lustre/mds/mds_fs.c

   1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
   2  * vim:expandtab:shiftwidth=8:tabstop=8:
   3  *
   4  *  mds/mds_fs.c
   5  *  Lustre Metadata Server (MDS) filesystem interface code
   6  *
   7  *  Copyright (C) 2002, 2003 Cluster File Systems, Inc.
   8  *   Author: Andreas Dilger <adilger@clusterfs.com>
   9  *
  10  *   This file is part of Lustre, http://www.lustre.org.
  11  *
  12  *   Lustre is free software; you can redistribute it and/or
  13  *   modify it under the terms of version 2 of the GNU General Public
  14  *   License as published by the Free Software Foundation.
  15  *
  16  *   Lustre is distributed in the hope that it will be useful,
  17  *   but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19  *   GNU General Public License for more details.
  20  *
  21  *   You should have received a copy of the GNU General Public License
  22  *   along with Lustre; if not, write to the Free Software
  23  *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  24  */
  25
  26 #ifndef EXPORT_SYMTAB
  27 # define EXPORT_SYMTAB
  28 #endif
  29 #define DEBUG_SUBSYSTEM S_MDS
  30
  31 #include <linux/module.h>
  32 #include <linux/kmod.h>
  33 #include <linux/version.h>
  34 #if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,0))
  35 #include <linux/mount.h>
  36 #endif
  37 #include <linux/lustre_mds.h>
  38 #include <linux/obd_class.h>
  39 #include <linux/obd_support.h>
  40 #include <linux/lustre_lib.h>
  41 #include <linux/lustre_fsfilt.h>
  42 #include <libcfs/list.h>
  43
  44 #include <linux/lustre_smfs.h>
  45 #include "mds_internal.h"
  46
  47 /* This limit is arbitrary, but for now we fit it in 1 page (32k clients) */
  48 #define MDS_MAX_CLIENTS (PAGE_SIZE * 8)
  49
  50 #define LAST_RCVD "last_rcvd"
  51 #define LOV_OBJID "lov_objid"
  52 #define LAST_FID  "last_fid"
  53 #define VIRT_FID  "virt_fid"
  54
  55 struct fidmap_entry {
  56         struct hlist_node fm_hash;
  57         struct lustre_id  fm_id;
  58 };
  59
  60 int mds_fidmap_init(struct obd_device *obd, int size)
  61 {
  62         struct mds_obd *mds = &obd->u.mds;
  63         struct hlist_head *head;
  64         int i = 0;
  65         ENTRY;
  66
  67         LASSERT(size > sizeof(sizeof(struct hlist_head)));
  68         mds->mds_fidmap_size = size / sizeof(struct hlist_head);
  69
  70         CWARN("allocating %lu fid mapping entries\n",
  71               (unsigned long)mds->mds_fidmap_size);
  72
  73         OBD_ALLOC(mds->mds_fidmap_table, size);
  74         if (!mds->mds_fidmap_table)
  75                 RETURN(-ENOMEM);
  76
  77         i = mds->mds_fidmap_size;
  78         head = mds->mds_fidmap_table;
  79         do {
  80                 INIT_HLIST_HEAD(head);
  81                 head++;
  82                 i--;
  83         } while(i);
  84
  85         RETURN(0);
  86 }
  87
  88 int mds_fidmap_cleanup(struct obd_device *obd)
  89 {
  90         struct hlist_node *node = NULL, *tmp = NULL;
  91         struct mds_obd *mds = &obd->u.mds;
  92         struct fidmap_entry *entry;
  93         struct hlist_head *head;
  94         int i = 0;
  95         ENTRY;
  96
  97         spin_lock(&mds->mds_fidmap_lock);
  98         for (i = 0, head = mds->mds_fidmap_table;
  99              i < mds->mds_fidmap_size; i++, head++) {
 100                 hlist_for_each_safe(node, tmp, head) {
 101                         entry = hlist_entry(node, struct fidmap_entry, fm_hash);
 102                         hlist_del_init(&entry->fm_hash);
 103                         OBD_FREE(entry, sizeof(*entry));
 104                 }
 105         }
 106         spin_unlock(&mds->mds_fidmap_lock);
 107         OBD_FREE(mds->mds_fidmap_table, mds->mds_fidmap_size *
 108                  sizeof(struct hlist_head));
 109         RETURN(0);
 110 }
 111
 112 static inline unsigned long
 113 const hashfn(struct obd_device *obd, __u64 fid)
 114 {
 115         struct mds_obd *mds = &obd->u.mds;
 116         return (unsigned long)(fid & (mds->mds_fidmap_size - 1));
 117 }
 118
 119 static struct fidmap_entry *
 120 __mds_fidmap_find(struct obd_device *obd, __u64 fid)
 121 {
 122         struct fidmap_entry *entry = NULL;
 123         struct mds_obd *mds = &obd->u.mds;
 124         struct hlist_node *node = NULL;
 125         struct hlist_head *head;
 126         ENTRY;
 127
 128         head = mds->mds_fidmap_table + hashfn(obd, fid);
 129         hlist_for_each(node, head) {
 130                 entry = hlist_entry(node, struct fidmap_entry, fm_hash);
 131                 if (id_fid(&entry->fm_id) == fid)
 132                         RETURN(entry);
 133         }
 134         RETURN(NULL);
 135 }
 136
 137 struct fidmap_entry *
 138 mds_fidmap_find(struct obd_device *obd, __u64 fid)
 139 {
 140         struct mds_obd *mds = &obd->u.mds;
 141         struct fidmap_entry *entry;
 142         ENTRY;
 143
 144         spin_lock(&mds->mds_fidmap_lock);
 145         entry = __mds_fidmap_find(obd, fid);
 146         spin_unlock(&mds->mds_fidmap_lock);
 147
 148         RETURN(entry);
 149 }
 150
 151 static void __mds_fidmap_insert(struct obd_device *obd,
 152                                 struct fidmap_entry *entry)
 153 {
 154         struct mds_obd *mds = &obd->u.mds;
 155         struct hlist_head *head;
 156         unsigned long idx;
 157         ENTRY;
 158
 159         idx = hashfn(obd, id_fid(&entry->fm_id));
 160         head = mds->mds_fidmap_table + idx;
 161         hlist_add_head(&entry->fm_hash, head);
 162
 163         EXIT;
 164 }
 165
 166 void mds_fidmap_insert(struct obd_device *obd,
 167                        struct fidmap_entry *entry)
 168 {
 169         struct mds_obd *mds = &obd->u.mds;
 170         ENTRY;
 171
 172         spin_lock(&mds->mds_fidmap_lock);
 173         __mds_fidmap_insert(obd, entry);
 174         spin_unlock(&mds->mds_fidmap_lock);
 175
 176         EXIT;
 177 }
 178
 179 static void __mds_fidmap_remove(struct obd_device *obd,
 180                                 struct fidmap_entry *entry)
 181 {
 182         ENTRY;
 183         hlist_del_init(&entry->fm_hash);
 184         EXIT;
 185 }
 186
 187 void mds_fidmap_remove(struct obd_device *obd,
 188                        struct fidmap_entry *entry)
 189 {
 190         struct mds_obd *mds = &obd->u.mds;
 191         ENTRY;
 192
 193         spin_lock(&mds->mds_fidmap_lock);
 194         __mds_fidmap_remove(obd, entry);
 195         spin_unlock(&mds->mds_fidmap_lock);
 196
 197         EXIT;
 198 }
 199
 200 /* creates new mapping remote fid -> local inode store cookie. Both are saved in
