfs). */
 501 static struct file *
 502 llog_object_create_alone(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 503 {
 504         struct file *filp;
 505         int rc = 0;
 506         ENTRY;
 507
 508         LASSERT(lgh_id != NULL);
 509         if (lgh_id->lgl_oid) {
 510                 struct dentry *dchild;
 511                 char fidname[LL_FID_NAMELEN];
 512                 int fidlen = 0;
 513
 514                 down(&ctxt->loc_objects_dir->d_inode->i_sem);
 515                 fidlen = ll_fid2str(fidname, lgh_id->lgl_oid, lgh_id->lgl_ogen);
 516                 dchild = lookup_one_len(fidname, ctxt->loc_objects_dir, fidlen);
 517                 if (IS_ERR(dchild)) {
 518                         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 519                         RETURN((struct file *)dchild);
 520                 }
 521                 if (dchild->d_inode == NULL) {
 522                         struct dentry_params dp;
 523                         struct inode *inode;
 524
 525                         dchild->d_fsdata = (void *) &dp;
 526                         dp.p_ptr = NULL;
 527                         dp.p_inum = lgh_id->lgl_oid;
 528                         rc = ll_vfs_create(ctxt->loc_objects_dir->d_inode,
 529                                            dchild, S_IFREG, NULL);
 530                         if (dchild->d_fsdata == (void *)(unsigned long)lgh_id->lgl_oid)
 531                                 dchild->d_fsdata = NULL;
 532                         if (rc) {
 533                                 CDEBUG(D_INODE, "err during create: %d\n", rc);
 534                                 dput(dchild);
 535                                 up(&ctxt->loc_objects_dir->d_inode->i_sem);
 536                                 RETURN(ERR_PTR(rc));
 537                         }
 538                         inode = dchild->d_inode;
 539                         LASSERT(inode->i_ino == lgh_id->lgl_oid);
 540                         inode->i_generation = lgh_id->lgl_ogen;
 541                         CDEBUG(D_HA, "recreated ino %lu with gen %u\n",
 542                                inode->i_ino, inode->i_generation);
 543                         mark_inode_dirty(inode);
 544                 }
 545
 546                 mntget(ctxt->loc_lvfs_ctxt->pwdmnt);
 547                 filp = dentry_open(dchild, ctxt->loc_lvfs_ctxt->pwdmnt,
 548                                     O_RDWR | O_LARGEFILE);
 549                 if (IS_ERR(filp)) {
 550                         dput(dchild);
 551                         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 552                         RETURN(filp);
 553                 }
 554                 if (!S_ISREG(filp->f_dentry->d_inode->i_mode)) {
 555                         CERROR("%s is not a regular file!: mode = %o\n", fidname,
 556                                filp->f_dentry->d_inode->i_mode);
 557                         filp_close(filp, 0);
 558                         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 559                         RETURN(ERR_PTR(-ENOENT));
 560                 }
 561
 562                 up(&ctxt->loc_objects_dir->d_inode->i_sem);
 563                 RETURN(filp);
 564
 565         } else {
 566                 unsigned int tmpname = ll_insecure_random_int();
 567                 char fidname[LL_FID_NAMELEN];
 568                 struct dentry *new_child, *parent;
 569                 void *handle;
 570                 int err, namelen;
 571
 572                 sprintf(fidname, "OBJECTS/%u", tmpname);
 573                 filp = filp_open(fidname, O_CREAT | O_EXCL, 0644);
