drivers/staging/lustre/lustre/osc/osc_request.c

   1 /*
   2  * GPL HEADER START
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 only,
   8  * as published by the Free Software Foundation.
   9  *
  10  * This program is distributed in the hope that it will be useful, but
  11  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * General Public License version 2 for more details (a copy is included
  14  * in the LICENSE file that accompanied this code).
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * version 2 along with this program; If not, see
  18  * http://www.gnu.org/licenses/gpl-2.0.html
  19  *
  20  * GPL HEADER END
  21  */
  22 /*
  23  * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved.
  24  * Use is subject to license terms.
  25  *
  26  * Copyright (c) 2011, 2015, Intel Corporation.
  27  */
  28 /*
  29  * This file is part of Lustre, http://www.lustre.org/
  30  * Lustre is a trademark of Sun Microsystems, Inc.
  31  */
  32
  33 #define DEBUG_SUBSYSTEM S_OSC
  34
  35 #include "../../include/linux/libcfs/libcfs.h"
  36
  37 #include "../include/lustre_dlm.h"
  38 #include "../include/lustre_net.h"
  39 #include "../include/lustre/lustre_user.h"
  40 #include "../include/obd_cksum.h"
  41
  42 #include "../include/lustre_ha.h"
  43 #include "../include/lprocfs_status.h"
  44 #include "../include/lustre_debug.h"
  45 #include "../include/lustre_param.h"
  46 #include "../include/lustre_fid.h"
  47 #include "../include/obd_class.h"
  48 #include "../include/obd.h"
  49 #include "osc_internal.h"
  50 #include "osc_cl_internal.h"
  51
  52 atomic_t osc_pool_req_count;
  53 unsigned int osc_reqpool_maxreqcount;
  54 struct ptlrpc_request_pool *osc_rq_pool;
  55
  56 /* max memory used for request pool, unit is MB */
  57 static unsigned int osc_reqpool_mem_max = 5;
  58 module_param(osc_reqpool_mem_max, uint, 0444);
  59
  60 struct osc_brw_async_args {
  61         struct obdo       *aa_oa;
  62         int             aa_requested_nob;
  63         int             aa_nio_count;
  64         u32             aa_page_count;
  65         int             aa_resends;
  66         struct brw_page  **aa_ppga;
  67         struct client_obd *aa_cli;
  68         struct list_head         aa_oaps;
  69         struct list_head         aa_exts;
  70         struct cl_req     *aa_clerq;
  71 };
  72
  73 struct osc_async_args {
  74         struct obd_info   *aa_oi;
  75 };
  76
  77 struct osc_setattr_args {
  78         struct obdo      *sa_oa;
  79         obd_enqueue_update_f sa_upcall;
  80         void            *sa_cookie;
  81 };
  82
  83 struct osc_fsync_args {
  84         struct obd_info     *fa_oi;
  85         obd_enqueue_update_f fa_upcall;
  86         void            *fa_cookie;
  87 };
  88
  89 struct osc_enqueue_args {
  90         struct obd_export       *oa_exp;
  91         enum ldlm_type          oa_type;
  92         enum ldlm_mode          oa_mode;
  93         __u64               *oa_flags;
  94         osc_enqueue_upcall_f    oa_upcall;
  95         void                 *oa_cookie;
  96         struct ost_lvb     *oa_lvb;
  97         struct lustre_handle    oa_lockh;
  98         unsigned int          oa_agl:1;
  99 };
 100
 101 static void osc_release_ppga(struct brw_page **ppga, u32 count);
 102 static int brw_interpret(const struct lu_env *env,
 103                          struct ptlrpc_request *req, void *data, int rc);
 104
 105 /* Pack OSC object metadata for disk storage (LE byte order). */
 106 static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
 107                       struct lov_stripe_md *lsm)
 108 {
 109         int lmm_size;
 110
 111         lmm_size = sizeof(**lmmp);
 112         if (!lmmp)
 113                 return lmm_size;
 114
 115         if (*lmmp && !lsm) {
 116                 kfree(*lmmp);
 117                 *lmmp = NULL;
 118                 return 0;
 119         } else if (unlikely(lsm && ostid_id(&lsm->lsm_oi) == 0)) {
 120                 return -EBADF;
 121         }
 122
 123         if (!*lmmp) {
 124                 *lmmp = kzalloc(lmm_size, GFP_NOFS);
 125                 if (!*lmmp)
 126                         return -ENOMEM;
 127         }
 128
 129         if (lsm)
 130                 ostid_cpu_to_le(&lsm->lsm_oi, &(*lmmp)->lmm_oi);
 131
 132         return lmm_size;
 133 }
 134
 135 /* Unpack OSC object metadata from disk storage (LE byte order). */
 136 static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
 137                         struct lov_mds_md *lmm, int lmm_bytes)
 138 {
 139         int lsm_size;
 140         struct obd_import *imp = class_exp2cliimp(exp);
 141
 142         if (lmm) {
 143                 if (lmm_bytes < sizeof(*lmm)) {
 144                         CERROR("%s: lov_mds_md too small: %d, need %d\n",
 145                                exp->exp_obd->obd_name, lmm_bytes,
 146                                (int)sizeof(*lmm));
 147                         return -EINVAL;
 148                 }
 149                 /* XXX LOV_MAGIC etc check? */
 150
 151                 if (unlikely(ostid_id(&lmm->lmm_oi) == 0)) {
 152                         CERROR("%s: zero lmm_object_id: rc = %d\n",
 153                                exp->exp_obd->obd_name, -EINVAL);
 154                         return -EINVAL;
 155                 }
 156         }
 157
 158         lsm_size = lov_stripe_md_size(1);
 159         if (!lsmp)
 160                 return lsm_size;
 161
 162         if (*lsmp && !lmm) {
 163                 kfree((*lsmp)->lsm_oinfo[0]);
 164                 kfree(*lsmp);
 165                 *lsmp = NULL;
 166                 return 0;
 167         }
 168
 169         if (!*lsmp) {
 170                 *lsmp = kzalloc(lsm_size, GFP_NOFS);
 171                 if (unlikely(!*lsmp))
 172                         return -ENOMEM;
 173                 (*lsmp)->lsm_oinfo[0] = kzalloc(sizeof(struct lov_oinfo),
 174                                                 GFP_NOFS);
 175                 if (unlikely(!(*lsmp)->lsm_oinfo[0])) {
 176                         kfree(*lsmp);
 177                         return -ENOMEM;
 178                 }
 179                 loi_init((*lsmp)->lsm_oinfo[0]);
 180         } else if (unlikely(ostid_id(&(*lsmp)->lsm_oi) == 0)) {
 181                 return -EBADF;
 182         }
 183
 184         if (lmm)
 185                 /* XXX zero *lsmp? */
 186                 ostid_le_to_cpu(&lmm->lmm_oi, &(*lsmp)->lsm_oi);
 187
 188         if (imp &&
 189             (imp->imp_connect_data.ocd_connect_flags & OBD_CONNECT_MAXBYTES))
 190                 (*lsmp)->lsm_maxbytes = imp->imp_connect_data.ocd_maxbytes;
 191         else
 192                 (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;
 193
 194         return lsm_size;
 195 }
 196
 197 static inline void osc_pack_req_body(struct ptlrpc_request *req,
 198                                      struct obd_info *oinfo)
 199 {
 200         struct ost_body *body;
 201
 202         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
 203         LASSERT(body);
 204
 205         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa,
 206                              oinfo->oi_oa);
 207 }
 208
 209 static int osc_getattr_interpret(const struct lu_env *env,
 210                                  struct ptlrpc_request *req,
 211                                  struct osc_async_args *aa, int rc)
 212 {
 213         struct ost_body *body;
 214
 215         if (rc != 0)
 216                 goto out;
 217
 218         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 219         if (body) {
 220                 CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
 221                 lustre_get_wire_obdo(&req->rq_import->imp_connect_data,
 222                                      aa->aa_oi->oi_oa, &body->oa);
 223
 224                 /* This should really be sent by the OST */
 225                 aa->aa_oi->oi_oa->o_blksize = DT_MAX_BRW_SIZE;
 226                 aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
 227         } else {
 228                 CDEBUG(D_INFO, "can't unpack ost_body\n");
 229                 rc = -EPROTO;
 230                 aa->aa_oi->oi_oa->o_valid = 0;
 231         }
 232 out:
 233         rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
 234         return rc;
 235 }
 236
 237 static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
 238                              struct ptlrpc_request_set *set)
 239 {
 240         struct ptlrpc_request *req;
 241         struct osc_async_args *aa;
 242         int rc;
 243
 244         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
 245         if (!req)
 246                 return -ENOMEM;
 247
 248         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
 249         if (rc) {
 250                 ptlrpc_request_free(req);
 251                 return rc;
 252         }
 253
 254         osc_pack_req_body(req, oinfo);
 255
 256         ptlrpc_request_set_replen(req);
 257         req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_getattr_interpret;
 258
 259         CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
 260         aa = ptlrpc_req_async_args(req);
 261         aa->aa_oi = oinfo;
 262
 263         ptlrpc_set_add_req(set, req);
 264         return 0;
 265 }
 266
 267 static int osc_getattr(const struct lu_env *env, struct obd_export *exp,
 268                        struct obd_info *oinfo)
 269 {
 270         struct ptlrpc_request *req;
 271         struct ost_body *body;
 272         int rc;
 273
 274         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_GETATTR);
 275         if (!req)
 276                 return -ENOMEM;
 277
 278         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GETATTR);
 279         if (rc) {
 280                 ptlrpc_request_free(req);
 281                 return rc;
 282         }
 283
 284         osc_pack_req_body(req, oinfo);
 285
 286         ptlrpc_request_set_replen(req);
 287
 288         rc = ptlrpc_queue_wait(req);
 289         if (rc)
 290                 goto out;
 291
 292         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 293         if (!body) {
 294                 rc = -EPROTO;
 295                 goto out;
 296         }
 297
 298         CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
 299         lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oinfo->oi_oa,
 300                              &body->oa);
 301
 302         oinfo->oi_oa->o_blksize = cli_brw_size(exp->exp_obd);
 303         oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
 304
 305  out:
 306         ptlrpc_req_finished(req);
 307         return rc;
 308 }
 309
 310 static int osc_setattr(const struct lu_env *env, struct obd_export *exp,
 311                        struct obd_info *oinfo, struct obd_trans_info *oti)
 312 {
 313         struct ptlrpc_request *req;
 314         struct ost_body *body;
 315         int rc;
 316
 317         LASSERT(oinfo->oi_oa->o_valid & OBD_MD_FLGROUP);
 318
 319         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
 320         if (!req)
 321                 return -ENOMEM;
 322
 323         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
 324         if (rc) {
 325                 ptlrpc_request_free(req);
 326                 return rc;
 327         }
 328
 329         osc_pack_req_body(req, oinfo);
 330
 331         ptlrpc_request_set_replen(req);
 332
 333         rc = ptlrpc_queue_wait(req);
 334         if (rc)
 335                 goto out;
 336
 337         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 338         if (!body) {
 339                 rc = -EPROTO;
 340                 goto out;
 341         }
 342
 343         lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oinfo->oi_oa,
 344                              &body->oa);
 345
 346 out:
 347         ptlrpc_req_finished(req);
 348         return rc;
 349 }
 350
 351 static int osc_setattr_interpret(const struct lu_env *env,
 352                                  struct ptlrpc_request *req,
 353                                  struct osc_setattr_args *sa, int rc)
 354 {
 355         struct ost_body *body;
 356
 357         if (rc != 0)
 358                 goto out;
 359
 360         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 361         if (!body) {
 362                 rc = -EPROTO;
 363                 goto out;
 364         }
 365
 366         lustre_get_wire_obdo(&req->rq_import->imp_connect_data, sa->sa_oa,
 367                              &body->oa);
 368 out:
 369         rc = sa->sa_upcall(sa->sa_cookie, rc);
 370         return rc;
 371 }
 372
 373 int osc_setattr_async_base(struct obd_export *exp, struct obd_info *oinfo,
 374                            struct obd_trans_info *oti,
 375                            obd_enqueue_update_f upcall, void *cookie,
 376                            struct ptlrpc_request_set *rqset)
 377 {
 378         struct ptlrpc_request *req;
 379         struct osc_setattr_args *sa;
 380         int rc;
 381
 382         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SETATTR);
 383         if (!req)
 384                 return -ENOMEM;
 385
 386         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SETATTR);
 387         if (rc) {
 388                 ptlrpc_request_free(req);
 389                 return rc;
 390         }
 391
 392         if (oti && oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE)
 393                 oinfo->oi_oa->o_lcookie = *oti->oti_logcookies;
 394
 395         osc_pack_req_body(req, oinfo);
 396
 397         ptlrpc_request_set_replen(req);
 398
 399         /* do mds to ost setattr asynchronously */
 400         if (!rqset) {
 401                 /* Do not wait for response. */
 402                 ptlrpcd_add_req(req);
 403         } else {
 404                 req->rq_interpret_reply =
 405                         (ptlrpc_interpterer_t)osc_setattr_interpret;
 406
 407                 CLASSERT(sizeof(*sa) <= sizeof(req->rq_async_args));
 408                 sa = ptlrpc_req_async_args(req);
 409                 sa->sa_oa = oinfo->oi_oa;
 410                 sa->sa_upcall = upcall;
 411                 sa->sa_cookie = cookie;
 412
 413                 if (rqset == PTLRPCD_SET)
 414                         ptlrpcd_add_req(req);
 415                 else
 416                         ptlrpc_set_add_req(rqset, req);
 417         }
 418
 419         return 0;
 420 }
 421
 422 static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
 423                              struct obd_trans_info *oti,
 424                              struct ptlrpc_request_set *rqset)
 425 {
 426         return osc_setattr_async_base(exp, oinfo, oti,
 427                                       oinfo->oi_cb_up, oinfo, rqset);
 428 }
 429
 430 static int osc_real_create(struct obd_export *exp, struct obdo *oa,
 431                            struct lov_stripe_md **ea,
 432                            struct obd_trans_info *oti)
 433 {
 434         struct ptlrpc_request *req;
 435         struct ost_body *body;
 436         struct lov_stripe_md *lsm;
 437         int rc;
 438
 439         LASSERT(oa);
 440         LASSERT(ea);
 441
 442         lsm = *ea;
 443         if (!lsm) {
 444                 rc = obd_alloc_memmd(exp, &lsm);
 445                 if (rc < 0)
 446                         return rc;
 447         }
 448
 449         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_CREATE);
 450         if (!req) {
 451                 rc = -ENOMEM;
 452                 goto out;
 453         }
 454
 455         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_CREATE);
 456         if (rc) {
 457                 ptlrpc_request_free(req);
 458                 goto out;
 459         }
 460
 461         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
 462         LASSERT(body);
 463
 464         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
 465
 466         ptlrpc_request_set_replen(req);
 467
 468         if ((oa->o_valid & OBD_MD_FLFLAGS) &&
 469             oa->o_flags == OBD_FL_DELORPHAN) {
 470                 DEBUG_REQ(D_HA, req,
 471                           "delorphan from OST integration");
 472                 /* Don't resend the delorphan req */
 473                 req->rq_no_resend = 1;
 474                 req->rq_no_delay = 1;
 475         }
 476
 477         rc = ptlrpc_queue_wait(req);
 478         if (rc)
 479                 goto out_req;
 480
 481         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 482         if (!body) {
 483                 rc = -EPROTO;
 484                 goto out_req;
 485         }
 486
 487         CDEBUG(D_INFO, "oa flags %x\n", oa->o_flags);
 488         lustre_get_wire_obdo(&req->rq_import->imp_connect_data, oa, &body->oa);
 489
 490         oa->o_blksize = cli_brw_size(exp->exp_obd);
 491         oa->o_valid |= OBD_MD_FLBLKSZ;
 492
 493         /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
 494          * have valid lsm_oinfo data structs, so don't go touching that.
 495          * This needs to be fixed in a big way.
