lustre/mdt/mdt_coordinator.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,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13  * GNU General Public License version 2 for more details.  A copy is
  14  * included in the COPYING file that accompanied this code.
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * along with this program; if not, write to the Free Software
  18  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
  19  *
  20  * GPL HEADER END
  21  */
  22 /*
  23  * Copyright (c) 2012, 2013, Intel Corporation.
  24  * Use is subject to license terms.
  25  * Copyright (c) 2011, 2012 Commissariat a l'energie atomique et aux energies
  26  *                          alternatives
  27  */
  28 /*
  29  * lustre/mdt/mdt_coordinator.c
  30  *
  31  * Lustre HSM Coordinator
  32  *
  33  * Author: Jacques-Charles Lafoucriere <jacques-charles.lafoucriere@cea.fr>
  34  * Author: Aurelien Degremont <aurelien.degremont@cea.fr>
  35  * Author: Thomas Leibovici <thomas.leibovici@cea.fr>
  36  */
  37
  38 #define DEBUG_SUBSYSTEM S_MDS
  39
  40 #include <obd_support.h>
  41 #include <lustre_net.h>
  42 #include <lustre_export.h>
  43 #include <obd.h>
  44 #include <lprocfs_status.h>
  45 #include <lustre_log.h>
  46 #include "mdt_internal.h"
  47
  48 static struct lprocfs_seq_vars lprocfs_mdt_hsm_vars[];
  49
  50 /**
  51  * get obj and HSM attributes on a fid
  52  * \param mti [IN] context
  53  * \param fid [IN] object fid
  54  * \param hsm [OUT] HSM meta data
  55  * \retval obj or error (-ENOENT if not found)
  56  */
  57 struct mdt_object *mdt_hsm_get_md_hsm(struct mdt_thread_info *mti,
  58                                       const struct lu_fid *fid,
  59                                       struct md_hsm *hsm)
  60 {
  61         struct md_attr          *ma;
  62         struct mdt_object       *obj;
  63         int                      rc;
  64         ENTRY;
  65
  66         ma = &mti->mti_attr;
  67         ma->ma_need = MA_HSM;
  68         ma->ma_valid = 0;
  69
  70         /* find object by FID */
  71         obj = mdt_object_find(mti->mti_env, mti->mti_mdt, fid);
  72         if (IS_ERR(obj))
  73                 RETURN(obj);
  74
  75         if (!mdt_object_exists(obj)) {
  76                 /* no more object */
  77                 mdt_object_put(mti->mti_env, obj);
  78                 RETURN(ERR_PTR(-ENOENT));
  79         }
  80
  81         rc = mdt_attr_get_complex(mti, obj, ma);
  82         if (rc) {
  83                 mdt_object_put(mti->mti_env, obj);
  84                 RETURN(ERR_PTR(rc));
  85         }
  86
  87         if (ma->ma_valid & MA_HSM)
  88                 *hsm = ma->ma_hsm;
  89         else
  90                 memset(hsm, 0, sizeof(*hsm));
  91         ma->ma_valid = 0;
  92         RETURN(obj);
  93 }
  94
  95 void mdt_hsm_dump_hal(int level, const char *prefix,
  96                       struct hsm_action_list *hal)
  97 {
  98         int                      i, sz;
  99         struct hsm_action_item  *hai;
 100         char                     buf[12];
 101
 102         CDEBUG(level, "%s: HAL header: version %X count %d compound "LPX64
 103                       " archive_id %d flags "LPX64"\n",
 104                prefix, hal->hal_version, hal->hal_count,
 105                hal->hal_compound_id, hal->hal_archive_id, hal->hal_flags);
 106
 107         hai = hai_first(hal);
 108         for (i = 0; i < hal->hal_count; i++) {
 109                 sz = hai->hai_len - sizeof(*hai);
 110                 CDEBUG(level, "%s %d: fid="DFID" dfid="DFID
 111                        " compound/cookie="LPX64"/"LPX64
 112                        " action=%s extent="LPX64"-"LPX64" gid="LPX64
 113                        " datalen=%d data=[%s]\n",
 114                        prefix, i,
 115                        PFID(&hai->hai_fid), PFID(&hai->hai_dfid),
 116                        hal->hal_compound_id, hai->hai_cookie,
 117                        hsm_copytool_action2name(hai->hai_action),
 118                        hai->hai_extent.offset,
 119                        hai->hai_extent.length,
 120                        hai->hai_gid, sz,
 121                        hai_dump_data_field(hai, buf, sizeof(buf)));
 122                 hai = hai_next(hai);
 123         }
 124 }
 125
 126 /**
 127  * data passed to llog_cat_process() callback
 128  * to scan requests and take actions
 129  */
 130 struct hsm_scan_data {
 131         struct mdt_thread_info          *mti;
 132         char                             fs_name[MTI_NAME_MAXLEN+1];
 133         /* request to be send to agents */
 134         int                              request_sz;    /** allocated size */
 135         int                              max_requests;  /** vector size */
 136         int                              request_cnt;   /** used count */
 137         struct {
 138                 int                      hal_sz;
 139                 int                      hal_used_sz;
 140                 struct hsm_action_list  *hal;
 141         } *request;
 142         /* records to be canceled */
 143         int                              max_cookie;    /** vector size */
 144         int                              cookie_cnt;    /** used count */
 145         __u64                           *cookies;
 146 };
 147
 148 /**
 149  *  llog_cat_process() callback, used to:
 150  *  - find waiting request and start action
 151  *  - purge canceled and done requests
 152  * \param env [IN] environment
 153  * \param llh [IN] llog handle
 154  * \param hdr [IN] llog record
 155  * \param data [IN/OUT] cb data = struct hsm_scan_data
 156  * \retval 0 success
 157  * \retval -ve failure
 158  */
 159 static int mdt_coordinator_cb(const struct lu_env *env,
 160                               struct llog_handle *llh,
 161                               struct llog_rec_hdr *hdr,
 162                               void *data)
 163 {
 164         const struct llog_agent_req_rec *larr;
 165         struct hsm_scan_data            *hsd;
 166         struct hsm_action_item          *hai;
 167         struct mdt_device               *mdt;
 168         struct coordinator              *cdt;
 169         int                              rc;
 170         ENTRY;
 171
 172         hsd = data;
 173         mdt = hsd->mti->mti_mdt;
 174         cdt = &mdt->mdt_coordinator;
 175
 176         larr = (struct llog_agent_req_rec *)hdr;
 177         dump_llog_agent_req_rec("mdt_coordinator_cb(): ", larr);
 178         switch (larr->arr_status) {
 179         case ARS_WAITING: {
 180                 int i, empty_slot, found;
 181
 182                 /* Are agents full? */
 183                 if (atomic_read(&cdt->cdt_request_count) ==
 184                     cdt->cdt_max_requests)
 185                         break;
 186
 187                 /* first search if the request if known in the list we have
 188                  * build and if there is room in the request vector */
 189                 empty_slot = -1;
 190                 found = -1;
 191                 for (i = 0; i < hsd->max_requests &&
 192                             (empty_slot == -1 || found == -1); i++) {
 193                         if (hsd->request[i].hal == NULL) {
 194                                 empty_slot = i;
 195                                 continue;
 196                         }
 197                         if (hsd->request[i].hal->hal_compound_id ==
 198                                 larr->arr_compound_id) {
 199                                 found = i;
 200                                 continue;
 201                         }
 202                 }
 203                 if (found == -1 && empty_slot == -1)
 204                         /* unknown request and no more room for new request,
 205                          * continue scan for to find other entries for
 206                          * already found request
 207                          */
 208                         RETURN(0);
 209
 210                 if (found == -1) {
 211                         struct hsm_action_list *hal;
 212
 213                         /* request is not already known */
 214                         /* allocates hai vector size just needs to be large
 215                          * enough */
 216                         hsd->request[empty_slot].hal_sz =
 217                                      sizeof(*hsd->request[empty_slot].hal) +
 218                                      cfs_size_round(MTI_NAME_MAXLEN+1) +
 219                                      2 * cfs_size_round(larr->arr_hai.hai_len);
 220                         OBD_ALLOC(hal, hsd->request[empty_slot].hal_sz);
 221                         if (!hal) {
 222                                 CERROR("%s: Cannot allocate memory (%d o)"
 223                                        "for compound "LPX64"\n",
 224                                        mdt_obd_name(mdt),
 225                                        hsd->request[i].hal_sz,
 226                                        larr->arr_compound_id);
 227                                 RETURN(-ENOMEM);
 228                         }
 229                         hal->hal_version = HAL_VERSION;
 230                         strncpy(hal->hal_fsname, hsd->fs_name,
 231                                 MTI_NAME_MAXLEN);
 232                         hal->hal_fsname[MTI_NAME_MAXLEN] = '\0';
 233                         hal->hal_compound_id = larr->arr_compound_id;
 234                         hal->hal_archive_id = larr->arr_archive_id;
 235                         hal->hal_flags = larr->arr_flags;
 236                         hal->hal_count = 0;
 237                         hsd->request[empty_slot].hal_used_sz = hal_size(hal);
 238                         hsd->request[empty_slot].hal = hal;
 239                         hsd->request_cnt++;
 240                         found = empty_slot;
 241                         hai = hai_first(hal);
 242                 } else {
 243                         /* request is known */
 244                         /* we check if record archive num is the same as the
 245                          * known request, if not we will serve it in multiple
 246                          * time because we do not know if the agent can serve
 247                          * multiple backend
 248                          * a use case is a compound made of multiple restore
 249                          * where the files are not archived in the same backend
 250                          */
 251                         if (larr->arr_archive_id !=
 252                             hsd->request[found].hal->hal_archive_id)
 253                                 RETURN(0);
 254
 255                         if (hsd->request[found].hal_sz <
 256                             hsd->request[found].hal_used_sz +
 257                              cfs_size_round(larr->arr_hai.hai_len)) {
 258                                 /* Not enough room, need an extension */
 259                                 void *hal_buffer;
 260                                 int sz;
 261
 262                                 sz = 2 * hsd->request[found].hal_sz;
 263                                 OBD_ALLOC(hal_buffer, sz);
 264                                 if (!hal_buffer) {
 265                                         CERROR("%s: Cannot allocate memory "
 266                                                "(%d o) for compound "LPX64"\n",
 267                                                mdt_obd_name(mdt), sz,
 268                                                larr->arr_compound_id);
 269                                         RETURN(-ENOMEM);
 270                                 }
 271                                 memcpy(hal_buffer, hsd->request[found].hal,
 272                                        hsd->request[found].hal_used_sz);
 273                                 OBD_FREE(hsd->request[found].hal,
 274                                          hsd->request[found].hal_sz);
 275                                 hsd->request[found].hal = hal_buffer;
 276                                 hsd->request[found].hal_sz = sz;
 277                         }
 278                         hai = hai_first(hsd->request[found].hal);
 279                         for (i = 0; i < hsd->request[found].hal->hal_count;
 280                              i++)
 281                                 hai = hai_next(hai);
 282                 }
 283                 memcpy(hai, &larr->arr_hai, larr->arr_hai.hai_len);
 284                 hai->hai_cookie = larr->arr_hai.hai_cookie;
 285                 hai->hai_gid = larr->arr_hai.hai_gid;
 286
 287                 hsd->request[found].hal_used_sz +=
 288                                                    cfs_size_round(hai->hai_len);