 201  * lustre_id for better usability, as all mds function use lustre_id as input
 202  * params.*/
 203 int mds_fidmap_add(struct obd_device *obd,
 204                    struct lustre_id *id)
 205 {
 206         struct mds_obd *mds = &obd->u.mds;
 207         struct fidmap_entry *entry;
 208         ENTRY;
 209
 210         OBD_ALLOC(entry, sizeof(*entry));
 211         if (!entry)
 212                 RETURN(-ENOMEM);
 213
 214         entry->fm_id = *id;
 215
 216         spin_lock(&mds->mds_fidmap_lock);
 217         if (!__mds_fidmap_find(obd, id_fid(id))) {
 218                 __mds_fidmap_insert(obd, entry);
 219                 spin_unlock(&mds->mds_fidmap_lock);
 220                 CDEBUG(D_INODE, "added mapping to "DLID4"\n",
 221                        OLID4(id));
 222                 RETURN(1);
 223         }
 224         spin_unlock(&mds->mds_fidmap_lock);
 225         OBD_FREE(entry, sizeof(*entry));
 226
 227         RETURN(0);
 228 }
 229
 230 /* removes mapping using fid component from passed @id */
 231 void mds_fidmap_del(struct obd_device *obd,
 232                     struct lustre_id *id)
 233 {
 234         struct mds_obd *mds = &obd->u.mds;
 235         struct fidmap_entry *entry;
 236         ENTRY;
 237
 238         spin_lock(&mds->mds_fidmap_lock);
 239         entry = __mds_fidmap_find(obd, id_fid(id));
 240         if (entry) {
 241                 __mds_fidmap_remove(obd, entry);
 242                 spin_unlock(&mds->mds_fidmap_lock);
 243                 OBD_FREE(entry, sizeof(*entry));
 244                 CDEBUG(D_INODE, "removed mapping to "DLID4"\n",
 245                        OLID4(id));
 246                 goto out;
 247         }
 248         spin_unlock(&mds->mds_fidmap_lock);
 249 out:
 250         EXIT;
 251 }
 252
 253 struct lustre_id *mds_fidmap_lookup(struct obd_device *obd,
 254                                     struct lustre_id *id)
 255 {
 256         struct mds_obd *mds = &obd->u.mds;
 257         struct fidmap_entry *entry;
 258         ENTRY;
 259
 260         spin_lock(&mds->mds_fidmap_lock);
 261         entry = __mds_fidmap_find(obd, id_fid(id));
 262         spin_unlock(&mds->mds_fidmap_lock);
 263
 264         if (!entry)
 265                 RETURN(NULL);
 266
 267         RETURN(&entry->fm_id);
 268 }
 269
 270 /* Add client data to the MDS.  We use a bitmap to locate a free space
 271  * in the last_rcvd file if cl_off is -1 (i.e. a new client).
 272  * Otherwise, we have just read the data from the last_rcvd file and
 273  * we know its offset.
 274  */
 275 int mds_client_add(struct obd_device *obd, struct mds_obd *mds,
 276                    struct mds_export_data *med, int cl_idx)
 277 {
 278         unsigned long *bitmap = mds->mds_client_bitmap;
 279         int new_client = (cl_idx == -1);
 280         ENTRY;
 281
 282         LASSERT(bitmap != NULL);
 283
 284         /* XXX if mcd_uuid were a real obd_uuid, I could use obd_uuid_equals */
 285         if (!strcmp((char *)med->med_mcd->mcd_uuid, (char *)obd->obd_uuid.uuid))
 286                 RETURN(0);
 287
 288         /* the bitmap operations can handle cl_idx > sizeof(long) * 8, so
 289          * there's no need for extra complication here
 290          */
 291         if (new_client) {
 292                 cl_idx = find_first_zero_bit(bitmap, MDS_MAX_CLIENTS);
 293         repeat:
 294                 if (cl_idx >= MDS_MAX_CLIENTS) {
 295                         CERROR("no room for clients - fix MDS_MAX_CLIENTS\n");
 296                         return -ENOMEM;
 297                 }
 298                 if (test_and_set_bit(cl_idx, bitmap)) {
 299                         cl_idx = find_next_zero_bit(bitmap, MDS_MAX_CLIENTS,
 300                                                     cl_idx);
 301                         goto repeat;
 302                 }
 303         } else {
 304                 if (test_and_set_bit(cl_idx, bitmap)) {
 305                         CERROR("MDS client %d: bit already set in bitmap!!\n",
 306                                cl_idx);
 307                         LBUG();
 308                 }
 309         }
 310
 311         CDEBUG(D_INFO, "client at idx %d with UUID '%s' added\n",
 312                cl_idx, med->med_mcd->mcd_uuid);
 313
 314         med->med_idx = cl_idx;
 315         med->med_off = le32_to_cpu(mds->mds_server_data->msd_client_start) +
 316                 (cl_idx * le16_to_cpu(mds->mds_server_data->msd_client_size));
 317
 318         if (new_client) {
 319                 struct file *file = mds->mds_rcvd_filp;
 320                 struct lvfs_run_ctxt saved;
 321                 loff_t off = med->med_off;
 322                 int rc;
 323
 324                 push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 325                 rc = fsfilt_write_record(obd, file, med->med_mcd,
 326                                          sizeof(*med->med_mcd), &off, 1);
 327                 pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 328
 329                 if (rc)
 330                         return rc;
 331                 CDEBUG(D_INFO, "wrote client mcd at idx %u off %llu (len %u)\n",
 332                        med->med_idx, med->med_off,
 333                        (unsigned int)sizeof(*med->med_mcd));
 334         }
 335         return 0;
 336 }
 337
 338 int mds_client_free(struct obd_export *exp, int clear_client)
 339 {
 340         struct mds_export_data *med = &exp->exp_mds_data;
 341         struct mds_obd *mds = &exp->exp_obd->u.mds;
 342         unsigned long *bitmap = mds->mds_client_bitmap;
 343         struct obd_device *obd = exp->exp_obd;
 344         struct mds_client_data zero_mcd;
 345         struct lvfs_run_ctxt saved;
 346         int rc;
 347
 348         if (!med->med_mcd)
 349                 RETURN(0);
 350
 351         /* XXX if mcd_uuid were a real obd_uuid, I could use obd_uuid_equals */
 352         if (!strcmp((char *)med->med_mcd->mcd_uuid, (char *)obd->obd_uuid.uuid))
 353                 GOTO(free_and_out, 0);
 354
 355         CDEBUG(D_INFO, "freeing client at idx %u (%lld)with UUID '%s'\n",
 356                med->med_idx, med->med_off, med->med_mcd->mcd_uuid);
 357
 358         LASSERT(bitmap);
 359
 360         /* Clear the bit _after_ zeroing out the client so we don't
 361            race with mds_client_add and zero out new clients.*/
 362         if (!test_bit(med->med_idx, bitmap)) {
 363                 CERROR("MDS client %u: bit already clear in bitmap!!\n",
 364                        med->med_idx);
 365                 LBUG();
