lustre/quota/qsd_lib.c

   1 /*
   2  * GPL HEADER START
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * This program is free software; you can redistribute it and/or modify
   7  * it under the terms of the GNU General Public License version 2 only,
   8  * as published by the Free Software Foundation.
   9  *
  10  * This program is distributed in the hope that it will be useful, but
  11  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * General Public License version 2 for more details (a copy is included
  14  * in the LICENSE file that accompanied this code).
  15  *
  16  * You should have received a copy of the GNU General Public License
  17  * version 2 along with this program; if not, write to the
  18  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  19  * Boston, MA 021110-1307, USA
  20  *
  21  * GPL HEADER END
  22  */
  23 /*
  24  * Copyright (c) 2012, 2017, Intel Corporation.
  25  * Use is subject to license terms.
  26  *
  27  * Author: Johann Lombardi <johann.lombardi@intel.com>
  28  * Author: Niu    Yawei    <yawei.niu@intel.com>
  29  */
  30
  31 /*
  32  * Quota Slave Driver (QSD) management.
  33  *
  34  * The quota slave feature is implemented under the form of a library called
  35  * QSD. Each OSD device should create a QSD instance via qsd_init() which will
  36  * be used to manage quota enforcement for this device. This implies:
  37  * - completing the reintegration procedure with the quota master (aka QMT, see
  38  *   qmt_dev.c) to retrieve the latest quota settings and space distribution.
  39  * - managing quota locks in order to be notified of configuration changes.
  40  * - acquiring space from the QMT when quota space for a given user/group is
  41  *   close to exhaustion.
  42  * - allocating quota space to service threads for local request processing.
  43  *
  44  * Once the QSD instance created, the OSD device should invoke qsd_start()
  45  * when recovery is completed. This notifies the QSD that we are about to
  46  * process new requests on which quota should be strictly enforced.
  47  * Then, qsd_op_begin/end can be used to reserve/release/pre-acquire quota space
  48  * for/after each operation until shutdown where the QSD instance should be
  49  * freed via qsd_fini().
  50  */
  51
  52 #define DEBUG_SUBSYSTEM S_LQUOTA
  53
  54 #include <obd_class.h>
  55 #include "qsd_internal.h"
  56
  57 struct kmem_cache *upd_kmem;
  58
  59 struct lu_kmem_descr qsd_caches[] = {
  60         {
  61                 .ckd_cache = &upd_kmem,
  62                 .ckd_name  = "upd_kmem",
  63                 .ckd_size  = sizeof(struct qsd_upd_rec)
  64         },
  65         {
  66                 .ckd_cache = NULL
  67         }
  68 };
  69
  70 /* define qsd thread key */
  71 LU_KEY_INIT_FINI(qsd, struct qsd_thread_info);
  72 LU_CONTEXT_KEY_DEFINE(qsd, LCT_MD_THREAD | LCT_MG_THREAD | LCT_DT_THREAD | LCT_LOCAL);
  73 LU_KEY_INIT_GENERIC(qsd);
  74
  75 /* some procfs helpers */
  76 static int qsd_state_seq_show(struct seq_file *m, void *data)
  77 {
  78         struct qsd_instance     *qsd = m->private;
  79         char                     enabled[5];
  80
  81         LASSERT(qsd != NULL);
  82
  83         memset(enabled, 0, sizeof(enabled));
  84         if (qsd_type_enabled(qsd, USRQUOTA))
  85                 strcat(enabled, "u");
  86         if (qsd_type_enabled(qsd, GRPQUOTA))
  87                 strcat(enabled, "g");
  88         if (qsd_type_enabled(qsd, PRJQUOTA))
  89                 strncat(enabled, "p", 1);
  90         if (strlen(enabled) == 0)
  91                 strcat(enabled, "none");
  92
  93         /* TODO: further pool ID should be removed or
  94          * replaced with pool Name */
  95         seq_printf(m, "target name:    %s\n"
  96                    "pool ID:        %d\n"
  97                    "type:           %s\n"
  98                    "quota enabled:  %s\n"
  99                    "conn to master: %s\n",
 100                    qsd->qsd_svname, 0,
 101                    qsd->qsd_is_md ? "md" : "dt", enabled,
 102                    qsd->qsd_exp_valid ? "setup" : "not setup yet");
 103
 104         if (qsd->qsd_prepared) {
 105                 memset(enabled, 0, sizeof(enabled));
 106                 if (qsd->qsd_type_array[USRQUOTA]->qqi_acct_obj != NULL)
 107                         strcat(enabled, "u");
 108                 if (qsd->qsd_type_array[GRPQUOTA]->qqi_acct_obj != NULL)
 109                         strcat(enabled, "g");
 110                 if (qsd->qsd_type_array[PRJQUOTA]->qqi_acct_obj != NULL)
 111                         strncat(enabled, "p", 1);
 112                 if (strlen(enabled) == 0)
 113                         strcat(enabled, "none");
 114                 seq_printf(m, "space acct:     %s\n"
