fs/ceph/mds_client.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 #include <linux/ceph/ceph_debug.h>
   3
   4 #include <linux/fs.h>
   5 #include <linux/wait.h>
   6 #include <linux/slab.h>
   7 #include <linux/gfp.h>
   8 #include <linux/sched.h>
   9 #include <linux/debugfs.h>
  10 #include <linux/seq_file.h>
  11 #include <linux/ratelimit.h>
  12 #include <linux/bits.h>
  13
  14 #include "super.h"
  15 #include "mds_client.h"
  16
  17 #include <linux/ceph/ceph_features.h>
  18 #include <linux/ceph/messenger.h>
  19 #include <linux/ceph/decode.h>
  20 #include <linux/ceph/pagelist.h>
  21 #include <linux/ceph/auth.h>
  22 #include <linux/ceph/debugfs.h>
  23
  24 #define RECONNECT_MAX_SIZE (INT_MAX - PAGE_SIZE)
  25
  26 /*
  27  * A cluster of MDS (metadata server) daemons is responsible for
  28  * managing the file system namespace (the directory hierarchy and
  29  * inodes) and for coordinating shared access to storage.  Metadata is
  30  * partitioning hierarchically across a number of servers, and that
  31  * partition varies over time as the cluster adjusts the distribution
  32  * in order to balance load.
  33  *
  34  * The MDS client is primarily responsible to managing synchronous
  35  * metadata requests for operations like open, unlink, and so forth.
  36  * If there is a MDS failure, we find out about it when we (possibly
  37  * request and) receive a new MDS map, and can resubmit affected
  38  * requests.
  39  *
  40  * For the most part, though, we take advantage of a lossless
  41  * communications channel to the MDS, and do not need to worry about
  42  * timing out or resubmitting requests.
  43  *
  44  * We maintain a stateful "session" with each MDS we interact with.
  45  * Within each session, we sent periodic heartbeat messages to ensure
  46  * any capabilities or leases we have been issues remain valid.  If
  47  * the session times out and goes stale, our leases and capabilities
  48  * are no longer valid.
  49  */
  50
  51 struct ceph_reconnect_state {
  52         struct ceph_mds_session *session;
  53         int nr_caps, nr_realms;
  54         struct ceph_pagelist *pagelist;
  55         unsigned msg_version;
  56         bool allow_multi;
  57 };
  58
  59 static void __wake_requests(struct ceph_mds_client *mdsc,
  60                             struct list_head *head);
  61 static void ceph_cap_release_work(struct work_struct *work);
  62 static void ceph_cap_reclaim_work(struct work_struct *work);
  63
  64 static const struct ceph_connection_operations mds_con_ops;
  65
  66
  67 /*
  68  * mds reply parsing
  69  */
  70
  71 static int parse_reply_info_quota(void **p, void *end,
  72                                   struct ceph_mds_reply_info_in *info)
  73 {
  74         u8 struct_v, struct_compat;
  75         u32 struct_len;
  76
  77         ceph_decode_8_safe(p, end, struct_v, bad);
  78         ceph_decode_8_safe(p, end, struct_compat, bad);
  79         /* struct_v is expected to be >= 1. we only
  80          * understand encoding with struct_compat == 1. */
  81         if (!struct_v || struct_compat != 1)
  82                 goto bad;
  83         ceph_decode_32_safe(p, end, struct_len, bad);
  84         ceph_decode_need(p, end, struct_len, bad);
  85         end = *p + struct_len;
  86         ceph_decode_64_safe(p, end, info->max_bytes, bad);
  87         ceph_decode_64_safe(p, end, info->max_files, bad);
  88         *p = end;
  89         return 0;
  90 bad:
  91         return -EIO;
  92 }
  93
  94 /*
  95  * parse individual inode info
  96  */
  97 static int parse_reply_info_in(void **p, void *end,
  98                                struct ceph_mds_reply_info_in *info,
  99                                u64 features)
 100 {
 101         int err = 0;
 102         u8 struct_v = 0;
 103
 104         if (features == (u64)-1) {
 105                 u32 struct_len;
 106                 u8 struct_compat;
 107                 ceph_decode_8_safe(p, end, struct_v, bad);
 108                 ceph_decode_8_safe(p, end, struct_compat, bad);
 109                 /* struct_v is expected to be >= 1. we only understand
 110                  * encoding with struct_compat == 1. */
 111                 if (!struct_v || struct_compat != 1)
 112                         goto bad;
 113                 ceph_decode_32_safe(p, end, struct_len, bad);
 114                 ceph_decode_need(p, end, struct_len, bad);
 115                 end = *p + struct_len;
 116         }
 117
 118         ceph_decode_need(p, end, sizeof(struct ceph_mds_reply_inode), bad);
 119         info->in = *p;
 120         *p += sizeof(struct ceph_mds_reply_inode) +
 121                 sizeof(*info->in->fragtree.splits) *
 122                 le32_to_cpu(info->in->fragtree.nsplits);
 123
 124         ceph_decode_32_safe(p, end, info->symlink_len, bad);
 125         ceph_decode_need(p, end, info->symlink_len, bad);
 126         info->symlink = *p;
 127         *p += info->symlink_len;
 128
 129         ceph_decode_copy_safe(p, end, &info->dir_layout,
 130                               sizeof(info->dir_layout), bad);
 131         ceph_decode_32_safe(p, end, info->xattr_len, bad);
 132         ceph_decode_need(p, end, info->xattr_len, bad);
 133         info->xattr_data = *p;
 134         *p += info->xattr_len;
 135
 136         if (features == (u64)-1) {
 137                 /* inline data */
 138                 ceph_decode_64_safe(p, end, info->inline_version, bad);
 139                 ceph_decode_32_safe(p, end, info->inline_len, bad);
 140                 ceph_decode_need(p, end, info->inline_len, bad);
 141                 info->inline_data = *p;
 142                 *p += info->inline_len;
 143                 /* quota */
 144                 err = parse_reply_info_quota(p, end, info);
 145                 if (err < 0)
 146                         goto out_bad;
 147                 /* pool namespace */
 148                 ceph_decode_32_safe(p, end, info->pool_ns_len, bad);
 149                 if (info->pool_ns_len > 0) {
 150                         ceph_decode_need(p, end, info->pool_ns_len, bad);
 151                         info->pool_ns_data = *p;
 152                         *p += info->pool_ns_len;
 153                 }
 154
 155                 /* btime */
 156                 ceph_decode_need(p, end, sizeof(info->btime), bad);
 157                 ceph_decode_copy(p, &info->btime, sizeof(info->btime));
 158
 159                 /* change attribute */
 160                 ceph_decode_64_safe(p, end, info->change_attr, bad);
 161
 162                 /* dir pin */
 163                 if (struct_v >= 2) {
 164                         ceph_decode_32_safe(p, end, info->dir_pin, bad);
 165                 } else {
 166                         info->dir_pin = -ENODATA;
 167                 }
 168
 169                 /* snapshot birth time, remains zero for v<=2 */
 170                 if (struct_v >= 3) {
 171                         ceph_decode_need(p, end, sizeof(info->snap_btime), bad);
 172                         ceph_decode_copy(p, &info->snap_btime,
 173                                          sizeof(info->snap_btime));
 174                 } else {
 175                         memset(&info->snap_btime, 0, sizeof(info->snap_btime));
 176                 }
 177
 178                 *p = end;
 179         } else {
 180                 if (features & CEPH_FEATURE_MDS_INLINE_DATA) {
 181                         ceph_decode_64_safe(p, end, info->inline_version, bad);
 182                         ceph_decode_32_safe(p, end, info->inline_len, bad);
 183                         ceph_decode_need(p, end, info->inline_len, bad);
 184                         info->inline_data = *p;
 185                         *p += info->inline_len;
 186                 } else
 187                         info->inline_version = CEPH_INLINE_NONE;
 188
 189                 if (features & CEPH_FEATURE_MDS_QUOTA) {
 190                         err = parse_reply_info_quota(p, end, info);
 191                         if (err < 0)
 192                                 goto out_bad;
 193                 } else {
 194                         info->max_bytes = 0;
 195                         info->max_files = 0;
 196                 }
 197
 198                 info->pool_ns_len = 0;
 199                 info->pool_ns_data = NULL;
 200                 if (features & CEPH_FEATURE_FS_FILE_LAYOUT_V2) {
 201                         ceph_decode_32_safe(p, end, info->pool_ns_len, bad);
 202                         if (info->pool_ns_len > 0) {
 203                                 ceph_decode_need(p, end, info->pool_ns_len, bad);
 204                                 info->pool_ns_data = *p;
 205                                 *p += info->pool_ns_len;
 206                         }
 207                 }
 208
 209                 if (features & CEPH_FEATURE_FS_BTIME) {
 210                         ceph_decode_need(p, end, sizeof(info->btime), bad);
 211                         ceph_decode_copy(p, &info->btime, sizeof(info->btime));
 212                         ceph_decode_64_safe(p, end, info->change_attr, bad);
 213                 }
 214
 215                 info->dir_pin = -ENODATA;
 216                 /* info->snap_btime remains zero */
 217         }
 218         return 0;
 219 bad:
 220         err = -EIO;
 221 out_bad:
 222         return err;
 223 }
 224
 225 static int parse_reply_info_dir(void **p, void *end,
 226                                 struct ceph_mds_reply_dirfrag **dirfrag,
 227                                 u64 features)
 228 {
 229         if (features == (u64)-1) {
 230                 u8 struct_v, struct_compat;
 231                 u32 struct_len;
 232                 ceph_decode_8_safe(p, end, struct_v, bad);
 233                 ceph_decode_8_safe(p, end, struct_compat, bad);
 234                 /* struct_v is expected to be >= 1. we only understand
 235                  * encoding whose struct_compat == 1. */
 236                 if (!struct_v || struct_compat != 1)
 237                         goto bad;
 238                 ceph_decode_32_safe(p, end, struct_len, bad);
 239                 ceph_decode_need(p, end, struct_len, bad);
 240                 end = *p + struct_len;
 241         }
 242
 243         ceph_decode_need(p, end, sizeof(**dirfrag), bad);
 244         *dirfrag = *p;
 245         *p += sizeof(**dirfrag) + sizeof(u32) * le32_to_cpu((*dirfrag)->ndist);
 246         if (unlikely(*p > end))
 247                 goto bad;
 248         if (features == (u64)-1)
 249                 *p = end;
 250         return 0;
 251 bad:
 252         return -EIO;
 253 }
 254
 255 static int parse_reply_info_lease(void **p, void *end,
 256                                   struct ceph_mds_reply_lease **lease,
 257                                   u64 features)
 258 {
 259         if (features == (u64)-1) {
 260                 u8 struct_v, struct_compat;
 261                 u32 struct_len;
 262                 ceph_decode_8_safe(p, end, struct_v, bad);
 263                 ceph_decode_8_safe(p, end, struct_compat, bad);
 264                 /* struct_v is expected to be >= 1. we only understand
 265                  * encoding whose struct_compat == 1. */
 266                 if (!struct_v || struct_compat != 1)
 267                         goto bad;
 268                 ceph_decode_32_safe(p, end, struct_len, bad);
 269                 ceph_decode_need(p, end, struct_len, bad);
 270                 end = *p + struct_len;
 271         }
 272
 273         ceph_decode_need(p, end, sizeof(**lease), bad);
 274         *lease = *p;
 275         *p += sizeof(**lease);
 276         if (features == (u64)-1)
 277                 *p = end;
 278         return 0;
 279 bad:
 280         return -EIO;
 281 }
 282
 283 /*
 284  * parse a normal reply, which may contain a (dir+)dentry and/or a
 285  * target inode.
 286  */
 287 static int parse_reply_info_trace(void **p, void *end,
 288                                   struct ceph_mds_reply_info_parsed *info,
 289                                   u64 features)
 290 {
 291         int err;
 292
 293         if (info->head->is_dentry) {
 294                 err = parse_reply_info_in(p, end, &info->diri, features);
 295                 if (err < 0)
 296                         goto out_bad;
 297
 298                 err = parse_reply_info_dir(p, end, &info->dirfrag, features);
 299                 if (err < 0)
 300                         goto out_bad;
 301
 302                 ceph_decode_32_safe(p, end, info->dname_len, bad);
 303                 ceph_decode_need(p, end, info->dname_len, bad);
 304                 info->dname = *p;
 305                 *p += info->dname_len;
 306
 307                 err = parse_reply_info_lease(p, end, &info->dlease, features);
 308                 if (err < 0)
 309                         goto out_bad;
 310         }
 311
 312         if (info->head->is_target) {
 313                 err = parse_reply_info_in(p, end, &info->targeti, features);
 314                 if (err < 0)
 315                         goto out_bad;
 316         }
 317
 318         if (unlikely(*p != end))
 319                 goto bad;
 320         return 0;
 321
 322 bad:
 323         err = -EIO;
 324 out_bad:
 325         pr_err("problem parsing mds trace %d\n", err);
 326         return err;
 327 }
 328
 329 /*
 330  * parse readdir results
 331  */
 332 static int parse_reply_info_readdir(void **p, void *end,
 333                                 struct ceph_mds_reply_info_parsed *info,
 334                                 u64 features)
 335 {
 336         u32 num, i = 0;
 337         int err;
 338
 339         err = parse_reply_info_dir(p, end, &info->dir_dir, features);
 340         if (err < 0)
 341                 goto out_bad;
 342
 343         ceph_decode_need(p, end, sizeof(num) + 2, bad);
 344         num = ceph_decode_32(p);
 345         {
 346                 u16 flags = ceph_decode_16(p);
 347                 info->dir_end = !!(flags & CEPH_READDIR_FRAG_END);
 348                 info->dir_complete = !!(flags & CEPH_READDIR_FRAG_COMPLETE);
 349                 info->hash_order = !!(flags & CEPH_READDIR_HASH_ORDER);
 350                 info->offset_hash = !!(flags & CEPH_READDIR_OFFSET_HASH);
 351         }
 352         if (num == 0)
 353                 goto done;
 354
 355         BUG_ON(!info->dir_entries);
 356         if ((unsigned long)(info->dir_entries + num) >
 357             (unsigned long)info->dir_entries + info->dir_buf_size) {
 358                 pr_err("dir contents are larger than expected\n");
 359                 WARN_ON(1);
 360                 goto bad;
 361         }
 362
 363         info->dir_nr = num;
 364         while (num) {
 365                 struct ceph_mds_reply_dir_entry *rde = info->dir_entries + i;
 366                 /* dentry */
 367                 ceph_decode_32_safe(p, end, rde->name_len, bad);
 368                 ceph_decode_need(p, end, rde->name_len, bad);
 369                 rde->name = *p;
 370                 *p += rde->name_len;
 371                 dout("parsed dir dname '%.*s'\n", rde->name_len, rde->name);
 372
 373                 /* dentry lease */
 374                 err = parse_reply_info_lease(p, end, &rde->lease, features);
 375                 if (err)
 376                         goto out_bad;
 377                 /* inode */
 378                 err = parse_reply_info_in(p, end, &rde->inode, features);
 379                 if (err < 0)
 380                         goto out_bad;
 381                 /* ceph_readdir_prepopulate() will update it */
 382                 rde->offset = 0;
 383                 i++;
 384                 num--;
 385         }
 386
 387 done:
 388         /* Skip over any unrecognized fields */
 389         *p = end;
 390         return 0;
 391
 392 bad:
 393         err = -EIO;
 394 out_bad:
 395         pr_err("problem parsing dir contents %d\n", err);
 396         return err;
 397 }
 398
 399 /*
 400  * parse fcntl F_GETLK results
 401  */
 402 static int parse_reply_info_filelock(void **p, void *end,
 403                                      struct ceph_mds_reply_info_parsed *info,
 404                                      u64 features)
 405 {
 406         if (*p + sizeof(*info->filelock_reply) > end)
 407                 goto bad;
 408
 409         info->filelock_reply = *p;
 410
 411         /* Skip over any unrecognized fields */
 412         *p = end;
 413         return 0;
 414 bad:
 415         return -EIO;
 416 }
 417
 418 /*
 419  * parse create results
 420  */
 421 static int parse_reply_info_create(void **p, void *end,
 422                                   struct ceph_mds_reply_info_parsed *info,
 423                                   u64 features)
 424 {
 425         if (features == (u64)-1 ||
 426             (features & CEPH_FEATURE_REPLY_CREATE_INODE)) {
 427                 /* Malformed reply? */
 428                 if (*p == end) {
 429                         info->has_create_ino = false;
 430                 } else {
 431                         info->has_create_ino = true;
 432                         ceph_decode_64_safe(p, end, info->ino, bad);
 433                 }
 434         } else {
 435                 if (*p != end)
 436                         goto bad;
 437         }
 438
 439         /* Skip over any unrecognized fields */
 440         *p = end;
 441         return 0;
 442 bad:
 443         return -EIO;
 444 }
 445
 446 /*
 447  * parse extra results
 448  */
 449 static int parse_reply_info_extra(void **p, void *end,
 450                                   struct ceph_mds_reply_info_parsed *info,
 451                                   u64 features)
 452 {
 453         u32 op = le32_to_cpu(info->head->op);
 454
 455         if (op == CEPH_MDS_OP_GETFILELOCK)
 456                 return parse_reply_info_filelock(p, end, info, features);
 457         else if (op == CEPH_MDS_OP_READDIR || op == CEPH_MDS_OP_LSSNAP)
 458                 return parse_reply_info_readdir(p, end, info, features);
 459         else if (op == CEPH_MDS_OP_CREATE)
 460                 return parse_reply_info_create(p, end, info, features);
 461         else
 462                 return -EIO;
 463 }
 464
 465 /*
 466  * parse entire mds reply
 467  */
 468 static int parse_reply_info(struct ceph_msg *msg,
 469                             struct ceph_mds_reply_info_parsed *info,
 470                             u64 features)
 471 {
 472         void *p, *end;
 473         u32 len;
 474         int err;
 475
 476         info->head = msg->front.iov_base;
 477         p = msg->front.iov_base + sizeof(struct ceph_mds_reply_head);
 478         end = p + msg->front.iov_len - sizeof(struct ceph_mds_reply_head);
 479
 480         /* trace */
 481         ceph_decode_32_safe(&p, end, len, bad);
 482         if (len > 0) {
 483                 ceph_decode_need(&p, end, len, bad);
 484                 err = parse_reply_info_trace(&p, p+len, info, features);
 485                 if (err < 0)
 486                         goto out_bad;
 487         }
 488
 489         /* extra */
 490         ceph_decode_32_safe(&p, end, len, bad);
 491         if (len > 0) {
 492                 ceph_decode_need(&p, end, len, bad);
 493                 err = parse_reply_info_extra(&p, p+len, info, features);
 494                 if (err < 0)
 495                         goto out_bad;
 496         }
 497
 498         /* snap blob */
 499         ceph_decode_32_safe(&p, end, len, bad);
 500         info->snapblob_len = len;
 501         info->snapblob = p;
 502         p += len;
 503
 504         if (p != end)
 505                 goto bad;
 506         return 0;
 507
 508 bad:
 509         err = -EIO;
 510 out_bad:
 511         pr_err("mds parse_reply err %d\n", err);
 512         return err;
 513 }
 514
 515 static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info)
 516 {
 517         if (!info->dir_entries)
 518                 return;
 519         free_pages((unsigned long)info->dir_entries, get_order(info->dir_buf_size));
 520 }
 521
 522
 523 /*
 524  * sessions
 525  */
 526 const char *ceph_session_state_name(int s)
 527 {
 528         switch (s) {
 529         case CEPH_MDS_SESSION_NEW: return "new";
 530         case CEPH_MDS_SESSION_OPENING: return "opening";
 531         case CEPH_MDS_SESSION_OPEN: return "open";
 532         case CEPH_MDS_SESSION_HUNG: return "hung";
 533         case CEPH_MDS_SESSION_CLOSING: return "closing";
 534         case CEPH_MDS_SESSION_CLOSED: return "closed";
 535         case CEPH_MDS_SESSION_RESTARTING: return "restarting";
 536         case CEPH_MDS_SESSION_RECONNECTING: return "reconnecting";
 537         case CEPH_MDS_SESSION_REJECTED: return "rejected";
 538         default: return "???";
 539         }
 540 }
 541
 542 struct ceph_mds_session *ceph_get_mds_session(struct ceph_mds_session *s)
 543 {
 544         if (refcount_inc_not_zero(&s->s_ref)) {
 545                 dout("mdsc get_session %p %d -> %d\n", s,
 546                      refcount_read(&s->s_ref)-1, refcount_read(&s->s_ref));
 547                 return s;
 548         } else {
 549                 dout("mdsc get_session %p 0 -- FAIL\n", s);
 550                 return NULL;
 551         }
 552 }
 553
 554 void ceph_put_mds_session(struct ceph_mds_session *s)
 555 {
 556         dout("mdsc put_session %p %d -> %d\n", s,
 557              refcount_read(&s->s_ref), refcount_read(&s->s_ref)-1);
 558         if (refcount_dec_and_test(&s->s_ref)) {
 559                 if (s->s_auth.authorizer)
 560                         ceph_auth_destroy_authorizer(s->s_auth.authorizer);
 561                 kfree(s);
 562         }
 563 }
 564
 565 /*
 566  * called under mdsc->mutex
 567  */
 568 struct ceph_mds_session *__ceph_lookup_mds_session(struct ceph_mds_client *mdsc,
 569                                                    int mds)
 570 {
 571         if (mds >= mdsc->max_sessions || !mdsc->sessions[mds])
 572                 return NULL;
 573         return ceph_get_mds_session(mdsc->sessions[mds]);
 574 }
 575
 576 static bool __have_session(struct ceph_mds_client *mdsc, int mds)
 577 {
 578         if (mds >= mdsc->max_sessions || !mdsc->sessions[mds])
 579                 return false;
 580         else
 581                 return true;
 582 }
 583
 584 static int __verify_registered_session(struct ceph_mds_client *mdsc,
 585                                        struct ceph_mds_session *s)
 586 {
 587         if (s->s_mds >= mdsc->max_sessions ||
 588             mdsc->sessions[s->s_mds] != s)
 589                 return -ENOENT;
 590         return 0;
 591 }
 592
 593 /*
 594  * create+register a new session for given mds.
 595  * called under mdsc->mutex.