 574                 if (IS_ERR(filp)) {
 575                         rc = PTR_ERR(filp);
 576                         if (rc == -EEXIST) {
 577                                 CERROR("impossible object name collision %u\n",
 578                                         tmpname);
 579                                 LBUG();
 580                         }
 581                         CERROR("error creating tmp object %u: rc %d\n", tmpname, rc);
 582                         RETURN(filp);
 583                 }
 584
 585                 namelen = ll_fid2str(fidname, filp->f_dentry->d_inode->i_ino,
 586                                      filp->f_dentry->d_inode->i_generation);
 587                 parent = filp->f_dentry->d_parent;
 588                 down(&parent->d_inode->i_sem);
 589                 new_child = lookup_one_len(fidname, parent, namelen);
 590                 if (IS_ERR(new_child)) {
 591                         CERROR("getting neg dentry for obj rename: %d\n", rc);
 592                         GOTO(out_close, rc = PTR_ERR(new_child));
 593                 }
 594                 if (new_child->d_inode != NULL) {
 595                         CERROR("impossible non-negative obj dentry %lu:%u!\n",
 596                                 filp->f_dentry->d_inode->i_ino,
 597                                 filp->f_dentry->d_inode->i_generation);
 598                         LBUG();
 599                 }
 600
 601                 handle = llog_fsfilt_start(ctxt, parent->d_inode, FSFILT_OP_RENAME, NULL);
 602                 if (IS_ERR(handle))
 603                         GOTO(out_dput, rc = PTR_ERR(handle));
 604
 605                 lock_kernel();
 606                 rc = vfs_rename(parent->d_inode, filp->f_dentry,
 607                                 parent->d_inode, new_child);
 608                 unlock_kernel();
 609                 if (rc)
 610                         CERROR("error renaming new object %lu:%u: rc %d\n",
 611                                 filp->f_dentry->d_inode->i_ino,
 612                                 filp->f_dentry->d_inode->i_generation, rc);
 613
 614                 err = llog_fsfilt_commit(ctxt, parent->d_inode, handle, 0);
 615                 if (!rc)
 616                         rc = err;
 617
 618         out_dput:
 619                 dput(new_child);
 620         out_close:
 621                 up(&parent->d_inode->i_sem);
 622                 if (rc) {
 623                         filp_close(filp, 0);
 624                         filp = ERR_PTR(rc);
 625                 } else {
 626                         /* FIXME: is this group 1 is correct? */
 627                         lgh_id->lgl_ogr = 1;
 628                         lgh_id->lgl_oid = filp->f_dentry->d_inode->i_ino;
 629                         lgh_id->lgl_ogen = filp->f_dentry->d_inode->i_generation;
 630                 }
 631                 RETURN(filp);
 632         }
 633 }
 634
 635 /* creates object for generic case (obd exists) */
 636 static struct file *
 637 llog_object_create_generic(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 638 {
 639         struct file *filp;
 640         struct dentry *dchild;
 641         struct obd_device *obd;
 642         struct obdo *oa = NULL;
 643         int open_flags = O_RDWR | O_LARGEFILE;
 644         int rc = 0;
 645         ENTRY;
 646
 647         obd = ctxt->loc_exp->exp_obd;
 648         LASSERT(obd != NULL);
 649
 650         if (lgh_id->lgl_oid) {
 651                 dchild = obd_lvfs_fid2dentry(ctxt->loc_exp, lgh_id->lgl_oid,
 652                                              lgh_id->lgl_ogen, lgh_id->lgl_ogr);