 366         }
 367
 368         if (clear_client) {
 369                 memset(&zero_mcd, 0, sizeof zero_mcd);
 370                 push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 371                 rc = fsfilt_write_record(obd, mds->mds_rcvd_filp, &zero_mcd,
 372                                          sizeof(zero_mcd), &med->med_off, 1);
 373                 pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 374
 375                 CDEBUG(rc == 0 ? D_INFO : D_ERROR,
 376                        "zeroing out client %s idx %u in %s rc %d\n",
 377                        med->med_mcd->mcd_uuid, med->med_idx, LAST_RCVD, rc);
 378         }
 379
 380         if (!test_and_clear_bit(med->med_idx, bitmap)) {
 381                 CERROR("MDS client %u: bit already clear in bitmap!!\n",
 382                        med->med_idx);
 383                 LBUG();
 384         }
 385
 386
 387         /* Make sure the server's last_transno is up to date. Do this
 388          * after the client is freed so we know all the client's
 389          * transactions have been committed. */
 390         mds_update_server_data(exp->exp_obd, 1);
 391
 392 free_and_out:
 393         OBD_FREE(med->med_mcd, sizeof(*med->med_mcd));
 394         med->med_mcd = NULL;
 395         return 0;
 396 }
 397
 398 static int mds_server_free_data(struct mds_obd *mds)
 399 {
 400         OBD_FREE(mds->mds_client_bitmap, MDS_MAX_CLIENTS / 8);
 401         OBD_FREE(mds->mds_server_data, sizeof(*mds->mds_server_data));
 402         mds->mds_server_data = NULL;
 403
 404         return 0;
 405 }
 406
 407 static int mds_read_last_fid(struct obd_device *obd, struct file *file)
 408 {
 409         int rc = 0;
 410         loff_t off = 0;
 411         struct mds_obd *mds = &obd->u.mds;
 412         unsigned long last_fid_size = file->f_dentry->d_inode->i_size;
 413         ENTRY;
 414
 415         if (last_fid_size == 0) {
 416                 CWARN("%s: initializing new %s\n", obd->obd_name,
 417                       file->f_dentry->d_name.name);
 418
 419                 /*
 420                  * as fid is used for forming res_id for locking, it should not
 421                  * be zero. This will keep us out of lots possible problems,
 422                  * asserts, etc.
 423                  */
 424                 mds_set_last_fid(obd, 0);
 425         } else {
 426                 __u64 lastfid;
 427
 428                 rc = fsfilt_read_record(obd, file, &lastfid,
 429                                         sizeof(lastfid), &off);
 430                 if (rc) {
 431                         CERROR("error reading MDS %s: rc = %d\n",
 432                                file->f_dentry->d_name.name, rc);
 433                         RETURN(rc);
 434                 }
 435
 436                 /*
 437                  * make sure, that fid is up-to-date.
 438                  */
 439                 mds_set_last_fid(obd, lastfid);
 440         }
 441
 442         CDEBUG(D_INODE, "%s: server last_fid: "LPU64"\n",
 443                obd->obd_name, mds->mds_last_fid);
 444
 445         rc = mds_update_last_fid(obd, NULL, 1);
 446         RETURN(rc);
 447 }
 448
 449 static int mds_read_last_rcvd(struct obd_device *obd, struct file *file)
 450 {
 451         unsigned long last_rcvd_size = file->f_dentry->d_inode->i_size;
 452         struct mds_obd *mds = &obd->u.mds;
 453         struct mds_server_data *msd = NULL;
 454         struct mds_client_data *mcd = NULL;
 455         loff_t off = 0;
 456         __u64 mount_count;
 457         int cl_idx, rc = 0;
 458         ENTRY;
 459
 460         /* ensure padding in the struct is the correct size */
 461         LASSERT(offsetof(struct mds_server_data, msd_padding) +
 462                 sizeof(msd->msd_padding) == MDS_LR_SERVER_SIZE);
 463         LASSERT(offsetof(struct mds_client_data, mcd_padding) +
 464                 sizeof(mcd->mcd_padding) == MDS_LR_CLIENT_SIZE);
 465
 466         OBD_ALLOC_WAIT(msd, sizeof(*msd));
 467         if (!msd)
 468                 RETURN(-ENOMEM);
 469
 470         OBD_ALLOC_WAIT(mds->mds_client_bitmap, MDS_MAX_CLIENTS / 8);
 471         if (!mds->mds_client_bitmap) {
 472                 OBD_FREE(msd, sizeof(*msd));
 473                 RETURN(-ENOMEM);
 474         }
 475
 476         mds->mds_server_data = msd;
 477
 478         if (last_rcvd_size == 0) {
 479                 CWARN("%s: initializing new %s\n", obd->obd_name,
 480                       file->f_dentry->d_name.name);
 481
 482                 memcpy(msd->msd_uuid, obd->obd_uuid.uuid,sizeof(msd->msd_uuid));
 483                 msd->msd_last_transno = 0;
 484                 mount_count = msd->msd_mount_count = 0;
 485                 msd->msd_server_size = cpu_to_le32(MDS_LR_SERVER_SIZE);
 486                 msd->msd_client_start = cpu_to_le32(MDS_LR_CLIENT_START);
 487                 msd->msd_client_size = cpu_to_le16(MDS_LR_CLIENT_SIZE);
 488                 msd->msd_feature_rocompat = cpu_to_le32(MDS_ROCOMPAT_LOVOBJID);
 489         } else {
 490                 rc = fsfilt_read_record(obd, file, msd, sizeof(*msd), &off);
 491                 if (rc) {
 492                         CERROR("error reading MDS %s: rc = %d\n",
 493                                file->f_dentry->d_name.name, rc);
 494                         GOTO(err_msd, rc);
 495                 }
 496                 if (strcmp((char *)msd->msd_uuid, (char *)obd->obd_uuid.uuid)) {
 497                         CERROR("OBD UUID %s does not match last_rcvd UUID %s\n",
 498                                obd->obd_uuid.uuid, msd->msd_uuid);
 499                         GOTO(err_msd, rc = -EINVAL);
 500                 }
 501                 mount_count = le64_to_cpu(msd->msd_mount_count);
 502         }
 503         if (msd->msd_feature_incompat & ~cpu_to_le32(MDS_INCOMPAT_SUPP)) {
 504                 CERROR("unsupported incompat feature %x\n",
 505                        le32_to_cpu(msd->msd_feature_incompat) &
 506                        ~MDS_INCOMPAT_SUPP);
 507                 GOTO(err_msd, rc = -EINVAL);
 508         }
 509         /* XXX updating existing b_devel fs only, can be removed in future */
 510         msd->msd_feature_rocompat = cpu_to_le32(MDS_ROCOMPAT_LOVOBJID);
 511         if (msd->msd_feature_rocompat & ~cpu_to_le32(MDS_ROCOMPAT_SUPP)) {
 512                 CERROR("unsupported read-only feature %x\n",
 513                        le32_to_cpu(msd->msd_feature_rocompat) &
 514                        ~MDS_ROCOMPAT_SUPP);
 515                 /* Do something like remount filesystem read-only */