 596  */
 597 static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
 598                                                  int mds)
 599 {
 600         struct ceph_mds_session *s;
 601
 602         if (mds >= mdsc->mdsmap->possible_max_rank)
 603                 return ERR_PTR(-EINVAL);
 604
 605         s = kzalloc(sizeof(*s), GFP_NOFS);
 606         if (!s)
 607                 return ERR_PTR(-ENOMEM);
 608
 609         if (mds >= mdsc->max_sessions) {
 610                 int newmax = 1 << get_count_order(mds + 1);
 611                 struct ceph_mds_session **sa;
 612
 613                 dout("%s: realloc to %d\n", __func__, newmax);
 614                 sa = kcalloc(newmax, sizeof(void *), GFP_NOFS);
 615                 if (!sa)
 616                         goto fail_realloc;
 617                 if (mdsc->sessions) {
 618                         memcpy(sa, mdsc->sessions,
 619                                mdsc->max_sessions * sizeof(void *));
 620                         kfree(mdsc->sessions);
 621                 }
 622                 mdsc->sessions = sa;
 623                 mdsc->max_sessions = newmax;
 624         }
 625
 626         dout("%s: mds%d\n", __func__, mds);
 627         s->s_mdsc = mdsc;
 628         s->s_mds = mds;
 629         s->s_state = CEPH_MDS_SESSION_NEW;
 630         s->s_ttl = 0;
 631         s->s_seq = 0;
 632         mutex_init(&s->s_mutex);
 633
 634         ceph_con_init(&s->s_con, s, &mds_con_ops, &mdsc->fsc->client->msgr);
 635
 636         spin_lock_init(&s->s_gen_ttl_lock);
 637         s->s_cap_gen = 1;
 638         s->s_cap_ttl = jiffies - 1;
 639
 640         spin_lock_init(&s->s_cap_lock);
 641         s->s_renew_requested = 0;
 642         s->s_renew_seq = 0;
 643         INIT_LIST_HEAD(&s->s_caps);
 644         s->s_nr_caps = 0;
 645         refcount_set(&s->s_ref, 1);
 646         INIT_LIST_HEAD(&s->s_waiting);
 647         INIT_LIST_HEAD(&s->s_unsafe);
 648         s->s_num_cap_releases = 0;
 649         s->s_cap_reconnect = 0;
 650         s->s_cap_iterator = NULL;
 651         INIT_LIST_HEAD(&s->s_cap_releases);
 652         INIT_WORK(&s->s_cap_release_work, ceph_cap_release_work);
 653
 654         INIT_LIST_HEAD(&s->s_cap_flushing);
 655
 656         mdsc->sessions[mds] = s;
 657         atomic_inc(&mdsc->num_sessions);
 658         refcount_inc(&s->s_ref);  /* one ref to sessions[], one to caller */
 659
 660         ceph_con_open(&s->s_con, CEPH_ENTITY_TYPE_MDS, mds,
 661                       ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
 662
 663         return s;
 664
 665 fail_realloc:
 666         kfree(s);
 667         return ERR_PTR(-ENOMEM);
 668 }
 669
 670 /*
 671  * called under mdsc->mutex
 672  */
 673 static void __unregister_session(struct ceph_mds_client *mdsc,
 674                                struct ceph_mds_session *s)
 675 {
 676         dout("__unregister_session mds%d %p\n", s->s_mds, s);
 677         BUG_ON(mdsc->sessions[s->s_mds] != s);
 678         mdsc->sessions[s->s_mds] = NULL;
 679         ceph_con_close(&s->s_con);
 680         ceph_put_mds_session(s);
 681         atomic_dec(&mdsc->num_sessions);
 682 }
 683
 684 /*
 685  * drop session refs in request.
 686  *
 687  * should be last request ref, or hold mdsc->mutex
 688  */
 689 static void put_request_session(struct ceph_mds_request *req)
 690 {
 691         if (req->r_session) {
 692                 ceph_put_mds_session(req->r_session);
 693                 req->r_session = NULL;
 694         }
 695 }
 696
 697 void ceph_mdsc_release_request(struct kref *kref)
 698 {
 699         struct ceph_mds_request *req = container_of(kref,
 700                                                     struct ceph_mds_request,
 701                                                     r_kref);
 702         destroy_reply_info(&req->r_reply_info);
 703         if (req->r_request)
 704                 ceph_msg_put(req->r_request);
 705         if (req->r_reply)
 706                 ceph_msg_put(req->r_reply);
 707         if (req->r_inode) {
 708                 ceph_put_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
 709                 /* avoid calling iput_final() in mds dispatch threads */
 710                 ceph_async_iput(req->r_inode);
 711         }
 712         if (req->r_parent) {
 713                 ceph_put_cap_refs(ceph_inode(req->r_parent), CEPH_CAP_PIN);
 714                 ceph_async_iput(req->r_parent);
 715         }
 716         ceph_async_iput(req->r_target_inode);
 717         if (req->r_dentry)
 718                 dput(req->r_dentry);
 719         if (req->r_old_dentry)
 720                 dput(req->r_old_dentry);
 721         if (req->r_old_dentry_dir) {
 722                 /*
 723                  * track (and drop pins for) r_old_dentry_dir
 724                  * separately, since r_old_dentry's d_parent may have
 725                  * changed between the dir mutex being dropped and
 726                  * this request being freed.
 727                  */
 728                 ceph_put_cap_refs(ceph_inode(req->r_old_dentry_dir),
 729                                   CEPH_CAP_PIN);
 730                 ceph_async_iput(req->r_old_dentry_dir);
 731         }
 732         kfree(req->r_path1);
 733         kfree(req->r_path2);
 734         if (req->r_pagelist)
 735                 ceph_pagelist_release(req->r_pagelist);
 736         put_request_session(req);
 737         ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation);
 738         WARN_ON_ONCE(!list_empty(&req->r_wait));
 739         kfree(req);
 740 }
 741
 742 DEFINE_RB_FUNCS(request, struct ceph_mds_request, r_tid, r_node)
 743
 744 /*
 745  * lookup session, bump ref if found.
 746  *
 747  * called under mdsc->mutex.
 748  */
 749 static struct ceph_mds_request *
 750 lookup_get_request(struct ceph_mds_client *mdsc, u64 tid)
 751 {
 752         struct ceph_mds_request *req;
 753
 754         req = lookup_request(&mdsc->request_tree, tid);
 755         if (req)
 756                 ceph_mdsc_get_request(req);
 757
 758         return req;
 759 }
 760
 761 /*
 762  * Register an in-flight request, and assign a tid.  Link to directory
 763  * are modifying (if any).
 764  *
 765  * Called under mdsc->mutex.
 766  */
 767 static void __register_request(struct ceph_mds_client *mdsc,
 768                                struct ceph_mds_request *req,
 769                                struct inode *dir)
 770 {
 771         int ret = 0;
 772
 773         req->r_tid = ++mdsc->last_tid;
 774         if (req->r_num_caps) {
 775                 ret = ceph_reserve_caps(mdsc, &req->r_caps_reservation,
 776                                         req->r_num_caps);
 777                 if (ret < 0) {
 778                         pr_err("__register_request %p "
 779                                "failed to reserve caps: %d\n", req, ret);
 780                         /* set req->r_err to fail early from __do_request */
 781                         req->r_err = ret;
 782                         return;
 783                 }
 784         }
 785         dout("__register_request %p tid %lld\n", req, req->r_tid);
 786         ceph_mdsc_get_request(req);
 787         insert_request(&mdsc->request_tree, req);
 788
 789         req->r_uid = current_fsuid();
 790         req->r_gid = current_fsgid();
 791
 792         if (mdsc->oldest_tid == 0 && req->r_op != CEPH_MDS_OP_SETFILELOCK)
 793                 mdsc->oldest_tid = req->r_tid;
 794
 795         if (dir) {
 796                 ihold(dir);
 797                 req->r_unsafe_dir = dir;
 798         }
 799 }
 800
 801 static void __unregister_request(struct ceph_mds_client *mdsc,
 802                                  struct ceph_mds_request *req)
 803 {
 804         dout("__unregister_request %p tid %lld\n", req, req->r_tid);
 805
 806         /* Never leave an unregistered request on an unsafe list! */
 807         list_del_init(&req->r_unsafe_item);
 808
 809         if (req->r_tid == mdsc->oldest_tid) {
 810                 struct rb_node *p = rb_next(&req->r_node);
 811                 mdsc->oldest_tid = 0;
 812                 while (p) {
 813                         struct ceph_mds_request *next_req =
 814                                 rb_entry(p, struct ceph_mds_request, r_node);
 815                         if (next_req->r_op != CEPH_MDS_OP_SETFILELOCK) {
 816                                 mdsc->oldest_tid = next_req->r_tid;
 817                                 break;
 818                         }
 819                         p = rb_next(p);
 820                 }
 821         }
 822
 823         erase_request(&mdsc->request_tree, req);
 824
 825         if (req->r_unsafe_dir  &&
 826             test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
 827                 struct ceph_inode_info *ci = ceph_inode(req->r_unsafe_dir);
 828                 spin_lock(&ci->i_unsafe_lock);
 829                 list_del_init(&req->r_unsafe_dir_item);
 830                 spin_unlock(&ci->i_unsafe_lock);
 831         }
 832         if (req->r_target_inode &&
 833             test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
 834                 struct ceph_inode_info *ci = ceph_inode(req->r_target_inode);
 835                 spin_lock(&ci->i_unsafe_lock);
 836                 list_del_init(&req->r_unsafe_target_item);
 837                 spin_unlock(&ci->i_unsafe_lock);
 838         }
 839
 840         if (req->r_unsafe_dir) {
 841                 /* avoid calling iput_final() in mds dispatch threads */
 842                 ceph_async_iput(req->r_unsafe_dir);
 843                 req->r_unsafe_dir = NULL;
 844         }
 845
 846         complete_all(&req->r_safe_completion);
 847
 848         ceph_mdsc_put_request(req);
 849 }
 850
 851 /*
 852  * Walk back up the dentry tree until we hit a dentry representing a
 853  * non-snapshot inode. We do this using the rcu_read_lock (which must be held
 854  * when calling this) to ensure that the objects won't disappear while we're
 855  * working with them. Once we hit a candidate dentry, we attempt to take a
 856  * reference to it, and return that as the result.
 857  */
 858 static struct inode *get_nonsnap_parent(struct dentry *dentry)
 859 {
 860         struct inode *inode = NULL;
 861
 862         while (dentry && !IS_ROOT(dentry)) {
 863                 inode = d_inode_rcu(dentry);
 864                 if (!inode || ceph_snap(inode) == CEPH_NOSNAP)
 865                         break;
 866                 dentry = dentry->d_parent;
 867         }
 868         if (inode)
 869                 inode = igrab(inode);
 870         return inode;
 871 }
 872
 873 /*
 874  * Choose mds to send request to next.  If there is a hint set in the
 875  * request (e.g., due to a prior forward hint from the mds), use that.
 876  * Otherwise, consult frag tree and/or caps to identify the
 877  * appropriate mds.  If all else fails, choose randomly.
 878  *
 879  * Called under mdsc->mutex.
 880  */
 881 static int __choose_mds(struct ceph_mds_client *mdsc,
 882                         struct ceph_mds_request *req,
 883                         bool *random)
 884 {
 885         struct inode *inode;
 886         struct ceph_inode_info *ci;
 887         struct ceph_cap *cap;
 888         int mode = req->r_direct_mode;
 889         int mds = -1;
 890         u32 hash = req->r_direct_hash;
 891         bool is_hash = test_bit(CEPH_MDS_R_DIRECT_IS_HASH, &req->r_req_flags);
 892
 893         if (random)
 894                 *random = false;
 895
 896         /*
 897          * is there a specific mds we should try?  ignore hint if we have
 898          * no session and the mds is not up (active or recovering).
 899          */
 900         if (req->r_resend_mds >= 0 &&
 901             (__have_session(mdsc, req->r_resend_mds) ||
 902              ceph_mdsmap_get_state(mdsc->mdsmap, req->r_resend_mds) > 0)) {
 903                 dout("%s using resend_mds mds%d\n", __func__,
 904                      req->r_resend_mds);
 905                 return req->r_resend_mds;
 906         }
 907
 908         if (mode == USE_RANDOM_MDS)
 909                 goto random;
 910
 911         inode = NULL;
 912         if (req->r_inode) {
 913                 if (ceph_snap(req->r_inode) != CEPH_SNAPDIR) {
 914                         inode = req->r_inode;
 915                         ihold(inode);
 916                 } else {
 917                         /* req->r_dentry is non-null for LSSNAP request */
 918                         rcu_read_lock();
 919                         inode = get_nonsnap_parent(req->r_dentry);
 920                         rcu_read_unlock();
 921                         dout("%s using snapdir's parent %p\n", __func__, inode);
 922                 }
 923         } else if (req->r_dentry) {
 924                 /* ignore race with rename; old or new d_parent is okay */
 925                 struct dentry *parent;
 926                 struct inode *dir;
 927
 928                 rcu_read_lock();
 929                 parent = READ_ONCE(req->r_dentry->d_parent);
 930                 dir = req->r_parent ? : d_inode_rcu(parent);
 931
 932                 if (!dir || dir->i_sb != mdsc->fsc->sb) {
 933                         /*  not this fs or parent went negative */
 934                         inode = d_inode(req->r_dentry);
 935                         if (inode)
 936                                 ihold(inode);
 937                 } else if (ceph_snap(dir) != CEPH_NOSNAP) {
 938                         /* direct snapped/virtual snapdir requests
 939                          * based on parent dir inode */
 940                         inode = get_nonsnap_parent(parent);
 941                         dout("%s using nonsnap parent %p\n", __func__, inode);
 942                 } else {
 943                         /* dentry target */
 944                         inode = d_inode(req->r_dentry);
 945                         if (!inode || mode == USE_AUTH_MDS) {
 946                                 /* dir + name */
 947                                 inode = igrab(dir);
 948                                 hash = ceph_dentry_hash(dir, req->r_dentry);
 949                                 is_hash = true;
 950                         } else {
 951                                 ihold(inode);
 952                         }
 953                 }
 954                 rcu_read_unlock();
 955         }
 956
 957         dout("%s %p is_hash=%d (0x%x) mode %d\n", __func__, inode, (int)is_hash,
 958              hash, mode);
 959         if (!inode)
 960                 goto random;
 961         ci = ceph_inode(inode);
 962
 963         if (is_hash && S_ISDIR(inode->i_mode)) {
 964                 struct ceph_inode_frag frag;
 965                 int found;
 966
 967                 ceph_choose_frag(ci, hash, &frag, &found);
 968                 if (found) {
 969                         if (mode == USE_ANY_MDS && frag.ndist > 0) {
 970                                 u8 r;
 971
 972                                 /* choose a random replica */
 973                                 get_random_bytes(&r, 1);
 974                                 r %= frag.ndist;
 975                                 mds = frag.dist[r];
 976                                 dout("%s %p %llx.%llx frag %u mds%d (%d/%d)\n",
 977                                      __func__, inode, ceph_vinop(inode),
 978                                      frag.frag, mds, (int)r, frag.ndist);
 979                                 if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
 980                                     CEPH_MDS_STATE_ACTIVE &&
 981                                     !ceph_mdsmap_is_laggy(mdsc->mdsmap, mds))
 982                                         goto out;
 983                         }
 984
 985                         /* since this file/dir wasn't known to be
 986                          * replicated, then we want to look for the
 987                          * authoritative mds. */
 988                         if (frag.mds >= 0) {
 989                                 /* choose auth mds */
 990                                 mds = frag.mds;
 991                                 dout("%s %p %llx.%llx frag %u mds%d (auth)\n",
 992                                      __func__, inode, ceph_vinop(inode),
 993                                      frag.frag, mds);
 994                                 if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
 995                                     CEPH_MDS_STATE_ACTIVE) {
 996                                         if (mode == USE_ANY_MDS &&
 997                                             !ceph_mdsmap_is_laggy(mdsc->mdsmap,
 998                                                                   mds))
 999                                                 goto out;
1000                                 }
1001                         }
1002                         mode = USE_AUTH_MDS;
1003                 }
1004         }
1005
1006         spin_lock(&ci->i_ceph_lock);
1007         cap = NULL;
1008         if (mode == USE_AUTH_MDS)
1009                 cap = ci->i_auth_cap;
1010         if (!cap && !RB_EMPTY_ROOT(&ci->i_caps))
1011                 cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node);
1012         if (!cap) {
1013                 spin_unlock(&ci->i_ceph_lock);
1014                 ceph_async_iput(inode);
1015                 goto random;
1016         }
1017         mds = cap->session->s_mds;
1018         dout("%s %p %llx.%llx mds%d (%scap %p)\n", __func__,
1019              inode, ceph_vinop(inode), mds,
1020              cap == ci->i_auth_cap ? "auth " : "", cap);
1021         spin_unlock(&ci->i_ceph_lock);
1022 out:
1023         /* avoid calling iput_final() while holding mdsc->mutex or
1024          * in mds dispatch threads */
1025         ceph_async_iput(inode);
1026         return mds;
1027
1028 random:
1029         if (random)
1030                 *random = true;
1031
1032         mds = ceph_mdsmap_get_random_mds(mdsc->mdsmap);
1033         dout("%s chose random mds%d\n", __func__, mds);
1034         return mds;
1035 }
1036
1037
1038 /*
1039  * session messages
1040  */
1041 static struct ceph_msg *create_session_msg(u32 op, u64 seq)
1042 {
1043         struct ceph_msg *msg;
1044         struct ceph_mds_session_head *h;
1045
1046         msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS,
1047                            false);
1048         if (!msg) {
1049                 pr_err("create_session_msg ENOMEM creating msg\n");
1050                 return NULL;
1051         }
1052         h = msg->front.iov_base;
1053         h->op = cpu_to_le32(op);
1054         h->seq = cpu_to_le64(seq);
1055
1056         return msg;
1057 }
1058
1059 static const unsigned char feature_bits[] = CEPHFS_FEATURES_CLIENT_SUPPORTED;
1060 #define FEATURE_BYTES(c) (DIV_ROUND_UP((size_t)feature_bits[c - 1] + 1, 64) * 8)
1061 static void encode_supported_features(void **p, void *end)
1062 {
1063         static const size_t count = ARRAY_SIZE(feature_bits);
1064
1065         if (count > 0) {
1066                 size_t i;
1067                 size_t size = FEATURE_BYTES(count);
1068
1069                 BUG_ON(*p + 4 + size > end);
1070                 ceph_encode_32(p, size);
1071                 memset(*p, 0, size);
1072                 for (i = 0; i < count; i++)
1073                         ((unsigned char*)(*p))[i / 8] |= BIT(feature_bits[i] % 8);
1074                 *p += size;
1075         } else {
1076                 BUG_ON(*p + 4 > end);
1077                 ceph_encode_32(p, 0);
1078         }
1079 }
1080
1081 /*
1082  * session message, specialization for CEPH_SESSION_REQUEST_OPEN
1083  * to include additional client metadata fields.
1084  */
1085 static struct ceph_msg *create_session_open_msg(struct ceph_mds_client *mdsc, u64 seq)
1086 {
1087         struct ceph_msg *msg;
1088         struct ceph_mds_session_head *h;
1089         int i = -1;
1090         int extra_bytes = 0;
1091         int metadata_key_count = 0;
1092         struct ceph_options *opt = mdsc->fsc->client->options;
1093         struct ceph_mount_options *fsopt = mdsc->fsc->mount_options;
1094         size_t size, count;
1095         void *p, *end;
1096
1097         const char* metadata[][2] = {
1098                 {"hostname", mdsc->nodename},
1099                 {"kernel_version", init_utsname()->release},
1100                 {"entity_id", opt->name ? : ""},
1101                 {"root", fsopt->server_path ? : "/"},
1102                 {NULL, NULL}
1103         };
1104
1105         /* Calculate serialized length of metadata */
1106         extra_bytes = 4;  /* map length */
1107         for (i = 0; metadata[i][0]; ++i) {
1108                 extra_bytes += 8 + strlen(metadata[i][0]) +
1109                         strlen(metadata[i][1]);
1110                 metadata_key_count++;
1111         }
1112
1113         /* supported feature */
1114         size = 0;
1115         count = ARRAY_SIZE(feature_bits);
1116         if (count > 0)
1117                 size = FEATURE_BYTES(count);
1118         extra_bytes += 4 + size;
1119
1120         /* Allocate the message */
1121         msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h) + extra_bytes,
1122                            GFP_NOFS, false);
1123         if (!msg) {
1124                 pr_err("create_session_msg ENOMEM creating msg\n");
1125                 return NULL;
1126         }
1127         p = msg->front.iov_base;
1128         end = p + msg->front.iov_len;
1129
1130         h = p;
1131         h->op = cpu_to_le32(CEPH_SESSION_REQUEST_OPEN);
1132         h->seq = cpu_to_le64(seq);
1133
1134         /*
1135          * Serialize client metadata into waiting buffer space, using
1136          * the format that userspace expects for map<string, string>
1137          *
1138          * ClientSession messages with metadata are v3
1139          */
1140         msg->hdr.version = cpu_to_le16(3);
1141         msg->hdr.compat_version = cpu_to_le16(1);
1142
1143         /* The write pointer, following the session_head structure */
1144         p += sizeof(*h);
1145
1146         /* Number of entries in the map */
1147         ceph_encode_32(&p, metadata_key_count);
1148
1149         /* Two length-prefixed strings for each entry in the map */
1150         for (i = 0; metadata[i][0]; ++i) {
1151                 size_t const key_len = strlen(metadata[i][0]);
1152                 size_t const val_len = strlen(metadata[i][1]);
1153
1154                 ceph_encode_32(&p, key_len);
1155                 memcpy(p, metadata[i][0], key_len);
1156                 p += key_len;
1157                 ceph_encode_32(&p, val_len);
1158                 memcpy(p, metadata[i][1], val_len);
1159                 p += val_len;
1160         }
1161
1162         encode_supported_features(&p, end);
1163         msg->front.iov_len = p - msg->front.iov_base;
1164         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
1165
1166         return msg;
1167 }
1168
1169 /*
1170  * send session open request.