 289                 hsd->request[found].hal->hal_count++;
 290                 break;
 291         }
 292         case ARS_STARTED: {
 293                 struct cdt_agent_req *car;
 294                 cfs_time_t last;
 295
 296                 /* we search for a running request
 297                  * error may happen if coordinator crashes or stopped
 298                  * with running request
 299                  */
 300                 car = mdt_cdt_find_request(cdt, larr->arr_hai.hai_cookie, NULL);
 301                 if (car == NULL) {
 302                         last = larr->arr_req_create;
 303                 } else {
 304                         last = car->car_req_update;
 305                         mdt_cdt_put_request(car);
 306                 }
 307
 308                 /* test if request too long, if yes cancel it
 309                  * the same way the copy tool acknowledge a cancel request */
 310                 if ((last + cdt->cdt_active_req_timeout)
 311                      < cfs_time_current_sec()) {
 312                         struct hsm_progress_kernel pgs;
 313
 314                         dump_llog_agent_req_rec("mdt_coordinator_cb(): "
 315                                                 "request timeouted, start "
 316                                                 "cleaning", larr);
 317                         /* a too old cancel request just needs to be removed
 318                          * this can happen, if copy tool does not support cancel
 319                          * for other requests, we have to remove the running
 320                          * request and notify the copytool
 321                          */
 322                         pgs.hpk_fid = larr->arr_hai.hai_fid;
 323                         pgs.hpk_cookie = larr->arr_hai.hai_cookie;
 324                         pgs.hpk_extent = larr->arr_hai.hai_extent;
 325                         pgs.hpk_flags = HP_FLAG_COMPLETED;
 326                         pgs.hpk_errval = ENOSYS;
 327                         pgs.hpk_data_version = 0;
 328                         /* update request state, but do not record in llog, to
 329                          * avoid deadlock on cdt_llog_lock
 330                          */
 331                         rc = mdt_hsm_update_request_state(hsd->mti, &pgs, 0);
 332                         if (rc)
 333                                 CERROR("%s: Cannot cleanup timeouted request: "
 334                                        DFID" for cookie "LPX64" action=%s\n",
 335                                        mdt_obd_name(mdt),
 336                                        PFID(&pgs.hpk_fid), pgs.hpk_cookie,
 337                                        hsm_copytool_action2name(
 338                                                      larr->arr_hai.hai_action));
 339
 340                         /* add the cookie to the list of record to be
 341                          * canceled by caller */
 342                         if (hsd->max_cookie == (hsd->cookie_cnt - 1)) {
 343                                 __u64 *ptr, *old_ptr;
 344                                 int old_sz, new_sz, new_cnt;
 345
 346                                 /* need to increase vector size */
 347                                 old_sz = sizeof(__u64) * hsd->max_cookie;
 348                                 old_ptr = hsd->cookies;
 349
 350                                 new_cnt = 2 * hsd->max_cookie;
 351                                 new_sz = sizeof(__u64) * new_cnt;
 352
 353                                 OBD_ALLOC(ptr, new_sz);
 354                                 if (!ptr) {
 355                                         CERROR("%s: Cannot allocate memory "
 356                                                "(%d o) for cookie vector\n",
 357                                                mdt_obd_name(mdt), new_sz);
 358                                         RETURN(-ENOMEM);
 359                                 }
 360                                 memcpy(ptr, hsd->cookies, old_sz);
 361                                 hsd->cookies = ptr;
 362                                 hsd->max_cookie = new_cnt;
 363                                 OBD_FREE(old_ptr, old_sz);
 364                         }
 365                         hsd->cookies[hsd->cookie_cnt] =
 366                                                        larr->arr_hai.hai_cookie;
 367                         hsd->cookie_cnt++;
 368                 }
 369                 break;
 370         }
 371         case ARS_FAILED:
 372         case ARS_CANCELED:
 373         case ARS_SUCCEED:
 374                 if ((larr->arr_req_change + cdt->cdt_grace_delay) <
 375                     cfs_time_current_sec())
 376                         RETURN(LLOG_DEL_RECORD);
 377                 break;
 378         }
 379         RETURN(0);
 380 }
 381
 382 /**
 383  * create /proc entries for coordinator
 384  * \param mdt [IN]
 385  * \retval 0 success
 386  * \retval -ve failure
 387  */
 388 int hsm_cdt_procfs_init(struct mdt_device *mdt)
 389 {
 390         struct coordinator      *cdt = &mdt->mdt_coordinator;
 391         int                      rc = 0;
 392         ENTRY;
 393
 394         /* init /proc entries, failure is not critical */
 395         cdt->cdt_proc_dir = lprocfs_seq_register("hsm",
 396                                              mdt2obd_dev(mdt)->obd_proc_entry,
 397                                              lprocfs_mdt_hsm_vars, mdt);
 398         if (IS_ERR(cdt->cdt_proc_dir)) {
 399                 rc = PTR_ERR(cdt->cdt_proc_dir);
 400                 CERROR("%s: Cannot create 'hsm' directory in mdt proc dir,"
 401                        " rc=%d\n", mdt_obd_name(mdt), rc);
 402                 cdt->cdt_proc_dir = NULL;
 403                 RETURN(rc);
 404         }
 405
 406         RETURN(0);
 407 }
 408
 409 /**
 410  * remove /proc entries for coordinator
 411  * \param mdt [IN]
 412  */
 413 void  hsm_cdt_procfs_fini(struct mdt_device *mdt)
 414 {
 415         struct coordinator      *cdt = &mdt->mdt_coordinator;
 416
 417         LASSERT(cdt->cdt_state == CDT_STOPPED);
 418         if (cdt->cdt_proc_dir != NULL)
 419                 lprocfs_remove(&cdt->cdt_proc_dir);
 420 }
 421
 422 /**
 423  * get vector of hsm cdt /proc vars
 424  * \param none
 425  * \retval var vector
 426  */
 427 struct lprocfs_seq_vars *hsm_cdt_get_proc_vars(void)
 428 {
 429         return lprocfs_mdt_hsm_vars;
 430 }
 431
 432 /**
 433  * coordinator thread
 434  * \param data [IN] obd device
 435  * \retval 0 success
 436  * \retval -ve failure
 437  */
 438 static int mdt_coordinator(void *data)
 439 {
 440         struct mdt_thread_info  *mti = data;
 441         struct mdt_device       *mdt = mti->mti_mdt;
 442         struct coordinator      *cdt = &mdt->mdt_coordinator;
 443         struct hsm_scan_data     hsd = { 0 };
 444         int                      rc = 0;
 445         ENTRY;
 446
 447         cdt->cdt_thread.t_flags = SVC_RUNNING;
 448         wake_up(&cdt->cdt_thread.t_ctl_waitq);
 449
 450         CDEBUG(D_HSM, "%s: coordinator thread starting, pid=%d\n",
 451                mdt_obd_name(mdt), current_pid());
 452
 453         /* timeouted cookie vector initialization */
 454         hsd.max_cookie = 0;
 455         hsd.cookie_cnt = 0;
 456         hsd.cookies = NULL;
 457         /* we use a copy of cdt_max_requests in the cb, so if cdt_max_requests
 458          * increases due to a change from /proc we do not overflow the
 459          * hsd.request[] vector
 460          */
 461         hsd.max_requests = cdt->cdt_max_requests;
 462         hsd.request_sz = hsd.max_requests * sizeof(*hsd.request);
 463         OBD_ALLOC(hsd.request, hsd.request_sz);
 464         if (!hsd.request)
 465                 GOTO(out, rc = -ENOMEM);
 466
 467         hsd.mti = mti;
 468         obd_uuid2fsname(hsd.fs_name, mdt_obd_name(mdt), MTI_NAME_MAXLEN);
 469
 470         while (1) {
 471                 struct l_wait_info lwi;
 472                 int i;
 473
 474                 lwi = LWI_TIMEOUT(cfs_time_seconds(cdt->cdt_loop_period),
 475                                   NULL, NULL);
 476                 l_wait_event(cdt->cdt_thread.t_ctl_waitq,
 477                              (cdt->cdt_thread.t_flags &
 478                               (SVC_STOPPING|SVC_EVENT)),
 479                              &lwi);
 480
 481                 CDEBUG(D_HSM, "coordinator resumes\n");
 482
 483                 if (cdt->cdt_thread.t_flags & SVC_STOPPING ||
 484                     cdt->cdt_state == CDT_STOPPING) {
 485                         cdt->cdt_thread.t_flags &= ~SVC_STOPPING;
 486                         rc = 0;
 487                         break;
 488                 }
 489
 490                 /* wake up before timeout, new work arrives */
 491                 if (cdt->cdt_thread.t_flags & SVC_EVENT)
 492                         cdt->cdt_thread.t_flags &= ~SVC_EVENT;
 493
 494                 /* if coordinator is suspended continue to wait */
 495                 if (cdt->cdt_state == CDT_DISABLE) {
 496                         CDEBUG(D_HSM, "disable state, coordinator sleeps\n");
 497                         continue;
 498                 }
 499
 500                 CDEBUG(D_HSM, "coordinator starts reading llog\n");
 501
 502                 if (hsd.max_requests != cdt->cdt_max_requests) {
 503                         /* cdt_max_requests has changed,
 504                          * we need to allocate a new buffer
 505                          */
 506                         OBD_FREE(hsd.request, hsd.request_sz);
 507                         hsd.max_requests = cdt->cdt_max_requests;
 508                         hsd.request_sz =
 509                                    hsd.max_requests * sizeof(*hsd.request);
 510                         OBD_ALLOC(hsd.request, hsd.request_sz);
 511                         if (!hsd.request) {
 512                                 rc = -ENOMEM;
 513                                 break;
 514                         }
 515                 }
 516
 517                 /* create canceled cookie vector for an arbitrary size
 518                  * if needed, vector will grow during llog scan
 519                  */
 520                 hsd.max_cookie = 10;
 521                 hsd.cookie_cnt = 0;
 522                 OBD_ALLOC(hsd.cookies, hsd.max_cookie * sizeof(__u64));
 523                 if (!hsd.cookies) {
 524                         rc = -ENOMEM;
 525                         goto clean_cb_alloc;
 526                 }
 527                 hsd.request_cnt = 0;
 528
 529                 rc = cdt_llog_process(mti->mti_env, mdt,
 530                                       mdt_coordinator_cb, &hsd);
 531                 if (rc < 0)
 532                         goto clean_cb_alloc;
 533
 534                 CDEBUG(D_HSM, "Found %d requests to send and %d"
 535                               " requests to cancel\n",
 536                        hsd.request_cnt, hsd.cookie_cnt);
 537                 /* first we cancel llog records of the timeouted requests */
 538                 if (hsd.cookie_cnt > 0) {
 539                         rc = mdt_agent_record_update(mti->mti_env, mdt,
 540                                                      hsd.cookies,
 541                                                      hsd.cookie_cnt,
 542                                                      ARS_CANCELED);
 543                         if (rc)
 544                                 CERROR("%s: mdt_agent_record_update() failed, "
 545                                        "rc=%d, cannot update status to %s "
 546                                        "for %d cookies\n",
 547                                        mdt_obd_name(mdt), rc,
 548                                        agent_req_status2name(ARS_CANCELED),
 549                                        hsd.cookie_cnt);
 550                 }
 551
 552                 if (list_empty(&cdt->cdt_agents)) {
 553                         CDEBUG(D_HSM, "no agent available, "
 554                                       "coordinator sleeps\n");
 555                         goto clean_cb_alloc;
 556                 }
 557
 558                 /* here hsd contains a list of requests to be started */
 559                 for (i = 0; i < hsd.max_requests; i++) {
 560                         struct hsm_action_list  *hal;
 561                         struct hsm_action_item  *hai;
 562                         __u64                   *cookies;
 563                         int                      sz, j;
 564                         enum agent_req_status    status;
 565
 566                         /* still room for work ? */