 496          */
 497         lsm->lsm_oi = oa->o_oi;
 498         *ea = lsm;
 499
 500         if (oti) {
 501                 oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);
 502
 503                 if (oa->o_valid & OBD_MD_FLCOOKIE) {
 504                         if (!oti->oti_logcookies)
 505                                 oti_alloc_cookies(oti, 1);
 506                         *oti->oti_logcookies = oa->o_lcookie;
 507                 }
 508         }
 509
 510         CDEBUG(D_HA, "transno: %lld\n",
 511                lustre_msg_get_transno(req->rq_repmsg));
 512 out_req:
 513         ptlrpc_req_finished(req);
 514 out:
 515         if (rc && !*ea)
 516                 obd_free_memmd(exp, &lsm);
 517         return rc;
 518 }
 519
 520 int osc_punch_base(struct obd_export *exp, struct obd_info *oinfo,
 521                    obd_enqueue_update_f upcall, void *cookie,
 522                    struct ptlrpc_request_set *rqset)
 523 {
 524         struct ptlrpc_request *req;
 525         struct osc_setattr_args *sa;
 526         struct ost_body *body;
 527         int rc;
 528
 529         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_PUNCH);
 530         if (!req)
 531                 return -ENOMEM;
 532
 533         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_PUNCH);
 534         if (rc) {
 535                 ptlrpc_request_free(req);
 536                 return rc;
 537         }
 538         req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
 539         ptlrpc_at_set_req_timeout(req);
 540
 541         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
 542         LASSERT(body);
 543         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa,
 544                              oinfo->oi_oa);
 545
 546         ptlrpc_request_set_replen(req);
 547
 548         req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_setattr_interpret;
 549         CLASSERT(sizeof(*sa) <= sizeof(req->rq_async_args));
 550         sa = ptlrpc_req_async_args(req);
 551         sa->sa_oa = oinfo->oi_oa;
 552         sa->sa_upcall = upcall;
 553         sa->sa_cookie = cookie;
 554         if (rqset == PTLRPCD_SET)
 555                 ptlrpcd_add_req(req);
 556         else
 557                 ptlrpc_set_add_req(rqset, req);
 558
 559         return 0;
 560 }
 561
 562 static int osc_sync_interpret(const struct lu_env *env,
 563                               struct ptlrpc_request *req,
 564                               void *arg, int rc)
 565 {
 566         struct osc_fsync_args *fa = arg;
 567         struct ost_body *body;
 568
 569         if (rc)
 570                 goto out;
 571
 572         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 573         if (!body) {
 574                 CERROR("can't unpack ost_body\n");
 575                 rc = -EPROTO;
 576                 goto out;
 577         }
 578
 579         *fa->fa_oi->oi_oa = body->oa;
 580 out:
 581         rc = fa->fa_upcall(fa->fa_cookie, rc);
 582         return rc;
 583 }
 584
 585 int osc_sync_base(struct obd_export *exp, struct obd_info *oinfo,
 586                   obd_enqueue_update_f upcall, void *cookie,
 587                   struct ptlrpc_request_set *rqset)
 588 {
 589         struct ptlrpc_request *req;
 590         struct ost_body *body;
 591         struct osc_fsync_args *fa;
 592         int rc;
 593
 594         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_SYNC);
 595         if (!req)
 596                 return -ENOMEM;
 597
 598         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SYNC);
 599         if (rc) {
 600                 ptlrpc_request_free(req);
 601                 return rc;
 602         }
 603
 604         /* overload the size and blocks fields in the oa with start/end */
 605         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
 606         LASSERT(body);
 607         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa,
 608                              oinfo->oi_oa);
 609
 610         ptlrpc_request_set_replen(req);
 611         req->rq_interpret_reply = osc_sync_interpret;
 612
 613         CLASSERT(sizeof(*fa) <= sizeof(req->rq_async_args));
 614         fa = ptlrpc_req_async_args(req);
 615         fa->fa_oi = oinfo;
 616         fa->fa_upcall = upcall;
 617         fa->fa_cookie = cookie;
 618
 619         if (rqset == PTLRPCD_SET)
 620                 ptlrpcd_add_req(req);
 621         else
 622                 ptlrpc_set_add_req(rqset, req);
 623
 624         return 0;
 625 }
 626
 627 /* Find and cancel locally locks matched by @mode in the resource found by
 628  * @objid. Found locks are added into @cancel list. Returns the amount of
 629  * locks added to @cancels list.
 630  */
 631 static int osc_resource_get_unused(struct obd_export *exp, struct obdo *oa,
 632                                    struct list_head *cancels,
 633                                    enum ldlm_mode mode, __u64 lock_flags)
 634 {
 635         struct ldlm_namespace *ns = exp->exp_obd->obd_namespace;
 636         struct ldlm_res_id res_id;
 637         struct ldlm_resource *res;
 638         int count;
 639
 640         /* Return, i.e. cancel nothing, only if ELC is supported (flag in
 641          * export) but disabled through procfs (flag in NS).
 642          *
 643          * This distinguishes from a case when ELC is not supported originally,
 644          * when we still want to cancel locks in advance and just cancel them
 645          * locally, without sending any RPC.
 646          */
 647         if (exp_connect_cancelset(exp) && !ns_connect_cancelset(ns))
 648                 return 0;
 649
 650         ostid_build_res_name(&oa->o_oi, &res_id);
 651         res = ldlm_resource_get(ns, NULL, &res_id, 0, 0);
 652         if (!res)
 653                 return 0;
 654
 655         LDLM_RESOURCE_ADDREF(res);
 656         count = ldlm_cancel_resource_local(res, cancels, NULL, mode,
 657                                            lock_flags, 0, NULL);
 658         LDLM_RESOURCE_DELREF(res);
 659         ldlm_resource_putref(res);
 660         return count;
 661 }
 662
 663 static int osc_destroy_interpret(const struct lu_env *env,
 664                                  struct ptlrpc_request *req, void *data,
 665                                  int rc)
 666 {
 667         struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
 668
 669         atomic_dec(&cli->cl_destroy_in_flight);
 670         wake_up(&cli->cl_destroy_waitq);
 671         return 0;
 672 }
 673
 674 static int osc_can_send_destroy(struct client_obd *cli)
 675 {
 676         if (atomic_inc_return(&cli->cl_destroy_in_flight) <=
 677             cli->cl_max_rpcs_in_flight) {
 678                 /* The destroy request can be sent */
 679                 return 1;
 680         }
 681         if (atomic_dec_return(&cli->cl_destroy_in_flight) <
 682             cli->cl_max_rpcs_in_flight) {
 683                 /*
 684                  * The counter has been modified between the two atomic
 685                  * operations.
 686                  */
 687                 wake_up(&cli->cl_destroy_waitq);
 688         }
 689         return 0;
 690 }
 691
 692 static int osc_create(const struct lu_env *env, struct obd_export *exp,
 693                       struct obdo *oa, struct lov_stripe_md **ea,
 694                       struct obd_trans_info *oti)
 695 {
 696         int rc = 0;
 697
 698         LASSERT(oa);
 699         LASSERT(ea);
 700         LASSERT(oa->o_valid & OBD_MD_FLGROUP);
 701
 702         if ((oa->o_valid & OBD_MD_FLFLAGS) &&
 703             oa->o_flags == OBD_FL_RECREATE_OBJS) {
 704                 return osc_real_create(exp, oa, ea, oti);
 705         }
 706
 707         if (!fid_seq_is_mdt(ostid_seq(&oa->o_oi)))
 708                 return osc_real_create(exp, oa, ea, oti);
 709
 710         /* we should not get here anymore */
 711         LBUG();
 712
 713         return rc;
 714 }
 715
 716 /* Destroy requests can be async always on the client, and we don't even really
 717  * care about the return code since the client cannot do anything at all about
 718  * a destroy failure.
 719  * When the MDS is unlinking a filename, it saves the file objects into a
 720  * recovery llog, and these object records are cancelled when the OST reports
 721  * they were destroyed and sync'd to disk (i.e. transaction committed).
 722  * If the client dies, or the OST is down when the object should be destroyed,
 723  * the records are not cancelled, and when the OST reconnects to the MDS next,
 724  * it will retrieve the llog unlink logs and then sends the log cancellation
 725  * cookies to the MDS after committing destroy transactions.
 726  */
 727 static int osc_destroy(const struct lu_env *env, struct obd_export *exp,
 728                        struct obdo *oa, struct lov_stripe_md *ea,
 729                        struct obd_trans_info *oti, struct obd_export *md_export)
 730 {
 731         struct client_obd *cli = &exp->exp_obd->u.cli;
 732         struct ptlrpc_request *req;
 733         struct ost_body *body;
 734         LIST_HEAD(cancels);
 735         int rc, count;
 736
 737         if (!oa) {
 738                 CDEBUG(D_INFO, "oa NULL\n");
 739                 return -EINVAL;
 740         }
 741
 742         count = osc_resource_get_unused(exp, oa, &cancels, LCK_PW,
 743                                         LDLM_FL_DISCARD_DATA);
 744
 745         req = ptlrpc_request_alloc(class_exp2cliimp(exp), &RQF_OST_DESTROY);
 746         if (!req) {
 747                 ldlm_lock_list_put(&cancels, l_bl_ast, count);
 748                 return -ENOMEM;
 749         }
 750
 751         rc = ldlm_prep_elc_req(exp, req, LUSTRE_OST_VERSION, OST_DESTROY,
 752                                0, &cancels, count);
 753         if (rc) {
 754                 ptlrpc_request_free(req);
 755                 return rc;
 756         }
 757
 758         req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
 759         ptlrpc_at_set_req_timeout(req);
 760
 761         if (oti && oa->o_valid & OBD_MD_FLCOOKIE)
 762                 oa->o_lcookie = *oti->oti_logcookies;
 763         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
 764         LASSERT(body);
 765         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
 766
 767         ptlrpc_request_set_replen(req);
 768
 769         /* If osc_destroy is for destroying the unlink orphan,
 770          * sent from MDT to OST, which should not be blocked here,
 771          * because the process might be triggered by ptlrpcd, and
 772          * it is not good to block ptlrpcd thread (b=16006
 773          **/
 774         if (!(oa->o_flags & OBD_FL_DELORPHAN)) {
 775                 req->rq_interpret_reply = osc_destroy_interpret;
 776                 if (!osc_can_send_destroy(cli)) {
 777                         struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP,
 778                                                           NULL);
 779
 780                         /*
 781                          * Wait until the number of on-going destroy RPCs drops
 782                          * under max_rpc_in_flight
 783                          */
 784                         l_wait_event_exclusive(cli->cl_destroy_waitq,
 785                                                osc_can_send_destroy(cli), &lwi);
 786                 }
 787         }
 788
 789         /* Do not wait for response */
 790         ptlrpcd_add_req(req);
 791         return 0;
 792 }
 793
 794 static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
 795                                 long writing_bytes)
 796 {
 797         u32 bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;
 798
 799         LASSERT(!(oa->o_valid & bits));
 800
 801         oa->o_valid |= bits;
 802         spin_lock(&cli->cl_loi_list_lock);
 803         oa->o_dirty = cli->cl_dirty;
 804         if (unlikely(cli->cl_dirty - cli->cl_dirty_transit >
 805                      cli->cl_dirty_max)) {
 806                 CERROR("dirty %lu - %lu > dirty_max %lu\n",
 807                        cli->cl_dirty, cli->cl_dirty_transit, cli->cl_dirty_max);
 808                 oa->o_undirty = 0;
 809         } else if (unlikely(atomic_read(&obd_unstable_pages) +
 810                             atomic_read(&obd_dirty_pages) -
 811                             atomic_read(&obd_dirty_transit_pages) >
 812                             (long)(obd_max_dirty_pages + 1))) {
 813                 /* The atomic_read() allowing the atomic_inc() are
 814                  * not covered by a lock thus they may safely race and trip
 815                  * this CERROR() unless we add in a small fudge factor (+1).
 816                  */
 817                 CERROR("%s: dirty %d + %d - %d > system dirty_max %d\n",
 818                        cli->cl_import->imp_obd->obd_name,
 819                        atomic_read(&obd_unstable_pages),
 820                        atomic_read(&obd_dirty_pages),
 821                        atomic_read(&obd_dirty_transit_pages),
 822                        obd_max_dirty_pages);
 823                 oa->o_undirty = 0;
 824         } else if (unlikely(cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff)) {
 825                 CERROR("dirty %lu - dirty_max %lu too big???\n",
 826                        cli->cl_dirty, cli->cl_dirty_max);
 827                 oa->o_undirty = 0;
 828         } else {
 829                 long max_in_flight = (cli->cl_max_pages_per_rpc <<
 830                                       PAGE_SHIFT)*
 831                                      (cli->cl_max_rpcs_in_flight + 1);
 832                 oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
 833         }
 834         oa->o_grant = cli->cl_avail_grant + cli->cl_reserved_grant;
 835         oa->o_dropped = cli->cl_lost_grant;
 836         cli->cl_lost_grant = 0;
 837         spin_unlock(&cli->cl_loi_list_lock);
 838         CDEBUG(D_CACHE, "dirty: %llu undirty: %u dropped %u grant: %llu\n",
 839                oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
 840 }
 841
 842 void osc_update_next_shrink(struct client_obd *cli)
 843 {
 844         cli->cl_next_shrink_grant =
 845                 cfs_time_shift(cli->cl_grant_shrink_interval);
 846         CDEBUG(D_CACHE, "next time %ld to shrink grant\n",
 847                cli->cl_next_shrink_grant);
 848 }
 849
 850 static void __osc_update_grant(struct client_obd *cli, u64 grant)
 851 {
 852         spin_lock(&cli->cl_loi_list_lock);
 853         cli->cl_avail_grant += grant;
 854         spin_unlock(&cli->cl_loi_list_lock);
 855 }
 856
 857 static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
 858 {
 859         if (body->oa.o_valid & OBD_MD_FLGRANT) {
 860                 CDEBUG(D_CACHE, "got %llu extra grant\n", body->oa.o_grant);
 861                 __osc_update_grant(cli, body->oa.o_grant);
 862         }
 863 }
 864
 865 static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
 866                               u32 keylen, void *key, u32 vallen,
 867                               void *val, struct ptlrpc_request_set *set);
 868
 869 static int osc_shrink_grant_interpret(const struct lu_env *env,
 870                                       struct ptlrpc_request *req,
 871                                       void *aa, int rc)
 872 {
 873         struct client_obd *cli = &req->rq_import->imp_obd->u.cli;
 874         struct obdo *oa = ((struct osc_brw_async_args *)aa)->aa_oa;
 875         struct ost_body *body;
 876
 877         if (rc != 0) {
 878                 __osc_update_grant(cli, oa->o_grant);
 879                 goto out;
 880         }
 881
 882         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
 883         LASSERT(body);
 884         osc_update_grant(cli, body);
 885 out:
 886         kmem_cache_free(obdo_cachep, oa);
 887         return rc;
 888 }
 889
 890 static void osc_shrink_grant_local(struct client_obd *cli, struct obdo *oa)
 891 {
 892         spin_lock(&cli->cl_loi_list_lock);
 893         oa->o_grant = cli->cl_avail_grant / 4;
 894         cli->cl_avail_grant -= oa->o_grant;
 895         spin_unlock(&cli->cl_loi_list_lock);
 896         if (!(oa->o_valid & OBD_MD_FLFLAGS)) {
 897                 oa->o_valid |= OBD_MD_FLFLAGS;
 898                 oa->o_flags = 0;
 899         }
 900         oa->o_flags |= OBD_FL_SHRINK_GRANT;
 901         osc_update_next_shrink(cli);
 902 }
 903
 904 /* Shrink the current grant, either from some large amount to enough for a
 905  * full set of in-flight RPCs, or if we have already shrunk to that limit
 906  * then to enough for a single RPC.  This avoids keeping more grant than
 907  * needed, and avoids shrinking the grant piecemeal.
 908  */
 909 static int osc_shrink_grant(struct client_obd *cli)
 910 {
 911         __u64 target_bytes = (cli->cl_max_rpcs_in_flight + 1) *
 912                              (cli->cl_max_pages_per_rpc << PAGE_SHIFT);
 913
 914         spin_lock(&cli->cl_loi_list_lock);
 915         if (cli->cl_avail_grant <= target_bytes)
 916                 target_bytes = cli->cl_max_pages_per_rpc << PAGE_SHIFT;
 917         spin_unlock(&cli->cl_loi_list_lock);
 918
 919         return osc_shrink_grant_to_target(cli, target_bytes);
 920 }
 921
 922 int osc_shrink_grant_to_target(struct client_obd *cli, __u64 target_bytes)
 923 {
 924         int rc = 0;
 925         struct ost_body *body;
 926
 927         spin_lock(&cli->cl_loi_list_lock);
 928         /* Don't shrink if we are already above or below the desired limit
 929          * We don't want to shrink below a single RPC, as that will negatively
 930          * impact block allocation and long-term performance.