1171  *
1172  * called under mdsc->mutex
1173  */
1174 static int __open_session(struct ceph_mds_client *mdsc,
1175                           struct ceph_mds_session *session)
1176 {
1177         struct ceph_msg *msg;
1178         int mstate;
1179         int mds = session->s_mds;
1180
1181         /* wait for mds to go active? */
1182         mstate = ceph_mdsmap_get_state(mdsc->mdsmap, mds);
1183         dout("open_session to mds%d (%s)\n", mds,
1184              ceph_mds_state_name(mstate));
1185         session->s_state = CEPH_MDS_SESSION_OPENING;
1186         session->s_renew_requested = jiffies;
1187
1188         /* send connect message */
1189         msg = create_session_open_msg(mdsc, session->s_seq);
1190         if (!msg)
1191                 return -ENOMEM;
1192         ceph_con_send(&session->s_con, msg);
1193         return 0;
1194 }
1195
1196 /*
1197  * open sessions for any export targets for the given mds
1198  *
1199  * called under mdsc->mutex
1200  */
1201 static struct ceph_mds_session *
1202 __open_export_target_session(struct ceph_mds_client *mdsc, int target)
1203 {
1204         struct ceph_mds_session *session;
1205
1206         session = __ceph_lookup_mds_session(mdsc, target);
1207         if (!session) {
1208                 session = register_session(mdsc, target);
1209                 if (IS_ERR(session))
1210                         return session;
1211         }
1212         if (session->s_state == CEPH_MDS_SESSION_NEW ||
1213             session->s_state == CEPH_MDS_SESSION_CLOSING)
1214                 __open_session(mdsc, session);
1215
1216         return session;
1217 }
1218
1219 struct ceph_mds_session *
1220 ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target)
1221 {
1222         struct ceph_mds_session *session;
1223
1224         dout("open_export_target_session to mds%d\n", target);
1225
1226         mutex_lock(&mdsc->mutex);
1227         session = __open_export_target_session(mdsc, target);
1228         mutex_unlock(&mdsc->mutex);
1229
1230         return session;
1231 }
1232
1233 static void __open_export_target_sessions(struct ceph_mds_client *mdsc,
1234                                           struct ceph_mds_session *session)
1235 {
1236         struct ceph_mds_info *mi;
1237         struct ceph_mds_session *ts;
1238         int i, mds = session->s_mds;
1239
1240         if (mds >= mdsc->mdsmap->possible_max_rank)
1241                 return;
1242
1243         mi = &mdsc->mdsmap->m_info[mds];
1244         dout("open_export_target_sessions for mds%d (%d targets)\n",
1245              session->s_mds, mi->num_export_targets);
1246
1247         for (i = 0; i < mi->num_export_targets; i++) {
1248                 ts = __open_export_target_session(mdsc, mi->export_targets[i]);
1249                 if (!IS_ERR(ts))
1250                         ceph_put_mds_session(ts);
1251         }
1252 }
1253
1254 void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
1255                                            struct ceph_mds_session *session)
1256 {
1257         mutex_lock(&mdsc->mutex);
1258         __open_export_target_sessions(mdsc, session);
1259         mutex_unlock(&mdsc->mutex);
1260 }
1261
1262 /*
1263  * session caps
1264  */
1265
1266 static void detach_cap_releases(struct ceph_mds_session *session,
1267                                 struct list_head *target)
1268 {
1269         lockdep_assert_held(&session->s_cap_lock);
1270
1271         list_splice_init(&session->s_cap_releases, target);
1272         session->s_num_cap_releases = 0;
1273         dout("dispose_cap_releases mds%d\n", session->s_mds);
1274 }
1275
1276 static void dispose_cap_releases(struct ceph_mds_client *mdsc,
1277                                  struct list_head *dispose)
1278 {
1279         while (!list_empty(dispose)) {
1280                 struct ceph_cap *cap;
1281                 /* zero out the in-progress message */
1282                 cap = list_first_entry(dispose, struct ceph_cap, session_caps);
1283                 list_del(&cap->session_caps);
1284                 ceph_put_cap(mdsc, cap);
1285         }
1286 }
1287
1288 static void cleanup_session_requests(struct ceph_mds_client *mdsc,
1289                                      struct ceph_mds_session *session)
1290 {
1291         struct ceph_mds_request *req;
1292         struct rb_node *p;
1293         struct ceph_inode_info *ci;
1294
1295         dout("cleanup_session_requests mds%d\n", session->s_mds);
1296         mutex_lock(&mdsc->mutex);
1297         while (!list_empty(&session->s_unsafe)) {
1298                 req = list_first_entry(&session->s_unsafe,
1299                                        struct ceph_mds_request, r_unsafe_item);
1300                 pr_warn_ratelimited(" dropping unsafe request %llu\n",
1301                                     req->r_tid);
1302                 if (req->r_target_inode) {
1303                         /* dropping unsafe change of inode's attributes */
1304                         ci = ceph_inode(req->r_target_inode);
1305                         errseq_set(&ci->i_meta_err, -EIO);
1306                 }
1307                 if (req->r_unsafe_dir) {
1308                         /* dropping unsafe directory operation */
1309                         ci = ceph_inode(req->r_unsafe_dir);
1310                         errseq_set(&ci->i_meta_err, -EIO);
1311                 }
1312                 __unregister_request(mdsc, req);
1313         }
1314         /* zero r_attempts, so kick_requests() will re-send requests */
1315         p = rb_first(&mdsc->request_tree);
1316         while (p) {
1317                 req = rb_entry(p, struct ceph_mds_request, r_node);
1318                 p = rb_next(p);
1319                 if (req->r_session &&
1320                     req->r_session->s_mds == session->s_mds)
1321                         req->r_attempts = 0;
1322         }
1323         mutex_unlock(&mdsc->mutex);
1324 }
1325
1326 /*
1327  * Helper to safely iterate over all caps associated with a session, with
1328  * special care taken to handle a racing __ceph_remove_cap().
1329  *
1330  * Caller must hold session s_mutex.
1331  */
1332 int ceph_iterate_session_caps(struct ceph_mds_session *session,
1333                               int (*cb)(struct inode *, struct ceph_cap *,
1334                                         void *), void *arg)
1335 {
1336         struct list_head *p;
1337         struct ceph_cap *cap;
1338         struct inode *inode, *last_inode = NULL;
1339         struct ceph_cap *old_cap = NULL;
1340         int ret;
1341
1342         dout("iterate_session_caps %p mds%d\n", session, session->s_mds);
1343         spin_lock(&session->s_cap_lock);
1344         p = session->s_caps.next;
1345         while (p != &session->s_caps) {
1346                 cap = list_entry(p, struct ceph_cap, session_caps);
1347                 inode = igrab(&cap->ci->vfs_inode);
1348                 if (!inode) {
1349                         p = p->next;
1350                         continue;
1351                 }
1352                 session->s_cap_iterator = cap;
1353                 spin_unlock(&session->s_cap_lock);
1354
1355                 if (last_inode) {
1356                         /* avoid calling iput_final() while holding
1357                          * s_mutex or in mds dispatch threads */
1358                         ceph_async_iput(last_inode);
1359                         last_inode = NULL;
1360                 }
1361                 if (old_cap) {
1362                         ceph_put_cap(session->s_mdsc, old_cap);
1363                         old_cap = NULL;
1364                 }
1365
1366                 ret = cb(inode, cap, arg);
1367                 last_inode = inode;
1368
1369                 spin_lock(&session->s_cap_lock);
1370                 p = p->next;
1371                 if (!cap->ci) {
1372                         dout("iterate_session_caps  finishing cap %p removal\n",
1373                              cap);
1374                         BUG_ON(cap->session != session);
1375                         cap->session = NULL;
1376                         list_del_init(&cap->session_caps);
1377                         session->s_nr_caps--;
1378                         if (cap->queue_release)
1379                                 __ceph_queue_cap_release(session, cap);
1380                         else
1381                                 old_cap = cap;  /* put_cap it w/o locks held */
1382                 }
1383                 if (ret < 0)
1384                         goto out;
1385         }
1386         ret = 0;
1387 out:
1388         session->s_cap_iterator = NULL;
1389         spin_unlock(&session->s_cap_lock);
1390
1391         ceph_async_iput(last_inode);
1392         if (old_cap)
1393                 ceph_put_cap(session->s_mdsc, old_cap);
1394
1395         return ret;
1396 }
1397
1398 static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
1399                                   void *arg)
1400 {
1401         struct ceph_fs_client *fsc = (struct ceph_fs_client *)arg;
1402         struct ceph_inode_info *ci = ceph_inode(inode);
1403         LIST_HEAD(to_remove);
1404         bool dirty_dropped = false;
1405         bool invalidate = false;
1406
1407         dout("removing cap %p, ci is %p, inode is %p\n",
1408              cap, ci, &ci->vfs_inode);
1409         spin_lock(&ci->i_ceph_lock);
1410         if (cap->mds_wanted | cap->issued)
1411                 ci->i_ceph_flags |= CEPH_I_CAP_DROPPED;
1412         __ceph_remove_cap(cap, false);
1413         if (!ci->i_auth_cap) {
1414                 struct ceph_cap_flush *cf;
1415                 struct ceph_mds_client *mdsc = fsc->mdsc;
1416
1417                 if (READ_ONCE(fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
1418                         if (inode->i_data.nrpages > 0)
1419                                 invalidate = true;
1420                         if (ci->i_wrbuffer_ref > 0)
1421                                 mapping_set_error(&inode->i_data, -EIO);
1422                 }
1423
1424                 while (!list_empty(&ci->i_cap_flush_list)) {
1425                         cf = list_first_entry(&ci->i_cap_flush_list,
1426                                               struct ceph_cap_flush, i_list);
1427                         list_move(&cf->i_list, &to_remove);
1428                 }
1429
1430                 spin_lock(&mdsc->cap_dirty_lock);
1431
1432                 list_for_each_entry(cf, &to_remove, i_list)
1433                         list_del(&cf->g_list);
1434
1435                 if (!list_empty(&ci->i_dirty_item)) {
1436                         pr_warn_ratelimited(
1437                                 " dropping dirty %s state for %p %lld\n",
1438                                 ceph_cap_string(ci->i_dirty_caps),
1439                                 inode, ceph_ino(inode));
1440                         ci->i_dirty_caps = 0;
1441                         list_del_init(&ci->i_dirty_item);
1442                         dirty_dropped = true;
1443                 }
1444                 if (!list_empty(&ci->i_flushing_item)) {
1445                         pr_warn_ratelimited(
1446                                 " dropping dirty+flushing %s state for %p %lld\n",
1447                                 ceph_cap_string(ci->i_flushing_caps),
1448                                 inode, ceph_ino(inode));
1449                         ci->i_flushing_caps = 0;
1450                         list_del_init(&ci->i_flushing_item);
1451                         mdsc->num_cap_flushing--;
1452                         dirty_dropped = true;
1453                 }
1454                 spin_unlock(&mdsc->cap_dirty_lock);
1455
1456                 if (dirty_dropped) {
1457                         errseq_set(&ci->i_meta_err, -EIO);
1458
1459                         if (ci->i_wrbuffer_ref_head == 0 &&
1460                             ci->i_wr_ref == 0 &&
1461                             ci->i_dirty_caps == 0 &&
1462                             ci->i_flushing_caps == 0) {
1463                                 ceph_put_snap_context(ci->i_head_snapc);
1464                                 ci->i_head_snapc = NULL;
1465                         }
1466                 }
1467
1468                 if (atomic_read(&ci->i_filelock_ref) > 0) {
1469                         /* make further file lock syscall return -EIO */
1470                         ci->i_ceph_flags |= CEPH_I_ERROR_FILELOCK;
1471                         pr_warn_ratelimited(" dropping file locks for %p %lld\n",
1472                                             inode, ceph_ino(inode));
1473                 }
1474
1475                 if (!ci->i_dirty_caps && ci->i_prealloc_cap_flush) {
1476                         list_add(&ci->i_prealloc_cap_flush->i_list, &to_remove);
1477                         ci->i_prealloc_cap_flush = NULL;
1478                 }
1479         }
1480         spin_unlock(&ci->i_ceph_lock);
1481         while (!list_empty(&to_remove)) {
1482                 struct ceph_cap_flush *cf;
1483                 cf = list_first_entry(&to_remove,
1484                                       struct ceph_cap_flush, i_list);
1485                 list_del(&cf->i_list);
1486                 ceph_free_cap_flush(cf);
1487         }
1488
1489         wake_up_all(&ci->i_cap_wq);
1490         if (invalidate)
1491                 ceph_queue_invalidate(inode);
1492         if (dirty_dropped)
1493                 iput(inode);
1494         return 0;
1495 }
1496
1497 /*
1498  * caller must hold session s_mutex
1499  */
1500 static void remove_session_caps(struct ceph_mds_session *session)
1501 {
1502         struct ceph_fs_client *fsc = session->s_mdsc->fsc;
1503         struct super_block *sb = fsc->sb;
1504         LIST_HEAD(dispose);
1505
1506         dout("remove_session_caps on %p\n", session);
1507         ceph_iterate_session_caps(session, remove_session_caps_cb, fsc);
1508
1509         wake_up_all(&fsc->mdsc->cap_flushing_wq);
1510
1511         spin_lock(&session->s_cap_lock);
1512         if (session->s_nr_caps > 0) {
1513                 struct inode *inode;
1514                 struct ceph_cap *cap, *prev = NULL;
1515                 struct ceph_vino vino;
1516                 /*
1517                  * iterate_session_caps() skips inodes that are being
1518                  * deleted, we need to wait until deletions are complete.
1519                  * __wait_on_freeing_inode() is designed for the job,
1520                  * but it is not exported, so use lookup inode function
1521                  * to access it.
1522                  */
1523                 while (!list_empty(&session->s_caps)) {
1524                         cap = list_entry(session->s_caps.next,
1525                                          struct ceph_cap, session_caps);
1526                         if (cap == prev)
1527                                 break;
1528                         prev = cap;
1529                         vino = cap->ci->i_vino;
1530                         spin_unlock(&session->s_cap_lock);
1531
1532                         inode = ceph_find_inode(sb, vino);
1533                          /* avoid calling iput_final() while holding s_mutex */
1534                         ceph_async_iput(inode);
1535
1536                         spin_lock(&session->s_cap_lock);
1537                 }
1538         }
1539
1540         // drop cap expires and unlock s_cap_lock
1541         detach_cap_releases(session, &dispose);
1542
1543         BUG_ON(session->s_nr_caps > 0);
1544         BUG_ON(!list_empty(&session->s_cap_flushing));
1545         spin_unlock(&session->s_cap_lock);
1546         dispose_cap_releases(session->s_mdsc, &dispose);
1547 }
1548
1549 enum {
1550         RECONNECT,
1551         RENEWCAPS,
1552         FORCE_RO,
1553 };
1554
1555 /*
1556  * wake up any threads waiting on this session's caps.  if the cap is
1557  * old (didn't get renewed on the client reconnect), remove it now.
1558  *
1559  * caller must hold s_mutex.
1560  */
1561 static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap,
1562                               void *arg)
1563 {
1564         struct ceph_inode_info *ci = ceph_inode(inode);
1565         unsigned long ev = (unsigned long)arg;
1566
1567         if (ev == RECONNECT) {
1568                 spin_lock(&ci->i_ceph_lock);
1569                 ci->i_wanted_max_size = 0;
1570                 ci->i_requested_max_size = 0;
1571                 spin_unlock(&ci->i_ceph_lock);
1572         } else if (ev == RENEWCAPS) {
1573                 if (cap->cap_gen < cap->session->s_cap_gen) {
1574                         /* mds did not re-issue stale cap */
1575                         spin_lock(&ci->i_ceph_lock);
1576                         cap->issued = cap->implemented = CEPH_CAP_PIN;
1577                         /* make sure mds knows what we want */
1578                         if (__ceph_caps_file_wanted(ci) & ~cap->mds_wanted)
1579                                 ci->i_ceph_flags |= CEPH_I_CAP_DROPPED;
1580                         spin_unlock(&ci->i_ceph_lock);
1581                 }
1582         } else if (ev == FORCE_RO) {
1583         }
1584         wake_up_all(&ci->i_cap_wq);
1585         return 0;
1586 }
1587
1588 static void wake_up_session_caps(struct ceph_mds_session *session, int ev)
1589 {
1590         dout("wake_up_session_caps %p mds%d\n", session, session->s_mds);
1591         ceph_iterate_session_caps(session, wake_up_session_cb,
1592                                   (void *)(unsigned long)ev);
1593 }
1594
1595 /*
1596  * Send periodic message to MDS renewing all currently held caps.  The
1597  * ack will reset the expiration for all caps from this session.
1598  *
1599  * caller holds s_mutex
1600  */
1601 static int send_renew_caps(struct ceph_mds_client *mdsc,
1602                            struct ceph_mds_session *session)
1603 {
1604         struct ceph_msg *msg;
1605         int state;
1606
1607         if (time_after_eq(jiffies, session->s_cap_ttl) &&
1608             time_after_eq(session->s_cap_ttl, session->s_renew_requested))
1609                 pr_info("mds%d caps stale\n", session->s_mds);
1610         session->s_renew_requested = jiffies;
1611
1612         /* do not try to renew caps until a recovering mds has reconnected
1613          * with its clients. */
1614         state = ceph_mdsmap_get_state(mdsc->mdsmap, session->s_mds);
1615         if (state < CEPH_MDS_STATE_RECONNECT) {
1616                 dout("send_renew_caps ignoring mds%d (%s)\n",
1617                      session->s_mds, ceph_mds_state_name(state));
1618                 return 0;
1619         }
1620
1621         dout("send_renew_caps to mds%d (%s)\n", session->s_mds,
1622                 ceph_mds_state_name(state));
1623         msg = create_session_msg(CEPH_SESSION_REQUEST_RENEWCAPS,
1624                                  ++session->s_renew_seq);
1625         if (!msg)
1626                 return -ENOMEM;
1627         ceph_con_send(&session->s_con, msg);
1628         return 0;
1629 }
1630
1631 static int send_flushmsg_ack(struct ceph_mds_client *mdsc,
1632                              struct ceph_mds_session *session, u64 seq)
1633 {
1634         struct ceph_msg *msg;
1635
1636         dout("send_flushmsg_ack to mds%d (%s)s seq %lld\n",
1637              session->s_mds, ceph_session_state_name(session->s_state), seq);
1638         msg = create_session_msg(CEPH_SESSION_FLUSHMSG_ACK, seq);
1639         if (!msg)
1640                 return -ENOMEM;
1641         ceph_con_send(&session->s_con, msg);
1642         return 0;
1643 }
1644
1645
1646 /*
1647  * Note new cap ttl, and any transition from stale -> not stale (fresh?).
1648  *
1649  * Called under session->s_mutex
1650  */
1651 static void renewed_caps(struct ceph_mds_client *mdsc,
1652                          struct ceph_mds_session *session, int is_renew)
1653 {
1654         int was_stale;
1655         int wake = 0;
1656
1657         spin_lock(&session->s_cap_lock);
1658         was_stale = is_renew && time_after_eq(jiffies, session->s_cap_ttl);
1659
1660         session->s_cap_ttl = session->s_renew_requested +
1661                 mdsc->mdsmap->m_session_timeout*HZ;
1662
1663         if (was_stale) {
1664                 if (time_before(jiffies, session->s_cap_ttl)) {
1665                         pr_info("mds%d caps renewed\n", session->s_mds);
1666                         wake = 1;
1667                 } else {
1668                         pr_info("mds%d caps still stale\n", session->s_mds);
1669                 }
1670         }
1671         dout("renewed_caps mds%d ttl now %lu, was %s, now %s\n",
1672              session->s_mds, session->s_cap_ttl, was_stale ? "stale" : "fresh",
1673              time_before(jiffies, session->s_cap_ttl) ? "stale" : "fresh");
1674         spin_unlock(&session->s_cap_lock);
1675
1676         if (wake)
1677                 wake_up_session_caps(session, RENEWCAPS);
1678 }
1679
1680 /*
1681  * send a session close request
1682  */
1683 static int request_close_session(struct ceph_mds_client *mdsc,
1684                                  struct ceph_mds_session *session)
1685 {
1686         struct ceph_msg *msg;
1687
1688         dout("request_close_session mds%d state %s seq %lld\n",
1689              session->s_mds, ceph_session_state_name(session->s_state),
1690              session->s_seq);
1691         msg = create_session_msg(CEPH_SESSION_REQUEST_CLOSE, session->s_seq);
1692         if (!msg)
1693                 return -ENOMEM;
1694         ceph_con_send(&session->s_con, msg);
1695         return 1;
1696 }
1697
1698 /*
1699  * Called with s_mutex held.
1700  */
1701 static int __close_session(struct ceph_mds_client *mdsc,
1702                          struct ceph_mds_session *session)
1703 {
1704         if (session->s_state >= CEPH_MDS_SESSION_CLOSING)
1705                 return 0;
1706         session->s_state = CEPH_MDS_SESSION_CLOSING;
1707         return request_close_session(mdsc, session);
1708 }
1709
1710 static bool drop_negative_children(struct dentry *dentry)
1711 {
1712         struct dentry *child;
1713         bool all_negative = true;
1714
1715         if (!d_is_dir(dentry))
1716                 goto out;
1717
1718         spin_lock(&dentry->d_lock);
1719         list_for_each_entry(child, &dentry->d_subdirs, d_child) {
1720                 if (d_really_is_positive(child)) {
1721                         all_negative = false;
1722                         break;
1723                 }
1724         }
1725         spin_unlock(&dentry->d_lock);
1726
1727         if (all_negative)
1728                 shrink_dcache_parent(dentry);
1729 out:
1730         return all_negative;
1731 }
1732
1733 /*
1734  * Trim old(er) caps.
1735  *
1736  * Because we can't cache an inode without one or more caps, we do
1737  * this indirectly: if a cap is unused, we prune its aliases, at which
1738  * point the inode will hopefully get dropped to.
1739  *
1740  * Yes, this is a bit sloppy.  Our only real goal here is to respond to
1741  * memory pressure from the MDS, though, so it needn't be perfect.
1742  */
1743 static int trim_caps_cb(struct inode *inode, struct ceph_cap *cap, void *arg)
1744 {
1745         int *remaining = arg;
1746         struct ceph_inode_info *ci = ceph_inode(inode);
1747         int used, wanted, oissued, mine;
1748
1749         if (*remaining <= 0)
1750                 return -1;
1751
1752         spin_lock(&ci->i_ceph_lock);
1753         mine = cap->issued | cap->implemented;
1754         used = __ceph_caps_used(ci);
1755         wanted = __ceph_caps_file_wanted(ci);
1756         oissued = __ceph_caps_issued_other(ci, cap);
1757
1758         dout("trim_caps_cb %p cap %p mine %s oissued %s used %s wanted %s\n",
1759              inode, cap, ceph_cap_string(mine), ceph_cap_string(oissued),
1760              ceph_cap_string(used), ceph_cap_string(wanted));
1761         if (cap == ci->i_auth_cap) {
1762                 if (ci->i_dirty_caps || ci->i_flushing_caps ||
1763                     !list_empty(&ci->i_cap_snaps))
1764                         goto out;
1765                 if ((used | wanted) & CEPH_CAP_ANY_WR)
1766                         goto out;
1767                 /* Note: it's possible that i_filelock_ref becomes non-zero
1768                  * after dropping auth caps. It doesn't hurt because reply
1769                  * of lock mds request will re-add auth caps. */
1770                 if (atomic_read(&ci->i_filelock_ref) > 0)
1771                         goto out;
1772         }
1773         /* The inode has cached pages, but it's no longer used.
1774          * we can safely drop it */
1775         if (wanted == 0 && used == CEPH_CAP_FILE_CACHE &&
1776             !(oissued & CEPH_CAP_FILE_CACHE)) {
1777           used = 0;
1778           oissued = 0;
1779         }
1780         if ((used | wanted) & ~oissued & mine)
1781                 goto out;   /* we need these caps */
1782
1783         if (oissued) {
1784                 /* we aren't the only cap.. just remove us */
1785                 __ceph_remove_cap(cap, true);
1786                 (*remaining)--;
1787         } else {
1788                 struct dentry *dentry;
1789                 /* try dropping referring dentries */
1790                 spin_unlock(&ci->i_ceph_lock);
1791                 dentry = d_find_any_alias(inode);
1792                 if (dentry && drop_negative_children(dentry)) {
1793                         int count;
1794                         dput(dentry);
1795                         d_prune_aliases(inode);
1796                         count = atomic_read(&inode->i_count);
1797                         if (count == 1)
1798                                 (*remaining)--;
1799                         dout("trim_caps_cb %p cap %p pruned, count now %d\n",
1800                              inode, cap, count);
1801                 } else {
1802                         dput(dentry);
1803                 }
1804                 return 0;
1805         }
1806
1807 out:
1808         spin_unlock(&ci->i_ceph_lock);
1809         return 0;
1810 }
1811
1812 /*
1813  * Trim session cap count down to some max number.
