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
  42         if (req->rq_peer.peer_nid == 0) {
  43                 /* dst->addr is a user address, but in a different task! */
  44                 char *buf = (char *)(long)dst->addr;
  45
  46                 set_fs(KERNEL_DS);
  47                 rc = mds_fs_readpage(&req->rq_obd->u.mds, file, buf, PAGE_SIZE,
  48                                      &offset);
  49                 set_fs(oldfs);
  50
  51                 if (rc != PAGE_SIZE) {
  52                         rc = -EIO;
  53                         GOTO(out, rc);
  54                 }
  55                 EXIT;
  56         } else {
  57                 struct ptlrpc_bulk_desc *bulk;
  58                 char *buf;
  59
  60                 bulk = ptlrpc_prep_bulk(&req->rq_peer);
  61                 if (bulk == NULL) {
  62                         rc = -ENOMEM;
  63                         GOTO(out, rc);
  64                 }
  65
  66                 bulk->b_xid = req->rq_xid;
  67
  68                 OBD_ALLOC(buf, PAGE_SIZE);
  69                 if (!buf) {
  70                         rc = -ENOMEM;
  71                         GOTO(cleanup_bulk, rc);
  72                 }
  73
  74                 set_fs(KERNEL_DS);
  75                 rc = mds_fs_readpage(&req->rq_obd->u.mds, file, buf, PAGE_SIZE,
  76                                      &offset);
  77                 set_fs(oldfs);
  78
  79                 if (rc != PAGE_SIZE) {
  80                         rc = -EIO;
  81                         GOTO(cleanup_buf, rc);
  82                 }
  83
  84                 bulk->b_buf = buf;
  85                 bulk->b_buflen = PAGE_SIZE;
  86
  87                 rc = ptlrpc_send_bulk(bulk, MDS_BULK_PORTAL);
  88                 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_SENDPAGE)) {
  89                         CERROR("obd_fail_loc=%x, fail operation rc=%d\n",
  90                                OBD_FAIL_MDS_SENDPAGE, rc);
  91                         PtlMDUnlink(bulk->b_md_h);
  92                         GOTO(cleanup_buf, rc);
  93                 }
  94                 wait_event_interruptible(bulk->b_waitq,
  95                                          ptlrpc_check_bulk_sent(bulk));
  96
  97                 if (bulk->b_flags == PTL_RPC_INTR) {
  98                         rc = -EINTR;
  99                         GOTO(cleanup_buf, rc);
 100                 }
 101
 102                 EXIT;
 103         cleanup_buf:
 104                 OBD_FREE(buf, PAGE_SIZE);
 105         cleanup_bulk:
 106                 OBD_FREE(bulk, sizeof(*bulk));
 107         }
 108 out:
 109         return rc;
 110 }
 111
 112 struct dentry *mds_fid2dentry(struct mds_obd *mds, struct ll_fid *fid,
 113                               struct vfsmount **mnt)
 114 {
 115         /* stolen from NFS */
 116         struct super_block *sb = mds->mds_sb;
 117         unsigned long ino = fid->id;
 118         __u32 generation = fid->generation;
 119         struct inode *inode;
 120         struct list_head *lp;
 121         struct dentry *result;
 122
 123         if (ino == 0)
 124                 return ERR_PTR(-ESTALE);
 125
 126         inode = iget(sb, ino);
 127         if (inode == NULL)
 128                 return ERR_PTR(-ENOMEM);
 129
 130         CDEBUG(D_DENTRY, "--> mds_fid2dentry: sb %p\n", inode->i_sb);
 131
 132         if (is_bad_inode(inode) ||
 133             (generation && inode->i_generation != generation)) {
 134                 /* we didn't find the right inode.. */
 135                 CERROR("bad inode %lu, link: %d ct: %d or version  %u/%u\n",
 136                         inode->i_ino,
 137                         inode->i_nlink, atomic_read(&inode->i_count),
 138                         inode->i_generation,
 139                         generation);
 140                 LBUG();
 141                 iput(inode);
 142                 return ERR_PTR(-ESTALE);
 143         }
 144
 145         /* now to find a dentry.