 931          */
 932         if (target_bytes < cli->cl_max_pages_per_rpc << PAGE_SHIFT)
 933                 target_bytes = cli->cl_max_pages_per_rpc << PAGE_SHIFT;
 934
 935         if (target_bytes >= cli->cl_avail_grant) {
 936                 spin_unlock(&cli->cl_loi_list_lock);
 937                 return 0;
 938         }
 939         spin_unlock(&cli->cl_loi_list_lock);
 940
 941         body = kzalloc(sizeof(*body), GFP_NOFS);
 942         if (!body)
 943                 return -ENOMEM;
 944
 945         osc_announce_cached(cli, &body->oa, 0);
 946
 947         spin_lock(&cli->cl_loi_list_lock);
 948         body->oa.o_grant = cli->cl_avail_grant - target_bytes;
 949         cli->cl_avail_grant = target_bytes;
 950         spin_unlock(&cli->cl_loi_list_lock);
 951         if (!(body->oa.o_valid & OBD_MD_FLFLAGS)) {
 952                 body->oa.o_valid |= OBD_MD_FLFLAGS;
 953                 body->oa.o_flags = 0;
 954         }
 955         body->oa.o_flags |= OBD_FL_SHRINK_GRANT;
 956         osc_update_next_shrink(cli);
 957
 958         rc = osc_set_info_async(NULL, cli->cl_import->imp_obd->obd_self_export,
 959                                 sizeof(KEY_GRANT_SHRINK), KEY_GRANT_SHRINK,
 960                                 sizeof(*body), body, NULL);
 961         if (rc != 0)
 962                 __osc_update_grant(cli, body->oa.o_grant);
 963         kfree(body);
 964         return rc;
 965 }
 966
 967 static int osc_should_shrink_grant(struct client_obd *client)
 968 {
 969         unsigned long time = cfs_time_current();
 970         unsigned long next_shrink = client->cl_next_shrink_grant;
 971
 972         if ((client->cl_import->imp_connect_data.ocd_connect_flags &
 973              OBD_CONNECT_GRANT_SHRINK) == 0)
 974                 return 0;
 975
 976         if (cfs_time_aftereq(time, next_shrink - 5 * CFS_TICK)) {
 977                 /* Get the current RPC size directly, instead of going via:
 978                  * cli_brw_size(obd->u.cli.cl_import->imp_obd->obd_self_export)
 979                  * Keep comment here so that it can be found by searching.
 980                  */
 981                 int brw_size = client->cl_max_pages_per_rpc << PAGE_SHIFT;
 982
 983                 if (client->cl_import->imp_state == LUSTRE_IMP_FULL &&
 984                     client->cl_avail_grant > brw_size)
 985                         return 1;
 986
 987                 osc_update_next_shrink(client);
 988         }
 989         return 0;
 990 }
 991
 992 static int osc_grant_shrink_grant_cb(struct timeout_item *item, void *data)
 993 {
 994         struct client_obd *client;
 995
 996         list_for_each_entry(client, &item->ti_obd_list, cl_grant_shrink_list) {
 997                 if (osc_should_shrink_grant(client))
 998                         osc_shrink_grant(client);
 999         }
1000         return 0;
1001 }
1002
1003 static int osc_add_shrink_grant(struct client_obd *client)
1004 {
1005         int rc;
1006
1007         rc = ptlrpc_add_timeout_client(client->cl_grant_shrink_interval,
1008                                        TIMEOUT_GRANT,
1009                                        osc_grant_shrink_grant_cb, NULL,
1010                                        &client->cl_grant_shrink_list);
1011         if (rc) {
1012                 CERROR("add grant client %s error %d\n",
1013                        client->cl_import->imp_obd->obd_name, rc);
1014                 return rc;
1015         }
1016         CDEBUG(D_CACHE, "add grant client %s\n",
1017                client->cl_import->imp_obd->obd_name);
1018         osc_update_next_shrink(client);
1019         return 0;
1020 }
1021
1022 static int osc_del_shrink_grant(struct client_obd *client)
1023 {
1024         return ptlrpc_del_timeout_client(&client->cl_grant_shrink_list,
1025                                          TIMEOUT_GRANT);
1026 }
1027
1028 static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
1029 {
1030         /*
1031          * ocd_grant is the total grant amount we're expect to hold: if we've
1032          * been evicted, it's the new avail_grant amount, cl_dirty will drop
1033          * to 0 as inflight RPCs fail out; otherwise, it's avail_grant + dirty.
1034          *
1035          * race is tolerable here: if we're evicted, but imp_state already
1036          * left EVICTED state, then cl_dirty must be 0 already.
1037          */
1038         spin_lock(&cli->cl_loi_list_lock);
1039         if (cli->cl_import->imp_state == LUSTRE_IMP_EVICTED)
1040                 cli->cl_avail_grant = ocd->ocd_grant;
1041         else
1042                 cli->cl_avail_grant = ocd->ocd_grant - cli->cl_dirty;
1043
1044         if (cli->cl_avail_grant < 0) {
1045                 CWARN("%s: available grant < 0: avail/ocd/dirty %ld/%u/%ld\n",
1046                       cli->cl_import->imp_obd->obd_name, cli->cl_avail_grant,
1047                       ocd->ocd_grant, cli->cl_dirty);
1048                 /* workaround for servers which do not have the patch from
1049                  * LU-2679
1050                  */
1051                 cli->cl_avail_grant = ocd->ocd_grant;
1052         }
1053
1054         /* determine the appropriate chunk size used by osc_extent. */
1055         cli->cl_chunkbits = max_t(int, PAGE_SHIFT, ocd->ocd_blocksize);
1056         spin_unlock(&cli->cl_loi_list_lock);
1057
1058         CDEBUG(D_CACHE, "%s, setting cl_avail_grant: %ld cl_lost_grant: %ld chunk bits: %d\n",
1059                cli->cl_import->imp_obd->obd_name,
1060                cli->cl_avail_grant, cli->cl_lost_grant, cli->cl_chunkbits);
1061
1062         if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT_SHRINK &&
1063             list_empty(&cli->cl_grant_shrink_list))
1064                 osc_add_shrink_grant(cli);
1065 }
1066
1067 /* We assume that the reason this OSC got a short read is because it read
1068  * beyond the end of a stripe file; i.e. lustre is reading a sparse file
1069  * via the LOV, and it _knows_ it's reading inside the file, it's just that
1070  * this stripe never got written at or beyond this stripe offset yet.
1071  */
1072 static void handle_short_read(int nob_read, u32 page_count,
1073                               struct brw_page **pga)
1074 {
1075         char *ptr;
1076         int i = 0;
1077
1078         /* skip bytes read OK */
1079         while (nob_read > 0) {
1080                 LASSERT(page_count > 0);
1081
1082                 if (pga[i]->count > nob_read) {
1083                         /* EOF inside this page */
1084                         ptr = kmap(pga[i]->pg) +
1085                                 (pga[i]->off & ~PAGE_MASK);
1086                         memset(ptr + nob_read, 0, pga[i]->count - nob_read);
1087                         kunmap(pga[i]->pg);
1088                         page_count--;
1089                         i++;
1090                         break;
1091                 }
1092
1093                 nob_read -= pga[i]->count;
1094                 page_count--;
1095                 i++;
1096         }
1097
1098         /* zero remaining pages */
1099         while (page_count-- > 0) {
1100                 ptr = kmap(pga[i]->pg) + (pga[i]->off & ~PAGE_MASK);
1101                 memset(ptr, 0, pga[i]->count);
1102                 kunmap(pga[i]->pg);
1103                 i++;
1104         }
1105 }
1106
1107 static int check_write_rcs(struct ptlrpc_request *req,
1108                            int requested_nob, int niocount,
1109                            u32 page_count, struct brw_page **pga)
1110 {
1111         int i;
1112         __u32 *remote_rcs;
1113
1114         remote_rcs = req_capsule_server_sized_get(&req->rq_pill, &RMF_RCS,
1115                                                   sizeof(*remote_rcs) *
1116                                                   niocount);
1117         if (!remote_rcs) {
1118                 CDEBUG(D_INFO, "Missing/short RC vector on BRW_WRITE reply\n");
1119                 return -EPROTO;
1120         }
1121
1122         /* return error if any niobuf was in error */
1123         for (i = 0; i < niocount; i++) {
1124                 if ((int)remote_rcs[i] < 0)
1125                         return remote_rcs[i];
1126
1127                 if (remote_rcs[i] != 0) {
1128                         CDEBUG(D_INFO, "rc[%d] invalid (%d) req %p\n",
1129                                i, remote_rcs[i], req);
1130                         return -EPROTO;
1131                 }
1132         }
1133
1134         if (req->rq_bulk->bd_nob_transferred != requested_nob) {
1135                 CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
1136                        req->rq_bulk->bd_nob_transferred, requested_nob);
1137                 return -EPROTO;
1138         }
1139
1140         return 0;
1141 }
1142
1143 static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
1144 {
1145         if (p1->flag != p2->flag) {
1146                 unsigned mask = ~(OBD_BRW_FROM_GRANT | OBD_BRW_NOCACHE |
1147                                   OBD_BRW_SYNC | OBD_BRW_ASYNC |
1148                                   OBD_BRW_NOQUOTA | OBD_BRW_SOFT_SYNC);
1149
1150                 /* warn if we try to combine flags that we don't know to be
1151                  * safe to combine
1152                  */
1153                 if (unlikely((p1->flag & mask) != (p2->flag & mask))) {
1154                         CWARN("Saw flags 0x%x and 0x%x in the same brw, please report this at http://bugs.whamcloud.com/\n",
1155                               p1->flag, p2->flag);
1156                 }
1157                 return 0;
1158         }
1159
1160         return (p1->off + p1->count == p2->off);
1161 }
1162
1163 static u32 osc_checksum_bulk(int nob, u32 pg_count,
1164                              struct brw_page **pga, int opc,
1165                              enum cksum_type cksum_type)
1166 {
1167         __u32 cksum;
1168         int i = 0;
1169         struct cfs_crypto_hash_desc *hdesc;
1170         unsigned int bufsize;
1171         int err;
1172         unsigned char cfs_alg = cksum_obd2cfs(cksum_type);
1173
1174         LASSERT(pg_count > 0);
1175
1176         hdesc = cfs_crypto_hash_init(cfs_alg, NULL, 0);
1177         if (IS_ERR(hdesc)) {
1178                 CERROR("Unable to initialize checksum hash %s\n",
1179                        cfs_crypto_hash_name(cfs_alg));
1180                 return PTR_ERR(hdesc);
1181         }
1182
1183         while (nob > 0 && pg_count > 0) {
1184                 int count = pga[i]->count > nob ? nob : pga[i]->count;
1185
1186                 /* corrupt the data before we compute the checksum, to
1187                  * simulate an OST->client data error
1188                  */
1189                 if (i == 0 && opc == OST_READ &&
1190                     OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_RECEIVE)) {
1191                         unsigned char *ptr = kmap(pga[i]->pg);
1192                         int off = pga[i]->off & ~PAGE_MASK;
1193
1194                         memcpy(ptr + off, "bad1", min(4, nob));
1195                         kunmap(pga[i]->pg);
1196                 }
1197                 cfs_crypto_hash_update_page(hdesc, pga[i]->pg,
1198                                             pga[i]->off & ~PAGE_MASK,
1199                                   count);
1200                 CDEBUG(D_PAGE,
1201                        "page %p map %p index %lu flags %lx count %u priv %0lx: off %d\n",
1202                        pga[i]->pg, pga[i]->pg->mapping, pga[i]->pg->index,
1203                        (long)pga[i]->pg->flags, page_count(pga[i]->pg),
1204                        page_private(pga[i]->pg),
1205                        (int)(pga[i]->off & ~PAGE_MASK));
1206
1207                 nob -= pga[i]->count;
1208                 pg_count--;
1209                 i++;
1210         }
1211
1212         bufsize = sizeof(cksum);
1213         err = cfs_crypto_hash_final(hdesc, (unsigned char *)&cksum, &bufsize);
1214
1215         /* For sending we only compute the wrong checksum instead
1216          * of corrupting the data so it is still correct on a redo
1217          */
1218         if (opc == OST_WRITE && OBD_FAIL_CHECK(OBD_FAIL_OSC_CHECKSUM_SEND))
1219                 cksum++;
1220
1221         return cksum;
1222 }
1223
1224 static int osc_brw_prep_request(int cmd, struct client_obd *cli,
1225                                 struct obdo *oa,
1226                                 struct lov_stripe_md *lsm, u32 page_count,
1227                                 struct brw_page **pga,
1228                                 struct ptlrpc_request **reqp,
1229                                 int reserve,
1230                                 int resend)
1231 {
1232         struct ptlrpc_request *req;
1233         struct ptlrpc_bulk_desc *desc;
1234         struct ost_body *body;
1235         struct obd_ioobj *ioobj;
1236         struct niobuf_remote *niobuf;
1237         int niocount, i, requested_nob, opc, rc;
1238         struct osc_brw_async_args *aa;
1239         struct req_capsule *pill;
1240         struct brw_page *pg_prev;
1241
1242         if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ))
1243                 return -ENOMEM; /* Recoverable */
1244         if (OBD_FAIL_CHECK(OBD_FAIL_OSC_BRW_PREP_REQ2))
1245                 return -EINVAL; /* Fatal */
1246
1247         if ((cmd & OBD_BRW_WRITE) != 0) {
1248                 opc = OST_WRITE;
1249                 req = ptlrpc_request_alloc_pool(cli->cl_import,
1250                                                 osc_rq_pool,
1251                                                 &RQF_OST_BRW_WRITE);
1252         } else {
1253                 opc = OST_READ;
1254                 req = ptlrpc_request_alloc(cli->cl_import, &RQF_OST_BRW_READ);
1255         }
1256         if (!req)
1257                 return -ENOMEM;
1258
1259         for (niocount = i = 1; i < page_count; i++) {
1260                 if (!can_merge_pages(pga[i - 1], pga[i]))
1261                         niocount++;
1262         }
1263
1264         pill = &req->rq_pill;
1265         req_capsule_set_size(pill, &RMF_OBD_IOOBJ, RCL_CLIENT,
1266                              sizeof(*ioobj));
1267         req_capsule_set_size(pill, &RMF_NIOBUF_REMOTE, RCL_CLIENT,
1268                              niocount * sizeof(*niobuf));
1269
1270         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, opc);
1271         if (rc) {
1272                 ptlrpc_request_free(req);
1273                 return rc;
1274         }
1275         req->rq_request_portal = OST_IO_PORTAL; /* bug 7198 */
1276         ptlrpc_at_set_req_timeout(req);
1277         /* ask ptlrpc not to resend on EINPROGRESS since BRWs have their own
1278          * retry logic
1279          */
1280         req->rq_no_retry_einprogress = 1;
1281
1282         desc = ptlrpc_prep_bulk_imp(req, page_count,
1283                 cli->cl_import->imp_connect_data.ocd_brw_size >> LNET_MTU_BITS,
1284                 opc == OST_WRITE ? BULK_GET_SOURCE : BULK_PUT_SINK,
1285                 OST_BULK_PORTAL);
1286
1287         if (!desc) {
1288                 rc = -ENOMEM;
1289                 goto out;
1290         }
1291         /* NB request now owns desc and will free it when it gets freed */
1292
1293         body = req_capsule_client_get(pill, &RMF_OST_BODY);
1294         ioobj = req_capsule_client_get(pill, &RMF_OBD_IOOBJ);
1295         niobuf = req_capsule_client_get(pill, &RMF_NIOBUF_REMOTE);
1296         LASSERT(body && ioobj && niobuf);
1297
1298         lustre_set_wire_obdo(&req->rq_import->imp_connect_data, &body->oa, oa);
1299
1300         obdo_to_ioobj(oa, ioobj);
1301         ioobj->ioo_bufcnt = niocount;
1302         /* The high bits of ioo_max_brw tells server _maximum_ number of bulks
1303          * that might be send for this request.  The actual number is decided
1304          * when the RPC is finally sent in ptlrpc_register_bulk(). It sends
1305          * "max - 1" for old client compatibility sending "0", and also so the
1306          * the actual maximum is a power-of-two number, not one less. LU-1431
1307          */
1308         ioobj_max_brw_set(ioobj, desc->bd_md_max_brw);
1309         LASSERT(page_count > 0);
1310         pg_prev = pga[0];
1311         for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
1312                 struct brw_page *pg = pga[i];
1313                 int poff = pg->off & ~PAGE_MASK;
1314
1315                 LASSERT(pg->count > 0);
1316                 /* make sure there is no gap in the middle of page array */
1317                 LASSERTF(page_count == 1 ||
1318                          (ergo(i == 0, poff + pg->count == PAGE_SIZE) &&
1319                           ergo(i > 0 && i < page_count - 1,
1320                                poff == 0 && pg->count == PAGE_SIZE)   &&
1321                           ergo(i == page_count - 1, poff == 0)),
1322                          "i: %d/%d pg: %p off: %llu, count: %u\n",
1323                          i, page_count, pg, pg->off, pg->count);
1324                 LASSERTF(i == 0 || pg->off > pg_prev->off,
1325                          "i %d p_c %u pg %p [pri %lu ind %lu] off %llu prev_pg %p [pri %lu ind %lu] off %llu\n",
1326                          i, page_count,
1327                          pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
1328                          pg_prev->pg, page_private(pg_prev->pg),
1329                          pg_prev->pg->index, pg_prev->off);
1330                 LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
1331                         (pg->flag & OBD_BRW_SRVLOCK));
1332
1333                 ptlrpc_prep_bulk_page_pin(desc, pg->pg, poff, pg->count);
1334                 requested_nob += pg->count;
1335
1336                 if (i > 0 && can_merge_pages(pg_prev, pg)) {
1337                         niobuf--;
1338                         niobuf->len += pg->count;
1339                 } else {
1340                         niobuf->offset = pg->off;
1341                         niobuf->len = pg->count;
1342                         niobuf->flags = pg->flag;
1343                 }
1344                 pg_prev = pg;
1345         }
1346
1347         LASSERTF((void *)(niobuf - niocount) ==
1348                 req_capsule_client_get(&req->rq_pill, &RMF_NIOBUF_REMOTE),
1349                 "want %p - real %p\n", req_capsule_client_get(&req->rq_pill,
1350                 &RMF_NIOBUF_REMOTE), (void *)(niobuf - niocount));
1351
1352         osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);
1353         if (resend) {
1354                 if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1355                         body->oa.o_valid |= OBD_MD_FLFLAGS;
1356                         body->oa.o_flags = 0;
1357                 }
1358                 body->oa.o_flags |= OBD_FL_RECOV_RESEND;
1359         }
1360
1361         if (osc_should_shrink_grant(cli))
1362                 osc_shrink_grant_local(cli, &body->oa);
1363
1364         /* size[REQ_REC_OFF] still sizeof (*body) */
1365         if (opc == OST_WRITE) {
1366                 if (cli->cl_checksum &&
1367                     !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1368                         /* store cl_cksum_type in a local variable since
1369                          * it can be changed via lprocfs
1370                          */
1371                         enum cksum_type cksum_type = cli->cl_cksum_type;
1372
1373                         if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0) {
1374                                 oa->o_flags &= OBD_FL_LOCAL_MASK;
1375                                 body->oa.o_flags = 0;
1376                         }
1377                         body->oa.o_flags |= cksum_type_pack(cksum_type);
1378                         body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1379                         body->oa.o_cksum = osc_checksum_bulk(requested_nob,
1380                                                              page_count, pga,
1381                                                              OST_WRITE,
1382                                                              cksum_type);
1383                         CDEBUG(D_PAGE, "checksum at write origin: %x\n",
1384                                body->oa.o_cksum);
1385                         /* save this in 'oa', too, for later checking */
1386                         oa->o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1387                         oa->o_flags |= cksum_type_pack(cksum_type);
1388                 } else {
1389                         /* clear out the checksum flag, in case this is a
1390                          * resend but cl_checksum is no longer set. b=11238
1391                          */
1392                         oa->o_valid &= ~OBD_MD_FLCKSUM;
1393                 }
1394                 oa->o_cksum = body->oa.o_cksum;
1395                 /* 1 RC per niobuf */
1396                 req_capsule_set_size(pill, &RMF_RCS, RCL_SERVER,
1397                                      sizeof(__u32) * niocount);
1398         } else {
1399                 if (cli->cl_checksum &&
1400                     !sptlrpc_flavor_has_bulk(&req->rq_flvr)) {
1401                         if ((body->oa.o_valid & OBD_MD_FLFLAGS) == 0)
1402                                 body->oa.o_flags = 0;
1403                         body->oa.o_flags |= cksum_type_pack(cli->cl_cksum_type);
1404                         body->oa.o_valid |= OBD_MD_FLCKSUM | OBD_MD_FLFLAGS;
1405                 }
1406         }
1407         ptlrpc_request_set_replen(req);
1408
1409         CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
1410         aa = ptlrpc_req_async_args(req);
1411         aa->aa_oa = oa;
1412         aa->aa_requested_nob = requested_nob;
1413         aa->aa_nio_count = niocount;
1414         aa->aa_page_count = page_count;
1415         aa->aa_resends = 0;
1416         aa->aa_ppga = pga;
1417         aa->aa_cli = cli;
1418         INIT_LIST_HEAD(&aa->aa_oaps);
1419
1420         *reqp = req;
1421         return 0;
1422
1423  out:
1424         ptlrpc_req_finished(req);
1425         return rc;
1426 }
1427
1428 static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
1429                                 __u32 client_cksum, __u32 server_cksum, int nob,
1430                                 u32 page_count, struct brw_page **pga,
1431                                 enum cksum_type client_cksum_type)
1432 {
1433         __u32 new_cksum;
1434         char *msg;
1435         enum cksum_type cksum_type;
1436
1437         if (server_cksum == client_cksum) {
1438                 CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1439                 return 0;
1440         }
1441
1442         cksum_type = cksum_type_unpack(oa->o_valid & OBD_MD_FLFLAGS ?
