lustre/mds/handler.c

   1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
   2  * vim:expandtab:shiftwidth=8:tabstop=8:
   3  *
   4  *  linux/mds/handler.c
   5  *
   6  *  Lustre Metadata Server (mds) request handler
   7  *
   8  *  Copyright (C) 2001, 2002 Cluster File Systems, Inc.
   9  *
  10  *  This code is issued under the GNU General Public License.
  11  *  See the file COPYING in this distribution
  12  *
  13  *  by Peter Braam <braam@clusterfs.com>
  14  *
  15  *  This server is single threaded at present (but can easily be multi threaded)
  16  *
  17  */
  18
  19 #define EXPORT_SYMTAB
  20
  21 #include <linux/version.h>
  22 #include <linux/module.h>
  23 #include <linux/fs.h>
  24 #include <linux/stat.h>
  25 #include <linux/locks.h>
  26 #include <linux/quotaops.h>
  27 #include <asm/unistd.h>
  28 #include <asm/uaccess.h>
  29
  30 #define DEBUG_SUBSYSTEM S_MDS
  31
  32 #include <linux/lustre_mds.h>
  33 #include <linux/lustre_lib.h>
  34 #include <linux/lustre_net.h>
  35
  36 int mds_sendpage(struct ptlrpc_request *req, struct file *file,
  37                  __u64 offset, struct niobuf *dst)
  38 {
  39         int rc = 0;
  40         mm_segment_t oldfs = get_fs();
  41         struct ptlrpc_bulk_desc *bulk;
  42         char *buf;
  43
  44         bulk = ptlrpc_prep_bulk(req->rq_connection);
  45         if (bulk == NULL) {
  46                 rc = -ENOMEM;
  47                 GOTO(out, rc);
  48         }
  49
  50         bulk->b_xid = req->rq_reqmsg->xid;
  51
  52         OBD_ALLOC(buf, PAGE_SIZE);
  53         if (!buf) {
  54                 rc = -ENOMEM;
  55                 GOTO(cleanup_bulk, rc);
  56         }
  57
  58         set_fs(KERNEL_DS);
  59         rc = mds_fs_readpage(&req->rq_obd->u.mds, file, buf, PAGE_SIZE,
  60                              (loff_t *)&offset);
  61         set_fs(oldfs);
  62
  63         if (rc != PAGE_SIZE) {
  64                 rc = -EIO;
  65                 GOTO(cleanup_buf, rc);
  66         }
  67
  68         bulk->b_buf = buf;
  69         bulk->b_buflen = PAGE_SIZE;
  70
  71         rc = ptlrpc_send_bulk(bulk, MDS_BULK_PORTAL);
  72         if (OBD_FAIL_CHECK(OBD_FAIL_MDS_SENDPAGE)) {
  73                 CERROR("obd_fail_loc=%x, fail operation rc=%d\n",
  74                        OBD_FAIL_MDS_SENDPAGE, rc);
  75                 PtlMDUnlink(bulk->b_md_h);
  76                 GOTO(cleanup_buf, rc);
  77         }
  78         wait_event_interruptible(bulk->b_waitq,
  79                                  ptlrpc_check_bulk_sent(bulk));
  80
  81         if (bulk->b_flags & PTL_RPC_FL_INTR) {
  82                 rc = -EINTR;
  83                 GOTO(cleanup_buf, rc);
  84         }
  85
  86         EXIT;
  87  cleanup_buf:
  88         OBD_FREE(buf, PAGE_SIZE);
  89  cleanup_bulk:
  90         ptlrpc_free_bulk(bulk);
  91  out:
  92         return rc;
  93 }
  94
  95 struct dentry *mds_fid2dentry(struct mds_obd *mds, struct ll_fid *fid,
  96                               struct vfsmount **mnt)
  97 {
  98         /* stolen from NFS */
  99         struct super_block *sb = mds->mds_sb;
 100         unsigned long ino = fid->id;
 101         __u32 generation = fid->generation;
 102         struct inode *inode;
 103         struct list_head *lp;
 104         struct dentry *result;
 105
 106         if (ino == 0)
 107                 return ERR_PTR(-ESTALE);
 108
 109         inode = iget(sb, ino);
 110         if (inode == NULL)
 111                 return ERR_PTR(-ENOMEM);
 112
 113         CDEBUG(D_DENTRY, "--> mds_fid2dentry: sb %p\n", inode->i_sb);
 114
 115         if (is_bad_inode(inode) ||
 116             (generation && inode->i_generation != generation)) {
 117                 /* we didn't find the right inode.. */
 118                 CERROR("bad inode %lu, link: %d ct: %d or version  %u/%u\n",
 119                         inode->i_ino,
 120                         inode->i_nlink, atomic_read(&inode->i_count),
 121                         inode->i_generation,
 122                         generation);
 123                 LBUG();
 124                 iput(inode);
 125                 return ERR_PTR(-ESTALE);
 126         }
 127
 128         /* now to find a dentry.