 146          * If possible, get a well-connected one
 147          */
 148         if (mnt)
 149                 *mnt = mds->mds_vfsmnt;
 150         spin_lock(&dcache_lock);
 151         for (lp = inode->i_dentry.next; lp != &inode->i_dentry ; lp=lp->next) {
 152                 result = list_entry(lp,struct dentry, d_alias);
 153                 if (! (result->d_flags & DCACHE_NFSD_DISCONNECTED)) {
 154                         dget_locked(result);
 155                         result->d_vfs_flags |= DCACHE_REFERENCED;
 156                         spin_unlock(&dcache_lock);
 157                         iput(inode);
 158                         if (mnt)
 159                                 mntget(*mnt);
 160                         return result;
 161                 }
 162         }
 163         spin_unlock(&dcache_lock);
 164         result = d_alloc_root(inode);
 165         if (result == NULL) {
 166                 iput(inode);
 167                 return ERR_PTR(-ENOMEM);
 168         }
 169         if (mnt)
 170                 mntget(*mnt);
 171         result->d_flags |= DCACHE_NFSD_DISCONNECTED;
 172         return result;
 173 }
 174
 175 int mds_getattr(struct ptlrpc_request *req)
 176 {
 177         struct dentry *de;
 178         struct inode *inode;
 179         struct mds_rep *rep;
 180         struct mds_obd *mds = &req->rq_obd->u.mds;
 181         int rc;
 182
 183         rc = mds_pack_rep(NULL, 0, NULL, 0, &req->rq_rephdr, &req->rq_rep,
 184                           &req->rq_replen, &req->rq_repbuf);
 185         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_GETATTR_PACK)) {
 186                 CERROR("mds: out of memory\n");
 187                 req->rq_status = -ENOMEM;
 188                 RETURN(0);
 189         }
 190
 191         req->rq_rephdr->xid = req->rq_reqhdr->xid;
 192         rep = req->rq_rep.mds;
 193
 194         de = mds_fid2dentry(mds, &req->rq_req.mds->fid1, NULL);
 195         if (IS_ERR(de)) {
 196                 req->rq_rephdr->status = -ENOENT;
 197                 RETURN(0);
 198         }
 199
 200         inode = de->d_inode;
 201         rep->ino = inode->i_ino;
 202         rep->generation = inode->i_generation;
 203         rep->atime = inode->i_atime;
 204         rep->ctime = inode->i_ctime;
 205         rep->mtime = inode->i_mtime;
 206         rep->uid = inode->i_uid;
 207         rep->gid = inode->i_gid;
 208         rep->size = inode->i_size;
 209         rep->mode = inode->i_mode;
 210         rep->nlink = inode->i_nlink;
 211         rep->valid = ~0;
 212         mds_fs_get_objid(mds, inode, &rep->objid);
 213         l_dput(de);
 214         return 0;
 215 }
 216
 217 int mds_open(struct ptlrpc_request *req)
 218 {
 219         struct dentry *de;
 220         struct mds_rep *rep;
 221         struct file *file;
 222         struct vfsmount *mnt;
 223         __u32 flags;
 224         int rc;
 225
 226         rc = mds_pack_rep(NULL, 0, NULL, 0, &req->rq_rephdr, &req->rq_rep,
 227                           &req->rq_replen, &req->rq_repbuf);
 228         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_OPEN_PACK)) {
 229                 CERROR("mds: out of memory\n");
 230                 req->rq_status = -ENOMEM;
 231                 RETURN(0);
 232         }
 233
 234         req->rq_rephdr->xid = req->rq_reqhdr->xid;
 235         rep = req->rq_rep.mds;
 236
 237         de = mds_fid2dentry(&req->rq_obd->u.mds, &req->rq_req.mds->fid1, &mnt);
 238         if (IS_ERR(de)) {
 239                 req->rq_rephdr->status = -ENOENT;
 240                 RETURN(0);
 241         }
 242         flags = req->rq_req.mds->flags;
 243         file = dentry_open(de, mnt, flags);
 244         if (!file || IS_ERR(file)) {
 245                 req->rq_rephdr->status = -EINVAL;