1443                                        oa->o_flags : 0);
1444         new_cksum = osc_checksum_bulk(nob, page_count, pga, OST_WRITE,
1445                                       cksum_type);
1446
1447         if (cksum_type != client_cksum_type)
1448                 msg = "the server did not use the checksum type specified in the original request - likely a protocol problem"
1449                         ;
1450         else if (new_cksum == server_cksum)
1451                 msg = "changed on the client after we checksummed it - likely false positive due to mmap IO (bug 11742)"
1452                         ;
1453         else if (new_cksum == client_cksum)
1454                 msg = "changed in transit before arrival at OST";
1455         else
1456                 msg = "changed in transit AND doesn't match the original - likely false positive due to mmap IO (bug 11742)"
1457                         ;
1458
1459         LCONSOLE_ERROR_MSG(0x132, "BAD WRITE CHECKSUM: %s: from %s inode "DFID
1460                            " object "DOSTID" extent [%llu-%llu]\n",
1461                            msg, libcfs_nid2str(peer->nid),
1462                            oa->o_valid & OBD_MD_FLFID ? oa->o_parent_seq : (__u64)0,
1463                            oa->o_valid & OBD_MD_FLFID ? oa->o_parent_oid : 0,
1464                            oa->o_valid & OBD_MD_FLFID ? oa->o_parent_ver : 0,
1465                            POSTID(&oa->o_oi), pga[0]->off,
1466                            pga[page_count-1]->off + pga[page_count-1]->count - 1);
1467         CERROR("original client csum %x (type %x), server csum %x (type %x), client csum now %x\n",
1468                client_cksum, client_cksum_type,
1469                server_cksum, cksum_type, new_cksum);
1470         return 1;
1471 }
1472
1473 /* Note rc enters this function as number of bytes transferred */
1474 static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
1475 {
1476         struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
1477         const lnet_process_id_t *peer =
1478                         &req->rq_import->imp_connection->c_peer;
1479         struct client_obd *cli = aa->aa_cli;
1480         struct ost_body *body;
1481         __u32 client_cksum = 0;
1482
1483         if (rc < 0 && rc != -EDQUOT) {
1484                 DEBUG_REQ(D_INFO, req, "Failed request with rc = %d\n", rc);
1485                 return rc;
1486         }
1487
1488         LASSERTF(req->rq_repmsg, "rc = %d\n", rc);
1489         body = req_capsule_server_get(&req->rq_pill, &RMF_OST_BODY);
1490         if (!body) {
1491                 DEBUG_REQ(D_INFO, req, "Can't unpack body\n");
1492                 return -EPROTO;
1493         }
1494
1495         /* set/clear over quota flag for a uid/gid */
1496         if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
1497             body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA)) {
1498                 unsigned int qid[MAXQUOTAS] = { body->oa.o_uid, body->oa.o_gid };
1499
1500                 CDEBUG(D_QUOTA, "setdq for [%u %u] with valid %#llx, flags %x\n",
1501                        body->oa.o_uid, body->oa.o_gid, body->oa.o_valid,
1502                        body->oa.o_flags);
1503                 osc_quota_setdq(cli, qid, body->oa.o_valid, body->oa.o_flags);
1504         }
1505
1506         osc_update_grant(cli, body);
1507
1508         if (rc < 0)
1509                 return rc;
1510
1511         if (aa->aa_oa->o_valid & OBD_MD_FLCKSUM)
1512                 client_cksum = aa->aa_oa->o_cksum; /* save for later */
1513
1514         if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1515                 if (rc > 0) {
1516                         CERROR("Unexpected +ve rc %d\n", rc);
1517                         return -EPROTO;
1518                 }
1519                 LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);
1520
1521                 if (sptlrpc_cli_unwrap_bulk_write(req, req->rq_bulk))
1522                         return -EAGAIN;
1523
1524                 if ((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) && client_cksum &&
1525                     check_write_checksum(&body->oa, peer, client_cksum,
1526                                          body->oa.o_cksum, aa->aa_requested_nob,
1527                                          aa->aa_page_count, aa->aa_ppga,
1528                                          cksum_type_unpack(aa->aa_oa->o_flags)))
1529                         return -EAGAIN;
1530
1531                 rc = check_write_rcs(req, aa->aa_requested_nob,
1532                                      aa->aa_nio_count,
1533                                      aa->aa_page_count, aa->aa_ppga);
1534                 goto out;
1535         }
1536
1537         /* The rest of this function executes only for OST_READs */
1538
1539         /* if unwrap_bulk failed, return -EAGAIN to retry */
1540         rc = sptlrpc_cli_unwrap_bulk_read(req, req->rq_bulk, rc);
1541         if (rc < 0) {
1542                 rc = -EAGAIN;
1543                 goto out;
1544         }
1545
1546         if (rc > aa->aa_requested_nob) {
1547                 CERROR("Unexpected rc %d (%d requested)\n", rc,
1548                        aa->aa_requested_nob);
1549                 return -EPROTO;
1550         }
1551
1552         if (rc != req->rq_bulk->bd_nob_transferred) {
1553                 CERROR("Unexpected rc %d (%d transferred)\n",
1554                        rc, req->rq_bulk->bd_nob_transferred);
1555                 return -EPROTO;
1556         }
1557
1558         if (rc < aa->aa_requested_nob)
1559                 handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);
1560
1561         if (body->oa.o_valid & OBD_MD_FLCKSUM) {
1562                 static int cksum_counter;
1563                 __u32 server_cksum = body->oa.o_cksum;
1564                 char *via = "";
1565                 char *router = "";
1566                 enum cksum_type cksum_type;
1567
1568                 cksum_type = cksum_type_unpack(body->oa.o_valid&OBD_MD_FLFLAGS ?
1569                                                body->oa.o_flags : 0);
1570                 client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
1571                                                  aa->aa_ppga, OST_READ,
1572                                                  cksum_type);
1573
1574                 if (peer->nid != req->rq_bulk->bd_sender) {
1575                         via = " via ";
1576                         router = libcfs_nid2str(req->rq_bulk->bd_sender);
1577                 }
1578
1579                 if (server_cksum != client_cksum) {
1580                         LCONSOLE_ERROR_MSG(0x133, "%s: BAD READ CHECKSUM: from %s%s%s inode " DFID " object " DOSTID " extent [%llu-%llu]\n",
1581                                            req->rq_import->imp_obd->obd_name,
1582                                            libcfs_nid2str(peer->nid),
1583                                            via, router,
1584                                            body->oa.o_valid & OBD_MD_FLFID ?
1585                                            body->oa.o_parent_seq : (__u64)0,
1586                                            body->oa.o_valid & OBD_MD_FLFID ?
1587                                            body->oa.o_parent_oid : 0,
1588                                            body->oa.o_valid & OBD_MD_FLFID ?
1589                                            body->oa.o_parent_ver : 0,
1590                                            POSTID(&body->oa.o_oi),
1591                                            aa->aa_ppga[0]->off,
1592                                            aa->aa_ppga[aa->aa_page_count-1]->off +
1593                                            aa->aa_ppga[aa->aa_page_count-1]->count -
1594                                            1);
1595                         CERROR("client %x, server %x, cksum_type %x\n",
1596                                client_cksum, server_cksum, cksum_type);
1597                         cksum_counter = 0;
1598                         aa->aa_oa->o_cksum = client_cksum;
1599                         rc = -EAGAIN;
1600                 } else {
1601                         cksum_counter++;
1602                         CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
1603                         rc = 0;
1604                 }
1605         } else if (unlikely(client_cksum)) {
1606                 static int cksum_missed;
1607
1608                 cksum_missed++;
1609                 if ((cksum_missed & (-cksum_missed)) == cksum_missed)
1610                         CERROR("Checksum %u requested from %s but not sent\n",
1611                                cksum_missed, libcfs_nid2str(peer->nid));
1612         } else {
1613                 rc = 0;
1614         }
1615 out:
1616         if (rc >= 0)
1617                 lustre_get_wire_obdo(&req->rq_import->imp_connect_data,
1618                                      aa->aa_oa, &body->oa);
1619
1620         return rc;
1621 }
1622
1623 static int osc_brw_redo_request(struct ptlrpc_request *request,
1624                                 struct osc_brw_async_args *aa, int rc)
1625 {
1626         struct ptlrpc_request *new_req;
1627         struct osc_brw_async_args *new_aa;
1628         struct osc_async_page *oap;
1629
1630         DEBUG_REQ(rc == -EINPROGRESS ? D_RPCTRACE : D_ERROR, request,
1631                   "redo for recoverable error %d", rc);
1632
1633         rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
1634                                         OST_WRITE ? OBD_BRW_WRITE : OBD_BRW_READ,
1635                                   aa->aa_cli, aa->aa_oa,
1636                                   NULL /* lsm unused by osc currently */,
1637                                   aa->aa_page_count, aa->aa_ppga,
1638                                   &new_req, 0, 1);
1639         if (rc)
1640                 return rc;
1641
1642         list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
1643                 if (oap->oap_request) {
1644                         LASSERTF(request == oap->oap_request,
1645                                  "request %p != oap_request %p\n",
1646                                  request, oap->oap_request);
1647                         if (oap->oap_interrupted) {
1648                                 ptlrpc_req_finished(new_req);
1649                                 return -EINTR;
1650                         }
1651                 }
1652         }
1653         /* New request takes over pga and oaps from old request.
1654          * Note that copying a list_head doesn't work, need to move it...
1655          */
1656         aa->aa_resends++;
1657         new_req->rq_interpret_reply = request->rq_interpret_reply;
1658         new_req->rq_async_args = request->rq_async_args;
1659         new_req->rq_commit_cb = request->rq_commit_cb;
1660         /* cap resend delay to the current request timeout, this is similar to
1661          * what ptlrpc does (see after_reply())
1662          */
1663         if (aa->aa_resends > new_req->rq_timeout)
1664                 new_req->rq_sent = ktime_get_real_seconds() + new_req->rq_timeout;
1665         else
1666                 new_req->rq_sent = ktime_get_real_seconds() + aa->aa_resends;
1667         new_req->rq_generation_set = 1;
1668         new_req->rq_import_generation = request->rq_import_generation;
1669
1670         new_aa = ptlrpc_req_async_args(new_req);
1671
1672         INIT_LIST_HEAD(&new_aa->aa_oaps);
1673         list_splice_init(&aa->aa_oaps, &new_aa->aa_oaps);
1674         INIT_LIST_HEAD(&new_aa->aa_exts);
1675         list_splice_init(&aa->aa_exts, &new_aa->aa_exts);
1676         new_aa->aa_resends = aa->aa_resends;
1677
1678         list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
1679                 if (oap->oap_request) {
1680                         ptlrpc_req_finished(oap->oap_request);
1681                         oap->oap_request = ptlrpc_request_addref(new_req);
1682                 }
1683         }
1684
1685         /* XXX: This code will run into problem if we're going to support
1686          * to add a series of BRW RPCs into a self-defined ptlrpc_request_set
1687          * and wait for all of them to be finished. We should inherit request
1688          * set from old request.
1689          */
1690         ptlrpcd_add_req(new_req);
1691
1692         DEBUG_REQ(D_INFO, new_req, "new request");
1693         return 0;
1694 }
1695
1696 /*
1697  * ugh, we want disk allocation on the target to happen in offset order.  we'll
1698  * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
1699  * fine for our small page arrays and doesn't require allocation.  its an
1700  * insertion sort that swaps elements that are strides apart, shrinking the
1701  * stride down until its '1' and the array is sorted.