 567                         if (atomic_read(&cdt->cdt_request_count) ==
 568                             cdt->cdt_max_requests)
 569                                 break;
 570
 571                         if (hsd.request[i].hal == NULL)
 572                                 continue;
 573
 574                         /* found a request, we start it */
 575                         /* kuc payload allocation so we avoid an additionnal
 576                          * allocation in mdt_hsm_agent_send()
 577                          */
 578                         hal = kuc_alloc(hsd.request[i].hal_used_sz,
 579                                         KUC_TRANSPORT_HSM, HMT_ACTION_LIST);
 580                         if (IS_ERR(hal)) {
 581                                 CERROR("%s: Cannot allocate memory (%d o) "
 582                                        "for compound "LPX64"\n",
 583                                        mdt_obd_name(mdt),
 584                                        hsd.request[i].hal_used_sz,
 585                                        hsd.request[i].hal->hal_compound_id);
 586                                 continue;
 587                         }
 588                         memcpy(hal, hsd.request[i].hal,
 589                                hsd.request[i].hal_used_sz);
 590
 591                         rc = mdt_hsm_agent_send(mti, hal, 0);
 592                         /* if failure, we suppose it is temporary
 593                          * if the copy tool failed to do the request
 594                          * it has to use hsm_progress
 595                          */
 596                         status = (rc ? ARS_WAITING : ARS_STARTED);
 597
 598                         /* set up cookie vector to set records status
 599                          * after copy tools start or failed
 600                          */
 601                         sz = hsd.request[i].hal->hal_count * sizeof(__u64);
 602                         OBD_ALLOC(cookies, sz);
 603                         if (cookies == NULL) {
 604                                 CERROR("%s: Cannot allocate memory (%d o) "
 605                                        "for cookies vector "LPX64"\n",
 606                                        mdt_obd_name(mdt), sz,
 607                                        hsd.request[i].hal->hal_compound_id);
 608                                 kuc_free(hal, hsd.request[i].hal_used_sz);
 609                                 continue;
 610                         }
 611                         hai = hai_first(hal);
 612                         for (j = 0; j < hsd.request[i].hal->hal_count; j++) {
 613                                 cookies[j] = hai->hai_cookie;
 614                                 hai = hai_next(hai);
 615                         }
 616
 617                         rc = mdt_agent_record_update(mti->mti_env, mdt, cookies,
 618                                                 hsd.request[i].hal->hal_count,
 619                                                 status);
 620                         if (rc)
 621                                 CERROR("%s: mdt_agent_record_update() failed, "
 622                                        "rc=%d, cannot update status to %s "
 623                                        "for %d cookies\n",
 624                                        mdt_obd_name(mdt), rc,
 625                                        agent_req_status2name(status),
 626                                        hsd.request[i].hal->hal_count);
 627
 628                         OBD_FREE(cookies, sz);
 629                         kuc_free(hal, hsd.request[i].hal_used_sz);
 630                 }
 631 clean_cb_alloc:
 632                 /* free cookie vector allocated for/by callback */
 633                 if (hsd.cookies) {
 634                         OBD_FREE(hsd.cookies, hsd.max_cookie * sizeof(__u64));
 635                         hsd.max_cookie = 0;
 636                         hsd.cookie_cnt = 0;
 637                         hsd.cookies = NULL;
 638                 }
 639
 640                 /* free hal allocated by callback */
 641                 for (i = 0; i < hsd.max_requests; i++) {
 642                         if (hsd.request[i].hal) {
 643                                 OBD_FREE(hsd.request[i].hal,
 644                                          hsd.request[i].hal_sz);
 645                                 hsd.request[i].hal_sz = 0;
 646                                 hsd.request[i].hal = NULL;
 647                                 hsd.request_cnt--;
 648                         }
 649                 }
 650                 LASSERT(hsd.request_cnt == 0);
 651
 652                 /* reset callback data */
 653                 memset(hsd.request, 0, hsd.request_sz);
 654         }
 655         EXIT;
 656 out:
 657         if (hsd.request)
 658                 OBD_FREE(hsd.request, hsd.request_sz);
 659
 660         if (hsd.cookies)
 661                 OBD_FREE(hsd.cookies, hsd.max_cookie * sizeof(__u64));
 662
 663         if (cdt->cdt_state == CDT_STOPPING) {
 664                 /* request comes from /proc path, so we need to clean cdt
 665                  * struct */
 666                  mdt_hsm_cdt_stop(mdt);
 667                  mdt->mdt_opts.mo_coordinator = 0;
 668         } else {
 669                 /* request comes from a thread event, generated
 670                  * by mdt_stop_coordinator(), we have to ack
 671                  * and cdt cleaning will be done by event sender
 672                  */
 673                 cdt->cdt_thread.t_flags = SVC_STOPPED;
 674                 wake_up(&cdt->cdt_thread.t_ctl_waitq);
 675         }
 676
 677         if (rc != 0)
 678                 CERROR("%s: coordinator thread exiting, process=%d, rc=%d\n",
 679                        mdt_obd_name(mdt), current_pid(), rc);
 680         else
 681                 CDEBUG(D_HSM, "%s: coordinator thread exiting, process=%d,"
 682                               " no error\n",
 683                        mdt_obd_name(mdt), current_pid());
 684
 685         return rc;
 686 }
 687
 688 /**
 689  * lookup a restore handle by FID
 690  * caller needs to hold cdt_restore_lock
 691  * \param cdt [IN] coordinator
 692  * \param fid [IN] FID
 693  * \retval cdt_restore_handle found
 694  * \retval NULL not found
 695  */
 696 static struct cdt_restore_handle *hsm_restore_hdl_find(struct coordinator *cdt,
 697                                                        const struct lu_fid *fid)
 698 {
 699         struct cdt_restore_handle       *crh;
 700         ENTRY;
 701
 702         list_for_each_entry(crh, &cdt->cdt_restore_hdl, crh_list) {
 703                 if (lu_fid_eq(&crh->crh_fid, fid))
 704                         RETURN(crh);
 705         }
 706         RETURN(NULL);
 707 }
 708
 709 /**
 710  * data passed to llog_cat_process() callback
 711  * to scan requests and take actions
 712  */
 713 struct hsm_restore_data {
 714         struct mdt_thread_info  *hrd_mti;
 715 };
 716
 717 /**
 718  *  llog_cat_process() callback, used to:
 719  *  - find restore request and allocate the restore handle
 720  * \param env [IN] environment
 721  * \param llh [IN] llog handle
 722  * \param hdr [IN] llog record
 723  * \param data [IN/OUT] cb data = struct hsm_restore_data
 724  * \retval 0 success
 725  * \retval -ve failure
 726  */
 727 static int hsm_restore_cb(const struct lu_env *env,
 728                           struct llog_handle *llh,
 729                           struct llog_rec_hdr *hdr, void *data)
 730 {
 731         struct llog_agent_req_rec       *larr;
 732         struct hsm_restore_data         *hrd;
 733         struct cdt_restore_handle       *crh;
 734         struct hsm_action_item          *hai;
 735         struct mdt_thread_info          *mti;
 736         struct coordinator              *cdt;
 737         struct mdt_object               *child;
 738         int rc;
 739         ENTRY;
 740
 741         hrd = data;
 742         mti = hrd->hrd_mti;
 743         cdt = &mti->mti_mdt->mdt_coordinator;
 744
 745         larr = (struct llog_agent_req_rec *)hdr;
 746         hai = &larr->arr_hai;
 747         if (hai->hai_cookie > cdt->cdt_last_cookie)
 748                 /* update the cookie to avoid collision */
 749                 cdt->cdt_last_cookie = hai->hai_cookie + 1;
 750
 751         if (hai->hai_action != HSMA_RESTORE ||
 752             agent_req_in_final_state(larr->arr_status))
 753                 RETURN(0);
 754
 755         /* restore request not in a final state */
 756
 757         OBD_SLAB_ALLOC_PTR(crh, mdt_hsm_cdt_kmem);
 758         if (crh == NULL)
 759                 RETURN(-ENOMEM);
 760
 761         crh->crh_fid = hai->hai_fid;
 762         /* in V1 all file is restored
 763         crh->extent.start = hai->hai_extent.offset;
 764         crh->extent.end = hai->hai_extent.offset + hai->hai_extent.length;
 765         */
 766         crh->crh_extent.start = 0;
 767         crh->crh_extent.end = hai->hai_extent.length;
 768         /* get the layout lock */
 769         mdt_lock_reg_init(&crh->crh_lh, LCK_EX);
 770         child = mdt_object_find_lock(mti, &crh->crh_fid, &crh->crh_lh,
 771                                      MDS_INODELOCK_LAYOUT);
 772         if (IS_ERR(child))
 773                 GOTO(out, rc = PTR_ERR(child));
 774
 775         rc = 0;
 776         /* we choose to not keep a reference
 777          * on the object during the restore time which can be very long */
 778         mdt_object_put(mti->mti_env, child);
 779
 780         mutex_lock(&cdt->cdt_restore_lock);
 781         list_add_tail(&crh->crh_list, &cdt->cdt_restore_hdl);
 782         mutex_unlock(&cdt->cdt_restore_lock);
 783
 784 out:
 785         RETURN(rc);
 786 }
 787
 788 /**
 789  * restore coordinator state at startup
 790  * the goal is to take a layout lock for each registered restore request
 791  * \param mti [IN] context
 792  */
 793 static int mdt_hsm_pending_restore(struct mdt_thread_info *mti)
 794 {
 795         struct hsm_restore_data  hrd;
 796         int                      rc;
 797         ENTRY;
 798
 799         hrd.hrd_mti = mti;
 800
 801         rc = cdt_llog_process(mti->mti_env, mti->mti_mdt,
 802                               hsm_restore_cb, &hrd);
 803
 804         RETURN(rc);
 805 }
 806
 807 static int hsm_init_ucred(struct lu_ucred *uc)
 808 {
 809         ENTRY;
 810
 811         uc->uc_valid = UCRED_OLD;
 812         uc->uc_o_uid = 0;
 813         uc->uc_o_gid = 0;
 814         uc->uc_o_fsuid = 0;
 815         uc->uc_o_fsgid = 0;
 816         uc->uc_uid = 0;
 817         uc->uc_gid = 0;
 818         uc->uc_fsuid = 0;
 819         uc->uc_fsgid = 0;
 820         uc->uc_suppgids[0] = -1;
 821         uc->uc_suppgids[1] = -1;
 822         uc->uc_cap = CFS_CAP_FS_MASK;
 823         uc->uc_umask = 0777;
 824         uc->uc_ginfo = NULL;
 825         uc->uc_identity = NULL;
 826
 827         RETURN(0);
 828 }
 829
 830 /**
 831  * wake up coordinator thread
 832  * \param mdt [IN] device
 833  * \retval 0 success
 834  * \retval -ve failure
 835  */
 836 int mdt_hsm_cdt_wakeup(struct mdt_device *mdt)
 837 {
 838         struct coordinator      *cdt = &mdt->mdt_coordinator;
 839         ENTRY;
 840
 841         if (cdt->cdt_state == CDT_STOPPED)
 842                 RETURN(-ESRCH);
 843
 844         /* wake up coordinator */
 845         cdt->cdt_thread.t_flags = SVC_EVENT;
 846         wake_up(&cdt->cdt_thread.t_ctl_waitq);
 847
 848         RETURN(0);
 849 }
 850
 851 /**
 852  * initialize coordinator struct
 853  * \param mdt [IN] device
 854  * \retval 0 success
 855  * \retval -ve failure
 856  */
 857 int mdt_hsm_cdt_init(struct mdt_device *mdt)
 858 {
 859         struct coordinator      *cdt = &mdt->mdt_coordinator;
 860         struct mdt_thread_info  *cdt_mti = NULL;
 861         int                      rc;
 862         ENTRY;
 863
 864         cdt->cdt_state = CDT_STOPPED;
 865
 866         init_waitqueue_head(&cdt->cdt_thread.t_ctl_waitq);
 867         mutex_init(&cdt->cdt_llog_lock);
 868         init_rwsem(&cdt->cdt_agent_lock);
 869         init_rwsem(&cdt->cdt_request_lock);
 870         mutex_init(&cdt->cdt_restore_lock);
 871
 872         CFS_INIT_LIST_HEAD(&cdt->cdt_requests);
 873         CFS_INIT_LIST_HEAD(&cdt->cdt_agents);
 874         CFS_INIT_LIST_HEAD(&cdt->cdt_restore_hdl);
 875
 876         rc = lu_env_init(&cdt->cdt_env, LCT_MD_THREAD);
 877         if (rc < 0)
 878                 RETURN(rc);
 879
 880         /* for mdt_ucred(), lu_ucred stored in lu_ucred_key */
 881         rc = lu_context_init(&cdt->cdt_session, LCT_SERVER_SESSION);
 882         if (rc == 0) {
 883                 lu_context_enter(&cdt->cdt_session);
 884                 cdt->cdt_env.le_ses = &cdt->cdt_session;
 885         } else {
