lustre/mdc/mdc_request.c

   1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
   2  * vim:expandtab:shiftwidth=8:tabstop=8:
   3  *
   4  * Copyright (C) 2001, 2002 Cluster File Systems, Inc.
   5  *
   6  *   This file is part of Lustre, http://www.sf.net/projects/lustre/
   7  *
   8  *   Lustre is free software; you can redistribute it and/or
   9  *   modify it under the terms of version 2 of the GNU General Public
  10  *   License as published by the Free Software Foundation.
  11  *
  12  *   Lustre is distributed in the hope that it will be useful,
  13  *   but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  *   GNU General Public License for more details.
  16  *
  17  *   You should have received a copy of the GNU General Public License
  18  *   along with Lustre; if not, write to the Free Software
  19  *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  20  *
  21  */
  22
  23 #define EXPORT_SYMTAB
  24 #define DEBUG_SUBSYSTEM S_MDC
  25
  26 #include <linux/module.h>
  27 #include <linux/miscdevice.h>
  28 #include <linux/lustre_mds.h>
  29 #include <linux/lustre_lite.h>
  30 #include <linux/lustre_dlm.h>
  31 #include <linux/init.h>
  32 #include <linux/obd_lov.h>
  33
  34 #define REQUEST_MINOR 244
  35
  36 extern int mds_queue_req(struct ptlrpc_request *);
  37
  38 int mdc_getstatus(struct lustre_handle *conn, struct ll_fid *rootfid,
  39                   __u64 *last_committed, __u64 *last_xid,
  40                   struct ptlrpc_request **request)
  41 {
  42         struct ptlrpc_request *req;
  43         struct mds_body *body;
  44         int rc, size = sizeof(*body);
  45         ENTRY;
  46
  47         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_GETSTATUS, 1, &size,
  48                               NULL);
  49         if (!req)
  50                 GOTO(out, rc = -ENOMEM);
  51
  52         body = lustre_msg_buf(req->rq_reqmsg, 0);
  53         req->rq_level = LUSTRE_CONN_CON;
  54         req->rq_replen = lustre_msg_size(1, &size);
  55
  56         mds_pack_req_body(req);
  57         rc = ptlrpc_queue_wait(req);
  58         rc = ptlrpc_check_status(req, rc);
  59
  60         if (!rc) {
  61                 body = lustre_msg_buf(req->rq_repmsg, 0);
  62                 mds_unpack_body(body);
  63                 memcpy(rootfid, &body->fid1, sizeof(*rootfid));
  64                 *last_committed = req->rq_repmsg->last_committed;
  65                 *last_xid = req->rq_repmsg->last_xid;
  66
  67                 CDEBUG(D_NET,"root ino=%ld, last_committed=%Lu, last_xid=%Ld\n",
  68                        (unsigned long)rootfid->id,
  69                        (unsigned long long)*last_committed,
  70                        (unsigned long long)*last_xid);
  71         }
  72
  73         EXIT;
  74  out:
  75         ptlrpc_free_req(req);
  76         return rc;
  77 }
  78
  79 int mdc_getlovinfo(struct obd_device *obd, struct lustre_handle *mdc_connh,
  80                    struct ptlrpc_request **request)
  81 {
  82         struct ptlrpc_request *req;
  83         struct mds_status_req *streq;
  84         int rc, size[2] = {sizeof(*streq)};
  85         ENTRY;
  86
  87         req = ptlrpc_prep_req(class_conn2cliimp(mdc_connh), MDS_GETLOVINFO, 1,
  88                               size, NULL);
  89         if (!req)
  90                 GOTO(out, rc = -ENOMEM);
  91
  92         *request = req;
  93         streq = lustre_msg_buf(req->rq_reqmsg, 0);
  94         streq->flags = HTON__u32(MDS_STATUS_LOV);
  95         streq->repbuf = HTON__u32(8192);
  96
  97         /* prepare for reply */
  98         req->rq_level = LUSTRE_CONN_CON;
  99         size[0] = 512;
 100         size[1] = 8192;
 101         req->rq_replen = lustre_msg_size(2, size);
 102
 103         rc = ptlrpc_queue_wait(req);
 104         rc = ptlrpc_check_status(req, rc);