 115                            "user uptodate:  glb[%d],slv[%d],reint[%d]\n"
 116                            "group uptodate: glb[%d],slv[%d],reint[%d]\n"
 117                            "project uptodate: glb[%d],slv[%d],reint[%d]\n",
 118                            enabled,
 119                            qsd->qsd_type_array[USRQUOTA]->qqi_glb_uptodate,
 120                            qsd->qsd_type_array[USRQUOTA]->qqi_slv_uptodate,
 121                            qsd->qsd_type_array[USRQUOTA]->qqi_reint,
 122                            qsd->qsd_type_array[GRPQUOTA]->qqi_glb_uptodate,
 123                            qsd->qsd_type_array[GRPQUOTA]->qqi_slv_uptodate,
 124                            qsd->qsd_type_array[GRPQUOTA]->qqi_reint,
 125                            qsd->qsd_type_array[PRJQUOTA]->qqi_glb_uptodate,
 126                            qsd->qsd_type_array[PRJQUOTA]->qqi_slv_uptodate,
 127                            qsd->qsd_type_array[PRJQUOTA]->qqi_reint);
 128         }
 129         return 0;
 130 }
 131 LPROC_SEQ_FOPS_RO(qsd_state);
 132
 133 static int qsd_enabled_seq_show(struct seq_file *m, void *data)
 134 {
 135         struct qsd_instance     *qsd = m->private;
 136         char                     enabled[5];
 137
 138         LASSERT(qsd != NULL);
 139
 140         memset(enabled, 0, sizeof(enabled));
 141         if (qsd_type_enabled(qsd, USRQUOTA))
 142                 strncat(enabled, "u", sizeof(enabled) - strlen(enabled));
 143         if (qsd_type_enabled(qsd, GRPQUOTA))
 144                 strncat(enabled, "g", sizeof(enabled) - strlen(enabled));
 145         if (qsd_type_enabled(qsd, PRJQUOTA))
 146                 strncat(enabled, "p", sizeof(enabled) - strlen(enabled));
 147         if (strlen(enabled) == 0)
 148                 strncat(enabled, "none", sizeof(enabled) - strlen(enabled));
 149
 150         seq_printf(m, "%s\n", enabled);
 151         return 0;
 152 }
 153
 154 static ssize_t qsd_enabled_seq_write(struct file *file,
 155                                      const char __user *buffer,
 156                                      size_t count, loff_t *off)
 157 {
 158         struct seq_file *m = file->private_data;
 159         struct qsd_instance *qsd = m->private;
 160         char fsname[LUSTRE_MAXFSNAME + 1];
 161         int enabled = 0;
 162         char valstr[5];
 163         int pool, rc;
 164
 165         if (count > 4)
 166                 return -E2BIG;
 167
 168         if (copy_from_user(valstr, buffer, count))
 169                 GOTO(out, count = -EFAULT);
 170
 171         valstr[sizeof(valstr) - 1] = 0;
 172         if (strchr(valstr, 'u'))
 173                 enabled |= BIT(USRQUOTA);
 174         if (strchr(valstr, 'g'))
 175                 enabled |= BIT(GRPQUOTA);
 176         if (strchr(valstr, 'p'))
 177                 enabled |= BIT(PRJQUOTA);
 178
 179         if (enabled == 0 && strcmp(valstr, "none"))
 180                 GOTO(out, count = -EINVAL);
 181
 182         if (qsd->qsd_is_md)
 183                 pool = LQUOTA_RES_MD;
 184         else
 185                 pool = LQUOTA_RES_DT;
 186
 187         if (server_name2fsname(qsd->qsd_svname, fsname, NULL))
 188                 GOTO(out, count = -EINVAL);
 189
 190         rc = qsd_config(valstr, fsname, pool);
 191         if (rc)
 192                 count = rc;
 193 out:
 194         return count;
 195 }
 196 LPROC_SEQ_FOPS(qsd_enabled);
 197
 198 /* force reintegration procedure to be executed.
 199  * Used for test/debugging purpose */
 200 static ssize_t
 201 lprocfs_force_reint_seq_write(struct file *file, const char __user *buffer,
 202                                 size_t count, loff_t *off)
 203 {
 204         struct seq_file     *m = file->private_data;
 205         struct qsd_instance *qsd = m->private;
 206         int                  rc = 0, qtype;
 207
 208         LASSERT(qsd != NULL);
 209
 210         write_lock(&qsd->qsd_lock);
 211         if (qsd->qsd_stopping) {
 212                 /* don't mess up with shutdown procedure, it is already
 213                  * complicated enough */
 214                 rc = -ESHUTDOWN;
 215         } else if (!qsd->qsd_prepared) {
 216                 rc = -EAGAIN;
 217         } else {
 218                 /* mark all indexes as stale */
 219                 for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++) {
 220                         qsd->qsd_type_array[qtype]->qqi_glb_uptodate = false;
 221                         qsd->qsd_type_array[qtype]->qqi_slv_uptodate = false;
 222                 }
 223         }
 224         write_unlock(&qsd->qsd_lock);
 225
 226         if (rc)
 227                 return rc;
 228
 229         /* kick off reintegration */
 230         for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++) {