 886                 lu_env_fini(&cdt->cdt_env);
 887                 RETURN(rc);
 888         }
 889
 890         cdt_mti = lu_context_key_get(&cdt->cdt_env.le_ctx, &mdt_thread_key);
 891         LASSERT(cdt_mti != NULL);
 892
 893         cdt_mti->mti_env = &cdt->cdt_env;
 894         cdt_mti->mti_mdt = mdt;
 895
 896         hsm_init_ucred(mdt_ucred(cdt_mti));
 897
 898         /* default values for /proc tunnables
 899          * can be override by MGS conf */
 900         cdt->cdt_default_archive_id = 1;
 901         cdt->cdt_grace_delay = 60;
 902         cdt->cdt_loop_period = 10;
 903         cdt->cdt_max_requests = 3;
 904         cdt->cdt_policy = CDT_DEFAULT_POLICY;
 905         cdt->cdt_active_req_timeout = 3600;
 906
 907         RETURN(0);
 908 }
 909
 910 /**
 911  * free a coordinator thread
 912  * \param mdt [IN] device
 913  */
 914 int  mdt_hsm_cdt_fini(struct mdt_device *mdt)
 915 {
 916         struct coordinator *cdt = &mdt->mdt_coordinator;
 917         ENTRY;
 918
 919         lu_context_exit(cdt->cdt_env.le_ses);
 920         lu_context_fini(cdt->cdt_env.le_ses);
 921
 922         lu_env_fini(&cdt->cdt_env);
 923
 924         RETURN(0);
 925 }
 926
 927 /**
 928  * start a coordinator thread
 929  * \param mdt [IN] device
 930  * \retval 0 success
 931  * \retval -ve failure
 932  */
 933 int mdt_hsm_cdt_start(struct mdt_device *mdt)
 934 {
 935         struct coordinator      *cdt = &mdt->mdt_coordinator;
 936         int                      rc;
 937         void                    *ptr;
 938         struct mdt_thread_info  *cdt_mti;
 939         struct task_struct      *task;
 940         ENTRY;
 941
 942         /* functions defined but not yet used
 943          * this avoid compilation warning
 944          */
 945         ptr = dump_requests;
 946
 947         if (cdt->cdt_state != CDT_STOPPED) {
 948                 CERROR("%s: Coordinator already started\n",
 949                        mdt_obd_name(mdt));
 950                 RETURN(-EALREADY);
 951         }
 952
 953         CLASSERT(1 << (CDT_POLICY_SHIFT_COUNT - 1) == CDT_POLICY_LAST);
 954         cdt->cdt_policy = CDT_DEFAULT_POLICY;
 955
 956         cdt->cdt_state = CDT_INIT;
 957
 958         atomic_set(&cdt->cdt_compound_id, cfs_time_current_sec());
 959         /* just need to be larger than previous one */
 960         /* cdt_last_cookie is protected by cdt_llog_lock */
 961         cdt->cdt_last_cookie = cfs_time_current_sec();
 962         atomic_set(&cdt->cdt_request_count, 0);
 963         cdt->cdt_user_request_mask = (1UL << HSMA_RESTORE);
 964         cdt->cdt_group_request_mask = (1UL << HSMA_RESTORE);
 965         cdt->cdt_other_request_mask = (1UL << HSMA_RESTORE);
 966
 967         /* to avoid deadlock when start is made through /proc
 968          * /proc entries are created by the coordinator thread */
 969
 970         /* set up list of started restore requests */
 971         cdt_mti = lu_context_key_get(&cdt->cdt_env.le_ctx, &mdt_thread_key);
 972         rc = mdt_hsm_pending_restore(cdt_mti);
 973         if (rc)
 974                 CERROR("%s: cannot take the layout locks needed"
 975                        " for registered restore: %d\n",
 976                        mdt_obd_name(mdt), rc);
 977
 978         task = kthread_run(mdt_coordinator, cdt_mti, "hsm_cdtr");
 979         if (IS_ERR(task)) {
 980                 rc = PTR_ERR(task);
 981                 cdt->cdt_state = CDT_STOPPED;
 982                 CERROR("%s: error starting coordinator thread: %d\n",
 983                        mdt_obd_name(mdt), rc);
 984                 RETURN(rc);
 985         } else {
 986                 CDEBUG(D_HSM, "%s: coordinator thread started\n",
 987                        mdt_obd_name(mdt));
 988                 rc = 0;
 989         }
 990
 991         wait_event(cdt->cdt_thread.t_ctl_waitq,
 992                        (cdt->cdt_thread.t_flags & SVC_RUNNING));
 993
 994         cdt->cdt_state = CDT_RUNNING;
 995         mdt->mdt_opts.mo_coordinator = 1;
 996         RETURN(0);
 997 }
 998
 999 /**
1000  * stop a coordinator thread
1001  * \param mdt [IN] device
1002  */
1003 int mdt_hsm_cdt_stop(struct mdt_device *mdt)
1004 {
1005         struct coordinator              *cdt = &mdt->mdt_coordinator;
1006         struct cdt_agent_req            *car, *tmp1;
1007         struct hsm_agent                *ha, *tmp2;
1008         struct cdt_restore_handle       *crh, *tmp3;
1009         struct mdt_thread_info          *cdt_mti;
1010         ENTRY;
1011
1012         if (cdt->cdt_state == CDT_STOPPED) {
1013                 CERROR("%s: Coordinator already stopped\n",
1014                        mdt_obd_name(mdt));
1015                 RETURN(-EALREADY);
1016         }
1017
1018         if (cdt->cdt_state != CDT_STOPPING) {
1019                 /* stop coordinator thread before cleaning */
1020                 cdt->cdt_thread.t_flags = SVC_STOPPING;
1021                 wake_up(&cdt->cdt_thread.t_ctl_waitq);
1022                 wait_event(cdt->cdt_thread.t_ctl_waitq,
1023                            cdt->cdt_thread.t_flags & SVC_STOPPED);
1024         }
1025         cdt->cdt_state = CDT_STOPPED;
1026
1027         /* start cleaning */
1028         down_write(&cdt->cdt_request_lock);
1029         list_for_each_entry_safe(car, tmp1, &cdt->cdt_requests,
1030                                  car_request_list) {
1031                 list_del(&car->car_request_list);
1032                 mdt_cdt_free_request(car);
1033         }
1034         up_write(&cdt->cdt_request_lock);
1035
1036         down_write(&cdt->cdt_agent_lock);
1037         list_for_each_entry_safe(ha, tmp2, &cdt->cdt_agents, ha_list) {
1038                 list_del(&ha->ha_list);
1039                 OBD_FREE_PTR(ha);
1040         }
1041         up_write(&cdt->cdt_agent_lock);
1042
1043         cdt_mti = lu_context_key_get(&cdt->cdt_env.le_ctx, &mdt_thread_key);
1044         mutex_lock(&cdt->cdt_restore_lock);
1045         list_for_each_entry_safe(crh, tmp3, &cdt->cdt_restore_hdl, crh_list) {
1046                 struct mdt_object       *child;
1047
1048                 /* give back layout lock */
1049                 child = mdt_object_find(&cdt->cdt_env, mdt, &crh->crh_fid);
1050                 if (!IS_ERR(child))
1051                         mdt_object_unlock_put(cdt_mti, child, &crh->crh_lh, 1);
1052
1053                 list_del(&crh->crh_list);
1054
1055                 OBD_SLAB_FREE_PTR(crh, mdt_hsm_cdt_kmem);
1056         }
1057         mutex_unlock(&cdt->cdt_restore_lock);
1058
1059         mdt->mdt_opts.mo_coordinator = 0;
1060
1061         RETURN(0);
1062 }
1063
1064 /**
1065  * register all requests from an hal in the memory list
1066  * \param mti [IN] context
1067  * \param hal [IN] request
1068  * \param uuid [OUT] in case of CANCEL, the uuid of the agent
1069  *  which is running the CT
1070  * \retval 0 success
1071  * \retval -ve failure
1072  */
1073 int mdt_hsm_add_hal(struct mdt_thread_info *mti,
1074                     struct hsm_action_list *hal, struct obd_uuid *uuid)
1075 {
1076         struct mdt_device       *mdt = mti->mti_mdt;
1077         struct coordinator      *cdt = &mdt->mdt_coordinator;
1078         struct hsm_action_item  *hai;
1079         int                      rc = 0, i;
1080         ENTRY;
1081
1082         /* register request in memory list */
1083         hai = hai_first(hal);
1084         for (i = 0; i < hal->hal_count; i++, hai = hai_next(hai)) {
1085                 struct cdt_agent_req *car;
1086
1087                 /* in case of a cancel request, we first mark the ondisk
1088                  * record of the request we want to stop as canceled
1089                  * this does not change the cancel record
1090                  * it will be done when updating the request status
1091                  */
1092                 if (hai->hai_action == HSMA_CANCEL) {
1093                         rc = mdt_agent_record_update(mti->mti_env, mti->mti_mdt,
1094                                                      &hai->hai_cookie,
1095                                                      1, ARS_CANCELED);
1096                         if (rc) {
1097                                 CERROR("%s: mdt_agent_record_update() failed, "
1098                                        "rc=%d, cannot update status to %s "
1099                                        "for cookie "LPX64"\n",
1100                                        mdt_obd_name(mdt), rc,
1101                                        agent_req_status2name(ARS_CANCELED),
1102                                        hai->hai_cookie);
1103                                 GOTO(out, rc);
1104                         }
1105
1106                         /* find the running request to set it canceled */
1107                         car = mdt_cdt_find_request(cdt, hai->hai_cookie, NULL);
1108                         if (car != NULL) {
1109                                 car->car_canceled = 1;
1110                                 /* uuid has to be changed to the one running the
1111                                 * request to cancel */
1112                                 *uuid = car->car_uuid;
1113                                 mdt_cdt_put_request(car);
1114                         }
1115                         /* no need to memorize cancel request
1116                          * this also avoid a deadlock when we receive
1117                          * a purge all requests command
1118                          */
1119                         continue;
1120                 }
1121
1122                 if (hai->hai_action == HSMA_ARCHIVE) {
1123                         struct mdt_object *obj;
1124                         struct md_hsm hsm;
1125
1126                         obj = mdt_hsm_get_md_hsm(mti, &hai->hai_fid, &hsm);
1127                         if (IS_ERR(obj) && (PTR_ERR(obj) == -ENOENT))
1128                                 continue;
1129                         if (IS_ERR(obj))
1130                                 GOTO(out, rc = PTR_ERR(obj));
1131
1132                         hsm.mh_flags |= HS_EXISTS;
1133                         hsm.mh_arch_id = hal->hal_archive_id;
1134                         rc = mdt_hsm_attr_set(mti, obj, &hsm);
1135                         mdt_object_put(mti->mti_env, obj);
1136                         if (rc)
1137                                 GOTO(out, rc);
1138                 }
1139
1140                 car = mdt_cdt_alloc_request(hal->hal_compound_id,
1141                                             hal->hal_archive_id, hal->hal_flags,
1142                                             uuid, hai);
1143                 if (IS_ERR(car))
1144                         GOTO(out, rc = PTR_ERR(car));
1145
1146                 rc = mdt_cdt_add_request(cdt, car);
1147                 if (rc != 0)
1148                         mdt_cdt_free_request(car);
1149         }
1150 out:
1151         RETURN(rc);
1152 }
1153
1154 /**
1155  * swap layouts between 2 fids
1156  * \param mti [IN] context
1157  * \param fid1 [IN]
1158  * \param fid2 [IN]
1159  * \param mh_common [IN] MD HSM
1160  */
1161 static int hsm_swap_layouts(struct mdt_thread_info *mti,
1162                             const lustre_fid *fid, const lustre_fid *dfid,
1163                             struct md_hsm *mh_common)
1164 {
1165         struct mdt_device       *mdt = mti->mti_mdt;
1166         struct mdt_object       *child1, *child2;
1167         struct mdt_lock_handle  *lh2;
1168         int                      rc;
1169         ENTRY;
1170
1171         child1 = mdt_object_find(mti->mti_env, mdt, fid);
1172         if (IS_ERR(child1))
1173                 GOTO(out, rc = PTR_ERR(child1));
1174
1175         /* we already have layout lock on FID so take only
1176          * on dfid */
1177         lh2 = &mti->mti_lh[MDT_LH_OLD];
1178         mdt_lock_reg_init(lh2, LCK_EX);
1179         child2 = mdt_object_find_lock(mti, dfid, lh2, MDS_INODELOCK_LAYOUT);
1180         if (IS_ERR(child2))
1181                 GOTO(out_child1, rc = PTR_ERR(child2));
1182
1183         /* if copy tool closes the volatile before sending the final
1184          * progress through llapi_hsm_copy_end(), all the objects
1185          * are removed and mdd_swap_layout LBUG */
1186         if (!mdt_object_exists(child2)) {
1187                 CERROR("%s: Copytool has closed volatile file "DFID"\n",
1188                        mdt_obd_name(mti->mti_mdt), PFID(dfid));
1189                 GOTO(out_child2, rc = -ENOENT);
1190         }
1191         /* Since we only handle restores here, unconditionally use
1192          * SWAP_LAYOUTS_MDS_HSM flag to ensure original layout will
1193          * be preserved in case of failure during swap_layout and not
1194          * leave a file in an intermediate but incoherent state.