 129          * If possible, get a well-connected one
 130          */
 131         if (mnt)
 132                 *mnt = mds->mds_vfsmnt;
 133         spin_lock(&dcache_lock);
 134         for (lp = inode->i_dentry.next; lp != &inode->i_dentry ; lp=lp->next) {
 135                 result = list_entry(lp,struct dentry, d_alias);
 136                 if (! (result->d_flags & DCACHE_NFSD_DISCONNECTED)) {
 137                         dget_locked(result);
 138                         result->d_vfs_flags |= DCACHE_REFERENCED;
 139                         spin_unlock(&dcache_lock);
 140                         iput(inode);
 141                         if (mnt)
 142                                 mntget(*mnt);
 143                         return result;
 144                 }
 145         }
 146         spin_unlock(&dcache_lock);
 147         result = d_alloc_root(inode);
 148         if (result == NULL) {
 149                 iput(inode);
 150                 return ERR_PTR(-ENOMEM);
 151         }
 152         if (mnt)
 153                 mntget(*mnt);
 154         result->d_flags |= DCACHE_NFSD_DISCONNECTED;
 155         return result;
 156 }
 157
 158 int mds_connect(struct ptlrpc_request *req)
 159 {
 160         struct mds_body *body;
 161         struct mds_obd *mds = &req->rq_obd->u.mds;
 162         int rc, size = sizeof(*body);
 163         ENTRY;
 164
 165         rc = lustre_pack_msg(1, &size, NULL, &req->rq_replen, &req->rq_repmsg);
 166         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_CONNECT_PACK)) {
 167                 CERROR("mds: out of memory\n");
 168                 req->rq_status = -ENOMEM;
 169                 RETURN(0);
 170         }
 171
 172         body = lustre_msg_buf(req->rq_reqmsg, 0);
 173         mds_unpack_req_body(req);
 174         /* Anything we need to do here with the client's trans no or so? */
 175
 176         body = lustre_msg_buf(req->rq_repmsg, 0);
 177         memcpy(&body->fid1, &mds->mds_rootfid  , sizeof(body->fid1));
 178         mds_pack_rep_body(req);
 179         RETURN(0);
 180 }
 181
 182 int mds_getattr(struct ptlrpc_request *req)
 183 {
 184         struct dentry *de;
 185         struct inode *inode;
 186         struct mds_body *body;
 187         struct mds_obd *mds = &req->rq_obd->u.mds;
 188         int rc, size = sizeof(*body);
 189         ENTRY;
 190
 191         rc = lustre_pack_msg(1, &size, NULL, &req->rq_replen, &req->rq_repmsg);
 192         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_GETATTR_PACK)) {
 193                 CERROR("mds: out of memory\n");
 194                 req->rq_status = -ENOMEM;
 195                 RETURN(0);
 196         }
 197
 198         body = lustre_msg_buf(req->rq_reqmsg, 0);
 199         de = mds_fid2dentry(mds, &body->fid1, NULL);
 200         if (IS_ERR(de)) {
 201                 req->rq_status = -ENOENT;
 202                 RETURN(0);
 203         }
 204
 205         body = lustre_msg_buf(req->rq_repmsg, 0);
 206         inode = de->d_inode;
 207         body->ino = inode->i_ino;
 208         body->generation = inode->i_generation;
 209         body->atime = inode->i_atime;
 210         body->ctime = inode->i_ctime;
 211         body->mtime = inode->i_mtime;
 212         body->uid = inode->i_uid;
 213         body->gid = inode->i_gid;
 214         body->size = inode->i_size;
 215         body->mode = inode->i_mode;
 216         body->nlink = inode->i_nlink;
 217         body->valid = ~0;
 218         mds_fs_get_objid(mds, inode, &body->objid);
 219         l_dput(de);
 220         RETURN(0);
 221 }
 222