 231                 rc = qsd_start_reint_thread(qsd->qsd_type_array[qtype]);
 232                 if (rc)
 233                         break;
 234         }
 235         return rc == 0 ? count : rc;
 236 }
 237 LPROC_SEQ_FOPS_WR_ONLY(qsd, force_reint);
 238
 239 static int qsd_timeout_seq_show(struct seq_file *m, void *data)
 240 {
 241         struct qsd_instance *qsd = m->private;
 242         LASSERT(qsd != NULL);
 243
 244         seq_printf(m, "%d\n", qsd_wait_timeout(qsd));
 245         return 0;
 246 }
 247
 248 static ssize_t
 249 qsd_timeout_seq_write(struct file *file, const char __user *buffer,
 250                         size_t count, loff_t *off)
 251 {
 252         struct seq_file *m = file->private_data;
 253         struct qsd_instance *qsd = m->private;
 254         time64_t timeout;
 255         int rc;
 256
 257         LASSERT(qsd != NULL);
 258         rc = kstrtoll_from_user(buffer, count, 0, &timeout);
 259         if (rc)
 260                 return rc;
 261
 262         if (timeout < 0)
 263                 return -EINVAL;
 264
 265         qsd->qsd_timeout = timeout;
 266         return count;
 267 }
 268 LPROC_SEQ_FOPS(qsd_timeout);
 269
 270 static int qsd_root_prj_enable_seq_show(struct seq_file *m, void *data)
 271 {
 272         struct qsd_instance *qsd = m->private;
 273
 274         LASSERT(qsd != NULL);
 275         seq_printf(m, "%d\n", qsd->qsd_root_prj_enable);
 276         return 0;
 277 }
 278
 279 static ssize_t
 280 qsd_root_prj_enable_seq_write(struct file *file, const char __user *buffer,
 281                         size_t count, loff_t *off)
 282 {
 283         struct seq_file *m = file->private_data;
 284         struct qsd_instance *qsd = m->private;
 285         bool enable;
 286         int rc;
 287
 288         LASSERT(qsd != NULL);
 289         rc = kstrtobool_from_user(buffer, count, &enable);
 290         if (rc)
 291                 return rc;
 292
 293         qsd->qsd_root_prj_enable = enable;
 294         return count;
 295 }
 296 LPROC_SEQ_FOPS(qsd_root_prj_enable);
 297
 298 static struct lprocfs_vars lprocfs_quota_qsd_vars[] = {
 299         { .name =       "info",
 300           .fops =       &qsd_state_fops         },
 301         { .name =       "enabled",
 302           .fops =       &qsd_enabled_fops       },
 303         { .name =       "force_reint",
 304           .fops =       &qsd_force_reint_fops   },
 305         { .name =       "timeout",
 306           .fops =       &qsd_timeout_fops       },
 307         { .name =       "root_prj_enable",
 308           .fops =       &qsd_root_prj_enable_fops       },
 309         { NULL }
 310 };
 311
 312 /*
 313  * Callback function invoked by the OSP layer when the connection to the master
 314  * has been set up.
 315  *
 316  * \param data - is a pointer to the qsd_instance
 317  *
 318  * \retval - 0 on success, appropriate error on failure
 319  */
 320 static int qsd_conn_callback(void *data)
 321 {
 322         struct qsd_instance *qsd = (struct qsd_instance *)data;
 323         int                  type;
 324         ENTRY;
 325
 326         /* qsd_exp should now be valid */
 327         LASSERT(qsd->qsd_exp);
 328
 329         qsd->qsd_ns = class_exp2obd(qsd->qsd_exp)->obd_namespace;
 330
 331         write_lock(&qsd->qsd_lock);
 332         /* notify that qsd_exp is now valid */
 333         qsd->qsd_exp_valid = true;
 334         write_unlock(&qsd->qsd_lock);
 335
 336         /* Now that the connection to master is setup, we can initiate the
 337          * reintegration procedure for quota types which are enabled.
 338          * It is worth noting that, if the qsd_instance hasn't been started
 339          * already, then we can only complete the first two steps of the
 340          * reintegration procedure (i.e. global lock enqueue and slave
 341          * index transfer) since the space usage reconciliation (i.e.
 342          * step 3) will have to wait for qsd_start() to be called */
 343         for (type = USRQUOTA; type < LL_MAXQUOTAS; type++) {
 344                 struct qsd_qtype_info *qqi = qsd->qsd_type_array[type];
 345                 struct task_struct *t;
 346
 347                 /* qqi_reint_task can be set to NULL at any time,
 348                  * so we need to be careful.
 349                  */
 350                 rcu_read_lock();
 351                 t = rcu_dereference(qqi->qqi_reint_task);
 352                 if (t)
 353                         wake_up_process(t);
 354                 rcu_read_unlock();
 355         }
 356
 357         RETURN(0);
 358 }
 359
 360 /*
 361  * Release qsd_qtype_info structure which contains data associated with a
 362  * given quota type. This releases the accounting objects.
 363  * It's called on OSD cleanup when the qsd instance is released.