 516                 GOTO(err_msd, rc = -EINVAL);
 517         }
 518
 519         mds->mds_last_transno = le64_to_cpu(msd->msd_last_transno);
 520
 521         CDEBUG(D_INODE, "%s: server last_transno: "LPU64"\n",
 522                obd->obd_name, mds->mds_last_transno);
 523         CDEBUG(D_INODE, "%s: server mount_count: "LPU64"\n",
 524                obd->obd_name, mount_count + 1);
 525         CDEBUG(D_INODE, "%s: server data size: %u\n",
 526                obd->obd_name, le32_to_cpu(msd->msd_server_size));
 527         CDEBUG(D_INODE, "%s: per-client data start: %u\n",
 528                obd->obd_name, le32_to_cpu(msd->msd_client_start));
 529         CDEBUG(D_INODE, "%s: per-client data size: %u\n",
 530                obd->obd_name, le32_to_cpu(msd->msd_client_size));
 531         CDEBUG(D_INODE, "%s: last_rcvd size: %lu\n",
 532                obd->obd_name, last_rcvd_size);
 533         CDEBUG(D_INODE, "%s: last_rcvd clients: %lu\n", obd->obd_name,
 534                last_rcvd_size <= le32_to_cpu(msd->msd_client_start) ? 0 :
 535                (last_rcvd_size - le32_to_cpu(msd->msd_client_start)) /
 536                 le16_to_cpu(msd->msd_client_size));
 537
 538         /* When we do a clean MDS shutdown, we save the last_transno into
 539          * the header.  If we find clients with higher last_transno values
 540          * then those clients may need recovery done. */
 541         for (cl_idx = 0, off = le32_to_cpu(msd->msd_client_start);
 542              off < last_rcvd_size; cl_idx++) {
 543                 __u64 last_transno;
 544                 struct obd_export *exp;
 545                 struct mds_export_data *med;
 546
 547                 if (!mcd) {
 548                         OBD_ALLOC_WAIT(mcd, sizeof(*mcd));
 549                         if (!mcd)
 550                                 GOTO(err_client, rc = -ENOMEM);
 551                 }
 552
 553                 /* Don't assume off is incremented properly by
 554                  * fsfilt_read_record(), in case sizeof(*mcd)
 555                  * isn't the same as msd->msd_client_size.  */
 556                 off = le32_to_cpu(msd->msd_client_start) +
 557                         cl_idx * le16_to_cpu(msd->msd_client_size);
 558                 rc = fsfilt_read_record(obd, file, mcd, sizeof(*mcd), &off);
 559                 if (rc) {
 560                         CERROR("error reading MDS %s idx %d, off %llu: rc %d\n",
 561                                file->f_dentry->d_name.name, cl_idx, off, rc);
 562                         break; /* read error shouldn't cause startup to fail */
 563                 }
 564
 565                 if (mcd->mcd_uuid[0] == '\0') {
 566                         CDEBUG(D_INFO, "skipping zeroed client at offset %d\n",
 567                                cl_idx);
 568                         continue;
 569                 }
 570
 571                 last_transno = le64_to_cpu(mcd->mcd_last_transno) >
 572                                le64_to_cpu(mcd->mcd_last_close_transno) ?
 573                                le64_to_cpu(mcd->mcd_last_transno) :
 574                                le64_to_cpu(mcd->mcd_last_close_transno);
 575
 576                 /* These exports are cleaned up by mds_disconnect(), so they
 577                  * need to be set up like real exports as mds_connect() does.
 578                  */
 579                 CDEBUG(D_HA|D_WARNING,"RCVRNG CLIENT uuid: %s idx: %d lr: "LPU64
 580                        " srv lr: "LPU64" lx: "LPU64"\n", mcd->mcd_uuid, cl_idx,
 581                        last_transno, le64_to_cpu(msd->msd_last_transno),
 582                        mcd->mcd_last_xid);
 583
 584                 exp = class_new_export(obd);
 585                 if (exp == NULL)
 586                         GOTO(err_client, rc = -ENOMEM);
 587
 588                 memcpy(&exp->exp_client_uuid.uuid, mcd->mcd_uuid,
 589                        sizeof exp->exp_client_uuid.uuid);
 590                 med = &exp->exp_mds_data;
 591                 med->med_mcd = mcd;
 592                 mds_client_add(obd, mds, med, cl_idx);
 593                 /* create helper if export init gets more complex */
 594                 INIT_LIST_HEAD(&med->med_open_head);
 595                 spin_lock_init(&med->med_open_lock);
 596
 597                 mcd = NULL;
 598                 exp->exp_connected = 0;
 599                 exp->exp_req_replay_needed = 1;
 600                 obd->obd_recoverable_clients++;
 601                 obd->obd_max_recoverable_clients++;
 602
 603                 /* track clients to separate req replay
 604                  * from lock replay. bug 6063 */
 605                 atomic_inc(&obd->obd_req_replay_clients);
 606                 exp->exp_req_replay_needed = 1;
 607                 atomic_inc(&obd->obd_lock_replay_clients);
 608                 exp->exp_lock_replay_needed = 1;
 609
 610                 class_export_put(exp);
 611
 612                 CDEBUG(D_OTHER, "client at idx %d has last_transno = "LPU64"\n",
 613                        cl_idx, last_transno);
 614
 615                 if (last_transno > mds->mds_last_transno)
 616                        mds->mds_last_transno = last_transno;
 617         }
 618         if (mcd)
 619                 OBD_FREE(mcd, sizeof(*mcd));
 620         obd->obd_last_committed = mds->mds_last_transno;
 621         if (obd->obd_recoverable_clients) {
 622                 CWARN("RECOVERY: service %s, %d recoverable clients, "
 623                       "last_transno "LPU64"\n", obd->obd_name,
 624                       obd->obd_recoverable_clients, mds->mds_last_transno);
 625                 obd->obd_next_recovery_transno = obd->obd_last_committed + 1;
 626                 target_start_recovery_thread(obd, mds_handle);
 627                 obd->obd_recovery_start = LTIME_S(CURRENT_TIME);
 628         }
 629
 630         mds->mds_mount_count = mount_count + 1;
 631         msd->msd_mount_count = cpu_to_le64(mds->mds_mount_count);
 632
 633         /* save it, so mount count and last_transno is current */
 634         rc = mds_update_server_data(obd, 1);
 635         if (rc)
 636                 GOTO(err_client, rc);
 637
 638         RETURN(0);
 639
 640 err_client:
 641         class_disconnect_exports(obd, 0);
 642 err_msd:
 643         mds_server_free_data(mds);
 644         RETURN(rc);
 645 }
 646
 647 /*
 648  * sets up root inode lustre_id. It tries to read it first from root inode and
 649  * if it is not there, new rootid is allocated and saved there.