1814  */
1815 int ceph_trim_caps(struct ceph_mds_client *mdsc,
1816                    struct ceph_mds_session *session,
1817                    int max_caps)
1818 {
1819         int trim_caps = session->s_nr_caps - max_caps;
1820
1821         dout("trim_caps mds%d start: %d / %d, trim %d\n",
1822              session->s_mds, session->s_nr_caps, max_caps, trim_caps);
1823         if (trim_caps > 0) {
1824                 int remaining = trim_caps;
1825
1826                 ceph_iterate_session_caps(session, trim_caps_cb, &remaining);
1827                 dout("trim_caps mds%d done: %d / %d, trimmed %d\n",
1828                      session->s_mds, session->s_nr_caps, max_caps,
1829                         trim_caps - remaining);
1830         }
1831
1832         ceph_flush_cap_releases(mdsc, session);
1833         return 0;
1834 }
1835
1836 static int check_caps_flush(struct ceph_mds_client *mdsc,
1837                             u64 want_flush_tid)
1838 {
1839         int ret = 1;
1840
1841         spin_lock(&mdsc->cap_dirty_lock);
1842         if (!list_empty(&mdsc->cap_flush_list)) {
1843                 struct ceph_cap_flush *cf =
1844                         list_first_entry(&mdsc->cap_flush_list,
1845                                          struct ceph_cap_flush, g_list);
1846                 if (cf->tid <= want_flush_tid) {
1847                         dout("check_caps_flush still flushing tid "
1848                              "%llu <= %llu\n", cf->tid, want_flush_tid);
1849                         ret = 0;
1850                 }
1851         }
1852         spin_unlock(&mdsc->cap_dirty_lock);
1853         return ret;
1854 }
1855
1856 /*
1857  * flush all dirty inode data to disk.
1858  *
1859  * returns true if we've flushed through want_flush_tid
1860  */
1861 static void wait_caps_flush(struct ceph_mds_client *mdsc,
1862                             u64 want_flush_tid)
1863 {
1864         dout("check_caps_flush want %llu\n", want_flush_tid);
1865
1866         wait_event(mdsc->cap_flushing_wq,
1867                    check_caps_flush(mdsc, want_flush_tid));
1868
1869         dout("check_caps_flush ok, flushed thru %llu\n", want_flush_tid);
1870 }
1871
1872 /*
1873  * called under s_mutex
1874  */
1875 static void ceph_send_cap_releases(struct ceph_mds_client *mdsc,
1876                                    struct ceph_mds_session *session)
1877 {
1878         struct ceph_msg *msg = NULL;
1879         struct ceph_mds_cap_release *head;
1880         struct ceph_mds_cap_item *item;
1881         struct ceph_osd_client *osdc = &mdsc->fsc->client->osdc;
1882         struct ceph_cap *cap;
1883         LIST_HEAD(tmp_list);
1884         int num_cap_releases;
1885         __le32  barrier, *cap_barrier;
1886
1887         down_read(&osdc->lock);
1888         barrier = cpu_to_le32(osdc->epoch_barrier);
1889         up_read(&osdc->lock);
1890
1891         spin_lock(&session->s_cap_lock);
1892 again:
1893         list_splice_init(&session->s_cap_releases, &tmp_list);
1894         num_cap_releases = session->s_num_cap_releases;
1895         session->s_num_cap_releases = 0;
1896         spin_unlock(&session->s_cap_lock);
1897
1898         while (!list_empty(&tmp_list)) {
1899                 if (!msg) {
1900                         msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPRELEASE,
1901                                         PAGE_SIZE, GFP_NOFS, false);
1902                         if (!msg)
1903                                 goto out_err;
1904                         head = msg->front.iov_base;
1905                         head->num = cpu_to_le32(0);
1906                         msg->front.iov_len = sizeof(*head);
1907
1908                         msg->hdr.version = cpu_to_le16(2);
1909                         msg->hdr.compat_version = cpu_to_le16(1);
1910                 }
1911
1912                 cap = list_first_entry(&tmp_list, struct ceph_cap,
1913                                         session_caps);
1914                 list_del(&cap->session_caps);
1915                 num_cap_releases--;
1916
1917                 head = msg->front.iov_base;
1918                 put_unaligned_le32(get_unaligned_le32(&head->num) + 1,
1919                                    &head->num);
1920                 item = msg->front.iov_base + msg->front.iov_len;
1921                 item->ino = cpu_to_le64(cap->cap_ino);
1922                 item->cap_id = cpu_to_le64(cap->cap_id);
1923                 item->migrate_seq = cpu_to_le32(cap->mseq);
1924                 item->seq = cpu_to_le32(cap->issue_seq);
1925                 msg->front.iov_len += sizeof(*item);
1926
1927                 ceph_put_cap(mdsc, cap);
1928
1929                 if (le32_to_cpu(head->num) == CEPH_CAPS_PER_RELEASE) {
1930                         // Append cap_barrier field
1931                         cap_barrier = msg->front.iov_base + msg->front.iov_len;
1932                         *cap_barrier = barrier;
1933                         msg->front.iov_len += sizeof(*cap_barrier);
1934
1935                         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
1936                         dout("send_cap_releases mds%d %p\n", session->s_mds, msg);
1937                         ceph_con_send(&session->s_con, msg);
1938                         msg = NULL;
1939                 }
1940         }
1941
1942         BUG_ON(num_cap_releases != 0);
1943
1944         spin_lock(&session->s_cap_lock);
1945         if (!list_empty(&session->s_cap_releases))
1946                 goto again;
1947         spin_unlock(&session->s_cap_lock);
1948
1949         if (msg) {
1950                 // Append cap_barrier field
1951                 cap_barrier = msg->front.iov_base + msg->front.iov_len;
1952                 *cap_barrier = barrier;
1953                 msg->front.iov_len += sizeof(*cap_barrier);
1954
1955                 msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
1956                 dout("send_cap_releases mds%d %p\n", session->s_mds, msg);
1957                 ceph_con_send(&session->s_con, msg);
1958         }
1959         return;
1960 out_err:
1961         pr_err("send_cap_releases mds%d, failed to allocate message\n",
1962                 session->s_mds);
1963         spin_lock(&session->s_cap_lock);
1964         list_splice(&tmp_list, &session->s_cap_releases);
1965         session->s_num_cap_releases += num_cap_releases;
1966         spin_unlock(&session->s_cap_lock);
1967 }
1968
1969 static void ceph_cap_release_work(struct work_struct *work)
1970 {
1971         struct ceph_mds_session *session =
1972                 container_of(work, struct ceph_mds_session, s_cap_release_work);
1973
1974         mutex_lock(&session->s_mutex);
1975         if (session->s_state == CEPH_MDS_SESSION_OPEN ||
1976             session->s_state == CEPH_MDS_SESSION_HUNG)
1977                 ceph_send_cap_releases(session->s_mdsc, session);
1978         mutex_unlock(&session->s_mutex);
1979         ceph_put_mds_session(session);
1980 }
1981
1982 void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
1983                              struct ceph_mds_session *session)
1984 {
1985         if (mdsc->stopping)
1986                 return;
1987
1988         ceph_get_mds_session(session);
1989         if (queue_work(mdsc->fsc->cap_wq,
1990                        &session->s_cap_release_work)) {
1991                 dout("cap release work queued\n");
1992         } else {
1993                 ceph_put_mds_session(session);
1994                 dout("failed to queue cap release work\n");
1995         }
1996 }
1997
1998 /*
1999  * caller holds session->s_cap_lock
2000  */
2001 void __ceph_queue_cap_release(struct ceph_mds_session *session,
2002                               struct ceph_cap *cap)
2003 {
2004         list_add_tail(&cap->session_caps, &session->s_cap_releases);
2005         session->s_num_cap_releases++;
2006
2007         if (!(session->s_num_cap_releases % CEPH_CAPS_PER_RELEASE))
2008                 ceph_flush_cap_releases(session->s_mdsc, session);
2009 }
2010
2011 static void ceph_cap_reclaim_work(struct work_struct *work)
2012 {
2013         struct ceph_mds_client *mdsc =
2014                 container_of(work, struct ceph_mds_client, cap_reclaim_work);
2015         int ret = ceph_trim_dentries(mdsc);
2016         if (ret == -EAGAIN)
2017                 ceph_queue_cap_reclaim_work(mdsc);
2018 }
2019
2020 void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc)
2021 {
2022         if (mdsc->stopping)
2023                 return;
2024
2025         if (queue_work(mdsc->fsc->cap_wq, &mdsc->cap_reclaim_work)) {
2026                 dout("caps reclaim work queued\n");
2027         } else {
2028                 dout("failed to queue caps release work\n");
2029         }
2030 }
2031
2032 void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr)
2033 {
2034         int val;
2035         if (!nr)
2036                 return;
2037         val = atomic_add_return(nr, &mdsc->cap_reclaim_pending);
2038         if ((val % CEPH_CAPS_PER_RELEASE) < nr) {
2039                 atomic_set(&mdsc->cap_reclaim_pending, 0);
2040                 ceph_queue_cap_reclaim_work(mdsc);
2041         }
2042 }
2043
2044 /*
2045  * requests
2046  */
2047
2048 int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
2049                                     struct inode *dir)
2050 {
2051         struct ceph_inode_info *ci = ceph_inode(dir);
2052         struct ceph_mds_reply_info_parsed *rinfo = &req->r_reply_info;
2053         struct ceph_mount_options *opt = req->r_mdsc->fsc->mount_options;
2054         size_t size = sizeof(struct ceph_mds_reply_dir_entry);
2055         unsigned int num_entries;
2056         int order;
2057
2058         spin_lock(&ci->i_ceph_lock);
2059         num_entries = ci->i_files + ci->i_subdirs;
2060         spin_unlock(&ci->i_ceph_lock);
2061         num_entries = max(num_entries, 1U);
2062         num_entries = min(num_entries, opt->max_readdir);
2063
2064         order = get_order(size * num_entries);
2065         while (order >= 0) {
2066                 rinfo->dir_entries = (void*)__get_free_pages(GFP_KERNEL |
2067                                                              __GFP_NOWARN,
2068                                                              order);
2069                 if (rinfo->dir_entries)
2070                         break;
2071                 order--;
2072         }
2073         if (!rinfo->dir_entries)
2074                 return -ENOMEM;
2075
2076         num_entries = (PAGE_SIZE << order) / size;
2077         num_entries = min(num_entries, opt->max_readdir);
2078
2079         rinfo->dir_buf_size = PAGE_SIZE << order;
2080         req->r_num_caps = num_entries + 1;
2081         req->r_args.readdir.max_entries = cpu_to_le32(num_entries);
2082         req->r_args.readdir.max_bytes = cpu_to_le32(opt->max_readdir_bytes);
2083         return 0;
2084 }
2085
2086 /*
2087  * Create an mds request.
2088  */
2089 struct ceph_mds_request *
2090 ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode)
2091 {
2092         struct ceph_mds_request *req = kzalloc(sizeof(*req), GFP_NOFS);
2093
2094         if (!req)
2095                 return ERR_PTR(-ENOMEM);
2096
2097         mutex_init(&req->r_fill_mutex);
2098         req->r_mdsc = mdsc;
2099         req->r_started = jiffies;
2100         req->r_resend_mds = -1;
2101         INIT_LIST_HEAD(&req->r_unsafe_dir_item);
2102         INIT_LIST_HEAD(&req->r_unsafe_target_item);
2103         req->r_fmode = -1;
2104         kref_init(&req->r_kref);
2105         RB_CLEAR_NODE(&req->r_node);
2106         INIT_LIST_HEAD(&req->r_wait);
2107         init_completion(&req->r_completion);
2108         init_completion(&req->r_safe_completion);
2109         INIT_LIST_HEAD(&req->r_unsafe_item);
2110
2111         ktime_get_coarse_real_ts64(&req->r_stamp);
2112
2113         req->r_op = op;
2114         req->r_direct_mode = mode;
2115         return req;
2116 }
2117
2118 /*
2119  * return oldest (lowest) request, tid in request tree, 0 if none.
2120  *
2121  * called under mdsc->mutex.
2122  */
2123 static struct ceph_mds_request *__get_oldest_req(struct ceph_mds_client *mdsc)
2124 {
2125         if (RB_EMPTY_ROOT(&mdsc->request_tree))
2126                 return NULL;
2127         return rb_entry(rb_first(&mdsc->request_tree),
2128                         struct ceph_mds_request, r_node);
2129 }
2130
2131 static inline  u64 __get_oldest_tid(struct ceph_mds_client *mdsc)
2132 {
2133         return mdsc->oldest_tid;
2134 }
2135
2136 /*
2137  * Build a dentry's path.  Allocate on heap; caller must kfree.  Based
2138  * on build_path_from_dentry in fs/cifs/dir.c.
2139  *
2140  * If @stop_on_nosnap, generate path relative to the first non-snapped
2141  * inode.
2142  *
2143  * Encode hidden .snap dirs as a double /, i.e.
2144  *   foo/.snap/bar -> foo//bar
2145  */
2146 char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *pbase,
2147                            int stop_on_nosnap)
2148 {
2149         struct dentry *temp;
2150         char *path;
2151         int pos;
2152         unsigned seq;
2153         u64 base;
2154
2155         if (!dentry)
2156                 return ERR_PTR(-EINVAL);
2157
2158         path = __getname();
2159         if (!path)
2160                 return ERR_PTR(-ENOMEM);
2161 retry:
2162         pos = PATH_MAX - 1;
2163         path[pos] = '\0';
2164
2165         seq = read_seqbegin(&rename_lock);
2166         rcu_read_lock();
2167         temp = dentry;
2168         for (;;) {
2169                 struct inode *inode;
2170
2171                 spin_lock(&temp->d_lock);
2172                 inode = d_inode(temp);
2173                 if (inode && ceph_snap(inode) == CEPH_SNAPDIR) {
2174                         dout("build_path path+%d: %p SNAPDIR\n",
2175                              pos, temp);
2176                 } else if (stop_on_nosnap && inode && dentry != temp &&
2177                            ceph_snap(inode) == CEPH_NOSNAP) {
2178                         spin_unlock(&temp->d_lock);
2179                         pos++; /* get rid of any prepended '/' */
2180                         break;
2181                 } else {
2182                         pos -= temp->d_name.len;
2183                         if (pos < 0) {
2184                                 spin_unlock(&temp->d_lock);
2185                                 break;
2186                         }
2187                         memcpy(path + pos, temp->d_name.name, temp->d_name.len);
2188                 }
2189                 spin_unlock(&temp->d_lock);
2190                 temp = READ_ONCE(temp->d_parent);
2191
2192                 /* Are we at the root? */
2193                 if (IS_ROOT(temp))
2194                         break;
2195
2196                 /* Are we out of buffer? */
2197                 if (--pos < 0)
2198                         break;
2199
2200                 path[pos] = '/';
2201         }
2202         base = ceph_ino(d_inode(temp));
2203         rcu_read_unlock();
2204
2205         if (read_seqretry(&rename_lock, seq))
2206                 goto retry;
2207
2208         if (pos < 0) {
2209                 /*
2210                  * A rename didn't occur, but somehow we didn't end up where
2211                  * we thought we would. Throw a warning and try again.
2212                  */
2213                 pr_warn("build_path did not end path lookup where "
2214                         "expected, pos is %d\n", pos);
2215                 goto retry;
2216         }
2217
2218         *pbase = base;
2219         *plen = PATH_MAX - 1 - pos;
2220         dout("build_path on %p %d built %llx '%.*s'\n",
2221              dentry, d_count(dentry), base, *plen, path + pos);
2222         return path + pos;
2223 }
2224
2225 static int build_dentry_path(struct dentry *dentry, struct inode *dir,
2226                              const char **ppath, int *ppathlen, u64 *pino,
2227                              bool *pfreepath, bool parent_locked)
2228 {
2229         char *path;
2230
2231         rcu_read_lock();
2232         if (!dir)
2233                 dir = d_inode_rcu(dentry->d_parent);
2234         if (dir && parent_locked && ceph_snap(dir) == CEPH_NOSNAP) {
2235                 *pino = ceph_ino(dir);
2236                 rcu_read_unlock();
2237                 *ppath = dentry->d_name.name;
2238                 *ppathlen = dentry->d_name.len;
2239                 return 0;
2240         }
2241         rcu_read_unlock();
2242         path = ceph_mdsc_build_path(dentry, ppathlen, pino, 1);
2243         if (IS_ERR(path))
2244                 return PTR_ERR(path);
2245         *ppath = path;
2246         *pfreepath = true;
2247         return 0;
2248 }
2249
2250 static int build_inode_path(struct inode *inode,
2251                             const char **ppath, int *ppathlen, u64 *pino,
2252                             bool *pfreepath)
2253 {
2254         struct dentry *dentry;
2255         char *path;
2256
2257         if (ceph_snap(inode) == CEPH_NOSNAP) {
2258                 *pino = ceph_ino(inode);
2259                 *ppathlen = 0;
2260                 return 0;
2261         }
2262         dentry = d_find_alias(inode);
2263         path = ceph_mdsc_build_path(dentry, ppathlen, pino, 1);
2264         dput(dentry);
2265         if (IS_ERR(path))
2266                 return PTR_ERR(path);
2267         *ppath = path;
2268         *pfreepath = true;
2269         return 0;
2270 }
2271
2272 /*
2273  * request arguments may be specified via an inode *, a dentry *, or
2274  * an explicit ino+path.
2275  */
2276 static int set_request_path_attr(struct inode *rinode, struct dentry *rdentry,
2277                                   struct inode *rdiri, const char *rpath,
2278                                   u64 rino, const char **ppath, int *pathlen,
2279                                   u64 *ino, bool *freepath, bool parent_locked)
2280 {
2281         int r = 0;
2282
2283         if (rinode) {
2284                 r = build_inode_path(rinode, ppath, pathlen, ino, freepath);
2285                 dout(" inode %p %llx.%llx\n", rinode, ceph_ino(rinode),
2286                      ceph_snap(rinode));
2287         } else if (rdentry) {
2288                 r = build_dentry_path(rdentry, rdiri, ppath, pathlen, ino,
2289                                         freepath, parent_locked);
2290                 dout(" dentry %p %llx/%.*s\n", rdentry, *ino, *pathlen,
2291                      *ppath);
2292         } else if (rpath || rino) {
2293                 *ino = rino;
2294                 *ppath = rpath;
2295                 *pathlen = rpath ? strlen(rpath) : 0;
2296                 dout(" path %.*s\n", *pathlen, rpath);
2297         }
2298
2299         return r;
2300 }
2301
2302 /*
2303  * called under mdsc->mutex
2304  */
2305 static struct ceph_msg *create_request_message(struct ceph_mds_client *mdsc,
2306                                                struct ceph_mds_request *req,
2307                                                int mds, bool drop_cap_releases)
2308 {
2309         struct ceph_msg *msg;
2310         struct ceph_mds_request_head *head;
2311         const char *path1 = NULL;
2312         const char *path2 = NULL;
2313         u64 ino1 = 0, ino2 = 0;
2314         int pathlen1 = 0, pathlen2 = 0;
2315         bool freepath1 = false, freepath2 = false;
2316         int len;
2317         u16 releases;
2318         void *p, *end;
2319         int ret;
2320
2321         ret = set_request_path_attr(req->r_inode, req->r_dentry,
2322                               req->r_parent, req->r_path1, req->r_ino1.ino,
2323                               &path1, &pathlen1, &ino1, &freepath1,
2324                               test_bit(CEPH_MDS_R_PARENT_LOCKED,
2325                                         &req->r_req_flags));
2326         if (ret < 0) {
2327                 msg = ERR_PTR(ret);
2328                 goto out;
2329         }
2330
2331         /* If r_old_dentry is set, then assume that its parent is locked */
2332         ret = set_request_path_attr(NULL, req->r_old_dentry,
2333                               req->r_old_dentry_dir,
2334                               req->r_path2, req->r_ino2.ino,
2335                               &path2, &pathlen2, &ino2, &freepath2, true);
2336         if (ret < 0) {
2337                 msg = ERR_PTR(ret);
2338                 goto out_free1;
2339         }
2340
2341         len = sizeof(*head) +
2342                 pathlen1 + pathlen2 + 2*(1 + sizeof(u32) + sizeof(u64)) +
2343                 sizeof(struct ceph_timespec);
2344
2345         /* calculate (max) length for cap releases */
2346         len += sizeof(struct ceph_mds_request_release) *
2347                 (!!req->r_inode_drop + !!req->r_dentry_drop +
2348                  !!req->r_old_inode_drop + !!req->r_old_dentry_drop);
2349         if (req->r_dentry_drop)
2350                 len += pathlen1;
2351         if (req->r_old_dentry_drop)
2352                 len += pathlen2;
2353
2354         msg = ceph_msg_new2(CEPH_MSG_CLIENT_REQUEST, len, 1, GFP_NOFS, false);
2355         if (!msg) {
2356                 msg = ERR_PTR(-ENOMEM);
2357                 goto out_free2;
2358         }
2359
2360         msg->hdr.version = cpu_to_le16(2);
2361         msg->hdr.tid = cpu_to_le64(req->r_tid);
2362
2363         head = msg->front.iov_base;
2364         p = msg->front.iov_base + sizeof(*head);
2365         end = msg->front.iov_base + msg->front.iov_len;
2366
2367         head->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
2368         head->op = cpu_to_le32(req->r_op);
2369         head->caller_uid = cpu_to_le32(from_kuid(&init_user_ns, req->r_uid));
2370         head->caller_gid = cpu_to_le32(from_kgid(&init_user_ns, req->r_gid));
2371         head->ino = 0;
2372         head->args = req->r_args;
2373
2374         ceph_encode_filepath(&p, end, ino1, path1);
2375         ceph_encode_filepath(&p, end, ino2, path2);
2376
2377         /* make note of release offset, in case we need to replay */
2378         req->r_request_release_offset = p - msg->front.iov_base;
2379
2380         /* cap releases */
2381         releases = 0;
2382         if (req->r_inode_drop)
2383                 releases += ceph_encode_inode_release(&p,
2384                       req->r_inode ? req->r_inode : d_inode(req->r_dentry),
2385                       mds, req->r_inode_drop, req->r_inode_unless, 0);
2386         if (req->r_dentry_drop)
2387                 releases += ceph_encode_dentry_release(&p, req->r_dentry,
2388                                 req->r_parent, mds, req->r_dentry_drop,
2389                                 req->r_dentry_unless);
2390         if (req->r_old_dentry_drop)
2391                 releases += ceph_encode_dentry_release(&p, req->r_old_dentry,
2392                                 req->r_old_dentry_dir, mds,
2393                                 req->r_old_dentry_drop,
2394                                 req->r_old_dentry_unless);
2395         if (req->r_old_inode_drop)
2396                 releases += ceph_encode_inode_release(&p,
2397                       d_inode(req->r_old_dentry),
2398                       mds, req->r_old_inode_drop, req->r_old_inode_unless, 0);
2399
2400         if (drop_cap_releases) {
2401                 releases = 0;
2402                 p = msg->front.iov_base + req->r_request_release_offset;
2403         }
2404
2405         head->num_releases = cpu_to_le16(releases);
2406
2407         /* time stamp */
2408         {
2409                 struct ceph_timespec ts;
2410                 ceph_encode_timespec64(&ts, &req->r_stamp);
2411                 ceph_encode_copy(&p, &ts, sizeof(ts));
2412         }
2413
2414         BUG_ON(p > end);
2415         msg->front.iov_len = p - msg->front.iov_base;
2416         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
2417
2418         if (req->r_pagelist) {
2419                 struct ceph_pagelist *pagelist = req->r_pagelist;
2420                 ceph_msg_data_add_pagelist(msg, pagelist);
2421                 msg->hdr.data_len = cpu_to_le32(pagelist->length);
2422         } else {
2423                 msg->hdr.data_len = 0;
2424         }
2425
2426         msg->hdr.data_off = cpu_to_le16(0);
2427
2428 out_free2:
2429         if (freepath2)
2430                 ceph_mdsc_free_path((char *)path2, pathlen2);
2431 out_free1:
2432         if (freepath1)
2433                 ceph_mdsc_free_path((char *)path1, pathlen1);
2434 out:
2435         return msg;
2436 }
2437
2438 /*
2439  * called under mdsc->mutex if error, under no mutex if
2440  * success.