 364  *
 365  * \param env - is the environment passed by the caller
 366  * \param qsd - is the qsd instance managing the qsd_qtype_info structure
 367  *              to be released
 368  * \param qtype - is the quota type to be shutdown
 369  */
 370 static void qsd_qtype_fini(const struct lu_env *env, struct qsd_instance *qsd,
 371                            int qtype)
 372 {
 373         struct qsd_qtype_info   *qqi;
 374         int repeat = 0;
 375         ENTRY;
 376
 377         if (qsd->qsd_type_array[qtype] == NULL)
 378                 RETURN_EXIT;
 379         qqi = qsd->qsd_type_array[qtype];
 380         qsd->qsd_type_array[qtype] = NULL;
 381
 382         /* all deferred work lists should be empty */
 383         LASSERT(list_empty(&qqi->qqi_deferred_glb));
 384         LASSERT(list_empty(&qqi->qqi_deferred_slv));
 385
 386         /* shutdown lquota site */
 387         if (qqi->qqi_site != NULL && !IS_ERR(qqi->qqi_site)) {
 388                 lquota_site_free(env, qqi->qqi_site);
 389                 qqi->qqi_site = NULL;
 390         }
 391
 392         /* The qqi may still be holding by global locks which are being
 393          * canceled asynchronously (LU-4365), see the following steps:
 394          *
 395          * - On server umount, we try to clear all quota locks first by
 396          *   disconnecting LWP (which will invalidate import and cleanup
 397          *   all locks on it), however, if quota reint process is holding
 398          *   the global lock for reintegration at that time, global lock
 399          *   will fail to be cleared on LWP disconnection.
 400          *
 401          * - Umount process goes on and stops reint process, the global
 402          *   lock will be dropped on reint process exit, however, the lock
 403          *   cancel in done in asynchronous way, so the
 404          *   qsd_glb_blocking_ast() might haven't been called yet when we
 405          *   get here.
 406          */
 407         while (atomic_read(&qqi->qqi_ref) > 1) {
 408                 CDEBUG(D_QUOTA, "qqi reference count %u, repeat: %d\n",
 409                        atomic_read(&qqi->qqi_ref), repeat);
 410                 repeat++;
 411                 schedule_timeout_interruptible(cfs_time_seconds(1));
 412         }
 413
 414         /* by now, all qqi users should have gone away */
 415         LASSERT(atomic_read(&qqi->qqi_ref) == 1);
 416         lu_ref_fini(&qqi->qqi_reference);
 417
 418         /* release accounting object */
 419         if (qqi->qqi_acct_obj != NULL && !IS_ERR(qqi->qqi_acct_obj)) {
 420                 dt_object_put(env, qqi->qqi_acct_obj);
 421                 qqi->qqi_acct_obj = NULL;
 422         }
 423
 424         /* release slv index */
 425         if (qqi->qqi_slv_obj != NULL && !IS_ERR(qqi->qqi_slv_obj)) {
 426                 dt_object_put(env, qqi->qqi_slv_obj);
 427                 qqi->qqi_slv_obj = NULL;
 428                 qqi->qqi_slv_ver = 0;
 429         }
 430
 431         /* release global index */
 432         if (qqi->qqi_glb_obj != NULL && !IS_ERR(qqi->qqi_glb_obj)) {
 433                 dt_object_put(env, qqi->qqi_glb_obj);
 434                 qqi->qqi_glb_obj = NULL;
 435                 qqi->qqi_glb_ver = 0;
 436         }
 437
 438         OBD_FREE_PTR(qqi);
 439         EXIT;
 440 }
 441
 442 static const char *qtype2acct_name(int qtype)
 443 {
 444         static char unknown[24];
 445
 446         switch (qtype) {
 447         case USRQUOTA:
 448                 return "acct_user";
 449         case GRPQUOTA:
 450                 return "acct_group";
 451         case PRJQUOTA:
 452                 return "acct_project";
 453         }
 454
 455         snprintf(unknown, sizeof(unknown), "acct_unknown_%u", qtype);
 456         return unknown;
 457 }
 458
 459 static const char *qtype2glb_name(int qtype)
 460 {
 461         static char unknown[24];
 462
 463         switch (qtype) {
 464         case USRQUOTA:
 465                 return "limit_user";
 466         case GRPQUOTA:
 467                 return "limit_group";
 468         case PRJQUOTA:
 469                 return "limit_project";
 470         }
 471
 472         snprintf(unknown, sizeof(unknown), "acct_unknown_%u", qtype);
 473         return unknown;
 474 }
 475
 476 /*
 477  * Allocate and initialize a qsd_qtype_info structure for quota type \qtype.
 478  * This opens the accounting object and initializes the proc file.
 479  * It's called on OSD start when the qsd_prepare() is invoked on the qsd
 480  * instance.
 481  *
 482  * \param env  - the environment passed by the caller
 483  * \param qsd  - is the qsd instance which will be in charge of the new
 484  *               qsd_qtype_info instance.