 650  */
 651 int mds_fs_setup_rootid(struct obd_device *obd)
 652 {
 653         int rc = 0;
 654         void *handle;
 655         struct inode *inode;
 656         struct dentry *dentry;
 657         struct mds_obd *mds = &obd->u.mds;
 658         ENTRY;
 659
 660         /* getting root directory and setup its fid. */
 661         dentry = mds_id2dentry(obd, &mds->mds_rootid, NULL);
 662         if (IS_ERR(dentry)) {
 663                 CERROR("Can't find ROOT by "DLID4", err = %d\n",
 664                        OLID4(&mds->mds_rootid), (int)PTR_ERR(dentry));
 665                 RETURN(PTR_ERR(dentry));
 666         }
 667
 668         inode = dentry->d_inode;
 669         LASSERT(dentry->d_inode);
 670
 671         rc = mds_pack_inode2id(obd, &mds->mds_rootid, inode, 1);
 672         if (rc && rc != -ENODATA)
 673                 GOTO(out_dentry, rc);
 674
 675         if (rc) {
 676                 if (rc != -ENODATA)
 677                         GOTO(out_dentry, rc);
 678         } else {
 679                 /* rootid is filled by mds_read_inode_sid(), so we do not need
 680                  * to allocate it and update. */
 681                 LASSERT(id_group(&mds->mds_rootid) == mds->mds_num);
 682                 mds_set_last_fid(obd, id_fid(&mds->mds_rootid));
 683
 684                 rc = mds_fidmap_add(obd, &mds->mds_rootid);
 685                 if (rc > 0)
 686                         rc = 0;
 687
 688                 GOTO(out_dentry, rc);
 689         }
 690
 691         /* allocating new one, as it is not found in root inode. */
 692         handle = fsfilt_start(obd, inode,
 693                               FSFILT_OP_SETATTR, NULL);
 694
 695         if (IS_ERR(handle)) {
 696                 rc = PTR_ERR(handle);
 697                 CERROR("fsfilt_start() failed, rc = %d\n", rc);
 698                 GOTO(out_dentry, rc);
 699         }
 700
 701         mds_inode2id(obd, &mds->mds_rootid, inode, mds_alloc_fid(obd));
 702         rc = mds_update_inode_ids(obd, inode, handle, &mds->mds_rootid, NULL);
 703         if (rc) {
 704                 CERROR("mds_update_inode_ids() failed, rc = %d\n", rc);
 705                 GOTO(out_dentry, rc);
 706         }
 707
 708         rc = mds_fidmap_add(obd, &mds->mds_rootid);
 709         if (rc < 0)
 710                 GOTO(out_dentry, rc);
 711         else
 712                 rc = 0;
 713
 714         rc = fsfilt_commit(obd, mds->mds_sb, inode, handle, 0);
 715         if (rc)
 716                 CERROR("fsfilt_commit() failed, rc = %d\n", rc);
 717
 718         EXIT;
 719 out_dentry:
 720         l_dput(dentry);
 721         if (rc == 0)
 722                 CWARN("%s: rootid: "DLID4"\n", obd->obd_name,
 723                       OLID4(&mds->mds_rootid));
 724         return rc;
 725 }
 726
 727 static int mds_update_virtid_fid(struct obd_device *obd,
 728                                  void *handle, int force_sync)
 729 {
 730         struct mds_obd *mds = &obd->u.mds;
 731         struct file *filp = mds->mds_virtid_filp;
 732         struct lvfs_run_ctxt saved;
 733         loff_t off = 0;
 734         int rc = 0;
 735         ENTRY;
 736
 737         push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 738         rc = fsfilt_write_record(obd, filp, &mds->mds_virtid_fid,
 739                                  sizeof(mds->mds_virtid_fid),
 740                                  &off, force_sync);
 741         if (rc) {
 742                 CERROR("error writing MDS virtid_fid #"LPU64
 743                        ", err = %d\n", mds->mds_virtid_fid, rc);
 744         }
 745
 746         CDEBUG(D_SUPER, "wrote virtid fid #"LPU64" at idx "
 747                "%llu: err = %d\n", mds->mds_virtid_fid,
 748                off, rc);
 749         pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 750
 751         RETURN(rc);
 752 }
 753
 754 static int mds_read_virtid_fid(struct obd_device *obd,
 755                                struct file *file)
 756 {
 757         int rc = 0;
 758         loff_t off = 0;
 759         struct mds_obd *mds = &obd->u.mds;
 760         unsigned long virtid_fid_size = file->f_dentry->d_inode->i_size;
 761         ENTRY;
 762
 763         if (virtid_fid_size == 0) {
 764                 mds->mds_virtid_fid = mds_alloc_fid(obd);
 765         } else {
 766                 rc = fsfilt_read_record(obd, file, &mds->mds_virtid_fid,
 767                                         sizeof(mds->mds_virtid_fid), &off);
 768                 if (rc) {
 769                         CERROR("error reading MDS %s: rc = %d\n",
 770                                file->f_dentry->d_name.name, rc);
 771                         RETURN(rc);
 772                 }
 773         }
 774         rc = mds_update_virtid_fid(obd, NULL, 1);
 775
 776         RETURN(rc);
 777 }
 778
 779 /*
 780  * initializes lustre_id for virtual id directory, it is needed sometimes, as it
 781  * is possible that it will be the parent for object an operations is going to
 782  * be performed on.