1702  */
1703 static void sort_brw_pages(struct brw_page **array, int num)
1704 {
1705         int stride, i, j;
1706         struct brw_page *tmp;
1707
1708         if (num == 1)
1709                 return;
1710         for (stride = 1; stride < num ; stride = (stride * 3) + 1)
1711                 ;
1712
1713         do {
1714                 stride /= 3;
1715                 for (i = stride ; i < num ; i++) {
1716                         tmp = array[i];
1717                         j = i;
1718                         while (j >= stride && array[j - stride]->off > tmp->off) {
1719                                 array[j] = array[j - stride];
1720                                 j -= stride;
1721                         }
1722                         array[j] = tmp;
1723                 }
1724         } while (stride > 1);
1725 }
1726
1727 static void osc_release_ppga(struct brw_page **ppga, u32 count)
1728 {
1729         LASSERT(ppga);
1730         kfree(ppga);
1731 }
1732
1733 static int brw_interpret(const struct lu_env *env,
1734                          struct ptlrpc_request *req, void *data, int rc)
1735 {
1736         struct osc_brw_async_args *aa = data;
1737         struct osc_extent *ext;
1738         struct osc_extent *tmp;
1739         struct client_obd *cli = aa->aa_cli;
1740
1741         rc = osc_brw_fini_request(req, rc);
1742         CDEBUG(D_INODE, "request %p aa %p rc %d\n", req, aa, rc);
1743         /* When server return -EINPROGRESS, client should always retry
1744          * regardless of the number of times the bulk was resent already.
1745          */
1746         if (osc_recoverable_error(rc)) {
1747                 if (req->rq_import_generation !=
1748                     req->rq_import->imp_generation) {
1749                         CDEBUG(D_HA, "%s: resend cross eviction for object: " DOSTID ", rc = %d.\n",
1750                                req->rq_import->imp_obd->obd_name,
1751                                POSTID(&aa->aa_oa->o_oi), rc);
1752                 } else if (rc == -EINPROGRESS ||
1753                     client_should_resend(aa->aa_resends, aa->aa_cli)) {
1754                         rc = osc_brw_redo_request(req, aa, rc);
1755                 } else {
1756                         CERROR("%s: too many resent retries for object: %llu:%llu, rc = %d.\n",
1757                                req->rq_import->imp_obd->obd_name,
1758                                POSTID(&aa->aa_oa->o_oi), rc);
1759                 }
1760
1761                 if (rc == 0)
1762                         return 0;
1763                 else if (rc == -EAGAIN || rc == -EINPROGRESS)
1764                         rc = -EIO;
1765         }
1766
1767         if (rc == 0) {
1768                 struct obdo *oa = aa->aa_oa;
1769                 struct cl_attr *attr  = &osc_env_info(env)->oti_attr;
1770                 unsigned long valid = 0;
1771                 struct cl_object *obj;
1772                 struct osc_async_page *last;
1773
1774                 last = brw_page2oap(aa->aa_ppga[aa->aa_page_count - 1]);
1775                 obj = osc2cl(last->oap_obj);
1776
1777                 cl_object_attr_lock(obj);
1778                 if (oa->o_valid & OBD_MD_FLBLOCKS) {
1779                         attr->cat_blocks = oa->o_blocks;
1780                         valid |= CAT_BLOCKS;
1781                 }
1782                 if (oa->o_valid & OBD_MD_FLMTIME) {
1783                         attr->cat_mtime = oa->o_mtime;
1784                         valid |= CAT_MTIME;
1785                 }
1786                 if (oa->o_valid & OBD_MD_FLATIME) {
1787                         attr->cat_atime = oa->o_atime;
1788                         valid |= CAT_ATIME;
1789                 }
1790                 if (oa->o_valid & OBD_MD_FLCTIME) {
1791                         attr->cat_ctime = oa->o_ctime;
1792                         valid |= CAT_CTIME;
1793                 }
1794
1795                 if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
1796                         struct lov_oinfo *loi = cl2osc(obj)->oo_oinfo;
1797                         loff_t last_off = last->oap_count + last->oap_obj_off;
1798
1799                         /* Change file size if this is an out of quota or
1800                          * direct IO write and it extends the file size
1801                          */
1802                         if (loi->loi_lvb.lvb_size < last_off) {
1803                                 attr->cat_size = last_off;
1804                                 valid |= CAT_SIZE;
1805                         }
1806                         /* Extend KMS if it's not a lockless write */
1807                         if (loi->loi_kms < last_off &&
1808                             oap2osc_page(last)->ops_srvlock == 0) {
1809                                 attr->cat_kms = last_off;
1810                                 valid |= CAT_KMS;
1811                         }
1812                 }
1813
1814                 if (valid != 0)
1815                         cl_object_attr_set(env, obj, attr, valid);
1816                 cl_object_attr_unlock(obj);
1817         }
1818         kmem_cache_free(obdo_cachep, aa->aa_oa);
1819
1820         list_for_each_entry_safe(ext, tmp, &aa->aa_exts, oe_link) {
1821                 list_del_init(&ext->oe_link);
1822                 osc_extent_finish(env, ext, 1, rc);
1823         }
1824         LASSERT(list_empty(&aa->aa_exts));
1825         LASSERT(list_empty(&aa->aa_oaps));
1826
1827         cl_req_completion(env, aa->aa_clerq, rc < 0 ? rc :
1828                           req->rq_bulk->bd_nob_transferred);
1829         osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
1830         ptlrpc_lprocfs_brw(req, req->rq_bulk->bd_nob_transferred);
1831
1832         spin_lock(&cli->cl_loi_list_lock);
1833         /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
1834          * is called so we know whether to go to sync BRWs or wait for more
1835          * RPCs to complete
1836          */
1837         if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE)
1838                 cli->cl_w_in_flight--;
1839         else
1840                 cli->cl_r_in_flight--;
1841         osc_wake_cache_waiters(cli);
1842         spin_unlock(&cli->cl_loi_list_lock);
1843
1844         osc_io_unplug(env, cli, NULL);
1845         return rc;
1846 }
1847
1848 static void brw_commit(struct ptlrpc_request *req)
1849 {
1850         spin_lock(&req->rq_lock);
1851         /*
1852          * If osc_inc_unstable_pages (via osc_extent_finish) races with
1853          * this called via the rq_commit_cb, I need to ensure
1854          * osc_dec_unstable_pages is still called. Otherwise unstable
1855          * pages may be leaked.
1856          */
1857         if (req->rq_unstable) {
1858                 spin_unlock(&req->rq_lock);
1859                 osc_dec_unstable_pages(req);
1860                 spin_lock(&req->rq_lock);
1861         } else {
1862                 req->rq_committed = 1;
1863         }
1864         spin_unlock(&req->rq_lock);
1865 }
1866
1867 /**
1868  * Build an RPC by the list of extent @ext_list. The caller must ensure
1869  * that the total pages in this list are NOT over max pages per RPC.
1870  * Extents in the list must be in OES_RPC state.
1871  */
1872 int osc_build_rpc(const struct lu_env *env, struct client_obd *cli,
1873                   struct list_head *ext_list, int cmd)
1874 {
1875         struct ptlrpc_request *req = NULL;
1876         struct osc_extent *ext;
1877         struct brw_page **pga = NULL;
1878         struct osc_brw_async_args *aa = NULL;
1879         struct obdo *oa = NULL;
1880         struct osc_async_page *oap;
1881         struct osc_async_page *tmp;
1882         struct cl_req *clerq = NULL;
1883         enum cl_req_type crt = (cmd & OBD_BRW_WRITE) ? CRT_WRITE : CRT_READ;
1884         struct ldlm_lock *lock = NULL;
1885         struct cl_req_attr *crattr = NULL;
1886         u64 starting_offset = OBD_OBJECT_EOF;
1887         u64 ending_offset = 0;
1888         int mpflag = 0;
1889         int mem_tight = 0;
1890         int page_count = 0;
1891         int i;
1892         int rc;
1893         struct ost_body *body;
1894         LIST_HEAD(rpc_list);
1895
1896         LASSERT(!list_empty(ext_list));
1897
1898         /* add pages into rpc_list to build BRW rpc */
1899         list_for_each_entry(ext, ext_list, oe_link) {
1900                 LASSERT(ext->oe_state == OES_RPC);
1901                 mem_tight |= ext->oe_memalloc;
1902                 list_for_each_entry(oap, &ext->oe_pages, oap_pending_item) {
1903                         ++page_count;
1904                         list_add_tail(&oap->oap_rpc_item, &rpc_list);
1905                         if (starting_offset > oap->oap_obj_off)
1906                                 starting_offset = oap->oap_obj_off;
1907                         else
1908                                 LASSERT(oap->oap_page_off == 0);
1909                         if (ending_offset < oap->oap_obj_off + oap->oap_count)
1910                                 ending_offset = oap->oap_obj_off +
1911                                                 oap->oap_count;
1912                         else
1913                                 LASSERT(oap->oap_page_off + oap->oap_count ==
1914                                         PAGE_SIZE);
1915                 }
1916         }
1917
1918         if (mem_tight)
1919                 mpflag = cfs_memory_pressure_get_and_set();
1920
1921         crattr = kzalloc(sizeof(*crattr), GFP_NOFS);
1922         if (!crattr) {
1923                 rc = -ENOMEM;
1924                 goto out;
1925         }
1926
1927         pga = kcalloc(page_count, sizeof(*pga), GFP_NOFS);
1928         if (!pga) {
1929                 rc = -ENOMEM;
1930                 goto out;
1931         }
1932
1933         oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
1934         if (!oa) {
1935                 rc = -ENOMEM;
1936                 goto out;
1937         }
1938
1939         i = 0;
1940         list_for_each_entry(oap, &rpc_list, oap_rpc_item) {
1941                 struct cl_page *page = oap2cl_page(oap);
1942
1943                 if (!clerq) {
1944                         clerq = cl_req_alloc(env, page, crt,
1945                                              1 /* only 1-object rpcs for now */);
1946                         if (IS_ERR(clerq)) {
1947                                 rc = PTR_ERR(clerq);
1948                                 goto out;
1949                         }
1950                         lock = oap->oap_ldlm_lock;
1951                 }
1952                 if (mem_tight)
1953                         oap->oap_brw_flags |= OBD_BRW_MEMALLOC;
1954                 pga[i] = &oap->oap_brw_page;
1955                 pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
1956                 CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
1957                        pga[i]->pg, oap->oap_page->index, oap,
1958                        pga[i]->flag);
1959                 i++;
1960                 cl_req_page_add(env, clerq, page);
1961         }
1962
1963         /* always get the data for the obdo for the rpc */
1964         LASSERT(clerq);
1965         crattr->cra_oa = oa;
1966         cl_req_attr_set(env, clerq, crattr, ~0ULL);
1967         if (lock) {
1968                 oa->o_handle = lock->l_remote_handle;
1969                 oa->o_valid |= OBD_MD_FLHANDLE;
1970         }
1971
1972         rc = cl_req_prep(env, clerq);
1973         if (rc != 0) {
1974                 CERROR("cl_req_prep failed: %d\n", rc);
1975                 goto out;
1976         }
1977
1978         sort_brw_pages(pga, page_count);
1979         rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count,
1980                                   pga, &req, 1, 0);
1981         if (rc != 0) {
1982                 CERROR("prep_req failed: %d\n", rc);
1983                 goto out;
1984         }
1985
1986         req->rq_commit_cb = brw_commit;
1987         req->rq_interpret_reply = brw_interpret;
1988
1989         if (mem_tight != 0)
1990                 req->rq_memalloc = 1;
1991
1992         /* Need to update the timestamps after the request is built in case
1993          * we race with setattr (locally or in queue at OST).  If OST gets
1994          * later setattr before earlier BRW (as determined by the request xid),
1995          * the OST will not use BRW timestamps.  Sadly, there is no obvious
1996          * way to do this in a single call.  bug 10150
1997          */
1998         body = req_capsule_client_get(&req->rq_pill, &RMF_OST_BODY);
1999         crattr->cra_oa = &body->oa;
2000         cl_req_attr_set(env, clerq, crattr,
2001                         OBD_MD_FLMTIME|OBD_MD_FLCTIME|OBD_MD_FLATIME);
2002
2003         lustre_msg_set_jobid(req->rq_reqmsg, crattr->cra_jobid);
2004
2005         CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2006         aa = ptlrpc_req_async_args(req);
2007         INIT_LIST_HEAD(&aa->aa_oaps);
2008         list_splice_init(&rpc_list, &aa->aa_oaps);
2009         INIT_LIST_HEAD(&aa->aa_exts);
2010         list_splice_init(ext_list, &aa->aa_exts);
2011         aa->aa_clerq = clerq;
2012
2013         /* queued sync pages can be torn down while the pages
2014          * were between the pending list and the rpc
2015          */
2016         tmp = NULL;
2017         list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
2018                 /* only one oap gets a request reference */
2019                 if (!tmp)
2020                         tmp = oap;
2021                 if (oap->oap_interrupted && !req->rq_intr) {
2022                         CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
2023                                oap, req);
2024                         ptlrpc_mark_interrupted(req);
2025                 }
2026         }
2027         if (tmp)
2028                 tmp->oap_request = ptlrpc_request_addref(req);
2029
2030         spin_lock(&cli->cl_loi_list_lock);
2031         starting_offset >>= PAGE_SHIFT;
2032         if (cmd == OBD_BRW_READ) {
2033                 cli->cl_r_in_flight++;
2034                 lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
2035                 lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
2036                 lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
2037                                       starting_offset + 1);
2038         } else {
2039                 cli->cl_w_in_flight++;
2040                 lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
2041                 lprocfs_oh_tally(&cli->cl_write_rpc_hist, cli->cl_w_in_flight);
2042                 lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
2043                                       starting_offset + 1);
2044         }
2045         spin_unlock(&cli->cl_loi_list_lock);
2046
2047         DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
2048                   page_count, aa, cli->cl_r_in_flight,
2049                   cli->cl_w_in_flight);
2050
2051         ptlrpcd_add_req(req);
2052         rc = 0;
2053
2054 out:
2055         if (mem_tight != 0)
2056                 cfs_memory_pressure_restore(mpflag);
2057
2058         kfree(crattr);
2059
2060         if (rc != 0) {
2061                 LASSERT(!req);
2062
2063                 if (oa)
2064                         kmem_cache_free(obdo_cachep, oa);
2065                 kfree(pga);
2066                 /* this should happen rarely and is pretty bad, it makes the
2067                  * pending list not follow the dirty order
2068                  */
2069                 while (!list_empty(ext_list)) {
2070                         ext = list_entry(ext_list->next, struct osc_extent,
2071                                          oe_link);
2072                         list_del_init(&ext->oe_link);
2073                         osc_extent_finish(env, ext, 0, rc);
2074                 }
2075                 if (clerq && !IS_ERR(clerq))
2076                         cl_req_completion(env, clerq, rc);
2077         }
2078         return rc;
2079 }
2080
2081 static int osc_set_lock_data_with_check(struct ldlm_lock *lock,
2082                                         struct ldlm_enqueue_info *einfo)
2083 {
2084         void *data = einfo->ei_cbdata;
2085         int set = 0;
2086
2087         LASSERT(lock->l_blocking_ast == einfo->ei_cb_bl);
2088         LASSERT(lock->l_resource->lr_type == einfo->ei_type);
2089         LASSERT(lock->l_completion_ast == einfo->ei_cb_cp);
2090         LASSERT(lock->l_glimpse_ast == einfo->ei_cb_gl);
2091
2092         lock_res_and_lock(lock);
2093
2094         if (!lock->l_ast_data)
2095                 lock->l_ast_data = data;
2096         if (lock->l_ast_data == data)
2097                 set = 1;
2098
2099         unlock_res_and_lock(lock);
2100
2101         return set;
2102 }
2103
2104 static int osc_set_data_with_check(struct lustre_handle *lockh,
2105                                    struct ldlm_enqueue_info *einfo)
2106 {
2107         struct ldlm_lock *lock = ldlm_handle2lock(lockh);
2108         int set = 0;
2109
2110         if (lock) {
2111                 set = osc_set_lock_data_with_check(lock, einfo);
2112                 LDLM_LOCK_PUT(lock);
2113         } else
2114                 CERROR("lockh %p, data %p - client evicted?\n",
2115                        lockh, einfo->ei_cbdata);
2116         return set;
2117 }
2118
2119 /* find any ldlm lock of the inode in osc
2120  * return 0    not find
2121  *      1    find one
2122  *      < 0    error
2123  */
2124 static int osc_find_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
2125                            ldlm_iterator_t replace, void *data)
2126 {
2127         struct ldlm_res_id res_id;
2128         struct obd_device *obd = class_exp2obd(exp);
2129         int rc = 0;
2130
2131         ostid_build_res_name(&lsm->lsm_oi, &res_id);
2132         rc = ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
2133         if (rc == LDLM_ITER_STOP)
2134                 return 1;
2135         if (rc == LDLM_ITER_CONTINUE)
2136                 return 0;
2137         return rc;
2138 }
2139
2140 static int osc_enqueue_fini(struct ptlrpc_request *req,
2141                             osc_enqueue_upcall_f upcall, void *cookie,
2142                             struct lustre_handle *lockh, enum ldlm_mode mode,
2143                             __u64 *flags, int agl, int errcode)
2144 {
2145         bool intent = *flags & LDLM_FL_HAS_INTENT;
2146         int rc;
2147
2148         /* The request was created before ldlm_cli_enqueue call. */
2149         if (intent && errcode == ELDLM_LOCK_ABORTED) {
2150                 struct ldlm_reply *rep;
2151
2152                 rep = req_capsule_server_get(&req->rq_pill, &RMF_DLM_REP);
2153
2154                 rep->lock_policy_res1 =
2155                         ptlrpc_status_ntoh(rep->lock_policy_res1);
2156                 if (rep->lock_policy_res1)
2157                         errcode = rep->lock_policy_res1;
2158                 if (!agl)
2159                         *flags |= LDLM_FL_LVB_READY;
2160         } else if (errcode == ELDLM_OK) {
2161                 *flags |= LDLM_FL_LVB_READY;
2162         }
2163
2164         /* Call the update callback. */
2165         rc = (*upcall)(cookie, lockh, errcode);
2166         /* release the reference taken in ldlm_cli_enqueue() */
2167         if (errcode == ELDLM_LOCK_MATCHED)
2168                 errcode = ELDLM_OK;
2169         if (errcode == ELDLM_OK && lustre_handle_is_used(lockh))
2170                 ldlm_lock_decref(lockh, mode);
2171
2172         return rc;
2173 }
2174
2175 static int osc_enqueue_interpret(const struct lu_env *env,
2176                                  struct ptlrpc_request *req,
2177                                  struct osc_enqueue_args *aa, int rc)
2178 {
2179         struct ldlm_lock *lock;
2180         struct lustre_handle *lockh = &aa->oa_lockh;
2181         enum ldlm_mode mode = aa->oa_mode;
2182         struct ost_lvb *lvb = aa->oa_lvb;
2183         __u32 lvb_len = sizeof(*lvb);
2184         __u64 flags = 0;
2185
2186
2187         /* ldlm_cli_enqueue is holding a reference on the lock, so it must
2188          * be valid.