 485  * \param qtype - is quota type to set up
 486  *
 487  * \retval - 0 on success and qsd->qsd_type_array[qtype] is allocated,
 488  *           appropriate error on failure
 489  */
 490 static int qsd_qtype_init(const struct lu_env *env, struct qsd_instance *qsd,
 491                           int qtype)
 492 {
 493         struct qsd_qtype_info   *qqi;
 494         int                      rc;
 495         struct obd_uuid          uuid;
 496         ENTRY;
 497
 498         LASSERT(qsd->qsd_type_array[qtype] == NULL);
 499
 500         /* allocate structure for this quota type */
 501         OBD_ALLOC_PTR(qqi);
 502         if (qqi == NULL)
 503                 RETURN(-ENOMEM);
 504         qsd->qsd_type_array[qtype] = qqi;
 505         atomic_set(&qqi->qqi_ref, 1); /* referenced from qsd */
 506
 507         /* set backpointer and other parameters */
 508         qqi->qqi_qsd   = qsd;
 509         qqi->qqi_qtype = qtype;
 510         lu_ref_init(&qqi->qqi_reference);
 511         qqi->qqi_glb_uptodate = false;
 512         qqi->qqi_slv_uptodate = false;
 513         qqi->qqi_reint        = false;
 514         INIT_LIST_HEAD(&qqi->qqi_deferred_glb);
 515         INIT_LIST_HEAD(&qqi->qqi_deferred_slv);
 516         lquota_generate_fid(&qqi->qqi_fid, QSD_RES_TYPE(qsd), qtype);
 517
 518         /* open accounting object */
 519         LASSERT(qqi->qqi_acct_obj == NULL);
 520         qqi->qqi_acct_obj = acct_obj_lookup(env, qsd->qsd_dev, qtype);
 521         if (IS_ERR(qqi->qqi_acct_obj)) {
 522                 CDEBUG(D_QUOTA, "%s: no %s space accounting support: rc = %ld\n",
 523                        qsd->qsd_svname, qtype_name(qtype),
 524                        PTR_ERR(qqi->qqi_acct_obj));
 525                 qqi->qqi_acct_obj = NULL;
 526                 qqi->qqi_acct_failed = true;
 527         }
 528
 529         /* open global index copy */
 530         LASSERT(qqi->qqi_glb_obj == NULL);
 531         qqi->qqi_glb_obj = lquota_disk_glb_find_create(env, qsd->qsd_dev,
 532                                                        qsd->qsd_root,
 533                                                        &qqi->qqi_fid, true);
 534         if (IS_ERR(qqi->qqi_glb_obj)) {
 535                 CERROR("%s: can't open global index copy "DFID" %ld\n",
 536                        qsd->qsd_svname, PFID(&qqi->qqi_fid),
 537                        PTR_ERR(qqi->qqi_glb_obj));
 538                 GOTO(out, rc = PTR_ERR(qqi->qqi_glb_obj));
 539         }
 540         qqi->qqi_glb_ver = dt_version_get(env, qqi->qqi_glb_obj);
 541
 542         /* open slave index copy */
 543         LASSERT(qqi->qqi_slv_obj == NULL);
 544         obd_str2uuid(&uuid, qsd->qsd_svname);
 545         qqi->qqi_slv_obj = lquota_disk_slv_find_create(env, qsd->qsd_dev,
 546                                                        qsd->qsd_root,
 547                                                        &qqi->qqi_fid, &uuid,
 548                                                        true);
 549         if (IS_ERR(qqi->qqi_slv_obj)) {
 550                 CERROR("%s: can't open slave index copy "DFID" %ld\n",
 551                        qsd->qsd_svname, PFID(&qqi->qqi_fid),
 552                        PTR_ERR(qqi->qqi_slv_obj));
 553                 GOTO(out, rc = PTR_ERR(qqi->qqi_slv_obj));
 554         }
 555         qqi->qqi_slv_ver = dt_version_get(env, qqi->qqi_slv_obj);
 556
 557         /* allocate site */
 558         qqi->qqi_site = lquota_site_alloc(env, qqi, false, qtype, &qsd_lqe_ops);
 559         if (IS_ERR(qqi->qqi_site)) {
 560                 CERROR("%s: can't allocate site "DFID" %ld\n", qsd->qsd_svname,
 561                        PFID(&qqi->qqi_fid), PTR_ERR(qqi->qqi_site));
 562                 GOTO(out, rc = PTR_ERR(qqi->qqi_site));
 563         }
 564
 565         /* register proc entry for accounting & global index copy objects */
 566         rc = lprocfs_seq_create(qsd->qsd_proc, qtype2acct_name(qtype),
 567                                 0444, &lprocfs_quota_seq_fops,
 568                                 qqi->qqi_acct_obj);
 569         if (rc) {
 570                 CERROR("%s: can't add procfs entry for accounting file %d\n",
 571                        qsd->qsd_svname, rc);
 572                 GOTO(out, rc);
 573         }
 574
 575         rc = lprocfs_seq_create(qsd->qsd_proc, qtype2glb_name(qtype),
 576                                 0444, &lprocfs_quota_seq_fops,
 577                                 qqi->qqi_glb_obj);
 578         if (rc) {
 579                 CERROR("%s: can't add procfs entry for global index copy %d\n",
 580                        qsd->qsd_svname, rc);
 581                 GOTO(out, rc);
 582         }
 583         EXIT;
 584 out:
 585         if (rc)
 586                 qsd_qtype_fini(env, qsd, qtype);
 587         return rc;
 588 }
 589
 590 /*
 591  * Release a qsd_instance. Companion of qsd_init(). This releases all data
 592  * structures associated with the quota slave (on-disk objects, lquota entry
 593  * tables, ...).
 594  * This function should be called when the OSD is shutting down.