 223 int mds_open(struct ptlrpc_request *req)
 224 {
 225         struct dentry *de;
 226         struct mds_body *body;
 227         struct file *file;
 228         struct vfsmount *mnt;
 229         __u32 flags;
 230         int rc, size = sizeof(*body);
 231         ENTRY;
 232
 233         rc = lustre_pack_msg(1, &size, NULL, &req->rq_replen, &req->rq_repmsg);
 234         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_OPEN_PACK)) {
 235                 CERROR("mds: out of memory\n");
 236                 req->rq_status = -ENOMEM;
 237                 RETURN(0);
 238         }
 239
 240         body = lustre_msg_buf(req->rq_reqmsg, 0);
 241         de = mds_fid2dentry(&req->rq_obd->u.mds, &body->fid1, &mnt);
 242         if (IS_ERR(de)) {
 243                 req->rq_status = -ENOENT;
 244                 RETURN(0);
 245         }
 246         flags = body->flags;
 247         file = dentry_open(de, mnt, flags);
 248         if (!file || IS_ERR(file)) {
 249                 req->rq_status = -EINVAL;
 250                 RETURN(0);
 251         }
 252
 253         body = lustre_msg_buf(req->rq_repmsg, 0);
 254         body->objid = (__u64) (unsigned long)file;
 255         RETURN(0);
 256 }
 257
 258 int mds_close(struct ptlrpc_request *req)
 259 {
 260         struct dentry *de;
 261         struct mds_body *body;
 262         struct file *file;
 263         struct vfsmount *mnt;
 264         int rc;
 265         ENTRY;
 266
 267         rc = lustre_pack_msg(0, NULL, NULL, &req->rq_replen, &req->rq_repmsg);
 268         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_CLOSE_PACK)) {
 269                 CERROR("mds: out of memory\n");
 270                 req->rq_status = -ENOMEM;
 271                 RETURN(0);
 272         }
 273
 274         body = lustre_msg_buf(req->rq_reqmsg, 0);
 275         de = mds_fid2dentry(&req->rq_obd->u.mds, &body->fid1, &mnt);
 276         if (IS_ERR(de)) {
 277                 req->rq_status = -ENOENT;
 278                 RETURN(0);
 279         }
 280
 281         file = (struct file *)(unsigned long)body->objid;
 282         req->rq_status = filp_close(file, 0);
 283         l_dput(de);
 284         mntput(mnt);
 285
 286         RETURN(0);
 287 }
 288
 289 int mds_readpage(struct ptlrpc_request *req)
 290 {
 291         struct vfsmount *mnt;
 292         struct dentry *de;
 293         struct file *file;
 294         struct niobuf *niobuf;
 295         struct mds_body *body;
 296         int rc, size = sizeof(*body);
 297         ENTRY;
 298
 299         rc = lustre_pack_msg(1, &size, NULL, &req->rq_replen, &req->rq_repmsg);
 300         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_READPAGE_PACK)) {
 301                 CERROR("mds: out of memory\n");
 302                 req->rq_status = -ENOMEM;
 303                 RETURN(0);
 304         }
 305
 306         body = lustre_msg_buf(req->rq_reqmsg, 0);
 307         de = mds_fid2dentry(&req->rq_obd->u.mds, &body->fid1, &mnt);
 308         if (IS_ERR(de)) {
 309                 req->rq_status = PTR_ERR(de);
 310                 RETURN(0);
 311         }
 312
 313         CDEBUG(D_INODE, "ino %ld\n", de->d_inode->i_ino);
 314
 315         file = dentry_open(de, mnt, O_RDONLY | O_LARGEFILE);
 316         /* note: in case of an error, dentry_open puts dentry */
 317         if (IS_ERR(file)) {
 318                 req->rq_status = PTR_ERR(file);
 319                 RETURN(0);
 320         }
 321
 322         niobuf = lustre_msg_buf(req->rq_reqmsg, 1);
 323         if (!niobuf) {
 324                 req->rq_status = -EINVAL;