1195          * But need to setup HSM xattr of data FID before, reuse
1196          * mti and mh presets for FID in hsm_cdt_request_completed(),
1197          * only need to clear RELEASED and DIRTY.
1198          */
1199         mh_common->mh_flags &= ~(HS_RELEASED | HS_DIRTY);
1200         rc = mdt_hsm_attr_set(mti, child2, mh_common);
1201         if (rc == 0)
1202                 rc = mo_swap_layouts(mti->mti_env,
1203                                      mdt_object_child(child1),
1204                                      mdt_object_child(child2),
1205                                      SWAP_LAYOUTS_MDS_HSM);
1206
1207 out_child2:
1208         mdt_object_unlock_put(mti, child2, lh2, 1);
1209 out_child1:
1210         mdt_object_put(mti->mti_env, child1);
1211 out:
1212         RETURN(rc);
1213 }
1214
1215 /**
1216  * update status of a completed request
1217  * \param mti [IN] context
1218  * \param pgs [IN] progress of the copy tool
1219  * \param update_record [IN] update llog record
1220  * \retval 0 success
1221  * \retval -ve failure
1222  */
1223 static int hsm_cdt_request_completed(struct mdt_thread_info *mti,
1224                                      struct hsm_progress_kernel *pgs,
1225                                      const struct cdt_agent_req *car,
1226                                      enum agent_req_status *status)
1227 {
1228         const struct lu_env     *env = mti->mti_env;
1229         struct mdt_device       *mdt = mti->mti_mdt;
1230         struct coordinator      *cdt = &mdt->mdt_coordinator;
1231         struct mdt_object       *obj = NULL;
1232         int                      cl_flags = 0, rc = 0;
1233         struct md_hsm            mh;
1234         bool                     is_mh_changed;
1235         ENTRY;
1236
1237         /* default is to retry */
1238         *status = ARS_WAITING;
1239
1240         /* find object by FID */
1241         obj = mdt_hsm_get_md_hsm(mti, &car->car_hai->hai_fid, &mh);
1242         /* we will update MD HSM only if needed */
1243         is_mh_changed = false;
1244         if (IS_ERR(obj)) {
1245                 /* object removed */
1246                 *status = ARS_SUCCEED;
1247                 goto unlock;
1248         }
1249
1250         /* no need to change mh->mh_arch_id
1251          * mdt_hsm_get_md_hsm() got it from disk and it is still valid
1252          */
1253         if (pgs->hpk_errval != 0) {
1254                 switch (pgs->hpk_errval) {
1255                 case ENOSYS:
1256                         /* the copy tool does not support cancel
1257                          * so the cancel request is failed
1258                          * As we cannot distinguish a cancel progress
1259                          * from another action progress (they have the
1260                          * same cookie), we suppose here the CT returns
1261                          * ENOSYS only if does not support cancel
1262                          */
1263                         /* this can also happen when cdt calls it to
1264                          * for a timeouted request */
1265                         *status = ARS_FAILED;
1266                         /* to have a cancel event in changelog */
1267                         pgs->hpk_errval = ECANCELED;
1268                         break;
1269                 case ECANCELED:
1270                         /* the request record has already been set to
1271                          * ARS_CANCELED, this set the cancel request
1272                          * to ARS_SUCCEED */
1273                         *status = ARS_SUCCEED;
1274                         break;
1275                 default:
1276                         *status = (cdt->cdt_policy & CDT_NORETRY_ACTION ||
1277                                    !(pgs->hpk_flags & HP_FLAG_RETRY) ?
1278                                    ARS_FAILED : ARS_WAITING);
1279                         break;
1280                 }
1281
1282                 if (pgs->hpk_errval > CLF_HSM_MAXERROR) {
1283                         CERROR("%s: Request "LPX64" on "DFID
1284                                " failed, error code %d too large\n",
1285                                mdt_obd_name(mdt),
1286                                pgs->hpk_cookie, PFID(&pgs->hpk_fid),
1287                                pgs->hpk_errval);
1288                         hsm_set_cl_error(&cl_flags,
1289                                          CLF_HSM_ERROVERFLOW);
1290                         rc = -EINVAL;
1291                 } else {
1292                         hsm_set_cl_error(&cl_flags, pgs->hpk_errval);
1293                 }
1294
1295                 switch (car->car_hai->hai_action) {
1296                 case HSMA_ARCHIVE:
1297                         hsm_set_cl_event(&cl_flags, HE_ARCHIVE);
1298                         break;
1299                 case HSMA_RESTORE:
1300                         hsm_set_cl_event(&cl_flags, HE_RESTORE);
1301                         break;
1302                 case HSMA_REMOVE:
1303                         hsm_set_cl_event(&cl_flags, HE_REMOVE);
1304                         break;
1305                 case HSMA_CANCEL:
1306                         hsm_set_cl_event(&cl_flags, HE_CANCEL);
1307                         CERROR("%s: Failed request "LPX64" on "DFID
1308                                " cannot be a CANCEL\n",
1309                                mdt_obd_name(mdt),
1310                                pgs->hpk_cookie,
1311                                PFID(&pgs->hpk_fid));
1312                         break;
1313                 default:
1314                         CERROR("%s: Failed request "LPX64" on "DFID
1315                                " %d is an unknown action\n",
1316                                mdt_obd_name(mdt),
1317                                pgs->hpk_cookie, PFID(&pgs->hpk_fid),
1318                                car->car_hai->hai_action);
1319                         rc = -EINVAL;
1320                         break;
1321                 }
1322         } else {
1323                 *status = ARS_SUCCEED;
1324                 switch (car->car_hai->hai_action) {
1325                 case HSMA_ARCHIVE:
1326                         hsm_set_cl_event(&cl_flags, HE_ARCHIVE);
1327                         /* set ARCHIVE keep EXIST and clear LOST and
1328                          * DIRTY */
1329                         mh.mh_arch_ver = pgs->hpk_data_version;
1330                         mh.mh_flags |= HS_ARCHIVED;
1331                         mh.mh_flags &= ~(HS_LOST|HS_DIRTY);
1332                         is_mh_changed = true;
1333                         break;
1334                 case HSMA_RESTORE:
1335                         hsm_set_cl_event(&cl_flags, HE_RESTORE);
1336
1337                         /* do not clear RELEASED and DIRTY here
1338                          * this will occur in hsm_swap_layouts()
1339                          */
1340
1341                         /* Restoring has changed the file version on
1342                          * disk. */
1343                         mh.mh_arch_ver = pgs->hpk_data_version;
1344                         is_mh_changed = true;
1345                         break;
1346                 case HSMA_REMOVE:
1347                         hsm_set_cl_event(&cl_flags, HE_REMOVE);
1348                         /* clear ARCHIVED EXISTS and LOST */
1349                         mh.mh_flags &= ~(HS_ARCHIVED | HS_EXISTS | HS_LOST);
1350                         is_mh_changed = true;
1351                         break;
1352                 case HSMA_CANCEL:
1353                         hsm_set_cl_event(&cl_flags, HE_CANCEL);
1354                         CERROR("%s: Successful request "LPX64
1355                                " on "DFID
1356                                " cannot be a CANCEL\n",
1357                                mdt_obd_name(mdt),
1358                                pgs->hpk_cookie,
1359                                PFID(&pgs->hpk_fid));
1360                         break;
1361                 default:
1362                         CERROR("%s: Successful request "LPX64
1363                                " on "DFID
1364                                " %d is an unknown action\n",
1365                                mdt_obd_name(mdt),
1366                                pgs->hpk_cookie, PFID(&pgs->hpk_fid),
1367                                car->car_hai->hai_action);
1368                         rc = -EINVAL;
1369                         break;
1370                 }
1371         }
1372
1373         /* rc != 0 means error when analysing action, it may come from
1374          * a crasy CT no need to manage DIRTY
1375          */
1376         if (rc == 0)
1377                 hsm_set_cl_flags(&cl_flags,
1378                                  mh.mh_flags & HS_DIRTY ? CLF_HSM_DIRTY : 0);
1379
1380         /* unlock is done later, after layout lock management */
1381         if (is_mh_changed)
1382                 rc = mdt_hsm_attr_set(mti, obj, &mh);
1383
1384 unlock:
1385         /* we give back layout lock only if restore was successful or
1386          * if restore was canceled or if policy is to not retry
1387          * in other cases we just unlock the object */
1388         if (car->car_hai->hai_action == HSMA_RESTORE &&
1389             (pgs->hpk_errval == 0 || pgs->hpk_errval == ECANCELED ||
1390              cdt->cdt_policy & CDT_NORETRY_ACTION)) {
1391                 struct cdt_restore_handle       *crh;
1392
1393                 /* restore in data FID done, we swap the layouts
1394                  * only if restore is successfull */
1395                 if (pgs->hpk_errval == 0) {
1396                         rc = hsm_swap_layouts(mti, &car->car_hai->hai_fid,
1397                                               &car->car_hai->hai_dfid, &mh);
1398                         if (rc) {
1399                                 if (cdt->cdt_policy & CDT_NORETRY_ACTION)
1400                                         *status = ARS_FAILED;
1401                                 pgs->hpk_errval = -rc;
1402                         }
1403                 }
1404                 /* we have to retry, so keep layout lock */
1405                 if (*status == ARS_WAITING)
1406                         GOTO(out, rc);
1407
1408                 /* give back layout lock */
1409                 mutex_lock(&cdt->cdt_restore_lock);
1410                 crh = hsm_restore_hdl_find(cdt, &car->car_hai->hai_fid);
1411                 if (crh != NULL)
1412                         list_del(&crh->crh_list);
1413                 mutex_unlock(&cdt->cdt_restore_lock);
1414                 /* just give back layout lock, we keep
1415                  * the reference which is given back
1416                  * later with the lock for HSM flags */
1417                 if (!IS_ERR(obj) && crh != NULL)
1418                         mdt_object_unlock(mti, obj, &crh->crh_lh, 1);
1419
1420                 if (crh != NULL)
1421                         OBD_SLAB_FREE_PTR(crh, mdt_hsm_cdt_kmem);
1422         }
1423