 595  *
 596  * \param env - is the environment passed by the caller
 597  * \param qsd - is the qsd instance to shutdown
 598  */
 599 void qsd_fini(const struct lu_env *env, struct qsd_instance *qsd)
 600 {
 601         int     qtype;
 602         ENTRY;
 603
 604         if (unlikely(qsd == NULL))
 605                 RETURN_EXIT;
 606
 607         CDEBUG(D_QUOTA, "%s: initiating QSD shutdown\n", qsd->qsd_svname);
 608         write_lock(&qsd->qsd_lock);
 609         qsd->qsd_stopping = true;
 610         write_unlock(&qsd->qsd_lock);
 611
 612         /* remove qsd proc entry */
 613         if (qsd->qsd_proc != NULL) {
 614                 lprocfs_remove(&qsd->qsd_proc);
 615                 qsd->qsd_proc = NULL;
 616         }
 617
 618         /* stop the writeback thread */
 619         qsd_stop_upd_thread(qsd);
 620
 621         /* shutdown the reintegration threads */
 622         for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++) {
 623                 if (qsd->qsd_type_array[qtype] == NULL)
 624                         continue;
 625                 qsd_stop_reint_thread(qsd->qsd_type_array[qtype]);
 626         }
 627
 628         if (qsd->qsd_ns != NULL) {
 629                 qsd->qsd_ns = NULL;
 630         }
 631
 632         /* release per-filesystem information */
 633         if (qsd->qsd_fsinfo != NULL) {
 634                 mutex_lock(&qsd->qsd_fsinfo->qfs_mutex);
 635                 /* remove from the list of fsinfo */
 636                 list_del_init(&qsd->qsd_link);
 637                 mutex_unlock(&qsd->qsd_fsinfo->qfs_mutex);
 638                 qsd_put_fsinfo(qsd->qsd_fsinfo);
 639                 qsd->qsd_fsinfo = NULL;
 640         }
 641
 642         /* free per-quota type data */
 643         for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++)
 644                 qsd_qtype_fini(env, qsd, qtype);
 645
 646         /* deregister connection to the quota master */
 647         qsd->qsd_exp_valid = false;
 648         lustre_deregister_lwp_item(&qsd->qsd_exp);
 649
 650         /* release quota root directory */
 651         if (qsd->qsd_root != NULL) {
 652                 dt_object_put(env, qsd->qsd_root);
 653                 qsd->qsd_root = NULL;
 654         }
 655
 656         /* release reference on dt_device */
 657         if (qsd->qsd_dev != NULL) {
 658                 lu_ref_del(&qsd->qsd_dev->dd_lu_dev.ld_reference, "qsd", qsd);
 659                 lu_device_put(&qsd->qsd_dev->dd_lu_dev);
 660                 qsd->qsd_dev = NULL;
 661         }
 662
 663         CDEBUG(D_QUOTA, "%s: QSD shutdown completed\n", qsd->qsd_svname);
 664         OBD_FREE_PTR(qsd);
 665         EXIT;
 666 }
 667 EXPORT_SYMBOL(qsd_fini);
 668
 669 /*
 670  * Create a new qsd_instance to be associated with backend osd device
 671  * identified by \dev.
 672  *
 673  * \param env    - the environment passed by the caller
 674  * \param svname - is the service name of the OSD device creating this instance
 675  * \param dev    - is the dt_device where to store quota index files
 676  * \param osd_proc - is the procfs parent directory where to create procfs file
 677  *                   related to this new qsd instance
 678  *
 679  * \retval - pointer to new qsd_instance associated with dev \dev on success,
 680  *           appropriate error on failure
 681  */
 682 struct qsd_instance *qsd_init(const struct lu_env *env, char *svname,
 683                               struct dt_device *dev,
 684                               struct proc_dir_entry *osd_proc,
 685                               bool is_md, bool excl)
 686 {
 687         struct qsd_thread_info  *qti = qsd_info(env);
 688         struct qsd_instance     *qsd;
 689         int                      rc, type, idx;
 690         ENTRY;
 691
 692         /* only configure qsd for MDT & OST */
 693         type = server_name2index(svname, &idx, NULL);
 694         if (type != LDD_F_SV_TYPE_MDT && type != LDD_F_SV_TYPE_OST)
 695                 RETURN(NULL);
 696
 697         /* allocate qsd instance */
 698         OBD_ALLOC_PTR(qsd);
 699         if (qsd == NULL)
 700                 RETURN(ERR_PTR(-ENOMEM));
 701
 702         /* generic initializations */
 703         rwlock_init(&qsd->qsd_lock);
 704         INIT_LIST_HEAD(&qsd->qsd_link);
 705         INIT_LIST_HEAD(&qsd->qsd_upd_list);
 706         spin_lock_init(&qsd->qsd_adjust_lock);
 707         INIT_LIST_HEAD(&qsd->qsd_adjust_list);
 708         qsd->qsd_prepared = false;
 709         qsd->qsd_started = false;
 710         qsd->qsd_is_md = is_md;
 711         qsd->qsd_updating = false;
 712         qsd->qsd_exclusive = excl;
 713
 714         /* copy service name */
 715         if (strlcpy(qsd->qsd_svname, svname, sizeof(qsd->qsd_svname))
 716             >= sizeof(qsd->qsd_svname))
 717                 GOTO(out, rc = -E2BIG);
 718
 719         /* grab reference on osd device */
 720         lu_device_get(&dev->dd_lu_dev);
 721         lu_ref_add(&dev->dd_lu_dev.ld_reference, "qsd", qsd);
 722         qsd->qsd_dev = dev;
 723
 724         /* get fsname from svname */