 783  */
 784 int mds_fs_setup_virtid(struct obd_device *obd)
 785 {
 786         int rc = 0;
 787         void *handle;
 788         struct lustre_id sid;
 789         struct mds_obd *mds = &obd->u.mds;
 790         struct inode *inode = mds->mds_id_dir->d_inode;
 791         ENTRY;
 792
 793         handle = fsfilt_start(obd, inode,
 794                               FSFILT_OP_SETATTR, NULL);
 795
 796         if (IS_ERR(handle)) {
 797                 rc = PTR_ERR(handle);
 798                 CERROR("fsfilt_start() failed, rc = %d\n", rc);
 799                 RETURN(rc);
 800         }
 801
 802         id_group(&sid) = mds->mds_num;
 803         id_fid(&sid) = mds->mds_virtid_fid;
 804
 805         id_ino(&sid) = inode->i_ino;
 806         id_gen(&sid) = inode->i_generation;
 807         id_type(&sid) = (S_IFMT & inode->i_mode);
 808
 809         rc = mds_update_inode_ids(obd, inode, handle, &sid, NULL);
 810
 811         if (rc) {
 812                 CERROR("mds_update_inode_ids() failed, rc = %d\n", rc);
 813                 RETURN(rc);
 814         }
 815
 816         rc = mds_fidmap_add(obd, &sid);
 817         if (rc < 0)
 818                 RETURN(rc);
 819         else
 820                 rc = 0;
 821
 822         rc = fsfilt_commit(obd, mds->mds_sb, inode, handle, 0);
 823         if (rc) {
 824                 CERROR("fsfilt_commit() failed, rc = %d\n", rc);
 825                 RETURN(rc);
 826         }
 827
 828         RETURN(rc);
 829 }
 830
 831 #define MDS_FIDMAP_SIZE (2*PAGE_SIZE)
 832
 833 int mds_fs_setup(struct obd_device *obd, struct vfsmount *mnt)
 834 {
 835         struct mds_obd *mds = &obd->u.mds;
 836         struct lvfs_run_ctxt saved;
 837         struct dentry *dentry;
 838         struct file *file;
 839         int rc;
 840         ENTRY;
 841
 842         rc = cleanup_group_info();
 843         if (rc)
 844                 RETURN(rc);
 845
 846         mds->mds_vfsmnt = mnt;
 847         mds->mds_sb = mnt->mnt_root->d_inode->i_sb;
 848
 849         fsfilt_setup(obd, mds->mds_sb);
 850
 851         OBD_SET_CTXT_MAGIC(&obd->obd_lvfs_ctxt);
 852         obd->obd_lvfs_ctxt.pwdmnt = mnt;
 853         obd->obd_lvfs_ctxt.pwd = mnt->mnt_root;
 854         obd->obd_lvfs_ctxt.fs = get_ds();
 855         obd->obd_lvfs_ctxt.cb_ops = mds_lvfs_ops;
 856
 857         /* setup the directory tree */
 858         push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 859         dentry = simple_mkdir(current->fs->pwd, "ROOT", 0755, 0);
 860         if (IS_ERR(dentry)) {
 861                 rc = PTR_ERR(dentry);
 862                 CERROR("cannot create ROOT directory: rc = %d\n", rc);
 863                 GOTO(err_pop, rc);
 864         }
 865
 866         mdc_pack_id(&mds->mds_rootid, dentry->d_inode->i_ino,
 867                     dentry->d_inode->i_generation, S_IFDIR, 0, 0);
 868
 869         dput(dentry);
 870
 871         dentry = lookup_one_len("__iopen__", current->fs->pwd,
 872                                 strlen("__iopen__"));
 873         if (IS_ERR(dentry)) {
 874                 rc = PTR_ERR(dentry);
 875                 CERROR("cannot lookup __iopen__ directory: rc = %d\n", rc);
 876                 GOTO(err_pop, rc);
 877         }
 878         mds->mds_id_de = dentry;
 879         if (!dentry->d_inode || is_bad_inode(dentry->d_inode)) {
 880                 rc = -ENOENT;
 881                 CERROR("__iopen__ directory has no inode? rc = %d\n", rc);
 882                 GOTO(err_id_de, rc);
 883         }
 884
 885         dentry = simple_mkdir(current->fs->pwd, "PENDING", 0777, 1);
 886         if (IS_ERR(dentry)) {
 887                 rc = PTR_ERR(dentry);
 888                 CERROR("cannot create PENDING directory: rc = %d\n", rc);
 889                 GOTO(err_id_de, rc);
 890         }
 891         mds->mds_pending_dir = dentry;
 892
 893         dentry = simple_mkdir(current->fs->pwd, "LOGS", 0777, 1);
 894         if (IS_ERR(dentry)) {
 895                 rc = PTR_ERR(dentry);
 896                 CERROR("cannot create LOGS directory: rc = %d\n", rc);
 897                 GOTO(err_pending, rc);
 898         }
 899         mds->mds_logs_dir = dentry;
 900
 901         dentry = simple_mkdir(current->fs->pwd, "OBJECTS", 0777, 1);
 902         if (IS_ERR(dentry)) {
 903                 rc = PTR_ERR(dentry);
 904                 CERROR("cannot create OBJECTS directory: rc = %d\n", rc);
 905                 GOTO(err_logs, rc);
 906         }
 907         mds->mds_objects_dir = dentry;
 908
 909         dentry = simple_mkdir(current->fs->pwd, "FIDS", 0777, 1);
 910         if (IS_ERR(dentry)) {
 911                 rc = PTR_ERR(dentry);
 912                 CERROR("cannot create FIDS directory: rc = %d\n", rc);
 913                 GOTO(err_objects, rc);
 914         }
 915         mds->mds_id_dir = dentry;
 916
 917         dentry = simple_mkdir(current->fs->pwd, "UNNAMED", 0777, 1);
 918         if (IS_ERR(dentry)) {
 919                 rc = PTR_ERR(dentry);
 920                 CERROR("cannot create UNNAMED directory: rc = %d\n", rc);
 921                 GOTO(err_id_dir, rc);
 922         }
 923         mds->mds_unnamed_dir = dentry;
 924
 925         /* open and test the last rcvd file */
 926         file = filp_open(LAST_RCVD, O_RDWR | O_CREAT, 0644);
 927         if (IS_ERR(file)) {
 928                 rc = PTR_ERR(file);
 929                 CERROR("cannot open/create %s file: rc = %d\n", LAST_RCVD, rc);
 930                 GOTO(err_unnamed, rc = PTR_ERR(file));
 931         }
 932         mds->mds_rcvd_filp = file;
 933
 934         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 935                 CERROR("%s is not a regular file!: mode = %o\n", LAST_RCVD,
 936                        file->f_dentry->d_inode->i_mode);
 937                 GOTO(err_last_rcvd, rc = -ENOENT);
 938         }
 939
 940         rc = mds_read_last_rcvd(obd, file);
 941         if (rc) {
 942                 CERROR("cannot read %s: rc = %d\n", LAST_RCVD, rc);
 943                 GOTO(err_last_rcvd, rc);
 944         }
 945
 946         /* open and test last fid file */
 947         file = filp_open(LAST_FID, O_RDWR | O_CREAT, 0644);
 948         if (IS_ERR(file)) {
 949                 rc = PTR_ERR(file);
 950                 CERROR("cannot open/create %s file: rc = %d\n",
 951                        LAST_FID, rc);
 952                 GOTO(err_client, rc = PTR_ERR(file));
 953         }
 954         mds->mds_fid_filp = file;
 955         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 956                 CERROR("%s is not a regular file!: mode = %o\n",
 957                        LAST_FID, file->f_dentry->d_inode->i_mode);
 958                 GOTO(err_last_fid, rc = -ENOENT);
 959         }
 960
 961         rc = mds_read_last_fid(obd, file);
 962         if (rc) {
 963                 CERROR("cannot read %s: rc = %d\n", LAST_FID, rc);
 964                 GOTO(err_last_fid, rc);