 105
 106  out:
 107         RETURN(rc);
 108 }
 109
 110
 111 int mdc_getattr(struct lustre_handle *conn,
 112                 obd_id ino, int type, unsigned long valid, size_t ea_size,
 113                 struct ptlrpc_request **request)
 114 {
 115         struct ptlrpc_request *req;
 116         struct mds_body *body;
 117         int rc, size[2] = {sizeof(*body), 0}, bufcount = 1;
 118         ENTRY;
 119
 120         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_GETATTR, 1, size,
 121                               NULL);
 122         if (!req)
 123                 GOTO(out, rc = -ENOMEM);
 124
 125         body = lustre_msg_buf(req->rq_reqmsg, 0);
 126         ll_ino2fid(&body->fid1, ino, 0, type);
 127         body->valid = valid;
 128         mds_pack_req_body(req);
 129
 130         if (S_ISREG(type)) {
 131                 struct client_obd *mdc = &class_conn2obd(conn)->u.cli;
 132                 bufcount = 2;
 133                 size[1] = mdc->cl_max_mds_easize;
 134         } else if (valid & OBD_MD_LINKNAME) {
 135                 bufcount = 2;
 136                 size[1] = ea_size;
 137         }
 138         req->rq_replen = lustre_msg_size(bufcount, size);
 139
 140         rc = ptlrpc_queue_wait(req);
 141         rc = ptlrpc_check_status(req, rc);
 142
 143         if (!rc) {
 144                 body = lustre_msg_buf(req->rq_repmsg, 0);
 145                 mds_unpack_body(body);
 146                 CDEBUG(D_NET, "mode: %o\n", body->mode);
 147         }
 148
 149         EXIT;
 150  out:
 151         *request = req;
 152         return rc;
 153 }
 154
 155 static int mdc_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc,
 156                             void *data, __u32 data_len)
 157 {
 158         int rc;
 159         struct inode *inode = data;
 160         struct lustre_handle lockh;
 161         ENTRY;
 162
 163         if (data_len != sizeof(*inode)) {
 164                 CERROR("data_len should be %d, but is %d\n", sizeof(*inode),
 165                        data_len);
 166                 LBUG();
 167                 RETURN(-EINVAL);
 168         }
 169
 170         if (inode == NULL)
 171                 LBUG();
 172
 173         /* FIXME: do something better than throwing away everything */
 174         if (S_ISDIR(inode->i_mode)) {
 175                 CDEBUG(D_INODE, "invalidating inode %ld\n",
 176                        inode->i_ino);
 177                 invalidate_inode_pages(inode);
 178         }
 179
 180         ldlm_lock2handle(lock, &lockh);
 181         rc = ldlm_cli_cancel(&lockh);
 182         if (rc < 0) {
 183                 CERROR("ldlm_cli_cancel: %d\n", rc);
 184                 LBUG();
 185         }
 186         RETURN(0);
 187 }
 188
 189 int mdc_enqueue(struct lustre_handle *conn, int lock_type,
 190                 struct lookup_intent *it, int lock_mode, struct inode *dir,
 191                 struct dentry *de, struct lustre_handle *lockh,
 192                 char *tgt, int tgtlen, void *data, int datalen)
 193 {
 194         struct ptlrpc_request *req;
 195         struct obd_device *obddev = class_conn2obd(conn);
 196         __u64 res_id[RES_NAME_SIZE] = {dir->i_ino};
 197         int size[5] = {sizeof(struct ldlm_request), sizeof(struct ldlm_intent)};
 198         int rc, flags;
 199         int repsize[3] = {sizeof(struct ldlm_reply),
 200                           sizeof(struct mds_body),
 201                           obddev->u.cli.cl_max_mds_easize};
 202         struct ldlm_reply *dlm_rep;
 203         struct ldlm_intent *lit;
 204         ENTRY;
 205
 206         LDLM_DEBUG_NOLOCK("mdsintent %s dir %ld", ldlm_it2str(it->it_op),
 207                           dir->i_ino);
 208
 209         switch (it->it_op) {
 210         case IT_MKDIR:
 211                 it->it_mode = (it->it_mode | S_IFDIR) & ~current->fs->umask;