2441  */
2442 static void complete_request(struct ceph_mds_client *mdsc,
2443                              struct ceph_mds_request *req)
2444 {
2445         if (req->r_callback)
2446                 req->r_callback(mdsc, req);
2447         complete_all(&req->r_completion);
2448 }
2449
2450 /*
2451  * called under mdsc->mutex
2452  */
2453 static int __prepare_send_request(struct ceph_mds_client *mdsc,
2454                                   struct ceph_mds_request *req,
2455                                   int mds, bool drop_cap_releases)
2456 {
2457         struct ceph_mds_request_head *rhead;
2458         struct ceph_msg *msg;
2459         int flags = 0;
2460
2461         req->r_attempts++;
2462         if (req->r_inode) {
2463                 struct ceph_cap *cap =
2464                         ceph_get_cap_for_mds(ceph_inode(req->r_inode), mds);
2465
2466                 if (cap)
2467                         req->r_sent_on_mseq = cap->mseq;
2468                 else
2469                         req->r_sent_on_mseq = -1;
2470         }
2471         dout("prepare_send_request %p tid %lld %s (attempt %d)\n", req,
2472              req->r_tid, ceph_mds_op_name(req->r_op), req->r_attempts);
2473
2474         if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
2475                 void *p;
2476                 /*
2477                  * Replay.  Do not regenerate message (and rebuild
2478                  * paths, etc.); just use the original message.
2479                  * Rebuilding paths will break for renames because
2480                  * d_move mangles the src name.
2481                  */
2482                 msg = req->r_request;
2483                 rhead = msg->front.iov_base;
2484
2485                 flags = le32_to_cpu(rhead->flags);
2486                 flags |= CEPH_MDS_FLAG_REPLAY;
2487                 rhead->flags = cpu_to_le32(flags);
2488
2489                 if (req->r_target_inode)
2490                         rhead->ino = cpu_to_le64(ceph_ino(req->r_target_inode));
2491
2492                 rhead->num_retry = req->r_attempts - 1;
2493
2494                 /* remove cap/dentry releases from message */
2495                 rhead->num_releases = 0;
2496
2497                 /* time stamp */
2498                 p = msg->front.iov_base + req->r_request_release_offset;
2499                 {
2500                         struct ceph_timespec ts;
2501                         ceph_encode_timespec64(&ts, &req->r_stamp);
2502                         ceph_encode_copy(&p, &ts, sizeof(ts));
2503                 }
2504
2505                 msg->front.iov_len = p - msg->front.iov_base;
2506                 msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
2507                 return 0;
2508         }
2509
2510         if (req->r_request) {
2511                 ceph_msg_put(req->r_request);
2512                 req->r_request = NULL;
2513         }
2514         msg = create_request_message(mdsc, req, mds, drop_cap_releases);
2515         if (IS_ERR(msg)) {
2516                 req->r_err = PTR_ERR(msg);
2517                 return PTR_ERR(msg);
2518         }
2519         req->r_request = msg;
2520
2521         rhead = msg->front.iov_base;
2522         rhead->oldest_client_tid = cpu_to_le64(__get_oldest_tid(mdsc));
2523         if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
2524                 flags |= CEPH_MDS_FLAG_REPLAY;
2525         if (req->r_parent)
2526                 flags |= CEPH_MDS_FLAG_WANT_DENTRY;
2527         rhead->flags = cpu_to_le32(flags);
2528         rhead->num_fwd = req->r_num_fwd;
2529         rhead->num_retry = req->r_attempts - 1;
2530         rhead->ino = 0;
2531
2532         dout(" r_parent = %p\n", req->r_parent);
2533         return 0;
2534 }
2535
2536 /*
2537  * called under mdsc->mutex
2538  */
2539 static int __send_request(struct ceph_mds_client *mdsc,
2540                           struct ceph_mds_session *session,
2541                           struct ceph_mds_request *req,
2542                           bool drop_cap_releases)
2543 {
2544         int err;
2545
2546         err = __prepare_send_request(mdsc, req, session->s_mds,
2547                                      drop_cap_releases);
2548         if (!err) {
2549                 ceph_msg_get(req->r_request);
2550                 ceph_con_send(&session->s_con, req->r_request);
2551         }
2552
2553         return err;
2554 }
2555
2556 /*
2557  * send request, or put it on the appropriate wait list.
2558  */
2559 static void __do_request(struct ceph_mds_client *mdsc,
2560                         struct ceph_mds_request *req)
2561 {
2562         struct ceph_mds_session *session = NULL;
2563         int mds = -1;
2564         int err = 0;
2565         bool random;
2566
2567         if (req->r_err || test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags)) {
2568                 if (test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags))
2569                         __unregister_request(mdsc, req);
2570                 return;
2571         }
2572
2573         if (req->r_timeout &&
2574             time_after_eq(jiffies, req->r_started + req->r_timeout)) {
2575                 dout("do_request timed out\n");
2576                 err = -EIO;
2577                 goto finish;
2578         }
2579         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
2580                 dout("do_request forced umount\n");
2581                 err = -EIO;
2582                 goto finish;
2583         }
2584         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_MOUNTING) {
2585                 if (mdsc->mdsmap_err) {
2586                         err = mdsc->mdsmap_err;
2587                         dout("do_request mdsmap err %d\n", err);
2588                         goto finish;
2589                 }
2590                 if (mdsc->mdsmap->m_epoch == 0) {
2591                         dout("do_request no mdsmap, waiting for map\n");
2592                         list_add(&req->r_wait, &mdsc->waiting_for_map);
2593                         return;
2594                 }
2595                 if (!(mdsc->fsc->mount_options->flags &
2596                       CEPH_MOUNT_OPT_MOUNTWAIT) &&
2597                     !ceph_mdsmap_is_cluster_available(mdsc->mdsmap)) {
2598                         err = -EHOSTUNREACH;
2599                         goto finish;
2600                 }
2601         }
2602
2603         put_request_session(req);
2604
2605         mds = __choose_mds(mdsc, req, &random);
2606         if (mds < 0 ||
2607             ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) {
2608                 dout("do_request no mds or not active, waiting for map\n");
2609                 list_add(&req->r_wait, &mdsc->waiting_for_map);
2610                 return;
2611         }
2612
2613         /* get, open session */
2614         session = __ceph_lookup_mds_session(mdsc, mds);
2615         if (!session) {
2616                 session = register_session(mdsc, mds);
2617                 if (IS_ERR(session)) {
2618                         err = PTR_ERR(session);
2619                         goto finish;
2620                 }
2621         }
2622         req->r_session = ceph_get_mds_session(session);
2623
2624         dout("do_request mds%d session %p state %s\n", mds, session,
2625              ceph_session_state_name(session->s_state));
2626         if (session->s_state != CEPH_MDS_SESSION_OPEN &&
2627             session->s_state != CEPH_MDS_SESSION_HUNG) {
2628                 if (session->s_state == CEPH_MDS_SESSION_REJECTED) {
2629                         err = -EACCES;
2630                         goto out_session;
2631                 }
2632                 if (session->s_state == CEPH_MDS_SESSION_NEW ||
2633                     session->s_state == CEPH_MDS_SESSION_CLOSING) {
2634                         __open_session(mdsc, session);
2635                         /* retry the same mds later */
2636                         if (random)
2637                                 req->r_resend_mds = mds;
2638                 }
2639                 list_add(&req->r_wait, &session->s_waiting);
2640                 goto out_session;
2641         }
2642
2643         /* send request */
2644         req->r_resend_mds = -1;   /* forget any previous mds hint */
2645
2646         if (req->r_request_started == 0)   /* note request start time */
2647                 req->r_request_started = jiffies;
2648
2649         err = __send_request(mdsc, session, req, false);
2650
2651 out_session:
2652         ceph_put_mds_session(session);
2653 finish:
2654         if (err) {
2655                 dout("__do_request early error %d\n", err);
2656                 req->r_err = err;
2657                 complete_request(mdsc, req);
2658                 __unregister_request(mdsc, req);
2659         }
2660         return;
2661 }
2662
2663 /*
2664  * called under mdsc->mutex
2665  */
2666 static void __wake_requests(struct ceph_mds_client *mdsc,
2667                             struct list_head *head)
2668 {
2669         struct ceph_mds_request *req;
2670         LIST_HEAD(tmp_list);
2671
2672         list_splice_init(head, &tmp_list);
2673
2674         while (!list_empty(&tmp_list)) {
2675                 req = list_entry(tmp_list.next,
2676                                  struct ceph_mds_request, r_wait);
2677                 list_del_init(&req->r_wait);
2678                 dout(" wake request %p tid %llu\n", req, req->r_tid);
2679                 __do_request(mdsc, req);
2680         }
2681 }
2682
2683 /*
2684  * Wake up threads with requests pending for @mds, so that they can
2685  * resubmit their requests to a possibly different mds.
2686  */
2687 static void kick_requests(struct ceph_mds_client *mdsc, int mds)
2688 {
2689         struct ceph_mds_request *req;
2690         struct rb_node *p = rb_first(&mdsc->request_tree);
2691
2692         dout("kick_requests mds%d\n", mds);
2693         while (p) {
2694                 req = rb_entry(p, struct ceph_mds_request, r_node);
2695                 p = rb_next(p);
2696                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
2697                         continue;
2698                 if (req->r_attempts > 0)
2699                         continue; /* only new requests */
2700                 if (req->r_session &&
2701                     req->r_session->s_mds == mds) {
2702                         dout(" kicking tid %llu\n", req->r_tid);
2703                         list_del_init(&req->r_wait);
2704                         __do_request(mdsc, req);
2705                 }
2706         }
2707 }
2708
2709 int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, struct inode *dir,
2710                               struct ceph_mds_request *req)
2711 {
2712         int err;
2713
2714         /* take CAP_PIN refs for r_inode, r_parent, r_old_dentry */
2715         if (req->r_inode)
2716                 ceph_get_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
2717         if (req->r_parent) {
2718                 ceph_get_cap_refs(ceph_inode(req->r_parent), CEPH_CAP_PIN);
2719                 ihold(req->r_parent);
2720         }
2721         if (req->r_old_dentry_dir)
2722                 ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir),
2723                                   CEPH_CAP_PIN);
2724
2725         dout("submit_request on %p for inode %p\n", req, dir);
2726         mutex_lock(&mdsc->mutex);
2727         __register_request(mdsc, req, dir);
2728         __do_request(mdsc, req);
2729         err = req->r_err;
2730         mutex_unlock(&mdsc->mutex);
2731         return err;
2732 }
2733
2734 static int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc,
2735                                   struct ceph_mds_request *req)
2736 {
2737         int err;
2738
2739         /* wait */
2740         dout("do_request waiting\n");
2741         if (!req->r_timeout && req->r_wait_for_completion) {
2742                 err = req->r_wait_for_completion(mdsc, req);
2743         } else {
2744                 long timeleft = wait_for_completion_killable_timeout(
2745                                         &req->r_completion,
2746                                         ceph_timeout_jiffies(req->r_timeout));
2747                 if (timeleft > 0)
2748                         err = 0;
2749                 else if (!timeleft)
2750                         err = -EIO;  /* timed out */
2751                 else
2752                         err = timeleft;  /* killed */
2753         }
2754         dout("do_request waited, got %d\n", err);
2755         mutex_lock(&mdsc->mutex);
2756
2757         /* only abort if we didn't race with a real reply */
2758         if (test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags)) {
2759                 err = le32_to_cpu(req->r_reply_info.head->result);
2760         } else if (err < 0) {
2761                 dout("aborted request %lld with %d\n", req->r_tid, err);
2762
2763                 /*
2764                  * ensure we aren't running concurrently with
2765                  * ceph_fill_trace or ceph_readdir_prepopulate, which
2766                  * rely on locks (dir mutex) held by our caller.
2767                  */
2768                 mutex_lock(&req->r_fill_mutex);
2769                 req->r_err = err;
2770                 set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
2771                 mutex_unlock(&req->r_fill_mutex);
2772
2773                 if (req->r_parent &&
2774                     (req->r_op & CEPH_MDS_OP_WRITE))
2775                         ceph_invalidate_dir_request(req);
2776         } else {
2777                 err = req->r_err;
2778         }
2779
2780         mutex_unlock(&mdsc->mutex);
2781         return err;
2782 }
2783
2784 /*
2785  * Synchrously perform an mds request.  Take care of all of the
2786  * session setup, forwarding, retry details.
2787  */
2788 int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
2789                          struct inode *dir,
2790                          struct ceph_mds_request *req)
2791 {
2792         int err;
2793
2794         dout("do_request on %p\n", req);
2795
2796         /* issue */
2797         err = ceph_mdsc_submit_request(mdsc, dir, req);
2798         if (!err)
2799                 err = ceph_mdsc_wait_request(mdsc, req);
2800         dout("do_request %p done, result %d\n", req, err);
2801         return err;
2802 }
2803
2804 /*
2805  * Invalidate dir's completeness, dentry lease state on an aborted MDS
2806  * namespace request.
2807  */
2808 void ceph_invalidate_dir_request(struct ceph_mds_request *req)
2809 {
2810         struct inode *dir = req->r_parent;
2811         struct inode *old_dir = req->r_old_dentry_dir;
2812
2813         dout("invalidate_dir_request %p %p (complete, lease(s))\n", dir, old_dir);
2814
2815         ceph_dir_clear_complete(dir);
2816         if (old_dir)
2817                 ceph_dir_clear_complete(old_dir);
2818         if (req->r_dentry)
2819                 ceph_invalidate_dentry_lease(req->r_dentry);
2820         if (req->r_old_dentry)
2821                 ceph_invalidate_dentry_lease(req->r_old_dentry);
2822 }
2823
2824 /*
2825  * Handle mds reply.
2826  *
2827  * We take the session mutex and parse and process the reply immediately.
2828  * This preserves the logical ordering of replies, capabilities, etc., sent
2829  * by the MDS as they are applied to our local cache.
2830  */
2831 static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
2832 {
2833         struct ceph_mds_client *mdsc = session->s_mdsc;
2834         struct ceph_mds_request *req;
2835         struct ceph_mds_reply_head *head = msg->front.iov_base;
2836         struct ceph_mds_reply_info_parsed *rinfo;  /* parsed reply info */
2837         struct ceph_snap_realm *realm;
2838         u64 tid;
2839         int err, result;
2840         int mds = session->s_mds;
2841
2842         if (msg->front.iov_len < sizeof(*head)) {
2843                 pr_err("mdsc_handle_reply got corrupt (short) reply\n");
2844                 ceph_msg_dump(msg);
2845                 return;
2846         }
2847
2848         /* get request, session */
2849         tid = le64_to_cpu(msg->hdr.tid);
2850         mutex_lock(&mdsc->mutex);
2851         req = lookup_get_request(mdsc, tid);
2852         if (!req) {
2853                 dout("handle_reply on unknown tid %llu\n", tid);
2854                 mutex_unlock(&mdsc->mutex);
2855                 return;
2856         }
2857         dout("handle_reply %p\n", req);
2858
2859         /* correct session? */
2860         if (req->r_session != session) {
2861                 pr_err("mdsc_handle_reply got %llu on session mds%d"
2862                        " not mds%d\n", tid, session->s_mds,
2863                        req->r_session ? req->r_session->s_mds : -1);
2864                 mutex_unlock(&mdsc->mutex);
2865                 goto out;
2866         }
2867
2868         /* dup? */
2869         if ((test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags) && !head->safe) ||
2870             (test_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags) && head->safe)) {
2871                 pr_warn("got a dup %s reply on %llu from mds%d\n",
2872                            head->safe ? "safe" : "unsafe", tid, mds);
2873                 mutex_unlock(&mdsc->mutex);
2874                 goto out;
2875         }
2876         if (test_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags)) {
2877                 pr_warn("got unsafe after safe on %llu from mds%d\n",
2878                            tid, mds);
2879                 mutex_unlock(&mdsc->mutex);
2880                 goto out;
2881         }
2882
2883         result = le32_to_cpu(head->result);
2884
2885         /*
2886          * Handle an ESTALE
2887          * if we're not talking to the authority, send to them
2888          * if the authority has changed while we weren't looking,
2889          * send to new authority
2890          * Otherwise we just have to return an ESTALE
2891          */
2892         if (result == -ESTALE) {
2893                 dout("got ESTALE on request %llu\n", req->r_tid);
2894                 req->r_resend_mds = -1;
2895                 if (req->r_direct_mode != USE_AUTH_MDS) {
2896                         dout("not using auth, setting for that now\n");
2897                         req->r_direct_mode = USE_AUTH_MDS;
2898                         __do_request(mdsc, req);
2899                         mutex_unlock(&mdsc->mutex);
2900                         goto out;
2901                 } else  {
2902                         int mds = __choose_mds(mdsc, req, NULL);
2903                         if (mds >= 0 && mds != req->r_session->s_mds) {
2904                                 dout("but auth changed, so resending\n");
2905                                 __do_request(mdsc, req);
2906                                 mutex_unlock(&mdsc->mutex);
2907                                 goto out;
2908                         }
2909                 }
2910                 dout("have to return ESTALE on request %llu\n", req->r_tid);
2911         }
2912
2913
2914         if (head->safe) {
2915                 set_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags);
2916                 __unregister_request(mdsc, req);
2917
2918                 /* last request during umount? */
2919                 if (mdsc->stopping && !__get_oldest_req(mdsc))
2920                         complete_all(&mdsc->safe_umount_waiters);
2921
2922                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
2923                         /*
2924                          * We already handled the unsafe response, now do the
2925                          * cleanup.  No need to examine the response; the MDS
2926                          * doesn't include any result info in the safe
2927                          * response.  And even if it did, there is nothing
2928                          * useful we could do with a revised return value.
2929                          */
2930                         dout("got safe reply %llu, mds%d\n", tid, mds);
2931
2932                         mutex_unlock(&mdsc->mutex);
2933                         goto out;
2934                 }
2935         } else {
2936                 set_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags);
2937                 list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe);
2938                 if (req->r_unsafe_dir) {
2939                         struct ceph_inode_info *ci =
2940                                         ceph_inode(req->r_unsafe_dir);
2941                         spin_lock(&ci->i_unsafe_lock);
2942                         list_add_tail(&req->r_unsafe_dir_item,
2943                                       &ci->i_unsafe_dirops);
2944                         spin_unlock(&ci->i_unsafe_lock);
2945                 }
2946         }
2947
2948         dout("handle_reply tid %lld result %d\n", tid, result);
2949         rinfo = &req->r_reply_info;
2950         if (test_bit(CEPHFS_FEATURE_REPLY_ENCODING, &session->s_features))
2951                 err = parse_reply_info(msg, rinfo, (u64)-1);
2952         else
2953                 err = parse_reply_info(msg, rinfo, session->s_con.peer_features);
2954         mutex_unlock(&mdsc->mutex);
2955
2956         mutex_lock(&session->s_mutex);
2957         if (err < 0) {
2958                 pr_err("mdsc_handle_reply got corrupt reply mds%d(tid:%lld)\n", mds, tid);
2959                 ceph_msg_dump(msg);
2960                 goto out_err;
2961         }
2962
2963         /* snap trace */
2964         realm = NULL;
2965         if (rinfo->snapblob_len) {
2966                 down_write(&mdsc->snap_rwsem);
2967                 ceph_update_snap_trace(mdsc, rinfo->snapblob,
2968                                 rinfo->snapblob + rinfo->snapblob_len,
2969                                 le32_to_cpu(head->op) == CEPH_MDS_OP_RMSNAP,
2970                                 &realm);
2971                 downgrade_write(&mdsc->snap_rwsem);
2972         } else {
2973                 down_read(&mdsc->snap_rwsem);
2974         }
2975
2976         /* insert trace into our cache */
2977         mutex_lock(&req->r_fill_mutex);
2978         current->journal_info = req;
2979         err = ceph_fill_trace(mdsc->fsc->sb, req);
2980         if (err == 0) {
2981                 if (result == 0 && (req->r_op == CEPH_MDS_OP_READDIR ||
2982                                     req->r_op == CEPH_MDS_OP_LSSNAP))
2983                         ceph_readdir_prepopulate(req, req->r_session);
2984         }
2985         current->journal_info = NULL;
2986         mutex_unlock(&req->r_fill_mutex);
2987
2988         up_read(&mdsc->snap_rwsem);
2989         if (realm)
2990                 ceph_put_snap_realm(mdsc, realm);
2991
2992         if (err == 0) {
2993                 if (req->r_target_inode &&
2994                     test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
2995                         struct ceph_inode_info *ci =
2996                                 ceph_inode(req->r_target_inode);
2997                         spin_lock(&ci->i_unsafe_lock);
2998                         list_add_tail(&req->r_unsafe_target_item,
2999                                       &ci->i_unsafe_iops);
3000                         spin_unlock(&ci->i_unsafe_lock);
3001                 }
3002
3003                 ceph_unreserve_caps(mdsc, &req->r_caps_reservation);
3004         }
3005 out_err:
3006         mutex_lock(&mdsc->mutex);
3007         if (!test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) {
3008                 if (err) {
3009                         req->r_err = err;
3010                 } else {
3011                         req->r_reply =  ceph_msg_get(msg);
3012                         set_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags);
3013                 }
3014         } else {
3015                 dout("reply arrived after request %lld was aborted\n", tid);
3016         }
3017         mutex_unlock(&mdsc->mutex);
3018
3019         mutex_unlock(&session->s_mutex);
3020
3021         /* kick calling process */
3022         complete_request(mdsc, req);
3023 out:
3024         ceph_mdsc_put_request(req);
3025         return;
3026 }
3027
3028
3029
3030 /*
3031  * handle mds notification that our request has been forwarded.