 246                 RETURN(0);
 247         }
 248
 249         rep->objid = (__u64) (unsigned long)file;
 250         return 0;
 251 }
 252
 253 int mds_close(struct ptlrpc_request *req)
 254 {
 255         struct dentry *de;
 256         struct mds_rep *rep;
 257         struct file *file;
 258         struct vfsmount *mnt;
 259         int rc;
 260
 261         rc = mds_pack_rep(NULL, 0, NULL, 0, &req->rq_rephdr, &req->rq_rep,
 262                           &req->rq_replen, &req->rq_repbuf);
 263         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_CLOSE_PACK)) {
 264                 CERROR("mds: out of memory\n");
 265                 req->rq_status = -ENOMEM;
 266                 RETURN(0);
 267         }
 268
 269         req->rq_rephdr->xid = req->rq_reqhdr->xid;
 270         rep = req->rq_rep.mds;
 271
 272         de = mds_fid2dentry(&req->rq_obd->u.mds, &req->rq_req.mds->fid1, &mnt);
 273         if (IS_ERR(de)) {
 274                 req->rq_rephdr->status = -ENOENT;
 275                 RETURN(0);
 276         }
 277
 278         file = (struct file *)(unsigned long) req->rq_req.mds->objid;
 279
 280         req->rq_rephdr->status = filp_close(file, 0);
 281         l_dput(de);
 282         mntput(mnt);
 283         return 0;
 284 }
 285
 286 int mds_readpage(struct ptlrpc_request *req)
 287 {
 288         struct vfsmount *mnt;
 289         struct dentry *de;
 290         struct file *file;
 291         struct niobuf *niobuf;
 292         struct mds_rep *rep;
 293         int rc;
 294
 295         ENTRY;
 296
 297         rc = mds_pack_rep(NULL, 0, NULL, 0, &req->rq_rephdr, &req->rq_rep,
 298                           &req->rq_replen, &req->rq_repbuf);
 299         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_READPAGE_PACK)) {
 300                 CERROR("mds: out of memory\n");
 301                 req->rq_status = -ENOMEM;
 302                 RETURN(0);
 303         }
 304
 305         req->rq_rephdr->xid = req->rq_reqhdr->xid;
 306         rep = req->rq_rep.mds;
 307
 308         de = mds_fid2dentry(&req->rq_obd->u.mds, &req->rq_req.mds->fid1, &mnt);
 309         if (IS_ERR(de)) {
 310                 req->rq_rephdr->status = PTR_ERR(de);
 311                 RETURN(0);
 312         }
 313
 314         CDEBUG(D_INODE, "ino %ld\n", de->d_inode->i_ino);
 315
 316         file = dentry_open(de, mnt, O_RDONLY | O_LARGEFILE);
 317         /* note: in case of an error, dentry_open puts dentry */
 318         if (IS_ERR(file)) {
 319                 req->rq_rephdr->status = PTR_ERR(file);
 320                 RETURN(0);
 321         }
 322
 323         niobuf = mds_req_tgt(req->rq_req.mds);
 324
 325         /* to make this asynchronous make sure that the handling function
 326            doesn't send a reply when this function completes. Instead a
 327            callback function would send the reply */
 328         rc = mds_sendpage(req, file, req->rq_req.mds->size, niobuf);
 329
 330         filp_close(file, 0);
 331         req->rq_rephdr->status = rc;
 332         RETURN(0);
 333 }
 334
 335 int mds_reint(struct ptlrpc_request *req)
 336 {
 337         char *buf;
 338         int rc, len;
 339         struct mds_update_record rec;
 340
 341         buf = mds_req_tgt(req->rq_req.mds);
 342         len = req->rq_req.mds->tgtlen;
 343
 344         rc = mds_update_unpack(buf, len, &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         struct ptlreq_hdr *hdr;
 360
 361         ENTRY;
 362
 363         hdr = (struct ptlreq_hdr *)req->rq_reqbuf;
 364
 365         if (NTOH__u32(hdr->type) != PTL_RPC_REQUEST) {
 366                 CERROR("lustre_mds: wrong packet type sent %d\n",
 367                        NTOH__u32(hdr->type));
 368                 rc = -EINVAL;
 369                 GOTO(out, rc);