 653                 if (IS_ERR(dchild) == -ENOENT) {
 654                         OBD_ALLOC(oa, sizeof(*oa));
 655                         if (!oa)
 656                                 RETURN(ERR_PTR(-ENOMEM));
 657
 658                         oa->o_id = lgh_id->lgl_oid;
 659                         oa->o_generation = lgh_id->lgl_ogen;
 660                         oa->o_gr = lgh_id->lgl_ogr;
 661                         oa->o_valid = OBD_MD_FLGENER | OBD_MD_FLGROUP;
 662                         rc = obd_create(ctxt->loc_exp, oa, NULL, NULL);
 663                         if (rc) {
 664                                 CDEBUG(D_INODE, "err during create: %d\n", rc);
 665                                 GOTO(out_free_oa, rc);
 666                         }
 667                         CDEBUG(D_HA, "re-create log object "LPX64":0x%x:"LPX64"\n",
 668                                lgh_id->lgl_oid, lgh_id->lgl_ogen, lgh_id->lgl_ogr);
 669
 670                         dchild = obd_lvfs_fid2dentry(ctxt->loc_exp, lgh_id->lgl_oid,
 671                                                      lgh_id->lgl_ogen, lgh_id->lgl_ogr);
 672                 } else if (IS_ERR(dchild)) {
 673                         CERROR("error looking up logfile "LPX64":0x%x: rc %d\n",
 674                                lgh_id->lgl_oid, lgh_id->lgl_ogen, rc);
 675                         RETURN((struct file *)dchild);
 676                 }
 677
 678                 filp = l_dentry_open(&obd->obd_lvfs_ctxt, dchild, open_flags);
 679                 if (IS_ERR(filp)) {
 680                         l_dput(dchild);
 681                         rc = PTR_ERR(filp);
 682                         CERROR("error opening logfile "LPX64"0x%x: rc %d\n",
 683                                lgh_id->lgl_oid, lgh_id->lgl_ogen, rc);
 684                 }
 685                 GOTO(out_free_oa, rc);
 686         } else {
 687                 /* this is important to work here over obd_create() as it manages
 688                   groups and we need it. Yet another reason is that mds_obd_create()
 689                  is fully the same as old version of this function and this helps
 690                  us to avoid code duplicating and layering violating. */
 691                 OBD_ALLOC(oa, sizeof(*oa));
 692                 if (!oa)
 693                         RETURN(ERR_PTR(-ENOMEM));
 694
 695                 oa->o_gr = FILTER_GROUP_LLOG;
 696                 oa->o_valid = OBD_MD_FLGENER | OBD_MD_FLGROUP;
 697                 rc = obd_create(ctxt->loc_exp, oa, NULL, NULL);
 698                 if (rc)
 699                         GOTO(out_free_oa, rc);
 700
 701                 dchild = obd_lvfs_fid2dentry(ctxt->loc_exp, oa->o_id,
 702                                              oa->o_generation, oa->o_gr);
 703                 if (IS_ERR(dchild))
 704                         GOTO(out_free_oa, rc = PTR_ERR(dchild));
 705
 706                 filp = l_dentry_open(&obd->obd_lvfs_ctxt, dchild,
 707                                      open_flags);
 708                 if (IS_ERR(filp)) {
 709                         l_dput(dchild);
 710                         GOTO(out_free_oa, rc = PTR_ERR(filp));
 711                 }
 712
 713                 /* group 1 is not longer valid, we use the group which is set
 714                 by obd_create()->mds_obd_create(). */
 715                 lgh_id->lgl_ogr = oa->o_gr;
 716                 lgh_id->lgl_oid = oa->o_id;
 717                 lgh_id->lgl_ogen = oa->o_generation;
 718         }
 719
 720 out_free_oa:
 721         if (rc)
 722                 filp = ERR_PTR(rc);
 723         if (oa)
 724                 OBD_FREE(oa, sizeof(*oa));