 212                 break;
 213         case (IT_CREAT|IT_OPEN):
 214         case IT_CREAT:
 215                 it->it_mode |= S_IFREG; /* no break */
 216         case IT_MKNOD:
 217                 it->it_mode &= ~current->fs->umask;
 218                 break;
 219         case IT_SYMLINK:
 220                 it->it_mode = (it->it_mode | S_IFLNK) & ~current->fs->umask;
 221                 break;
 222         }
 223
 224         if (it->it_op & (IT_MKDIR | IT_CREAT | IT_SYMLINK | IT_MKNOD)) {
 225                 size[2] = sizeof(struct mds_rec_create);
 226                 size[3] = de->d_name.len + 1;
 227                 size[4] = tgtlen + 1;
 228                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 5,
 229                                       size, NULL);
 230                 if (!req)
 231                         RETURN(-ENOMEM);
 232
 233                 /* pack the intent */
 234                 lit = lustre_msg_buf(req->rq_reqmsg, 1);
 235                 lit->opc = NTOH__u64((__u64)it->it_op);
 236
 237                 /* pack the intended request */
 238                 mds_create_pack(req, 2, dir, it->it_mode, 0, current->fsuid,
 239                                 current->fsgid, CURRENT_TIME, de->d_name.name,
 240                                 de->d_name.len, tgt, tgtlen);
 241                 req->rq_replen = lustre_msg_size(3, repsize);
 242         } else if (it->it_op == IT_RENAME2) {
 243                 struct dentry *old_de = it->it_data;
 244
 245                 size[2] = sizeof(struct mds_rec_rename);
 246                 size[3] = old_de->d_name.len + 1;
 247                 size[4] = de->d_name.len + 1;
 248                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 5,
 249                                       size, NULL);
 250                 if (!req)
 251                         RETURN(-ENOMEM);
 252
 253                 /* pack the intent */
 254                 lit = lustre_msg_buf(req->rq_reqmsg, 1);
 255                 lit->opc = NTOH__u64((__u64)it->it_op);
 256
 257                 /* pack the intended request */
 258                 mds_rename_pack(req, 2, old_de->d_parent->d_inode, dir,
 259                                 old_de->d_name.name, old_de->d_name.len,
 260                                 de->d_name.name, de->d_name.len);
 261                 req->rq_replen = lustre_msg_size(3, repsize);
 262         } else if (it->it_op == IT_LINK2) {
 263                 struct dentry *old_de = it->it_data;
 264
 265                 size[2] = sizeof(struct mds_rec_link);
 266                 size[3] = de->d_name.len + 1;
 267                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 4,
 268                                       size, NULL);
 269                 if (!req)
 270                         RETURN(-ENOMEM);
 271
 272                 /* pack the intent */
 273                 lit = lustre_msg_buf(req->rq_reqmsg, 1);
 274                 lit->opc = NTOH__u64((__u64)it->it_op);
 275
 276                 /* pack the intended request */
 277                 mds_link_pack(req, 2, old_de->d_inode, dir,
 278                                 de->d_name.name, de->d_name.len);
 279                 req->rq_replen = lustre_msg_size(3, repsize);
 280         } else if (it->it_op == IT_UNLINK || it->it_op == IT_RMDIR) {
 281                 size[2] = sizeof(struct mds_rec_unlink);
 282                 size[3] = de->d_name.len + 1;
 283                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 4,
 284                                       size, NULL);
 285                 if (!req)
 286                         RETURN(-ENOMEM);
 287
 288                 /* pack the intent */
 289                 lit = lustre_msg_buf(req->rq_reqmsg, 1);
 290                 lit->opc = NTOH__u64((__u64)it->it_op);
