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 #include <linux/ktime.h>
  14 #include <linux/bitmap.h>
  15 #include <linux/mnt_idmapping.h>
  16
  17 #include "super.h"
  18 #include "mds_client.h"
  19 #include "crypto.h"
  20
  21 #include <linux/ceph/ceph_features.h>
  22 #include <linux/ceph/messenger.h>
  23 #include <linux/ceph/decode.h>
  24 #include <linux/ceph/pagelist.h>
  25 #include <linux/ceph/auth.h>
  26 #include <linux/ceph/debugfs.h>
  27
  28 #define RECONNECT_MAX_SIZE (INT_MAX - PAGE_SIZE)
  29
  30 /*
  31  * A cluster of MDS (metadata server) daemons is responsible for
  32  * managing the file system namespace (the directory hierarchy and
  33  * inodes) and for coordinating shared access to storage.  Metadata is
  34  * partitioning hierarchically across a number of servers, and that
  35  * partition varies over time as the cluster adjusts the distribution
  36  * in order to balance load.
  37  *
  38  * The MDS client is primarily responsible to managing synchronous
  39  * metadata requests for operations like open, unlink, and so forth.
  40  * If there is a MDS failure, we find out about it when we (possibly
  41  * request and) receive a new MDS map, and can resubmit affected
  42  * requests.
  43  *
  44  * For the most part, though, we take advantage of a lossless
  45  * communications channel to the MDS, and do not need to worry about
  46  * timing out or resubmitting requests.
  47  *
  48  * We maintain a stateful "session" with each MDS we interact with.
  49  * Within each session, we sent periodic heartbeat messages to ensure
  50  * any capabilities or leases we have been issues remain valid.  If
  51  * the session times out and goes stale, our leases and capabilities
  52  * are no longer valid.
  53  */
  54
  55 struct ceph_reconnect_state {
  56         struct ceph_mds_session *session;
  57         int nr_caps, nr_realms;
  58         struct ceph_pagelist *pagelist;
  59         unsigned msg_version;
  60         bool allow_multi;
  61 };
  62
  63 static void __wake_requests(struct ceph_mds_client *mdsc,
  64                             struct list_head *head);
  65 static void ceph_cap_release_work(struct work_struct *work);
  66 static void ceph_cap_reclaim_work(struct work_struct *work);
  67
  68 static const struct ceph_connection_operations mds_con_ops;
  69
  70
  71 /*
  72  * mds reply parsing
  73  */
  74
  75 static int parse_reply_info_quota(void **p, void *end,
  76                                   struct ceph_mds_reply_info_in *info)
  77 {
  78         u8 struct_v, struct_compat;
  79         u32 struct_len;
  80
  81         ceph_decode_8_safe(p, end, struct_v, bad);
  82         ceph_decode_8_safe(p, end, struct_compat, bad);
  83         /* struct_v is expected to be >= 1. we only
  84          * understand encoding with struct_compat == 1. */
  85         if (!struct_v || struct_compat != 1)
  86                 goto bad;
  87         ceph_decode_32_safe(p, end, struct_len, bad);
  88         ceph_decode_need(p, end, struct_len, bad);
  89         end = *p + struct_len;
  90         ceph_decode_64_safe(p, end, info->max_bytes, bad);
  91         ceph_decode_64_safe(p, end, info->max_files, bad);
  92         *p = end;
  93         return 0;
  94 bad:
  95         return -EIO;
  96 }
  97
  98 /*
  99  * parse individual inode info
 100  */
 101 static int parse_reply_info_in(void **p, void *end,
 102                                struct ceph_mds_reply_info_in *info,
 103                                u64 features)
 104 {
 105         int err = 0;
 106         u8 struct_v = 0;
 107
 108         if (features == (u64)-1) {
 109                 u32 struct_len;
 110                 u8 struct_compat;
 111                 ceph_decode_8_safe(p, end, struct_v, bad);
 112                 ceph_decode_8_safe(p, end, struct_compat, bad);
 113                 /* struct_v is expected to be >= 1. we only understand
 114                  * encoding with struct_compat == 1. */
 115                 if (!struct_v || struct_compat != 1)
 116                         goto bad;
 117                 ceph_decode_32_safe(p, end, struct_len, bad);
 118                 ceph_decode_need(p, end, struct_len, bad);
 119                 end = *p + struct_len;
 120         }
 121
 122         ceph_decode_need(p, end, sizeof(struct ceph_mds_reply_inode), bad);
 123         info->in = *p;
 124         *p += sizeof(struct ceph_mds_reply_inode) +
 125                 sizeof(*info->in->fragtree.splits) *
 126                 le32_to_cpu(info->in->fragtree.nsplits);
 127
 128         ceph_decode_32_safe(p, end, info->symlink_len, bad);
 129         ceph_decode_need(p, end, info->symlink_len, bad);
 130         info->symlink = *p;
 131         *p += info->symlink_len;
 132
 133         ceph_decode_copy_safe(p, end, &info->dir_layout,
 134                               sizeof(info->dir_layout), bad);
 135         ceph_decode_32_safe(p, end, info->xattr_len, bad);
 136         ceph_decode_need(p, end, info->xattr_len, bad);
 137         info->xattr_data = *p;
 138         *p += info->xattr_len;
 139
 140         if (features == (u64)-1) {
 141                 /* inline data */
 142                 ceph_decode_64_safe(p, end, info->inline_version, bad);
 143                 ceph_decode_32_safe(p, end, info->inline_len, bad);
 144                 ceph_decode_need(p, end, info->inline_len, bad);
 145                 info->inline_data = *p;
 146                 *p += info->inline_len;
 147                 /* quota */
 148                 err = parse_reply_info_quota(p, end, info);
 149                 if (err < 0)
 150                         goto out_bad;
 151                 /* pool namespace */
 152                 ceph_decode_32_safe(p, end, info->pool_ns_len, bad);
 153                 if (info->pool_ns_len > 0) {
 154                         ceph_decode_need(p, end, info->pool_ns_len, bad);
 155                         info->pool_ns_data = *p;
 156                         *p += info->pool_ns_len;
 157                 }
 158
 159                 /* btime */
 160                 ceph_decode_need(p, end, sizeof(info->btime), bad);
 161                 ceph_decode_copy(p, &info->btime, sizeof(info->btime));
 162
 163                 /* change attribute */
 164                 ceph_decode_64_safe(p, end, info->change_attr, bad);
 165
 166                 /* dir pin */
 167                 if (struct_v >= 2) {
 168                         ceph_decode_32_safe(p, end, info->dir_pin, bad);
 169                 } else {
 170                         info->dir_pin = -ENODATA;
 171                 }
 172
 173                 /* snapshot birth time, remains zero for v<=2 */
 174                 if (struct_v >= 3) {
 175                         ceph_decode_need(p, end, sizeof(info->snap_btime), bad);
 176                         ceph_decode_copy(p, &info->snap_btime,
 177                                          sizeof(info->snap_btime));
 178                 } else {
 179                         memset(&info->snap_btime, 0, sizeof(info->snap_btime));
 180                 }
 181
 182                 /* snapshot count, remains zero for v<=3 */
 183                 if (struct_v >= 4) {
 184                         ceph_decode_64_safe(p, end, info->rsnaps, bad);
 185                 } else {
 186                         info->rsnaps = 0;
 187                 }
 188
 189                 if (struct_v >= 5) {
 190                         u32 alen;
 191
 192                         ceph_decode_32_safe(p, end, alen, bad);
 193
 194                         while (alen--) {
 195                                 u32 len;
 196
 197                                 /* key */
 198                                 ceph_decode_32_safe(p, end, len, bad);
 199                                 ceph_decode_skip_n(p, end, len, bad);
 200                                 /* value */
 201                                 ceph_decode_32_safe(p, end, len, bad);
 202                                 ceph_decode_skip_n(p, end, len, bad);
 203                         }
 204                 }
 205
 206                 /* fscrypt flag -- ignore */
 207                 if (struct_v >= 6)
 208                         ceph_decode_skip_8(p, end, bad);
 209
 210                 info->fscrypt_auth = NULL;
 211                 info->fscrypt_auth_len = 0;
 212                 info->fscrypt_file = NULL;
 213                 info->fscrypt_file_len = 0;
 214                 if (struct_v >= 7) {
 215                         ceph_decode_32_safe(p, end, info->fscrypt_auth_len, bad);
 216                         if (info->fscrypt_auth_len) {
 217                                 info->fscrypt_auth = kmalloc(info->fscrypt_auth_len,
 218                                                              GFP_KERNEL);
 219                                 if (!info->fscrypt_auth)
 220                                         return -ENOMEM;
 221                                 ceph_decode_copy_safe(p, end, info->fscrypt_auth,
 222                                                       info->fscrypt_auth_len, bad);
 223                         }
 224                         ceph_decode_32_safe(p, end, info->fscrypt_file_len, bad);
 225                         if (info->fscrypt_file_len) {
 226                                 info->fscrypt_file = kmalloc(info->fscrypt_file_len,
 227                                                              GFP_KERNEL);
 228                                 if (!info->fscrypt_file)
 229                                         return -ENOMEM;
 230                                 ceph_decode_copy_safe(p, end, info->fscrypt_file,
 231                                                       info->fscrypt_file_len, bad);
 232                         }
 233                 }
 234                 *p = end;
 235         } else {
 236                 /* legacy (unversioned) struct */
 237                 if (features & CEPH_FEATURE_MDS_INLINE_DATA) {
 238                         ceph_decode_64_safe(p, end, info->inline_version, bad);
 239                         ceph_decode_32_safe(p, end, info->inline_len, bad);
 240                         ceph_decode_need(p, end, info->inline_len, bad);
 241                         info->inline_data = *p;
 242                         *p += info->inline_len;
 243                 } else
 244                         info->inline_version = CEPH_INLINE_NONE;
 245
 246                 if (features & CEPH_FEATURE_MDS_QUOTA) {
 247                         err = parse_reply_info_quota(p, end, info);
 248                         if (err < 0)
 249                                 goto out_bad;
 250                 } else {
 251                         info->max_bytes = 0;
 252                         info->max_files = 0;
 253                 }
 254
 255                 info->pool_ns_len = 0;
 256                 info->pool_ns_data = NULL;
 257                 if (features & CEPH_FEATURE_FS_FILE_LAYOUT_V2) {
 258                         ceph_decode_32_safe(p, end, info->pool_ns_len, bad);
 259                         if (info->pool_ns_len > 0) {
 260                                 ceph_decode_need(p, end, info->pool_ns_len, bad);
 261                                 info->pool_ns_data = *p;
 262                                 *p += info->pool_ns_len;
 263                         }
 264                 }
 265
 266                 if (features & CEPH_FEATURE_FS_BTIME) {
 267                         ceph_decode_need(p, end, sizeof(info->btime), bad);
 268                         ceph_decode_copy(p, &info->btime, sizeof(info->btime));
 269                         ceph_decode_64_safe(p, end, info->change_attr, bad);
 270                 }
 271
 272                 info->dir_pin = -ENODATA;
 273                 /* info->snap_btime and info->rsnaps remain zero */
 274         }
 275         return 0;
 276 bad:
 277         err = -EIO;
 278 out_bad:
 279         return err;
 280 }
 281
 282 static int parse_reply_info_dir(void **p, void *end,
 283                                 struct ceph_mds_reply_dirfrag **dirfrag,
 284                                 u64 features)
 285 {
 286         if (features == (u64)-1) {
 287                 u8 struct_v, struct_compat;
 288                 u32 struct_len;
 289                 ceph_decode_8_safe(p, end, struct_v, bad);
 290                 ceph_decode_8_safe(p, end, struct_compat, bad);
 291                 /* struct_v is expected to be >= 1. we only understand
 292                  * encoding whose struct_compat == 1. */
 293                 if (!struct_v || struct_compat != 1)
 294                         goto bad;
 295                 ceph_decode_32_safe(p, end, struct_len, bad);
 296                 ceph_decode_need(p, end, struct_len, bad);
 297                 end = *p + struct_len;
 298         }
 299
 300         ceph_decode_need(p, end, sizeof(**dirfrag), bad);
 301         *dirfrag = *p;
 302         *p += sizeof(**dirfrag) + sizeof(u32) * le32_to_cpu((*dirfrag)->ndist);
 303         if (unlikely(*p > end))
 304                 goto bad;
 305         if (features == (u64)-1)
 306                 *p = end;
 307         return 0;
 308 bad:
 309         return -EIO;
 310 }
 311
 312 static int parse_reply_info_lease(void **p, void *end,
 313                                   struct ceph_mds_reply_lease **lease,
 314                                   u64 features, u32 *altname_len, u8 **altname)
 315 {
 316         u8 struct_v;
 317         u32 struct_len;
 318         void *lend;
 319
 320         if (features == (u64)-1) {
 321                 u8 struct_compat;
 322
 323                 ceph_decode_8_safe(p, end, struct_v, bad);
 324                 ceph_decode_8_safe(p, end, struct_compat, bad);
 325
 326                 /* struct_v is expected to be >= 1. we only understand
 327                  * encoding whose struct_compat == 1. */
 328                 if (!struct_v || struct_compat != 1)
 329                         goto bad;
 330
 331                 ceph_decode_32_safe(p, end, struct_len, bad);
 332         } else {
 333                 struct_len = sizeof(**lease);
 334                 *altname_len = 0;
 335                 *altname = NULL;
 336         }
 337
 338         lend = *p + struct_len;
 339         ceph_decode_need(p, end, struct_len, bad);
 340         *lease = *p;
 341         *p += sizeof(**lease);
 342
 343         if (features == (u64)-1) {
 344                 if (struct_v >= 2) {
 345                         ceph_decode_32_safe(p, end, *altname_len, bad);
 346                         ceph_decode_need(p, end, *altname_len, bad);
 347                         *altname = *p;
 348                         *p += *altname_len;
 349                 } else {
 350                         *altname = NULL;
 351                         *altname_len = 0;
 352                 }
 353         }
 354         *p = lend;
 355         return 0;
 356 bad:
 357         return -EIO;
 358 }
 359
 360 /*
 361  * parse a normal reply, which may contain a (dir+)dentry and/or a
 362  * target inode.
 363  */
 364 static int parse_reply_info_trace(void **p, void *end,
 365                                   struct ceph_mds_reply_info_parsed *info,
 366                                   u64 features)
 367 {
 368         int err;
 369
 370         if (info->head->is_dentry) {
 371                 err = parse_reply_info_in(p, end, &info->diri, features);
 372                 if (err < 0)
 373                         goto out_bad;
 374
 375                 err = parse_reply_info_dir(p, end, &info->dirfrag, features);
 376                 if (err < 0)
 377                         goto out_bad;
 378
 379                 ceph_decode_32_safe(p, end, info->dname_len, bad);
 380                 ceph_decode_need(p, end, info->dname_len, bad);
 381                 info->dname = *p;
 382                 *p += info->dname_len;
 383
 384                 err = parse_reply_info_lease(p, end, &info->dlease, features,
 385                                              &info->altname_len, &info->altname);
 386                 if (err < 0)
 387                         goto out_bad;
 388         }
 389
 390         if (info->head->is_target) {
 391                 err = parse_reply_info_in(p, end, &info->targeti, features);
 392                 if (err < 0)
 393                         goto out_bad;
 394         }
 395
 396         if (unlikely(*p != end))
 397                 goto bad;
 398         return 0;
 399
 400 bad:
 401         err = -EIO;
 402 out_bad:
 403         pr_err("problem parsing mds trace %d\n", err);
 404         return err;
 405 }
 406
 407 /*
 408  * parse readdir results
 409  */
 410 static int parse_reply_info_readdir(void **p, void *end,
 411                                     struct ceph_mds_request *req,
 412                                     u64 features)
 413 {
 414         struct ceph_mds_reply_info_parsed *info = &req->r_reply_info;
 415         struct ceph_client *cl = req->r_mdsc->fsc->client;
 416         u32 num, i = 0;
 417         int err;
 418
 419         err = parse_reply_info_dir(p, end, &info->dir_dir, features);
 420         if (err < 0)
 421                 goto out_bad;
 422
 423         ceph_decode_need(p, end, sizeof(num) + 2, bad);
 424         num = ceph_decode_32(p);
 425         {
 426                 u16 flags = ceph_decode_16(p);
 427                 info->dir_end = !!(flags & CEPH_READDIR_FRAG_END);
 428                 info->dir_complete = !!(flags & CEPH_READDIR_FRAG_COMPLETE);
 429                 info->hash_order = !!(flags & CEPH_READDIR_HASH_ORDER);
 430                 info->offset_hash = !!(flags & CEPH_READDIR_OFFSET_HASH);
 431         }
 432         if (num == 0)
 433                 goto done;
 434
 435         BUG_ON(!info->dir_entries);
 436         if ((unsigned long)(info->dir_entries + num) >
 437             (unsigned long)info->dir_entries + info->dir_buf_size) {
 438                 pr_err_client(cl, "dir contents are larger than expected\n");
 439                 WARN_ON(1);
 440                 goto bad;
 441         }
 442
 443         info->dir_nr = num;
 444         while (num) {
 445                 struct inode *inode = d_inode(req->r_dentry);
 446                 struct ceph_inode_info *ci = ceph_inode(inode);
 447                 struct ceph_mds_reply_dir_entry *rde = info->dir_entries + i;
 448                 struct fscrypt_str tname = FSTR_INIT(NULL, 0);
 449                 struct fscrypt_str oname = FSTR_INIT(NULL, 0);
 450                 struct ceph_fname fname;
 451                 u32 altname_len, _name_len;
 452                 u8 *altname, *_name;
 453
 454                 /* dentry */
 455                 ceph_decode_32_safe(p, end, _name_len, bad);
 456                 ceph_decode_need(p, end, _name_len, bad);
 457                 _name = *p;
 458                 *p += _name_len;
 459                 doutc(cl, "parsed dir dname '%.*s'\n", _name_len, _name);
 460
 461                 if (info->hash_order)
 462                         rde->raw_hash = ceph_str_hash(ci->i_dir_layout.dl_dir_hash,
 463                                                       _name, _name_len);
 464
 465                 /* dentry lease */
 466                 err = parse_reply_info_lease(p, end, &rde->lease, features,
 467                                              &altname_len, &altname);
 468                 if (err)
 469                         goto out_bad;
 470
 471                 /*
 472                  * Try to dencrypt the dentry names and update them
 473                  * in the ceph_mds_reply_dir_entry struct.
 474                  */
 475                 fname.dir = inode;
 476                 fname.name = _name;
 477                 fname.name_len = _name_len;
 478                 fname.ctext = altname;
 479                 fname.ctext_len = altname_len;
 480                 /*
 481                  * The _name_len maybe larger than altname_len, such as
 482                  * when the human readable name length is in range of
 483                  * (CEPH_NOHASH_NAME_MAX, CEPH_NOHASH_NAME_MAX + SHA256_DIGEST_SIZE),
 484                  * then the copy in ceph_fname_to_usr will corrupt the
 485                  * data if there has no encryption key.
 486                  *
 487                  * Just set the no_copy flag and then if there has no
 488                  * encryption key the oname.name will be assigned to
 489                  * _name always.
 490                  */
 491                 fname.no_copy = true;
 492                 if (altname_len == 0) {
 493                         /*
 494                          * Set tname to _name, and this will be used
 495                          * to do the base64_decode in-place. It's
 496                          * safe because the decoded string should
 497                          * always be shorter, which is 3/4 of origin
 498                          * string.
 499                          */
 500                         tname.name = _name;
 501
 502                         /*
 503                          * Set oname to _name too, and this will be
 504                          * used to do the dencryption in-place.
 505                          */
 506                         oname.name = _name;
 507                         oname.len = _name_len;
 508                 } else {
 509                         /*
 510                          * This will do the decryption only in-place
 511                          * from altname cryptext directly.
 512                          */
 513                         oname.name = altname;
 514                         oname.len = altname_len;
 515                 }
 516                 rde->is_nokey = false;
 517                 err = ceph_fname_to_usr(&fname, &tname, &oname, &rde->is_nokey);
 518                 if (err) {
 519                         pr_err_client(cl, "unable to decode %.*s, got %d\n",
 520                                       _name_len, _name, err);
 521                         goto out_bad;
 522                 }
 523                 rde->name = oname.name;
 524                 rde->name_len = oname.len;
 525
 526                 /* inode */
 527                 err = parse_reply_info_in(p, end, &rde->inode, features);
 528                 if (err < 0)
 529                         goto out_bad;
 530                 /* ceph_readdir_prepopulate() will update it */
 531                 rde->offset = 0;
 532                 i++;
 533                 num--;
 534         }
 535
 536 done:
 537         /* Skip over any unrecognized fields */
 538         *p = end;
 539         return 0;
 540
 541 bad:
 542         err = -EIO;
 543 out_bad:
 544         pr_err_client(cl, "problem parsing dir contents %d\n", err);
 545         return err;
 546 }
 547
 548 /*
 549  * parse fcntl F_GETLK results
 550  */
 551 static int parse_reply_info_filelock(void **p, void *end,
 552                                      struct ceph_mds_reply_info_parsed *info,
 553                                      u64 features)
 554 {
 555         if (*p + sizeof(*info->filelock_reply) > end)
 556                 goto bad;
 557
 558         info->filelock_reply = *p;
 559
 560         /* Skip over any unrecognized fields */
 561         *p = end;
 562         return 0;
 563 bad:
 564         return -EIO;
 565 }
 566
 567
 568 #if BITS_PER_LONG == 64
 569
 570 #define DELEGATED_INO_AVAILABLE         xa_mk_value(1)
 571
 572 static int ceph_parse_deleg_inos(void **p, void *end,
 573                                  struct ceph_mds_session *s)
 574 {
 575         struct ceph_client *cl = s->s_mdsc->fsc->client;
 576         u32 sets;
 577
 578         ceph_decode_32_safe(p, end, sets, bad);
 579         doutc(cl, "got %u sets of delegated inodes\n", sets);
 580         while (sets--) {
 581                 u64 start, len;
 582
 583                 ceph_decode_64_safe(p, end, start, bad);
 584                 ceph_decode_64_safe(p, end, len, bad);
 585
 586                 /* Don't accept a delegation of system inodes */
 587                 if (start < CEPH_INO_SYSTEM_BASE) {
 588                         pr_warn_ratelimited_client(cl,
 589                                 "ignoring reserved inode range delegation (start=0x%llx len=0x%llx)\n",
 590                                 start, len);
 591                         continue;
 592                 }
 593                 while (len--) {
 594                         int err = xa_insert(&s->s_delegated_inos, start++,
 595                                             DELEGATED_INO_AVAILABLE,
 596                                             GFP_KERNEL);
 597                         if (!err) {
 598                                 doutc(cl, "added delegated inode 0x%llx\n", start - 1);
 599                         } else if (err == -EBUSY) {
 600                                 pr_warn_client(cl,
 601                                         "MDS delegated inode 0x%llx more than once.\n",
 602                                         start - 1);
 603                         } else {
 604                                 return err;
 605                         }
 606                 }
 607         }
 608         return 0;
 609 bad:
 610         return -EIO;
 611 }
 612
 613 u64 ceph_get_deleg_ino(struct ceph_mds_session *s)
 614 {
 615         unsigned long ino;
 616         void *val;
 617
 618         xa_for_each(&s->s_delegated_inos, ino, val) {
 619                 val = xa_erase(&s->s_delegated_inos, ino);
 620                 if (val == DELEGATED_INO_AVAILABLE)
 621                         return ino;
 622         }
 623         return 0;
 624 }
 625
 626 int ceph_restore_deleg_ino(struct ceph_mds_session *s, u64 ino)
 627 {
 628         return xa_insert(&s->s_delegated_inos, ino, DELEGATED_INO_AVAILABLE,
 629                          GFP_KERNEL);
 630 }
 631 #else /* BITS_PER_LONG == 64 */
 632 /*
 633  * FIXME: xarrays can't handle 64-bit indexes on a 32-bit arch. For now, just
 634  * ignore delegated_inos on 32 bit arch. Maybe eventually add xarrays for top
 635  * and bottom words?
 636  */
 637 static int ceph_parse_deleg_inos(void **p, void *end,
 638                                  struct ceph_mds_session *s)
 639 {
 640         u32 sets;
 641
 642         ceph_decode_32_safe(p, end, sets, bad);
 643         if (sets)
 644                 ceph_decode_skip_n(p, end, sets * 2 * sizeof(__le64), bad);
 645         return 0;
 646 bad:
 647         return -EIO;
 648 }
 649
 650 u64 ceph_get_deleg_ino(struct ceph_mds_session *s)
 651 {
 652         return 0;
 653 }
 654
 655 int ceph_restore_deleg_ino(struct ceph_mds_session *s, u64 ino)
 656 {
 657         return 0;
 658 }
 659 #endif /* BITS_PER_LONG == 64 */
 660
 661 /*
 662  * parse create results
 663  */
 664 static int parse_reply_info_create(void **p, void *end,
 665                                   struct ceph_mds_reply_info_parsed *info,
 666                                   u64 features, struct ceph_mds_session *s)
 667 {
 668         int ret;
 669
 670         if (features == (u64)-1 ||
 671             (features & CEPH_FEATURE_REPLY_CREATE_INODE)) {
 672                 if (*p == end) {
 673                         /* Malformed reply? */
 674                         info->has_create_ino = false;
 675                 } else if (test_bit(CEPHFS_FEATURE_DELEG_INO, &s->s_features)) {
 676                         info->has_create_ino = true;
 677                         /* struct_v, struct_compat, and len */
 678                         ceph_decode_skip_n(p, end, 2 + sizeof(u32), bad);
 679                         ceph_decode_64_safe(p, end, info->ino, bad);
 680                         ret = ceph_parse_deleg_inos(p, end, s);
 681                         if (ret)
 682                                 return ret;
 683                 } else {
 684                         /* legacy */
 685                         ceph_decode_64_safe(p, end, info->ino, bad);
 686                         info->has_create_ino = true;
 687                 }
 688         } else {
 689                 if (*p != end)
 690                         goto bad;
 691         }
 692
 693         /* Skip over any unrecognized fields */
 694         *p = end;
 695         return 0;
 696 bad:
 697         return -EIO;
 698 }
 699
 700 static int parse_reply_info_getvxattr(void **p, void *end,
 701                                       struct ceph_mds_reply_info_parsed *info,
 702                                       u64 features)
 703 {
 704         u32 value_len;
 705
 706         ceph_decode_skip_8(p, end, bad); /* skip current version: 1 */
 707         ceph_decode_skip_8(p, end, bad); /* skip first version: 1 */
 708         ceph_decode_skip_32(p, end, bad); /* skip payload length */
 709
 710         ceph_decode_32_safe(p, end, value_len, bad);
 711
 712         if (value_len == end - *p) {
 713           info->xattr_info.xattr_value = *p;
 714           info->xattr_info.xattr_value_len = value_len;
 715           *p = end;
 716           return value_len;
 717         }
 718 bad:
 719         return -EIO;
 720 }
 721
 722 /*
 723  * parse extra results
 724  */
 725 static int parse_reply_info_extra(void **p, void *end,
 726                                   struct ceph_mds_request *req,
 727                                   u64 features, struct ceph_mds_session *s)
 728 {
 729         struct ceph_mds_reply_info_parsed *info = &req->r_reply_info;
 730         u32 op = le32_to_cpu(info->head->op);
 731
 732         if (op == CEPH_MDS_OP_GETFILELOCK)
 733                 return parse_reply_info_filelock(p, end, info, features);
 734         else if (op == CEPH_MDS_OP_READDIR || op == CEPH_MDS_OP_LSSNAP)
 735                 return parse_reply_info_readdir(p, end, req, features);
 736         else if (op == CEPH_MDS_OP_CREATE)
 737                 return parse_reply_info_create(p, end, info, features, s);
 738         else if (op == CEPH_MDS_OP_GETVXATTR)
 739                 return parse_reply_info_getvxattr(p, end, info, features);
 740         else
 741                 return -EIO;
 742 }
 743
 744 /*
 745  * parse entire mds reply
 746  */
 747 static int parse_reply_info(struct ceph_mds_session *s, struct ceph_msg *msg,
 748                             struct ceph_mds_request *req, u64 features)
 749 {
 750         struct ceph_mds_reply_info_parsed *info = &req->r_reply_info;
 751         struct ceph_client *cl = s->s_mdsc->fsc->client;
 752         void *p, *end;
 753         u32 len;
 754         int err;
 755
 756         info->head = msg->front.iov_base;
 757         p = msg->front.iov_base + sizeof(struct ceph_mds_reply_head);
 758         end = p + msg->front.iov_len - sizeof(struct ceph_mds_reply_head);
 759
 760         /* trace */
 761         ceph_decode_32_safe(&p, end, len, bad);
 762         if (len > 0) {
 763                 ceph_decode_need(&p, end, len, bad);
 764                 err = parse_reply_info_trace(&p, p+len, info, features);
 765                 if (err < 0)
 766                         goto out_bad;
 767         }
 768
 769         /* extra */
 770         ceph_decode_32_safe(&p, end, len, bad);
 771         if (len > 0) {
 772                 ceph_decode_need(&p, end, len, bad);
 773                 err = parse_reply_info_extra(&p, p+len, req, features, s);
 774                 if (err < 0)
 775                         goto out_bad;
 776         }
 777
 778         /* snap blob */
 779         ceph_decode_32_safe(&p, end, len, bad);
 780         info->snapblob_len = len;
 781         info->snapblob = p;
 782         p += len;
 783
 784         if (p != end)
 785                 goto bad;
 786         return 0;
 787
 788 bad:
 789         err = -EIO;
 790 out_bad:
 791         pr_err_client(cl, "mds parse_reply err %d\n", err);
 792         ceph_msg_dump(msg);
 793         return err;
 794 }
 795
 796 static void destroy_reply_info(struct ceph_mds_reply_info_parsed *info)
 797 {
 798         int i;
 799
 800         kfree(info->diri.fscrypt_auth);
 801         kfree(info->diri.fscrypt_file);
 802         kfree(info->targeti.fscrypt_auth);
 803         kfree(info->targeti.fscrypt_file);
 804         if (!info->dir_entries)
 805                 return;
 806
 807         for (i = 0; i < info->dir_nr; i++) {
 808                 struct ceph_mds_reply_dir_entry *rde = info->dir_entries + i;
 809
 810                 kfree(rde->inode.fscrypt_auth);
 811                 kfree(rde->inode.fscrypt_file);
 812         }
 813         free_pages((unsigned long)info->dir_entries, get_order(info->dir_buf_size));
 814 }
 815
 816 /*
 817  * In async unlink case the kclient won't wait for the first reply
 818  * from MDS and just drop all the links and unhash the dentry and then
 819  * succeeds immediately.