 725         RETURN(filp);
 726 }
 727
 728 static struct file *
 729 llog_object_create(struct llog_ctxt *ctxt, struct llog_logid *lgh_id)
 730 {
 731         if (ctxt->loc_alone)
 732                 return llog_object_create_alone(ctxt, lgh_id);
 733         else
 734                 return llog_object_create_generic(ctxt, lgh_id);
 735 }
 736
 737 static int llog_add_link_object(struct llog_ctxt *ctxt, struct llog_logid logid,
 738                                 struct dentry *dentry)
 739 {
 740         struct dentry *new_child;
 741         char fidname[LL_FID_NAMELEN];
 742         void *handle;
 743         int namelen, rc = 0, err;
 744         ENTRY;
 745
 746         namelen = ll_fid2str(fidname, logid.lgl_oid, logid.lgl_ogen);
 747         down(&ctxt->loc_objects_dir->d_inode->i_sem);
 748         new_child = lookup_one_len(fidname, ctxt->loc_objects_dir, namelen);
 749         if (IS_ERR(new_child)) {
 750                 CERROR("getting neg dentry for obj rename: %d\n", rc);
 751                 GOTO(out, rc = PTR_ERR(new_child));
 752         }
 753         if (new_child->d_inode == dentry->d_inode)
 754                 GOTO(out_dput, rc);
 755         if (new_child->d_inode != NULL) {
 756                 CERROR("impossible non-negative obj dentry "LPX64":%u!\n",
 757                        logid.lgl_oid, logid.lgl_ogen);
 758                 LBUG();
 759         }
 760         handle = llog_fsfilt_start(ctxt, ctxt->loc_objects_dir->d_inode,
 761                                    FSFILT_OP_LINK, NULL);
 762         if (IS_ERR(handle))
 763                 GOTO(out_dput, rc = PTR_ERR(handle));
 764
 765         lock_kernel();
 766         rc = vfs_link(dentry, ctxt->loc_objects_dir->d_inode, new_child);
 767         unlock_kernel();
 768         if (rc)
 769                 CERROR("error link new object "LPX64":%u: rc %d\n",
 770                        logid.lgl_oid, logid.lgl_ogen, rc);
 771         err = llog_fsfilt_commit(ctxt, ctxt->loc_objects_dir->d_inode, handle, 0);
 772 out_dput:
 773         l_dput(new_child);
 774 out:
 775         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 776         RETURN(rc);
 777 }
 778
 779 static int llog_lvfs_open(struct llog_ctxt *ctxt, struct llog_handle **res,
 780                           struct llog_logid *logid, char *name, int flags)
 781 {
 782         struct llog_handle *handle;
 783         struct lvfs_run_ctxt saved;
 784         int rc = 0;
 785         int open_flags = O_RDWR | O_LARGEFILE;
 786         ENTRY;
 787
 788         if (flags & OBD_LLOG_FL_CREATE)
 789                 open_flags |= O_CREAT;
 790
 791         handle = llog_alloc_handle();
 792         if (handle == NULL)
 793                 RETURN(-ENOMEM);
 794         *res = handle;
 795
 796         LASSERT(ctxt);
 797         if (ctxt->loc_lvfs_ctxt)
 798                 push_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 799
 800         if (logid != NULL) {
 801                 handle->lgh_file = llog_object_create(ctxt, logid);
 802                 if (IS_ERR(handle->lgh_file)) {
 803                         CERROR("cannot create/open llog object "LPX64":%x "
 804                                "error = %ld", logid->lgl_oid, logid->lgl_ogen,
 805                                PTR_ERR(handle->lgh_file));
 806                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 807                 }
 808                 handle->lgh_id = *logid;
 809
 810         } else if (name) {
 811                 handle->lgh_file = llog_filp_open(name, open_flags, 0644);
 812                 if (IS_ERR(handle->lgh_file)) {