1424         GOTO(out, rc);
1425
1426 out:
1427         if (obj != NULL && !IS_ERR(obj)) {
1428                 mo_changelog(env, CL_HSM, cl_flags,
1429                              mdt_object_child(obj));
1430                 mdt_object_put(mti->mti_env, obj);
1431         }
1432
1433         RETURN(rc);
1434 }
1435
1436 /**
1437  * update status of a request
1438  * \param mti [IN] context
1439  * \param pgs [IN] progress of the copy tool
1440  * \param update_record [IN] update llog record
1441  * \retval 0 success
1442  * \retval -ve failure
1443  */
1444 int mdt_hsm_update_request_state(struct mdt_thread_info *mti,
1445                                  struct hsm_progress_kernel *pgs,
1446                                  const int update_record)
1447 {
1448         struct mdt_device       *mdt = mti->mti_mdt;
1449         struct coordinator      *cdt = &mdt->mdt_coordinator;
1450         struct cdt_agent_req    *car;
1451         int                      rc = 0;
1452         ENTRY;
1453
1454         /* no coordinator started, so we cannot serve requests */
1455         if (cdt->cdt_state == CDT_STOPPED)
1456                 RETURN(-EAGAIN);
1457
1458         /* first do sanity checks */
1459         car = mdt_cdt_update_request(cdt, pgs);
1460         if (IS_ERR(car)) {
1461                 CERROR("%s: Cannot find running request for cookie "LPX64
1462                        " on fid="DFID"\n",
1463                        mdt_obd_name(mdt),
1464                        pgs->hpk_cookie, PFID(&pgs->hpk_fid));
1465                 if (car == NULL)
1466                         RETURN(-ENOENT);
1467                 RETURN(PTR_ERR(car));
1468         }
1469
1470         CDEBUG(D_HSM, "Progress received for fid="DFID" cookie="LPX64
1471                       " action=%s flags=%d err=%d fid="DFID" dfid="DFID"\n",
1472                       PFID(&pgs->hpk_fid), pgs->hpk_cookie,
1473                       hsm_copytool_action2name(car->car_hai->hai_action),
1474                       pgs->hpk_flags, pgs->hpk_errval,
1475                       PFID(&car->car_hai->hai_fid),
1476                       PFID(&car->car_hai->hai_dfid));
1477
1478         /* progress is done on FID or data FID depending of the action and
1479          * of the copy progress */
1480         /* for restore progress is used to send back the data FID to cdt */
1481         if (car->car_hai->hai_action == HSMA_RESTORE &&
1482             lu_fid_eq(&car->car_hai->hai_fid, &car->car_hai->hai_dfid))
1483                 car->car_hai->hai_dfid = pgs->hpk_fid;
1484
1485         if ((car->car_hai->hai_action == HSMA_RESTORE ||
1486              car->car_hai->hai_action == HSMA_ARCHIVE) &&
1487             (!lu_fid_eq(&pgs->hpk_fid, &car->car_hai->hai_dfid) &&
1488              !lu_fid_eq(&pgs->hpk_fid, &car->car_hai->hai_fid))) {
1489                 CERROR("%s: Progress on "DFID" for cookie "LPX64
1490                        " does not match request FID "DFID" nor data FID "
1491                        DFID"\n",
1492                        mdt_obd_name(mdt),
1493                        PFID(&pgs->hpk_fid), pgs->hpk_cookie,
1494                        PFID(&car->car_hai->hai_fid),
1495                        PFID(&car->car_hai->hai_dfid));
1496                 GOTO(out, rc = -EINVAL);
1497         }
1498
1499         if (pgs->hpk_errval != 0 && !(pgs->hpk_flags & HP_FLAG_COMPLETED)) {
1500                 CERROR("%s: Progress on "DFID" for cookie "LPX64" action=%s"
1501                        " is not coherent (err=%d and not completed"
1502                        " (flags=%d))\n",
1503                        mdt_obd_name(mdt),
1504                        PFID(&pgs->hpk_fid), pgs->hpk_cookie,
1505                        hsm_copytool_action2name(car->car_hai->hai_action),
1506                        pgs->hpk_errval, pgs->hpk_flags);
1507                 GOTO(out, rc = -EINVAL);
1508         }
1509
1510         /* now progress is valid */
1511
1512         /* we use a root like ucred */
1513         hsm_init_ucred(mdt_ucred(mti));
1514
1515         if (pgs->hpk_flags & HP_FLAG_COMPLETED) {
1516                 enum agent_req_status    status;
1517
1518                 rc = hsm_cdt_request_completed(mti, pgs, car, &status);
1519
1520                 /* remove request from memory list */
1521                 mdt_cdt_remove_request(cdt, pgs->hpk_cookie);
1522
1523                 CDEBUG(D_HSM, "Updating record: fid="DFID" cookie="LPX64
1524                               " action=%s status=%s\n", PFID(&pgs->hpk_fid),
1525                        pgs->hpk_cookie,
1526                        hsm_copytool_action2name(car->car_hai->hai_action),
1527                        agent_req_status2name(status));
1528
1529                 if (update_record) {
1530                         int rc1;
1531
1532                         rc1 = mdt_agent_record_update(mti->mti_env, mdt,
1533                                                      &pgs->hpk_cookie, 1,
1534                                                      status);
1535                         if (rc1)
1536                                 CERROR("%s: mdt_agent_record_update() failed,"
1537                                        " rc=%d, cannot update status to %s"
1538                                        " for cookie "LPX64"\n",
1539                                        mdt_obd_name(mdt), rc1,
1540                                        agent_req_status2name(status),
1541                                        pgs->hpk_cookie);
1542                         rc = (rc != 0 ? rc : rc1);
1543                 }
1544                 /* ct has completed a request, so a slot is available, wakeup
1545                  * cdt to find new work */
1546                 mdt_hsm_cdt_wakeup(mdt);
1547         } else {
1548                 /* if copytool send a progress on a canceled request
1549                  * we inform copytool it should stop
1550                  */
1551                 if (car->car_canceled == 1)
1552                         rc = -ECANCELED;
1553         }
1554         GOTO(out, rc);
1555
1556 out:
1557         /* remove ref got from mdt_cdt_update_request() */
1558         mdt_cdt_put_request(car);
1559
1560         return rc;
1561 }
1562
1563
1564 /**
1565  * data passed to llog_cat_process() callback
1566  * to cancel requests
1567  */
1568 struct hsm_cancel_all_data {
1569         struct mdt_device       *mdt;
1570 };
1571
1572 /**
1573  *  llog_cat_process() callback, used to:
1574  *  - purge all requests
1575  * \param env [IN] environment
1576  * \param llh [IN] llog handle
1577  * \param hdr [IN] llog record
1578  * \param data [IN] cb data = struct hsm_cancel_all_data
1579  * \retval 0 success
1580  * \retval -ve failure
1581  */
1582 static int mdt_cancel_all_cb(const struct lu_env *env,
1583                              struct llog_handle *llh,
1584                              struct llog_rec_hdr *hdr, void *data)
1585 {
1586         struct llog_agent_req_rec       *larr;
1587         struct hsm_cancel_all_data      *hcad;
1588         int                              rc = 0;
1589         ENTRY;
1590
1591         larr = (struct llog_agent_req_rec *)hdr;
1592         hcad = data;
1593         if (larr->arr_status == ARS_WAITING ||
1594             larr->arr_status == ARS_STARTED) {
1595                 larr->arr_status = ARS_CANCELED;
1596                 larr->arr_req_change = cfs_time_current_sec();
1597                 rc = mdt_agent_llog_update_rec(env, hcad->mdt, llh, larr);
1598                 if (rc == 0)
1599                         RETURN(LLOG_DEL_RECORD);
1600         }
1601         RETURN(rc);
1602 }
1603
1604 /**
1605  * cancel all actions
1606  * \param obd [IN] MDT device
1607  */
1608 static int hsm_cancel_all_actions(struct mdt_device *mdt)
1609 {
1610         struct mdt_thread_info          *mti;
1611         struct coordinator              *cdt = &mdt->mdt_coordinator;
1612         struct cdt_agent_req            *car;
1613         struct hsm_action_list          *hal = NULL;
1614         struct hsm_action_item          *hai;
1615         struct hsm_cancel_all_data       hcad;
1616         int                              hal_sz = 0, hal_len, rc;
1617         enum cdt_states                  save_state;
1618         ENTRY;
1619
1620         /* retrieve coordinator context */
1621         mti = lu_context_key_get(&cdt->cdt_env.le_ctx, &mdt_thread_key);
1622
1623         /* disable coordinator */
1624         save_state = cdt->cdt_state;
1625         cdt->cdt_state = CDT_DISABLE;
1626
1627         /* send cancel to all running requests */
1628         down_read(&cdt->cdt_request_lock);
1629         list_for_each_entry(car, &cdt->cdt_requests, car_request_list) {
1630                 mdt_cdt_get_request(car);
1631                 /* request is not yet removed from list, it will be done
1632                  * when copytool will return progress
1633                  */
1634
1635                 if (car->car_hai->hai_action == HSMA_CANCEL) {
1636                         mdt_cdt_put_request(car);
1637                         continue;
1638                 }
1639
1640                 /* needed size */
1641                 hal_len = sizeof(*hal) + cfs_size_round(MTI_NAME_MAXLEN + 1) +
1642                           cfs_size_round(car->car_hai->hai_len);
1643
1644                 if (hal_len > hal_sz && hal_sz > 0) {
1645                         /* not enough room, free old buffer */
1646                         OBD_FREE(hal, hal_sz);
1647                         hal = NULL;
1648                 }
1649
1650                 /* empty buffer, allocate one */
1651                 if (hal == NULL) {
1652                         hal_sz = hal_len;
1653                         OBD_ALLOC(hal, hal_sz);
1654                         if (hal == NULL) {
1655                                 mdt_cdt_put_request(car);
1656                                 up_read(&cdt->cdt_request_lock);
1657                                 GOTO(out, rc = -ENOMEM);
1658                         }
1659                 }
1660
1661                 hal->hal_version = HAL_VERSION;
1662                 obd_uuid2fsname(hal->hal_fsname, mdt_obd_name(mdt),
1663                                 MTI_NAME_MAXLEN);
1664                 hal->hal_fsname[MTI_NAME_MAXLEN] = '\0';
1665                 hal->hal_compound_id = car->car_compound_id;
1666                 hal->hal_archive_id = car->car_archive_id;
1667                 hal->hal_flags = car->car_flags;
1668                 hal->hal_count = 0;
1669
1670                 hai = hai_first(hal);
1671                 memcpy(hai, car->car_hai, car->car_hai->hai_len);
1672                 hai->hai_action = HSMA_CANCEL;
1673                 hal->hal_count = 1;
1674
1675                 /* it is possible to safely call mdt_hsm_agent_send()
1676                  * (ie without a deadlock on cdt_request_lock), because the
1677                  * write lock is taken only if we are not in purge mode
1678                  * (mdt_hsm_agent_send() does not call mdt_cdt_add_request()
1679                  *   nor mdt_cdt_remove_request())