 820  *
 821  * For any new create/link/rename,etc requests followed by using the
 822  * same file names we must wait for the first reply of the inflight
 823  * unlink request, or the MDS possibly will fail these following
 824  * requests with -EEXIST if the inflight async unlink request was
 825  * delayed for some reasons.
 826  *
 827  * And the worst case is that for the none async openc request it will
 828  * successfully open the file if the CDentry hasn't been unlinked yet,
 829  * but later the previous delayed async unlink request will remove the
 830  * CDenty. That means the just created file is possiblly deleted later
 831  * by accident.
 832  *
 833  * We need to wait for the inflight async unlink requests to finish
 834  * when creating new files/directories by using the same file names.
 835  */
 836 int ceph_wait_on_conflict_unlink(struct dentry *dentry)
 837 {
 838         struct ceph_fs_client *fsc = ceph_sb_to_fs_client(dentry->d_sb);
 839         struct ceph_client *cl = fsc->client;
 840         struct dentry *pdentry = dentry->d_parent;
 841         struct dentry *udentry, *found = NULL;
 842         struct ceph_dentry_info *di;
 843         struct qstr dname;
 844         u32 hash = dentry->d_name.hash;
 845         int err;
 846
 847         dname.name = dentry->d_name.name;
 848         dname.len = dentry->d_name.len;
 849
 850         rcu_read_lock();
 851         hash_for_each_possible_rcu(fsc->async_unlink_conflict, di,
 852                                    hnode, hash) {
 853                 udentry = di->dentry;
 854
 855                 spin_lock(&udentry->d_lock);
 856                 if (udentry->d_name.hash != hash)
 857                         goto next;
 858                 if (unlikely(udentry->d_parent != pdentry))
 859                         goto next;
 860                 if (!hash_hashed(&di->hnode))
 861                         goto next;
 862
 863                 if (!test_bit(CEPH_DENTRY_ASYNC_UNLINK_BIT, &di->flags))
 864                         pr_warn_client(cl, "dentry %p:%pd async unlink bit is not set\n",
 865                                        dentry, dentry);
 866
 867                 if (!d_same_name(udentry, pdentry, &dname))
 868                         goto next;
 869
 870                 found = dget_dlock(udentry);
 871                 spin_unlock(&udentry->d_lock);
 872                 break;
 873 next:
 874                 spin_unlock(&udentry->d_lock);
 875         }
 876         rcu_read_unlock();
 877
 878         if (likely(!found))
 879                 return 0;
 880
 881         doutc(cl, "dentry %p:%pd conflict with old %p:%pd\n", dentry, dentry,
 882               found, found);
 883
 884         err = wait_on_bit(&di->flags, CEPH_DENTRY_ASYNC_UNLINK_BIT,
 885                           TASK_KILLABLE);
 886         dput(found);
 887         return err;
 888 }
 889
 890
 891 /*
 892  * sessions
 893  */
 894 const char *ceph_session_state_name(int s)
 895 {
 896         switch (s) {
 897         case CEPH_MDS_SESSION_NEW: return "new";
 898         case CEPH_MDS_SESSION_OPENING: return "opening";
 899         case CEPH_MDS_SESSION_OPEN: return "open";
 900         case CEPH_MDS_SESSION_HUNG: return "hung";
 901         case CEPH_MDS_SESSION_CLOSING: return "closing";
 902         case CEPH_MDS_SESSION_CLOSED: return "closed";
 903         case CEPH_MDS_SESSION_RESTARTING: return "restarting";
 904         case CEPH_MDS_SESSION_RECONNECTING: return "reconnecting";
 905         case CEPH_MDS_SESSION_REJECTED: return "rejected";
 906         default: return "???";
 907         }
 908 }
 909
 910 struct ceph_mds_session *ceph_get_mds_session(struct ceph_mds_session *s)
 911 {
 912         if (refcount_inc_not_zero(&s->s_ref))
 913                 return s;
 914         return NULL;
 915 }
 916
 917 void ceph_put_mds_session(struct ceph_mds_session *s)
 918 {
 919         if (IS_ERR_OR_NULL(s))
 920                 return;
 921
 922         if (refcount_dec_and_test(&s->s_ref)) {
 923                 if (s->s_auth.authorizer)
 924                         ceph_auth_destroy_authorizer(s->s_auth.authorizer);
 925                 WARN_ON(mutex_is_locked(&s->s_mutex));
 926                 xa_destroy(&s->s_delegated_inos);
 927                 kfree(s);
 928         }
 929 }
 930
 931 /*
 932  * called under mdsc->mutex
 933  */
 934 struct ceph_mds_session *__ceph_lookup_mds_session(struct ceph_mds_client *mdsc,
 935                                                    int mds)
 936 {
 937         if (mds >= mdsc->max_sessions || !mdsc->sessions[mds])
 938                 return NULL;
 939         return ceph_get_mds_session(mdsc->sessions[mds]);
 940 }
 941
 942 static bool __have_session(struct ceph_mds_client *mdsc, int mds)
 943 {
 944         if (mds >= mdsc->max_sessions || !mdsc->sessions[mds])
 945                 return false;
 946         else
 947                 return true;
 948 }
 949
 950 static int __verify_registered_session(struct ceph_mds_client *mdsc,
 951                                        struct ceph_mds_session *s)
 952 {
 953         if (s->s_mds >= mdsc->max_sessions ||
 954             mdsc->sessions[s->s_mds] != s)
 955                 return -ENOENT;
 956         return 0;
 957 }
 958
 959 /*
 960  * create+register a new session for given mds.
 961  * called under mdsc->mutex.
 962  */
 963 static struct ceph_mds_session *register_session(struct ceph_mds_client *mdsc,
 964                                                  int mds)
 965 {
 966         struct ceph_client *cl = mdsc->fsc->client;
 967         struct ceph_mds_session *s;
 968
 969         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_FENCE_IO)
 970                 return ERR_PTR(-EIO);
 971
 972         if (mds >= mdsc->mdsmap->possible_max_rank)
 973                 return ERR_PTR(-EINVAL);
 974
 975         s = kzalloc(sizeof(*s), GFP_NOFS);
 976         if (!s)
 977                 return ERR_PTR(-ENOMEM);
 978
 979         if (mds >= mdsc->max_sessions) {
 980                 int newmax = 1 << get_count_order(mds + 1);
 981                 struct ceph_mds_session **sa;
 982
 983                 doutc(cl, "realloc to %d\n", newmax);
 984                 sa = kcalloc(newmax, sizeof(void *), GFP_NOFS);
 985                 if (!sa)
 986                         goto fail_realloc;
 987                 if (mdsc->sessions) {
 988                         memcpy(sa, mdsc->sessions,
 989                                mdsc->max_sessions * sizeof(void *));
 990                         kfree(mdsc->sessions);
 991                 }
 992                 mdsc->sessions = sa;
 993                 mdsc->max_sessions = newmax;
 994         }
 995
 996         doutc(cl, "mds%d\n", mds);
 997         s->s_mdsc = mdsc;
 998         s->s_mds = mds;
 999         s->s_state = CEPH_MDS_SESSION_NEW;
1000         mutex_init(&s->s_mutex);
1001
1002         ceph_con_init(&s->s_con, s, &mds_con_ops, &mdsc->fsc->client->msgr);
1003
1004         atomic_set(&s->s_cap_gen, 1);
1005         s->s_cap_ttl = jiffies - 1;
1006
1007         spin_lock_init(&s->s_cap_lock);
1008         INIT_LIST_HEAD(&s->s_caps);
1009         refcount_set(&s->s_ref, 1);
1010         INIT_LIST_HEAD(&s->s_waiting);
1011         INIT_LIST_HEAD(&s->s_unsafe);
1012         xa_init(&s->s_delegated_inos);
1013         INIT_LIST_HEAD(&s->s_cap_releases);
1014         INIT_WORK(&s->s_cap_release_work, ceph_cap_release_work);
1015
1016         INIT_LIST_HEAD(&s->s_cap_dirty);
1017         INIT_LIST_HEAD(&s->s_cap_flushing);
1018
1019         mdsc->sessions[mds] = s;
1020         atomic_inc(&mdsc->num_sessions);
1021         refcount_inc(&s->s_ref);  /* one ref to sessions[], one to caller */
1022
1023         ceph_con_open(&s->s_con, CEPH_ENTITY_TYPE_MDS, mds,
1024                       ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
1025
1026         return s;
1027
1028 fail_realloc:
1029         kfree(s);
1030         return ERR_PTR(-ENOMEM);
1031 }
1032
1033 /*
1034  * called under mdsc->mutex
1035  */
1036 static void __unregister_session(struct ceph_mds_client *mdsc,
1037                                struct ceph_mds_session *s)
1038 {
1039         doutc(mdsc->fsc->client, "mds%d %p\n", s->s_mds, s);
1040         BUG_ON(mdsc->sessions[s->s_mds] != s);
1041         mdsc->sessions[s->s_mds] = NULL;
1042         ceph_con_close(&s->s_con);
1043         ceph_put_mds_session(s);
1044         atomic_dec(&mdsc->num_sessions);
1045 }
1046
1047 /*
1048  * drop session refs in request.
1049  *
1050  * should be last request ref, or hold mdsc->mutex
1051  */
1052 static void put_request_session(struct ceph_mds_request *req)
1053 {
1054         if (req->r_session) {
1055                 ceph_put_mds_session(req->r_session);
1056                 req->r_session = NULL;
1057         }
1058 }
1059
1060 void ceph_mdsc_iterate_sessions(struct ceph_mds_client *mdsc,
1061                                 void (*cb)(struct ceph_mds_session *),
1062                                 bool check_state)
1063 {
1064         int mds;
1065
1066         mutex_lock(&mdsc->mutex);
1067         for (mds = 0; mds < mdsc->max_sessions; ++mds) {
1068                 struct ceph_mds_session *s;
1069
1070                 s = __ceph_lookup_mds_session(mdsc, mds);
1071                 if (!s)
1072                         continue;
1073
1074                 if (check_state && !check_session_state(s)) {
1075                         ceph_put_mds_session(s);
1076                         continue;
1077                 }
1078
1079                 mutex_unlock(&mdsc->mutex);
1080                 cb(s);
1081                 ceph_put_mds_session(s);
1082                 mutex_lock(&mdsc->mutex);
1083         }
1084         mutex_unlock(&mdsc->mutex);
1085 }
1086
1087 void ceph_mdsc_release_request(struct kref *kref)
1088 {
1089         struct ceph_mds_request *req = container_of(kref,
1090                                                     struct ceph_mds_request,
1091                                                     r_kref);
1092         ceph_mdsc_release_dir_caps_async(req);
1093         destroy_reply_info(&req->r_reply_info);
1094         if (req->r_request)
1095                 ceph_msg_put(req->r_request);
1096         if (req->r_reply)
1097                 ceph_msg_put(req->r_reply);
1098         if (req->r_inode) {
1099                 ceph_put_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
1100                 iput(req->r_inode);
1101         }
1102         if (req->r_parent) {
1103                 ceph_put_cap_refs(ceph_inode(req->r_parent), CEPH_CAP_PIN);
1104                 iput(req->r_parent);
1105         }
1106         iput(req->r_target_inode);
1107         iput(req->r_new_inode);
1108         if (req->r_dentry)
1109                 dput(req->r_dentry);
1110         if (req->r_old_dentry)
1111                 dput(req->r_old_dentry);
1112         if (req->r_old_dentry_dir) {
1113                 /*
1114                  * track (and drop pins for) r_old_dentry_dir
1115                  * separately, since r_old_dentry's d_parent may have
1116                  * changed between the dir mutex being dropped and
1117                  * this request being freed.
1118                  */
1119                 ceph_put_cap_refs(ceph_inode(req->r_old_dentry_dir),
1120                                   CEPH_CAP_PIN);
1121                 iput(req->r_old_dentry_dir);
1122         }
1123         kfree(req->r_path1);
1124         kfree(req->r_path2);
1125         put_cred(req->r_cred);
1126         if (req->r_mnt_idmap)
1127                 mnt_idmap_put(req->r_mnt_idmap);
1128         if (req->r_pagelist)
1129                 ceph_pagelist_release(req->r_pagelist);
1130         kfree(req->r_fscrypt_auth);
1131         kfree(req->r_altname);
1132         put_request_session(req);
1133         ceph_unreserve_caps(req->r_mdsc, &req->r_caps_reservation);
1134         WARN_ON_ONCE(!list_empty(&req->r_wait));
1135         kmem_cache_free(ceph_mds_request_cachep, req);
1136 }
1137
1138 DEFINE_RB_FUNCS(request, struct ceph_mds_request, r_tid, r_node)
1139
1140 /*
1141  * lookup session, bump ref if found.
1142  *
1143  * called under mdsc->mutex.
1144  */
1145 static struct ceph_mds_request *
1146 lookup_get_request(struct ceph_mds_client *mdsc, u64 tid)
1147 {
1148         struct ceph_mds_request *req;
1149
1150         req = lookup_request(&mdsc->request_tree, tid);
1151         if (req)
1152                 ceph_mdsc_get_request(req);
1153
1154         return req;
1155 }
1156
1157 /*
1158  * Register an in-flight request, and assign a tid.  Link to directory
1159  * are modifying (if any).
1160  *
1161  * Called under mdsc->mutex.
1162  */
1163 static void __register_request(struct ceph_mds_client *mdsc,
1164                                struct ceph_mds_request *req,
1165                                struct inode *dir)
1166 {
1167         struct ceph_client *cl = mdsc->fsc->client;
1168         int ret = 0;
1169
1170         req->r_tid = ++mdsc->last_tid;
1171         if (req->r_num_caps) {
1172                 ret = ceph_reserve_caps(mdsc, &req->r_caps_reservation,
1173                                         req->r_num_caps);
1174                 if (ret < 0) {
1175                         pr_err_client(cl, "%p failed to reserve caps: %d\n",
1176                                       req, ret);
1177                         /* set req->r_err to fail early from __do_request */
1178                         req->r_err = ret;
1179                         return;
1180                 }
1181         }
1182         doutc(cl, "%p tid %lld\n", req, req->r_tid);
1183         ceph_mdsc_get_request(req);
1184         insert_request(&mdsc->request_tree, req);
1185
1186         req->r_cred = get_current_cred();
1187         if (!req->r_mnt_idmap)
1188                 req->r_mnt_idmap = &nop_mnt_idmap;
1189
1190         if (mdsc->oldest_tid == 0 && req->r_op != CEPH_MDS_OP_SETFILELOCK)
1191                 mdsc->oldest_tid = req->r_tid;
1192
1193         if (dir) {
1194                 struct ceph_inode_info *ci = ceph_inode(dir);
1195
1196                 ihold(dir);
1197                 req->r_unsafe_dir = dir;
1198                 spin_lock(&ci->i_unsafe_lock);
1199                 list_add_tail(&req->r_unsafe_dir_item, &ci->i_unsafe_dirops);
1200                 spin_unlock(&ci->i_unsafe_lock);
1201         }
1202 }
1203
1204 static void __unregister_request(struct ceph_mds_client *mdsc,
1205                                  struct ceph_mds_request *req)
1206 {
1207         doutc(mdsc->fsc->client, "%p tid %lld\n", req, req->r_tid);
1208
1209         /* Never leave an unregistered request on an unsafe list! */
1210         list_del_init(&req->r_unsafe_item);
1211
1212         if (req->r_tid == mdsc->oldest_tid) {
1213                 struct rb_node *p = rb_next(&req->r_node);
1214                 mdsc->oldest_tid = 0;
1215                 while (p) {
1216                         struct ceph_mds_request *next_req =
1217                                 rb_entry(p, struct ceph_mds_request, r_node);
1218                         if (next_req->r_op != CEPH_MDS_OP_SETFILELOCK) {
1219                                 mdsc->oldest_tid = next_req->r_tid;
1220                                 break;
1221                         }
1222                         p = rb_next(p);
1223                 }
1224         }
1225
1226         erase_request(&mdsc->request_tree, req);
1227
1228         if (req->r_unsafe_dir) {
1229                 struct ceph_inode_info *ci = ceph_inode(req->r_unsafe_dir);
1230                 spin_lock(&ci->i_unsafe_lock);
1231                 list_del_init(&req->r_unsafe_dir_item);
1232                 spin_unlock(&ci->i_unsafe_lock);
1233         }
1234         if (req->r_target_inode &&
1235             test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
1236                 struct ceph_inode_info *ci = ceph_inode(req->r_target_inode);
1237                 spin_lock(&ci->i_unsafe_lock);
1238                 list_del_init(&req->r_unsafe_target_item);
1239                 spin_unlock(&ci->i_unsafe_lock);
1240         }
1241
1242         if (req->r_unsafe_dir) {
1243                 iput(req->r_unsafe_dir);
1244                 req->r_unsafe_dir = NULL;
1245         }
1246
1247         complete_all(&req->r_safe_completion);
1248
1249         ceph_mdsc_put_request(req);
1250 }
1251
1252 /*
1253  * Walk back up the dentry tree until we hit a dentry representing a
1254  * non-snapshot inode. We do this using the rcu_read_lock (which must be held
1255  * when calling this) to ensure that the objects won't disappear while we're
1256  * working with them. Once we hit a candidate dentry, we attempt to take a
1257  * reference to it, and return that as the result.
1258  */
1259 static struct inode *get_nonsnap_parent(struct dentry *dentry)
1260 {
1261         struct inode *inode = NULL;
1262
1263         while (dentry && !IS_ROOT(dentry)) {
1264                 inode = d_inode_rcu(dentry);
1265                 if (!inode || ceph_snap(inode) == CEPH_NOSNAP)
1266                         break;
1267                 dentry = dentry->d_parent;
1268         }
1269         if (inode)
1270                 inode = igrab(inode);
1271         return inode;
1272 }
1273
1274 /*
1275  * Choose mds to send request to next.  If there is a hint set in the
1276  * request (e.g., due to a prior forward hint from the mds), use that.
1277  * Otherwise, consult frag tree and/or caps to identify the
1278  * appropriate mds.  If all else fails, choose randomly.
1279  *
1280  * Called under mdsc->mutex.
1281  */
1282 static int __choose_mds(struct ceph_mds_client *mdsc,
1283                         struct ceph_mds_request *req,
1284                         bool *random)
1285 {
1286         struct inode *inode;
1287         struct ceph_inode_info *ci;
1288         struct ceph_cap *cap;
1289         int mode = req->r_direct_mode;
1290         int mds = -1;
1291         u32 hash = req->r_direct_hash;
1292         bool is_hash = test_bit(CEPH_MDS_R_DIRECT_IS_HASH, &req->r_req_flags);
1293         struct ceph_client *cl = mdsc->fsc->client;
1294
1295         if (random)
1296                 *random = false;
1297
1298         /*
1299          * is there a specific mds we should try?  ignore hint if we have
1300          * no session and the mds is not up (active or recovering).
1301          */
1302         if (req->r_resend_mds >= 0 &&
1303             (__have_session(mdsc, req->r_resend_mds) ||
1304              ceph_mdsmap_get_state(mdsc->mdsmap, req->r_resend_mds) > 0)) {
1305                 doutc(cl, "using resend_mds mds%d\n", req->r_resend_mds);
1306                 return req->r_resend_mds;
1307         }
1308
1309         if (mode == USE_RANDOM_MDS)
1310                 goto random;
1311
1312         inode = NULL;
1313         if (req->r_inode) {
1314                 if (ceph_snap(req->r_inode) != CEPH_SNAPDIR) {
1315                         inode = req->r_inode;
1316                         ihold(inode);
1317                 } else {
1318                         /* req->r_dentry is non-null for LSSNAP request */
1319                         rcu_read_lock();
1320                         inode = get_nonsnap_parent(req->r_dentry);
1321                         rcu_read_unlock();
1322                         doutc(cl, "using snapdir's parent %p %llx.%llx\n",
1323                               inode, ceph_vinop(inode));
1324                 }
1325         } else if (req->r_dentry) {
1326                 /* ignore race with rename; old or new d_parent is okay */
1327                 struct dentry *parent;
1328                 struct inode *dir;
1329
1330                 rcu_read_lock();
1331                 parent = READ_ONCE(req->r_dentry->d_parent);
1332                 dir = req->r_parent ? : d_inode_rcu(parent);
1333
1334                 if (!dir || dir->i_sb != mdsc->fsc->sb) {
1335                         /*  not this fs or parent went negative */
1336                         inode = d_inode(req->r_dentry);
1337                         if (inode)
1338                                 ihold(inode);
1339                 } else if (ceph_snap(dir) != CEPH_NOSNAP) {
1340                         /* direct snapped/virtual snapdir requests
1341                          * based on parent dir inode */
1342                         inode = get_nonsnap_parent(parent);
1343                         doutc(cl, "using nonsnap parent %p %llx.%llx\n",
1344                               inode, ceph_vinop(inode));
1345                 } else {
1346                         /* dentry target */
1347                         inode = d_inode(req->r_dentry);
1348                         if (!inode || mode == USE_AUTH_MDS) {
1349                                 /* dir + name */
1350                                 inode = igrab(dir);
1351                                 hash = ceph_dentry_hash(dir, req->r_dentry);
1352                                 is_hash = true;
1353                         } else {
1354                                 ihold(inode);
1355                         }
1356                 }
1357                 rcu_read_unlock();
1358         }
1359
1360         if (!inode)
1361                 goto random;
1362
1363         doutc(cl, "%p %llx.%llx is_hash=%d (0x%x) mode %d\n", inode,
1364               ceph_vinop(inode), (int)is_hash, hash, mode);
1365         ci = ceph_inode(inode);
1366
1367         if (is_hash && S_ISDIR(inode->i_mode)) {
1368                 struct ceph_inode_frag frag;
1369                 int found;
1370
1371                 ceph_choose_frag(ci, hash, &frag, &found);
1372                 if (found) {
1373                         if (mode == USE_ANY_MDS && frag.ndist > 0) {
1374                                 u8 r;
1375
1376                                 /* choose a random replica */
1377                                 get_random_bytes(&r, 1);
1378                                 r %= frag.ndist;
1379                                 mds = frag.dist[r];
1380                                 doutc(cl, "%p %llx.%llx frag %u mds%d (%d/%d)\n",
1381                                       inode, ceph_vinop(inode), frag.frag,
1382                                       mds, (int)r, frag.ndist);
1383                                 if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
1384                                     CEPH_MDS_STATE_ACTIVE &&
1385                                     !ceph_mdsmap_is_laggy(mdsc->mdsmap, mds))
1386                                         goto out;
1387                         }
1388
1389                         /* since this file/dir wasn't known to be
1390                          * replicated, then we want to look for the
1391                          * authoritative mds. */
1392                         if (frag.mds >= 0) {
1393                                 /* choose auth mds */
1394                                 mds = frag.mds;
1395                                 doutc(cl, "%p %llx.%llx frag %u mds%d (auth)\n",
1396                                       inode, ceph_vinop(inode), frag.frag, mds);
1397                                 if (ceph_mdsmap_get_state(mdsc->mdsmap, mds) >=
1398                                     CEPH_MDS_STATE_ACTIVE) {
1399                                         if (!ceph_mdsmap_is_laggy(mdsc->mdsmap,
1400                                                                   mds))
1401                                                 goto out;
1402                                 }
1403                         }
1404                         mode = USE_AUTH_MDS;
1405                 }
1406         }
1407
1408         spin_lock(&ci->i_ceph_lock);
1409         cap = NULL;
1410         if (mode == USE_AUTH_MDS)
1411                 cap = ci->i_auth_cap;
1412         if (!cap && !RB_EMPTY_ROOT(&ci->i_caps))
1413                 cap = rb_entry(rb_first(&ci->i_caps), struct ceph_cap, ci_node);
1414         if (!cap) {
1415                 spin_unlock(&ci->i_ceph_lock);
1416                 iput(inode);
1417                 goto random;
1418         }
1419         mds = cap->session->s_mds;
1420         doutc(cl, "%p %llx.%llx mds%d (%scap %p)\n", inode,
1421               ceph_vinop(inode), mds,
1422               cap == ci->i_auth_cap ? "auth " : "", cap);
1423         spin_unlock(&ci->i_ceph_lock);
1424 out:
1425         iput(inode);
1426         return mds;
1427
1428 random:
1429         if (random)
1430                 *random = true;
1431
1432         mds = ceph_mdsmap_get_random_mds(mdsc->mdsmap);
1433         doutc(cl, "chose random mds%d\n", mds);
1434         return mds;
1435 }
1436
1437
1438 /*
1439  * session messages
1440  */
1441 struct ceph_msg *ceph_create_session_msg(u32 op, u64 seq)
1442 {
1443         struct ceph_msg *msg;
1444         struct ceph_mds_session_head *h;
1445
1446         msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h), GFP_NOFS,
1447                            false);
1448         if (!msg) {
1449                 pr_err("ENOMEM creating session %s msg\n",
1450                        ceph_session_op_name(op));
1451                 return NULL;
1452         }
1453         h = msg->front.iov_base;
1454         h->op = cpu_to_le32(op);
1455         h->seq = cpu_to_le64(seq);
1456
1457         return msg;
1458 }
1459
1460 static const unsigned char feature_bits[] = CEPHFS_FEATURES_CLIENT_SUPPORTED;
1461 #define FEATURE_BYTES(c) (DIV_ROUND_UP((size_t)feature_bits[c - 1] + 1, 64) * 8)
1462 static int encode_supported_features(void **p, void *end)
1463 {
1464         static const size_t count = ARRAY_SIZE(feature_bits);
1465
1466         if (count > 0) {
1467                 size_t i;
1468                 size_t size = FEATURE_BYTES(count);
1469                 unsigned long bit;
1470
1471                 if (WARN_ON_ONCE(*p + 4 + size > end))
1472                         return -ERANGE;
1473
1474                 ceph_encode_32(p, size);
1475                 memset(*p, 0, size);
1476                 for (i = 0; i < count; i++) {
1477                         bit = feature_bits[i];
1478                         ((unsigned char *)(*p))[bit / 8] |= BIT(bit % 8);
1479                 }
1480                 *p += size;
1481         } else {
1482                 if (WARN_ON_ONCE(*p + 4 > end))
1483                         return -ERANGE;
1484
1485                 ceph_encode_32(p, 0);
1486         }
1487
1488         return 0;
1489 }
1490
1491 static const unsigned char metric_bits[] = CEPHFS_METRIC_SPEC_CLIENT_SUPPORTED;
1492 #define METRIC_BYTES(cnt) (DIV_ROUND_UP((size_t)metric_bits[cnt - 1] + 1, 64) * 8)
1493 static int encode_metric_spec(void **p, void *end)
1494 {
1495         static const size_t count = ARRAY_SIZE(metric_bits);
1496
1497         /* header */
1498         if (WARN_ON_ONCE(*p + 2 > end))
1499                 return -ERANGE;
1500
1501         ceph_encode_8(p, 1); /* version */
1502         ceph_encode_8(p, 1); /* compat */
1503
1504         if (count > 0) {
1505                 size_t i;
1506                 size_t size = METRIC_BYTES(count);
1507
1508                 if (WARN_ON_ONCE(*p + 4 + 4 + size > end))
1509                         return -ERANGE;
1510
1511                 /* metric spec info length */
1512                 ceph_encode_32(p, 4 + size);
1513
1514                 /* metric spec */
1515                 ceph_encode_32(p, size);
1516                 memset(*p, 0, size);
1517                 for (i = 0; i < count; i++)
1518                         ((unsigned char *)(*p))[i / 8] |= BIT(metric_bits[i] % 8);
1519                 *p += size;
1520         } else {
1521                 if (WARN_ON_ONCE(*p + 4 + 4 > end))
1522                         return -ERANGE;
1523
1524                 /* metric spec info length */
1525                 ceph_encode_32(p, 4);
1526                 /* metric spec */
1527                 ceph_encode_32(p, 0);
1528         }
1529
1530         return 0;
1531 }
1532
1533 /*
1534  * session message, specialization for CEPH_SESSION_REQUEST_OPEN
1535  * to include additional client metadata fields.