 325                 LBUG();
 326                 RETURN(0);
 327         }
 328
 329         /* to make this asynchronous make sure that the handling function
 330            doesn't send a reply when this function completes. Instead a
 331            callback function would send the reply */
 332         rc = mds_sendpage(req, file, body->size, niobuf);
 333
 334         filp_close(file, 0);
 335         req->rq_status = rc;
 336         RETURN(0);
 337 }
 338
 339 int mds_reint(struct ptlrpc_request *req)
 340 {
 341         int rc;
 342         struct mds_update_record rec;
 343
 344         rc = mds_update_unpack(req, &rec);
 345         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_REINT_UNPACK)) {
 346                 CERROR("invalid record\n");
 347                 req->rq_status = -EINVAL;
 348                 RETURN(0);
 349         }
 350         /* rc will be used to interrupt a for loop over multiple records */
 351         rc = mds_reint_rec(&rec, req);
 352         return 0;
 353 }
 354
 355 int mds_handle(struct obd_device *dev, struct ptlrpc_service *svc,
 356                struct ptlrpc_request *req)
 357 {
 358         int rc;
 359         ENTRY;
 360
 361         rc = lustre_unpack_msg(req->rq_reqmsg, req->rq_reqlen);
 362         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_HANDLE_UNPACK)) {
 363                 CERROR("lustre_mds: Invalid request\n");
 364                 GOTO(out, rc);
 365         }
 366
 367         if (req->rq_reqmsg->type != PTL_RPC_MSG_REQUEST) {
 368                 CERROR("lustre_mds: wrong packet type sent %d\n",
 369                        req->rq_reqmsg->type);
 370                 GOTO(out, rc = -EINVAL);
 371         }
 372
 373         switch (req->rq_reqmsg->opc) {
 374         case MDS_CONNECT:
 375                 CDEBUG(D_INODE, "getattr\n");
 376                 OBD_FAIL_RETURN(OBD_FAIL_MDS_CONNECT_NET, 0);
 377                 rc = mds_connect(req);
 378                 break;
 379
 380         case MDS_GETATTR:
 381                 CDEBUG(D_INODE, "getattr\n");
 382                 OBD_FAIL_RETURN(OBD_FAIL_MDS_GETATTR_NET, 0);
 383                 rc = mds_getattr(req);
 384                 break;
 385
 386         case MDS_READPAGE:
 387                 CDEBUG(D_INODE, "readpage\n");
 388                 OBD_FAIL_RETURN(OBD_FAIL_MDS_READPAGE_NET, 0);
 389                 rc = mds_readpage(req);
 390
 391                 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_SENDPAGE))
 392                         return 0;
 393                 break;
 394
 395         case MDS_REINT:
 396                 CDEBUG(D_INODE, "reint\n");
 397                 OBD_FAIL_RETURN(OBD_FAIL_MDS_REINT_NET, 0);
 398                 rc = mds_reint(req);
 399                 break;
 400
 401         case MDS_OPEN:
 402                 CDEBUG(D_INODE, "open\n");
 403                 OBD_FAIL_RETURN(OBD_FAIL_MDS_OPEN_NET, 0);
 404                 rc = mds_open(req);
 405                 break;
 406
 407         case MDS_CLOSE:
 408                 CDEBUG(D_INODE, "close\n");
 409                 OBD_FAIL_RETURN(OBD_FAIL_MDS_CLOSE_NET, 0);
 410                 rc = mds_close(req);
 411                 break;
 412
 413         default:
 414                 rc = ptlrpc_error(svc, req);
 415                 RETURN(rc);
 416         }
 417
 418         EXIT;
 419 out:
 420         if (rc) {
 421                 ptlrpc_error(svc, req);
 422         } else {
 423                 CDEBUG(D_NET, "sending reply\n");
 424                 ptlrpc_reply(svc, req);
 425         }