1680                  */
1681                 /* no conflict with cdt thread because cdt is disable and we
1682                  * have the request lock */
1683                 mdt_hsm_agent_send(mti, hal, 1);
1684
1685                 mdt_cdt_put_request(car);
1686         }
1687         up_read(&cdt->cdt_request_lock);
1688
1689         if (hal != NULL)
1690                 OBD_FREE(hal, hal_sz);
1691
1692         /* cancel all on-disk records */
1693         hcad.mdt = mdt;
1694
1695         rc = cdt_llog_process(mti->mti_env, mti->mti_mdt,
1696                               mdt_cancel_all_cb, &hcad);
1697 out:
1698         /* enable coordinator */
1699         cdt->cdt_state = save_state;
1700
1701         RETURN(rc);
1702 }
1703
1704 /**
1705  * check if a request is comptaible with file status
1706  * \param hai [IN] request description
1707  * \param hal_an [IN] request archive number (not used)
1708  * \param rq_flags [IN] request flags
1709  * \param hsm [IN] file HSM metadata
1710  * \retval boolean
1711  */
1712 bool mdt_hsm_is_action_compat(const struct hsm_action_item *hai,
1713                               const int hal_an, const __u64 rq_flags,
1714                               const struct md_hsm *hsm)
1715 {
1716         int      is_compat = false;
1717         int      hsm_flags;
1718         ENTRY;
1719
1720         hsm_flags = hsm->mh_flags;
1721         switch (hai->hai_action) {
1722         case HSMA_ARCHIVE:
1723                 if (!(hsm_flags & HS_NOARCHIVE) &&
1724                     (hsm_flags & HS_DIRTY || !(hsm_flags & HS_ARCHIVED)))
1725                         is_compat = true;
1726                 break;
1727         case HSMA_RESTORE:
1728                 if (!(hsm_flags & HS_DIRTY) && (hsm_flags & HS_RELEASED) &&
1729                     hsm_flags & HS_ARCHIVED && !(hsm_flags & HS_LOST))
1730                         is_compat = true;
1731                 break;
1732         case HSMA_REMOVE:
1733                 if (!(hsm_flags & HS_RELEASED) &&
1734                     (hsm_flags & (HS_ARCHIVED | HS_EXISTS)))
1735                         is_compat = true;
1736                 break;
1737         case HSMA_CANCEL:
1738                 is_compat = true;
1739                 break;
1740         }
1741         CDEBUG(D_HSM, "fid="DFID" action=%s flags="LPX64
1742                       " extent="LPX64"-"LPX64" hsm_flags=%.8X %s\n",
1743                       PFID(&hai->hai_fid),
1744                       hsm_copytool_action2name(hai->hai_action), rq_flags,
1745                       hai->hai_extent.offset, hai->hai_extent.length,
1746                       hsm->mh_flags,
1747                       (is_compat ? "compatible" : "uncompatible"));
1748
1749         RETURN(is_compat);
1750 }
1751
1752 /*
1753  * /proc interface used to get/set HSM behaviour (cdt->cdt_policy)
1754  */
1755 static const struct {
1756         __u64            bit;
1757         char            *name;
1758         char            *nickname;
1759 } hsm_policy_names[] = {
1760         { CDT_NONBLOCKING_RESTORE,      "NonBlockingRestore",   "NBR"},
1761         { CDT_NORETRY_ACTION,           "NoRetryAction",        "NRA"},
1762         { 0 },
1763 };
1764
1765 /**
1766  * convert a policy name to a bit
1767  * \param name [IN] policy name
1768  * \retval 0 unknown
1769  * \retval   policy bit
1770  */
1771 static __u64 hsm_policy_str2bit(const char *name)
1772 {
1773         int      i;
1774
1775         for (i = 0; hsm_policy_names[i].bit != 0; i++)
1776                 if (strcmp(hsm_policy_names[i].nickname, name) == 0 ||
1777                     strcmp(hsm_policy_names[i].name, name) == 0)
1778                         return hsm_policy_names[i].bit;
1779         return 0;
1780 }
1781
1782 /**
1783  * convert a policy bit field to a string
1784  * \param mask [IN] policy bit field
1785  * \param hexa [IN] print mask before bit names
1786  * \param buffer [OUT] string
1787  * \param count [IN] size of buffer
1788  */
1789 static void hsm_policy_bit2str(struct seq_file *m, const __u64 mask,
1790                                 const bool hexa)
1791 {
1792         int      i, j;
1793         __u64    bit;
1794         ENTRY;
1795
1796         if (hexa)
1797                 seq_printf(m, "("LPX64") ", mask);
1798
1799         for (i = 0; i < CDT_POLICY_SHIFT_COUNT; i++) {
1800                 bit = (1ULL << i);
1801
1802                 for (j = 0; hsm_policy_names[j].bit != 0; j++) {
1803                         if (hsm_policy_names[j].bit == bit)
1804                                 break;
1805                 }
1806                 if (bit & mask)
1807                         seq_printf(m, "[%s] ", hsm_policy_names[j].name);
1808                 else
1809                         seq_printf(m, "%s ", hsm_policy_names[j].name);
1810         }
1811         /* remove last ' ' */
1812         m->count--;
1813         seq_putc(m, '\0');
1814 }
1815
1816 /* methods to read/write HSM policy flags */
1817 static int mdt_hsm_policy_seq_show(struct seq_file *m, void *data)
1818 {
1819         struct mdt_device       *mdt = m->private;
1820         struct coordinator      *cdt = &mdt->mdt_coordinator;
1821         ENTRY;
1822
1823         hsm_policy_bit2str(m, cdt->cdt_policy, false);
1824         RETURN(0);
1825 }
1826
1827 static ssize_t
1828 mdt_hsm_policy_seq_write(struct file *file, const char __user *buffer,
1829                          size_t count, loff_t *off)
1830 {
1831         struct seq_file         *m = file->private_data;
1832         struct mdt_device       *mdt = m->private;
1833         struct coordinator      *cdt = &mdt->mdt_coordinator;
1834         char                    *start, *token, sign;
1835         char                    *buf;
1836         __u64                    policy;
1837         __u64                    add_mask, remove_mask, set_mask;
1838         int                      rc;
1839         ENTRY;
1840
1841         if (count + 1 > PAGE_SIZE)
1842                 RETURN(-EINVAL);
1843
1844         OBD_ALLOC(buf, count + 1);
1845         if (buf == NULL)
1846                 RETURN(-ENOMEM);
1847
1848         if (copy_from_user(buf, buffer, count))
1849                 GOTO(out, rc = -EFAULT);
1850
1851         buf[count] = '\0';
1852
1853         start = buf;
1854         CDEBUG(D_HSM, "%s: receive new policy: '%s'\n", mdt_obd_name(mdt),
1855                start);
1856
1857         add_mask = remove_mask = set_mask = 0;
1858         do {
1859                 token = strsep(&start, "\n ");
1860                 sign = *token;
1861
1862                 if (sign == '\0')
1863                         continue;
1864
1865                 if (sign == '-' || sign == '+')
1866                         token++;
1867
1868                 policy = hsm_policy_str2bit(token);
1869                 if (policy == 0) {
1870                         CWARN("%s: '%s' is unknown, "
1871                               "supported policies are:\n", mdt_obd_name(mdt),
1872                               token);
1873                         hsm_policy_bit2str(m, 0, false);
1874                         GOTO(out, rc = -EINVAL);
1875                 }
1876                 switch (sign) {
1877                 case '-':
1878                         remove_mask |= policy;
1879                         break;
1880                 case '+':
1881                         add_mask |= policy;
1882                         break;
1883                 default:
1884                         set_mask |= policy;
1885                         break;
1886                 }
1887
1888         } while (start != NULL);
1889
1890         CDEBUG(D_HSM, "%s: new policy: rm="LPX64" add="LPX64" set="LPX64"\n",
1891                mdt_obd_name(mdt), remove_mask, add_mask, set_mask);
1892
1893         /* if no sign in all string, it is a clear and set
1894          * if some sign found, all unsigned are converted
1895          * to add
1896          * P1 P2 = set to P1 and P2
1897          * P1 -P2 = add P1 clear P2 same as +P1 -P2
1898          */
1899         if (remove_mask == 0 && add_mask == 0) {
1900                 cdt->cdt_policy = set_mask;
1901         } else {
1902                 cdt->cdt_policy |= set_mask | add_mask;
1903                 cdt->cdt_policy &= ~remove_mask;
1904         }
1905
1906         GOTO(out, rc = count);
1907
1908 out:
1909         OBD_FREE(buf, count + 1);
1910         RETURN(rc);
1911 }
1912 LPROC_SEQ_FOPS(mdt_hsm_policy);
1913
1914 #define GENERATE_PROC_METHOD(VAR)                                       \
1915 static int mdt_hsm_##VAR##_seq_show(struct seq_file *m, void *data)     \
1916 {                                                                       \
1917         struct mdt_device       *mdt = m->private;                      \
1918         struct coordinator      *cdt = &mdt->mdt_coordinator;           \
1919         ENTRY;                                                          \
1920                                                                         \
1921         seq_printf(m, LPU64"\n", (__u64)cdt->VAR);                      \
1922         RETURN(0);                                                      \
1923 }                                                                       \
1924 static ssize_t                                                          \
1925 mdt_hsm_##VAR##_seq_write(struct file *file, const char __user *buffer, \
1926                           size_t count, loff_t *off)                    \
1927                                                                         \
1928 {                                                                       \
1929         struct seq_file         *m = file->private_data;                \
1930         struct mdt_device       *mdt = m->private;                      \
1931         struct coordinator      *cdt = &mdt->mdt_coordinator;           \
1932         int                      val;                                   \
1933         int                      rc;                                    \
1934         ENTRY;                                                          \
1935                                                                         \
1936         rc = lprocfs_write_helper(buffer, count, &val);                 \
1937         if (rc)                                                         \
1938                 RETURN(rc);                                             \
1939         if (val > 0) {                                                  \
1940                 cdt->VAR = val;                                         \
1941                 RETURN(count);                                          \
1942         }                                                               \
1943         RETURN(-EINVAL);                                                \
1944 }                                                                       \
1945
1946 GENERATE_PROC_METHOD(cdt_loop_period)
1947 GENERATE_PROC_METHOD(cdt_grace_delay)
1948 GENERATE_PROC_METHOD(cdt_active_req_timeout)