 370         }
 371
 372         rc = mds_unpack_req(req->rq_reqbuf, req->rq_reqlen,
 373                             &req->rq_reqhdr, &req->rq_req);
 374         if (rc || OBD_FAIL_CHECK(OBD_FAIL_MDS_HANDLE_UNPACK)) {
 375                 CERROR("lustre_mds: Invalid request\n");
 376                 GOTO(out, rc);
 377         }
 378
 379         switch (req->rq_reqhdr->opc) {
 380
 381         case MDS_GETATTR:
 382                 CDEBUG(D_INODE, "getattr\n");
 383                 OBD_FAIL_RETURN(OBD_FAIL_MDS_GETATTR_NET, 0);
 384                 rc = mds_getattr(req);
 385                 break;
 386
 387         case MDS_READPAGE:
 388                 CDEBUG(D_INODE, "readpage\n");
 389                 OBD_FAIL_RETURN(OBD_FAIL_MDS_READPAGE_NET, 0);
 390                 rc = mds_readpage(req);
 391
 392                 if (OBD_FAIL_CHECK(OBD_FAIL_MDS_SENDPAGE))
 393                         return 0;
 394                 break;
 395
 396         case MDS_REINT:
 397                 CDEBUG(D_INODE, "reint\n");
 398                 OBD_FAIL_RETURN(OBD_FAIL_MDS_REINT_NET, 0);
 399                 rc = mds_reint(req);
 400                 break;
 401
 402         case MDS_OPEN:
 403                 CDEBUG(D_INODE, "open\n");
 404                 OBD_FAIL_RETURN(OBD_FAIL_MDS_OPEN_NET, 0);
 405                 rc = mds_open(req);
 406                 break;
 407
 408         case MDS_CLOSE:
 409                 CDEBUG(D_INODE, "close\n");
 410                 OBD_FAIL_RETURN(OBD_FAIL_MDS_CLOSE_NET, 0);
 411                 rc = mds_close(req);
 412                 break;
 413
 414         default:
 415                 rc = ptlrpc_error(dev, svc, req);
 416                 RETURN(rc);
 417         }
 418
 419         EXIT;
 420 out:
 421         if (rc) {
 422                 CERROR("no header\n");
 423                 LBUG();
 424                 return 0;
 425         }
 426
 427         if( req->rq_status) {
 428                 ptlrpc_error(dev, svc, req);
 429         } else {
 430                 CDEBUG(D_NET, "sending reply\n");
 431                 ptlrpc_reply(dev, svc, req);
 432         }
 433
 434         return 0;
 435 }
 436
 437 static int mds_prep(struct obd_device *obddev)
 438 {
 439         struct obd_run_ctxt saved;
 440         struct mds_obd *mds = &obddev->u.mds;
 441         struct super_operations *s_ops;
 442         int err;
 443
 444         mds->mds_service = ptlrpc_init_svc(128 * 1024,
 445                                            MDS_REQUEST_PORTAL, MDC_REPLY_PORTAL,
 446                                            "self", mds_handle);
 447
 448         if (!mds->mds_service) {
 449                 CERROR("failed to start service\n");
 450                 RETURN(-EINVAL);
 451         }
 452
 453         err = ptlrpc_start_thread(obddev, mds->mds_service, "lustre_mds");
 454         if (err) {
 455                 CERROR("cannot start thread\n");
 456                 GOTO(err_svc, err);
 457         }
 458
 459         push_ctxt(&saved, &mds->mds_ctxt);
 460         err = simple_mkdir(current->fs->pwd, "ROOT", 0700);
 461         err = simple_mkdir(current->fs->pwd, "FH", 0700);
 462         pop_ctxt(&saved);
 463
 464         /*
 465          * Replace the client filesystem delete_inode method with our own,
 466          * so that we can clear the object ID before the inode is deleted.
 467          * The fs_delete_inode method will call cl_delete_inode for us.
 468          *
 469          * We need to do this for the MDS superblock only, hence we install
 470          * a modified copy of the original superblock method table.
 471          *
 472          * We still assume that there is only a single MDS client filesystem
 473          * type, as we don't have access to the mds struct in * delete_inode.