 813                         CERROR("cannot open %s file, error = %ld\n",
 814                                name, PTR_ERR(handle->lgh_file));
 815                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 816                 }
 817                 LASSERT(handle->lgh_file->f_dentry->d_parent == ctxt->loc_logs_dir);
 818
 819                 handle->lgh_id.lgl_ogr = 1;
 820                 handle->lgh_id.lgl_oid = handle->lgh_file->f_dentry->d_inode->i_ino;
 821                 handle->lgh_id.lgl_ogen = handle->lgh_file->f_dentry->d_inode->i_generation;
 822                 rc = llog_add_link_object(ctxt, handle->lgh_id, handle->lgh_file->f_dentry);
 823                 if (rc)
 824                         GOTO(cleanup, rc);
 825
 826         } else {
 827                 handle->lgh_file = llog_object_create(ctxt, &handle->lgh_id);
 828                 if (IS_ERR(handle->lgh_file)) {
 829                         CERROR("cannot create llog object, error = %ld\n",
 830                                PTR_ERR(handle->lgh_file));
 831                         GOTO(cleanup, rc = PTR_ERR(handle->lgh_file));
 832                 }
 833         }
 834
 835         handle->lgh_ctxt = ctxt;
 836 finish:
 837         if (ctxt->loc_lvfs_ctxt)
 838                 pop_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 839         RETURN(rc);
 840 cleanup:
 841         llog_free_handle(handle);
 842         goto finish;
 843 }
 844
 845 static int llog_lvfs_close(struct llog_handle *handle)
 846 {
 847         int rc;
 848         ENTRY;
 849
 850         rc = filp_close(handle->lgh_file, 0);
 851         if (rc)
 852                 CERROR("error closing log: rc %d\n", rc);
 853         RETURN(rc);
 854 }
 855
 856 static int llog_lvfs_destroy(struct llog_handle *loghandle)
 857 {
 858         struct llog_ctxt *ctxt = loghandle->lgh_ctxt;
 859         struct lvfs_run_ctxt saved;
 860         struct dentry *fdentry;
 861         struct inode *parent_inode;
 862         char fidname[LL_FID_NAMELEN];
 863         void *handle;
 864         int rc = -EINVAL, err, namelen;
 865         ENTRY;
 866
 867         if (ctxt->loc_lvfs_ctxt)
 868                 push_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 869
 870         fdentry = loghandle->lgh_file->f_dentry;
 871         parent_inode = fdentry->d_parent->d_inode;
 872
 873         if (!strcmp(fdentry->d_parent->d_name.name, "LOGS")) {
 874                 LASSERT(parent_inode == ctxt->loc_logs_dir->d_inode);
 875
 876                 namelen = ll_fid2str(fidname, fdentry->d_inode->i_ino,
 877                                      fdentry->d_inode->i_generation);
 878                 dget(fdentry);
 879                 rc = llog_lvfs_close(loghandle);
 880                 if (rc) {
 881                         dput(fdentry);
 882                         GOTO(out, rc);
 883                 }
 884
 885                 handle = llog_fsfilt_start(ctxt, parent_inode,
 886                                            FSFILT_OP_UNLINK, NULL);
 887                 if (IS_ERR(handle)) {
 888                         dput(fdentry);
 889                         GOTO(out, rc = PTR_ERR(handle));
 890                 }
 891
 892                 down(&parent_inode->i_sem);
 893                 rc = vfs_unlink(parent_inode, fdentry);
 894                 up(&parent_inode->i_sem);
 895                 dput(fdentry);
 896
 897                 if (!rc) {
 898                         down(&ctxt->loc_objects_dir->d_inode->i_sem);
 899                         fdentry = lookup_one_len(fidname, ctxt->loc_objects_dir,
 900                                                  namelen);