1536  */
1537 static struct ceph_msg *
1538 create_session_full_msg(struct ceph_mds_client *mdsc, int op, u64 seq)
1539 {
1540         struct ceph_msg *msg;
1541         struct ceph_mds_session_head *h;
1542         int i;
1543         int extra_bytes = 0;
1544         int metadata_key_count = 0;
1545         struct ceph_options *opt = mdsc->fsc->client->options;
1546         struct ceph_mount_options *fsopt = mdsc->fsc->mount_options;
1547         struct ceph_client *cl = mdsc->fsc->client;
1548         size_t size, count;
1549         void *p, *end;
1550         int ret;
1551
1552         const char* metadata[][2] = {
1553                 {"hostname", mdsc->nodename},
1554                 {"kernel_version", init_utsname()->release},
1555                 {"entity_id", opt->name ? : ""},
1556                 {"root", fsopt->server_path ? : "/"},
1557                 {NULL, NULL}
1558         };
1559
1560         /* Calculate serialized length of metadata */
1561         extra_bytes = 4;  /* map length */
1562         for (i = 0; metadata[i][0]; ++i) {
1563                 extra_bytes += 8 + strlen(metadata[i][0]) +
1564                         strlen(metadata[i][1]);
1565                 metadata_key_count++;
1566         }
1567
1568         /* supported feature */
1569         size = 0;
1570         count = ARRAY_SIZE(feature_bits);
1571         if (count > 0)
1572                 size = FEATURE_BYTES(count);
1573         extra_bytes += 4 + size;
1574
1575         /* metric spec */
1576         size = 0;
1577         count = ARRAY_SIZE(metric_bits);
1578         if (count > 0)
1579                 size = METRIC_BYTES(count);
1580         extra_bytes += 2 + 4 + 4 + size;
1581
1582         /* flags, mds auth caps and oldest_client_tid */
1583         extra_bytes += 4 + 4 + 8;
1584
1585         /* Allocate the message */
1586         msg = ceph_msg_new(CEPH_MSG_CLIENT_SESSION, sizeof(*h) + extra_bytes,
1587                            GFP_NOFS, false);
1588         if (!msg) {
1589                 pr_err_client(cl, "ENOMEM creating session open msg\n");
1590                 return ERR_PTR(-ENOMEM);
1591         }
1592         p = msg->front.iov_base;
1593         end = p + msg->front.iov_len;
1594
1595         h = p;
1596         h->op = cpu_to_le32(op);
1597         h->seq = cpu_to_le64(seq);
1598
1599         /*
1600          * Serialize client metadata into waiting buffer space, using
1601          * the format that userspace expects for map<string, string>
1602          *
1603          * ClientSession messages with metadata are v7
1604          */
1605         msg->hdr.version = cpu_to_le16(7);
1606         msg->hdr.compat_version = cpu_to_le16(1);
1607
1608         /* The write pointer, following the session_head structure */
1609         p += sizeof(*h);
1610
1611         /* Number of entries in the map */
1612         ceph_encode_32(&p, metadata_key_count);
1613
1614         /* Two length-prefixed strings for each entry in the map */
1615         for (i = 0; metadata[i][0]; ++i) {
1616                 size_t const key_len = strlen(metadata[i][0]);
1617                 size_t const val_len = strlen(metadata[i][1]);
1618
1619                 ceph_encode_32(&p, key_len);
1620                 memcpy(p, metadata[i][0], key_len);
1621                 p += key_len;
1622                 ceph_encode_32(&p, val_len);
1623                 memcpy(p, metadata[i][1], val_len);
1624                 p += val_len;
1625         }
1626
1627         ret = encode_supported_features(&p, end);
1628         if (ret) {
1629                 pr_err_client(cl, "encode_supported_features failed!\n");
1630                 ceph_msg_put(msg);
1631                 return ERR_PTR(ret);
1632         }
1633
1634         ret = encode_metric_spec(&p, end);
1635         if (ret) {
1636                 pr_err_client(cl, "encode_metric_spec failed!\n");
1637                 ceph_msg_put(msg);
1638                 return ERR_PTR(ret);
1639         }
1640
1641         /* version == 5, flags */
1642         ceph_encode_32(&p, 0);
1643
1644         /* version == 6, mds auth caps */
1645         ceph_encode_32(&p, 0);
1646
1647         /* version == 7, oldest_client_tid */
1648         ceph_encode_64(&p, mdsc->oldest_tid);
1649
1650         msg->front.iov_len = p - msg->front.iov_base;
1651         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
1652
1653         return msg;
1654 }
1655
1656 /*
1657  * send session open request.
1658  *
1659  * called under mdsc->mutex
1660  */
1661 static int __open_session(struct ceph_mds_client *mdsc,
1662                           struct ceph_mds_session *session)
1663 {
1664         struct ceph_msg *msg;
1665         int mstate;
1666         int mds = session->s_mds;
1667
1668         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_FENCE_IO)
1669                 return -EIO;
1670
1671         /* wait for mds to go active? */
1672         mstate = ceph_mdsmap_get_state(mdsc->mdsmap, mds);
1673         doutc(mdsc->fsc->client, "open_session to mds%d (%s)\n", mds,
1674               ceph_mds_state_name(mstate));
1675         session->s_state = CEPH_MDS_SESSION_OPENING;
1676         session->s_renew_requested = jiffies;
1677
1678         /* send connect message */
1679         msg = create_session_full_msg(mdsc, CEPH_SESSION_REQUEST_OPEN,
1680                                       session->s_seq);
1681         if (IS_ERR(msg))
1682                 return PTR_ERR(msg);
1683         ceph_con_send(&session->s_con, msg);
1684         return 0;
1685 }
1686
1687 /*
1688  * open sessions for any export targets for the given mds
1689  *
1690  * called under mdsc->mutex
1691  */
1692 static struct ceph_mds_session *
1693 __open_export_target_session(struct ceph_mds_client *mdsc, int target)
1694 {
1695         struct ceph_mds_session *session;
1696         int ret;
1697
1698         session = __ceph_lookup_mds_session(mdsc, target);
1699         if (!session) {
1700                 session = register_session(mdsc, target);
1701                 if (IS_ERR(session))
1702                         return session;
1703         }
1704         if (session->s_state == CEPH_MDS_SESSION_NEW ||
1705             session->s_state == CEPH_MDS_SESSION_CLOSING) {
1706                 ret = __open_session(mdsc, session);
1707                 if (ret)
1708                         return ERR_PTR(ret);
1709         }
1710
1711         return session;
1712 }
1713
1714 struct ceph_mds_session *
1715 ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target)
1716 {
1717         struct ceph_mds_session *session;
1718         struct ceph_client *cl = mdsc->fsc->client;
1719
1720         doutc(cl, "to mds%d\n", target);
1721
1722         mutex_lock(&mdsc->mutex);
1723         session = __open_export_target_session(mdsc, target);
1724         mutex_unlock(&mdsc->mutex);
1725
1726         return session;
1727 }
1728
1729 static void __open_export_target_sessions(struct ceph_mds_client *mdsc,
1730                                           struct ceph_mds_session *session)
1731 {
1732         struct ceph_mds_info *mi;
1733         struct ceph_mds_session *ts;
1734         int i, mds = session->s_mds;
1735         struct ceph_client *cl = mdsc->fsc->client;
1736
1737         if (mds >= mdsc->mdsmap->possible_max_rank)
1738                 return;
1739
1740         mi = &mdsc->mdsmap->m_info[mds];
1741         doutc(cl, "for mds%d (%d targets)\n", session->s_mds,
1742               mi->num_export_targets);
1743
1744         for (i = 0; i < mi->num_export_targets; i++) {
1745                 ts = __open_export_target_session(mdsc, mi->export_targets[i]);
1746                 ceph_put_mds_session(ts);
1747         }
1748 }
1749
1750 void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
1751                                            struct ceph_mds_session *session)
1752 {
1753         mutex_lock(&mdsc->mutex);
1754         __open_export_target_sessions(mdsc, session);
1755         mutex_unlock(&mdsc->mutex);
1756 }
1757
1758 /*
1759  * session caps
1760  */
1761
1762 static void detach_cap_releases(struct ceph_mds_session *session,
1763                                 struct list_head *target)
1764 {
1765         struct ceph_client *cl = session->s_mdsc->fsc->client;
1766
1767         lockdep_assert_held(&session->s_cap_lock);
1768
1769         list_splice_init(&session->s_cap_releases, target);
1770         session->s_num_cap_releases = 0;
1771         doutc(cl, "mds%d\n", session->s_mds);
1772 }
1773
1774 static void dispose_cap_releases(struct ceph_mds_client *mdsc,
1775                                  struct list_head *dispose)
1776 {
1777         while (!list_empty(dispose)) {
1778                 struct ceph_cap *cap;
1779                 /* zero out the in-progress message */
1780                 cap = list_first_entry(dispose, struct ceph_cap, session_caps);
1781                 list_del(&cap->session_caps);
1782                 ceph_put_cap(mdsc, cap);
1783         }
1784 }
1785
1786 static void cleanup_session_requests(struct ceph_mds_client *mdsc,
1787                                      struct ceph_mds_session *session)
1788 {
1789         struct ceph_client *cl = mdsc->fsc->client;
1790         struct ceph_mds_request *req;
1791         struct rb_node *p;
1792
1793         doutc(cl, "mds%d\n", session->s_mds);
1794         mutex_lock(&mdsc->mutex);
1795         while (!list_empty(&session->s_unsafe)) {
1796                 req = list_first_entry(&session->s_unsafe,
1797                                        struct ceph_mds_request, r_unsafe_item);
1798                 pr_warn_ratelimited_client(cl, " dropping unsafe request %llu\n",
1799                                            req->r_tid);
1800                 if (req->r_target_inode)
1801                         mapping_set_error(req->r_target_inode->i_mapping, -EIO);
1802                 if (req->r_unsafe_dir)
1803                         mapping_set_error(req->r_unsafe_dir->i_mapping, -EIO);
1804                 __unregister_request(mdsc, req);
1805         }
1806         /* zero r_attempts, so kick_requests() will re-send requests */
1807         p = rb_first(&mdsc->request_tree);
1808         while (p) {
1809                 req = rb_entry(p, struct ceph_mds_request, r_node);
1810                 p = rb_next(p);
1811                 if (req->r_session &&
1812                     req->r_session->s_mds == session->s_mds)
1813                         req->r_attempts = 0;
1814         }
1815         mutex_unlock(&mdsc->mutex);
1816 }
1817
1818 /*
1819  * Helper to safely iterate over all caps associated with a session, with
1820  * special care taken to handle a racing __ceph_remove_cap().
1821  *
1822  * Caller must hold session s_mutex.
1823  */
1824 int ceph_iterate_session_caps(struct ceph_mds_session *session,
1825                               int (*cb)(struct inode *, int mds, void *),
1826                               void *arg)
1827 {
1828         struct ceph_client *cl = session->s_mdsc->fsc->client;
1829         struct list_head *p;
1830         struct ceph_cap *cap;
1831         struct inode *inode, *last_inode = NULL;
1832         struct ceph_cap *old_cap = NULL;
1833         int ret;
1834
1835         doutc(cl, "%p mds%d\n", session, session->s_mds);
1836         spin_lock(&session->s_cap_lock);
1837         p = session->s_caps.next;
1838         while (p != &session->s_caps) {
1839                 int mds;
1840
1841                 cap = list_entry(p, struct ceph_cap, session_caps);
1842                 inode = igrab(&cap->ci->netfs.inode);
1843                 if (!inode) {
1844                         p = p->next;
1845                         continue;
1846                 }
1847                 session->s_cap_iterator = cap;
1848                 mds = cap->mds;
1849                 spin_unlock(&session->s_cap_lock);
1850
1851                 if (last_inode) {
1852                         iput(last_inode);
1853                         last_inode = NULL;
1854                 }
1855                 if (old_cap) {
1856                         ceph_put_cap(session->s_mdsc, old_cap);
1857                         old_cap = NULL;
1858                 }
1859
1860                 ret = cb(inode, mds, arg);
1861                 last_inode = inode;
1862
1863                 spin_lock(&session->s_cap_lock);
1864                 p = p->next;
1865                 if (!cap->ci) {
1866                         doutc(cl, "finishing cap %p removal\n", cap);
1867                         BUG_ON(cap->session != session);
1868                         cap->session = NULL;
1869                         list_del_init(&cap->session_caps);
1870                         session->s_nr_caps--;
1871                         atomic64_dec(&session->s_mdsc->metric.total_caps);
1872                         if (cap->queue_release)
1873                                 __ceph_queue_cap_release(session, cap);
1874                         else
1875                                 old_cap = cap;  /* put_cap it w/o locks held */
1876                 }
1877                 if (ret < 0)
1878                         goto out;
1879         }
1880         ret = 0;
1881 out:
1882         session->s_cap_iterator = NULL;
1883         spin_unlock(&session->s_cap_lock);
1884
1885         iput(last_inode);
1886         if (old_cap)
1887                 ceph_put_cap(session->s_mdsc, old_cap);
1888
1889         return ret;
1890 }
1891
1892 static int remove_session_caps_cb(struct inode *inode, int mds, void *arg)
1893 {
1894         struct ceph_inode_info *ci = ceph_inode(inode);
1895         struct ceph_client *cl = ceph_inode_to_client(inode);
1896         bool invalidate = false;
1897         struct ceph_cap *cap;
1898         int iputs = 0;
1899
1900         spin_lock(&ci->i_ceph_lock);
1901         cap = __get_cap_for_mds(ci, mds);
1902         if (cap) {
1903                 doutc(cl, " removing cap %p, ci is %p, inode is %p\n",
1904                       cap, ci, &ci->netfs.inode);
1905
1906                 iputs = ceph_purge_inode_cap(inode, cap, &invalidate);
1907         }
1908         spin_unlock(&ci->i_ceph_lock);
1909
1910         if (cap)
1911                 wake_up_all(&ci->i_cap_wq);
1912         if (invalidate)
1913                 ceph_queue_invalidate(inode);
1914         while (iputs--)
1915                 iput(inode);
1916         return 0;
1917 }
1918
1919 /*
1920  * caller must hold session s_mutex
1921  */
1922 static void remove_session_caps(struct ceph_mds_session *session)
1923 {
1924         struct ceph_fs_client *fsc = session->s_mdsc->fsc;
1925         struct super_block *sb = fsc->sb;
1926         LIST_HEAD(dispose);
1927
1928         doutc(fsc->client, "on %p\n", session);
1929         ceph_iterate_session_caps(session, remove_session_caps_cb, fsc);
1930
1931         wake_up_all(&fsc->mdsc->cap_flushing_wq);
1932
1933         spin_lock(&session->s_cap_lock);
1934         if (session->s_nr_caps > 0) {
1935                 struct inode *inode;
1936                 struct ceph_cap *cap, *prev = NULL;
1937                 struct ceph_vino vino;
1938                 /*
1939                  * iterate_session_caps() skips inodes that are being
1940                  * deleted, we need to wait until deletions are complete.
1941                  * __wait_on_freeing_inode() is designed for the job,
1942                  * but it is not exported, so use lookup inode function
1943                  * to access it.
1944                  */
1945                 while (!list_empty(&session->s_caps)) {
1946                         cap = list_entry(session->s_caps.next,
1947                                          struct ceph_cap, session_caps);
1948                         if (cap == prev)
1949                                 break;
1950                         prev = cap;
1951                         vino = cap->ci->i_vino;
1952                         spin_unlock(&session->s_cap_lock);
1953
1954                         inode = ceph_find_inode(sb, vino);
1955                         iput(inode);
1956
1957                         spin_lock(&session->s_cap_lock);
1958                 }
1959         }
1960
1961         // drop cap expires and unlock s_cap_lock
1962         detach_cap_releases(session, &dispose);
1963
1964         BUG_ON(session->s_nr_caps > 0);
1965         BUG_ON(!list_empty(&session->s_cap_flushing));
1966         spin_unlock(&session->s_cap_lock);
1967         dispose_cap_releases(session->s_mdsc, &dispose);
1968 }
1969
1970 enum {
1971         RECONNECT,
1972         RENEWCAPS,
1973         FORCE_RO,
1974 };
1975
1976 /*
1977  * wake up any threads waiting on this session's caps.  if the cap is
1978  * old (didn't get renewed on the client reconnect), remove it now.
1979  *
1980  * caller must hold s_mutex.
1981  */
1982 static int wake_up_session_cb(struct inode *inode, int mds, void *arg)
1983 {
1984         struct ceph_inode_info *ci = ceph_inode(inode);
1985         unsigned long ev = (unsigned long)arg;
1986
1987         if (ev == RECONNECT) {
1988                 spin_lock(&ci->i_ceph_lock);
1989                 ci->i_wanted_max_size = 0;
1990                 ci->i_requested_max_size = 0;
1991                 spin_unlock(&ci->i_ceph_lock);
1992         } else if (ev == RENEWCAPS) {
1993                 struct ceph_cap *cap;
1994
1995                 spin_lock(&ci->i_ceph_lock);
1996                 cap = __get_cap_for_mds(ci, mds);
1997                 /* mds did not re-issue stale cap */
1998                 if (cap && cap->cap_gen < atomic_read(&cap->session->s_cap_gen))
1999                         cap->issued = cap->implemented = CEPH_CAP_PIN;
2000                 spin_unlock(&ci->i_ceph_lock);
2001         } else if (ev == FORCE_RO) {
2002         }
2003         wake_up_all(&ci->i_cap_wq);
2004         return 0;
2005 }
2006
2007 static void wake_up_session_caps(struct ceph_mds_session *session, int ev)
2008 {
2009         struct ceph_client *cl = session->s_mdsc->fsc->client;
2010
2011         doutc(cl, "session %p mds%d\n", session, session->s_mds);
2012         ceph_iterate_session_caps(session, wake_up_session_cb,
2013                                   (void *)(unsigned long)ev);
2014 }
2015
2016 /*
2017  * Send periodic message to MDS renewing all currently held caps.  The
2018  * ack will reset the expiration for all caps from this session.
2019  *
2020  * caller holds s_mutex
2021  */
2022 static int send_renew_caps(struct ceph_mds_client *mdsc,
2023                            struct ceph_mds_session *session)
2024 {
2025         struct ceph_client *cl = mdsc->fsc->client;
2026         struct ceph_msg *msg;
2027         int state;
2028
2029         if (time_after_eq(jiffies, session->s_cap_ttl) &&
2030             time_after_eq(session->s_cap_ttl, session->s_renew_requested))
2031                 pr_info_client(cl, "mds%d caps stale\n", session->s_mds);
2032         session->s_renew_requested = jiffies;
2033
2034         /* do not try to renew caps until a recovering mds has reconnected
2035          * with its clients. */
2036         state = ceph_mdsmap_get_state(mdsc->mdsmap, session->s_mds);
2037         if (state < CEPH_MDS_STATE_RECONNECT) {
2038                 doutc(cl, "ignoring mds%d (%s)\n", session->s_mds,
2039                       ceph_mds_state_name(state));
2040                 return 0;
2041         }
2042
2043         doutc(cl, "to mds%d (%s)\n", session->s_mds,
2044               ceph_mds_state_name(state));
2045         msg = create_session_full_msg(mdsc, CEPH_SESSION_REQUEST_RENEWCAPS,
2046                                       ++session->s_renew_seq);
2047         if (IS_ERR(msg))
2048                 return PTR_ERR(msg);
2049         ceph_con_send(&session->s_con, msg);
2050         return 0;
2051 }
2052
2053 static int send_flushmsg_ack(struct ceph_mds_client *mdsc,
2054                              struct ceph_mds_session *session, u64 seq)
2055 {
2056         struct ceph_client *cl = mdsc->fsc->client;
2057         struct ceph_msg *msg;
2058
2059         doutc(cl, "to mds%d (%s)s seq %lld\n", session->s_mds,
2060               ceph_session_state_name(session->s_state), seq);
2061         msg = ceph_create_session_msg(CEPH_SESSION_FLUSHMSG_ACK, seq);
2062         if (!msg)
2063                 return -ENOMEM;
2064         ceph_con_send(&session->s_con, msg);
2065         return 0;
2066 }
2067
2068
2069 /*
2070  * Note new cap ttl, and any transition from stale -> not stale (fresh?).
2071  *
2072  * Called under session->s_mutex
2073  */
2074 static void renewed_caps(struct ceph_mds_client *mdsc,
2075                          struct ceph_mds_session *session, int is_renew)
2076 {
2077         struct ceph_client *cl = mdsc->fsc->client;
2078         int was_stale;
2079         int wake = 0;
2080
2081         spin_lock(&session->s_cap_lock);
2082         was_stale = is_renew && time_after_eq(jiffies, session->s_cap_ttl);
2083
2084         session->s_cap_ttl = session->s_renew_requested +
2085                 mdsc->mdsmap->m_session_timeout*HZ;
2086
2087         if (was_stale) {
2088                 if (time_before(jiffies, session->s_cap_ttl)) {
2089                         pr_info_client(cl, "mds%d caps renewed\n",
2090                                        session->s_mds);
2091                         wake = 1;
2092                 } else {
2093                         pr_info_client(cl, "mds%d caps still stale\n",
2094                                        session->s_mds);
2095                 }
2096         }
2097         doutc(cl, "mds%d ttl now %lu, was %s, now %s\n", session->s_mds,
2098               session->s_cap_ttl, was_stale ? "stale" : "fresh",
2099               time_before(jiffies, session->s_cap_ttl) ? "stale" : "fresh");
2100         spin_unlock(&session->s_cap_lock);
2101
2102         if (wake)
2103                 wake_up_session_caps(session, RENEWCAPS);
2104 }
2105
2106 /*
2107  * send a session close request
2108  */
2109 static int request_close_session(struct ceph_mds_session *session)
2110 {
2111         struct ceph_client *cl = session->s_mdsc->fsc->client;
2112         struct ceph_msg *msg;
2113
2114         doutc(cl, "mds%d state %s seq %lld\n", session->s_mds,
2115               ceph_session_state_name(session->s_state), session->s_seq);
2116         msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_CLOSE,
2117                                       session->s_seq);
2118         if (!msg)
2119                 return -ENOMEM;
2120         ceph_con_send(&session->s_con, msg);
2121         return 1;
2122 }
2123
2124 /*
2125  * Called with s_mutex held.
2126  */
2127 static int __close_session(struct ceph_mds_client *mdsc,
2128                          struct ceph_mds_session *session)
2129 {
2130         if (session->s_state >= CEPH_MDS_SESSION_CLOSING)
2131                 return 0;
2132         session->s_state = CEPH_MDS_SESSION_CLOSING;
2133         return request_close_session(session);
2134 }
2135
2136 static bool drop_negative_children(struct dentry *dentry)
2137 {
2138         struct dentry *child;
2139         bool all_negative = true;
2140
2141         if (!d_is_dir(dentry))
2142                 goto out;
2143
2144         spin_lock(&dentry->d_lock);
2145         hlist_for_each_entry(child, &dentry->d_children, d_sib) {
2146                 if (d_really_is_positive(child)) {
2147                         all_negative = false;
2148                         break;
2149                 }
2150         }
2151         spin_unlock(&dentry->d_lock);
2152
2153         if (all_negative)
2154                 shrink_dcache_parent(dentry);
2155 out:
2156         return all_negative;
2157 }
2158
2159 /*
2160  * Trim old(er) caps.
2161  *
2162  * Because we can't cache an inode without one or more caps, we do
2163  * this indirectly: if a cap is unused, we prune its aliases, at which
2164  * point the inode will hopefully get dropped to.