3032  */
3033 static void handle_forward(struct ceph_mds_client *mdsc,
3034                            struct ceph_mds_session *session,
3035                            struct ceph_msg *msg)
3036 {
3037         struct ceph_mds_request *req;
3038         u64 tid = le64_to_cpu(msg->hdr.tid);
3039         u32 next_mds;
3040         u32 fwd_seq;
3041         int err = -EINVAL;
3042         void *p = msg->front.iov_base;
3043         void *end = p + msg->front.iov_len;
3044
3045         ceph_decode_need(&p, end, 2*sizeof(u32), bad);
3046         next_mds = ceph_decode_32(&p);
3047         fwd_seq = ceph_decode_32(&p);
3048
3049         mutex_lock(&mdsc->mutex);
3050         req = lookup_get_request(mdsc, tid);
3051         if (!req) {
3052                 dout("forward tid %llu to mds%d - req dne\n", tid, next_mds);
3053                 goto out;  /* dup reply? */
3054         }
3055
3056         if (test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) {
3057                 dout("forward tid %llu aborted, unregistering\n", tid);
3058                 __unregister_request(mdsc, req);
3059         } else if (fwd_seq <= req->r_num_fwd) {
3060                 dout("forward tid %llu to mds%d - old seq %d <= %d\n",
3061                      tid, next_mds, req->r_num_fwd, fwd_seq);
3062         } else {
3063                 /* resend. forward race not possible; mds would drop */
3064                 dout("forward tid %llu to mds%d (we resend)\n", tid, next_mds);
3065                 BUG_ON(req->r_err);
3066                 BUG_ON(test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags));
3067                 req->r_attempts = 0;
3068                 req->r_num_fwd = fwd_seq;
3069                 req->r_resend_mds = next_mds;
3070                 put_request_session(req);
3071                 __do_request(mdsc, req);
3072         }
3073         ceph_mdsc_put_request(req);
3074 out:
3075         mutex_unlock(&mdsc->mutex);
3076         return;
3077
3078 bad:
3079         pr_err("mdsc_handle_forward decode error err=%d\n", err);
3080 }
3081
3082 static int __decode_session_metadata(void **p, void *end,
3083                                      bool *blacklisted)
3084 {
3085         /* map<string,string> */
3086         u32 n;
3087         bool err_str;
3088         ceph_decode_32_safe(p, end, n, bad);
3089         while (n-- > 0) {
3090                 u32 len;
3091                 ceph_decode_32_safe(p, end, len, bad);
3092                 ceph_decode_need(p, end, len, bad);
3093                 err_str = !strncmp(*p, "error_string", len);
3094                 *p += len;
3095                 ceph_decode_32_safe(p, end, len, bad);
3096                 ceph_decode_need(p, end, len, bad);
3097                 if (err_str && strnstr(*p, "blacklisted", len))
3098                         *blacklisted = true;
3099                 *p += len;
3100         }
3101         return 0;
3102 bad:
3103         return -1;
3104 }
3105
3106 /*
3107  * handle a mds session control message
3108  */
3109 static void handle_session(struct ceph_mds_session *session,
3110                            struct ceph_msg *msg)
3111 {
3112         struct ceph_mds_client *mdsc = session->s_mdsc;
3113         int mds = session->s_mds;
3114         int msg_version = le16_to_cpu(msg->hdr.version);
3115         void *p = msg->front.iov_base;
3116         void *end = p + msg->front.iov_len;
3117         struct ceph_mds_session_head *h;
3118         u32 op;
3119         u64 seq;
3120         unsigned long features = 0;
3121         int wake = 0;
3122         bool blacklisted = false;
3123
3124         /* decode */
3125         ceph_decode_need(&p, end, sizeof(*h), bad);
3126         h = p;
3127         p += sizeof(*h);
3128
3129         op = le32_to_cpu(h->op);
3130         seq = le64_to_cpu(h->seq);
3131
3132         if (msg_version >= 3) {
3133                 u32 len;
3134                 /* version >= 2, metadata */
3135                 if (__decode_session_metadata(&p, end, &blacklisted) < 0)
3136                         goto bad;
3137                 /* version >= 3, feature bits */
3138                 ceph_decode_32_safe(&p, end, len, bad);
3139                 ceph_decode_need(&p, end, len, bad);
3140                 memcpy(&features, p, min_t(size_t, len, sizeof(features)));
3141                 p += len;
3142         }
3143
3144         mutex_lock(&mdsc->mutex);
3145         if (op == CEPH_SESSION_CLOSE) {
3146                 ceph_get_mds_session(session);
3147                 __unregister_session(mdsc, session);
3148         }
3149         /* FIXME: this ttl calculation is generous */
3150         session->s_ttl = jiffies + HZ*mdsc->mdsmap->m_session_autoclose;
3151         mutex_unlock(&mdsc->mutex);
3152
3153         mutex_lock(&session->s_mutex);
3154
3155         dout("handle_session mds%d %s %p state %s seq %llu\n",
3156              mds, ceph_session_op_name(op), session,
3157              ceph_session_state_name(session->s_state), seq);
3158
3159         if (session->s_state == CEPH_MDS_SESSION_HUNG) {
3160                 session->s_state = CEPH_MDS_SESSION_OPEN;
3161                 pr_info("mds%d came back\n", session->s_mds);
3162         }
3163
3164         switch (op) {
3165         case CEPH_SESSION_OPEN:
3166                 if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
3167                         pr_info("mds%d reconnect success\n", session->s_mds);
3168                 session->s_state = CEPH_MDS_SESSION_OPEN;
3169                 session->s_features = features;
3170                 renewed_caps(mdsc, session, 0);
3171                 wake = 1;
3172                 if (mdsc->stopping)
3173                         __close_session(mdsc, session);
3174                 break;
3175
3176         case CEPH_SESSION_RENEWCAPS:
3177                 if (session->s_renew_seq == seq)
3178                         renewed_caps(mdsc, session, 1);
3179                 break;
3180
3181         case CEPH_SESSION_CLOSE:
3182                 if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
3183                         pr_info("mds%d reconnect denied\n", session->s_mds);
3184                 session->s_state = CEPH_MDS_SESSION_CLOSED;
3185                 cleanup_session_requests(mdsc, session);
3186                 remove_session_caps(session);
3187                 wake = 2; /* for good measure */
3188                 wake_up_all(&mdsc->session_close_wq);
3189                 break;
3190
3191         case CEPH_SESSION_STALE:
3192                 pr_info("mds%d caps went stale, renewing\n",
3193                         session->s_mds);
3194                 spin_lock(&session->s_gen_ttl_lock);
3195                 session->s_cap_gen++;
3196                 session->s_cap_ttl = jiffies - 1;
3197                 spin_unlock(&session->s_gen_ttl_lock);
3198                 send_renew_caps(mdsc, session);
3199                 break;
3200
3201         case CEPH_SESSION_RECALL_STATE:
3202                 ceph_trim_caps(mdsc, session, le32_to_cpu(h->max_caps));
3203                 break;
3204
3205         case CEPH_SESSION_FLUSHMSG:
3206                 send_flushmsg_ack(mdsc, session, seq);
3207                 break;
3208
3209         case CEPH_SESSION_FORCE_RO:
3210                 dout("force_session_readonly %p\n", session);
3211                 spin_lock(&session->s_cap_lock);
3212                 session->s_readonly = true;
3213                 spin_unlock(&session->s_cap_lock);
3214                 wake_up_session_caps(session, FORCE_RO);
3215                 break;
3216
3217         case CEPH_SESSION_REJECT:
3218                 WARN_ON(session->s_state != CEPH_MDS_SESSION_OPENING);
3219                 pr_info("mds%d rejected session\n", session->s_mds);
3220                 session->s_state = CEPH_MDS_SESSION_REJECTED;
3221                 cleanup_session_requests(mdsc, session);
3222                 remove_session_caps(session);
3223                 if (blacklisted)
3224                         mdsc->fsc->blacklisted = true;
3225                 wake = 2; /* for good measure */
3226                 break;
3227
3228         default:
3229                 pr_err("mdsc_handle_session bad op %d mds%d\n", op, mds);
3230                 WARN_ON(1);
3231         }
3232
3233         mutex_unlock(&session->s_mutex);
3234         if (wake) {
3235                 mutex_lock(&mdsc->mutex);
3236                 __wake_requests(mdsc, &session->s_waiting);
3237                 if (wake == 2)
3238                         kick_requests(mdsc, mds);
3239                 mutex_unlock(&mdsc->mutex);
3240         }
3241         if (op == CEPH_SESSION_CLOSE)
3242                 ceph_put_mds_session(session);
3243         return;
3244
3245 bad:
3246         pr_err("mdsc_handle_session corrupt message mds%d len %d\n", mds,
3247                (int)msg->front.iov_len);
3248         ceph_msg_dump(msg);
3249         return;
3250 }
3251
3252 /*
3253  * called under session->mutex.
3254  */
3255 static void replay_unsafe_requests(struct ceph_mds_client *mdsc,
3256                                    struct ceph_mds_session *session)
3257 {
3258         struct ceph_mds_request *req, *nreq;
3259         struct rb_node *p;
3260
3261         dout("replay_unsafe_requests mds%d\n", session->s_mds);
3262
3263         mutex_lock(&mdsc->mutex);
3264         list_for_each_entry_safe(req, nreq, &session->s_unsafe, r_unsafe_item)
3265                 __send_request(mdsc, session, req, true);
3266
3267         /*
3268          * also re-send old requests when MDS enters reconnect stage. So that MDS
3269          * can process completed request in clientreplay stage.
3270          */
3271         p = rb_first(&mdsc->request_tree);
3272         while (p) {
3273                 req = rb_entry(p, struct ceph_mds_request, r_node);
3274                 p = rb_next(p);
3275                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
3276                         continue;
3277                 if (req->r_attempts == 0)
3278                         continue; /* only old requests */
3279                 if (req->r_session &&
3280                     req->r_session->s_mds == session->s_mds)
3281                         __send_request(mdsc, session, req, true);
3282         }
3283         mutex_unlock(&mdsc->mutex);
3284 }
3285
3286 static int send_reconnect_partial(struct ceph_reconnect_state *recon_state)
3287 {
3288         struct ceph_msg *reply;
3289         struct ceph_pagelist *_pagelist;
3290         struct page *page;
3291         __le32 *addr;
3292         int err = -ENOMEM;
3293
3294         if (!recon_state->allow_multi)
3295                 return -ENOSPC;
3296
3297         /* can't handle message that contains both caps and realm */
3298         BUG_ON(!recon_state->nr_caps == !recon_state->nr_realms);
3299
3300         /* pre-allocate new pagelist */
3301         _pagelist = ceph_pagelist_alloc(GFP_NOFS);
3302         if (!_pagelist)
3303                 return -ENOMEM;
3304
3305         reply = ceph_msg_new2(CEPH_MSG_CLIENT_RECONNECT, 0, 1, GFP_NOFS, false);
3306         if (!reply)
3307                 goto fail_msg;
3308
3309         /* placeholder for nr_caps */
3310         err = ceph_pagelist_encode_32(_pagelist, 0);
3311         if (err < 0)
3312                 goto fail;
3313
3314         if (recon_state->nr_caps) {
3315                 /* currently encoding caps */
3316                 err = ceph_pagelist_encode_32(recon_state->pagelist, 0);
3317                 if (err)
3318                         goto fail;
3319         } else {
3320                 /* placeholder for nr_realms (currently encoding relams) */
3321                 err = ceph_pagelist_encode_32(_pagelist, 0);
3322                 if (err < 0)
3323                         goto fail;
3324         }
3325
3326         err = ceph_pagelist_encode_8(recon_state->pagelist, 1);
3327         if (err)
3328                 goto fail;
3329
3330         page = list_first_entry(&recon_state->pagelist->head, struct page, lru);
3331         addr = kmap_atomic(page);
3332         if (recon_state->nr_caps) {
3333                 /* currently encoding caps */
3334                 *addr = cpu_to_le32(recon_state->nr_caps);
3335         } else {
3336                 /* currently encoding relams */
3337                 *(addr + 1) = cpu_to_le32(recon_state->nr_realms);
3338         }
3339         kunmap_atomic(addr);
3340
3341         reply->hdr.version = cpu_to_le16(5);
3342         reply->hdr.compat_version = cpu_to_le16(4);
3343
3344         reply->hdr.data_len = cpu_to_le32(recon_state->pagelist->length);
3345         ceph_msg_data_add_pagelist(reply, recon_state->pagelist);
3346
3347         ceph_con_send(&recon_state->session->s_con, reply);
3348         ceph_pagelist_release(recon_state->pagelist);
3349
3350         recon_state->pagelist = _pagelist;
3351         recon_state->nr_caps = 0;
3352         recon_state->nr_realms = 0;
3353         recon_state->msg_version = 5;
3354         return 0;
3355 fail:
3356         ceph_msg_put(reply);
3357 fail_msg:
3358         ceph_pagelist_release(_pagelist);
3359         return err;
3360 }
3361
3362 /*
3363  * Encode information about a cap for a reconnect with the MDS.
3364  */
3365 static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap,
3366                           void *arg)
3367 {
3368         union {
3369                 struct ceph_mds_cap_reconnect v2;
3370                 struct ceph_mds_cap_reconnect_v1 v1;
3371         } rec;
3372         struct ceph_inode_info *ci = cap->ci;
3373         struct ceph_reconnect_state *recon_state = arg;
3374         struct ceph_pagelist *pagelist = recon_state->pagelist;
3375         int err;
3376         u64 snap_follows;
3377
3378         dout(" adding %p ino %llx.%llx cap %p %lld %s\n",
3379              inode, ceph_vinop(inode), cap, cap->cap_id,
3380              ceph_cap_string(cap->issued));
3381
3382         spin_lock(&ci->i_ceph_lock);
3383         cap->seq = 0;        /* reset cap seq */
3384         cap->issue_seq = 0;  /* and issue_seq */
3385         cap->mseq = 0;       /* and migrate_seq */
3386         cap->cap_gen = cap->session->s_cap_gen;
3387
3388         if (recon_state->msg_version >= 2) {
3389                 rec.v2.cap_id = cpu_to_le64(cap->cap_id);
3390                 rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
3391                 rec.v2.issued = cpu_to_le32(cap->issued);
3392                 rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
3393                 rec.v2.pathbase = 0;
3394                 rec.v2.flock_len = (__force __le32)
3395                         ((ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) ? 0 : 1);
3396         } else {
3397                 rec.v1.cap_id = cpu_to_le64(cap->cap_id);
3398                 rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
3399                 rec.v1.issued = cpu_to_le32(cap->issued);
3400                 rec.v1.size = cpu_to_le64(inode->i_size);
3401                 ceph_encode_timespec64(&rec.v1.mtime, &inode->i_mtime);
3402                 ceph_encode_timespec64(&rec.v1.atime, &inode->i_atime);
3403                 rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
3404                 rec.v1.pathbase = 0;
3405         }
3406
3407         if (list_empty(&ci->i_cap_snaps)) {
3408                 snap_follows = ci->i_head_snapc ? ci->i_head_snapc->seq : 0;
3409         } else {
3410                 struct ceph_cap_snap *capsnap =
3411                         list_first_entry(&ci->i_cap_snaps,
3412                                          struct ceph_cap_snap, ci_item);
3413                 snap_follows = capsnap->follows;
3414         }
3415         spin_unlock(&ci->i_ceph_lock);
3416
3417         if (recon_state->msg_version >= 2) {
3418                 int num_fcntl_locks, num_flock_locks;
3419                 struct ceph_filelock *flocks = NULL;
3420                 size_t struct_len, total_len = sizeof(u64);
3421                 u8 struct_v = 0;
3422
3423 encode_again:
3424                 if (rec.v2.flock_len) {
3425                         ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks);
3426                 } else {
3427                         num_fcntl_locks = 0;
3428                         num_flock_locks = 0;
3429                 }
3430                 if (num_fcntl_locks + num_flock_locks > 0) {
3431                         flocks = kmalloc_array(num_fcntl_locks + num_flock_locks,
3432                                                sizeof(struct ceph_filelock),
3433                                                GFP_NOFS);
3434                         if (!flocks) {
3435                                 err = -ENOMEM;
3436                                 goto out_err;
3437                         }
3438                         err = ceph_encode_locks_to_buffer(inode, flocks,
3439                                                           num_fcntl_locks,
3440                                                           num_flock_locks);
3441                         if (err) {
3442                                 kfree(flocks);
3443                                 flocks = NULL;
3444                                 if (err == -ENOSPC)
3445                                         goto encode_again;
3446                                 goto out_err;
3447                         }
3448                 } else {
3449                         kfree(flocks);
3450                         flocks = NULL;
3451                 }
3452
3453                 if (recon_state->msg_version >= 3) {
3454                         /* version, compat_version and struct_len */
3455                         total_len += 2 * sizeof(u8) + sizeof(u32);
3456                         struct_v = 2;
3457                 }
3458                 /*
3459                  * number of encoded locks is stable, so copy to pagelist
3460                  */
3461                 struct_len = 2 * sizeof(u32) +
3462                             (num_fcntl_locks + num_flock_locks) *
3463                             sizeof(struct ceph_filelock);
3464                 rec.v2.flock_len = cpu_to_le32(struct_len);
3465
3466                 struct_len += sizeof(u32) + sizeof(rec.v2);
3467
3468                 if (struct_v >= 2)
3469                         struct_len += sizeof(u64); /* snap_follows */
3470
3471                 total_len += struct_len;
3472
3473                 if (pagelist->length + total_len > RECONNECT_MAX_SIZE) {
3474                         err = send_reconnect_partial(recon_state);
3475                         if (err)
3476                                 goto out_freeflocks;
3477                         pagelist = recon_state->pagelist;
3478                 }
3479
3480                 err = ceph_pagelist_reserve(pagelist, total_len);
3481                 if (err)
3482                         goto out_freeflocks;
3483
3484                 ceph_pagelist_encode_64(pagelist, ceph_ino(inode));
3485                 if (recon_state->msg_version >= 3) {
3486                         ceph_pagelist_encode_8(pagelist, struct_v);
3487                         ceph_pagelist_encode_8(pagelist, 1);
3488                         ceph_pagelist_encode_32(pagelist, struct_len);
3489                 }
3490                 ceph_pagelist_encode_string(pagelist, NULL, 0);
3491                 ceph_pagelist_append(pagelist, &rec, sizeof(rec.v2));
3492                 ceph_locks_to_pagelist(flocks, pagelist,
3493                                        num_fcntl_locks, num_flock_locks);
3494                 if (struct_v >= 2)
3495                         ceph_pagelist_encode_64(pagelist, snap_follows);
3496 out_freeflocks:
3497                 kfree(flocks);
3498         } else {
3499                 u64 pathbase = 0;
3500                 int pathlen = 0;
3501                 char *path = NULL;
3502                 struct dentry *dentry;
3503
3504                 dentry = d_find_alias(inode);
3505                 if (dentry) {
3506                         path = ceph_mdsc_build_path(dentry,
3507                                                 &pathlen, &pathbase, 0);
3508                         dput(dentry);
3509                         if (IS_ERR(path)) {
3510                                 err = PTR_ERR(path);
3511                                 goto out_err;
3512                         }
3513                         rec.v1.pathbase = cpu_to_le64(pathbase);
3514                 }
3515
3516                 err = ceph_pagelist_reserve(pagelist,
3517                                             sizeof(u64) + sizeof(u32) +
3518                                             pathlen + sizeof(rec.v1));
3519                 if (err) {
3520                         goto out_freepath;
3521                 }
3522
3523                 ceph_pagelist_encode_64(pagelist, ceph_ino(inode));
3524                 ceph_pagelist_encode_string(pagelist, path, pathlen);
3525                 ceph_pagelist_append(pagelist, &rec, sizeof(rec.v1));
3526 out_freepath:
3527                 ceph_mdsc_free_path(path, pathlen);
3528         }
3529
3530 out_err:
3531         if (err >= 0)
3532                 recon_state->nr_caps++;
3533         return err;
3534 }
3535
3536 static int encode_snap_realms(struct ceph_mds_client *mdsc,
3537                               struct ceph_reconnect_state *recon_state)
3538 {
3539         struct rb_node *p;
3540         struct ceph_pagelist *pagelist = recon_state->pagelist;
3541         int err = 0;
3542
3543         if (recon_state->msg_version >= 4) {
3544                 err = ceph_pagelist_encode_32(pagelist, mdsc->num_snap_realms);
3545                 if (err < 0)
3546                         goto fail;
3547         }
3548
3549         /*
3550          * snaprealms.  we provide mds with the ino, seq (version), and
3551          * parent for all of our realms.  If the mds has any newer info,
3552          * it will tell us.
3553          */
3554         for (p = rb_first(&mdsc->snap_realms); p; p = rb_next(p)) {
3555                 struct ceph_snap_realm *realm =
3556                        rb_entry(p, struct ceph_snap_realm, node);
3557                 struct ceph_mds_snaprealm_reconnect sr_rec;
3558
3559                 if (recon_state->msg_version >= 4) {
3560                         size_t need = sizeof(u8) * 2 + sizeof(u32) +
3561                                       sizeof(sr_rec);
3562
3563                         if (pagelist->length + need > RECONNECT_MAX_SIZE) {
3564                                 err = send_reconnect_partial(recon_state);
3565                                 if (err)
3566                                         goto fail;
3567                                 pagelist = recon_state->pagelist;
3568                         }
3569
3570                         err = ceph_pagelist_reserve(pagelist, need);
3571                         if (err)
3572                                 goto fail;
3573
3574                         ceph_pagelist_encode_8(pagelist, 1);
3575                         ceph_pagelist_encode_8(pagelist, 1);
3576                         ceph_pagelist_encode_32(pagelist, sizeof(sr_rec));
3577                 }
3578
3579                 dout(" adding snap realm %llx seq %lld parent %llx\n",
3580                      realm->ino, realm->seq, realm->parent_ino);
3581                 sr_rec.ino = cpu_to_le64(realm->ino);
3582                 sr_rec.seq = cpu_to_le64(realm->seq);
3583                 sr_rec.parent = cpu_to_le64(realm->parent_ino);
3584
3585                 err = ceph_pagelist_append(pagelist, &sr_rec, sizeof(sr_rec));
3586                 if (err)
3587                         goto fail;
3588
3589                 recon_state->nr_realms++;
3590         }
3591 fail:
3592         return err;
3593 }
3594
3595
3596 /*
3597  * If an MDS fails and recovers, clients need to reconnect in order to
3598  * reestablish shared state.  This includes all caps issued through
3599  * this session _and_ the snap_realm hierarchy.  Because it's not
3600  * clear which snap realms the mds cares about, we send everything we
3601  * know about.. that ensures we'll then get any new info the
3602  * recovering MDS might have.
3603  *
3604  * This is a relatively heavyweight operation, but it's rare.
3605  *
3606  * called with mdsc->mutex held.
3607  */
3608 static void send_mds_reconnect(struct ceph_mds_client *mdsc,
3609                                struct ceph_mds_session *session)
3610 {
3611         struct ceph_msg *reply;
3612         int mds = session->s_mds;
3613         int err = -ENOMEM;
3614         struct ceph_reconnect_state recon_state = {
3615                 .session = session,
3616         };
3617         LIST_HEAD(dispose);
3618
3619         pr_info("mds%d reconnect start\n", mds);
3620
3621         recon_state.pagelist = ceph_pagelist_alloc(GFP_NOFS);
3622         if (!recon_state.pagelist)
3623                 goto fail_nopagelist;
3624
3625         reply = ceph_msg_new2(CEPH_MSG_CLIENT_RECONNECT, 0, 1, GFP_NOFS, false);
3626         if (!reply)
3627                 goto fail_nomsg;
3628
3629         mutex_lock(&session->s_mutex);
3630         session->s_state = CEPH_MDS_SESSION_RECONNECTING;
3631         session->s_seq = 0;
3632
3633         dout("session %p state %s\n", session,
3634              ceph_session_state_name(session->s_state));
3635
3636         spin_lock(&session->s_gen_ttl_lock);
3637         session->s_cap_gen++;
3638         spin_unlock(&session->s_gen_ttl_lock);
3639
3640         spin_lock(&session->s_cap_lock);
3641         /* don't know if session is readonly */
3642         session->s_readonly = 0;
3643         /*
3644          * notify __ceph_remove_cap() that we are composing cap reconnect.