 725         rc = server_name2fsname(svname, qti->qti_buf, NULL);
 726         if (rc) {
 727                 CERROR("%s: fail to extract filesystem name\n", svname);
 728                 GOTO(out, rc);
 729         }
 730
 731         /* look up quota setting for the filesystem the target belongs to */
 732         qsd->qsd_fsinfo = qsd_get_fsinfo(qti->qti_buf, 1);
 733         if (qsd->qsd_fsinfo == NULL) {
 734                 CERROR("%s: failed to locate filesystem information\n", svname);
 735                 GOTO(out, rc = -EINVAL);
 736         }
 737
 738         /* add in the list of lquota_fsinfo */
 739         mutex_lock(&qsd->qsd_fsinfo->qfs_mutex);
 740         list_add_tail(&qsd->qsd_link, &qsd->qsd_fsinfo->qfs_qsd_list);
 741         mutex_unlock(&qsd->qsd_fsinfo->qfs_mutex);
 742
 743         /* register procfs directory */
 744         if (qsd->qsd_is_md)
 745                 qsd->qsd_proc = lprocfs_register(QSD_DIR_MD, osd_proc,
 746                                                  lprocfs_quota_qsd_vars, qsd);
 747         else
 748                 qsd->qsd_proc = lprocfs_register(QSD_DIR_DT, osd_proc,
 749                                                  lprocfs_quota_qsd_vars, qsd);
 750
 751         if (type == LDD_F_SV_TYPE_MDT && qsd->qsd_is_md)
 752                 lprocfs_add_symlink(QSD_DIR, osd_proc, "./%s", QSD_DIR_MD);
 753         else if (type == LDD_F_SV_TYPE_OST && !qsd->qsd_is_md)
 754                 lprocfs_add_symlink(QSD_DIR, osd_proc, "./%s", QSD_DIR_DT);
 755
 756         if (IS_ERR(qsd->qsd_proc)) {
 757                 rc = PTR_ERR(qsd->qsd_proc);
 758                 qsd->qsd_proc = NULL;
 759                 CERROR("%s: fail to create quota slave proc entry (%d)\n",
 760                        svname, rc);
 761                 GOTO(out, rc);
 762         }
 763         EXIT;
 764 out:
 765         if (rc) {
 766                 qsd_fini(env, qsd);
 767                 return ERR_PTR(rc);
 768         }
 769         RETURN(qsd);
 770 }
 771 EXPORT_SYMBOL(qsd_init);
 772
 773 /*
 774  * Initialize on-disk structures in order to manage quota enforcement for
 775  * the target associated with the qsd instance \qsd and starts the reintegration
 776  * procedure for each quota type as soon as possible.
 777  * The last step of the reintegration will be completed once qsd_start() is
 778  * called, at which points the space reconciliation with the master will be
 779  * executed.
 780  * This function must be called when the server stack is fully configured,
 781  * typically when ->ldo_prepare is called across the stack.
 782  *
 783  * \param env - the environment passed by the caller
 784  * \param qsd - is qsd_instance to prepare
 785  *
 786  * \retval - 0 on success, appropriate error on failure
 787  */
 788 int qsd_prepare(const struct lu_env *env, struct qsd_instance *qsd)
 789 {
 790         struct qsd_thread_info  *qti = qsd_info(env);
 791         int                      qtype, rc = 0;
 792         ENTRY;
 793
 794         if (unlikely(qsd == NULL))
 795                 RETURN(0);
 796
 797         read_lock(&qsd->qsd_lock);
 798         if (qsd->qsd_prepared) {
 799                 CERROR("%s: qsd instance already prepared\n", qsd->qsd_svname);
 800                 rc = -EALREADY;
 801         }
 802         read_unlock(&qsd->qsd_lock);
 803         if (rc)
 804                 RETURN(rc);
 805
 806         /* Record whether this qsd instance is managing quota enforcement for a
 807          * MDT (i.e. inode quota) or OST (block quota) */
 808         if (qsd->qsd_is_md)
 809                 qsd->qsd_sync_threshold = LQUOTA_LEAST_QUNIT(LQUOTA_RES_MD);
 810         else
 811                 qsd->qsd_sync_threshold = LQUOTA_LEAST_QUNIT(LQUOTA_RES_DT);
 812
 813         /* look-up on-disk directory for the quota slave */
 814         qsd->qsd_root = lquota_disk_dir_find_create(env, qsd->qsd_dev, NULL,
 815                                                     QSD_DIR);
 816         if (IS_ERR(qsd->qsd_root)) {
 817                 rc = PTR_ERR(qsd->qsd_root);
 818                 qsd->qsd_root = NULL;
 819                 CERROR("%s: failed to create quota slave root dir (%d)\n",
 820                        qsd->qsd_svname, rc);
 821                 RETURN(rc);
 822         }
 823
 824         /* initialize per-quota type data */
 825         for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++) {
 826                 rc = qsd_qtype_init(env, qsd, qtype);
 827                 if (rc)
 828                         RETURN(rc);
 829         }
 830
 831         /* pools successfully setup, mark the qsd as prepared */
 832         write_lock(&qsd->qsd_lock);
 833         qsd->qsd_prepared = true;
 834         write_unlock(&qsd->qsd_lock);
 835
 836         if (qsd->qsd_dev->dd_rdonly)
 837                 RETURN(0);
 838
 839         /* start reintegration thread for each type, if required */
 840         for (qtype = USRQUOTA; qtype < LL_MAXQUOTAS; qtype++) {
 841                 struct qsd_qtype_info   *qqi = qsd->qsd_type_array[qtype];
 842