 901                         if (fdentry == NULL || fdentry->d_inode == NULL) {
 902                                 CERROR("destroy non_existent object %s\n", fidname);
 903                                 GOTO(out_err, rc = IS_ERR(fdentry) ?
 904                                      PTR_ERR(fdentry) : -ENOENT);
 905                         }
 906                         rc = vfs_unlink(ctxt->loc_objects_dir->d_inode, fdentry);
 907                         l_dput(fdentry);
 908 out_err:
 909                         up(&ctxt->loc_objects_dir->d_inode->i_sem);
 910                 }
 911                 err = llog_fsfilt_commit(ctxt, parent_inode, handle, 0);
 912                 if (err && !rc)
 913                         err = rc;
 914
 915                 GOTO(out, rc);
 916         }
 917         if (ctxt->loc_alone) {
 918                 if (!strcmp(fdentry->d_parent->d_name.name, "OBJECTS")) {
 919                         LASSERT(parent_inode == ctxt->loc_objects_dir->d_inode);
 920
 921                         dget(fdentry);
 922                         rc = llog_lvfs_close(loghandle);
 923                         if (rc == 0) {
 924                                 down(&parent_inode->i_sem);
 925                                 rc = vfs_unlink(parent_inode, fdentry);
 926                                 up(&parent_inode->i_sem);
 927                         }
 928                         dput(fdentry);
 929                 }
 930         } else {
 931                 struct obdo *oa = NULL;
 932
 933                 OBD_ALLOC(oa, sizeof(*oa));
 934                 if (!oa)
 935                         GOTO(out, rc = -ENOMEM);
 936
 937                 oa->o_id = loghandle->lgh_id.lgl_oid;
 938                 oa->o_gr = loghandle->lgh_id.lgl_ogr;
 939                 oa->o_generation = loghandle->lgh_id.lgl_ogen;
 940                 oa->o_valid = OBD_MD_FLID | OBD_MD_FLGROUP | OBD_MD_FLGENER;
 941
 942                 rc = llog_lvfs_close(loghandle);
 943                 if (rc)
 944                         GOTO(out_free_oa, rc);
 945
 946                 rc = obd_destroy(loghandle->lgh_ctxt->loc_exp, oa, NULL, NULL);
 947 out_free_oa:
 948                 OBD_FREE(oa, sizeof(*oa));
 949         }
 950 out:
 951         if (ctxt->loc_lvfs_ctxt)
 952                 pop_ctxt(&saved, ctxt->loc_lvfs_ctxt, NULL);
 953         RETURN(rc);
 954 }
 955
 956 /* reads the catalog list */
 957 int llog_get_cat_list(struct lvfs_run_ctxt *ctxt,
 958                       struct fsfilt_operations *fsops, char *name,
 959                       int count, struct llog_catid *idarray)
 960 {
 961         struct lvfs_run_ctxt saved;
 962         struct l_file *file;
 963         int size = sizeof(*idarray) * count;
 964         loff_t off = 0;
 965         int rc;
 966
 967         LASSERT(count);
 968
 969         if (ctxt)
 970                 push_ctxt(&saved, ctxt, NULL);
 971         file = l_filp_open(name, O_RDWR | O_CREAT | O_LARGEFILE, 0700);
 972         if (!file || IS_ERR(file)) {
 973                 rc = PTR_ERR(file);
 974                 CERROR("OBD filter: cannot open/create %s: rc = %d\n",
 975                        name, rc);
 976                 GOTO(out, rc);
 977         }
 978
 979         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 980                 CERROR("%s is not a regular file!: mode = %o\n", name,
 981                        file->f_dentry->d_inode->i_mode);
 982                 GOTO(out, rc = -ENOENT);
 983         }
 984
 985         rc = fsops->fs_read_record(file, idarray, size, &off);
 986         if (rc) {
 987                 CDEBUG(D_INODE,"OBD filter: error reading %s: rc %d\n",
 988                        name, rc);
 989                 GOTO(out, rc);
 990         }
 991
 992  out:
 993         if (file && !IS_ERR(file))
 994                 rc = filp_close(file, 0);
 995         if (ctxt)
 996                 pop_ctxt(&saved, ctxt, NULL);
 997         RETURN(rc);
 998 }
 999 EXPORT_SYMBOL(llog_get_cat_list);
1000
1001 /* writes the cat list */
1002 int llog_put_cat_list(struct lvfs_run_ctxt *ctxt,
1003                       struct fsfilt_operations *fsops, char *name,