2189          */
2190         lock = ldlm_handle2lock(lockh);
2191         LASSERTF(lock, "lockh %llx, req %p, aa %p - client evicted?\n",
2192                  lockh->cookie, req, aa);
2193
2194         /* Take an additional reference so that a blocking AST that
2195          * ldlm_cli_enqueue_fini() might post for a failed lock, is guaranteed
2196          * to arrive after an upcall has been executed by
2197          * osc_enqueue_fini().
2198          */
2199         ldlm_lock_addref(lockh, mode);
2200
2201         /* Let cl_lock_state_wait fail with -ERESTARTSYS to unuse sublocks. */
2202         OBD_FAIL_TIMEOUT(OBD_FAIL_LDLM_ENQUEUE_HANG, 2);
2203
2204         /* Let CP AST to grant the lock first. */
2205         OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_ENQ_RACE, 1);
2206
2207         if (aa->oa_agl) {
2208                 LASSERT(!aa->oa_lvb);
2209                 LASSERT(!aa->oa_flags);
2210                 aa->oa_flags = &flags;
2211         }
2212
2213         /* Complete obtaining the lock procedure. */
2214         rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_type, 1,
2215                                    aa->oa_mode, aa->oa_flags, lvb, lvb_len,
2216                                    lockh, rc);
2217         /* Complete osc stuff. */
2218         rc = osc_enqueue_fini(req, aa->oa_upcall, aa->oa_cookie, lockh, mode,
2219                               aa->oa_flags, aa->oa_agl, rc);
2220
2221         OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_CP_CANCEL_RACE, 10);
2222
2223         ldlm_lock_decref(lockh, mode);
2224         LDLM_LOCK_PUT(lock);
2225         return rc;
2226 }
2227
2228 struct ptlrpc_request_set *PTLRPCD_SET = (void *)1;
2229
2230 /* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
2231  * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
2232  * other synchronous requests, however keeping some locks and trying to obtain
2233  * others may take a considerable amount of time in a case of ost failure; and
2234  * when other sync requests do not get released lock from a client, the client
2235  * is evicted from the cluster -- such scenaries make the life difficult, so
2236  * release locks just after they are obtained.
2237  */
2238 int osc_enqueue_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2239                      __u64 *flags, ldlm_policy_data_t *policy,
2240                      struct ost_lvb *lvb, int kms_valid,
2241                      osc_enqueue_upcall_f upcall, void *cookie,
2242                      struct ldlm_enqueue_info *einfo,
2243                      struct ptlrpc_request_set *rqset, int async, int agl)
2244 {
2245         struct obd_device *obd = exp->exp_obd;
2246         struct lustre_handle lockh = { 0 };
2247         struct ptlrpc_request *req = NULL;
2248         int intent = *flags & LDLM_FL_HAS_INTENT;
2249         __u64 match_flags = *flags;
2250         enum ldlm_mode mode;
2251         int rc;
2252
2253         /* Filesystem lock extents are extended to page boundaries so that
2254          * dealing with the page cache is a little smoother.
2255          */
2256         policy->l_extent.start -= policy->l_extent.start & ~PAGE_MASK;
2257         policy->l_extent.end |= ~PAGE_MASK;
2258
2259         /*
2260          * kms is not valid when either object is completely fresh (so that no
2261          * locks are cached), or object was evicted. In the latter case cached
2262          * lock cannot be used, because it would prime inode state with
2263          * potentially stale LVB.
2264          */
2265         if (!kms_valid)
2266                 goto no_match;
2267
2268         /* Next, search for already existing extent locks that will cover us */
2269         /* If we're trying to read, we also search for an existing PW lock.  The
2270          * VFS and page cache already protect us locally, so lots of readers/
2271          * writers can share a single PW lock.
2272          *
2273          * There are problems with conversion deadlocks, so instead of
2274          * converting a read lock to a write lock, we'll just enqueue a new
2275          * one.
2276          *
2277          * At some point we should cancel the read lock instead of making them
2278          * send us a blocking callback, but there are problems with canceling
2279          * locks out from other users right now, too.
2280          */
2281         mode = einfo->ei_mode;
2282         if (einfo->ei_mode == LCK_PR)
2283                 mode |= LCK_PW;
2284         if (agl == 0)
2285                 match_flags |= LDLM_FL_LVB_READY;
2286         if (intent != 0)
2287                 match_flags |= LDLM_FL_BLOCK_GRANTED;
2288         mode = ldlm_lock_match(obd->obd_namespace, match_flags, res_id,
2289                                einfo->ei_type, policy, mode, &lockh, 0);
2290         if (mode) {
2291                 struct ldlm_lock *matched;
2292
2293                 if (*flags & LDLM_FL_TEST_LOCK)
2294                         return ELDLM_OK;
2295
2296                 matched = ldlm_handle2lock(&lockh);
2297                 if (agl) {
2298                         /* AGL enqueues DLM locks speculatively. Therefore if
2299                          * it already exists a DLM lock, it wll just inform the
2300                          * caller to cancel the AGL process for this stripe.
2301                          */
2302                         ldlm_lock_decref(&lockh, mode);
2303                         LDLM_LOCK_PUT(matched);
2304                         return -ECANCELED;
2305                 } else if (osc_set_lock_data_with_check(matched, einfo)) {
2306                         *flags |= LDLM_FL_LVB_READY;
2307                         /* We already have a lock, and it's referenced. */
2308                         (*upcall)(cookie, &lockh, ELDLM_LOCK_MATCHED);
2309
2310                         ldlm_lock_decref(&lockh, mode);
2311                         LDLM_LOCK_PUT(matched);
2312                         return ELDLM_OK;
2313                 } else {
2314                         ldlm_lock_decref(&lockh, mode);
2315                         LDLM_LOCK_PUT(matched);
2316                 }
2317         }
2318
2319 no_match:
2320         if (*flags & LDLM_FL_TEST_LOCK)
2321                 return -ENOLCK;
2322         if (intent) {
2323                 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2324                                            &RQF_LDLM_ENQUEUE_LVB);
2325                 if (!req)
2326                         return -ENOMEM;
2327
2328                 rc = ldlm_prep_enqueue_req(exp, req, NULL, 0);
2329                 if (rc) {
2330                         ptlrpc_request_free(req);
2331                         return rc;
2332                 }
2333
2334                 req_capsule_set_size(&req->rq_pill, &RMF_DLM_LVB, RCL_SERVER,
2335                                      sizeof(*lvb));
2336                 ptlrpc_request_set_replen(req);
2337         }
2338
2339         /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
2340         *flags &= ~LDLM_FL_BLOCK_GRANTED;
2341
2342         rc = ldlm_cli_enqueue(exp, &req, einfo, res_id, policy, flags, lvb,
2343                               sizeof(*lvb), LVB_T_OST, &lockh, async);
2344         if (async) {
2345                 if (!rc) {
2346                         struct osc_enqueue_args *aa;
2347
2348                         CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2349                         aa = ptlrpc_req_async_args(req);
2350                         aa->oa_exp = exp;
2351                         aa->oa_mode = einfo->ei_mode;
2352                         aa->oa_type = einfo->ei_type;
2353                         lustre_handle_copy(&aa->oa_lockh, &lockh);
2354                         aa->oa_upcall = upcall;
2355                         aa->oa_cookie = cookie;
2356                         aa->oa_agl    = !!agl;
2357                         if (!agl) {
2358                                 aa->oa_flags = flags;
2359                                 aa->oa_lvb = lvb;
2360                         } else {
2361                                 /* AGL is essentially to enqueue an DLM lock
2362                                 * in advance, so we don't care about the
2363                                 * result of AGL enqueue.
2364                                 */
2365                                 aa->oa_lvb = NULL;
2366                                 aa->oa_flags = NULL;
2367                         }
2368
2369                         req->rq_interpret_reply =
2370                                 (ptlrpc_interpterer_t)osc_enqueue_interpret;
2371                         if (rqset == PTLRPCD_SET)
2372                                 ptlrpcd_add_req(req);
2373                         else
2374                                 ptlrpc_set_add_req(rqset, req);
2375                 } else if (intent) {
2376                         ptlrpc_req_finished(req);
2377                 }
2378                 return rc;
2379         }
2380
2381         rc = osc_enqueue_fini(req, upcall, cookie, &lockh, einfo->ei_mode,
2382                               flags, agl, rc);
2383         if (intent)
2384                 ptlrpc_req_finished(req);
2385
2386         return rc;
2387 }
2388
2389 int osc_match_base(struct obd_export *exp, struct ldlm_res_id *res_id,
2390                    __u32 type, ldlm_policy_data_t *policy, __u32 mode,
2391                    __u64 *flags, void *data, struct lustre_handle *lockh,
2392                    int unref)
2393 {
2394         struct obd_device *obd = exp->exp_obd;
2395         __u64 lflags = *flags;
2396         enum ldlm_mode rc;
2397
2398         if (OBD_FAIL_CHECK(OBD_FAIL_OSC_MATCH))
2399                 return -EIO;
2400
2401         /* Filesystem lock extents are extended to page boundaries so that
2402          * dealing with the page cache is a little smoother
2403          */
2404         policy->l_extent.start -= policy->l_extent.start & ~PAGE_MASK;
2405         policy->l_extent.end |= ~PAGE_MASK;
2406
2407         /* Next, search for already existing extent locks that will cover us */
2408         /* If we're trying to read, we also search for an existing PW lock.  The
2409          * VFS and page cache already protect us locally, so lots of readers/
2410          * writers can share a single PW lock.
2411          */
2412         rc = mode;
2413         if (mode == LCK_PR)
2414                 rc |= LCK_PW;
2415         rc = ldlm_lock_match(obd->obd_namespace, lflags,
2416                              res_id, type, policy, rc, lockh, unref);
2417         if (rc) {
2418                 if (data) {
2419                         if (!osc_set_data_with_check(lockh, data)) {
2420                                 if (!(lflags & LDLM_FL_TEST_LOCK))
2421                                         ldlm_lock_decref(lockh, rc);
2422                                 return 0;
2423                         }
2424                 }
2425                 if (!(lflags & LDLM_FL_TEST_LOCK) && mode != rc) {
2426                         ldlm_lock_addref(lockh, LCK_PR);
2427                         ldlm_lock_decref(lockh, LCK_PW);
2428                 }
2429                 return rc;
2430         }
2431         return rc;
2432 }
2433
2434 int osc_cancel_base(struct lustre_handle *lockh, __u32 mode)
2435 {
2436         if (unlikely(mode == LCK_GROUP))
2437                 ldlm_lock_decref_and_cancel(lockh, mode);
2438         else
2439                 ldlm_lock_decref(lockh, mode);
2440
2441         return 0;
2442 }
2443
2444 static int osc_statfs_interpret(const struct lu_env *env,
2445                                 struct ptlrpc_request *req,
2446                                 struct osc_async_args *aa, int rc)
2447 {
2448         struct obd_statfs *msfs;
2449
2450         if (rc == -EBADR)
2451                 /* The request has in fact never been sent
2452                  * due to issues at a higher level (LOV).
2453                  * Exit immediately since the caller is
2454                  * aware of the problem and takes care
2455                  * of the clean up
2456                  */
2457                 return rc;
2458
2459         if ((rc == -ENOTCONN || rc == -EAGAIN) &&
2460             (aa->aa_oi->oi_flags & OBD_STATFS_NODELAY)) {
2461                 rc = 0;
2462                 goto out;
2463         }
2464
2465         if (rc != 0)
2466                 goto out;
2467
2468         msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2469         if (!msfs) {
2470                 rc = -EPROTO;
2471                 goto out;
2472         }
2473
2474         *aa->aa_oi->oi_osfs = *msfs;
2475 out:
2476         rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
2477         return rc;
2478 }
2479
2480 static int osc_statfs_async(struct obd_export *exp,
2481                             struct obd_info *oinfo, __u64 max_age,
2482                             struct ptlrpc_request_set *rqset)
2483 {
2484         struct obd_device *obd = class_exp2obd(exp);
2485         struct ptlrpc_request *req;
2486         struct osc_async_args *aa;
2487         int rc;
2488
2489         /* We could possibly pass max_age in the request (as an absolute
2490          * timestamp or a "seconds.usec ago") so the target can avoid doing
2491          * extra calls into the filesystem if that isn't necessary (e.g.
2492          * during mount that would help a bit).  Having relative timestamps
2493          * is not so great if request processing is slow, while absolute
2494          * timestamps are not ideal because they need time synchronization.
2495          */
2496         req = ptlrpc_request_alloc(obd->u.cli.cl_import, &RQF_OST_STATFS);
2497         if (!req)
2498                 return -ENOMEM;
2499
2500         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2501         if (rc) {
2502                 ptlrpc_request_free(req);
2503                 return rc;
2504         }
2505         ptlrpc_request_set_replen(req);
2506         req->rq_request_portal = OST_CREATE_PORTAL;
2507         ptlrpc_at_set_req_timeout(req);
2508
2509         if (oinfo->oi_flags & OBD_STATFS_NODELAY) {
2510                 /* procfs requests not want stat in wait for avoid deadlock */
2511                 req->rq_no_resend = 1;
2512                 req->rq_no_delay = 1;
2513         }
2514
2515         req->rq_interpret_reply = (ptlrpc_interpterer_t)osc_statfs_interpret;
2516         CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2517         aa = ptlrpc_req_async_args(req);
2518         aa->aa_oi = oinfo;
2519
2520         ptlrpc_set_add_req(rqset, req);
2521         return 0;
2522 }
2523
2524 static int osc_statfs(const struct lu_env *env, struct obd_export *exp,
2525                       struct obd_statfs *osfs, __u64 max_age, __u32 flags)
2526 {
2527         struct obd_device *obd = class_exp2obd(exp);
2528         struct obd_statfs *msfs;
2529         struct ptlrpc_request *req;
2530         struct obd_import *imp = NULL;
2531         int rc;
2532
2533         /* Since the request might also come from lprocfs, so we need
2534          * sync this with client_disconnect_export Bug15684
2535          */
2536         down_read(&obd->u.cli.cl_sem);
2537         if (obd->u.cli.cl_import)
2538                 imp = class_import_get(obd->u.cli.cl_import);
2539         up_read(&obd->u.cli.cl_sem);
2540         if (!imp)
2541                 return -ENODEV;
2542
2543         /* We could possibly pass max_age in the request (as an absolute
2544          * timestamp or a "seconds.usec ago") so the target can avoid doing
2545          * extra calls into the filesystem if that isn't necessary (e.g.
2546          * during mount that would help a bit).  Having relative timestamps
2547          * is not so great if request processing is slow, while absolute
2548          * timestamps are not ideal because they need time synchronization.