2165  *
2166  * Yes, this is a bit sloppy.  Our only real goal here is to respond to
2167  * memory pressure from the MDS, though, so it needn't be perfect.
2168  */
2169 static int trim_caps_cb(struct inode *inode, int mds, void *arg)
2170 {
2171         struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb);
2172         struct ceph_client *cl = mdsc->fsc->client;
2173         int *remaining = arg;
2174         struct ceph_inode_info *ci = ceph_inode(inode);
2175         int used, wanted, oissued, mine;
2176         struct ceph_cap *cap;
2177
2178         if (*remaining <= 0)
2179                 return -1;
2180
2181         spin_lock(&ci->i_ceph_lock);
2182         cap = __get_cap_for_mds(ci, mds);
2183         if (!cap) {
2184                 spin_unlock(&ci->i_ceph_lock);
2185                 return 0;
2186         }
2187         mine = cap->issued | cap->implemented;
2188         used = __ceph_caps_used(ci);
2189         wanted = __ceph_caps_file_wanted(ci);
2190         oissued = __ceph_caps_issued_other(ci, cap);
2191
2192         doutc(cl, "%p %llx.%llx cap %p mine %s oissued %s used %s wanted %s\n",
2193               inode, ceph_vinop(inode), cap, ceph_cap_string(mine),
2194               ceph_cap_string(oissued), ceph_cap_string(used),
2195               ceph_cap_string(wanted));
2196         if (cap == ci->i_auth_cap) {
2197                 if (ci->i_dirty_caps || ci->i_flushing_caps ||
2198                     !list_empty(&ci->i_cap_snaps))
2199                         goto out;
2200                 if ((used | wanted) & CEPH_CAP_ANY_WR)
2201                         goto out;
2202                 /* Note: it's possible that i_filelock_ref becomes non-zero
2203                  * after dropping auth caps. It doesn't hurt because reply
2204                  * of lock mds request will re-add auth caps. */
2205                 if (atomic_read(&ci->i_filelock_ref) > 0)
2206                         goto out;
2207         }
2208         /* The inode has cached pages, but it's no longer used.
2209          * we can safely drop it */
2210         if (S_ISREG(inode->i_mode) &&
2211             wanted == 0 && used == CEPH_CAP_FILE_CACHE &&
2212             !(oissued & CEPH_CAP_FILE_CACHE)) {
2213           used = 0;
2214           oissued = 0;
2215         }
2216         if ((used | wanted) & ~oissued & mine)
2217                 goto out;   /* we need these caps */
2218
2219         if (oissued) {
2220                 /* we aren't the only cap.. just remove us */
2221                 ceph_remove_cap(mdsc, cap, true);
2222                 (*remaining)--;
2223         } else {
2224                 struct dentry *dentry;
2225                 /* try dropping referring dentries */
2226                 spin_unlock(&ci->i_ceph_lock);
2227                 dentry = d_find_any_alias(inode);
2228                 if (dentry && drop_negative_children(dentry)) {
2229                         int count;
2230                         dput(dentry);
2231                         d_prune_aliases(inode);
2232                         count = atomic_read(&inode->i_count);
2233                         if (count == 1)
2234                                 (*remaining)--;
2235                         doutc(cl, "%p %llx.%llx cap %p pruned, count now %d\n",
2236                               inode, ceph_vinop(inode), cap, count);
2237                 } else {
2238                         dput(dentry);
2239                 }
2240                 return 0;
2241         }
2242
2243 out:
2244         spin_unlock(&ci->i_ceph_lock);
2245         return 0;
2246 }
2247
2248 /*
2249  * Trim session cap count down to some max number.
2250  */
2251 int ceph_trim_caps(struct ceph_mds_client *mdsc,
2252                    struct ceph_mds_session *session,
2253                    int max_caps)
2254 {
2255         struct ceph_client *cl = mdsc->fsc->client;
2256         int trim_caps = session->s_nr_caps - max_caps;
2257
2258         doutc(cl, "mds%d start: %d / %d, trim %d\n", session->s_mds,
2259               session->s_nr_caps, max_caps, trim_caps);
2260         if (trim_caps > 0) {
2261                 int remaining = trim_caps;
2262
2263                 ceph_iterate_session_caps(session, trim_caps_cb, &remaining);
2264                 doutc(cl, "mds%d done: %d / %d, trimmed %d\n",
2265                       session->s_mds, session->s_nr_caps, max_caps,
2266                       trim_caps - remaining);
2267         }
2268
2269         ceph_flush_cap_releases(mdsc, session);
2270         return 0;
2271 }
2272
2273 static int check_caps_flush(struct ceph_mds_client *mdsc,
2274                             u64 want_flush_tid)
2275 {
2276         struct ceph_client *cl = mdsc->fsc->client;
2277         int ret = 1;
2278
2279         spin_lock(&mdsc->cap_dirty_lock);
2280         if (!list_empty(&mdsc->cap_flush_list)) {
2281                 struct ceph_cap_flush *cf =
2282                         list_first_entry(&mdsc->cap_flush_list,
2283                                          struct ceph_cap_flush, g_list);
2284                 if (cf->tid <= want_flush_tid) {
2285                         doutc(cl, "still flushing tid %llu <= %llu\n",
2286                               cf->tid, want_flush_tid);
2287                         ret = 0;
2288                 }
2289         }
2290         spin_unlock(&mdsc->cap_dirty_lock);
2291         return ret;
2292 }
2293
2294 /*
2295  * flush all dirty inode data to disk.
2296  *
2297  * returns true if we've flushed through want_flush_tid
2298  */
2299 static void wait_caps_flush(struct ceph_mds_client *mdsc,
2300                             u64 want_flush_tid)
2301 {
2302         struct ceph_client *cl = mdsc->fsc->client;
2303
2304         doutc(cl, "want %llu\n", want_flush_tid);
2305
2306         wait_event(mdsc->cap_flushing_wq,
2307                    check_caps_flush(mdsc, want_flush_tid));
2308
2309         doutc(cl, "ok, flushed thru %llu\n", want_flush_tid);
2310 }
2311
2312 /*
2313  * called under s_mutex
2314  */
2315 static void ceph_send_cap_releases(struct ceph_mds_client *mdsc,
2316                                    struct ceph_mds_session *session)
2317 {
2318         struct ceph_client *cl = mdsc->fsc->client;
2319         struct ceph_msg *msg = NULL;
2320         struct ceph_mds_cap_release *head;
2321         struct ceph_mds_cap_item *item;
2322         struct ceph_osd_client *osdc = &mdsc->fsc->client->osdc;
2323         struct ceph_cap *cap;
2324         LIST_HEAD(tmp_list);
2325         int num_cap_releases;
2326         __le32  barrier, *cap_barrier;
2327
2328         down_read(&osdc->lock);
2329         barrier = cpu_to_le32(osdc->epoch_barrier);
2330         up_read(&osdc->lock);
2331
2332         spin_lock(&session->s_cap_lock);
2333 again:
2334         list_splice_init(&session->s_cap_releases, &tmp_list);
2335         num_cap_releases = session->s_num_cap_releases;
2336         session->s_num_cap_releases = 0;
2337         spin_unlock(&session->s_cap_lock);
2338
2339         while (!list_empty(&tmp_list)) {
2340                 if (!msg) {
2341                         msg = ceph_msg_new(CEPH_MSG_CLIENT_CAPRELEASE,
2342                                         PAGE_SIZE, GFP_NOFS, false);
2343                         if (!msg)
2344                                 goto out_err;
2345                         head = msg->front.iov_base;
2346                         head->num = cpu_to_le32(0);
2347                         msg->front.iov_len = sizeof(*head);
2348
2349                         msg->hdr.version = cpu_to_le16(2);
2350                         msg->hdr.compat_version = cpu_to_le16(1);
2351                 }
2352
2353                 cap = list_first_entry(&tmp_list, struct ceph_cap,
2354                                         session_caps);
2355                 list_del(&cap->session_caps);
2356                 num_cap_releases--;
2357
2358                 head = msg->front.iov_base;
2359                 put_unaligned_le32(get_unaligned_le32(&head->num) + 1,
2360                                    &head->num);
2361                 item = msg->front.iov_base + msg->front.iov_len;
2362                 item->ino = cpu_to_le64(cap->cap_ino);
2363                 item->cap_id = cpu_to_le64(cap->cap_id);
2364                 item->migrate_seq = cpu_to_le32(cap->mseq);
2365                 item->seq = cpu_to_le32(cap->issue_seq);
2366                 msg->front.iov_len += sizeof(*item);
2367
2368                 ceph_put_cap(mdsc, cap);
2369
2370                 if (le32_to_cpu(head->num) == CEPH_CAPS_PER_RELEASE) {
2371                         // Append cap_barrier field
2372                         cap_barrier = msg->front.iov_base + msg->front.iov_len;
2373                         *cap_barrier = barrier;
2374                         msg->front.iov_len += sizeof(*cap_barrier);
2375
2376                         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
2377                         doutc(cl, "mds%d %p\n", session->s_mds, msg);
2378                         ceph_con_send(&session->s_con, msg);
2379                         msg = NULL;
2380                 }
2381         }
2382
2383         BUG_ON(num_cap_releases != 0);
2384
2385         spin_lock(&session->s_cap_lock);
2386         if (!list_empty(&session->s_cap_releases))
2387                 goto again;
2388         spin_unlock(&session->s_cap_lock);
2389
2390         if (msg) {
2391                 // Append cap_barrier field
2392                 cap_barrier = msg->front.iov_base + msg->front.iov_len;
2393                 *cap_barrier = barrier;
2394                 msg->front.iov_len += sizeof(*cap_barrier);
2395
2396                 msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
2397                 doutc(cl, "mds%d %p\n", session->s_mds, msg);
2398                 ceph_con_send(&session->s_con, msg);
2399         }
2400         return;
2401 out_err:
2402         pr_err_client(cl, "mds%d, failed to allocate message\n",
2403                       session->s_mds);
2404         spin_lock(&session->s_cap_lock);
2405         list_splice(&tmp_list, &session->s_cap_releases);
2406         session->s_num_cap_releases += num_cap_releases;
2407         spin_unlock(&session->s_cap_lock);
2408 }
2409
2410 static void ceph_cap_release_work(struct work_struct *work)
2411 {
2412         struct ceph_mds_session *session =
2413                 container_of(work, struct ceph_mds_session, s_cap_release_work);
2414
2415         mutex_lock(&session->s_mutex);
2416         if (session->s_state == CEPH_MDS_SESSION_OPEN ||
2417             session->s_state == CEPH_MDS_SESSION_HUNG)
2418                 ceph_send_cap_releases(session->s_mdsc, session);
2419         mutex_unlock(&session->s_mutex);
2420         ceph_put_mds_session(session);
2421 }
2422
2423 void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
2424                              struct ceph_mds_session *session)
2425 {
2426         struct ceph_client *cl = mdsc->fsc->client;
2427         if (mdsc->stopping)
2428                 return;
2429
2430         ceph_get_mds_session(session);
2431         if (queue_work(mdsc->fsc->cap_wq,
2432                        &session->s_cap_release_work)) {
2433                 doutc(cl, "cap release work queued\n");
2434         } else {
2435                 ceph_put_mds_session(session);
2436                 doutc(cl, "failed to queue cap release work\n");
2437         }
2438 }
2439
2440 /*
2441  * caller holds session->s_cap_lock
2442  */
2443 void __ceph_queue_cap_release(struct ceph_mds_session *session,
2444                               struct ceph_cap *cap)
2445 {
2446         list_add_tail(&cap->session_caps, &session->s_cap_releases);
2447         session->s_num_cap_releases++;
2448
2449         if (!(session->s_num_cap_releases % CEPH_CAPS_PER_RELEASE))
2450                 ceph_flush_cap_releases(session->s_mdsc, session);
2451 }
2452
2453 static void ceph_cap_reclaim_work(struct work_struct *work)
2454 {
2455         struct ceph_mds_client *mdsc =
2456                 container_of(work, struct ceph_mds_client, cap_reclaim_work);
2457         int ret = ceph_trim_dentries(mdsc);
2458         if (ret == -EAGAIN)
2459                 ceph_queue_cap_reclaim_work(mdsc);
2460 }
2461
2462 void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc)
2463 {
2464         struct ceph_client *cl = mdsc->fsc->client;
2465         if (mdsc->stopping)
2466                 return;
2467
2468         if (queue_work(mdsc->fsc->cap_wq, &mdsc->cap_reclaim_work)) {
2469                 doutc(cl, "caps reclaim work queued\n");
2470         } else {
2471                 doutc(cl, "failed to queue caps release work\n");
2472         }
2473 }
2474
2475 void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr)
2476 {
2477         int val;
2478         if (!nr)
2479                 return;
2480         val = atomic_add_return(nr, &mdsc->cap_reclaim_pending);
2481         if ((val % CEPH_CAPS_PER_RELEASE) < nr) {
2482                 atomic_set(&mdsc->cap_reclaim_pending, 0);
2483                 ceph_queue_cap_reclaim_work(mdsc);
2484         }
2485 }
2486
2487 void ceph_queue_cap_unlink_work(struct ceph_mds_client *mdsc)
2488 {
2489         struct ceph_client *cl = mdsc->fsc->client;
2490         if (mdsc->stopping)
2491                 return;
2492
2493         if (queue_work(mdsc->fsc->cap_wq, &mdsc->cap_unlink_work)) {
2494                 doutc(cl, "caps unlink work queued\n");
2495         } else {
2496                 doutc(cl, "failed to queue caps unlink work\n");
2497         }
2498 }
2499
2500 static void ceph_cap_unlink_work(struct work_struct *work)
2501 {
2502         struct ceph_mds_client *mdsc =
2503                 container_of(work, struct ceph_mds_client, cap_unlink_work);
2504         struct ceph_client *cl = mdsc->fsc->client;
2505
2506         doutc(cl, "begin\n");
2507         spin_lock(&mdsc->cap_unlink_delay_lock);
2508         while (!list_empty(&mdsc->cap_unlink_delay_list)) {
2509                 struct ceph_inode_info *ci;
2510                 struct inode *inode;
2511
2512                 ci = list_first_entry(&mdsc->cap_unlink_delay_list,
2513                                       struct ceph_inode_info,
2514                                       i_cap_delay_list);
2515                 list_del_init(&ci->i_cap_delay_list);
2516
2517                 inode = igrab(&ci->netfs.inode);
2518                 if (inode) {
2519                         spin_unlock(&mdsc->cap_unlink_delay_lock);
2520                         doutc(cl, "on %p %llx.%llx\n", inode,
2521                               ceph_vinop(inode));
2522                         ceph_check_caps(ci, CHECK_CAPS_FLUSH);
2523                         iput(inode);
2524                         spin_lock(&mdsc->cap_unlink_delay_lock);
2525                 }
2526         }
2527         spin_unlock(&mdsc->cap_unlink_delay_lock);
2528         doutc(cl, "done\n");
2529 }
2530
2531 /*
2532  * requests
2533  */
2534
2535 int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
2536                                     struct inode *dir)
2537 {
2538         struct ceph_inode_info *ci = ceph_inode(dir);
2539         struct ceph_mds_reply_info_parsed *rinfo = &req->r_reply_info;
2540         struct ceph_mount_options *opt = req->r_mdsc->fsc->mount_options;
2541         size_t size = sizeof(struct ceph_mds_reply_dir_entry);
2542         unsigned int num_entries;
2543         int order;
2544
2545         spin_lock(&ci->i_ceph_lock);
2546         num_entries = ci->i_files + ci->i_subdirs;
2547         spin_unlock(&ci->i_ceph_lock);
2548         num_entries = max(num_entries, 1U);
2549         num_entries = min(num_entries, opt->max_readdir);
2550
2551         order = get_order(size * num_entries);
2552         while (order >= 0) {
2553                 rinfo->dir_entries = (void*)__get_free_pages(GFP_KERNEL |
2554                                                              __GFP_NOWARN |
2555                                                              __GFP_ZERO,
2556                                                              order);
2557                 if (rinfo->dir_entries)
2558                         break;
2559                 order--;
2560         }
2561         if (!rinfo->dir_entries)
2562                 return -ENOMEM;
2563
2564         num_entries = (PAGE_SIZE << order) / size;
2565         num_entries = min(num_entries, opt->max_readdir);
2566
2567         rinfo->dir_buf_size = PAGE_SIZE << order;
2568         req->r_num_caps = num_entries + 1;
2569         req->r_args.readdir.max_entries = cpu_to_le32(num_entries);
2570         req->r_args.readdir.max_bytes = cpu_to_le32(opt->max_readdir_bytes);
2571         return 0;
2572 }
2573
2574 /*
2575  * Create an mds request.
2576  */
2577 struct ceph_mds_request *
2578 ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode)
2579 {
2580         struct ceph_mds_request *req;
2581
2582         req = kmem_cache_zalloc(ceph_mds_request_cachep, GFP_NOFS);
2583         if (!req)
2584                 return ERR_PTR(-ENOMEM);
2585
2586         mutex_init(&req->r_fill_mutex);
2587         req->r_mdsc = mdsc;
2588         req->r_started = jiffies;
2589         req->r_start_latency = ktime_get();
2590         req->r_resend_mds = -1;
2591         INIT_LIST_HEAD(&req->r_unsafe_dir_item);
2592         INIT_LIST_HEAD(&req->r_unsafe_target_item);
2593         req->r_fmode = -1;
2594         req->r_feature_needed = -1;
2595         kref_init(&req->r_kref);
2596         RB_CLEAR_NODE(&req->r_node);
2597         INIT_LIST_HEAD(&req->r_wait);
2598         init_completion(&req->r_completion);
2599         init_completion(&req->r_safe_completion);
2600         INIT_LIST_HEAD(&req->r_unsafe_item);
2601
2602         ktime_get_coarse_real_ts64(&req->r_stamp);
2603
2604         req->r_op = op;
2605         req->r_direct_mode = mode;
2606         return req;
2607 }
2608
2609 /*
2610  * return oldest (lowest) request, tid in request tree, 0 if none.
2611  *
2612  * called under mdsc->mutex.
2613  */
2614 static struct ceph_mds_request *__get_oldest_req(struct ceph_mds_client *mdsc)
2615 {
2616         if (RB_EMPTY_ROOT(&mdsc->request_tree))
2617                 return NULL;
2618         return rb_entry(rb_first(&mdsc->request_tree),
2619                         struct ceph_mds_request, r_node);
2620 }
2621
2622 static inline  u64 __get_oldest_tid(struct ceph_mds_client *mdsc)
2623 {
2624         return mdsc->oldest_tid;
2625 }
2626
2627 #if IS_ENABLED(CONFIG_FS_ENCRYPTION)
2628 static u8 *get_fscrypt_altname(const struct ceph_mds_request *req, u32 *plen)
2629 {
2630         struct inode *dir = req->r_parent;
2631         struct dentry *dentry = req->r_dentry;
2632         u8 *cryptbuf = NULL;
2633         u32 len = 0;
2634         int ret = 0;
2635
2636         /* only encode if we have parent and dentry */
2637         if (!dir || !dentry)
2638                 goto success;
2639
2640         /* No-op unless this is encrypted */
2641         if (!IS_ENCRYPTED(dir))
2642                 goto success;
2643
2644         ret = ceph_fscrypt_prepare_readdir(dir);
2645         if (ret < 0)
2646                 return ERR_PTR(ret);
2647
2648         /* No key? Just ignore it. */
2649         if (!fscrypt_has_encryption_key(dir))
2650                 goto success;
2651
2652         if (!fscrypt_fname_encrypted_size(dir, dentry->d_name.len, NAME_MAX,
2653                                           &len)) {
2654                 WARN_ON_ONCE(1);
2655                 return ERR_PTR(-ENAMETOOLONG);
2656         }
2657
2658         /* No need to append altname if name is short enough */
2659         if (len <= CEPH_NOHASH_NAME_MAX) {
2660                 len = 0;
2661                 goto success;
2662         }
2663
2664         cryptbuf = kmalloc(len, GFP_KERNEL);
2665         if (!cryptbuf)
2666                 return ERR_PTR(-ENOMEM);
2667
2668         ret = fscrypt_fname_encrypt(dir, &dentry->d_name, cryptbuf, len);
2669         if (ret) {
2670                 kfree(cryptbuf);
2671                 return ERR_PTR(ret);
2672         }
2673 success:
2674         *plen = len;
2675         return cryptbuf;
2676 }
2677 #else
2678 static u8 *get_fscrypt_altname(const struct ceph_mds_request *req, u32 *plen)
2679 {
2680         *plen = 0;
2681         return NULL;
2682 }
2683 #endif
2684
2685 /**
2686  * ceph_mdsc_build_path - build a path string to a given dentry
2687  * @mdsc: mds client
2688  * @dentry: dentry to which path should be built
2689  * @plen: returned length of string
2690  * @pbase: returned base inode number
2691  * @for_wire: is this path going to be sent to the MDS?
2692  *
2693  * Build a string that represents the path to the dentry. This is mostly called
2694  * for two different purposes:
2695  *
2696  * 1) we need to build a path string to send to the MDS (for_wire == true)
2697  * 2) we need a path string for local presentation (e.g. debugfs)
2698  *    (for_wire == false)
2699  *
2700  * The path is built in reverse, starting with the dentry. Walk back up toward
2701  * the root, building the path until the first non-snapped inode is reached
2702  * (for_wire) or the root inode is reached (!for_wire).
2703  *
2704  * Encode hidden .snap dirs as a double /, i.e.
2705  *   foo/.snap/bar -> foo//bar
2706  */
2707 char *ceph_mdsc_build_path(struct ceph_mds_client *mdsc, struct dentry *dentry,
2708                            int *plen, u64 *pbase, int for_wire)
2709 {
2710         struct ceph_client *cl = mdsc->fsc->client;
2711         struct dentry *cur;
2712         struct inode *inode;
2713         char *path;
2714         int pos;
2715         unsigned seq;
2716         u64 base;
2717
2718         if (!dentry)
2719                 return ERR_PTR(-EINVAL);
2720
2721         path = __getname();
2722         if (!path)
2723                 return ERR_PTR(-ENOMEM);
2724 retry:
2725         pos = PATH_MAX - 1;
2726         path[pos] = '\0';
2727
2728         seq = read_seqbegin(&rename_lock);
2729         cur = dget(dentry);
2730         for (;;) {
2731                 struct dentry *parent;
2732
2733                 spin_lock(&cur->d_lock);
2734                 inode = d_inode(cur);
2735                 if (inode && ceph_snap(inode) == CEPH_SNAPDIR) {
2736                         doutc(cl, "path+%d: %p SNAPDIR\n", pos, cur);
2737                         spin_unlock(&cur->d_lock);
2738                         parent = dget_parent(cur);
2739                 } else if (for_wire && inode && dentry != cur &&
2740                            ceph_snap(inode) == CEPH_NOSNAP) {
2741                         spin_unlock(&cur->d_lock);
2742                         pos++; /* get rid of any prepended '/' */
2743                         break;
2744                 } else if (!for_wire || !IS_ENCRYPTED(d_inode(cur->d_parent))) {
2745                         pos -= cur->d_name.len;
2746                         if (pos < 0) {
2747                                 spin_unlock(&cur->d_lock);
2748                                 break;
2749                         }
2750                         memcpy(path + pos, cur->d_name.name, cur->d_name.len);
2751                         spin_unlock(&cur->d_lock);
2752                         parent = dget_parent(cur);
2753                 } else {
2754                         int len, ret;
2755                         char buf[NAME_MAX];
2756
2757                         /*
2758                          * Proactively copy name into buf, in case we need to
2759                          * present it as-is.
2760                          */
2761                         memcpy(buf, cur->d_name.name, cur->d_name.len);
2762                         len = cur->d_name.len;
2763                         spin_unlock(&cur->d_lock);
2764                         parent = dget_parent(cur);
2765
2766                         ret = ceph_fscrypt_prepare_readdir(d_inode(parent));
2767                         if (ret < 0) {
2768                                 dput(parent);
2769                                 dput(cur);
2770                                 return ERR_PTR(ret);
2771                         }
2772
2773                         if (fscrypt_has_encryption_key(d_inode(parent))) {
2774                                 len = ceph_encode_encrypted_fname(d_inode(parent),
2775                                                                   cur, buf);
2776                                 if (len < 0) {
2777                                         dput(parent);
2778                                         dput(cur);
2779                                         return ERR_PTR(len);
2780                                 }
2781                         }
2782                         pos -= len;
2783                         if (pos < 0) {
2784                                 dput(parent);
2785                                 break;
2786                         }
2787                         memcpy(path + pos, buf, len);
2788                 }
2789                 dput(cur);
2790                 cur = parent;
2791
2792                 /* Are we at the root? */
2793                 if (IS_ROOT(cur))
2794                         break;
2795
2796                 /* Are we out of buffer? */
2797                 if (--pos < 0)
2798                         break;
2799
2800                 path[pos] = '/';
2801         }
2802         inode = d_inode(cur);
2803         base = inode ? ceph_ino(inode) : 0;
2804         dput(cur);
2805
2806         if (read_seqretry(&rename_lock, seq))
2807                 goto retry;
2808
2809         if (pos < 0) {
2810                 /*
2811                  * A rename didn't occur, but somehow we didn't end up where
2812                  * we thought we would. Throw a warning and try again.
2813                  */
2814                 pr_warn_client(cl, "did not end path lookup where expected (pos = %d)\n",
2815                                pos);
2816                 goto retry;
2817         }
2818
2819         *pbase = base;
2820         *plen = PATH_MAX - 1 - pos;
2821         doutc(cl, "on %p %d built %llx '%.*s'\n", dentry, d_count(dentry),
2822               base, *plen, path + pos);
2823         return path + pos;
2824 }
2825
2826 static int build_dentry_path(struct ceph_mds_client *mdsc, struct dentry *dentry,
2827                              struct inode *dir, const char **ppath, int *ppathlen,
2828                              u64 *pino, bool *pfreepath, bool parent_locked)
2829 {
2830         char *path;
2831
2832         rcu_read_lock();
2833         if (!dir)
2834                 dir = d_inode_rcu(dentry->d_parent);
2835         if (dir && parent_locked && ceph_snap(dir) == CEPH_NOSNAP &&
2836             !IS_ENCRYPTED(dir)) {
2837                 *pino = ceph_ino(dir);
2838                 rcu_read_unlock();
2839                 *ppath = dentry->d_name.name;
2840                 *ppathlen = dentry->d_name.len;
2841                 return 0;
2842         }
2843         rcu_read_unlock();
2844         path = ceph_mdsc_build_path(mdsc, dentry, ppathlen, pino, 1);
2845         if (IS_ERR(path))
2846                 return PTR_ERR(path);
2847         *ppath = path;
2848         *pfreepath = true;
2849         return 0;
2850 }
2851
2852 static int build_inode_path(struct inode *inode,
2853                             const char **ppath, int *ppathlen, u64 *pino,
2854                             bool *pfreepath)
2855 {
2856         struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb);
2857         struct dentry *dentry;
2858         char *path;
2859
2860         if (ceph_snap(inode) == CEPH_NOSNAP) {
2861                 *pino = ceph_ino(inode);
2862                 *ppathlen = 0;
2863                 return 0;
2864         }
2865         dentry = d_find_alias(inode);
2866         path = ceph_mdsc_build_path(mdsc, dentry, ppathlen, pino, 1);
2867         dput(dentry);
2868         if (IS_ERR(path))
2869                 return PTR_ERR(path);
2870         *ppath = path;
2871         *pfreepath = true;
2872         return 0;
2873 }
2874
2875 /*
2876  * request arguments may be specified via an inode *, a dentry *, or
2877  * an explicit ino+path.