 291
 292                 /* pack the intended request */
 293                 mds_unlink_pack(req, 2, dir, NULL,
 294                                 it->it_op == IT_UNLINK ? S_IFREG : S_IFDIR,
 295                                 de->d_name.name, de->d_name.len);
 296
 297                 req->rq_replen = lustre_msg_size(3, repsize);
 298         } else if (it->it_op  & (IT_GETATTR | IT_RENAME | IT_LINK |
 299                    IT_OPEN |  IT_SETATTR | IT_LOOKUP | IT_READLINK)) {
 300                 size[2] = sizeof(struct mds_body);
 301                 size[3] = de->d_name.len + 1;
 302
 303                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 4,
 304                                       size, NULL);
 305                 if (!req)
 306                         RETURN(-ENOMEM);
 307
 308                 /* pack the intent */
 309                 lit = lustre_msg_buf(req->rq_reqmsg, 1);
 310                 lit->opc = NTOH__u64((__u64)it->it_op);
 311
 312                 /* pack the intended request */
 313                 mds_getattr_pack(req, 2, dir, de->d_name.name, de->d_name.len);
 314
 315                 /* we need to replay opens */
 316                 if (it->it_op == IT_OPEN)
 317                         req->rq_flags |= PTL_RPC_FL_REPLAY;
 318
 319                 /* get ready for the reply */
 320                 req->rq_replen = lustre_msg_size(3, repsize);
 321         } else if (it->it_op == IT_READDIR) {
 322                 req = ptlrpc_prep_req(class_conn2cliimp(conn), LDLM_ENQUEUE, 1,
 323                                       size, NULL);
 324                 if (!req)
 325                         RETURN(-ENOMEM);
 326
 327                 /* get ready for the reply */
 328                 req->rq_replen = lustre_msg_size(1, repsize);
 329         } else {
 330                 LBUG();
 331                 RETURN(-EINVAL);
 332         }
 333 #warning FIXME: the data here needs to be different if a lock was granted for a different inode
 334         rc = ldlm_cli_enqueue(conn, req, obddev->obd_namespace, NULL, res_id,
 335                               lock_type, NULL, 0, lock_mode, &flags,
 336                               ldlm_completion_ast, mdc_blocking_ast, data,
 337                               datalen, lockh);
 338         if (rc == -ENOENT) {
 339                 /* This can go when we're sure that this can never happen */
 340                 LBUG();
 341         }
 342         if (rc == ELDLM_LOCK_ABORTED) {
 343                 lock_mode = 0;
 344                 memset(lockh, 0, sizeof(*lockh));
 345                 /* rc = 0 */
 346         } else if (rc != 0) {
 347                 CERROR("ldlm_cli_enqueue: %d\n", rc);
 348                 RETURN(rc);
 349         }
 350
 351         dlm_rep = lustre_msg_buf(req->rq_repmsg, 0);
 352         it->it_disposition = (int) dlm_rep->lock_policy_res1;
 353         it->it_status = (int) dlm_rep->lock_policy_res2;
 354         it->it_lock_mode = lock_mode;
 355         it->it_data = req;
 356
 357         RETURN(0);
 358 }
 359
 360 static void mdc_replay_open(struct ptlrpc_request *req, void *data)
 361 {
 362         __u64 *fh = data;
 363         struct mds_body *body;
 364
 365         body = lustre_msg_buf(req->rq_repmsg, 0);
 366         mds_unpack_body(body);
 367         *fh = body->extra;
 368 }
 369
 370 int mdc_open(struct lustre_handle *conn, obd_id ino, int type, int flags,
 371              struct lov_stripe_md *lsm, __u64 cookie, __u64 *fh,
 372              struct ptlrpc_request **request)
 373 {
 374         struct mds_body *body;
 375         int rc, size[2] = {sizeof(*body)}, bufcount = 1;
 376         struct ptlrpc_request *req;
 377         ENTRY;
 378
 379         if (lsm) {
 380                 bufcount = 2;
 381                 // size[1] = mdc->cl_max_mds_easize; soon...