1004                       int count, struct llog_catid *idarray)
1005 {
1006         struct lvfs_run_ctxt saved;
1007         struct l_file *file;
1008         int size = sizeof(*idarray) * count;
1009         loff_t off = 0;
1010         int rc;
1011
1012         LASSERT(count);
1013
1014         if (ctxt)
1015                 push_ctxt(&saved, ctxt, NULL);
1016         file = filp_open(name, O_RDWR | O_CREAT | O_LARGEFILE, 0700);
1017         if (!file || IS_ERR(file)) {
1018                 rc = PTR_ERR(file);
1019                 CERROR("OBD filter: cannot open/create %s: rc = %d\n",
1020                        name, rc);
1021                 GOTO(out, rc);
1022         }
1023
1024         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
1025                 CERROR("%s is not a regular file!: mode = %o\n", name,
1026                        file->f_dentry->d_inode->i_mode);
1027                 GOTO(out, rc = -ENOENT);
1028         }
1029
1030         rc = fsops->fs_write_record(file, idarray, size, &off, 1);
1031         if (rc) {
1032                 CDEBUG(D_INODE,"OBD filter: error reading %s: rc %d\n",
1033                        name, rc);
1034                 GOTO(out, rc);
1035         }
1036
1037  out:
1038         if (file && !IS_ERR(file))
1039                 rc = filp_close(file, 0);
1040         if (ctxt)
1041                 pop_ctxt(&saved, ctxt, NULL);
1042         RETURN(rc);
1043 }
1044 EXPORT_SYMBOL(llog_put_cat_list);
1045
1046 struct llog_operations llog_lvfs_ops = {
1047         lop_open:        llog_lvfs_open,
1048         lop_destroy:     llog_lvfs_destroy,
1049         lop_close:       llog_lvfs_close,
1050         lop_read_header: llog_lvfs_read_header,
1051         lop_write_rec:   llog_lvfs_write_rec,
1052         lop_next_block:  llog_lvfs_next_block,
1053         lop_prev_block:  llog_lvfs_prev_block,
1054 };
1055 EXPORT_SYMBOL(llog_lvfs_ops);
1056
1057 #else /* !__KERNEL__ */
1058
1059 static int llog_lvfs_read_header(struct llog_handle *handle)
1060 {
1061         LBUG();
1062         return 0;
1063 }
1064
1065 static int llog_lvfs_write_rec(struct llog_handle *loghandle,
1066                                struct llog_rec_hdr *rec,
1067                                struct llog_cookie *reccookie, int cookiecount,
1068                                void *buf, int idx)
1069 {
1070         LBUG();
1071         return 0;
1072 }
1073
1074 static int llog_lvfs_open(struct llog_ctxt *ctxt, struct llog_handle **res,
1075                           struct llog_logid *logid, char *name, int flags)
1076 {
1077         LBUG();
1078         return 0;
1079 }
1080
1081 static int llog_lvfs_close(struct llog_handle *handle)
1082 {
1083         LBUG();
1084         return 0;
1085 }
1086
1087 static int llog_lvfs_destroy(struct llog_handle *handle)
1088 {
1089         LBUG();
1090         return 0;
1091 }
1092
1093 int llog_get_cat_list(struct lvfs_run_ctxt *ctxt,
1094                       struct fsfilt_operations *fsops, char *name,
1095                       int count, struct llog_catid *idarray)
1096 {
1097         LBUG();
1098         return 0;
1099 }
1100
1101 int llog_put_cat_list(struct lvfs_run_ctxt *ctxt,
1102                       struct fsfilt_operations *fsops, char *name,
1103                       int count, struct llog_catid *idarray)
1104 {
1105         LBUG();
1106         return 0;
1107 }
1108
1109 int llog_lvfs_prev_block(struct llog_handle *loghandle,
1110                          int prev_idx, void *buf, int len)
1111 {
1112         LBUG();
1113         return 0;
1114 }
1115
1116 int llog_lvfs_next_block(struct llog_handle *loghandle, int *curr_idx,
1117                          int next_idx, __u64 *offset, void *buf, int len)
1118 {
1119         LBUG();
1120         return 0;
1121 }
1122
1123 struct llog_operations llog_lvfs_ops = {
1124         lop_open:        llog_lvfs_open,
1125         lop_destroy:     llog_lvfs_destroy,
1126         lop_close:       llog_lvfs_close,
1127         lop_read_header: llog_lvfs_read_header,
1128         lop_write_rec:   llog_lvfs_write_rec,
1129         lop_next_block:  llog_lvfs_next_block,
1130         lop_prev_block:  llog_lvfs_prev_block,
1131 };
1132 #endif