2549          */
2550         req = ptlrpc_request_alloc(imp, &RQF_OST_STATFS);
2551
2552         class_import_put(imp);
2553
2554         if (!req)
2555                 return -ENOMEM;
2556
2557         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_STATFS);
2558         if (rc) {
2559                 ptlrpc_request_free(req);
2560                 return rc;
2561         }
2562         ptlrpc_request_set_replen(req);
2563         req->rq_request_portal = OST_CREATE_PORTAL;
2564         ptlrpc_at_set_req_timeout(req);
2565
2566         if (flags & OBD_STATFS_NODELAY) {
2567                 /* procfs requests not want stat in wait for avoid deadlock */
2568                 req->rq_no_resend = 1;
2569                 req->rq_no_delay = 1;
2570         }
2571
2572         rc = ptlrpc_queue_wait(req);
2573         if (rc)
2574                 goto out;
2575
2576         msfs = req_capsule_server_get(&req->rq_pill, &RMF_OBD_STATFS);
2577         if (!msfs) {
2578                 rc = -EPROTO;
2579                 goto out;
2580         }
2581
2582         *osfs = *msfs;
2583
2584  out:
2585         ptlrpc_req_finished(req);
2586         return rc;
2587 }
2588
2589 /* Retrieve object striping information.
2590  *
2591  * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
2592  * the maximum number of OST indices which will fit in the user buffer.
2593  * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
2594  */
2595 static int osc_getstripe(struct lov_stripe_md *lsm,
2596                          struct lov_user_md __user *lump)
2597 {
2598         /* we use lov_user_md_v3 because it is larger than lov_user_md_v1 */
2599         struct lov_user_md_v3 lum, *lumk;
2600         struct lov_user_ost_data_v1 *lmm_objects;
2601         int rc = 0, lum_size;
2602
2603         if (!lsm)
2604                 return -ENODATA;
2605
2606         /* we only need the header part from user space to get lmm_magic and
2607          * lmm_stripe_count, (the header part is common to v1 and v3)
2608          */
2609         lum_size = sizeof(struct lov_user_md_v1);
2610         if (copy_from_user(&lum, lump, lum_size))
2611                 return -EFAULT;
2612
2613         if ((lum.lmm_magic != LOV_USER_MAGIC_V1) &&
2614             (lum.lmm_magic != LOV_USER_MAGIC_V3))
2615                 return -EINVAL;
2616
2617         /* lov_user_md_vX and lov_mds_md_vX must have the same size */
2618         LASSERT(sizeof(struct lov_user_md_v1) == sizeof(struct lov_mds_md_v1));
2619         LASSERT(sizeof(struct lov_user_md_v3) == sizeof(struct lov_mds_md_v3));
2620         LASSERT(sizeof(lum.lmm_objects[0]) == sizeof(lumk->lmm_objects[0]));
2621
2622         /* we can use lov_mds_md_size() to compute lum_size
2623          * because lov_user_md_vX and lov_mds_md_vX have the same size
2624          */
2625         if (lum.lmm_stripe_count > 0) {
2626                 lum_size = lov_mds_md_size(lum.lmm_stripe_count, lum.lmm_magic);
2627                 lumk = kzalloc(lum_size, GFP_NOFS);
2628                 if (!lumk)
2629                         return -ENOMEM;
2630
2631                 if (lum.lmm_magic == LOV_USER_MAGIC_V1)
2632                         lmm_objects =
2633                             &(((struct lov_user_md_v1 *)lumk)->lmm_objects[0]);
2634                 else
2635                         lmm_objects = &(lumk->lmm_objects[0]);
2636                 lmm_objects->l_ost_oi = lsm->lsm_oi;
2637         } else {
2638                 lum_size = lov_mds_md_size(0, lum.lmm_magic);
2639                 lumk = &lum;
2640         }
2641
2642         lumk->lmm_oi = lsm->lsm_oi;
2643         lumk->lmm_stripe_count = 1;
2644
2645         if (copy_to_user(lump, lumk, lum_size))
2646                 rc = -EFAULT;
2647
2648         if (lumk != &lum)
2649                 kfree(lumk);
2650
2651         return rc;
2652 }
2653
2654 static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
2655                          void *karg, void __user *uarg)
2656 {
2657         struct obd_device *obd = exp->exp_obd;
2658         struct obd_ioctl_data *data = karg;
2659         int err = 0;
2660
2661         if (!try_module_get(THIS_MODULE)) {
2662                 CERROR("%s: cannot get module '%s'\n", obd->obd_name,
2663                        module_name(THIS_MODULE));
2664                 return -EINVAL;
2665         }
2666         switch (cmd) {
2667         case OBD_IOC_LOV_GET_CONFIG: {
2668                 char *buf;
2669                 struct lov_desc *desc;
2670                 struct obd_uuid uuid;
2671
2672                 buf = NULL;
2673                 len = 0;
2674                 if (obd_ioctl_getdata(&buf, &len, uarg)) {
2675                         err = -EINVAL;
2676                         goto out;
2677                 }
2678
2679                 data = (struct obd_ioctl_data *)buf;
2680
2681                 if (sizeof(*desc) > data->ioc_inllen1) {
2682                         obd_ioctl_freedata(buf, len);
2683                         err = -EINVAL;
2684                         goto out;
2685                 }
2686
2687                 if (data->ioc_inllen2 < sizeof(uuid)) {
2688                         obd_ioctl_freedata(buf, len);
2689                         err = -EINVAL;
2690                         goto out;
2691                 }
2692
2693                 desc = (struct lov_desc *)data->ioc_inlbuf1;
2694                 desc->ld_tgt_count = 1;
2695                 desc->ld_active_tgt_count = 1;
2696                 desc->ld_default_stripe_count = 1;
2697                 desc->ld_default_stripe_size = 0;
2698                 desc->ld_default_stripe_offset = 0;
2699                 desc->ld_pattern = 0;
2700                 memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));
2701
2702                 memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));
2703
2704                 err = copy_to_user(uarg, buf, len);
2705                 if (err)
2706                         err = -EFAULT;
2707                 obd_ioctl_freedata(buf, len);
2708                 goto out;
2709         }
2710         case LL_IOC_LOV_SETSTRIPE:
2711                 err = obd_alloc_memmd(exp, karg);
2712                 if (err > 0)
2713                         err = 0;
2714                 goto out;
2715         case LL_IOC_LOV_GETSTRIPE:
2716                 err = osc_getstripe(karg, uarg);
2717                 goto out;
2718         case OBD_IOC_CLIENT_RECOVER:
2719                 err = ptlrpc_recover_import(obd->u.cli.cl_import,
2720                                             data->ioc_inlbuf1, 0);
2721                 if (err > 0)
2722                         err = 0;
2723                 goto out;
2724         case IOC_OSC_SET_ACTIVE:
2725                 err = ptlrpc_set_import_active(obd->u.cli.cl_import,
2726                                                data->ioc_offset);
2727                 goto out;
2728         case OBD_IOC_POLL_QUOTACHECK:
2729                 err = osc_quota_poll_check(exp, karg);
2730                 goto out;
2731         case OBD_IOC_PING_TARGET:
2732                 err = ptlrpc_obd_ping(obd);
2733                 goto out;
2734         default:
2735                 CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
2736                        cmd, current_comm());
2737                 err = -ENOTTY;
2738                 goto out;
2739         }
2740 out:
2741         module_put(THIS_MODULE);
2742         return err;
2743 }
2744
2745 static int osc_get_info(const struct lu_env *env, struct obd_export *exp,
2746                         u32 keylen, void *key, __u32 *vallen, void *val,
2747                         struct lov_stripe_md *lsm)
2748 {
2749         if (!vallen || !val)
2750                 return -EFAULT;
2751
2752         if (KEY_IS(KEY_LOCK_TO_STRIPE)) {
2753                 __u32 *stripe = val;
2754                 *vallen = sizeof(*stripe);
2755                 *stripe = 0;
2756                 return 0;
2757         } else if (KEY_IS(KEY_LAST_ID)) {
2758                 struct ptlrpc_request *req;
2759                 u64 *reply;
2760                 char *tmp;
2761                 int rc;
2762
2763                 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2764                                            &RQF_OST_GET_INFO_LAST_ID);
2765                 if (!req)
2766                         return -ENOMEM;
2767
2768                 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
2769                                      RCL_CLIENT, keylen);
2770                 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
2771                 if (rc) {
2772                         ptlrpc_request_free(req);
2773                         return rc;
2774                 }
2775
2776                 tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
2777                 memcpy(tmp, key, keylen);
2778
2779                 req->rq_no_delay = 1;
2780                 req->rq_no_resend = 1;
2781                 ptlrpc_request_set_replen(req);
2782                 rc = ptlrpc_queue_wait(req);
2783                 if (rc)
2784                         goto out;
2785
2786                 reply = req_capsule_server_get(&req->rq_pill, &RMF_OBD_ID);
2787                 if (!reply) {
2788                         rc = -EPROTO;
2789                         goto out;
2790                 }
2791
2792                 *((u64 *)val) = *reply;
2793 out:
2794                 ptlrpc_req_finished(req);
2795                 return rc;
2796         } else if (KEY_IS(KEY_FIEMAP)) {
2797                 struct ll_fiemap_info_key *fm_key = key;
2798                 struct ldlm_res_id res_id;
2799                 ldlm_policy_data_t policy;
2800                 struct lustre_handle lockh;
2801                 enum ldlm_mode mode = 0;
2802                 struct ptlrpc_request *req;
2803                 struct ll_user_fiemap *reply;
2804                 char *tmp;
2805                 int rc;
2806
2807                 if (!(fm_key->fiemap.fm_flags & FIEMAP_FLAG_SYNC))
2808                         goto skip_locking;
2809
2810                 policy.l_extent.start = fm_key->fiemap.fm_start &
2811                                                 PAGE_MASK;
2812
2813                 if (OBD_OBJECT_EOF - fm_key->fiemap.fm_length <=
2814                     fm_key->fiemap.fm_start + PAGE_SIZE - 1)
2815                         policy.l_extent.end = OBD_OBJECT_EOF;
2816                 else
2817                         policy.l_extent.end = (fm_key->fiemap.fm_start +
2818                                 fm_key->fiemap.fm_length +
2819                                 PAGE_SIZE - 1) & PAGE_MASK;
2820
2821                 ostid_build_res_name(&fm_key->oa.o_oi, &res_id);
2822                 mode = ldlm_lock_match(exp->exp_obd->obd_namespace,
2823                                        LDLM_FL_BLOCK_GRANTED |
2824                                        LDLM_FL_LVB_READY,
2825                                        &res_id, LDLM_EXTENT, &policy,
2826                                        LCK_PR | LCK_PW, &lockh, 0);
2827                 if (mode) { /* lock is cached on client */
2828                         if (mode != LCK_PR) {
2829                                 ldlm_lock_addref(&lockh, LCK_PR);
2830                                 ldlm_lock_decref(&lockh, LCK_PW);
2831                         }
2832                 } else { /* no cached lock, needs acquire lock on server side */
2833                         fm_key->oa.o_valid |= OBD_MD_FLFLAGS;
2834                         fm_key->oa.o_flags |= OBD_FL_SRVLOCK;
2835                 }
2836
2837 skip_locking:
2838                 req = ptlrpc_request_alloc(class_exp2cliimp(exp),
2839                                            &RQF_OST_GET_INFO_FIEMAP);
2840                 if (!req) {
2841                         rc = -ENOMEM;
2842                         goto drop_lock;
2843                 }
2844
2845                 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_KEY,
2846                                      RCL_CLIENT, keylen);
2847                 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
2848                                      RCL_CLIENT, *vallen);
2849                 req_capsule_set_size(&req->rq_pill, &RMF_FIEMAP_VAL,
2850                                      RCL_SERVER, *vallen);
2851
2852                 rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_GET_INFO);
2853                 if (rc) {
2854                         ptlrpc_request_free(req);
2855                         goto drop_lock;
2856                 }
2857
2858                 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_KEY);
2859                 memcpy(tmp, key, keylen);
2860                 tmp = req_capsule_client_get(&req->rq_pill, &RMF_FIEMAP_VAL);
2861                 memcpy(tmp, val, *vallen);
2862
2863                 ptlrpc_request_set_replen(req);
2864                 rc = ptlrpc_queue_wait(req);
2865                 if (rc)
2866                         goto fini_req;
2867
2868                 reply = req_capsule_server_get(&req->rq_pill, &RMF_FIEMAP_VAL);
2869                 if (!reply) {
2870                         rc = -EPROTO;
2871                         goto fini_req;
2872                 }
2873
2874                 memcpy(val, reply, *vallen);
2875 fini_req:
2876                 ptlrpc_req_finished(req);
2877 drop_lock:
2878                 if (mode)
2879                         ldlm_lock_decref(&lockh, LCK_PR);
2880                 return rc;
2881         }
2882
2883         return -EINVAL;
2884 }
2885
2886 static int osc_set_info_async(const struct lu_env *env, struct obd_export *exp,
2887                               u32 keylen, void *key, u32 vallen,
2888                               void *val, struct ptlrpc_request_set *set)
2889 {
2890         struct ptlrpc_request *req;
2891         struct obd_device *obd = exp->exp_obd;
2892         struct obd_import *imp = class_exp2cliimp(exp);
2893         char *tmp;
2894         int rc;
2895
2896         OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);
2897
2898         if (KEY_IS(KEY_CHECKSUM)) {
2899                 if (vallen != sizeof(int))
2900                         return -EINVAL;
2901                 exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
2902                 return 0;
2903         }
2904
2905         if (KEY_IS(KEY_SPTLRPC_CONF)) {
2906                 sptlrpc_conf_client_adapt(obd);
2907                 return 0;
2908         }
2909
2910         if (KEY_IS(KEY_FLUSH_CTX)) {
2911                 sptlrpc_import_flush_my_ctx(imp);
2912                 return 0;
2913         }
2914
2915         if (KEY_IS(KEY_CACHE_SET)) {
2916                 struct client_obd *cli = &obd->u.cli;
2917
2918                 LASSERT(!cli->cl_cache); /* only once */
2919                 cli->cl_cache = val;
2920                 cl_cache_incref(cli->cl_cache);
2921                 cli->cl_lru_left = &cli->cl_cache->ccc_lru_left;
2922
2923                 /* add this osc into entity list */
2924                 LASSERT(list_empty(&cli->cl_lru_osc));
2925                 spin_lock(&cli->cl_cache->ccc_lru_lock);
2926                 list_add(&cli->cl_lru_osc, &cli->cl_cache->ccc_lru);
2927                 spin_unlock(&cli->cl_cache->ccc_lru_lock);
2928
2929                 return 0;
2930         }
2931
2932         if (KEY_IS(KEY_CACHE_LRU_SHRINK)) {
2933                 struct client_obd *cli = &obd->u.cli;
2934                 int nr = atomic_read(&cli->cl_lru_in_list) >> 1;
2935                 int target = *(int *)val;
2936
2937                 nr = osc_lru_shrink(env, cli, min(nr, target), true);
2938                 *(int *)val -= nr;
2939                 return 0;
2940         }
2941
2942         if (!set && !KEY_IS(KEY_GRANT_SHRINK))
2943                 return -EINVAL;
2944
2945         /* We pass all other commands directly to OST. Since nobody calls osc
2946          * methods directly and everybody is supposed to go through LOV, we
2947          * assume lov checked invalid values for us.
2948          * The only recognised values so far are evict_by_nid and mds_conn.
2949          * Even if something bad goes through, we'd get a -EINVAL from OST
2950          * anyway.
2951          */
2952
2953         req = ptlrpc_request_alloc(imp, KEY_IS(KEY_GRANT_SHRINK) ?
2954                                                 &RQF_OST_SET_GRANT_INFO :
2955                                                 &RQF_OBD_SET_INFO);
2956         if (!req)
2957                 return -ENOMEM;
2958
2959         req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_KEY,
2960                              RCL_CLIENT, keylen);
2961         if (!KEY_IS(KEY_GRANT_SHRINK))
2962                 req_capsule_set_size(&req->rq_pill, &RMF_SETINFO_VAL,
2963                                      RCL_CLIENT, vallen);
2964         rc = ptlrpc_request_pack(req, LUSTRE_OST_VERSION, OST_SET_INFO);
2965         if (rc) {
2966                 ptlrpc_request_free(req);
2967                 return rc;
2968         }
2969
2970         tmp = req_capsule_client_get(&req->rq_pill, &RMF_SETINFO_KEY);
2971         memcpy(tmp, key, keylen);
2972         tmp = req_capsule_client_get(&req->rq_pill, KEY_IS(KEY_GRANT_SHRINK) ?