2878  */
2879 static int set_request_path_attr(struct ceph_mds_client *mdsc, struct inode *rinode,
2880                                  struct dentry *rdentry, struct inode *rdiri,
2881                                  const char *rpath, u64 rino, const char **ppath,
2882                                  int *pathlen, u64 *ino, bool *freepath,
2883                                  bool parent_locked)
2884 {
2885         struct ceph_client *cl = mdsc->fsc->client;
2886         int r = 0;
2887
2888         if (rinode) {
2889                 r = build_inode_path(rinode, ppath, pathlen, ino, freepath);
2890                 doutc(cl, " inode %p %llx.%llx\n", rinode, ceph_ino(rinode),
2891                       ceph_snap(rinode));
2892         } else if (rdentry) {
2893                 r = build_dentry_path(mdsc, rdentry, rdiri, ppath, pathlen, ino,
2894                                         freepath, parent_locked);
2895                 doutc(cl, " dentry %p %llx/%.*s\n", rdentry, *ino, *pathlen, *ppath);
2896         } else if (rpath || rino) {
2897                 *ino = rino;
2898                 *ppath = rpath;
2899                 *pathlen = rpath ? strlen(rpath) : 0;
2900                 doutc(cl, " path %.*s\n", *pathlen, rpath);
2901         }
2902
2903         return r;
2904 }
2905
2906 static void encode_mclientrequest_tail(void **p,
2907                                        const struct ceph_mds_request *req)
2908 {
2909         struct ceph_timespec ts;
2910         int i;
2911
2912         ceph_encode_timespec64(&ts, &req->r_stamp);
2913         ceph_encode_copy(p, &ts, sizeof(ts));
2914
2915         /* v4: gid_list */
2916         ceph_encode_32(p, req->r_cred->group_info->ngroups);
2917         for (i = 0; i < req->r_cred->group_info->ngroups; i++)
2918                 ceph_encode_64(p, from_kgid(&init_user_ns,
2919                                             req->r_cred->group_info->gid[i]));
2920
2921         /* v5: altname */
2922         ceph_encode_32(p, req->r_altname_len);
2923         ceph_encode_copy(p, req->r_altname, req->r_altname_len);
2924
2925         /* v6: fscrypt_auth and fscrypt_file */
2926         if (req->r_fscrypt_auth) {
2927                 u32 authlen = ceph_fscrypt_auth_len(req->r_fscrypt_auth);
2928
2929                 ceph_encode_32(p, authlen);
2930                 ceph_encode_copy(p, req->r_fscrypt_auth, authlen);
2931         } else {
2932                 ceph_encode_32(p, 0);
2933         }
2934         if (test_bit(CEPH_MDS_R_FSCRYPT_FILE, &req->r_req_flags)) {
2935                 ceph_encode_32(p, sizeof(__le64));
2936                 ceph_encode_64(p, req->r_fscrypt_file);
2937         } else {
2938                 ceph_encode_32(p, 0);
2939         }
2940 }
2941
2942 static inline u16 mds_supported_head_version(struct ceph_mds_session *session)
2943 {
2944         if (!test_bit(CEPHFS_FEATURE_32BITS_RETRY_FWD, &session->s_features))
2945                 return 1;
2946
2947         if (!test_bit(CEPHFS_FEATURE_HAS_OWNER_UIDGID, &session->s_features))
2948                 return 2;
2949
2950         return CEPH_MDS_REQUEST_HEAD_VERSION;
2951 }
2952
2953 static struct ceph_mds_request_head_legacy *
2954 find_legacy_request_head(void *p, u64 features)
2955 {
2956         bool legacy = !(features & CEPH_FEATURE_FS_BTIME);
2957         struct ceph_mds_request_head_old *ohead;
2958
2959         if (legacy)
2960                 return (struct ceph_mds_request_head_legacy *)p;
2961         ohead = (struct ceph_mds_request_head_old *)p;
2962         return (struct ceph_mds_request_head_legacy *)&ohead->oldest_client_tid;
2963 }
2964
2965 /*
2966  * called under mdsc->mutex
2967  */
2968 static struct ceph_msg *create_request_message(struct ceph_mds_session *session,
2969                                                struct ceph_mds_request *req,
2970                                                bool drop_cap_releases)
2971 {
2972         int mds = session->s_mds;
2973         struct ceph_mds_client *mdsc = session->s_mdsc;
2974         struct ceph_client *cl = mdsc->fsc->client;
2975         struct ceph_msg *msg;
2976         struct ceph_mds_request_head_legacy *lhead;
2977         const char *path1 = NULL;
2978         const char *path2 = NULL;
2979         u64 ino1 = 0, ino2 = 0;
2980         int pathlen1 = 0, pathlen2 = 0;
2981         bool freepath1 = false, freepath2 = false;
2982         struct dentry *old_dentry = NULL;
2983         int len;
2984         u16 releases;
2985         void *p, *end;
2986         int ret;
2987         bool legacy = !(session->s_con.peer_features & CEPH_FEATURE_FS_BTIME);
2988         u16 request_head_version = mds_supported_head_version(session);
2989         kuid_t caller_fsuid = req->r_cred->fsuid;
2990         kgid_t caller_fsgid = req->r_cred->fsgid;
2991
2992         ret = set_request_path_attr(mdsc, req->r_inode, req->r_dentry,
2993                               req->r_parent, req->r_path1, req->r_ino1.ino,
2994                               &path1, &pathlen1, &ino1, &freepath1,
2995                               test_bit(CEPH_MDS_R_PARENT_LOCKED,
2996                                         &req->r_req_flags));
2997         if (ret < 0) {
2998                 msg = ERR_PTR(ret);
2999                 goto out;
3000         }
3001
3002         /* If r_old_dentry is set, then assume that its parent is locked */
3003         if (req->r_old_dentry &&
3004             !(req->r_old_dentry->d_flags & DCACHE_DISCONNECTED))
3005                 old_dentry = req->r_old_dentry;
3006         ret = set_request_path_attr(mdsc, NULL, old_dentry,
3007                               req->r_old_dentry_dir,
3008                               req->r_path2, req->r_ino2.ino,
3009                               &path2, &pathlen2, &ino2, &freepath2, true);
3010         if (ret < 0) {
3011                 msg = ERR_PTR(ret);
3012                 goto out_free1;
3013         }
3014
3015         req->r_altname = get_fscrypt_altname(req, &req->r_altname_len);
3016         if (IS_ERR(req->r_altname)) {
3017                 msg = ERR_CAST(req->r_altname);
3018                 req->r_altname = NULL;
3019                 goto out_free2;
3020         }
3021
3022         /*
3023          * For old cephs without supporting the 32bit retry/fwd feature
3024          * it will copy the raw memories directly when decoding the
3025          * requests. While new cephs will decode the head depending the
3026          * version member, so we need to make sure it will be compatible
3027          * with them both.
3028          */
3029         if (legacy)
3030                 len = sizeof(struct ceph_mds_request_head_legacy);
3031         else if (request_head_version == 1)
3032                 len = sizeof(struct ceph_mds_request_head_old);
3033         else if (request_head_version == 2)
3034                 len = offsetofend(struct ceph_mds_request_head, ext_num_fwd);
3035         else
3036                 len = sizeof(struct ceph_mds_request_head);
3037
3038         /* filepaths */
3039         len += 2 * (1 + sizeof(u32) + sizeof(u64));
3040         len += pathlen1 + pathlen2;
3041
3042         /* cap releases */
3043         len += sizeof(struct ceph_mds_request_release) *
3044                 (!!req->r_inode_drop + !!req->r_dentry_drop +
3045                  !!req->r_old_inode_drop + !!req->r_old_dentry_drop);
3046
3047         if (req->r_dentry_drop)
3048                 len += pathlen1;
3049         if (req->r_old_dentry_drop)
3050                 len += pathlen2;
3051
3052         /* MClientRequest tail */
3053
3054         /* req->r_stamp */
3055         len += sizeof(struct ceph_timespec);
3056
3057         /* gid list */
3058         len += sizeof(u32) + (sizeof(u64) * req->r_cred->group_info->ngroups);
3059
3060         /* alternate name */
3061         len += sizeof(u32) + req->r_altname_len;
3062
3063         /* fscrypt_auth */
3064         len += sizeof(u32); // fscrypt_auth
3065         if (req->r_fscrypt_auth)
3066                 len += ceph_fscrypt_auth_len(req->r_fscrypt_auth);
3067
3068         /* fscrypt_file */
3069         len += sizeof(u32);
3070         if (test_bit(CEPH_MDS_R_FSCRYPT_FILE, &req->r_req_flags))
3071                 len += sizeof(__le64);
3072
3073         msg = ceph_msg_new2(CEPH_MSG_CLIENT_REQUEST, len, 1, GFP_NOFS, false);
3074         if (!msg) {
3075                 msg = ERR_PTR(-ENOMEM);
3076                 goto out_free2;
3077         }
3078
3079         msg->hdr.tid = cpu_to_le64(req->r_tid);
3080
3081         lhead = find_legacy_request_head(msg->front.iov_base,
3082                                          session->s_con.peer_features);
3083
3084         if ((req->r_mnt_idmap != &nop_mnt_idmap) &&
3085             !test_bit(CEPHFS_FEATURE_HAS_OWNER_UIDGID, &session->s_features)) {
3086                 WARN_ON_ONCE(!IS_CEPH_MDS_OP_NEWINODE(req->r_op));
3087
3088                 if (enable_unsafe_idmap) {
3089                         pr_warn_once_client(cl,
3090                                 "idmapped mount is used and CEPHFS_FEATURE_HAS_OWNER_UIDGID"
3091                                 " is not supported by MDS. UID/GID-based restrictions may"
3092                                 " not work properly.\n");
3093
3094                         caller_fsuid = from_vfsuid(req->r_mnt_idmap, &init_user_ns,
3095                                                    VFSUIDT_INIT(req->r_cred->fsuid));
3096                         caller_fsgid = from_vfsgid(req->r_mnt_idmap, &init_user_ns,
3097                                                    VFSGIDT_INIT(req->r_cred->fsgid));
3098                 } else {
3099                         pr_err_ratelimited_client(cl,
3100                                 "idmapped mount is used and CEPHFS_FEATURE_HAS_OWNER_UIDGID"
3101                                 " is not supported by MDS. Fail request with -EIO.\n");
3102
3103                         ret = -EIO;
3104                         goto out_err;
3105                 }
3106         }
3107
3108         /*
3109          * The ceph_mds_request_head_legacy didn't contain a version field, and
3110          * one was added when we moved the message version from 3->4.
3111          */
3112         if (legacy) {
3113                 msg->hdr.version = cpu_to_le16(3);
3114                 p = msg->front.iov_base + sizeof(*lhead);
3115         } else if (request_head_version == 1) {
3116                 struct ceph_mds_request_head_old *ohead = msg->front.iov_base;
3117
3118                 msg->hdr.version = cpu_to_le16(4);
3119                 ohead->version = cpu_to_le16(1);
3120                 p = msg->front.iov_base + sizeof(*ohead);
3121         } else if (request_head_version == 2) {
3122                 struct ceph_mds_request_head *nhead = msg->front.iov_base;
3123
3124                 msg->hdr.version = cpu_to_le16(6);
3125                 nhead->version = cpu_to_le16(2);
3126
3127                 p = msg->front.iov_base + offsetofend(struct ceph_mds_request_head, ext_num_fwd);
3128         } else {
3129                 struct ceph_mds_request_head *nhead = msg->front.iov_base;
3130                 kuid_t owner_fsuid;
3131                 kgid_t owner_fsgid;
3132
3133                 msg->hdr.version = cpu_to_le16(6);
3134                 nhead->version = cpu_to_le16(CEPH_MDS_REQUEST_HEAD_VERSION);
3135                 nhead->struct_len = cpu_to_le32(sizeof(struct ceph_mds_request_head));
3136
3137                 if (IS_CEPH_MDS_OP_NEWINODE(req->r_op)) {
3138                         owner_fsuid = from_vfsuid(req->r_mnt_idmap, &init_user_ns,
3139                                                 VFSUIDT_INIT(req->r_cred->fsuid));
3140                         owner_fsgid = from_vfsgid(req->r_mnt_idmap, &init_user_ns,
3141                                                 VFSGIDT_INIT(req->r_cred->fsgid));
3142                         nhead->owner_uid = cpu_to_le32(from_kuid(&init_user_ns, owner_fsuid));
3143                         nhead->owner_gid = cpu_to_le32(from_kgid(&init_user_ns, owner_fsgid));
3144                 } else {
3145                         nhead->owner_uid = cpu_to_le32(-1);
3146                         nhead->owner_gid = cpu_to_le32(-1);
3147                 }
3148
3149                 p = msg->front.iov_base + sizeof(*nhead);
3150         }
3151
3152         end = msg->front.iov_base + msg->front.iov_len;
3153
3154         lhead->mdsmap_epoch = cpu_to_le32(mdsc->mdsmap->m_epoch);
3155         lhead->op = cpu_to_le32(req->r_op);
3156         lhead->caller_uid = cpu_to_le32(from_kuid(&init_user_ns,
3157                                                   caller_fsuid));
3158         lhead->caller_gid = cpu_to_le32(from_kgid(&init_user_ns,
3159                                                   caller_fsgid));
3160         lhead->ino = cpu_to_le64(req->r_deleg_ino);
3161         lhead->args = req->r_args;
3162
3163         ceph_encode_filepath(&p, end, ino1, path1);
3164         ceph_encode_filepath(&p, end, ino2, path2);
3165
3166         /* make note of release offset, in case we need to replay */
3167         req->r_request_release_offset = p - msg->front.iov_base;
3168
3169         /* cap releases */
3170         releases = 0;
3171         if (req->r_inode_drop)
3172                 releases += ceph_encode_inode_release(&p,
3173                       req->r_inode ? req->r_inode : d_inode(req->r_dentry),
3174                       mds, req->r_inode_drop, req->r_inode_unless,
3175                       req->r_op == CEPH_MDS_OP_READDIR);
3176         if (req->r_dentry_drop) {
3177                 ret = ceph_encode_dentry_release(&p, req->r_dentry,
3178                                 req->r_parent, mds, req->r_dentry_drop,
3179                                 req->r_dentry_unless);
3180                 if (ret < 0)
3181                         goto out_err;
3182                 releases += ret;
3183         }
3184         if (req->r_old_dentry_drop) {
3185                 ret = ceph_encode_dentry_release(&p, req->r_old_dentry,
3186                                 req->r_old_dentry_dir, mds,
3187                                 req->r_old_dentry_drop,
3188                                 req->r_old_dentry_unless);
3189                 if (ret < 0)
3190                         goto out_err;
3191                 releases += ret;
3192         }
3193         if (req->r_old_inode_drop)
3194                 releases += ceph_encode_inode_release(&p,
3195                       d_inode(req->r_old_dentry),
3196                       mds, req->r_old_inode_drop, req->r_old_inode_unless, 0);
3197
3198         if (drop_cap_releases) {
3199                 releases = 0;
3200                 p = msg->front.iov_base + req->r_request_release_offset;
3201         }
3202
3203         lhead->num_releases = cpu_to_le16(releases);
3204
3205         encode_mclientrequest_tail(&p, req);
3206
3207         if (WARN_ON_ONCE(p > end)) {
3208                 ceph_msg_put(msg);
3209                 msg = ERR_PTR(-ERANGE);
3210                 goto out_free2;
3211         }
3212
3213         msg->front.iov_len = p - msg->front.iov_base;
3214         msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
3215
3216         if (req->r_pagelist) {
3217                 struct ceph_pagelist *pagelist = req->r_pagelist;
3218                 ceph_msg_data_add_pagelist(msg, pagelist);
3219                 msg->hdr.data_len = cpu_to_le32(pagelist->length);
3220         } else {
3221                 msg->hdr.data_len = 0;
3222         }
3223
3224         msg->hdr.data_off = cpu_to_le16(0);
3225
3226 out_free2:
3227         if (freepath2)
3228                 ceph_mdsc_free_path((char *)path2, pathlen2);
3229 out_free1:
3230         if (freepath1)
3231                 ceph_mdsc_free_path((char *)path1, pathlen1);
3232 out:
3233         return msg;
3234 out_err:
3235         ceph_msg_put(msg);
3236         msg = ERR_PTR(ret);
3237         goto out_free2;
3238 }
3239
3240 /*
3241  * called under mdsc->mutex if error, under no mutex if
3242  * success.
3243  */
3244 static void complete_request(struct ceph_mds_client *mdsc,
3245                              struct ceph_mds_request *req)
3246 {
3247         req->r_end_latency = ktime_get();
3248
3249         if (req->r_callback)
3250                 req->r_callback(mdsc, req);
3251         complete_all(&req->r_completion);
3252 }
3253
3254 /*
3255  * called under mdsc->mutex
3256  */
3257 static int __prepare_send_request(struct ceph_mds_session *session,
3258                                   struct ceph_mds_request *req,
3259                                   bool drop_cap_releases)
3260 {
3261         int mds = session->s_mds;
3262         struct ceph_mds_client *mdsc = session->s_mdsc;
3263         struct ceph_client *cl = mdsc->fsc->client;
3264         struct ceph_mds_request_head_legacy *lhead;
3265         struct ceph_mds_request_head *nhead;
3266         struct ceph_msg *msg;
3267         int flags = 0, old_max_retry;
3268         bool old_version = !test_bit(CEPHFS_FEATURE_32BITS_RETRY_FWD,
3269                                      &session->s_features);
3270
3271         /*
3272          * Avoid inifinite retrying after overflow. The client will
3273          * increase the retry count and if the MDS is old version,
3274          * so we limit to retry at most 256 times.
3275          */
3276         if (req->r_attempts) {
3277                old_max_retry = sizeof_field(struct ceph_mds_request_head_old,
3278                                             num_retry);
3279                old_max_retry = 1 << (old_max_retry * BITS_PER_BYTE);
3280                if ((old_version && req->r_attempts >= old_max_retry) ||
3281                    ((uint32_t)req->r_attempts >= U32_MAX)) {
3282                         pr_warn_ratelimited_client(cl, "request tid %llu seq overflow\n",
3283                                                    req->r_tid);
3284                         return -EMULTIHOP;
3285                }
3286         }
3287
3288         req->r_attempts++;
3289         if (req->r_inode) {
3290                 struct ceph_cap *cap =
3291                         ceph_get_cap_for_mds(ceph_inode(req->r_inode), mds);
3292
3293                 if (cap)
3294                         req->r_sent_on_mseq = cap->mseq;
3295                 else
3296                         req->r_sent_on_mseq = -1;
3297         }
3298         doutc(cl, "%p tid %lld %s (attempt %d)\n", req, req->r_tid,
3299               ceph_mds_op_name(req->r_op), req->r_attempts);
3300
3301         if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
3302                 void *p;
3303
3304                 /*
3305                  * Replay.  Do not regenerate message (and rebuild
3306                  * paths, etc.); just use the original message.
3307                  * Rebuilding paths will break for renames because
3308                  * d_move mangles the src name.
3309                  */
3310                 msg = req->r_request;
3311                 lhead = find_legacy_request_head(msg->front.iov_base,
3312                                                  session->s_con.peer_features);
3313
3314                 flags = le32_to_cpu(lhead->flags);
3315                 flags |= CEPH_MDS_FLAG_REPLAY;
3316                 lhead->flags = cpu_to_le32(flags);
3317
3318                 if (req->r_target_inode)
3319                         lhead->ino = cpu_to_le64(ceph_ino(req->r_target_inode));
3320
3321                 lhead->num_retry = req->r_attempts - 1;
3322                 if (!old_version) {
3323                         nhead = (struct ceph_mds_request_head*)msg->front.iov_base;
3324                         nhead->ext_num_retry = cpu_to_le32(req->r_attempts - 1);
3325                 }
3326
3327                 /* remove cap/dentry releases from message */
3328                 lhead->num_releases = 0;
3329
3330                 p = msg->front.iov_base + req->r_request_release_offset;
3331                 encode_mclientrequest_tail(&p, req);
3332
3333                 msg->front.iov_len = p - msg->front.iov_base;
3334                 msg->hdr.front_len = cpu_to_le32(msg->front.iov_len);
3335                 return 0;
3336         }
3337
3338         if (req->r_request) {
3339                 ceph_msg_put(req->r_request);
3340                 req->r_request = NULL;
3341         }
3342         msg = create_request_message(session, req, drop_cap_releases);
3343         if (IS_ERR(msg)) {
3344                 req->r_err = PTR_ERR(msg);
3345                 return PTR_ERR(msg);
3346         }
3347         req->r_request = msg;
3348
3349         lhead = find_legacy_request_head(msg->front.iov_base,
3350                                          session->s_con.peer_features);
3351         lhead->oldest_client_tid = cpu_to_le64(__get_oldest_tid(mdsc));
3352         if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
3353                 flags |= CEPH_MDS_FLAG_REPLAY;
3354         if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags))
3355                 flags |= CEPH_MDS_FLAG_ASYNC;
3356         if (req->r_parent)
3357                 flags |= CEPH_MDS_FLAG_WANT_DENTRY;
3358         lhead->flags = cpu_to_le32(flags);
3359         lhead->num_fwd = req->r_num_fwd;
3360         lhead->num_retry = req->r_attempts - 1;
3361         if (!old_version) {
3362                 nhead = (struct ceph_mds_request_head*)msg->front.iov_base;
3363                 nhead->ext_num_fwd = cpu_to_le32(req->r_num_fwd);
3364                 nhead->ext_num_retry = cpu_to_le32(req->r_attempts - 1);
3365         }
3366
3367         doutc(cl, " r_parent = %p\n", req->r_parent);
3368         return 0;
3369 }
3370
3371 /*
3372  * called under mdsc->mutex
3373  */
3374 static int __send_request(struct ceph_mds_session *session,
3375                           struct ceph_mds_request *req,
3376                           bool drop_cap_releases)
3377 {
3378         int err;
3379
3380         err = __prepare_send_request(session, req, drop_cap_releases);
3381         if (!err) {
3382                 ceph_msg_get(req->r_request);
3383                 ceph_con_send(&session->s_con, req->r_request);
3384         }
3385
3386         return err;
3387 }
3388
3389 /*
3390  * send request, or put it on the appropriate wait list.
3391  */
3392 static void __do_request(struct ceph_mds_client *mdsc,
3393                         struct ceph_mds_request *req)
3394 {
3395         struct ceph_client *cl = mdsc->fsc->client;
3396         struct ceph_mds_session *session = NULL;
3397         int mds = -1;
3398         int err = 0;
3399         bool random;
3400
3401         if (req->r_err || test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags)) {
3402                 if (test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags))
3403                         __unregister_request(mdsc, req);
3404                 return;
3405         }
3406
3407         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_FENCE_IO) {
3408                 doutc(cl, "metadata corrupted\n");
3409                 err = -EIO;
3410                 goto finish;
3411         }
3412         if (req->r_timeout &&
3413             time_after_eq(jiffies, req->r_started + req->r_timeout)) {
3414                 doutc(cl, "timed out\n");
3415                 err = -ETIMEDOUT;
3416                 goto finish;
3417         }
3418         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
3419                 doutc(cl, "forced umount\n");
3420                 err = -EIO;
3421                 goto finish;
3422         }
3423         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_MOUNTING) {
3424                 if (mdsc->mdsmap_err) {
3425                         err = mdsc->mdsmap_err;
3426                         doutc(cl, "mdsmap err %d\n", err);
3427                         goto finish;
3428                 }
3429                 if (mdsc->mdsmap->m_epoch == 0) {
3430                         doutc(cl, "no mdsmap, waiting for map\n");
3431                         list_add(&req->r_wait, &mdsc->waiting_for_map);
3432                         return;
3433                 }
3434                 if (!(mdsc->fsc->mount_options->flags &
3435                       CEPH_MOUNT_OPT_MOUNTWAIT) &&
3436                     !ceph_mdsmap_is_cluster_available(mdsc->mdsmap)) {
3437                         err = -EHOSTUNREACH;
3438                         goto finish;
3439                 }
3440         }
3441
3442         put_request_session(req);
3443
3444         mds = __choose_mds(mdsc, req, &random);
3445         if (mds < 0 ||
3446             ceph_mdsmap_get_state(mdsc->mdsmap, mds) < CEPH_MDS_STATE_ACTIVE) {
3447                 if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags)) {
3448                         err = -EJUKEBOX;
3449                         goto finish;
3450                 }
3451                 doutc(cl, "no mds or not active, waiting for map\n");
3452                 list_add(&req->r_wait, &mdsc->waiting_for_map);
3453                 return;
3454         }
3455
3456         /* get, open session */
3457         session = __ceph_lookup_mds_session(mdsc, mds);
3458         if (!session) {
3459                 session = register_session(mdsc, mds);
3460                 if (IS_ERR(session)) {
3461                         err = PTR_ERR(session);
3462                         goto finish;
3463                 }
3464         }
3465         req->r_session = ceph_get_mds_session(session);
3466
3467         doutc(cl, "mds%d session %p state %s\n", mds, session,
3468               ceph_session_state_name(session->s_state));
3469
3470         /*
3471          * The old ceph will crash the MDSs when see unknown OPs
3472          */
3473         if (req->r_feature_needed > 0 &&
3474             !test_bit(req->r_feature_needed, &session->s_features)) {
3475                 err = -EOPNOTSUPP;
3476                 goto out_session;
3477         }
3478
3479         if (session->s_state != CEPH_MDS_SESSION_OPEN &&
3480             session->s_state != CEPH_MDS_SESSION_HUNG) {
3481                 /*
3482                  * We cannot queue async requests since the caps and delegated
3483                  * inodes are bound to the session. Just return -EJUKEBOX and
3484                  * let the caller retry a sync request in that case.
3485                  */
3486                 if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags)) {
3487                         err = -EJUKEBOX;
3488                         goto out_session;
3489                 }
3490
3491                 /*
3492                  * If the session has been REJECTED, then return a hard error,
3493                  * unless it's a CLEANRECOVER mount, in which case we'll queue
3494                  * it to the mdsc queue.
3495                  */
3496                 if (session->s_state == CEPH_MDS_SESSION_REJECTED) {
3497                         if (ceph_test_mount_opt(mdsc->fsc, CLEANRECOVER))
3498                                 list_add(&req->r_wait, &mdsc->waiting_for_map);
3499                         else
3500                                 err = -EACCES;
3501                         goto out_session;
3502                 }
3503
3504                 if (session->s_state == CEPH_MDS_SESSION_NEW ||
3505                     session->s_state == CEPH_MDS_SESSION_CLOSING) {
3506                         err = __open_session(mdsc, session);
3507                         if (err)
3508                                 goto out_session;
3509                         /* retry the same mds later */
3510                         if (random)
3511                                 req->r_resend_mds = mds;
3512                 }
3513                 list_add(&req->r_wait, &session->s_waiting);
3514                 goto out_session;
3515         }
3516
3517         /* send request */
3518         req->r_resend_mds = -1;   /* forget any previous mds hint */
3519
3520         if (req->r_request_started == 0)   /* note request start time */
3521                 req->r_request_started = jiffies;
3522
3523         /*
3524          * For async create we will choose the auth MDS of frag in parent
3525          * directory to send the request and ususally this works fine, but
3526          * if the migrated the dirtory to another MDS before it could handle
3527          * it the request will be forwarded.