 965         }
 966
 967         /* open and test virtid fid file */
 968         file = filp_open(VIRT_FID, O_RDWR | O_CREAT, 0644);
 969         if (IS_ERR(file)) {
 970                 rc = PTR_ERR(file);
 971                 CERROR("cannot open/create %s file: rc = %d\n",
 972                        VIRT_FID, rc);
 973                 GOTO(err_last_fid, rc = PTR_ERR(file));
 974         }
 975         mds->mds_virtid_filp = file;
 976         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 977                 CERROR("%s is not a regular file!: mode = %o\n",
 978                        VIRT_FID, file->f_dentry->d_inode->i_mode);
 979                 GOTO(err_virtid_fid, rc = -ENOENT);
 980         }
 981
 982         rc = mds_read_virtid_fid(obd, file);
 983         if (rc) {
 984                 CERROR("cannot read %s: rc = %d\n", VIRT_FID, rc);
 985                 GOTO(err_virtid_fid, rc);
 986         }
 987
 988         /* open and test the lov objid file */
 989         file = filp_open(LOV_OBJID, O_RDWR | O_CREAT, 0644);
 990         if (IS_ERR(file)) {
 991                 rc = PTR_ERR(file);
 992                 CERROR("cannot open/create %s file: rc = %d\n", LOV_OBJID, rc);
 993                 GOTO(err_last_fid, rc = PTR_ERR(file));
 994         }
 995         mds->mds_dt_objid_filp = file;
 996         if (!S_ISREG(file->f_dentry->d_inode->i_mode)) {
 997                 CERROR("%s is not a regular file!: mode = %o\n", LOV_OBJID,
 998                        file->f_dentry->d_inode->i_mode);
 999                 GOTO(err_lov_objid, rc = -ENOENT);
1000         }
1001
1002         /* reint fidext thumb by last fid after root and virt are initialized */
1003         mds->mds_fidext_thumb = mds->mds_last_fid;
1004
1005         rc = mds_fidmap_init(obd, MDS_FIDMAP_SIZE);
1006         if (rc) {
1007                 CERROR("cannot init fid mapping tables, err %d\n", rc);
1008                 GOTO(err_lov_objid, rc);
1009         }
1010
1011 err_pop:
1012         pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
1013         return rc;
1014
1015 err_lov_objid:
1016         if (mds->mds_dt_objid_filp && filp_close(mds->mds_dt_objid_filp, 0))
1017                 CERROR("can't close %s after error\n", LOV_OBJID);
1018 err_virtid_fid:
1019         if (mds->mds_virtid_filp && filp_close(mds->mds_virtid_filp, 0))
1020                 CERROR("can't close %s after error\n", VIRT_FID);
1021 err_last_fid:
1022         if (mds->mds_fid_filp && filp_close(mds->mds_fid_filp, 0))
1023                 CERROR("can't close %s after error\n", LAST_FID);
1024 err_client:
1025         class_disconnect_exports(obd, 0);
1026 err_last_rcvd:
1027         if (mds->mds_rcvd_filp && filp_close(mds->mds_rcvd_filp, 0))
1028                 CERROR("can't close %s after error\n", LAST_RCVD);
1029 err_unnamed:
1030         dput(mds->mds_unnamed_dir);
1031 err_id_dir:
1032         dput(mds->mds_id_dir);
1033 err_objects:
1034         dput(mds->mds_objects_dir);
1035 err_logs:
1036         dput(mds->mds_logs_dir);
1037 err_pending:
1038         dput(mds->mds_pending_dir);
1039 err_id_de:
1040         dput(mds->mds_id_de);
1041         goto err_pop;
1042 }
1043
1044 static int  mds_fs_post_cleanup(struct obd_device *obd)
1045 {
1046         int    rc = 0;
1047         rc = fsfilt_post_cleanup(obd);
1048         return rc;
1049 }
1050
1051 int mds_fs_cleanup(struct obd_device *obd, int flags)
1052 {
1053         struct mds_obd *mds = &obd->u.mds;
1054         struct lvfs_run_ctxt saved;
1055         int rc = 0;
1056
1057         if (flags & OBD_OPT_FAILOVER)
1058                 CERROR("%s: shutting down for failover; client state will"
1059                        " be preserved.\n", obd->obd_name);
1060
1061         class_disconnect_exports(obd, flags); /* cleans up client info too */
1062         target_cleanup_recovery(obd);
1063         mds_server_free_data(mds);
1064         mds_fidmap_cleanup(obd);
1065
1066         push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
1067         if (mds->mds_virtid_filp) {
1068                 rc = filp_close(mds->mds_virtid_filp, 0);
1069                 mds->mds_virtid_filp = NULL;
1070                 if (rc)
1071                         CERROR("%s file won't close, rc = %d\n", VIRT_FID, rc);
1072         }
1073         if (mds->mds_fid_filp) {
1074                 rc = filp_close(mds->mds_fid_filp, 0);
1075                 mds->mds_fid_filp = NULL;
1076                 if (rc)
1077                         CERROR("%s file won't close, rc = %d\n", LAST_FID, rc);
1078         }
1079         if (mds->mds_rcvd_filp) {
1080                 rc = filp_close(mds->mds_rcvd_filp, 0);
1081                 mds->mds_rcvd_filp = NULL;
1082                 if (rc)
1083                         CERROR("%s file won't close, rc = %d\n", LAST_RCVD, rc);
1084         }
1085         if (mds->mds_dt_objid_filp) {
1086                 rc = filp_close(mds->mds_dt_objid_filp, 0);
1087                 mds->mds_dt_objid_filp = NULL;
1088                 if (rc)
1089                         CERROR("%s file won't close, rc=%d\n", LOV_OBJID, rc);
1090         }
1091         if (mds->mds_unnamed_dir != NULL) {
1092                 l_dput(mds->mds_unnamed_dir);
1093                 mds->mds_unnamed_dir = NULL;
1094         }
1095         if (mds->mds_id_dir != NULL) {
1096                 l_dput(mds->mds_id_dir);
1097                 mds->mds_id_dir = NULL;
1098         }
1099         if (mds->mds_objects_dir != NULL) {
1100                 l_dput(mds->mds_objects_dir);
1101                 mds->mds_objects_dir = NULL;
1102         }
1103         if (mds->mds_logs_dir) {
1104                 l_dput(mds->mds_logs_dir);
1105                 mds->mds_logs_dir = NULL;
1106         }
1107         if (mds->mds_pending_dir) {
1108                 l_dput(mds->mds_pending_dir);
1109                 mds->mds_pending_dir = NULL;
1110         }
1111         rc = mds_fs_post_cleanup(obd);
1112
1113         pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
1114         shrink_dcache_parent(mds->mds_id_de);
1115         dput(mds->mds_id_de);
1116
1117         return rc;
1118 }
1119
1120 /* Creates an object with the same name as its id.  Because this is not at all
1121  * performance sensitive, it is accomplished by creating a file, checking the
1122  * id, and renaming it. */
1123 int mds_obd_create(struct obd_export *exp, struct obdo *oa,
1124                    void *acl, int acl_size,
1125                    struct lov_stripe_md **ea, struct obd_trans_info *oti)
1126 {
1127         struct mds_obd *mds = &exp->exp_obd->u.mds;
1128         struct inode *parent_inode = mds->mds_objects_dir->d_inode;
1129         struct file *filp;
1130         struct dentry *dchild;
1131         struct lvfs_run_ctxt saved;
1132         char idname[LL_ID_NAMELEN];
1133         int rc = 0, err, idlen;
1134         void *handle;