2973                                                         &RMF_OST_BODY :
2974                                                         &RMF_SETINFO_VAL);
2975         memcpy(tmp, val, vallen);
2976
2977         if (KEY_IS(KEY_GRANT_SHRINK)) {
2978                 struct osc_brw_async_args *aa;
2979                 struct obdo *oa;
2980
2981                 CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
2982                 aa = ptlrpc_req_async_args(req);
2983                 oa = kmem_cache_zalloc(obdo_cachep, GFP_NOFS);
2984                 if (!oa) {
2985                         ptlrpc_req_finished(req);
2986                         return -ENOMEM;
2987                 }
2988                 *oa = ((struct ost_body *)val)->oa;
2989                 aa->aa_oa = oa;
2990                 req->rq_interpret_reply = osc_shrink_grant_interpret;
2991         }
2992
2993         ptlrpc_request_set_replen(req);
2994         if (!KEY_IS(KEY_GRANT_SHRINK)) {
2995                 LASSERT(set);
2996                 ptlrpc_set_add_req(set, req);
2997                 ptlrpc_check_set(NULL, set);
2998         } else {
2999                 ptlrpcd_add_req(req);
3000         }
3001
3002         return 0;
3003 }
3004
3005 static int osc_reconnect(const struct lu_env *env,
3006                          struct obd_export *exp, struct obd_device *obd,
3007                          struct obd_uuid *cluuid,
3008                          struct obd_connect_data *data,
3009                          void *localdata)
3010 {
3011         struct client_obd *cli = &obd->u.cli;
3012
3013         if (data && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
3014                 long lost_grant;
3015
3016                 spin_lock(&cli->cl_loi_list_lock);
3017                 data->ocd_grant = (cli->cl_avail_grant + cli->cl_dirty) ?:
3018                                 2 * cli_brw_size(obd);
3019                 lost_grant = cli->cl_lost_grant;
3020                 cli->cl_lost_grant = 0;
3021                 spin_unlock(&cli->cl_loi_list_lock);
3022
3023                 CDEBUG(D_RPCTRACE, "ocd_connect_flags: %#llx ocd_version: %d ocd_grant: %d, lost: %ld.\n",
3024                        data->ocd_connect_flags,
3025                        data->ocd_version, data->ocd_grant, lost_grant);
3026         }
3027
3028         return 0;
3029 }
3030
3031 static int osc_disconnect(struct obd_export *exp)
3032 {
3033         struct obd_device *obd = class_exp2obd(exp);
3034         int rc;
3035
3036         rc = client_disconnect_export(exp);
3037         /**
3038          * Initially we put del_shrink_grant before disconnect_export, but it
3039          * causes the following problem if setup (connect) and cleanup
3040          * (disconnect) are tangled together.
3041          *      connect p1                   disconnect p2
3042          *   ptlrpc_connect_import
3043          *     ...............         class_manual_cleanup
3044          *                                   osc_disconnect
3045          *                                   del_shrink_grant
3046          *   ptlrpc_connect_interrupt
3047          *     init_grant_shrink
3048          *   add this client to shrink list
3049          *                                    cleanup_osc
3050          * Bang! pinger trigger the shrink.
3051          * So the osc should be disconnected from the shrink list, after we
3052          * are sure the import has been destroyed. BUG18662
3053          */
3054         if (!obd->u.cli.cl_import)
3055                 osc_del_shrink_grant(&obd->u.cli);
3056         return rc;
3057 }
3058
3059 static int osc_import_event(struct obd_device *obd,
3060                             struct obd_import *imp,
3061                             enum obd_import_event event)
3062 {
3063         struct client_obd *cli;
3064         int rc = 0;
3065
3066         LASSERT(imp->imp_obd == obd);
3067
3068         switch (event) {
3069         case IMP_EVENT_DISCON: {
3070                 cli = &obd->u.cli;
3071                 spin_lock(&cli->cl_loi_list_lock);
3072                 cli->cl_avail_grant = 0;
3073                 cli->cl_lost_grant = 0;
3074                 spin_unlock(&cli->cl_loi_list_lock);
3075                 break;
3076         }
3077         case IMP_EVENT_INACTIVE: {
3078                 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
3079                 break;
3080         }
3081         case IMP_EVENT_INVALIDATE: {
3082                 struct ldlm_namespace *ns = obd->obd_namespace;
3083                 struct lu_env *env;
3084                 int refcheck;
3085
3086                 env = cl_env_get(&refcheck);
3087                 if (!IS_ERR(env)) {
3088                         /* Reset grants */
3089                         cli = &obd->u.cli;
3090                         /* all pages go to failing rpcs due to the invalid
3091                          * import
3092                          */
3093                         osc_io_unplug(env, cli, NULL);
3094
3095                         ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);
3096                         cl_env_put(env, &refcheck);
3097                 } else {
3098                         rc = PTR_ERR(env);
3099                 }
3100                 break;
3101         }
3102         case IMP_EVENT_ACTIVE: {
3103                 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
3104                 break;
3105         }
3106         case IMP_EVENT_OCD: {
3107                 struct obd_connect_data *ocd = &imp->imp_connect_data;
3108
3109                 if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
3110                         osc_init_grant(&obd->u.cli, ocd);
3111
3112                 /* See bug 7198 */
3113                 if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
3114                         imp->imp_client->cli_request_portal = OST_REQUEST_PORTAL;
3115
3116                 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
3117                 break;
3118         }
3119         case IMP_EVENT_DEACTIVATE: {
3120                 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_DEACTIVATE, NULL);
3121                 break;
3122         }
3123         case IMP_EVENT_ACTIVATE: {
3124                 rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVATE, NULL);
3125                 break;
3126         }
3127         default:
3128                 CERROR("Unknown import event %d\n", event);
3129                 LBUG();
3130         }
3131         return rc;
3132 }
3133
3134 /**
3135  * Determine whether the lock can be canceled before replaying the lock
3136  * during recovery, see bug16774 for detailed information.
3137  *
3138  * \retval zero the lock can't be canceled
3139  * \retval other ok to cancel
3140  */
3141 static int osc_cancel_weight(struct ldlm_lock *lock)
3142 {
3143         /*
3144          * Cancel all unused and granted extent lock.
3145          */
3146         if (lock->l_resource->lr_type == LDLM_EXTENT &&
3147             lock->l_granted_mode == lock->l_req_mode &&
3148             osc_ldlm_weigh_ast(lock) == 0)
3149                 return 1;
3150
3151         return 0;
3152 }
3153
3154 static int brw_queue_work(const struct lu_env *env, void *data)
3155 {
3156         struct client_obd *cli = data;
3157
3158         CDEBUG(D_CACHE, "Run writeback work for client obd %p.\n", cli);
3159
3160         osc_io_unplug(env, cli, NULL);
3161         return 0;
3162 }
3163
3164 int osc_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
3165 {
3166         struct lprocfs_static_vars lvars = { NULL };
3167         struct client_obd *cli = &obd->u.cli;
3168         void *handler;
3169         int rc;
3170         int adding;
3171         int added;
3172         int req_count;
3173
3174         rc = ptlrpcd_addref();
3175         if (rc)
3176                 return rc;
3177
3178         rc = client_obd_setup(obd, lcfg);
3179         if (rc)
3180                 goto out_ptlrpcd;
3181
3182         handler = ptlrpcd_alloc_work(cli->cl_import, brw_queue_work, cli);
3183         if (IS_ERR(handler)) {
3184                 rc = PTR_ERR(handler);
3185                 goto out_client_setup;
3186         }
3187         cli->cl_writeback_work = handler;
3188
3189         handler = ptlrpcd_alloc_work(cli->cl_import, lru_queue_work, cli);
3190         if (IS_ERR(handler)) {
3191                 rc = PTR_ERR(handler);
3192                 goto out_ptlrpcd_work;
3193         }
3194
3195         cli->cl_lru_work = handler;
3196
3197         rc = osc_quota_setup(obd);
3198         if (rc)
3199                 goto out_ptlrpcd_work;
3200
3201         cli->cl_grant_shrink_interval = GRANT_SHRINK_INTERVAL;
3202         lprocfs_osc_init_vars(&lvars);
3203         if (lprocfs_obd_setup(obd, lvars.obd_vars, lvars.sysfs_vars) == 0) {
3204                 lproc_osc_attach_seqstat(obd);
3205                 sptlrpc_lprocfs_cliobd_attach(obd);
3206                 ptlrpc_lprocfs_register_obd(obd);
3207         }
3208
3209         /*
3210          * We try to control the total number of requests with a upper limit
3211          * osc_reqpool_maxreqcount. There might be some race which will cause
3212          * over-limit allocation, but it is fine.
3213          */
3214         req_count = atomic_read(&osc_pool_req_count);
3215         if (req_count < osc_reqpool_maxreqcount) {
3216                 adding = cli->cl_max_rpcs_in_flight + 2;
3217                 if (req_count + adding > osc_reqpool_maxreqcount)
3218                         adding = osc_reqpool_maxreqcount - req_count;
3219
3220                 added = ptlrpc_add_rqs_to_pool(osc_rq_pool, adding);
3221                 atomic_add(added, &osc_pool_req_count);
3222         }
3223
3224         INIT_LIST_HEAD(&cli->cl_grant_shrink_list);
3225         ns_register_cancel(obd->obd_namespace, osc_cancel_weight);
3226         return rc;
3227
3228 out_ptlrpcd_work:
3229         if (cli->cl_writeback_work) {
3230                 ptlrpcd_destroy_work(cli->cl_writeback_work);
3231                 cli->cl_writeback_work = NULL;
3232         }
3233         if (cli->cl_lru_work) {
3234                 ptlrpcd_destroy_work(cli->cl_lru_work);
3235                 cli->cl_lru_work = NULL;
3236         }
3237 out_client_setup:
3238         client_obd_cleanup(obd);
3239 out_ptlrpcd:
3240         ptlrpcd_decref();
3241         return rc;
3242 }
3243
3244 static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
3245 {
3246         switch (stage) {
3247         case OBD_CLEANUP_EARLY: {
3248                 struct obd_import *imp;
3249
3250                 imp = obd->u.cli.cl_import;
3251                 CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
3252                 /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
3253                 ptlrpc_deactivate_import(imp);
3254                 spin_lock(&imp->imp_lock);
3255                 imp->imp_pingable = 0;
3256                 spin_unlock(&imp->imp_lock);
3257                 break;
3258         }
3259         case OBD_CLEANUP_EXPORTS: {
3260                 struct client_obd *cli = &obd->u.cli;
3261                 /* LU-464
3262                  * for echo client, export may be on zombie list, wait for
3263                  * zombie thread to cull it, because cli.cl_import will be
3264                  * cleared in client_disconnect_export():
3265                  *   class_export_destroy() -> obd_cleanup() ->
3266                  *   echo_device_free() -> echo_client_cleanup() ->
3267                  *   obd_disconnect() -> osc_disconnect() ->
3268                  *   client_disconnect_export()
3269                  */
3270                 obd_zombie_barrier();
3271                 if (cli->cl_writeback_work) {
3272                         ptlrpcd_destroy_work(cli->cl_writeback_work);
3273                         cli->cl_writeback_work = NULL;
3274                 }
3275                 if (cli->cl_lru_work) {
3276                         ptlrpcd_destroy_work(cli->cl_lru_work);
3277                         cli->cl_lru_work = NULL;
3278                 }
3279                 obd_cleanup_client_import(obd);
3280                 ptlrpc_lprocfs_unregister_obd(obd);
3281                 lprocfs_obd_cleanup(obd);
3282                 break;
3283                 }
3284         }
3285         return 0;
3286 }
3287
3288 static int osc_cleanup(struct obd_device *obd)
3289 {
3290         struct client_obd *cli = &obd->u.cli;
3291         int rc;
3292
3293         /* lru cleanup */
3294         if (cli->cl_cache) {
3295                 LASSERT(atomic_read(&cli->cl_cache->ccc_users) > 0);
3296                 spin_lock(&cli->cl_cache->ccc_lru_lock);
3297                 list_del_init(&cli->cl_lru_osc);
3298                 spin_unlock(&cli->cl_cache->ccc_lru_lock);
3299                 cli->cl_lru_left = NULL;
3300                 cl_cache_decref(cli->cl_cache);
3301                 cli->cl_cache = NULL;
3302         }
3303
3304         /* free memory of osc quota cache */
3305         osc_quota_cleanup(obd);
3306
3307         rc = client_obd_cleanup(obd);
3308
3309         ptlrpcd_decref();
3310         return rc;
3311 }
3312
3313 int osc_process_config_base(struct obd_device *obd, struct lustre_cfg *lcfg)
3314 {
3315         struct lprocfs_static_vars lvars = { NULL };
3316         int rc = 0;
3317
3318         lprocfs_osc_init_vars(&lvars);
3319
3320         switch (lcfg->lcfg_command) {
3321         default:
3322                 rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars,
3323                                               lcfg, obd);
3324                 if (rc > 0)
3325                         rc = 0;
3326                 break;
3327         }
3328
3329         return rc;
3330 }
3331
3332 static int osc_process_config(struct obd_device *obd, u32 len, void *buf)
3333 {
3334         return osc_process_config_base(obd, buf);
3335 }
3336
3337 static struct obd_ops osc_obd_ops = {
3338         .owner          = THIS_MODULE,
3339         .setup          = osc_setup,
3340         .precleanup     = osc_precleanup,
3341         .cleanup        = osc_cleanup,
3342         .add_conn       = client_import_add_conn,
3343         .del_conn       = client_import_del_conn,
3344         .connect        = client_connect_import,
3345         .reconnect      = osc_reconnect,
3346         .disconnect     = osc_disconnect,
3347         .statfs         = osc_statfs,
3348         .statfs_async   = osc_statfs_async,
3349         .packmd         = osc_packmd,
3350         .unpackmd       = osc_unpackmd,
3351         .create         = osc_create,
3352         .destroy        = osc_destroy,
3353         .getattr        = osc_getattr,
3354         .getattr_async  = osc_getattr_async,
3355         .setattr        = osc_setattr,
3356         .setattr_async  = osc_setattr_async,
3357         .find_cbdata    = osc_find_cbdata,
3358         .iocontrol      = osc_iocontrol,
3359         .get_info       = osc_get_info,
3360         .set_info_async = osc_set_info_async,
3361         .import_event   = osc_import_event,
3362         .process_config = osc_process_config,
3363         .quotactl       = osc_quotactl,
3364         .quotacheck     = osc_quotacheck,
3365 };
3366
3367 extern struct lu_kmem_descr osc_caches[];
3368 extern struct lock_class_key osc_ast_guard_class;
3369
3370 static int __init osc_init(void)
3371 {
3372         struct lprocfs_static_vars lvars = { NULL };
3373         unsigned int reqpool_size;
3374         unsigned int reqsize;
3375         int rc;
3376
3377         /* print an address of _any_ initialized kernel symbol from this
3378          * module, to allow debugging with gdb that doesn't support data
3379          * symbols from modules.
3380          */
3381         CDEBUG(D_INFO, "Lustre OSC module (%p).\n", &osc_caches);
3382
3383         rc = lu_kmem_init(osc_caches);
3384         if (rc)
3385                 return rc;
3386
3387         lprocfs_osc_init_vars(&lvars);
3388
3389         rc = class_register_type(&osc_obd_ops, NULL,
3390                                  LUSTRE_OSC_NAME, &osc_device_type);
3391         if (rc)
3392                 goto out_kmem;
3393
3394         /* This is obviously too much memory, only prevent overflow here */
3395         if (osc_reqpool_mem_max >= 1 << 12 || osc_reqpool_mem_max == 0) {
3396                 rc = -EINVAL;
3397                 goto out_type;
3398         }
3399
3400         reqpool_size = osc_reqpool_mem_max << 20;
3401
3402         reqsize = 1;
3403         while (reqsize < OST_MAXREQSIZE)
3404                 reqsize = reqsize << 1;
3405
3406         /*
3407          * We don't enlarge the request count in OSC pool according to
3408          * cl_max_rpcs_in_flight. The allocation from the pool will only be
3409          * tried after normal allocation failed. So a small OSC pool won't
3410          * cause much performance degression in most of cases.
3411          */
3412         osc_reqpool_maxreqcount = reqpool_size / reqsize;
3413
3414         atomic_set(&osc_pool_req_count, 0);
3415         osc_rq_pool = ptlrpc_init_rq_pool(0, OST_MAXREQSIZE,
3416                                           ptlrpc_add_rqs_to_pool);
3417
3418         if (osc_rq_pool)
3419                 return 0;
3420
3421         rc = -ENOMEM;
3422
3423 out_type:
3424         class_unregister_type(LUSTRE_OSC_NAME);
3425 out_kmem:
3426         lu_kmem_fini(osc_caches);
3427         return rc;
3428 }
3429
3430 static void /*__exit*/ osc_exit(void)
3431 {
3432         class_unregister_type(LUSTRE_OSC_NAME);
3433         lu_kmem_fini(osc_caches);
3434         ptlrpc_free_rq_pool(osc_rq_pool);
3435 }
3436
3437 MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>");
3438 MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
3439 MODULE_LICENSE("GPL");
3440 MODULE_VERSION(LUSTRE_VERSION_STRING);
3441
3442 module_init(osc_init);
3443 module_exit(osc_exit);