3528          *
3529          * And then the auth cap will be changed.
3530          */
3531         if (test_bit(CEPH_MDS_R_ASYNC, &req->r_req_flags) && req->r_num_fwd) {
3532                 struct ceph_dentry_info *di = ceph_dentry(req->r_dentry);
3533                 struct ceph_inode_info *ci;
3534                 struct ceph_cap *cap;
3535
3536                 /*
3537                  * The request maybe handled very fast and the new inode
3538                  * hasn't been linked to the dentry yet. We need to wait
3539                  * for the ceph_finish_async_create(), which shouldn't be
3540                  * stuck too long or fail in thoery, to finish when forwarding
3541                  * the request.
3542                  */
3543                 if (!d_inode(req->r_dentry)) {
3544                         err = wait_on_bit(&di->flags, CEPH_DENTRY_ASYNC_CREATE_BIT,
3545                                           TASK_KILLABLE);
3546                         if (err) {
3547                                 mutex_lock(&req->r_fill_mutex);
3548                                 set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
3549                                 mutex_unlock(&req->r_fill_mutex);
3550                                 goto out_session;
3551                         }
3552                 }
3553
3554                 ci = ceph_inode(d_inode(req->r_dentry));
3555
3556                 spin_lock(&ci->i_ceph_lock);
3557                 cap = ci->i_auth_cap;
3558                 if (ci->i_ceph_flags & CEPH_I_ASYNC_CREATE && mds != cap->mds) {
3559                         doutc(cl, "session changed for auth cap %d -> %d\n",
3560                               cap->session->s_mds, session->s_mds);
3561
3562                         /* Remove the auth cap from old session */
3563                         spin_lock(&cap->session->s_cap_lock);
3564                         cap->session->s_nr_caps--;
3565                         list_del_init(&cap->session_caps);
3566                         spin_unlock(&cap->session->s_cap_lock);
3567
3568                         /* Add the auth cap to the new session */
3569                         cap->mds = mds;
3570                         cap->session = session;
3571                         spin_lock(&session->s_cap_lock);
3572                         session->s_nr_caps++;
3573                         list_add_tail(&cap->session_caps, &session->s_caps);
3574                         spin_unlock(&session->s_cap_lock);
3575
3576                         change_auth_cap_ses(ci, session);
3577                 }
3578                 spin_unlock(&ci->i_ceph_lock);
3579         }
3580
3581         err = __send_request(session, req, false);
3582
3583 out_session:
3584         ceph_put_mds_session(session);
3585 finish:
3586         if (err) {
3587                 doutc(cl, "early error %d\n", err);
3588                 req->r_err = err;
3589                 complete_request(mdsc, req);
3590                 __unregister_request(mdsc, req);
3591         }
3592         return;
3593 }
3594
3595 /*
3596  * called under mdsc->mutex
3597  */
3598 static void __wake_requests(struct ceph_mds_client *mdsc,
3599                             struct list_head *head)
3600 {
3601         struct ceph_client *cl = mdsc->fsc->client;
3602         struct ceph_mds_request *req;
3603         LIST_HEAD(tmp_list);
3604
3605         list_splice_init(head, &tmp_list);
3606
3607         while (!list_empty(&tmp_list)) {
3608                 req = list_entry(tmp_list.next,
3609                                  struct ceph_mds_request, r_wait);
3610                 list_del_init(&req->r_wait);
3611                 doutc(cl, " wake request %p tid %llu\n", req,
3612                       req->r_tid);
3613                 __do_request(mdsc, req);
3614         }
3615 }
3616
3617 /*
3618  * Wake up threads with requests pending for @mds, so that they can
3619  * resubmit their requests to a possibly different mds.
3620  */
3621 static void kick_requests(struct ceph_mds_client *mdsc, int mds)
3622 {
3623         struct ceph_client *cl = mdsc->fsc->client;
3624         struct ceph_mds_request *req;
3625         struct rb_node *p = rb_first(&mdsc->request_tree);
3626
3627         doutc(cl, "kick_requests mds%d\n", mds);
3628         while (p) {
3629                 req = rb_entry(p, struct ceph_mds_request, r_node);
3630                 p = rb_next(p);
3631                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
3632                         continue;
3633                 if (req->r_attempts > 0)
3634                         continue; /* only new requests */
3635                 if (req->r_session &&
3636                     req->r_session->s_mds == mds) {
3637                         doutc(cl, " kicking tid %llu\n", req->r_tid);
3638                         list_del_init(&req->r_wait);
3639                         __do_request(mdsc, req);
3640                 }
3641         }
3642 }
3643
3644 int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc, struct inode *dir,
3645                               struct ceph_mds_request *req)
3646 {
3647         struct ceph_client *cl = mdsc->fsc->client;
3648         int err = 0;
3649
3650         /* take CAP_PIN refs for r_inode, r_parent, r_old_dentry */
3651         if (req->r_inode)
3652                 ceph_get_cap_refs(ceph_inode(req->r_inode), CEPH_CAP_PIN);
3653         if (req->r_parent) {
3654                 struct ceph_inode_info *ci = ceph_inode(req->r_parent);
3655                 int fmode = (req->r_op & CEPH_MDS_OP_WRITE) ?
3656                             CEPH_FILE_MODE_WR : CEPH_FILE_MODE_RD;
3657                 spin_lock(&ci->i_ceph_lock);
3658                 ceph_take_cap_refs(ci, CEPH_CAP_PIN, false);
3659                 __ceph_touch_fmode(ci, mdsc, fmode);
3660                 spin_unlock(&ci->i_ceph_lock);
3661         }
3662         if (req->r_old_dentry_dir)
3663                 ceph_get_cap_refs(ceph_inode(req->r_old_dentry_dir),
3664                                   CEPH_CAP_PIN);
3665
3666         if (req->r_inode) {
3667                 err = ceph_wait_on_async_create(req->r_inode);
3668                 if (err) {
3669                         doutc(cl, "wait for async create returned: %d\n", err);
3670                         return err;
3671                 }
3672         }
3673
3674         if (!err && req->r_old_inode) {
3675                 err = ceph_wait_on_async_create(req->r_old_inode);
3676                 if (err) {
3677                         doutc(cl, "wait for async create returned: %d\n", err);
3678                         return err;
3679                 }
3680         }
3681
3682         doutc(cl, "submit_request on %p for inode %p\n", req, dir);
3683         mutex_lock(&mdsc->mutex);
3684         __register_request(mdsc, req, dir);
3685         __do_request(mdsc, req);
3686         err = req->r_err;
3687         mutex_unlock(&mdsc->mutex);
3688         return err;
3689 }
3690
3691 int ceph_mdsc_wait_request(struct ceph_mds_client *mdsc,
3692                            struct ceph_mds_request *req,
3693                            ceph_mds_request_wait_callback_t wait_func)
3694 {
3695         struct ceph_client *cl = mdsc->fsc->client;
3696         int err;
3697
3698         /* wait */
3699         doutc(cl, "do_request waiting\n");
3700         if (wait_func) {
3701                 err = wait_func(mdsc, req);
3702         } else {
3703                 long timeleft = wait_for_completion_killable_timeout(
3704                                         &req->r_completion,
3705                                         ceph_timeout_jiffies(req->r_timeout));
3706                 if (timeleft > 0)
3707                         err = 0;
3708                 else if (!timeleft)
3709                         err = -ETIMEDOUT;  /* timed out */
3710                 else
3711                         err = timeleft;  /* killed */
3712         }
3713         doutc(cl, "do_request waited, got %d\n", err);
3714         mutex_lock(&mdsc->mutex);
3715
3716         /* only abort if we didn't race with a real reply */
3717         if (test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags)) {
3718                 err = le32_to_cpu(req->r_reply_info.head->result);
3719         } else if (err < 0) {
3720                 doutc(cl, "aborted request %lld with %d\n", req->r_tid, err);
3721
3722                 /*
3723                  * ensure we aren't running concurrently with
3724                  * ceph_fill_trace or ceph_readdir_prepopulate, which
3725                  * rely on locks (dir mutex) held by our caller.
3726                  */
3727                 mutex_lock(&req->r_fill_mutex);
3728                 req->r_err = err;
3729                 set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
3730                 mutex_unlock(&req->r_fill_mutex);
3731
3732                 if (req->r_parent &&
3733                     (req->r_op & CEPH_MDS_OP_WRITE))
3734                         ceph_invalidate_dir_request(req);
3735         } else {
3736                 err = req->r_err;
3737         }
3738
3739         mutex_unlock(&mdsc->mutex);
3740         return err;
3741 }
3742
3743 /*
3744  * Synchrously perform an mds request.  Take care of all of the
3745  * session setup, forwarding, retry details.
3746  */
3747 int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
3748                          struct inode *dir,
3749                          struct ceph_mds_request *req)
3750 {
3751         struct ceph_client *cl = mdsc->fsc->client;
3752         int err;
3753
3754         doutc(cl, "do_request on %p\n", req);
3755
3756         /* issue */
3757         err = ceph_mdsc_submit_request(mdsc, dir, req);
3758         if (!err)
3759                 err = ceph_mdsc_wait_request(mdsc, req, NULL);
3760         doutc(cl, "do_request %p done, result %d\n", req, err);
3761         return err;
3762 }
3763
3764 /*
3765  * Invalidate dir's completeness, dentry lease state on an aborted MDS
3766  * namespace request.
3767  */
3768 void ceph_invalidate_dir_request(struct ceph_mds_request *req)
3769 {
3770         struct inode *dir = req->r_parent;
3771         struct inode *old_dir = req->r_old_dentry_dir;
3772         struct ceph_client *cl = req->r_mdsc->fsc->client;
3773
3774         doutc(cl, "invalidate_dir_request %p %p (complete, lease(s))\n",
3775               dir, old_dir);
3776
3777         ceph_dir_clear_complete(dir);
3778         if (old_dir)
3779                 ceph_dir_clear_complete(old_dir);
3780         if (req->r_dentry)
3781                 ceph_invalidate_dentry_lease(req->r_dentry);
3782         if (req->r_old_dentry)
3783                 ceph_invalidate_dentry_lease(req->r_old_dentry);
3784 }
3785
3786 /*
3787  * Handle mds reply.
3788  *
3789  * We take the session mutex and parse and process the reply immediately.
3790  * This preserves the logical ordering of replies, capabilities, etc., sent
3791  * by the MDS as they are applied to our local cache.
3792  */
3793 static void handle_reply(struct ceph_mds_session *session, struct ceph_msg *msg)
3794 {
3795         struct ceph_mds_client *mdsc = session->s_mdsc;
3796         struct ceph_client *cl = mdsc->fsc->client;
3797         struct ceph_mds_request *req;
3798         struct ceph_mds_reply_head *head = msg->front.iov_base;
3799         struct ceph_mds_reply_info_parsed *rinfo;  /* parsed reply info */
3800         struct ceph_snap_realm *realm;
3801         u64 tid;
3802         int err, result;
3803         int mds = session->s_mds;
3804         bool close_sessions = false;
3805
3806         if (msg->front.iov_len < sizeof(*head)) {
3807                 pr_err_client(cl, "got corrupt (short) reply\n");
3808                 ceph_msg_dump(msg);
3809                 return;
3810         }
3811
3812         /* get request, session */
3813         tid = le64_to_cpu(msg->hdr.tid);
3814         mutex_lock(&mdsc->mutex);
3815         req = lookup_get_request(mdsc, tid);
3816         if (!req) {
3817                 doutc(cl, "on unknown tid %llu\n", tid);
3818                 mutex_unlock(&mdsc->mutex);
3819                 return;
3820         }
3821         doutc(cl, "handle_reply %p\n", req);
3822
3823         /* correct session? */
3824         if (req->r_session != session) {
3825                 pr_err_client(cl, "got %llu on session mds%d not mds%d\n",
3826                               tid, session->s_mds,
3827                               req->r_session ? req->r_session->s_mds : -1);
3828                 mutex_unlock(&mdsc->mutex);
3829                 goto out;
3830         }
3831
3832         /* dup? */
3833         if ((test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags) && !head->safe) ||
3834             (test_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags) && head->safe)) {
3835                 pr_warn_client(cl, "got a dup %s reply on %llu from mds%d\n",
3836                                head->safe ? "safe" : "unsafe", tid, mds);
3837                 mutex_unlock(&mdsc->mutex);
3838                 goto out;
3839         }
3840         if (test_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags)) {
3841                 pr_warn_client(cl, "got unsafe after safe on %llu from mds%d\n",
3842                                tid, mds);
3843                 mutex_unlock(&mdsc->mutex);
3844                 goto out;
3845         }
3846
3847         result = le32_to_cpu(head->result);
3848
3849         if (head->safe) {
3850                 set_bit(CEPH_MDS_R_GOT_SAFE, &req->r_req_flags);
3851                 __unregister_request(mdsc, req);
3852
3853                 /* last request during umount? */
3854                 if (mdsc->stopping && !__get_oldest_req(mdsc))
3855                         complete_all(&mdsc->safe_umount_waiters);
3856
3857                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
3858                         /*
3859                          * We already handled the unsafe response, now do the
3860                          * cleanup.  No need to examine the response; the MDS
3861                          * doesn't include any result info in the safe
3862                          * response.  And even if it did, there is nothing
3863                          * useful we could do with a revised return value.
3864                          */
3865                         doutc(cl, "got safe reply %llu, mds%d\n", tid, mds);
3866
3867                         mutex_unlock(&mdsc->mutex);
3868                         goto out;
3869                 }
3870         } else {
3871                 set_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags);
3872                 list_add_tail(&req->r_unsafe_item, &req->r_session->s_unsafe);
3873         }
3874
3875         doutc(cl, "tid %lld result %d\n", tid, result);
3876         if (test_bit(CEPHFS_FEATURE_REPLY_ENCODING, &session->s_features))
3877                 err = parse_reply_info(session, msg, req, (u64)-1);
3878         else
3879                 err = parse_reply_info(session, msg, req,
3880                                        session->s_con.peer_features);
3881         mutex_unlock(&mdsc->mutex);
3882
3883         /* Must find target inode outside of mutexes to avoid deadlocks */
3884         rinfo = &req->r_reply_info;
3885         if ((err >= 0) && rinfo->head->is_target) {
3886                 struct inode *in = xchg(&req->r_new_inode, NULL);
3887                 struct ceph_vino tvino = {
3888                         .ino  = le64_to_cpu(rinfo->targeti.in->ino),
3889                         .snap = le64_to_cpu(rinfo->targeti.in->snapid)
3890                 };
3891
3892                 /*
3893                  * If we ended up opening an existing inode, discard
3894                  * r_new_inode
3895                  */
3896                 if (req->r_op == CEPH_MDS_OP_CREATE &&
3897                     !req->r_reply_info.has_create_ino) {
3898                         /* This should never happen on an async create */
3899                         WARN_ON_ONCE(req->r_deleg_ino);
3900                         iput(in);
3901                         in = NULL;
3902                 }
3903
3904                 in = ceph_get_inode(mdsc->fsc->sb, tvino, in);
3905                 if (IS_ERR(in)) {
3906                         err = PTR_ERR(in);
3907                         mutex_lock(&session->s_mutex);
3908                         goto out_err;
3909                 }
3910                 req->r_target_inode = in;
3911         }
3912
3913         mutex_lock(&session->s_mutex);
3914         if (err < 0) {
3915                 pr_err_client(cl, "got corrupt reply mds%d(tid:%lld)\n",
3916                               mds, tid);
3917                 ceph_msg_dump(msg);
3918                 goto out_err;
3919         }
3920
3921         /* snap trace */
3922         realm = NULL;
3923         if (rinfo->snapblob_len) {
3924                 down_write(&mdsc->snap_rwsem);
3925                 err = ceph_update_snap_trace(mdsc, rinfo->snapblob,
3926                                 rinfo->snapblob + rinfo->snapblob_len,
3927                                 le32_to_cpu(head->op) == CEPH_MDS_OP_RMSNAP,
3928                                 &realm);
3929                 if (err) {
3930                         up_write(&mdsc->snap_rwsem);
3931                         close_sessions = true;
3932                         if (err == -EIO)
3933                                 ceph_msg_dump(msg);
3934                         goto out_err;
3935                 }
3936                 downgrade_write(&mdsc->snap_rwsem);
3937         } else {
3938                 down_read(&mdsc->snap_rwsem);
3939         }
3940
3941         /* insert trace into our cache */
3942         mutex_lock(&req->r_fill_mutex);
3943         current->journal_info = req;
3944         err = ceph_fill_trace(mdsc->fsc->sb, req);
3945         if (err == 0) {
3946                 if (result == 0 && (req->r_op == CEPH_MDS_OP_READDIR ||
3947                                     req->r_op == CEPH_MDS_OP_LSSNAP))
3948                         err = ceph_readdir_prepopulate(req, req->r_session);
3949         }
3950         current->journal_info = NULL;
3951         mutex_unlock(&req->r_fill_mutex);
3952
3953         up_read(&mdsc->snap_rwsem);
3954         if (realm)
3955                 ceph_put_snap_realm(mdsc, realm);
3956
3957         if (err == 0) {
3958                 if (req->r_target_inode &&
3959                     test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags)) {
3960                         struct ceph_inode_info *ci =
3961                                 ceph_inode(req->r_target_inode);
3962                         spin_lock(&ci->i_unsafe_lock);
3963                         list_add_tail(&req->r_unsafe_target_item,
3964                                       &ci->i_unsafe_iops);
3965                         spin_unlock(&ci->i_unsafe_lock);
3966                 }
3967
3968                 ceph_unreserve_caps(mdsc, &req->r_caps_reservation);
3969         }
3970 out_err:
3971         mutex_lock(&mdsc->mutex);
3972         if (!test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) {
3973                 if (err) {
3974                         req->r_err = err;
3975                 } else {
3976                         req->r_reply =  ceph_msg_get(msg);
3977                         set_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags);
3978                 }
3979         } else {
3980                 doutc(cl, "reply arrived after request %lld was aborted\n", tid);
3981         }
3982         mutex_unlock(&mdsc->mutex);
3983
3984         mutex_unlock(&session->s_mutex);
3985
3986         /* kick calling process */
3987         complete_request(mdsc, req);
3988
3989         ceph_update_metadata_metrics(&mdsc->metric, req->r_start_latency,
3990                                      req->r_end_latency, err);
3991 out:
3992         ceph_mdsc_put_request(req);
3993
3994         /* Defer closing the sessions after s_mutex lock being released */
3995         if (close_sessions)
3996                 ceph_mdsc_close_sessions(mdsc);
3997         return;
3998 }
3999
4000
4001
4002 /*
4003  * handle mds notification that our request has been forwarded.
4004  */
4005 static void handle_forward(struct ceph_mds_client *mdsc,
4006                            struct ceph_mds_session *session,
4007                            struct ceph_msg *msg)
4008 {
4009         struct ceph_client *cl = mdsc->fsc->client;
4010         struct ceph_mds_request *req;
4011         u64 tid = le64_to_cpu(msg->hdr.tid);
4012         u32 next_mds;
4013         u32 fwd_seq;
4014         int err = -EINVAL;
4015         void *p = msg->front.iov_base;
4016         void *end = p + msg->front.iov_len;
4017         bool aborted = false;
4018
4019         ceph_decode_need(&p, end, 2*sizeof(u32), bad);
4020         next_mds = ceph_decode_32(&p);
4021         fwd_seq = ceph_decode_32(&p);
4022
4023         mutex_lock(&mdsc->mutex);
4024         req = lookup_get_request(mdsc, tid);
4025         if (!req) {
4026                 mutex_unlock(&mdsc->mutex);
4027                 doutc(cl, "forward tid %llu to mds%d - req dne\n", tid, next_mds);
4028                 return;  /* dup reply? */
4029         }
4030
4031         if (test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) {
4032                 doutc(cl, "forward tid %llu aborted, unregistering\n", tid);
4033                 __unregister_request(mdsc, req);
4034         } else if (fwd_seq <= req->r_num_fwd || (uint32_t)fwd_seq >= U32_MAX) {
4035                 /*
4036                  * Avoid inifinite retrying after overflow.
4037                  *
4038                  * The MDS will increase the fwd count and in client side
4039                  * if the num_fwd is less than the one saved in request
4040                  * that means the MDS is an old version and overflowed of
4041                  * 8 bits.
4042                  */
4043                 mutex_lock(&req->r_fill_mutex);
4044                 req->r_err = -EMULTIHOP;
4045                 set_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags);
4046                 mutex_unlock(&req->r_fill_mutex);
4047                 aborted = true;
4048                 pr_warn_ratelimited_client(cl, "forward tid %llu seq overflow\n",
4049                                            tid);
4050         } else {
4051                 /* resend. forward race not possible; mds would drop */
4052                 doutc(cl, "forward tid %llu to mds%d (we resend)\n", tid, next_mds);
4053                 BUG_ON(req->r_err);
4054                 BUG_ON(test_bit(CEPH_MDS_R_GOT_RESULT, &req->r_req_flags));
4055                 req->r_attempts = 0;
4056                 req->r_num_fwd = fwd_seq;
4057                 req->r_resend_mds = next_mds;
4058                 put_request_session(req);
4059                 __do_request(mdsc, req);
4060         }
4061         mutex_unlock(&mdsc->mutex);
4062
4063         /* kick calling process */
4064         if (aborted)
4065                 complete_request(mdsc, req);
4066         ceph_mdsc_put_request(req);
4067         return;
4068
4069 bad:
4070         pr_err_client(cl, "decode error err=%d\n", err);
4071         ceph_msg_dump(msg);
4072 }
4073
4074 static int __decode_session_metadata(void **p, void *end,
4075                                      bool *blocklisted)
4076 {
4077         /* map<string,string> */
4078         u32 n;
4079         bool err_str;
4080         ceph_decode_32_safe(p, end, n, bad);
4081         while (n-- > 0) {
4082                 u32 len;
4083                 ceph_decode_32_safe(p, end, len, bad);
4084                 ceph_decode_need(p, end, len, bad);
4085                 err_str = !strncmp(*p, "error_string", len);
4086                 *p += len;
4087                 ceph_decode_32_safe(p, end, len, bad);
4088                 ceph_decode_need(p, end, len, bad);
4089                 /*
4090                  * Match "blocklisted (blacklisted)" from newer MDSes,
4091                  * or "blacklisted" from older MDSes.
4092                  */
4093                 if (err_str && strnstr(*p, "blacklisted", len))
4094                         *blocklisted = true;
4095                 *p += len;
4096         }
4097         return 0;
4098 bad:
4099         return -1;
4100 }
4101
4102 /*
4103  * handle a mds session control message
4104  */
4105 static void handle_session(struct ceph_mds_session *session,
4106                            struct ceph_msg *msg)
4107 {
4108         struct ceph_mds_client *mdsc = session->s_mdsc;
4109         struct ceph_client *cl = mdsc->fsc->client;
4110         int mds = session->s_mds;
4111         int msg_version = le16_to_cpu(msg->hdr.version);
4112         void *p = msg->front.iov_base;
4113         void *end = p + msg->front.iov_len;
4114         struct ceph_mds_session_head *h;
4115         u32 op;
4116         u64 seq, features = 0;
4117         int wake = 0;
4118         bool blocklisted = false;
4119
4120         /* decode */
4121         ceph_decode_need(&p, end, sizeof(*h), bad);
4122         h = p;
4123         p += sizeof(*h);
4124
4125         op = le32_to_cpu(h->op);
4126         seq = le64_to_cpu(h->seq);
4127
4128         if (msg_version >= 3) {
4129                 u32 len;
4130                 /* version >= 2 and < 5, decode metadata, skip otherwise
4131                  * as it's handled via flags.
4132                  */
4133                 if (msg_version >= 5)
4134                         ceph_decode_skip_map(&p, end, string, string, bad);
4135                 else if (__decode_session_metadata(&p, end, &blocklisted) < 0)
4136                         goto bad;
4137
4138                 /* version >= 3, feature bits */
4139                 ceph_decode_32_safe(&p, end, len, bad);
4140                 if (len) {
4141                         ceph_decode_64_safe(&p, end, features, bad);
4142                         p += len - sizeof(features);
4143                 }
4144         }
4145
4146         if (msg_version >= 5) {
4147                 u32 flags, len;
4148
4149                 /* version >= 4 */
4150                 ceph_decode_skip_16(&p, end, bad); /* struct_v, struct_cv */
4151                 ceph_decode_32_safe(&p, end, len, bad); /* len */
4152                 ceph_decode_skip_n(&p, end, len, bad); /* metric_spec */
4153
4154                 /* version >= 5, flags   */
4155                 ceph_decode_32_safe(&p, end, flags, bad);
4156                 if (flags & CEPH_SESSION_BLOCKLISTED) {
4157                         pr_warn_client(cl, "mds%d session blocklisted\n",
4158                                        session->s_mds);
4159                         blocklisted = true;
4160                 }
4161         }
4162
4163         mutex_lock(&mdsc->mutex);
4164         if (op == CEPH_SESSION_CLOSE) {
4165                 ceph_get_mds_session(session);
4166                 __unregister_session(mdsc, session);
4167         }
4168         /* FIXME: this ttl calculation is generous */
4169         session->s_ttl = jiffies + HZ*mdsc->mdsmap->m_session_autoclose;
4170         mutex_unlock(&mdsc->mutex);
4171
4172         mutex_lock(&session->s_mutex);
4173
4174         doutc(cl, "mds%d %s %p state %s seq %llu\n", mds,
4175               ceph_session_op_name(op), session,
4176               ceph_session_state_name(session->s_state), seq);
4177
4178         if (session->s_state == CEPH_MDS_SESSION_HUNG) {
4179                 session->s_state = CEPH_MDS_SESSION_OPEN;
4180                 pr_info_client(cl, "mds%d came back\n", session->s_mds);
4181         }
4182
4183         switch (op) {
4184         case CEPH_SESSION_OPEN:
4185                 if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
4186                         pr_info_client(cl, "mds%d reconnect success\n",
4187                                        session->s_mds);
4188
4189                 session->s_features = features;
4190                 if (session->s_state == CEPH_MDS_SESSION_OPEN) {
4191                         pr_notice_client(cl, "mds%d is already opened\n",
4192                                          session->s_mds);
4193                 } else {
4194                         session->s_state = CEPH_MDS_SESSION_OPEN;
4195                         renewed_caps(mdsc, session, 0);
4196                         if (test_bit(CEPHFS_FEATURE_METRIC_COLLECT,
4197                                      &session->s_features))
4198                                 metric_schedule_delayed(&mdsc->metric);
4199                 }
4200
4201                 /*
4202                  * The connection maybe broken and the session in client
4203                  * side has been reinitialized, need to update the seq
4204                  * anyway.