 426
 427         return 0;
 428 }
 429
 430 #define LAST_RCVD "last_rcvd"
 431
 432 static int mds_prep(struct obd_device *obddev)
 433 {
 434         struct obd_run_ctxt saved;
 435         struct mds_obd *mds = &obddev->u.mds;
 436         struct super_operations *s_ops;
 437         struct file *f;
 438         loff_t off = 0;
 439         __u64 mount_count;
 440         int rc;
 441
 442         push_ctxt(&saved, &mds->mds_ctxt);
 443         rc = simple_mkdir(current->fs->pwd, "ROOT", 0700);
 444         if (rc && rc != -EEXIST) {
 445                 CERROR("cannot create ROOT directory: rc = %d\n", rc);
 446                 GOTO(err_pop, rc);
 447         }
 448         f = filp_open("ROOT", O_RDONLY, 0);
 449         if (IS_ERR(f)) {
 450                 rc = PTR_ERR(f);
 451                 CERROR("cannot open ROOT: rc = %d\n", rc);
 452                 LBUG();
 453                 GOTO(err_pop, rc);
 454         }
 455
 456         mds->mds_rootfid.id = f->f_dentry->d_inode->i_ino;
 457         mds->mds_rootfid.generation = f->f_dentry->d_inode->i_generation;
 458         mds->mds_rootfid.f_type = S_IFDIR;
 459
 460         rc = filp_close(f, 0);
 461         if (rc) {
 462                 CERROR("cannot close ROOT: rc = %d\n", rc);
 463                 LBUG();
 464         }
 465
 466         rc = simple_mkdir(current->fs->pwd, "FH", 0700);
 467         if (rc && rc != -EEXIST) {
 468                 CERROR("cannot create FH directory: rc = %d\n", rc);
 469                 GOTO(err_pop, rc);
 470         }
 471
 472         f = filp_open("mount_count", O_RDWR | O_CREAT, 0644);
 473         if (IS_ERR(f)) {
 474                 rc = PTR_ERR(f);
 475                 CERROR("cannot open/create mount_count file, rc = %d\n", rc);
 476                 GOTO(err_pop, rc = PTR_ERR(f));
 477         }
 478         rc = lustre_fread(f, (char *)&mount_count, sizeof(mount_count), &off);
 479         if (rc == 0) {
 480                 CERROR("empty MDS mount_count, new MDS?\n");
 481                 /* XXX maybe this should just be a random number? */
 482                 mds->mds_mount_count = 0;
 483         } else if (rc != sizeof(mount_count)) {
 484                 CERROR("error reading mount_count: rc = %d\n", rc);
 485                 /* XXX maybe this should just be a random number? */
 486                 mds->mds_mount_count = 0;
 487         } else {
 488                 mds->mds_mount_count = le64_to_cpu(mount_count);
 489         }
 490
 491         mds->mds_mount_count++;
 492         CDEBUG(D_SUPER, "MDS mount_count is %Ld\n",
 493                (unsigned long long)mds->mds_mount_count);
 494         off = 0;
 495         mount_count = cpu_to_le64(mds->mds_mount_count);
 496         rc = lustre_fwrite(f, (char *)&mount_count, sizeof(mount_count), &off);
 497         if (rc != sizeof(mount_count))
 498                 CERROR("error writing mount_count: rc = %d\n", rc);
 499         rc = filp_close(f, 0);
 500         if (rc)
 501                 CERROR("error closing mount_count: rc = %d\n", rc);
 502
 503         f = filp_open("last_rcvd", O_RDWR | O_CREAT, 0644);
 504         if (IS_ERR(f)) {
 505                 rc = PTR_ERR(f);
 506                 CERROR("cannot open/create %s file: rc = %d\n", LAST_RCVD, rc);
 507                 GOTO(err_pop, rc);
 508         }
 509         mds->mds_rcvd_filp = f;
 510         pop_ctxt(&saved);
 511
 512         /*
 513          * Replace the client filesystem delete_inode method with our own,
 514          * so that we can clear the object ID before the inode is deleted.