 382                 size[1] = lsm->lsm_mds_easize;
 383         }
 384
 385         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_OPEN, bufcount, size,
 386                               NULL);
 387         if (!req)
 388                 GOTO(out, rc = -ENOMEM);
 389
 390         req->rq_flags |= PTL_RPC_FL_REPLAY;
 391         body = lustre_msg_buf(req->rq_reqmsg, 0);
 392
 393         ll_ino2fid(&body->fid1, ino, 0, type);
 394         body->flags = HTON__u32(flags);
 395         body->extra = cookie;
 396
 397         if (lsm)
 398                 lov_packmd(lustre_msg_buf(req->rq_reqmsg, 1), lsm);
 399
 400         req->rq_replen = lustre_msg_size(1, size);
 401
 402         rc = ptlrpc_queue_wait(req);
 403         rc = ptlrpc_check_status(req, rc);
 404         if (!rc) {
 405                 body = lustre_msg_buf(req->rq_repmsg, 0);
 406                 mds_unpack_body(body);
 407                 *fh = body->extra;
 408         }
 409
 410         /* If open is replayed, we need to fix up the fh. */
 411         req->rq_replay_cb = mdc_replay_open;
 412         req->rq_replay_cb_data = fh;
 413
 414         EXIT;
 415  out:
 416         *request = req;
 417         return rc;
 418 }
 419
 420 int mdc_close(struct lustre_handle *conn,
 421               obd_id ino, int type, __u64 fh, struct ptlrpc_request **request)
 422 {
 423         struct mds_body *body;
 424         int rc, size = sizeof(*body);
 425         struct ptlrpc_request *req;
 426
 427         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_CLOSE, 1, &size,
 428                               NULL);
 429         if (!req)
 430                 GOTO(out, rc = -ENOMEM);
 431
 432         body = lustre_msg_buf(req->rq_reqmsg, 0);
 433         ll_ino2fid(&body->fid1, ino, 0, type);
 434         body->extra = fh;
 435
 436         req->rq_replen = lustre_msg_size(0, NULL);
 437
 438         rc = ptlrpc_queue_wait(req);
 439         rc = ptlrpc_check_status(req, rc);
 440
 441         EXIT;
 442  out:
 443         *request = req;
 444         return rc;
 445 }
 446
 447 int mdc_readpage(struct lustre_handle *conn, obd_id ino, int type, __u64 offset,
 448                  char *addr, struct ptlrpc_request **request)
 449 {
 450         struct ptlrpc_connection *connection =
 451                 client_conn2cli(conn)->cl_import.imp_connection;
 452         struct ptlrpc_request *req = NULL;
 453         struct ptlrpc_bulk_desc *desc = NULL;
 454         struct ptlrpc_bulk_page *bulk = NULL;
 455         struct mds_body *body;
 456         int rc, size = sizeof(*body);
 457         ENTRY;
 458
 459         CDEBUG(D_INODE, "inode: %ld\n", (long)ino);
 460
 461         desc = ptlrpc_prep_bulk(connection);
 462         if (desc == NULL)
 463                 GOTO(out, rc = -ENOMEM);
 464
 465         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_READPAGE, 1, &size,
 466                               NULL);
 467         if (!req)
 468                 GOTO(out2, rc = -ENOMEM);
 469
 470         bulk = ptlrpc_prep_bulk_page(desc);