 843                 if (qsd_type_enabled(qsd, qtype) &&
 844                     qqi->qqi_acct_failed) {
 845                         LCONSOLE_ERROR("%s: can't enable quota enforcement "
 846                                        "since space accounting isn't functional"
 847                                        ". Please run tunefs.lustre --quota on "
 848                                        "an unmounted filesystem if not done "
 849                                        "already\n", qsd->qsd_svname);
 850                         continue;
 851                 }
 852
 853                 rc = qsd_start_reint_thread(qqi);
 854                 if (rc) {
 855                         CERROR("%s: failed to start reint thread for type %s: rc = %d\n",
 856                                 qsd->qsd_svname, qtype_name(qtype), rc);
 857                         RETURN(rc);
 858                 }
 859         }
 860
 861         /* start writeback thread */
 862         rc = qsd_start_upd_thread(qsd);
 863         if (rc) {
 864                 CERROR("%s: failed to start writeback thread (%d)\n",
 865                        qsd->qsd_svname, rc);
 866                 RETURN(rc);
 867         }
 868
 869         /* generate osp name */
 870         rc = tgt_name2lwp_name(qsd->qsd_svname, qti->qti_buf,
 871                                MTI_NAME_MAXLEN, 0);
 872         if (rc) {
 873                 CERROR("%s: failed to generate ospname (%d)\n",
 874                        qsd->qsd_svname, rc);
 875                 RETURN(rc);
 876         }
 877
 878         /* the connection callback will start the reintegration
 879          * procedure if quota is enabled */
 880         rc = lustre_register_lwp_item(qti->qti_buf, &qsd->qsd_exp,
 881                                       qsd_conn_callback, (void *)qsd);
 882         if (rc) {
 883                 CERROR("%s: fail to get connection to master (%d)\n",
 884                        qsd->qsd_svname, rc);
 885                 RETURN(rc);
 886         }
 887
 888         RETURN(0);
 889 }
 890 EXPORT_SYMBOL(qsd_prepare);
 891
 892 /*
 893  * Start a qsd instance. This will complete the last step of the reintegration
 894  * procedure as soon as possible (provided that the master is reachable).
 895  * This should be called when recovery has been completed and quota should now
 896  * be enforced on every operations.
 897  *
 898  * \param env - the environment passed by the caller
 899  * \param qsd - is the qsd instance associated with the osd device to start
 900  */
 901 int qsd_start(const struct lu_env *env, struct qsd_instance *qsd)
 902 {
 903         int     type, rc = 0;
 904         ENTRY;
 905
 906         if (unlikely(qsd == NULL))
 907                 RETURN(0);
 908
 909         write_lock(&qsd->qsd_lock);
 910         if (!qsd->qsd_prepared) {
 911                 CERROR("%s: can't start qsd instance since it wasn't properly "
 912                        "initialized\n", qsd->qsd_svname);
 913                 rc = -EFAULT;
 914         } else if (qsd->qsd_started) {
 915                 CERROR("%s: qsd instance already started\n", qsd->qsd_svname);
 916                 rc = -EALREADY;
 917         } else {
 918                 /* notify that the qsd_instance is now started */
 919                 qsd->qsd_started = true;
 920         }
 921         write_unlock(&qsd->qsd_lock);
 922
 923         if (rc)
 924                 RETURN(rc);
 925
 926         /* Trigger the 3rd step of reintegration: If usage > granted, acquire
 927          * up to usage; If usage < granted, release down to usage.  */
 928         for (type = USRQUOTA; type < LL_MAXQUOTAS; type++) {
 929                 struct qsd_qtype_info   *qqi = qsd->qsd_type_array[type];
 930                 struct task_struct *t;
 931
 932                 /* qqi_reint_task can be set to NULL at any time,
 933                  * so we need to be careful.
 934                  */
 935                 rcu_read_lock();
 936                 t = rcu_dereference(qqi->qqi_reint_task);
 937                 if (t)
 938                         wake_up_process(t);
 939                 rcu_read_unlock();
 940         }
 941
 942         RETURN(rc);
 943 }
 944 EXPORT_SYMBOL(qsd_start);
 945
 946 void lustre_register_quota_process_config(int (*qpc)(struct lustre_cfg *lcfg));
 947
 948 /*
 949  * Global initialization performed at module load time
 950  */
 951 int qsd_glb_init(void)
 952 {
 953         int     rc;
 954
 955         rc = lu_kmem_init(qsd_caches);
 956         if (rc)
 957                 return rc;
 958
 959         qsd_key_init_generic(&qsd_thread_key, NULL);
 960         lu_context_key_register(&qsd_thread_key);
 961         lustre_register_quota_process_config(qsd_process_config);
 962
 963         return 0;
 964 }
 965
 966 /*
 967  * Companion of qsd_glb_init() called at module unload time
 968  */
 969 void qsd_glb_fini(void)
 970 {
 971         lustre_register_quota_process_config(NULL);
 972         lu_kmem_fini(qsd_caches);
 973         lu_context_key_degister(&qsd_thread_key);
 974 }