 515          * The fs_delete_inode method will call cl_delete_inode for us.
 516          *
 517          * We need to do this for the MDS superblock only, hence we install
 518          * a modified copy of the original superblock method table.
 519          *
 520          * We still assume that there is only a single MDS client filesystem
 521          * type, as we don't have access to the mds struct in delete_inode
 522          * and store the client delete_inode method in a global table.  This
 523          * will only become a problem when multiple MDSs are running on a
 524          * single host with different client filesystems.
 525          */
 526         OBD_ALLOC(s_ops, sizeof(*s_ops));
 527         if (!s_ops)
 528                 GOTO(err_filp, rc = -ENOMEM);
 529
 530         memcpy(s_ops, mds->mds_sb->s_op, sizeof(*s_ops));
 531         mds->mds_fsops->cl_delete_inode = s_ops->delete_inode;
 532         s_ops->delete_inode = mds->mds_fsops->fs_delete_inode;
 533         mds->mds_sb->s_op = s_ops;
 534
 535         mds->mds_service = ptlrpc_init_svc(128 * 1024,
 536                                            MDS_REQUEST_PORTAL, MDC_REPLY_PORTAL,
 537                                            "self", mds_handle);
 538
 539         if (!mds->mds_service) {
 540                 CERROR("failed to start service\n");
 541                 GOTO(err_filp, rc = -EINVAL);
 542         }
 543
 544         rc = ptlrpc_start_thread(obddev, mds->mds_service, "lustre_mds");
 545         if (rc) {
 546                 CERROR("cannot start thread\n");
 547                 GOTO(err_svc, rc);
 548         }
 549
 550         RETURN(0);
 551
 552 err_svc:
 553         rpc_unregister_service(mds->mds_service);
 554         OBD_FREE(mds->mds_service, sizeof(*mds->mds_service));
 555 err_filp:
 556         if (filp_close(f, 0))
 557                 CERROR("can't close %s after error\n", LAST_RCVD);
 558 err_pop:
 559         pop_ctxt(&saved);
 560
 561         return rc;
 562 }
 563
 564 /* mount the file system (secretly) */
 565 static int mds_setup(struct obd_device *obddev, obd_count len, void *buf)
 566 {
 567         struct obd_ioctl_data* data = buf;
 568         struct mds_obd *mds = &obddev->u.mds;
 569         struct vfsmount *mnt;
 570         int rc = 0;
 571         ENTRY;
 572
 573 #ifdef CONFIG_DEV_RDONLY
 574         dev_clear_rdonly(2);
 575 #endif
 576         if (!data->ioc_inlbuf1 || !data->ioc_inlbuf2)
 577                 RETURN(-EINVAL);
 578
 579         mds->mds_fstype = strdup(data->ioc_inlbuf2);
 580
 581         if (!strcmp(mds->mds_fstype, "ext3"))
 582                 mds->mds_fsops = &mds_ext3_fs_ops;
 583         else if (!strcmp(mds->mds_fstype, "ext2"))
 584                 mds->mds_fsops = &mds_ext2_fs_ops;
 585         else {
 586                 CERROR("unsupported MDS filesystem type %s\n", mds->mds_fstype);
 587                 GOTO(err_kfree, rc = -EPERM);
 588         }
 589
 590         MOD_INC_USE_COUNT;
 591         mnt = do_kern_mount(mds->mds_fstype, 0, data->ioc_inlbuf1, NULL);
 592         if (IS_ERR(mnt)) {