3645          * If a cap get released before being added to the cap reconnect,
3646          * __ceph_remove_cap() should skip queuing cap release.
3647          */
3648         session->s_cap_reconnect = 1;
3649         /* drop old cap expires; we're about to reestablish that state */
3650         detach_cap_releases(session, &dispose);
3651         spin_unlock(&session->s_cap_lock);
3652         dispose_cap_releases(mdsc, &dispose);
3653
3654         /* trim unused caps to reduce MDS's cache rejoin time */
3655         if (mdsc->fsc->sb->s_root)
3656                 shrink_dcache_parent(mdsc->fsc->sb->s_root);
3657
3658         ceph_con_close(&session->s_con);
3659         ceph_con_open(&session->s_con,
3660                       CEPH_ENTITY_TYPE_MDS, mds,
3661                       ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
3662
3663         /* replay unsafe requests */
3664         replay_unsafe_requests(mdsc, session);
3665
3666         ceph_early_kick_flushing_caps(mdsc, session);
3667
3668         down_read(&mdsc->snap_rwsem);
3669
3670         /* placeholder for nr_caps */
3671         err = ceph_pagelist_encode_32(recon_state.pagelist, 0);
3672         if (err)
3673                 goto fail;
3674
3675         if (test_bit(CEPHFS_FEATURE_MULTI_RECONNECT, &session->s_features)) {
3676                 recon_state.msg_version = 3;
3677                 recon_state.allow_multi = true;
3678         } else if (session->s_con.peer_features & CEPH_FEATURE_MDSENC) {
3679                 recon_state.msg_version = 3;
3680         } else {
3681                 recon_state.msg_version = 2;
3682         }
3683         /* trsaverse this session's caps */
3684         err = ceph_iterate_session_caps(session, encode_caps_cb, &recon_state);
3685
3686         spin_lock(&session->s_cap_lock);
3687         session->s_cap_reconnect = 0;
3688         spin_unlock(&session->s_cap_lock);
3689
3690         if (err < 0)
3691                 goto fail;
3692
3693         /* check if all realms can be encoded into current message */
3694         if (mdsc->num_snap_realms) {
3695                 size_t total_len =
3696                         recon_state.pagelist->length +
3697                         mdsc->num_snap_realms *
3698                         sizeof(struct ceph_mds_snaprealm_reconnect);
3699                 if (recon_state.msg_version >= 4) {
3700                         /* number of realms */
3701                         total_len += sizeof(u32);
3702                         /* version, compat_version and struct_len */
3703                         total_len += mdsc->num_snap_realms *
3704                                      (2 * sizeof(u8) + sizeof(u32));
3705                 }
3706                 if (total_len > RECONNECT_MAX_SIZE) {
3707                         if (!recon_state.allow_multi) {
3708                                 err = -ENOSPC;
3709                                 goto fail;
3710                         }
3711                         if (recon_state.nr_caps) {
3712                                 err = send_reconnect_partial(&recon_state);
3713                                 if (err)
3714                                         goto fail;
3715                         }
3716                         recon_state.msg_version = 5;
3717                 }
3718         }
3719
3720         err = encode_snap_realms(mdsc, &recon_state);
3721         if (err < 0)
3722                 goto fail;
3723
3724         if (recon_state.msg_version >= 5) {
3725                 err = ceph_pagelist_encode_8(recon_state.pagelist, 0);
3726                 if (err < 0)
3727                         goto fail;
3728         }
3729
3730         if (recon_state.nr_caps || recon_state.nr_realms) {
3731                 struct page *page =
3732                         list_first_entry(&recon_state.pagelist->head,
3733                                         struct page, lru);
3734                 __le32 *addr = kmap_atomic(page);
3735                 if (recon_state.nr_caps) {
3736                         WARN_ON(recon_state.nr_realms != mdsc->num_snap_realms);
3737                         *addr = cpu_to_le32(recon_state.nr_caps);
3738                 } else if (recon_state.msg_version >= 4) {
3739                         *(addr + 1) = cpu_to_le32(recon_state.nr_realms);
3740                 }
3741                 kunmap_atomic(addr);
3742         }
3743
3744         reply->hdr.version = cpu_to_le16(recon_state.msg_version);
3745         if (recon_state.msg_version >= 4)
3746                 reply->hdr.compat_version = cpu_to_le16(4);
3747
3748         reply->hdr.data_len = cpu_to_le32(recon_state.pagelist->length);
3749         ceph_msg_data_add_pagelist(reply, recon_state.pagelist);
3750
3751         ceph_con_send(&session->s_con, reply);
3752
3753         mutex_unlock(&session->s_mutex);
3754
3755         mutex_lock(&mdsc->mutex);
3756         __wake_requests(mdsc, &session->s_waiting);
3757         mutex_unlock(&mdsc->mutex);
3758
3759         up_read(&mdsc->snap_rwsem);
3760         ceph_pagelist_release(recon_state.pagelist);
3761         return;
3762
3763 fail:
3764         ceph_msg_put(reply);
3765         up_read(&mdsc->snap_rwsem);
3766         mutex_unlock(&session->s_mutex);
3767 fail_nomsg:
3768         ceph_pagelist_release(recon_state.pagelist);
3769 fail_nopagelist:
3770         pr_err("error %d preparing reconnect for mds%d\n", err, mds);
3771         return;
3772 }
3773
3774
3775 /*
3776  * compare old and new mdsmaps, kicking requests
3777  * and closing out old connections as necessary
3778  *
3779  * called under mdsc->mutex.
3780  */
3781 static void check_new_map(struct ceph_mds_client *mdsc,
3782                           struct ceph_mdsmap *newmap,
3783                           struct ceph_mdsmap *oldmap)
3784 {
3785         int i;
3786         int oldstate, newstate;
3787         struct ceph_mds_session *s;
3788
3789         dout("check_new_map new %u old %u\n",
3790              newmap->m_epoch, oldmap->m_epoch);
3791
3792         for (i = 0; i < oldmap->possible_max_rank && i < mdsc->max_sessions; i++) {
3793                 if (!mdsc->sessions[i])
3794                         continue;
3795                 s = mdsc->sessions[i];
3796                 oldstate = ceph_mdsmap_get_state(oldmap, i);
3797                 newstate = ceph_mdsmap_get_state(newmap, i);
3798
3799                 dout("check_new_map mds%d state %s%s -> %s%s (session %s)\n",
3800                      i, ceph_mds_state_name(oldstate),
3801                      ceph_mdsmap_is_laggy(oldmap, i) ? " (laggy)" : "",
3802                      ceph_mds_state_name(newstate),
3803                      ceph_mdsmap_is_laggy(newmap, i) ? " (laggy)" : "",
3804                      ceph_session_state_name(s->s_state));
3805
3806                 if (i >= newmap->possible_max_rank) {
3807                         /* force close session for stopped mds */
3808                         ceph_get_mds_session(s);
3809                         __unregister_session(mdsc, s);
3810                         __wake_requests(mdsc, &s->s_waiting);
3811                         mutex_unlock(&mdsc->mutex);
3812
3813                         mutex_lock(&s->s_mutex);
3814                         cleanup_session_requests(mdsc, s);
3815                         remove_session_caps(s);
3816                         mutex_unlock(&s->s_mutex);
3817
3818                         ceph_put_mds_session(s);
3819
3820                         mutex_lock(&mdsc->mutex);
3821                         kick_requests(mdsc, i);
3822                         continue;
3823                 }
3824
3825                 if (memcmp(ceph_mdsmap_get_addr(oldmap, i),
3826                            ceph_mdsmap_get_addr(newmap, i),
3827                            sizeof(struct ceph_entity_addr))) {
3828                         /* just close it */
3829                         mutex_unlock(&mdsc->mutex);
3830                         mutex_lock(&s->s_mutex);
3831                         mutex_lock(&mdsc->mutex);
3832                         ceph_con_close(&s->s_con);
3833                         mutex_unlock(&s->s_mutex);
3834                         s->s_state = CEPH_MDS_SESSION_RESTARTING;
3835                 } else if (oldstate == newstate) {
3836                         continue;  /* nothing new with this mds */
3837                 }
3838
3839                 /*
3840                  * send reconnect?
3841                  */
3842                 if (s->s_state == CEPH_MDS_SESSION_RESTARTING &&
3843                     newstate >= CEPH_MDS_STATE_RECONNECT) {
3844                         mutex_unlock(&mdsc->mutex);
3845                         send_mds_reconnect(mdsc, s);
3846                         mutex_lock(&mdsc->mutex);
3847                 }
3848
3849                 /*
3850                  * kick request on any mds that has gone active.
3851                  */
3852                 if (oldstate < CEPH_MDS_STATE_ACTIVE &&
3853                     newstate >= CEPH_MDS_STATE_ACTIVE) {
3854                         if (oldstate != CEPH_MDS_STATE_CREATING &&
3855                             oldstate != CEPH_MDS_STATE_STARTING)
3856                                 pr_info("mds%d recovery completed\n", s->s_mds);
3857                         kick_requests(mdsc, i);
3858                         ceph_kick_flushing_caps(mdsc, s);
3859                         wake_up_session_caps(s, RECONNECT);
3860                 }
3861         }
3862
3863         for (i = 0; i < newmap->possible_max_rank && i < mdsc->max_sessions; i++) {
3864                 s = mdsc->sessions[i];
3865                 if (!s)
3866                         continue;
3867                 if (!ceph_mdsmap_is_laggy(newmap, i))
3868                         continue;
3869                 if (s->s_state == CEPH_MDS_SESSION_OPEN ||
3870                     s->s_state == CEPH_MDS_SESSION_HUNG ||
3871                     s->s_state == CEPH_MDS_SESSION_CLOSING) {
3872                         dout(" connecting to export targets of laggy mds%d\n",
3873                              i);
3874                         __open_export_target_sessions(mdsc, s);
3875                 }
3876         }
3877 }
3878
3879
3880
3881 /*
3882  * leases
3883  */
3884
3885 /*
3886  * caller must hold session s_mutex, dentry->d_lock
3887  */
3888 void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry)
3889 {
3890         struct ceph_dentry_info *di = ceph_dentry(dentry);
3891
3892         ceph_put_mds_session(di->lease_session);
3893         di->lease_session = NULL;
3894 }
3895
3896 static void handle_lease(struct ceph_mds_client *mdsc,
3897                          struct ceph_mds_session *session,
3898                          struct ceph_msg *msg)
3899 {
3900         struct super_block *sb = mdsc->fsc->sb;
3901         struct inode *inode;
3902         struct dentry *parent, *dentry;
3903         struct ceph_dentry_info *di;
3904         int mds = session->s_mds;
3905         struct ceph_mds_lease *h = msg->front.iov_base;
3906         u32 seq;
3907         struct ceph_vino vino;
3908         struct qstr dname;
3909         int release = 0;
3910
3911         dout("handle_lease from mds%d\n", mds);
3912
3913         /* decode */
3914         if (msg->front.iov_len < sizeof(*h) + sizeof(u32))
3915                 goto bad;
3916         vino.ino = le64_to_cpu(h->ino);
3917         vino.snap = CEPH_NOSNAP;
3918         seq = le32_to_cpu(h->seq);
3919         dname.len = get_unaligned_le32(h + 1);
3920         if (msg->front.iov_len < sizeof(*h) + sizeof(u32) + dname.len)
3921                 goto bad;
3922         dname.name = (void *)(h + 1) + sizeof(u32);
3923
3924         /* lookup inode */
3925         inode = ceph_find_inode(sb, vino);
3926         dout("handle_lease %s, ino %llx %p %.*s\n",
3927              ceph_lease_op_name(h->action), vino.ino, inode,
3928              dname.len, dname.name);
3929
3930         mutex_lock(&session->s_mutex);
3931         session->s_seq++;
3932
3933         if (!inode) {
3934                 dout("handle_lease no inode %llx\n", vino.ino);
3935                 goto release;
3936         }
3937
3938         /* dentry */
3939         parent = d_find_alias(inode);
3940         if (!parent) {
3941                 dout("no parent dentry on inode %p\n", inode);
3942                 WARN_ON(1);
3943                 goto release;  /* hrm... */
3944         }
3945         dname.hash = full_name_hash(parent, dname.name, dname.len);
3946         dentry = d_lookup(parent, &dname);
3947         dput(parent);
3948         if (!dentry)
3949                 goto release;
3950
3951         spin_lock(&dentry->d_lock);
3952         di = ceph_dentry(dentry);
3953         switch (h->action) {
3954         case CEPH_MDS_LEASE_REVOKE:
3955                 if (di->lease_session == session) {
3956                         if (ceph_seq_cmp(di->lease_seq, seq) > 0)
3957                                 h->seq = cpu_to_le32(di->lease_seq);
3958                         __ceph_mdsc_drop_dentry_lease(dentry);
3959                 }
3960                 release = 1;
3961                 break;
3962
3963         case CEPH_MDS_LEASE_RENEW:
3964                 if (di->lease_session == session &&
3965                     di->lease_gen == session->s_cap_gen &&
3966                     di->lease_renew_from &&
3967                     di->lease_renew_after == 0) {
3968                         unsigned long duration =
3969                                 msecs_to_jiffies(le32_to_cpu(h->duration_ms));
3970
3971                         di->lease_seq = seq;
3972                         di->time = di->lease_renew_from + duration;
3973                         di->lease_renew_after = di->lease_renew_from +
3974                                 (duration >> 1);
3975                         di->lease_renew_from = 0;
3976                 }
3977                 break;
3978         }
3979         spin_unlock(&dentry->d_lock);
3980         dput(dentry);
3981
3982         if (!release)
3983                 goto out;
3984
3985 release:
3986         /* let's just reuse the same message */
3987         h->action = CEPH_MDS_LEASE_REVOKE_ACK;
3988         ceph_msg_get(msg);
3989         ceph_con_send(&session->s_con, msg);
3990
3991 out:
3992         mutex_unlock(&session->s_mutex);
3993         /* avoid calling iput_final() in mds dispatch threads */
3994         ceph_async_iput(inode);
3995         return;
3996
3997 bad:
3998         pr_err("corrupt lease message\n");
3999         ceph_msg_dump(msg);
4000 }
4001
4002 void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
4003                               struct dentry *dentry, char action,
4004                               u32 seq)
4005 {
4006         struct ceph_msg *msg;
4007         struct ceph_mds_lease *lease;
4008         struct inode *dir;
4009         int len = sizeof(*lease) + sizeof(u32) + NAME_MAX;
4010
4011         dout("lease_send_msg identry %p %s to mds%d\n",
4012              dentry, ceph_lease_op_name(action), session->s_mds);
4013
4014         msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, GFP_NOFS, false);
4015         if (!msg)
4016                 return;
4017         lease = msg->front.iov_base;
4018         lease->action = action;
4019         lease->seq = cpu_to_le32(seq);
4020
4021         spin_lock(&dentry->d_lock);
4022         dir = d_inode(dentry->d_parent);
4023         lease->ino = cpu_to_le64(ceph_ino(dir));
4024         lease->first = lease->last = cpu_to_le64(ceph_snap(dir));
4025
4026         put_unaligned_le32(dentry->d_name.len, lease + 1);
4027         memcpy((void *)(lease + 1) + 4,
4028                dentry->d_name.name, dentry->d_name.len);
4029         spin_unlock(&dentry->d_lock);
4030         /*
4031          * if this is a preemptive lease RELEASE, no need to
4032          * flush request stream, since the actual request will
4033          * soon follow.
4034          */
4035         msg->more_to_follow = (action == CEPH_MDS_LEASE_RELEASE);
4036
4037         ceph_con_send(&session->s_con, msg);
4038 }
4039
4040 /*
4041  * lock unlock sessions, to wait ongoing session activities
4042  */
4043 static void lock_unlock_sessions(struct ceph_mds_client *mdsc)
4044 {
4045         int i;
4046
4047         mutex_lock(&mdsc->mutex);
4048         for (i = 0; i < mdsc->max_sessions; i++) {
4049                 struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i);
4050                 if (!s)
4051                         continue;
4052                 mutex_unlock(&mdsc->mutex);
4053                 mutex_lock(&s->s_mutex);
4054                 mutex_unlock(&s->s_mutex);
4055                 ceph_put_mds_session(s);
4056                 mutex_lock(&mdsc->mutex);
4057         }
4058         mutex_unlock(&mdsc->mutex);
4059 }
4060
4061 static void maybe_recover_session(struct ceph_mds_client *mdsc)
4062 {
4063         struct ceph_fs_client *fsc = mdsc->fsc;
4064
4065         if (!ceph_test_mount_opt(fsc, CLEANRECOVER))
4066                 return;
4067
4068         if (READ_ONCE(fsc->mount_state) != CEPH_MOUNT_MOUNTED)
4069                 return;
4070
4071         if (!READ_ONCE(fsc->blacklisted))
4072                 return;
4073
4074         if (fsc->last_auto_reconnect &&
4075             time_before(jiffies, fsc->last_auto_reconnect + HZ * 60 * 30))
4076                 return;
4077
4078         pr_info("auto reconnect after blacklisted\n");
4079         fsc->last_auto_reconnect = jiffies;
4080         ceph_force_reconnect(fsc->sb);
4081 }
4082
4083 /*
4084  * delayed work -- periodically trim expired leases, renew caps with mds
4085  */
4086 static void schedule_delayed(struct ceph_mds_client *mdsc)
4087 {
4088         int delay = 5;
4089         unsigned hz = round_jiffies_relative(HZ * delay);
4090         schedule_delayed_work(&mdsc->delayed_work, hz);
4091 }
4092
4093 static void delayed_work(struct work_struct *work)
4094 {
4095         int i;
4096         struct ceph_mds_client *mdsc =
4097                 container_of(work, struct ceph_mds_client, delayed_work.work);
4098         int renew_interval;
4099         int renew_caps;
4100
4101         dout("mdsc delayed_work\n");
4102
4103         mutex_lock(&mdsc->mutex);
4104         renew_interval = mdsc->mdsmap->m_session_timeout >> 2;
4105         renew_caps = time_after_eq(jiffies, HZ*renew_interval +
4106                                    mdsc->last_renew_caps);
4107         if (renew_caps)
4108                 mdsc->last_renew_caps = jiffies;
4109
4110         for (i = 0; i < mdsc->max_sessions; i++) {
4111                 struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i);
4112                 if (!s)
4113                         continue;
4114                 if (s->s_state == CEPH_MDS_SESSION_CLOSING) {
4115                         dout("resending session close request for mds%d\n",
4116                              s->s_mds);
4117                         request_close_session(mdsc, s);
4118                         ceph_put_mds_session(s);
4119                         continue;
4120                 }
4121                 if (s->s_ttl && time_after(jiffies, s->s_ttl)) {
4122                         if (s->s_state == CEPH_MDS_SESSION_OPEN) {
4123                                 s->s_state = CEPH_MDS_SESSION_HUNG;
4124                                 pr_info("mds%d hung\n", s->s_mds);
4125                         }
4126                 }
4127                 if (s->s_state == CEPH_MDS_SESSION_NEW ||
4128                     s->s_state == CEPH_MDS_SESSION_RESTARTING ||
4129                     s->s_state == CEPH_MDS_SESSION_REJECTED) {
4130                         /* this mds is failed or recovering, just wait */
4131                         ceph_put_mds_session(s);
4132                         continue;
4133                 }
4134                 mutex_unlock(&mdsc->mutex);
4135
4136                 mutex_lock(&s->s_mutex);
4137                 if (renew_caps)
4138                         send_renew_caps(mdsc, s);
4139                 else
4140                         ceph_con_keepalive(&s->s_con);
4141                 if (s->s_state == CEPH_MDS_SESSION_OPEN ||
4142                     s->s_state == CEPH_MDS_SESSION_HUNG)
4143                         ceph_send_cap_releases(mdsc, s);
4144                 mutex_unlock(&s->s_mutex);
4145                 ceph_put_mds_session(s);
4146
4147                 mutex_lock(&mdsc->mutex);
4148         }
4149         mutex_unlock(&mdsc->mutex);
4150
4151         ceph_check_delayed_caps(mdsc);
4152
4153         ceph_queue_cap_reclaim_work(mdsc);
4154
4155         ceph_trim_snapid_map(mdsc);
4156
4157         maybe_recover_session(mdsc);
4158
4159         schedule_delayed(mdsc);
4160 }
4161
4162 int ceph_mdsc_init(struct ceph_fs_client *fsc)
4163
4164 {
4165         struct ceph_mds_client *mdsc;
4166
4167         mdsc = kzalloc(sizeof(struct ceph_mds_client), GFP_NOFS);
4168         if (!mdsc)
4169                 return -ENOMEM;
4170         mdsc->fsc = fsc;
4171         mutex_init(&mdsc->mutex);
4172         mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS);
4173         if (!mdsc->mdsmap) {
4174                 kfree(mdsc);
4175                 return -ENOMEM;
4176         }
4177
4178         fsc->mdsc = mdsc;
4179         init_completion(&mdsc->safe_umount_waiters);
4180         init_waitqueue_head(&mdsc->session_close_wq);
4181         INIT_LIST_HEAD(&mdsc->waiting_for_map);
4182         mdsc->sessions = NULL;
4183         atomic_set(&mdsc->num_sessions, 0);
4184         mdsc->max_sessions = 0;
4185         mdsc->stopping = 0;
4186         atomic64_set(&mdsc->quotarealms_count, 0);
4187         mdsc->quotarealms_inodes = RB_ROOT;
4188         mutex_init(&mdsc->quotarealms_inodes_mutex);
4189         mdsc->last_snap_seq = 0;
4190         init_rwsem(&mdsc->snap_rwsem);
4191         mdsc->snap_realms = RB_ROOT;
4192         INIT_LIST_HEAD(&mdsc->snap_empty);
4193         mdsc->num_snap_realms = 0;
4194         spin_lock_init(&mdsc->snap_empty_lock);
4195         mdsc->last_tid = 0;
4196         mdsc->oldest_tid = 0;
4197         mdsc->request_tree = RB_ROOT;
4198         INIT_DELAYED_WORK(&mdsc->delayed_work, delayed_work);
4199         mdsc->last_renew_caps = jiffies;
4200         INIT_LIST_HEAD(&mdsc->cap_delay_list);
4201         INIT_LIST_HEAD(&mdsc->cap_wait_list);
4202         spin_lock_init(&mdsc->cap_delay_lock);
4203         INIT_LIST_HEAD(&mdsc->snap_flush_list);
4204         spin_lock_init(&mdsc->snap_flush_lock);
4205         mdsc->last_cap_flush_tid = 1;
4206         INIT_LIST_HEAD(&mdsc->cap_flush_list);
4207         INIT_LIST_HEAD(&mdsc->cap_dirty);
4208         INIT_LIST_HEAD(&mdsc->cap_dirty_migrating);
4209         mdsc->num_cap_flushing = 0;
4210         spin_lock_init(&mdsc->cap_dirty_lock);
4211         init_waitqueue_head(&mdsc->cap_flushing_wq);
4212         INIT_WORK(&mdsc->cap_reclaim_work, ceph_cap_reclaim_work);
4213         atomic_set(&mdsc->cap_reclaim_pending, 0);
4214
4215         spin_lock_init(&mdsc->dentry_list_lock);
4216         INIT_LIST_HEAD(&mdsc->dentry_leases);
4217         INIT_LIST_HEAD(&mdsc->dentry_dir_leases);
4218
4219         ceph_caps_init(mdsc);
4220         ceph_adjust_caps_max_min(mdsc, fsc->mount_options);
4221
4222         spin_lock_init(&mdsc->snapid_map_lock);
4223         mdsc->snapid_map_tree = RB_ROOT;
4224         INIT_LIST_HEAD(&mdsc->snapid_map_lru);
4225
4226         init_rwsem(&mdsc->pool_perm_rwsem);
4227         mdsc->pool_perm_tree = RB_ROOT;
4228
4229         strscpy(mdsc->nodename, utsname()->nodename,
4230                 sizeof(mdsc->nodename));
4231         return 0;
4232 }
4233
4234 /*
4235  * Wait for safe replies on open mds requests.  If we time out, drop
4236  * all requests from the tree to avoid dangling dentry refs.