4205                  */
4206                 if (!session->s_seq && seq)
4207                         session->s_seq = seq;
4208
4209                 wake = 1;
4210                 if (mdsc->stopping)
4211                         __close_session(mdsc, session);
4212                 break;
4213
4214         case CEPH_SESSION_RENEWCAPS:
4215                 if (session->s_renew_seq == seq)
4216                         renewed_caps(mdsc, session, 1);
4217                 break;
4218
4219         case CEPH_SESSION_CLOSE:
4220                 if (session->s_state == CEPH_MDS_SESSION_RECONNECTING)
4221                         pr_info_client(cl, "mds%d reconnect denied\n",
4222                                        session->s_mds);
4223                 session->s_state = CEPH_MDS_SESSION_CLOSED;
4224                 cleanup_session_requests(mdsc, session);
4225                 remove_session_caps(session);
4226                 wake = 2; /* for good measure */
4227                 wake_up_all(&mdsc->session_close_wq);
4228                 break;
4229
4230         case CEPH_SESSION_STALE:
4231                 pr_info_client(cl, "mds%d caps went stale, renewing\n",
4232                                session->s_mds);
4233                 atomic_inc(&session->s_cap_gen);
4234                 session->s_cap_ttl = jiffies - 1;
4235                 send_renew_caps(mdsc, session);
4236                 break;
4237
4238         case CEPH_SESSION_RECALL_STATE:
4239                 ceph_trim_caps(mdsc, session, le32_to_cpu(h->max_caps));
4240                 break;
4241
4242         case CEPH_SESSION_FLUSHMSG:
4243                 /* flush cap releases */
4244                 spin_lock(&session->s_cap_lock);
4245                 if (session->s_num_cap_releases)
4246                         ceph_flush_cap_releases(mdsc, session);
4247                 spin_unlock(&session->s_cap_lock);
4248
4249                 send_flushmsg_ack(mdsc, session, seq);
4250                 break;
4251
4252         case CEPH_SESSION_FORCE_RO:
4253                 doutc(cl, "force_session_readonly %p\n", session);
4254                 spin_lock(&session->s_cap_lock);
4255                 session->s_readonly = true;
4256                 spin_unlock(&session->s_cap_lock);
4257                 wake_up_session_caps(session, FORCE_RO);
4258                 break;
4259
4260         case CEPH_SESSION_REJECT:
4261                 WARN_ON(session->s_state != CEPH_MDS_SESSION_OPENING);
4262                 pr_info_client(cl, "mds%d rejected session\n",
4263                                session->s_mds);
4264                 session->s_state = CEPH_MDS_SESSION_REJECTED;
4265                 cleanup_session_requests(mdsc, session);
4266                 remove_session_caps(session);
4267                 if (blocklisted)
4268                         mdsc->fsc->blocklisted = true;
4269                 wake = 2; /* for good measure */
4270                 break;
4271
4272         default:
4273                 pr_err_client(cl, "bad op %d mds%d\n", op, mds);
4274                 WARN_ON(1);
4275         }
4276
4277         mutex_unlock(&session->s_mutex);
4278         if (wake) {
4279                 mutex_lock(&mdsc->mutex);
4280                 __wake_requests(mdsc, &session->s_waiting);
4281                 if (wake == 2)
4282                         kick_requests(mdsc, mds);
4283                 mutex_unlock(&mdsc->mutex);
4284         }
4285         if (op == CEPH_SESSION_CLOSE)
4286                 ceph_put_mds_session(session);
4287         return;
4288
4289 bad:
4290         pr_err_client(cl, "corrupt message mds%d len %d\n", mds,
4291                       (int)msg->front.iov_len);
4292         ceph_msg_dump(msg);
4293         return;
4294 }
4295
4296 void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req)
4297 {
4298         struct ceph_client *cl = req->r_mdsc->fsc->client;
4299         int dcaps;
4300
4301         dcaps = xchg(&req->r_dir_caps, 0);
4302         if (dcaps) {
4303                 doutc(cl, "releasing r_dir_caps=%s\n", ceph_cap_string(dcaps));
4304                 ceph_put_cap_refs(ceph_inode(req->r_parent), dcaps);
4305         }
4306 }
4307
4308 void ceph_mdsc_release_dir_caps_async(struct ceph_mds_request *req)
4309 {
4310         struct ceph_client *cl = req->r_mdsc->fsc->client;
4311         int dcaps;
4312
4313         dcaps = xchg(&req->r_dir_caps, 0);
4314         if (dcaps) {
4315                 doutc(cl, "releasing r_dir_caps=%s\n", ceph_cap_string(dcaps));
4316                 ceph_put_cap_refs_async(ceph_inode(req->r_parent), dcaps);
4317         }
4318 }
4319
4320 /*
4321  * called under session->mutex.
4322  */
4323 static void replay_unsafe_requests(struct ceph_mds_client *mdsc,
4324                                    struct ceph_mds_session *session)
4325 {
4326         struct ceph_mds_request *req, *nreq;
4327         struct rb_node *p;
4328
4329         doutc(mdsc->fsc->client, "mds%d\n", session->s_mds);
4330
4331         mutex_lock(&mdsc->mutex);
4332         list_for_each_entry_safe(req, nreq, &session->s_unsafe, r_unsafe_item)
4333                 __send_request(session, req, true);
4334
4335         /*
4336          * also re-send old requests when MDS enters reconnect stage. So that MDS
4337          * can process completed request in clientreplay stage.
4338          */
4339         p = rb_first(&mdsc->request_tree);
4340         while (p) {
4341                 req = rb_entry(p, struct ceph_mds_request, r_node);
4342                 p = rb_next(p);
4343                 if (test_bit(CEPH_MDS_R_GOT_UNSAFE, &req->r_req_flags))
4344                         continue;
4345                 if (req->r_attempts == 0)
4346                         continue; /* only old requests */
4347                 if (!req->r_session)
4348                         continue;
4349                 if (req->r_session->s_mds != session->s_mds)
4350                         continue;
4351
4352                 ceph_mdsc_release_dir_caps_async(req);
4353
4354                 __send_request(session, req, true);
4355         }
4356         mutex_unlock(&mdsc->mutex);
4357 }
4358
4359 static int send_reconnect_partial(struct ceph_reconnect_state *recon_state)
4360 {
4361         struct ceph_msg *reply;
4362         struct ceph_pagelist *_pagelist;
4363         struct page *page;
4364         __le32 *addr;
4365         int err = -ENOMEM;
4366
4367         if (!recon_state->allow_multi)
4368                 return -ENOSPC;
4369
4370         /* can't handle message that contains both caps and realm */
4371         BUG_ON(!recon_state->nr_caps == !recon_state->nr_realms);
4372
4373         /* pre-allocate new pagelist */
4374         _pagelist = ceph_pagelist_alloc(GFP_NOFS);
4375         if (!_pagelist)
4376                 return -ENOMEM;
4377
4378         reply = ceph_msg_new2(CEPH_MSG_CLIENT_RECONNECT, 0, 1, GFP_NOFS, false);
4379         if (!reply)
4380                 goto fail_msg;
4381
4382         /* placeholder for nr_caps */
4383         err = ceph_pagelist_encode_32(_pagelist, 0);
4384         if (err < 0)
4385                 goto fail;
4386
4387         if (recon_state->nr_caps) {
4388                 /* currently encoding caps */
4389                 err = ceph_pagelist_encode_32(recon_state->pagelist, 0);
4390                 if (err)
4391                         goto fail;
4392         } else {
4393                 /* placeholder for nr_realms (currently encoding relams) */
4394                 err = ceph_pagelist_encode_32(_pagelist, 0);
4395                 if (err < 0)
4396                         goto fail;
4397         }
4398
4399         err = ceph_pagelist_encode_8(recon_state->pagelist, 1);
4400         if (err)
4401                 goto fail;
4402
4403         page = list_first_entry(&recon_state->pagelist->head, struct page, lru);
4404         addr = kmap_atomic(page);
4405         if (recon_state->nr_caps) {
4406                 /* currently encoding caps */
4407                 *addr = cpu_to_le32(recon_state->nr_caps);
4408         } else {
4409                 /* currently encoding relams */
4410                 *(addr + 1) = cpu_to_le32(recon_state->nr_realms);
4411         }
4412         kunmap_atomic(addr);
4413
4414         reply->hdr.version = cpu_to_le16(5);
4415         reply->hdr.compat_version = cpu_to_le16(4);
4416
4417         reply->hdr.data_len = cpu_to_le32(recon_state->pagelist->length);
4418         ceph_msg_data_add_pagelist(reply, recon_state->pagelist);
4419
4420         ceph_con_send(&recon_state->session->s_con, reply);
4421         ceph_pagelist_release(recon_state->pagelist);
4422
4423         recon_state->pagelist = _pagelist;
4424         recon_state->nr_caps = 0;
4425         recon_state->nr_realms = 0;
4426         recon_state->msg_version = 5;
4427         return 0;
4428 fail:
4429         ceph_msg_put(reply);
4430 fail_msg:
4431         ceph_pagelist_release(_pagelist);
4432         return err;
4433 }
4434
4435 static struct dentry* d_find_primary(struct inode *inode)
4436 {
4437         struct dentry *alias, *dn = NULL;
4438
4439         if (hlist_empty(&inode->i_dentry))
4440                 return NULL;
4441
4442         spin_lock(&inode->i_lock);
4443         if (hlist_empty(&inode->i_dentry))
4444                 goto out_unlock;
4445
4446         if (S_ISDIR(inode->i_mode)) {
4447                 alias = hlist_entry(inode->i_dentry.first, struct dentry, d_u.d_alias);
4448                 if (!IS_ROOT(alias))
4449                         dn = dget(alias);
4450                 goto out_unlock;
4451         }
4452
4453         hlist_for_each_entry(alias, &inode->i_dentry, d_u.d_alias) {
4454                 spin_lock(&alias->d_lock);
4455                 if (!d_unhashed(alias) &&
4456                     (ceph_dentry(alias)->flags & CEPH_DENTRY_PRIMARY_LINK)) {
4457                         dn = dget_dlock(alias);
4458                 }
4459                 spin_unlock(&alias->d_lock);
4460                 if (dn)
4461                         break;
4462         }
4463 out_unlock:
4464         spin_unlock(&inode->i_lock);
4465         return dn;
4466 }
4467
4468 /*
4469  * Encode information about a cap for a reconnect with the MDS.
4470  */
4471 static int reconnect_caps_cb(struct inode *inode, int mds, void *arg)
4472 {
4473         struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb);
4474         struct ceph_client *cl = ceph_inode_to_client(inode);
4475         union {
4476                 struct ceph_mds_cap_reconnect v2;
4477                 struct ceph_mds_cap_reconnect_v1 v1;
4478         } rec;
4479         struct ceph_inode_info *ci = ceph_inode(inode);
4480         struct ceph_reconnect_state *recon_state = arg;
4481         struct ceph_pagelist *pagelist = recon_state->pagelist;
4482         struct dentry *dentry;
4483         struct ceph_cap *cap;
4484         char *path;
4485         int pathlen = 0, err;
4486         u64 pathbase;
4487         u64 snap_follows;
4488
4489         dentry = d_find_primary(inode);
4490         if (dentry) {
4491                 /* set pathbase to parent dir when msg_version >= 2 */
4492                 path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase,
4493                                             recon_state->msg_version >= 2);
4494                 dput(dentry);
4495                 if (IS_ERR(path)) {
4496                         err = PTR_ERR(path);
4497                         goto out_err;
4498                 }
4499         } else {
4500                 path = NULL;
4501                 pathbase = 0;
4502         }
4503
4504         spin_lock(&ci->i_ceph_lock);
4505         cap = __get_cap_for_mds(ci, mds);
4506         if (!cap) {
4507                 spin_unlock(&ci->i_ceph_lock);
4508                 err = 0;
4509                 goto out_err;
4510         }
4511         doutc(cl, " adding %p ino %llx.%llx cap %p %lld %s\n", inode,
4512               ceph_vinop(inode), cap, cap->cap_id,
4513               ceph_cap_string(cap->issued));
4514
4515         cap->seq = 0;        /* reset cap seq */
4516         cap->issue_seq = 0;  /* and issue_seq */
4517         cap->mseq = 0;       /* and migrate_seq */
4518         cap->cap_gen = atomic_read(&cap->session->s_cap_gen);
4519
4520         /* These are lost when the session goes away */
4521         if (S_ISDIR(inode->i_mode)) {
4522                 if (cap->issued & CEPH_CAP_DIR_CREATE) {
4523                         ceph_put_string(rcu_dereference_raw(ci->i_cached_layout.pool_ns));
4524                         memset(&ci->i_cached_layout, 0, sizeof(ci->i_cached_layout));
4525                 }
4526                 cap->issued &= ~CEPH_CAP_ANY_DIR_OPS;
4527         }
4528
4529         if (recon_state->msg_version >= 2) {
4530                 rec.v2.cap_id = cpu_to_le64(cap->cap_id);
4531                 rec.v2.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
4532                 rec.v2.issued = cpu_to_le32(cap->issued);
4533                 rec.v2.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
4534                 rec.v2.pathbase = cpu_to_le64(pathbase);
4535                 rec.v2.flock_len = (__force __le32)
4536                         ((ci->i_ceph_flags & CEPH_I_ERROR_FILELOCK) ? 0 : 1);
4537         } else {
4538                 struct timespec64 ts;
4539
4540                 rec.v1.cap_id = cpu_to_le64(cap->cap_id);
4541                 rec.v1.wanted = cpu_to_le32(__ceph_caps_wanted(ci));
4542                 rec.v1.issued = cpu_to_le32(cap->issued);
4543                 rec.v1.size = cpu_to_le64(i_size_read(inode));
4544                 ts = inode_get_mtime(inode);
4545                 ceph_encode_timespec64(&rec.v1.mtime, &ts);
4546                 ts = inode_get_atime(inode);
4547                 ceph_encode_timespec64(&rec.v1.atime, &ts);
4548                 rec.v1.snaprealm = cpu_to_le64(ci->i_snap_realm->ino);
4549                 rec.v1.pathbase = cpu_to_le64(pathbase);
4550         }
4551
4552         if (list_empty(&ci->i_cap_snaps)) {
4553                 snap_follows = ci->i_head_snapc ? ci->i_head_snapc->seq : 0;
4554         } else {
4555                 struct ceph_cap_snap *capsnap =
4556                         list_first_entry(&ci->i_cap_snaps,
4557                                          struct ceph_cap_snap, ci_item);
4558                 snap_follows = capsnap->follows;
4559         }
4560         spin_unlock(&ci->i_ceph_lock);
4561
4562         if (recon_state->msg_version >= 2) {
4563                 int num_fcntl_locks, num_flock_locks;
4564                 struct ceph_filelock *flocks = NULL;
4565                 size_t struct_len, total_len = sizeof(u64);
4566                 u8 struct_v = 0;
4567
4568 encode_again:
4569                 if (rec.v2.flock_len) {
4570                         ceph_count_locks(inode, &num_fcntl_locks, &num_flock_locks);
4571                 } else {
4572                         num_fcntl_locks = 0;
4573                         num_flock_locks = 0;
4574                 }
4575                 if (num_fcntl_locks + num_flock_locks > 0) {
4576                         flocks = kmalloc_array(num_fcntl_locks + num_flock_locks,
4577                                                sizeof(struct ceph_filelock),
4578                                                GFP_NOFS);
4579                         if (!flocks) {
4580                                 err = -ENOMEM;
4581                                 goto out_err;
4582                         }
4583                         err = ceph_encode_locks_to_buffer(inode, flocks,
4584                                                           num_fcntl_locks,
4585                                                           num_flock_locks);
4586                         if (err) {
4587                                 kfree(flocks);
4588                                 flocks = NULL;
4589                                 if (err == -ENOSPC)
4590                                         goto encode_again;
4591                                 goto out_err;
4592                         }
4593                 } else {
4594                         kfree(flocks);
4595                         flocks = NULL;
4596                 }
4597
4598                 if (recon_state->msg_version >= 3) {
4599                         /* version, compat_version and struct_len */
4600                         total_len += 2 * sizeof(u8) + sizeof(u32);
4601                         struct_v = 2;
4602                 }
4603                 /*
4604                  * number of encoded locks is stable, so copy to pagelist
4605                  */
4606                 struct_len = 2 * sizeof(u32) +
4607                             (num_fcntl_locks + num_flock_locks) *
4608                             sizeof(struct ceph_filelock);
4609                 rec.v2.flock_len = cpu_to_le32(struct_len);
4610
4611                 struct_len += sizeof(u32) + pathlen + sizeof(rec.v2);
4612
4613                 if (struct_v >= 2)
4614                         struct_len += sizeof(u64); /* snap_follows */
4615
4616                 total_len += struct_len;
4617
4618                 if (pagelist->length + total_len > RECONNECT_MAX_SIZE) {
4619                         err = send_reconnect_partial(recon_state);
4620                         if (err)
4621                                 goto out_freeflocks;
4622                         pagelist = recon_state->pagelist;
4623                 }
4624
4625                 err = ceph_pagelist_reserve(pagelist, total_len);
4626                 if (err)
4627                         goto out_freeflocks;
4628
4629                 ceph_pagelist_encode_64(pagelist, ceph_ino(inode));
4630                 if (recon_state->msg_version >= 3) {
4631                         ceph_pagelist_encode_8(pagelist, struct_v);
4632                         ceph_pagelist_encode_8(pagelist, 1);
4633                         ceph_pagelist_encode_32(pagelist, struct_len);
4634                 }
4635                 ceph_pagelist_encode_string(pagelist, path, pathlen);
4636                 ceph_pagelist_append(pagelist, &rec, sizeof(rec.v2));
4637                 ceph_locks_to_pagelist(flocks, pagelist,
4638                                        num_fcntl_locks, num_flock_locks);
4639                 if (struct_v >= 2)
4640                         ceph_pagelist_encode_64(pagelist, snap_follows);
4641 out_freeflocks:
4642                 kfree(flocks);
4643         } else {
4644                 err = ceph_pagelist_reserve(pagelist,
4645                                             sizeof(u64) + sizeof(u32) +
4646                                             pathlen + sizeof(rec.v1));
4647                 if (err)
4648                         goto out_err;
4649
4650                 ceph_pagelist_encode_64(pagelist, ceph_ino(inode));
4651                 ceph_pagelist_encode_string(pagelist, path, pathlen);
4652                 ceph_pagelist_append(pagelist, &rec, sizeof(rec.v1));
4653         }
4654
4655 out_err:
4656         ceph_mdsc_free_path(path, pathlen);
4657         if (!err)
4658                 recon_state->nr_caps++;
4659         return err;
4660 }
4661
4662 static int encode_snap_realms(struct ceph_mds_client *mdsc,
4663                               struct ceph_reconnect_state *recon_state)
4664 {
4665         struct rb_node *p;
4666         struct ceph_pagelist *pagelist = recon_state->pagelist;
4667         struct ceph_client *cl = mdsc->fsc->client;
4668         int err = 0;
4669
4670         if (recon_state->msg_version >= 4) {
4671                 err = ceph_pagelist_encode_32(pagelist, mdsc->num_snap_realms);
4672                 if (err < 0)
4673                         goto fail;
4674         }
4675
4676         /*
4677          * snaprealms.  we provide mds with the ino, seq (version), and
4678          * parent for all of our realms.  If the mds has any newer info,
4679          * it will tell us.
4680          */
4681         for (p = rb_first(&mdsc->snap_realms); p; p = rb_next(p)) {
4682                 struct ceph_snap_realm *realm =
4683                        rb_entry(p, struct ceph_snap_realm, node);
4684                 struct ceph_mds_snaprealm_reconnect sr_rec;
4685
4686                 if (recon_state->msg_version >= 4) {
4687                         size_t need = sizeof(u8) * 2 + sizeof(u32) +
4688                                       sizeof(sr_rec);
4689
4690                         if (pagelist->length + need > RECONNECT_MAX_SIZE) {
4691                                 err = send_reconnect_partial(recon_state);
4692                                 if (err)
4693                                         goto fail;
4694                                 pagelist = recon_state->pagelist;
4695                         }
4696
4697                         err = ceph_pagelist_reserve(pagelist, need);
4698                         if (err)
4699                                 goto fail;
4700
4701                         ceph_pagelist_encode_8(pagelist, 1);
4702                         ceph_pagelist_encode_8(pagelist, 1);
4703                         ceph_pagelist_encode_32(pagelist, sizeof(sr_rec));
4704                 }
4705
4706                 doutc(cl, " adding snap realm %llx seq %lld parent %llx\n",
4707                       realm->ino, realm->seq, realm->parent_ino);
4708                 sr_rec.ino = cpu_to_le64(realm->ino);
4709                 sr_rec.seq = cpu_to_le64(realm->seq);
4710                 sr_rec.parent = cpu_to_le64(realm->parent_ino);
4711
4712                 err = ceph_pagelist_append(pagelist, &sr_rec, sizeof(sr_rec));
4713                 if (err)
4714                         goto fail;
4715
4716                 recon_state->nr_realms++;
4717         }
4718 fail:
4719         return err;
4720 }
4721
4722
4723 /*
4724  * If an MDS fails and recovers, clients need to reconnect in order to
4725  * reestablish shared state.  This includes all caps issued through
4726  * this session _and_ the snap_realm hierarchy.  Because it's not
4727  * clear which snap realms the mds cares about, we send everything we
4728  * know about.. that ensures we'll then get any new info the
4729  * recovering MDS might have.
4730  *
4731  * This is a relatively heavyweight operation, but it's rare.
4732  */
4733 static void send_mds_reconnect(struct ceph_mds_client *mdsc,
4734                                struct ceph_mds_session *session)
4735 {
4736         struct ceph_client *cl = mdsc->fsc->client;
4737         struct ceph_msg *reply;
4738         int mds = session->s_mds;
4739         int err = -ENOMEM;
4740         struct ceph_reconnect_state recon_state = {
4741                 .session = session,
4742         };
4743         LIST_HEAD(dispose);
4744
4745         pr_info_client(cl, "mds%d reconnect start\n", mds);
4746
4747         recon_state.pagelist = ceph_pagelist_alloc(GFP_NOFS);
4748         if (!recon_state.pagelist)
4749                 goto fail_nopagelist;
4750
4751         reply = ceph_msg_new2(CEPH_MSG_CLIENT_RECONNECT, 0, 1, GFP_NOFS, false);
4752         if (!reply)
4753                 goto fail_nomsg;
4754
4755         xa_destroy(&session->s_delegated_inos);
4756
4757         mutex_lock(&session->s_mutex);
4758         session->s_state = CEPH_MDS_SESSION_RECONNECTING;
4759         session->s_seq = 0;
4760
4761         doutc(cl, "session %p state %s\n", session,
4762               ceph_session_state_name(session->s_state));
4763
4764         atomic_inc(&session->s_cap_gen);
4765
4766         spin_lock(&session->s_cap_lock);
4767         /* don't know if session is readonly */
4768         session->s_readonly = 0;
4769         /*
4770          * notify __ceph_remove_cap() that we are composing cap reconnect.
4771          * If a cap get released before being added to the cap reconnect,
4772          * __ceph_remove_cap() should skip queuing cap release.
4773          */
4774         session->s_cap_reconnect = 1;
4775         /* drop old cap expires; we're about to reestablish that state */
4776         detach_cap_releases(session, &dispose);
4777         spin_unlock(&session->s_cap_lock);
4778         dispose_cap_releases(mdsc, &dispose);
4779
4780         /* trim unused caps to reduce MDS's cache rejoin time */
4781         if (mdsc->fsc->sb->s_root)
4782                 shrink_dcache_parent(mdsc->fsc->sb->s_root);
4783
4784         ceph_con_close(&session->s_con);
4785         ceph_con_open(&session->s_con,
4786                       CEPH_ENTITY_TYPE_MDS, mds,
4787                       ceph_mdsmap_get_addr(mdsc->mdsmap, mds));
4788
4789         /* replay unsafe requests */
4790         replay_unsafe_requests(mdsc, session);
4791
4792         ceph_early_kick_flushing_caps(mdsc, session);
4793
4794         down_read(&mdsc->snap_rwsem);
4795
4796         /* placeholder for nr_caps */
4797         err = ceph_pagelist_encode_32(recon_state.pagelist, 0);
4798         if (err)
4799                 goto fail;
4800
4801         if (test_bit(CEPHFS_FEATURE_MULTI_RECONNECT, &session->s_features)) {
4802                 recon_state.msg_version = 3;
4803                 recon_state.allow_multi = true;
4804         } else if (session->s_con.peer_features & CEPH_FEATURE_MDSENC) {
4805                 recon_state.msg_version = 3;
4806         } else {
4807                 recon_state.msg_version = 2;
4808         }
4809         /* trsaverse this session's caps */
4810         err = ceph_iterate_session_caps(session, reconnect_caps_cb, &recon_state);
4811
4812         spin_lock(&session->s_cap_lock);
4813         session->s_cap_reconnect = 0;
4814         spin_unlock(&session->s_cap_lock);
4815
4816         if (err < 0)
4817                 goto fail;
4818
4819         /* check if all realms can be encoded into current message */
4820         if (mdsc->num_snap_realms) {
4821                 size_t total_len =
4822                         recon_state.pagelist->length +
4823                         mdsc->num_snap_realms *
4824                         sizeof(struct ceph_mds_snaprealm_reconnect);
4825                 if (recon_state.msg_version >= 4) {
4826                         /* number of realms */
4827                         total_len += sizeof(u32);
4828                         /* version, compat_version and struct_len */
4829                         total_len += mdsc->num_snap_realms *
4830                                      (2 * sizeof(u8) + sizeof(u32));
4831                 }
4832                 if (total_len > RECONNECT_MAX_SIZE) {
4833                         if (!recon_state.allow_multi) {
4834                                 err = -ENOSPC;
4835                                 goto fail;
4836                         }
4837                         if (recon_state.nr_caps) {
4838                                 err = send_reconnect_partial(&recon_state);
4839                                 if (err)
4840                                         goto fail;
4841                         }
4842                         recon_state.msg_version = 5;
4843                 }
4844         }
4845
4846         err = encode_snap_realms(mdsc, &recon_state);
4847         if (err < 0)
4848                 goto fail;
4849
4850         if (recon_state.msg_version >= 5) {
4851                 err = ceph_pagelist_encode_8(recon_state.pagelist, 0);
4852                 if (err < 0)
4853                         goto fail;
4854         }
4855
4856         if (recon_state.nr_caps || recon_state.nr_realms) {
4857                 struct page *page =
4858                         list_first_entry(&recon_state.pagelist->head,
4859                                         struct page, lru);
4860                 __le32 *addr = kmap_atomic(page);
4861                 if (recon_state.nr_caps) {
4862                         WARN_ON(recon_state.nr_realms != mdsc->num_snap_realms);
4863                         *addr = cpu_to_le32(recon_state.nr_caps);
4864                 } else if (recon_state.msg_version >= 4) {
4865                         *(addr + 1) = cpu_to_le32(recon_state.nr_realms);
4866                 }
4867                 kunmap_atomic(addr);
4868         }
4869
4870         reply->hdr.version = cpu_to_le16(recon_state.msg_version);
4871         if (recon_state.msg_version >= 4)
4872                 reply->hdr.compat_version = cpu_to_le16(4);
4873
4874         reply->hdr.data_len = cpu_to_le32(recon_state.pagelist->length);
4875         ceph_msg_data_add_pagelist(reply, recon_state.pagelist);
4876
4877         ceph_con_send(&session->s_con, reply);
4878
4879         mutex_unlock(&session->s_mutex);
4880
4881         mutex_lock(&mdsc->mutex);
4882         __wake_requests(mdsc, &session->s_waiting);
4883         mutex_unlock(&mdsc->mutex);
4884
4885         up_read(&mdsc->snap_rwsem);
4886         ceph_pagelist_release(recon_state.pagelist);
4887         return;
4888
4889 fail:
4890         ceph_msg_put(reply);
4891         up_read(&mdsc->snap_rwsem);
4892         mutex_unlock(&session->s_mutex);
4893 fail_nomsg:
4894         ceph_pagelist_release(recon_state.pagelist);
4895 fail_nopagelist:
4896         pr_err_client(cl, "error %d preparing reconnect for mds%d\n",
4897                       err, mds);
4898         return;
4899 }
4900
4901
4902 /*
4903  * compare old and new mdsmaps, kicking requests
4904  * and closing out old connections as necessary
4905  *
4906  * called under mdsc->mutex.