1949 GENERATE_PROC_METHOD(cdt_max_requests)
1950 GENERATE_PROC_METHOD(cdt_default_archive_id)
1951
1952 /*
1953  * procfs write method for MDT/hsm_control
1954  * proc entry is in mdt directory so data is mdt obd_device pointer
1955  */
1956 #define CDT_ENABLE_CMD   "enabled"
1957 #define CDT_STOP_CMD     "shutdown"
1958 #define CDT_DISABLE_CMD  "disabled"
1959 #define CDT_PURGE_CMD    "purge"
1960 #define CDT_HELP_CMD     "help"
1961 #define CDT_MAX_CMD_LEN  10
1962
1963 ssize_t
1964 mdt_hsm_cdt_control_seq_write(struct file *file, const char __user *buffer,
1965                               size_t count, loff_t *off)
1966 {
1967         struct seq_file         *m = file->private_data;
1968         struct obd_device       *obd = m->private;
1969         struct mdt_device       *mdt = mdt_dev(obd->obd_lu_dev);
1970         struct coordinator      *cdt = &(mdt->mdt_coordinator);
1971         int                      rc, usage = 0;
1972         char                     kernbuf[CDT_MAX_CMD_LEN];
1973         ENTRY;
1974
1975         if (count == 0 || count >= sizeof(kernbuf))
1976                 RETURN(-EINVAL);
1977
1978         if (copy_from_user(kernbuf, buffer, count))
1979                 RETURN(-EFAULT);
1980         kernbuf[count] = 0;
1981
1982         if (kernbuf[count - 1] == '\n')
1983                 kernbuf[count - 1] = 0;
1984
1985         rc = 0;
1986         if (strcmp(kernbuf, CDT_ENABLE_CMD) == 0) {
1987                 if (cdt->cdt_state == CDT_DISABLE) {
1988                         cdt->cdt_state = CDT_RUNNING;
1989                         mdt_hsm_cdt_wakeup(mdt);
1990                 } else {
1991                         rc = mdt_hsm_cdt_start(mdt);
1992                 }
1993         } else if (strcmp(kernbuf, CDT_STOP_CMD) == 0) {
1994                 if ((cdt->cdt_state == CDT_STOPPING) ||
1995                     (cdt->cdt_state == CDT_STOPPED)) {
1996                         CERROR("%s: Coordinator already stopped\n",
1997                                mdt_obd_name(mdt));
1998                         rc = -EALREADY;
1999                 } else {
2000                         cdt->cdt_state = CDT_STOPPING;
2001                 }
2002         } else if (strcmp(kernbuf, CDT_DISABLE_CMD) == 0) {
2003                 if ((cdt->cdt_state == CDT_STOPPING) ||
2004                     (cdt->cdt_state == CDT_STOPPED)) {
2005                         CERROR("%s: Coordinator is stopped\n",
2006                                mdt_obd_name(mdt));
2007                         rc = -EINVAL;
2008                 } else {
2009                         cdt->cdt_state = CDT_DISABLE;
2010                 }
2011         } else if (strcmp(kernbuf, CDT_PURGE_CMD) == 0) {
2012                 rc = hsm_cancel_all_actions(mdt);
2013         } else if (strcmp(kernbuf, CDT_HELP_CMD) == 0) {
2014                 usage = 1;
2015         } else {
2016                 usage = 1;
2017                 rc = -EINVAL;
2018         }
2019
2020         if (usage == 1)
2021                 CERROR("%s: Valid coordinator control commands are: "
2022                        "%s %s %s %s %s\n", mdt_obd_name(mdt),
2023                        CDT_ENABLE_CMD, CDT_STOP_CMD, CDT_DISABLE_CMD,
2024                        CDT_PURGE_CMD, CDT_HELP_CMD);
2025
2026         if (rc)
2027                 RETURN(rc);
2028
2029         RETURN(count);
2030 }
2031
2032 int mdt_hsm_cdt_control_seq_show(struct seq_file *m, void *data)
2033 {
2034         struct obd_device       *obd = m->private;
2035         struct coordinator      *cdt;
2036         ENTRY;
2037
2038         cdt = &(mdt_dev(obd->obd_lu_dev)->mdt_coordinator);
2039
2040         if (cdt->cdt_state == CDT_INIT)
2041                 seq_printf(m, "init\n");
2042         else if (cdt->cdt_state == CDT_RUNNING)
2043                 seq_printf(m, "enabled\n");
2044         else if (cdt->cdt_state == CDT_STOPPING)
2045                 seq_printf(m, "stopping\n");
2046         else if (cdt->cdt_state == CDT_STOPPED)
2047                 seq_printf(m, "stopped\n");
2048         else if (cdt->cdt_state == CDT_DISABLE)
2049                 seq_printf(m, "disabled\n");
2050         else
2051                 seq_printf(m, "unknown\n");
2052
2053         RETURN(0);
2054 }
2055
2056 static int
2057 mdt_hsm_request_mask_show(struct seq_file *m, __u64 mask)
2058 {
2059         int i, rc = 0;
2060         ENTRY;
2061
2062         for (i = 0; i < 8 * sizeof(mask); i++) {
2063                 if (mask & (1UL << i))
2064                         rc += seq_printf(m, "%s%s", rc == 0 ? "" : " ",
2065                                         hsm_copytool_action2name(i));
2066         }
2067         rc += seq_printf(m, "\n");
2068
2069         RETURN(rc);
2070 }
2071
2072 static int
2073 mdt_hsm_user_request_mask_seq_show(struct seq_file *m, void *data)
2074 {
2075         struct mdt_device *mdt = m->private;
2076         struct coordinator *cdt = &mdt->mdt_coordinator;
2077
2078         return mdt_hsm_request_mask_show(m, cdt->cdt_user_request_mask);
2079 }
2080
2081 static int
2082 mdt_hsm_group_request_mask_seq_show(struct seq_file *m, void *data)
2083 {
2084         struct mdt_device *mdt = m->private;
2085         struct coordinator *cdt = &mdt->mdt_coordinator;
2086
2087         return mdt_hsm_request_mask_show(m, cdt->cdt_group_request_mask);
2088 }
2089
2090 static int
2091 mdt_hsm_other_request_mask_seq_show(struct seq_file *m, void *data)
2092 {
2093         struct mdt_device *mdt = m->private;
2094         struct coordinator *cdt = &mdt->mdt_coordinator;
2095
2096         return mdt_hsm_request_mask_show(m, cdt->cdt_other_request_mask);
2097 }
2098
2099 static inline enum hsm_copytool_action
2100 hsm_copytool_name2action(const char *name)
2101 {
2102         if (strcasecmp(name, "NOOP") == 0)
2103                 return HSMA_NONE;
2104         else if (strcasecmp(name, "ARCHIVE") == 0)
2105                 return HSMA_ARCHIVE;
2106         else if (strcasecmp(name, "RESTORE") == 0)
2107                 return HSMA_RESTORE;
2108         else if (strcasecmp(name, "REMOVE") == 0)
2109                 return HSMA_REMOVE;
2110         else if (strcasecmp(name, "CANCEL") == 0)
2111                 return HSMA_CANCEL;
2112         else
2113                 return -1;
2114 }
2115
2116 static ssize_t
2117 mdt_write_hsm_request_mask(struct file *file, const char __user *user_buf,
2118                             size_t user_count, __u64 *mask)
2119 {
2120         char *buf, *pos, *name;
2121         size_t buf_size;
2122         __u64 new_mask = 0;
2123         int rc;
2124         ENTRY;
2125
2126         if (!(user_count < 4096))
2127                 RETURN(-ENOMEM);
2128
2129         buf_size = user_count + 1;
2130
2131         OBD_ALLOC(buf, buf_size);
2132         if (buf == NULL)
2133                 RETURN(-ENOMEM);
2134
2135         if (copy_from_user(buf, user_buf, buf_size - 1))
2136                 GOTO(out, rc = -EFAULT);
2137
2138         buf[buf_size - 1] = '\0';
2139
2140         pos = buf;
2141         while ((name = strsep(&pos, " \t\v\n")) != NULL) {
2142                 int action;
2143
2144                 if (*name == '\0')
2145                         continue;
2146
2147                 action = hsm_copytool_name2action(name);
2148                 if (action < 0)
2149                         GOTO(out, rc = -EINVAL);
2150
2151                 new_mask |= (1UL << action);
2152         }
2153
2154         *mask = new_mask;
2155         rc = user_count;
2156 out:
2157         OBD_FREE(buf, buf_size);
2158
2159         RETURN(rc);
2160 }
2161
2162 static ssize_t
2163 mdt_hsm_user_request_mask_seq_write(struct file *file, const char __user *buf,
2164                                         size_t count, loff_t *off)
2165 {
2166         struct seq_file         *m = file->private_data;
2167         struct mdt_device       *mdt = m->private;
2168         struct coordinator *cdt = &mdt->mdt_coordinator;
2169
2170         return mdt_write_hsm_request_mask(file, buf, count,
2171                                            &cdt->cdt_user_request_mask);
2172 }
2173
2174 static ssize_t
2175 mdt_hsm_group_request_mask_seq_write(struct file *file, const char __user *buf,
2176                                         size_t count, loff_t *off)
2177 {
2178         struct seq_file         *m = file->private_data;
2179         struct mdt_device       *mdt = m->private;
2180         struct coordinator      *cdt = &mdt->mdt_coordinator;
2181
2182         return mdt_write_hsm_request_mask(file, buf, count,
2183                                            &cdt->cdt_group_request_mask);
2184 }
2185
2186 static ssize_t
2187 mdt_hsm_other_request_mask_seq_write(struct file *file, const char __user *buf,
2188                                         size_t count, loff_t *off)
2189 {
2190         struct seq_file         *m = file->private_data;
2191         struct mdt_device       *mdt = m->private;
2192         struct coordinator      *cdt = &mdt->mdt_coordinator;
2193
2194         return mdt_write_hsm_request_mask(file, buf, count,
2195                                            &cdt->cdt_other_request_mask);
2196 }
2197
2198 LPROC_SEQ_FOPS(mdt_hsm_cdt_loop_period);
2199 LPROC_SEQ_FOPS(mdt_hsm_cdt_grace_delay);
2200 LPROC_SEQ_FOPS(mdt_hsm_cdt_active_req_timeout);
2201 LPROC_SEQ_FOPS(mdt_hsm_cdt_max_requests);
2202 LPROC_SEQ_FOPS(mdt_hsm_cdt_default_archive_id);
2203 LPROC_SEQ_FOPS(mdt_hsm_user_request_mask);
2204 LPROC_SEQ_FOPS(mdt_hsm_group_request_mask);
2205 LPROC_SEQ_FOPS(mdt_hsm_other_request_mask);
2206
2207 static struct lprocfs_seq_vars lprocfs_mdt_hsm_vars[] = {
2208         { .name =       "agents",
2209           .fops =       &mdt_hsm_agent_fops                     },
2210         { .name =       "actions",
2211           .fops =       &mdt_hsm_actions_fops,
2212           .proc_mode =  0444                                    },
2213         { .name =       "default_archive_id",
2214           .fops =       &mdt_hsm_cdt_default_archive_id_fops    },
2215         { .name =       "grace_delay",
2216           .fops =       &mdt_hsm_cdt_grace_delay_fops           },
2217         { .name =       "loop_period",
2218           .fops =       &mdt_hsm_cdt_loop_period_fops           },
2219         { .name =       "max_requests",
2220           .fops =       &mdt_hsm_cdt_max_requests_fops          },
2221         { .name =       "policy",
2222           .fops =       &mdt_hsm_policy_fops                    },
2223         { .name =       "active_request_timeout",
2224           .fops =       &mdt_hsm_cdt_active_req_timeout_fops    },
2225         { .name =       "active_requests",
2226           .fops =       &mdt_hsm_active_requests_fops           },
2227         { .name =       "user_request_mask",
2228           .fops =       &mdt_hsm_user_request_mask_fops,        },
2229         { .name =       "group_request_mask",
2230           .fops =       &mdt_hsm_group_request_mask_fops,       },
2231         { .name =       "other_request_mask",
2232           .fops =       &mdt_hsm_other_request_mask_fops,       },
2233         { 0 }
2234 };