 474          */
 475         OBD_ALLOC(s_ops, sizeof(*s_ops));
 476         memcpy(s_ops, mds->mds_sb->s_op, sizeof(*s_ops));
 477         mds->mds_fsops->cl_delete_inode = s_ops->delete_inode;
 478         s_ops->delete_inode = mds->mds_fsops->fs_delete_inode;
 479         mds->mds_sb->s_op = s_ops;
 480
 481         RETURN(0);
 482
 483 err_svc:
 484         rpc_unregister_service(mds->mds_service);
 485         OBD_FREE(mds->mds_service, sizeof(*mds->mds_service));
 486
 487         return(err);
 488 }
 489
 490 /* mount the file system (secretly) */
 491 static int mds_setup(struct obd_device *obddev, obd_count len, void *buf)
 492 {
 493         struct obd_ioctl_data* data = buf;
 494         struct mds_obd *mds = &obddev->u.mds;
 495         struct vfsmount *mnt;
 496         int err = 0;
 497         ENTRY;
 498
 499 #ifdef CONFIG_DEV_RDONLY
 500         dev_clear_rdonly(2);
 501 #endif
 502         mds->mds_fstype = strdup(data->ioc_inlbuf2);
 503
 504         if (!strcmp(mds->mds_fstype, "ext3"))
 505                 mds->mds_fsops = &mds_ext3_fs_ops;
 506         else if (!strcmp(mds->mds_fstype, "ext2"))
 507                 mds->mds_fsops = &mds_ext2_fs_ops;
 508         else {
 509                 CERROR("unsupported MDS filesystem type %s\n", mds->mds_fstype);
 510                 GOTO(err_kfree, (err = -EPERM));
 511         }
 512
 513         MOD_INC_USE_COUNT;
 514         mnt = do_kern_mount(mds->mds_fstype, 0, data->ioc_inlbuf1, NULL);
 515         if (IS_ERR(mnt)) {
 516                 CERROR("do_kern_mount failed: %d\n", err);
 517                 GOTO(err_dec, err = PTR_ERR(mnt));
 518         }
 519
 520         mds->mds_sb = mnt->mnt_root->d_inode->i_sb;
 521         if (!mds->mds_sb)
 522                 GOTO(err_put, (err = -ENODEV));
 523
 524         mds->mds_vfsmnt = mnt;
 525         mds->mds_ctxt.pwdmnt = mnt;
 526         mds->mds_ctxt.pwd = mnt->mnt_root;
 527         mds->mds_ctxt.fs = KERNEL_DS;
 528
 529         err = mds_prep(obddev);
 530         if (err)
 531                 GOTO(err_put, err);
 532
 533         RETURN(0);
 534
 535 err_put:
 536         unlock_kernel();
 537         mntput(mds->mds_vfsmnt);
 538         mds->mds_sb = 0;
 539         lock_kernel();
 540 err_dec:
 541         MOD_DEC_USE_COUNT;
 542 err_kfree:
 543         kfree(mds->mds_fstype);
 544         return err;
 545 }
 546
 547 static int mds_cleanup(struct obd_device * obddev)
 548 {
 549         struct super_operations *s_ops = NULL;
 550         struct super_block *sb;
 551         struct mds_obd *mds = &obddev->u.mds;
 552
 553         ENTRY;
 554
 555         if ( !list_empty(&obddev->obd_gen_clients) ) {
 556                 CERROR("still has clients!\n");
 557                 RETURN(-EBUSY);
 558         }
 559
 560         ptlrpc_stop_thread(mds->mds_service);
 561         rpc_unregister_service(mds->mds_service);
 562         if (!list_empty(&mds->mds_service->srv_reqs)) {
 563                 // XXX reply with errors and clean up
 564                 CERROR("Request list not empty!\n");
 565         }
 566         OBD_FREE(mds->mds_service, sizeof(*mds->mds_service));
 567
 568         sb = mds->mds_sb;
 569         if (!mds->mds_sb)
 570                 RETURN(0);
 571
 572         s_ops = sb->s_op;
 573
 574         unlock_kernel();
 575         mntput(mds->mds_vfsmnt);
 576         mds->mds_sb = 0;
 577         kfree(mds->mds_fstype);
 578         lock_kernel();
 579 #ifdef CONFIG_DEV_RDONLY
 580         dev_clear_rdonly(2);
 581 #endif
 582         OBD_FREE(s_ops, sizeof(*s_ops));
 583
 584         MOD_DEC_USE_COUNT;
 585         RETURN(0);
 586 }
 587
 588 /* use obd ops to offer management infrastructure */
 589 static struct obd_ops mds_obd_ops = {
 590         o_setup:       mds_setup,
 591         o_cleanup:     mds_cleanup,
 592 };
 593
 594 static int __init mds_init(void)
 595 {
 596         obd_register_type(&mds_obd_ops, LUSTRE_MDS_NAME);
 597         return 0;
 598 }
 599
 600 static void __exit mds_exit(void)
 601 {
 602         obd_unregister_type(LUSTRE_MDS_NAME);
 603 }
 604
 605 MODULE_AUTHOR("Peter J. Braam <braam@clusterfs.com>");
 606 MODULE_DESCRIPTION("Lustre Metadata Server (MDS) v0.01");
 607 MODULE_LICENSE("GPL");
 608
 609 module_init(mds_init);
 610 module_exit(mds_exit);