1135         ENTRY;
1136
1137         push_ctxt(&saved, &exp->exp_obd->obd_lvfs_ctxt, NULL);
1138         down(&parent_inode->i_sem);
1139         if (oa->o_id) {
1140                 idlen = ll_id2str(idname, oa->o_id, oa->o_generation);
1141                 dchild = lookup_one_len(idname, mds->mds_objects_dir, idlen);
1142                 if (IS_ERR(dchild))
1143                         GOTO(out_pop, rc = PTR_ERR(dchild));
1144
1145                 if (dchild->d_inode == NULL) {
1146                         struct dentry_params dp;
1147                         struct inode *inode;
1148
1149                         CWARN("creating log with ID "LPU64"\n", oa->o_id);
1150
1151                         dchild->d_fsdata = (void *) &dp;
1152                         dp.p_ptr = NULL;
1153                         dp.p_inum = oa->o_id;
1154                         rc = ll_vfs_create(parent_inode, dchild, S_IFREG, NULL);
1155                         if (dchild->d_fsdata == (void *)(unsigned long)oa->o_id)
1156                                 dchild->d_fsdata = NULL;
1157                         if (rc) {
1158                                 CDEBUG(D_INODE, "err during create: %d\n", rc);
1159                                 dput(dchild);
1160                                 GOTO(out_pop, rc);
1161                         }
1162                         inode = dchild->d_inode;
1163                         LASSERT(inode->i_ino == oa->o_id);
1164                         inode->i_generation = oa->o_generation;
1165                         CDEBUG(D_HA, "recreated ino %lu with gen %u\n",
1166                                inode->i_ino, inode->i_generation);
1167                         mark_inode_dirty(inode);
1168                 } else {
1169                         CWARN("it should be here!\n");
1170                 }
1171                 GOTO(out_pop, rc);
1172         }
1173
1174         sprintf(idname, "OBJECTS/%u.%u", ll_insecure_random_int(), current->pid);
1175         filp = filp_open(idname, O_CREAT | O_EXCL, 0644);
1176         if (IS_ERR(filp)) {
1177                 rc = PTR_ERR(filp);
1178                 if (rc == -EEXIST) {
1179                         CERROR("impossible object name collision %s\n",
1180                                idname);
1181                         LBUG();
1182                 }
1183                 CERROR("error creating tmp object %s: rc %d\n",
1184                        idname, rc);
1185                 GOTO(out_pop, rc);
1186         }
1187
1188         LASSERT(mds->mds_objects_dir == filp->f_dentry->d_parent);
1189
1190         oa->o_id = filp->f_dentry->d_inode->i_ino;
1191         oa->o_generation = filp->f_dentry->d_inode->i_generation;
1192         idlen = ll_id2str(idname, oa->o_id, oa->o_generation);
1193
1194         CWARN("created log anonymous "LPU64"/%u\n",
1195               oa->o_id, oa->o_generation);
1196
1197         dchild = lookup_one_len(idname, mds->mds_objects_dir, idlen);
1198         if (IS_ERR(dchild)) {
1199                 CERROR("getting neg dentry for obj rename: %d\n", rc);
1200                 GOTO(out_close, rc = PTR_ERR(dchild));
1201         }
1202         if (dchild->d_inode != NULL) {
1203                 CERROR("impossible non-negative obj dentry " LPU64":%u!\n",
1204                        oa->o_id, oa->o_generation);
1205                 LBUG();
1206         }
1207
1208         handle = fsfilt_start(exp->exp_obd, mds->mds_objects_dir->d_inode,
1209                               FSFILT_OP_RENAME, NULL);
1210         if (IS_ERR(handle))
1211                 GOTO(out_dput, rc = PTR_ERR(handle));
1212
1213         lock_kernel();
1214         rc = vfs_rename(mds->mds_objects_dir->d_inode, filp->f_dentry,
1215                         mds->mds_objects_dir->d_inode, dchild);
1216         unlock_kernel();
1217         if (rc)
1218                 CERROR("error renaming new object "LPU64":%u: rc %d\n",
1219                        oa->o_id, oa->o_generation, rc);
1220
1221         err = fsfilt_commit(exp->exp_obd, mds->mds_sb,
1222                             mds->mds_objects_dir->d_inode, handle, 0);
1223         if (!err) {
1224                 oa->o_gr = FILTER_GROUP_FIRST_MDS + mds->mds_num;
1225                 oa->o_valid |= OBD_MD_FLID | OBD_MD_FLGENER | OBD_MD_FLGROUP;
1226         } else if (!rc)
1227                 rc = err;
1228 out_dput:
1229         dput(dchild);
1230 out_close:
1231         err = filp_close(filp, 0);
1232         if (err) {
1233                 CERROR("closing tmpfile %s: rc %d\n", idname, rc);
1234                 if (!rc)
1235                         rc = err;
1236         }
1237 out_pop:
1238         up(&parent_inode->i_sem);
1239         pop_ctxt(&saved, &exp->exp_obd->obd_lvfs_ctxt, NULL);
1240         RETURN(rc);
1241 }
1242
1243 int mds_obd_destroy(struct obd_export *exp, struct obdo *oa,
1244                     struct lov_stripe_md *ea, struct obd_trans_info *oti)
1245 {
1246         struct mds_obd *mds = &exp->exp_obd->u.mds;
1247         struct inode *parent_inode = mds->mds_objects_dir->d_inode;
1248         struct obd_device *obd = exp->exp_obd;
1249         struct lvfs_run_ctxt saved;
1250         char idname[LL_ID_NAMELEN];
1251         struct dentry *de;
1252         void *handle;
1253         int err, idlen, rc = 0;
1254         ENTRY;
1255
1256         push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
1257
1258         idlen = ll_id2str(idname, oa->o_id, oa->o_generation);
1259
1260         down(&parent_inode->i_sem);
1261         de = lookup_one_len(idname, mds->mds_objects_dir, idlen);
1262         if (IS_ERR(de) || de->d_inode == NULL) {
1263                 rc = IS_ERR(de) ? PTR_ERR(de) : -ENOENT;
1264                 CERROR("destroying non-existent object "LPU64" %s: rc %d\n",
1265                        oa->o_id, idname, rc);
1266                 GOTO(out_dput, rc);
1267         }
1268         /* Stripe count is 1 here since this is some MDS specific stuff
1269            that is unlinked, not spanned across multiple OSTs */
1270         handle = fsfilt_start_log(obd, mds->mds_objects_dir->d_inode,
1271                                   FSFILT_OP_UNLINK, oti, 1);
1272
1273         if (IS_ERR(handle))
1274                 GOTO(out_dput, rc = PTR_ERR(handle));
1275
1276         rc = vfs_unlink(mds->mds_objects_dir->d_inode, de);
1277         if (rc)
1278                 CERROR("error destroying object "LPU64":%u: rc %d\n",
1279                        oa->o_id, oa->o_generation, rc);
1280
1281         err = fsfilt_commit(obd, mds->mds_sb, mds->mds_objects_dir->d_inode,
1282                             handle, exp->exp_sync);
1283         if (err && !rc)
1284                 rc = err;
1285 out_dput:
1286         if (de != NULL)
1287                 l_dput(de);
1288         up(&parent_inode->i_sem);
1289         pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
1290         RETURN(rc);
1291 }