 593                 rc = PTR_ERR(mnt);
 594                 CERROR("do_kern_mount failed: rc = %d\n", rc);
 595                 GOTO(err_dec, rc);
 596         }
 597
 598         mds->mds_sb = mnt->mnt_root->d_inode->i_sb;
 599         if (!mds->mds_sb)
 600                 GOTO(err_put, rc = -ENODEV);
 601
 602         mds->mds_vfsmnt = mnt;
 603         mds->mds_ctxt.pwdmnt = mnt;
 604         mds->mds_ctxt.pwd = mnt->mnt_root;
 605         mds->mds_ctxt.fs = KERNEL_DS;
 606
 607         rc = mds_prep(obddev);
 608         if (rc)
 609                 GOTO(err_put, rc);
 610
 611         RETURN(0);
 612
 613 err_put:
 614         unlock_kernel();
 615         mntput(mds->mds_vfsmnt);
 616         mds->mds_sb = 0;
 617         lock_kernel();
 618 err_dec:
 619         MOD_DEC_USE_COUNT;
 620 err_kfree:
 621         kfree(mds->mds_fstype);
 622         return rc;
 623 }
 624
 625 static int mds_cleanup(struct obd_device * obddev)
 626 {
 627         struct super_operations *s_ops = NULL;
 628         struct super_block *sb;
 629         struct mds_obd *mds = &obddev->u.mds;
 630
 631         ENTRY;
 632
 633         if ( !list_empty(&obddev->obd_gen_clients) ) {
 634                 CERROR("still has clients!\n");
 635                 RETURN(-EBUSY);
 636         }
 637
 638         ptlrpc_stop_thread(mds->mds_service);
 639         rpc_unregister_service(mds->mds_service);
 640         if (!list_empty(&mds->mds_service->srv_reqs)) {
 641                 // XXX reply with errors and clean up
 642                 CERROR("Request list not empty!\n");
 643         }
 644         OBD_FREE(mds->mds_service, sizeof(*mds->mds_service));
 645
 646         sb = mds->mds_sb;
 647         if (!mds->mds_sb)
 648                 RETURN(0);
 649
 650         if (mds->mds_rcvd_filp) {
 651                 int rc = filp_close(mds->mds_rcvd_filp, 0);
 652                 mds->mds_rcvd_filp = NULL;
 653
 654                 if (rc)
 655                         CERROR("last_rcvd file won't close, rc=%d\n", rc);
 656         }
 657         s_ops = sb->s_op;
 658
 659         unlock_kernel();
 660         mntput(mds->mds_vfsmnt);
 661         mds->mds_sb = 0;
 662         kfree(mds->mds_fstype);
 663         lock_kernel();
 664 #ifdef CONFIG_DEV_RDONLY
 665         dev_clear_rdonly(2);
 666 #endif
 667         OBD_FREE(s_ops, sizeof(*s_ops));
 668
 669         MOD_DEC_USE_COUNT;
 670         RETURN(0);
 671 }
 672
 673 /* use obd ops to offer management infrastructure */
 674 static struct obd_ops mds_obd_ops = {
 675         o_setup:       mds_setup,
 676         o_cleanup:     mds_cleanup,
 677 };
 678
 679 static int __init mds_init(void)
 680 {
 681         obd_register_type(&mds_obd_ops, LUSTRE_MDS_NAME);
 682         return 0;
 683 }
 684
 685 static void __exit mds_exit(void)
 686 {
 687         obd_unregister_type(LUSTRE_MDS_NAME);
 688 }
 689
 690 MODULE_AUTHOR("Peter J. Braam <braam@clusterfs.com>");
 691 MODULE_DESCRIPTION("Lustre Metadata Server (MDS) v0.01");
 692 MODULE_LICENSE("GPL");
 693
 694 module_init(mds_init);
 695 module_exit(mds_exit);