 471         bulk->bp_buflen = PAGE_SIZE;
 472         bulk->bp_buf = addr;
 473         bulk->bp_xid = req->rq_xid;
 474         desc->bd_portal = MDS_BULK_PORTAL;
 475
 476         rc = ptlrpc_register_bulk(desc);
 477         if (rc) {
 478                 CERROR("couldn't setup bulk sink: error %d.\n", rc);
 479                 GOTO(out2, rc);
 480         }
 481
 482         body = lustre_msg_buf(req->rq_reqmsg, 0);
 483         body->fid1.id = ino;
 484         body->fid1.f_type = type;
 485         body->size = offset;
 486
 487         req->rq_replen = lustre_msg_size(1, &size);
 488         rc = ptlrpc_queue_wait(req);
 489         rc = ptlrpc_check_status(req, rc);
 490         if (rc) {
 491                 ptlrpc_abort_bulk(desc);
 492                 GOTO(out2, rc);
 493         } else {
 494                 body = lustre_msg_buf(req->rq_repmsg, 0);
 495                 mds_unpack_body(body);
 496         }
 497
 498         EXIT;
 499  out2:
 500         ptlrpc_free_bulk(desc);
 501  out:
 502         *request = req;
 503         return rc;
 504 }
 505
 506 int mdc_statfs(struct lustre_handle *conn, struct obd_statfs *osfs,
 507                struct ptlrpc_request **request)
 508 {
 509         struct ptlrpc_request *req;
 510         int rc, size = sizeof(*osfs);
 511         ENTRY;
 512
 513         req = ptlrpc_prep_req(class_conn2cliimp(conn), MDS_STATFS, 0, NULL,
 514                               NULL);
 515         if (!req)
 516                 GOTO(out, rc = -ENOMEM);
 517         req->rq_replen = lustre_msg_size(1, &size);
 518
 519         rc = ptlrpc_queue_wait(req);
 520         rc = ptlrpc_check_status(req, rc);
 521
 522         if (rc)
 523                 GOTO(out, rc);
 524
 525         obd_statfs_unpack(osfs, lustre_msg_buf(req->rq_repmsg, 0));
 526
 527         EXIT;
 528 out:
 529         *request = req;
 530
 531         return rc;
 532 }
 533
 534 struct obd_ops mdc_obd_ops = {
 535         o_setup:   client_obd_setup,
 536         o_cleanup: client_obd_cleanup,
 537         o_connect: client_obd_connect,
 538         o_disconnect: client_obd_disconnect,
 539 };
 540
 541 static int __init ptlrpc_request_init(void)
 542 {
 543         return class_register_type(&mdc_obd_ops, LUSTRE_MDC_NAME);
 544 }
 545
 546 static void __exit ptlrpc_request_exit(void)
 547 {
 548         class_unregister_type(LUSTRE_MDC_NAME);
 549 }
 550
 551 MODULE_AUTHOR("Cluster File Systems <info@clusterfs.com>");
 552 MODULE_DESCRIPTION("Lustre Metadata Client v1.0");
 553 MODULE_LICENSE("GPL");
 554
 555 EXPORT_SYMBOL(mdc_getstatus);
 556 EXPORT_SYMBOL(mdc_getlovinfo);
 557 EXPORT_SYMBOL(mdc_enqueue);
 558 EXPORT_SYMBOL(mdc_getattr);
 559 EXPORT_SYMBOL(mdc_statfs);
 560 EXPORT_SYMBOL(mdc_create);
 561 EXPORT_SYMBOL(mdc_unlink);
 562 EXPORT_SYMBOL(mdc_rename);
 563 EXPORT_SYMBOL(mdc_link);
 564 EXPORT_SYMBOL(mdc_readpage);
 565 EXPORT_SYMBOL(mdc_setattr);
 566 EXPORT_SYMBOL(mdc_close);
 567 EXPORT_SYMBOL(mdc_open);
 568
 569 module_init(ptlrpc_request_init);
 570 module_exit(ptlrpc_request_exit);