4237  */
4238 static void wait_requests(struct ceph_mds_client *mdsc)
4239 {
4240         struct ceph_options *opts = mdsc->fsc->client->options;
4241         struct ceph_mds_request *req;
4242
4243         mutex_lock(&mdsc->mutex);
4244         if (__get_oldest_req(mdsc)) {
4245                 mutex_unlock(&mdsc->mutex);
4246
4247                 dout("wait_requests waiting for requests\n");
4248                 wait_for_completion_timeout(&mdsc->safe_umount_waiters,
4249                                     ceph_timeout_jiffies(opts->mount_timeout));
4250
4251                 /* tear down remaining requests */
4252                 mutex_lock(&mdsc->mutex);
4253                 while ((req = __get_oldest_req(mdsc))) {
4254                         dout("wait_requests timed out on tid %llu\n",
4255                              req->r_tid);
4256                         list_del_init(&req->r_wait);
4257                         __unregister_request(mdsc, req);
4258                 }
4259         }
4260         mutex_unlock(&mdsc->mutex);
4261         dout("wait_requests done\n");
4262 }
4263
4264 /*
4265  * called before mount is ro, and before dentries are torn down.
4266  * (hmm, does this still race with new lookups?)
4267  */
4268 void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc)
4269 {
4270         dout("pre_umount\n");
4271         mdsc->stopping = 1;
4272
4273         lock_unlock_sessions(mdsc);
4274         ceph_flush_dirty_caps(mdsc);
4275         wait_requests(mdsc);
4276
4277         /*
4278          * wait for reply handlers to drop their request refs and
4279          * their inode/dcache refs
4280          */
4281         ceph_msgr_flush();
4282
4283         ceph_cleanup_quotarealms_inodes(mdsc);
4284 }
4285
4286 /*
4287  * wait for all write mds requests to flush.
4288  */
4289 static void wait_unsafe_requests(struct ceph_mds_client *mdsc, u64 want_tid)
4290 {
4291         struct ceph_mds_request *req = NULL, *nextreq;
4292         struct rb_node *n;
4293
4294         mutex_lock(&mdsc->mutex);
4295         dout("wait_unsafe_requests want %lld\n", want_tid);
4296 restart:
4297         req = __get_oldest_req(mdsc);
4298         while (req && req->r_tid <= want_tid) {
4299                 /* find next request */
4300                 n = rb_next(&req->r_node);
4301                 if (n)
4302                         nextreq = rb_entry(n, struct ceph_mds_request, r_node);
4303                 else
4304                         nextreq = NULL;
4305                 if (req->r_op != CEPH_MDS_OP_SETFILELOCK &&
4306                     (req->r_op & CEPH_MDS_OP_WRITE)) {
4307                         /* write op */
4308                         ceph_mdsc_get_request(req);
4309                         if (nextreq)
4310                                 ceph_mdsc_get_request(nextreq);
4311                         mutex_unlock(&mdsc->mutex);
4312                         dout("wait_unsafe_requests  wait on %llu (want %llu)\n",
4313                              req->r_tid, want_tid);
4314                         wait_for_completion(&req->r_safe_completion);
4315                         mutex_lock(&mdsc->mutex);
4316                         ceph_mdsc_put_request(req);
4317                         if (!nextreq)
4318                                 break;  /* next dne before, so we're done! */
4319                         if (RB_EMPTY_NODE(&nextreq->r_node)) {
4320                                 /* next request was removed from tree */
4321                                 ceph_mdsc_put_request(nextreq);
4322                                 goto restart;
4323                         }
4324                         ceph_mdsc_put_request(nextreq);  /* won't go away */
4325                 }
4326                 req = nextreq;
4327         }
4328         mutex_unlock(&mdsc->mutex);
4329         dout("wait_unsafe_requests done\n");
4330 }
4331
4332 void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
4333 {
4334         u64 want_tid, want_flush;
4335
4336         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN)
4337                 return;
4338
4339         dout("sync\n");
4340         mutex_lock(&mdsc->mutex);
4341         want_tid = mdsc->last_tid;
4342         mutex_unlock(&mdsc->mutex);
4343
4344         ceph_flush_dirty_caps(mdsc);
4345         spin_lock(&mdsc->cap_dirty_lock);
4346         want_flush = mdsc->last_cap_flush_tid;
4347         if (!list_empty(&mdsc->cap_flush_list)) {
4348                 struct ceph_cap_flush *cf =
4349                         list_last_entry(&mdsc->cap_flush_list,
4350                                         struct ceph_cap_flush, g_list);
4351                 cf->wake = true;
4352         }
4353         spin_unlock(&mdsc->cap_dirty_lock);
4354
4355         dout("sync want tid %lld flush_seq %lld\n",
4356              want_tid, want_flush);
4357
4358         wait_unsafe_requests(mdsc, want_tid);
4359         wait_caps_flush(mdsc, want_flush);
4360 }
4361
4362 /*
4363  * true if all sessions are closed, or we force unmount
4364  */
4365 static bool done_closing_sessions(struct ceph_mds_client *mdsc, int skipped)
4366 {
4367         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN)
4368                 return true;
4369         return atomic_read(&mdsc->num_sessions) <= skipped;
4370 }
4371
4372 /*
4373  * called after sb is ro.
4374  */
4375 void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc)
4376 {
4377         struct ceph_options *opts = mdsc->fsc->client->options;
4378         struct ceph_mds_session *session;
4379         int i;
4380         int skipped = 0;
4381
4382         dout("close_sessions\n");
4383
4384         /* close sessions */
4385         mutex_lock(&mdsc->mutex);
4386         for (i = 0; i < mdsc->max_sessions; i++) {
4387                 session = __ceph_lookup_mds_session(mdsc, i);
4388                 if (!session)
4389                         continue;
4390                 mutex_unlock(&mdsc->mutex);
4391                 mutex_lock(&session->s_mutex);
4392                 if (__close_session(mdsc, session) <= 0)
4393                         skipped++;
4394                 mutex_unlock(&session->s_mutex);
4395                 ceph_put_mds_session(session);
4396                 mutex_lock(&mdsc->mutex);
4397         }
4398         mutex_unlock(&mdsc->mutex);
4399
4400         dout("waiting for sessions to close\n");
4401         wait_event_timeout(mdsc->session_close_wq,
4402                            done_closing_sessions(mdsc, skipped),
4403                            ceph_timeout_jiffies(opts->mount_timeout));
4404
4405         /* tear down remaining sessions */
4406         mutex_lock(&mdsc->mutex);
4407         for (i = 0; i < mdsc->max_sessions; i++) {
4408                 if (mdsc->sessions[i]) {
4409                         session = ceph_get_mds_session(mdsc->sessions[i]);
4410                         __unregister_session(mdsc, session);
4411                         mutex_unlock(&mdsc->mutex);
4412                         mutex_lock(&session->s_mutex);
4413                         remove_session_caps(session);
4414                         mutex_unlock(&session->s_mutex);
4415                         ceph_put_mds_session(session);
4416                         mutex_lock(&mdsc->mutex);
4417                 }
4418         }
4419         WARN_ON(!list_empty(&mdsc->cap_delay_list));
4420         mutex_unlock(&mdsc->mutex);
4421
4422         ceph_cleanup_snapid_map(mdsc);
4423         ceph_cleanup_empty_realms(mdsc);
4424
4425         cancel_work_sync(&mdsc->cap_reclaim_work);
4426         cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */
4427
4428         dout("stopped\n");
4429 }
4430
4431 void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc)
4432 {
4433         struct ceph_mds_session *session;
4434         int mds;
4435
4436         dout("force umount\n");
4437
4438         mutex_lock(&mdsc->mutex);
4439         for (mds = 0; mds < mdsc->max_sessions; mds++) {
4440                 session = __ceph_lookup_mds_session(mdsc, mds);
4441                 if (!session)
4442                         continue;
4443
4444                 if (session->s_state == CEPH_MDS_SESSION_REJECTED)
4445                         __unregister_session(mdsc, session);
4446                 __wake_requests(mdsc, &session->s_waiting);
4447                 mutex_unlock(&mdsc->mutex);
4448
4449                 mutex_lock(&session->s_mutex);
4450                 __close_session(mdsc, session);
4451                 if (session->s_state == CEPH_MDS_SESSION_CLOSING) {
4452                         cleanup_session_requests(mdsc, session);
4453                         remove_session_caps(session);
4454                 }
4455                 mutex_unlock(&session->s_mutex);
4456                 ceph_put_mds_session(session);
4457
4458                 mutex_lock(&mdsc->mutex);
4459                 kick_requests(mdsc, mds);
4460         }
4461         __wake_requests(mdsc, &mdsc->waiting_for_map);
4462         mutex_unlock(&mdsc->mutex);
4463 }
4464
4465 static void ceph_mdsc_stop(struct ceph_mds_client *mdsc)
4466 {
4467         dout("stop\n");
4468         cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */
4469         if (mdsc->mdsmap)
4470                 ceph_mdsmap_destroy(mdsc->mdsmap);
4471         kfree(mdsc->sessions);
4472         ceph_caps_finalize(mdsc);
4473         ceph_pool_perm_destroy(mdsc);
4474 }
4475
4476 void ceph_mdsc_destroy(struct ceph_fs_client *fsc)
4477 {
4478         struct ceph_mds_client *mdsc = fsc->mdsc;
4479         dout("mdsc_destroy %p\n", mdsc);
4480
4481         if (!mdsc)
4482                 return;
4483
4484         /* flush out any connection work with references to us */
4485         ceph_msgr_flush();
4486
4487         ceph_mdsc_stop(mdsc);
4488
4489         fsc->mdsc = NULL;
4490         kfree(mdsc);
4491         dout("mdsc_destroy %p done\n", mdsc);
4492 }
4493
4494 void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
4495 {
4496         struct ceph_fs_client *fsc = mdsc->fsc;
4497         const char *mds_namespace = fsc->mount_options->mds_namespace;
4498         void *p = msg->front.iov_base;
4499         void *end = p + msg->front.iov_len;
4500         u32 epoch;
4501         u32 map_len;
4502         u32 num_fs;
4503         u32 mount_fscid = (u32)-1;
4504         u8 struct_v, struct_cv;
4505         int err = -EINVAL;
4506
4507         ceph_decode_need(&p, end, sizeof(u32), bad);
4508         epoch = ceph_decode_32(&p);
4509
4510         dout("handle_fsmap epoch %u\n", epoch);
4511
4512         ceph_decode_need(&p, end, 2 + sizeof(u32), bad);
4513         struct_v = ceph_decode_8(&p);
4514         struct_cv = ceph_decode_8(&p);
4515         map_len = ceph_decode_32(&p);
4516
4517         ceph_decode_need(&p, end, sizeof(u32) * 3, bad);
4518         p += sizeof(u32) * 2; /* skip epoch and legacy_client_fscid */
4519
4520         num_fs = ceph_decode_32(&p);
4521         while (num_fs-- > 0) {
4522                 void *info_p, *info_end;
4523                 u32 info_len;
4524                 u8 info_v, info_cv;
4525                 u32 fscid, namelen;
4526
4527                 ceph_decode_need(&p, end, 2 + sizeof(u32), bad);
4528                 info_v = ceph_decode_8(&p);
4529                 info_cv = ceph_decode_8(&p);
4530                 info_len = ceph_decode_32(&p);
4531                 ceph_decode_need(&p, end, info_len, bad);
4532                 info_p = p;
4533                 info_end = p + info_len;
4534                 p = info_end;
4535
4536                 ceph_decode_need(&info_p, info_end, sizeof(u32) * 2, bad);
4537                 fscid = ceph_decode_32(&info_p);
4538                 namelen = ceph_decode_32(&info_p);
4539                 ceph_decode_need(&info_p, info_end, namelen, bad);
4540
4541                 if (mds_namespace &&
4542                     strlen(mds_namespace) == namelen &&
4543                     !strncmp(mds_namespace, (char *)info_p, namelen)) {
4544                         mount_fscid = fscid;
4545                         break;
4546                 }
4547         }
4548
4549         ceph_monc_got_map(&fsc->client->monc, CEPH_SUB_FSMAP, epoch);
4550         if (mount_fscid != (u32)-1) {
4551                 fsc->client->monc.fs_cluster_id = mount_fscid;
4552                 ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP,
4553                                    0, true);
4554                 ceph_monc_renew_subs(&fsc->client->monc);
4555         } else {
4556                 err = -ENOENT;
4557                 goto err_out;
4558         }
4559         return;
4560
4561 bad:
4562         pr_err("error decoding fsmap\n");
4563 err_out:
4564         mutex_lock(&mdsc->mutex);
4565         mdsc->mdsmap_err = err;
4566         __wake_requests(mdsc, &mdsc->waiting_for_map);
4567         mutex_unlock(&mdsc->mutex);
4568 }
4569
4570 /*
4571  * handle mds map update.
4572  */
4573 void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
4574 {
4575         u32 epoch;
4576         u32 maplen;
4577         void *p = msg->front.iov_base;
4578         void *end = p + msg->front.iov_len;
4579         struct ceph_mdsmap *newmap, *oldmap;
4580         struct ceph_fsid fsid;
4581         int err = -EINVAL;
4582
4583         ceph_decode_need(&p, end, sizeof(fsid)+2*sizeof(u32), bad);
4584         ceph_decode_copy(&p, &fsid, sizeof(fsid));
4585         if (ceph_check_fsid(mdsc->fsc->client, &fsid) < 0)
4586                 return;
4587         epoch = ceph_decode_32(&p);
4588         maplen = ceph_decode_32(&p);
4589         dout("handle_map epoch %u len %d\n", epoch, (int)maplen);
4590
4591         /* do we need it? */
4592         mutex_lock(&mdsc->mutex);
4593         if (mdsc->mdsmap && epoch <= mdsc->mdsmap->m_epoch) {
4594                 dout("handle_map epoch %u <= our %u\n",
4595                      epoch, mdsc->mdsmap->m_epoch);
4596                 mutex_unlock(&mdsc->mutex);
4597                 return;
4598         }
4599
4600         newmap = ceph_mdsmap_decode(&p, end);
4601         if (IS_ERR(newmap)) {
4602                 err = PTR_ERR(newmap);
4603                 goto bad_unlock;
4604         }
4605
4606         /* swap into place */
4607         if (mdsc->mdsmap) {
4608                 oldmap = mdsc->mdsmap;
4609                 mdsc->mdsmap = newmap;
4610                 check_new_map(mdsc, newmap, oldmap);
4611                 ceph_mdsmap_destroy(oldmap);
4612         } else {
4613                 mdsc->mdsmap = newmap;  /* first mds map */
4614         }
4615         mdsc->fsc->max_file_size = min((loff_t)mdsc->mdsmap->m_max_file_size,
4616                                         MAX_LFS_FILESIZE);
4617
4618         __wake_requests(mdsc, &mdsc->waiting_for_map);
4619         ceph_monc_got_map(&mdsc->fsc->client->monc, CEPH_SUB_MDSMAP,
4620                           mdsc->mdsmap->m_epoch);
4621
4622         mutex_unlock(&mdsc->mutex);
4623         schedule_delayed(mdsc);
4624         return;
4625
4626 bad_unlock:
4627         mutex_unlock(&mdsc->mutex);
4628 bad:
4629         pr_err("error decoding mdsmap %d\n", err);
4630         return;
4631 }
4632
4633 static struct ceph_connection *con_get(struct ceph_connection *con)
4634 {
4635         struct ceph_mds_session *s = con->private;
4636
4637         if (ceph_get_mds_session(s))
4638                 return con;
4639         return NULL;
4640 }
4641
4642 static void con_put(struct ceph_connection *con)
4643 {
4644         struct ceph_mds_session *s = con->private;
4645
4646         ceph_put_mds_session(s);
4647 }
4648
4649 /*
4650  * if the client is unresponsive for long enough, the mds will kill
4651  * the session entirely.
4652  */
4653 static void peer_reset(struct ceph_connection *con)
4654 {
4655         struct ceph_mds_session *s = con->private;
4656         struct ceph_mds_client *mdsc = s->s_mdsc;
4657
4658         pr_warn("mds%d closed our session\n", s->s_mds);
4659         send_mds_reconnect(mdsc, s);
4660 }
4661
4662 static void dispatch(struct ceph_connection *con, struct ceph_msg *msg)
4663 {
4664         struct ceph_mds_session *s = con->private;
4665         struct ceph_mds_client *mdsc = s->s_mdsc;
4666         int type = le16_to_cpu(msg->hdr.type);
4667
4668         mutex_lock(&mdsc->mutex);
4669         if (__verify_registered_session(mdsc, s) < 0) {
4670                 mutex_unlock(&mdsc->mutex);
4671                 goto out;
4672         }
4673         mutex_unlock(&mdsc->mutex);
4674
4675         switch (type) {
4676         case CEPH_MSG_MDS_MAP:
4677                 ceph_mdsc_handle_mdsmap(mdsc, msg);
4678                 break;
4679         case CEPH_MSG_FS_MAP_USER:
4680                 ceph_mdsc_handle_fsmap(mdsc, msg);
4681                 break;
4682         case CEPH_MSG_CLIENT_SESSION:
4683                 handle_session(s, msg);
4684                 break;
4685         case CEPH_MSG_CLIENT_REPLY:
4686                 handle_reply(s, msg);
4687                 break;
4688         case CEPH_MSG_CLIENT_REQUEST_FORWARD:
4689                 handle_forward(mdsc, s, msg);
4690                 break;
4691         case CEPH_MSG_CLIENT_CAPS:
4692                 ceph_handle_caps(s, msg);
4693                 break;
4694         case CEPH_MSG_CLIENT_SNAP:
4695                 ceph_handle_snap(mdsc, s, msg);
4696                 break;
4697         case CEPH_MSG_CLIENT_LEASE:
4698                 handle_lease(mdsc, s, msg);
4699                 break;
4700         case CEPH_MSG_CLIENT_QUOTA:
4701                 ceph_handle_quota(mdsc, s, msg);
4702                 break;
4703
4704         default:
4705                 pr_err("received unknown message type %d %s\n", type,
4706                        ceph_msg_type_name(type));
4707         }
4708 out:
4709         ceph_msg_put(msg);
4710 }
4711
4712 /*
4713  * authentication
4714  */
4715
4716 /*
4717  * Note: returned pointer is the address of a structure that's
4718  * managed separately.  Caller must *not* attempt to free it.
4719  */
4720 static struct ceph_auth_handshake *get_authorizer(struct ceph_connection *con,
4721                                         int *proto, int force_new)
4722 {
4723         struct ceph_mds_session *s = con->private;
4724         struct ceph_mds_client *mdsc = s->s_mdsc;
4725         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
4726         struct ceph_auth_handshake *auth = &s->s_auth;
4727
4728         if (force_new && auth->authorizer) {
4729                 ceph_auth_destroy_authorizer(auth->authorizer);
4730                 auth->authorizer = NULL;
4731         }
4732         if (!auth->authorizer) {
4733                 int ret = ceph_auth_create_authorizer(ac, CEPH_ENTITY_TYPE_MDS,
4734                                                       auth);
4735                 if (ret)
4736                         return ERR_PTR(ret);
4737         } else {
4738                 int ret = ceph_auth_update_authorizer(ac, CEPH_ENTITY_TYPE_MDS,
4739                                                       auth);
4740                 if (ret)
4741                         return ERR_PTR(ret);
4742         }
4743         *proto = ac->protocol;
4744
4745         return auth;
4746 }
4747
4748 static int add_authorizer_challenge(struct ceph_connection *con,
4749                                     void *challenge_buf, int challenge_buf_len)
4750 {
4751         struct ceph_mds_session *s = con->private;
4752         struct ceph_mds_client *mdsc = s->s_mdsc;
4753         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
4754
4755         return ceph_auth_add_authorizer_challenge(ac, s->s_auth.authorizer,
4756                                             challenge_buf, challenge_buf_len);
4757 }
4758
4759 static int verify_authorizer_reply(struct ceph_connection *con)
4760 {
4761         struct ceph_mds_session *s = con->private;
4762         struct ceph_mds_client *mdsc = s->s_mdsc;
4763         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
4764
4765         return ceph_auth_verify_authorizer_reply(ac, s->s_auth.authorizer);
4766 }
4767
4768 static int invalidate_authorizer(struct ceph_connection *con)
4769 {
4770         struct ceph_mds_session *s = con->private;
4771         struct ceph_mds_client *mdsc = s->s_mdsc;
4772         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
4773
4774         ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS);
4775
4776         return ceph_monc_validate_auth(&mdsc->fsc->client->monc);
4777 }
4778
4779 static struct ceph_msg *mds_alloc_msg(struct ceph_connection *con,
4780                                 struct ceph_msg_header *hdr, int *skip)
4781 {
4782         struct ceph_msg *msg;
4783         int type = (int) le16_to_cpu(hdr->type);
4784         int front_len = (int) le32_to_cpu(hdr->front_len);
4785
4786         if (con->in_msg)
4787                 return con->in_msg;
4788
4789         *skip = 0;
4790         msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
4791         if (!msg) {
4792                 pr_err("unable to allocate msg type %d len %d\n",
4793                        type, front_len);
4794                 return NULL;
4795         }
4796
4797         return msg;
4798 }
4799
4800 static int mds_sign_message(struct ceph_msg *msg)
4801 {
4802        struct ceph_mds_session *s = msg->con->private;
4803        struct ceph_auth_handshake *auth = &s->s_auth;
4804
4805        return ceph_auth_sign_message(auth, msg);
4806 }
4807
4808 static int mds_check_message_signature(struct ceph_msg *msg)
4809 {
4810        struct ceph_mds_session *s = msg->con->private;
4811        struct ceph_auth_handshake *auth = &s->s_auth;
4812
4813        return ceph_auth_check_message_signature(auth, msg);
4814 }
4815
4816 static const struct ceph_connection_operations mds_con_ops = {
4817         .get = con_get,
4818         .put = con_put,
4819         .dispatch = dispatch,
4820         .get_authorizer = get_authorizer,
4821         .add_authorizer_challenge = add_authorizer_challenge,
4822         .verify_authorizer_reply = verify_authorizer_reply,
4823         .invalidate_authorizer = invalidate_authorizer,
4824         .peer_reset = peer_reset,
4825         .alloc_msg = mds_alloc_msg,
4826         .sign_message = mds_sign_message,
4827         .check_message_signature = mds_check_message_signature,
4828 };
4829
4830 /* eof */