4907  */
4908 static void check_new_map(struct ceph_mds_client *mdsc,
4909                           struct ceph_mdsmap *newmap,
4910                           struct ceph_mdsmap *oldmap)
4911 {
4912         int i, j, err;
4913         int oldstate, newstate;
4914         struct ceph_mds_session *s;
4915         unsigned long targets[DIV_ROUND_UP(CEPH_MAX_MDS, sizeof(unsigned long))] = {0};
4916         struct ceph_client *cl = mdsc->fsc->client;
4917
4918         doutc(cl, "new %u old %u\n", newmap->m_epoch, oldmap->m_epoch);
4919
4920         if (newmap->m_info) {
4921                 for (i = 0; i < newmap->possible_max_rank; i++) {
4922                         for (j = 0; j < newmap->m_info[i].num_export_targets; j++)
4923                                 set_bit(newmap->m_info[i].export_targets[j], targets);
4924                 }
4925         }
4926
4927         for (i = 0; i < oldmap->possible_max_rank && i < mdsc->max_sessions; i++) {
4928                 if (!mdsc->sessions[i])
4929                         continue;
4930                 s = mdsc->sessions[i];
4931                 oldstate = ceph_mdsmap_get_state(oldmap, i);
4932                 newstate = ceph_mdsmap_get_state(newmap, i);
4933
4934                 doutc(cl, "mds%d state %s%s -> %s%s (session %s)\n",
4935                       i, ceph_mds_state_name(oldstate),
4936                       ceph_mdsmap_is_laggy(oldmap, i) ? " (laggy)" : "",
4937                       ceph_mds_state_name(newstate),
4938                       ceph_mdsmap_is_laggy(newmap, i) ? " (laggy)" : "",
4939                       ceph_session_state_name(s->s_state));
4940
4941                 if (i >= newmap->possible_max_rank) {
4942                         /* force close session for stopped mds */
4943                         ceph_get_mds_session(s);
4944                         __unregister_session(mdsc, s);
4945                         __wake_requests(mdsc, &s->s_waiting);
4946                         mutex_unlock(&mdsc->mutex);
4947
4948                         mutex_lock(&s->s_mutex);
4949                         cleanup_session_requests(mdsc, s);
4950                         remove_session_caps(s);
4951                         mutex_unlock(&s->s_mutex);
4952
4953                         ceph_put_mds_session(s);
4954
4955                         mutex_lock(&mdsc->mutex);
4956                         kick_requests(mdsc, i);
4957                         continue;
4958                 }
4959
4960                 if (memcmp(ceph_mdsmap_get_addr(oldmap, i),
4961                            ceph_mdsmap_get_addr(newmap, i),
4962                            sizeof(struct ceph_entity_addr))) {
4963                         /* just close it */
4964                         mutex_unlock(&mdsc->mutex);
4965                         mutex_lock(&s->s_mutex);
4966                         mutex_lock(&mdsc->mutex);
4967                         ceph_con_close(&s->s_con);
4968                         mutex_unlock(&s->s_mutex);
4969                         s->s_state = CEPH_MDS_SESSION_RESTARTING;
4970                 } else if (oldstate == newstate) {
4971                         continue;  /* nothing new with this mds */
4972                 }
4973
4974                 /*
4975                  * send reconnect?
4976                  */
4977                 if (s->s_state == CEPH_MDS_SESSION_RESTARTING &&
4978                     newstate >= CEPH_MDS_STATE_RECONNECT) {
4979                         mutex_unlock(&mdsc->mutex);
4980                         clear_bit(i, targets);
4981                         send_mds_reconnect(mdsc, s);
4982                         mutex_lock(&mdsc->mutex);
4983                 }
4984
4985                 /*
4986                  * kick request on any mds that has gone active.
4987                  */
4988                 if (oldstate < CEPH_MDS_STATE_ACTIVE &&
4989                     newstate >= CEPH_MDS_STATE_ACTIVE) {
4990                         if (oldstate != CEPH_MDS_STATE_CREATING &&
4991                             oldstate != CEPH_MDS_STATE_STARTING)
4992                                 pr_info_client(cl, "mds%d recovery completed\n",
4993                                                s->s_mds);
4994                         kick_requests(mdsc, i);
4995                         mutex_unlock(&mdsc->mutex);
4996                         mutex_lock(&s->s_mutex);
4997                         mutex_lock(&mdsc->mutex);
4998                         ceph_kick_flushing_caps(mdsc, s);
4999                         mutex_unlock(&s->s_mutex);
5000                         wake_up_session_caps(s, RECONNECT);
5001                 }
5002         }
5003
5004         /*
5005          * Only open and reconnect sessions that don't exist yet.
5006          */
5007         for (i = 0; i < newmap->possible_max_rank; i++) {
5008                 /*
5009                  * In case the import MDS is crashed just after
5010                  * the EImportStart journal is flushed, so when
5011                  * a standby MDS takes over it and is replaying
5012                  * the EImportStart journal the new MDS daemon
5013                  * will wait the client to reconnect it, but the
5014                  * client may never register/open the session yet.
5015                  *
5016                  * Will try to reconnect that MDS daemon if the
5017                  * rank number is in the export targets array and
5018                  * is the up:reconnect state.
5019                  */
5020                 newstate = ceph_mdsmap_get_state(newmap, i);
5021                 if (!test_bit(i, targets) || newstate != CEPH_MDS_STATE_RECONNECT)
5022                         continue;
5023
5024                 /*
5025                  * The session maybe registered and opened by some
5026                  * requests which were choosing random MDSes during
5027                  * the mdsc->mutex's unlock/lock gap below in rare
5028                  * case. But the related MDS daemon will just queue
5029                  * that requests and be still waiting for the client's
5030                  * reconnection request in up:reconnect state.
5031                  */
5032                 s = __ceph_lookup_mds_session(mdsc, i);
5033                 if (likely(!s)) {
5034                         s = __open_export_target_session(mdsc, i);
5035                         if (IS_ERR(s)) {
5036                                 err = PTR_ERR(s);
5037                                 pr_err_client(cl,
5038                                               "failed to open export target session, err %d\n",
5039                                               err);
5040                                 continue;
5041                         }
5042                 }
5043                 doutc(cl, "send reconnect to export target mds.%d\n", i);
5044                 mutex_unlock(&mdsc->mutex);
5045                 send_mds_reconnect(mdsc, s);
5046                 ceph_put_mds_session(s);
5047                 mutex_lock(&mdsc->mutex);
5048         }
5049
5050         for (i = 0; i < newmap->possible_max_rank && i < mdsc->max_sessions; i++) {
5051                 s = mdsc->sessions[i];
5052                 if (!s)
5053                         continue;
5054                 if (!ceph_mdsmap_is_laggy(newmap, i))
5055                         continue;
5056                 if (s->s_state == CEPH_MDS_SESSION_OPEN ||
5057                     s->s_state == CEPH_MDS_SESSION_HUNG ||
5058                     s->s_state == CEPH_MDS_SESSION_CLOSING) {
5059                         doutc(cl, " connecting to export targets of laggy mds%d\n", i);
5060                         __open_export_target_sessions(mdsc, s);
5061                 }
5062         }
5063 }
5064
5065
5066
5067 /*
5068  * leases
5069  */
5070
5071 /*
5072  * caller must hold session s_mutex, dentry->d_lock
5073  */
5074 void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry)
5075 {
5076         struct ceph_dentry_info *di = ceph_dentry(dentry);
5077
5078         ceph_put_mds_session(di->lease_session);
5079         di->lease_session = NULL;
5080 }
5081
5082 static void handle_lease(struct ceph_mds_client *mdsc,
5083                          struct ceph_mds_session *session,
5084                          struct ceph_msg *msg)
5085 {
5086         struct ceph_client *cl = mdsc->fsc->client;
5087         struct super_block *sb = mdsc->fsc->sb;
5088         struct inode *inode;
5089         struct dentry *parent, *dentry;
5090         struct ceph_dentry_info *di;
5091         int mds = session->s_mds;
5092         struct ceph_mds_lease *h = msg->front.iov_base;
5093         u32 seq;
5094         struct ceph_vino vino;
5095         struct qstr dname;
5096         int release = 0;
5097
5098         doutc(cl, "from mds%d\n", mds);
5099
5100         if (!ceph_inc_mds_stopping_blocker(mdsc, session))
5101                 return;
5102
5103         /* decode */
5104         if (msg->front.iov_len < sizeof(*h) + sizeof(u32))
5105                 goto bad;
5106         vino.ino = le64_to_cpu(h->ino);
5107         vino.snap = CEPH_NOSNAP;
5108         seq = le32_to_cpu(h->seq);
5109         dname.len = get_unaligned_le32(h + 1);
5110         if (msg->front.iov_len < sizeof(*h) + sizeof(u32) + dname.len)
5111                 goto bad;
5112         dname.name = (void *)(h + 1) + sizeof(u32);
5113
5114         /* lookup inode */
5115         inode = ceph_find_inode(sb, vino);
5116         doutc(cl, "%s, ino %llx %p %.*s\n", ceph_lease_op_name(h->action),
5117               vino.ino, inode, dname.len, dname.name);
5118
5119         mutex_lock(&session->s_mutex);
5120         if (!inode) {
5121                 doutc(cl, "no inode %llx\n", vino.ino);
5122                 goto release;
5123         }
5124
5125         /* dentry */
5126         parent = d_find_alias(inode);
5127         if (!parent) {
5128                 doutc(cl, "no parent dentry on inode %p\n", inode);
5129                 WARN_ON(1);
5130                 goto release;  /* hrm... */
5131         }
5132         dname.hash = full_name_hash(parent, dname.name, dname.len);
5133         dentry = d_lookup(parent, &dname);
5134         dput(parent);
5135         if (!dentry)
5136                 goto release;
5137
5138         spin_lock(&dentry->d_lock);
5139         di = ceph_dentry(dentry);
5140         switch (h->action) {
5141         case CEPH_MDS_LEASE_REVOKE:
5142                 if (di->lease_session == session) {
5143                         if (ceph_seq_cmp(di->lease_seq, seq) > 0)
5144                                 h->seq = cpu_to_le32(di->lease_seq);
5145                         __ceph_mdsc_drop_dentry_lease(dentry);
5146                 }
5147                 release = 1;
5148                 break;
5149
5150         case CEPH_MDS_LEASE_RENEW:
5151                 if (di->lease_session == session &&
5152                     di->lease_gen == atomic_read(&session->s_cap_gen) &&
5153                     di->lease_renew_from &&
5154                     di->lease_renew_after == 0) {
5155                         unsigned long duration =
5156                                 msecs_to_jiffies(le32_to_cpu(h->duration_ms));
5157
5158                         di->lease_seq = seq;
5159                         di->time = di->lease_renew_from + duration;
5160                         di->lease_renew_after = di->lease_renew_from +
5161                                 (duration >> 1);
5162                         di->lease_renew_from = 0;
5163                 }
5164                 break;
5165         }
5166         spin_unlock(&dentry->d_lock);
5167         dput(dentry);
5168
5169         if (!release)
5170                 goto out;
5171
5172 release:
5173         /* let's just reuse the same message */
5174         h->action = CEPH_MDS_LEASE_REVOKE_ACK;
5175         ceph_msg_get(msg);
5176         ceph_con_send(&session->s_con, msg);
5177
5178 out:
5179         mutex_unlock(&session->s_mutex);
5180         iput(inode);
5181
5182         ceph_dec_mds_stopping_blocker(mdsc);
5183         return;
5184
5185 bad:
5186         ceph_dec_mds_stopping_blocker(mdsc);
5187
5188         pr_err_client(cl, "corrupt lease message\n");
5189         ceph_msg_dump(msg);
5190 }
5191
5192 void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
5193                               struct dentry *dentry, char action,
5194                               u32 seq)
5195 {
5196         struct ceph_client *cl = session->s_mdsc->fsc->client;
5197         struct ceph_msg *msg;
5198         struct ceph_mds_lease *lease;
5199         struct inode *dir;
5200         int len = sizeof(*lease) + sizeof(u32) + NAME_MAX;
5201
5202         doutc(cl, "identry %p %s to mds%d\n", dentry, ceph_lease_op_name(action),
5203               session->s_mds);
5204
5205         msg = ceph_msg_new(CEPH_MSG_CLIENT_LEASE, len, GFP_NOFS, false);
5206         if (!msg)
5207                 return;
5208         lease = msg->front.iov_base;
5209         lease->action = action;
5210         lease->seq = cpu_to_le32(seq);
5211
5212         spin_lock(&dentry->d_lock);
5213         dir = d_inode(dentry->d_parent);
5214         lease->ino = cpu_to_le64(ceph_ino(dir));
5215         lease->first = lease->last = cpu_to_le64(ceph_snap(dir));
5216
5217         put_unaligned_le32(dentry->d_name.len, lease + 1);
5218         memcpy((void *)(lease + 1) + 4,
5219                dentry->d_name.name, dentry->d_name.len);
5220         spin_unlock(&dentry->d_lock);
5221
5222         ceph_con_send(&session->s_con, msg);
5223 }
5224
5225 /*
5226  * lock unlock the session, to wait ongoing session activities
5227  */
5228 static void lock_unlock_session(struct ceph_mds_session *s)
5229 {
5230         mutex_lock(&s->s_mutex);
5231         mutex_unlock(&s->s_mutex);
5232 }
5233
5234 static void maybe_recover_session(struct ceph_mds_client *mdsc)
5235 {
5236         struct ceph_client *cl = mdsc->fsc->client;
5237         struct ceph_fs_client *fsc = mdsc->fsc;
5238
5239         if (!ceph_test_mount_opt(fsc, CLEANRECOVER))
5240                 return;
5241
5242         if (READ_ONCE(fsc->mount_state) != CEPH_MOUNT_MOUNTED)
5243                 return;
5244
5245         if (!READ_ONCE(fsc->blocklisted))
5246                 return;
5247
5248         pr_info_client(cl, "auto reconnect after blocklisted\n");
5249         ceph_force_reconnect(fsc->sb);
5250 }
5251
5252 bool check_session_state(struct ceph_mds_session *s)
5253 {
5254         struct ceph_client *cl = s->s_mdsc->fsc->client;
5255
5256         switch (s->s_state) {
5257         case CEPH_MDS_SESSION_OPEN:
5258                 if (s->s_ttl && time_after(jiffies, s->s_ttl)) {
5259                         s->s_state = CEPH_MDS_SESSION_HUNG;
5260                         pr_info_client(cl, "mds%d hung\n", s->s_mds);
5261                 }
5262                 break;
5263         case CEPH_MDS_SESSION_CLOSING:
5264         case CEPH_MDS_SESSION_NEW:
5265         case CEPH_MDS_SESSION_RESTARTING:
5266         case CEPH_MDS_SESSION_CLOSED:
5267         case CEPH_MDS_SESSION_REJECTED:
5268                 return false;
5269         }
5270
5271         return true;
5272 }
5273
5274 /*
5275  * If the sequence is incremented while we're waiting on a REQUEST_CLOSE reply,
5276  * then we need to retransmit that request.
5277  */
5278 void inc_session_sequence(struct ceph_mds_session *s)
5279 {
5280         struct ceph_client *cl = s->s_mdsc->fsc->client;
5281
5282         lockdep_assert_held(&s->s_mutex);
5283
5284         s->s_seq++;
5285
5286         if (s->s_state == CEPH_MDS_SESSION_CLOSING) {
5287                 int ret;
5288
5289                 doutc(cl, "resending session close request for mds%d\n", s->s_mds);
5290                 ret = request_close_session(s);
5291                 if (ret < 0)
5292                         pr_err_client(cl, "unable to close session to mds%d: %d\n",
5293                                       s->s_mds, ret);
5294         }
5295 }
5296
5297 /*
5298  * delayed work -- periodically trim expired leases, renew caps with mds.  If
5299  * the @delay parameter is set to 0 or if it's more than 5 secs, the default
5300  * workqueue delay value of 5 secs will be used.
5301  */
5302 static void schedule_delayed(struct ceph_mds_client *mdsc, unsigned long delay)
5303 {
5304         unsigned long max_delay = HZ * 5;
5305
5306         /* 5 secs default delay */
5307         if (!delay || (delay > max_delay))
5308                 delay = max_delay;
5309         schedule_delayed_work(&mdsc->delayed_work,
5310                               round_jiffies_relative(delay));
5311 }
5312
5313 static void delayed_work(struct work_struct *work)
5314 {
5315         struct ceph_mds_client *mdsc =
5316                 container_of(work, struct ceph_mds_client, delayed_work.work);
5317         unsigned long delay;
5318         int renew_interval;
5319         int renew_caps;
5320         int i;
5321
5322         doutc(mdsc->fsc->client, "mdsc delayed_work\n");
5323
5324         if (mdsc->stopping >= CEPH_MDSC_STOPPING_FLUSHED)
5325                 return;
5326
5327         mutex_lock(&mdsc->mutex);
5328         renew_interval = mdsc->mdsmap->m_session_timeout >> 2;
5329         renew_caps = time_after_eq(jiffies, HZ*renew_interval +
5330                                    mdsc->last_renew_caps);
5331         if (renew_caps)
5332                 mdsc->last_renew_caps = jiffies;
5333
5334         for (i = 0; i < mdsc->max_sessions; i++) {
5335                 struct ceph_mds_session *s = __ceph_lookup_mds_session(mdsc, i);
5336                 if (!s)
5337                         continue;
5338
5339                 if (!check_session_state(s)) {
5340                         ceph_put_mds_session(s);
5341                         continue;
5342                 }
5343                 mutex_unlock(&mdsc->mutex);
5344
5345                 mutex_lock(&s->s_mutex);
5346                 if (renew_caps)
5347                         send_renew_caps(mdsc, s);
5348                 else
5349                         ceph_con_keepalive(&s->s_con);
5350                 if (s->s_state == CEPH_MDS_SESSION_OPEN ||
5351                     s->s_state == CEPH_MDS_SESSION_HUNG)
5352                         ceph_send_cap_releases(mdsc, s);
5353                 mutex_unlock(&s->s_mutex);
5354                 ceph_put_mds_session(s);
5355
5356                 mutex_lock(&mdsc->mutex);
5357         }
5358         mutex_unlock(&mdsc->mutex);
5359
5360         delay = ceph_check_delayed_caps(mdsc);
5361
5362         ceph_queue_cap_reclaim_work(mdsc);
5363
5364         ceph_trim_snapid_map(mdsc);
5365
5366         maybe_recover_session(mdsc);
5367
5368         schedule_delayed(mdsc, delay);
5369 }
5370
5371 int ceph_mdsc_init(struct ceph_fs_client *fsc)
5372
5373 {
5374         struct ceph_mds_client *mdsc;
5375         int err;
5376
5377         mdsc = kzalloc(sizeof(struct ceph_mds_client), GFP_NOFS);
5378         if (!mdsc)
5379                 return -ENOMEM;
5380         mdsc->fsc = fsc;
5381         mutex_init(&mdsc->mutex);
5382         mdsc->mdsmap = kzalloc(sizeof(*mdsc->mdsmap), GFP_NOFS);
5383         if (!mdsc->mdsmap) {
5384                 err = -ENOMEM;
5385                 goto err_mdsc;
5386         }
5387
5388         init_completion(&mdsc->safe_umount_waiters);
5389         spin_lock_init(&mdsc->stopping_lock);
5390         atomic_set(&mdsc->stopping_blockers, 0);
5391         init_completion(&mdsc->stopping_waiter);
5392         init_waitqueue_head(&mdsc->session_close_wq);
5393         INIT_LIST_HEAD(&mdsc->waiting_for_map);
5394         mdsc->quotarealms_inodes = RB_ROOT;
5395         mutex_init(&mdsc->quotarealms_inodes_mutex);
5396         init_rwsem(&mdsc->snap_rwsem);
5397         mdsc->snap_realms = RB_ROOT;
5398         INIT_LIST_HEAD(&mdsc->snap_empty);
5399         spin_lock_init(&mdsc->snap_empty_lock);
5400         mdsc->request_tree = RB_ROOT;
5401         INIT_DELAYED_WORK(&mdsc->delayed_work, delayed_work);
5402         mdsc->last_renew_caps = jiffies;
5403         INIT_LIST_HEAD(&mdsc->cap_delay_list);
5404         INIT_LIST_HEAD(&mdsc->cap_wait_list);
5405         spin_lock_init(&mdsc->cap_delay_lock);
5406         INIT_LIST_HEAD(&mdsc->cap_unlink_delay_list);
5407         spin_lock_init(&mdsc->cap_unlink_delay_lock);
5408         INIT_LIST_HEAD(&mdsc->snap_flush_list);
5409         spin_lock_init(&mdsc->snap_flush_lock);
5410         mdsc->last_cap_flush_tid = 1;
5411         INIT_LIST_HEAD(&mdsc->cap_flush_list);
5412         INIT_LIST_HEAD(&mdsc->cap_dirty_migrating);
5413         spin_lock_init(&mdsc->cap_dirty_lock);
5414         init_waitqueue_head(&mdsc->cap_flushing_wq);
5415         INIT_WORK(&mdsc->cap_reclaim_work, ceph_cap_reclaim_work);
5416         INIT_WORK(&mdsc->cap_unlink_work, ceph_cap_unlink_work);
5417         err = ceph_metric_init(&mdsc->metric);
5418         if (err)
5419                 goto err_mdsmap;
5420
5421         spin_lock_init(&mdsc->dentry_list_lock);
5422         INIT_LIST_HEAD(&mdsc->dentry_leases);
5423         INIT_LIST_HEAD(&mdsc->dentry_dir_leases);
5424
5425         ceph_caps_init(mdsc);
5426         ceph_adjust_caps_max_min(mdsc, fsc->mount_options);
5427
5428         spin_lock_init(&mdsc->snapid_map_lock);
5429         mdsc->snapid_map_tree = RB_ROOT;
5430         INIT_LIST_HEAD(&mdsc->snapid_map_lru);
5431
5432         init_rwsem(&mdsc->pool_perm_rwsem);
5433         mdsc->pool_perm_tree = RB_ROOT;
5434
5435         strscpy(mdsc->nodename, utsname()->nodename,
5436                 sizeof(mdsc->nodename));
5437
5438         fsc->mdsc = mdsc;
5439         return 0;
5440
5441 err_mdsmap:
5442         kfree(mdsc->mdsmap);
5443 err_mdsc:
5444         kfree(mdsc);
5445         return err;
5446 }
5447
5448 /*
5449  * Wait for safe replies on open mds requests.  If we time out, drop
5450  * all requests from the tree to avoid dangling dentry refs.
5451  */
5452 static void wait_requests(struct ceph_mds_client *mdsc)
5453 {
5454         struct ceph_client *cl = mdsc->fsc->client;
5455         struct ceph_options *opts = mdsc->fsc->client->options;
5456         struct ceph_mds_request *req;
5457
5458         mutex_lock(&mdsc->mutex);
5459         if (__get_oldest_req(mdsc)) {
5460                 mutex_unlock(&mdsc->mutex);
5461
5462                 doutc(cl, "waiting for requests\n");
5463                 wait_for_completion_timeout(&mdsc->safe_umount_waiters,
5464                                     ceph_timeout_jiffies(opts->mount_timeout));
5465
5466                 /* tear down remaining requests */
5467                 mutex_lock(&mdsc->mutex);
5468                 while ((req = __get_oldest_req(mdsc))) {
5469                         doutc(cl, "timed out on tid %llu\n", req->r_tid);
5470                         list_del_init(&req->r_wait);
5471                         __unregister_request(mdsc, req);
5472                 }
5473         }
5474         mutex_unlock(&mdsc->mutex);
5475         doutc(cl, "done\n");
5476 }
5477
5478 void send_flush_mdlog(struct ceph_mds_session *s)
5479 {
5480         struct ceph_client *cl = s->s_mdsc->fsc->client;
5481         struct ceph_msg *msg;
5482
5483         /*
5484          * Pre-luminous MDS crashes when it sees an unknown session request
5485          */
5486         if (!CEPH_HAVE_FEATURE(s->s_con.peer_features, SERVER_LUMINOUS))
5487                 return;
5488
5489         mutex_lock(&s->s_mutex);
5490         doutc(cl, "request mdlog flush to mds%d (%s)s seq %lld\n",
5491               s->s_mds, ceph_session_state_name(s->s_state), s->s_seq);
5492         msg = ceph_create_session_msg(CEPH_SESSION_REQUEST_FLUSH_MDLOG,
5493                                       s->s_seq);
5494         if (!msg) {
5495                 pr_err_client(cl, "failed to request mdlog flush to mds%d (%s) seq %lld\n",
5496                               s->s_mds, ceph_session_state_name(s->s_state), s->s_seq);
5497         } else {
5498                 ceph_con_send(&s->s_con, msg);
5499         }
5500         mutex_unlock(&s->s_mutex);
5501 }
5502
5503 /*
5504  * called before mount is ro, and before dentries are torn down.
5505  * (hmm, does this still race with new lookups?)
5506  */
5507 void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc)
5508 {
5509         doutc(mdsc->fsc->client, "begin\n");
5510         mdsc->stopping = CEPH_MDSC_STOPPING_BEGIN;
5511
5512         ceph_mdsc_iterate_sessions(mdsc, send_flush_mdlog, true);
5513         ceph_mdsc_iterate_sessions(mdsc, lock_unlock_session, false);
5514         ceph_flush_dirty_caps(mdsc);
5515         wait_requests(mdsc);
5516
5517         /*
5518          * wait for reply handlers to drop their request refs and
5519          * their inode/dcache refs
5520          */
5521         ceph_msgr_flush();
5522
5523         ceph_cleanup_quotarealms_inodes(mdsc);
5524         doutc(mdsc->fsc->client, "done\n");
5525 }
5526
5527 /*
5528  * flush the mdlog and wait for all write mds requests to flush.
5529  */
5530 static void flush_mdlog_and_wait_mdsc_unsafe_requests(struct ceph_mds_client *mdsc,
5531                                                  u64 want_tid)
5532 {
5533         struct ceph_client *cl = mdsc->fsc->client;
5534         struct ceph_mds_request *req = NULL, *nextreq;
5535         struct ceph_mds_session *last_session = NULL;
5536         struct rb_node *n;
5537
5538         mutex_lock(&mdsc->mutex);
5539         doutc(cl, "want %lld\n", want_tid);
5540 restart:
5541         req = __get_oldest_req(mdsc);
5542         while (req && req->r_tid <= want_tid) {
5543                 /* find next request */
5544                 n = rb_next(&req->r_node);
5545                 if (n)
5546                         nextreq = rb_entry(n, struct ceph_mds_request, r_node);
5547                 else
5548                         nextreq = NULL;
5549                 if (req->r_op != CEPH_MDS_OP_SETFILELOCK &&
5550                     (req->r_op & CEPH_MDS_OP_WRITE)) {
5551                         struct ceph_mds_session *s = req->r_session;
5552
5553                         if (!s) {
5554                                 req = nextreq;
5555                                 continue;
5556                         }
5557
5558                         /* write op */
5559                         ceph_mdsc_get_request(req);
5560                         if (nextreq)
5561                                 ceph_mdsc_get_request(nextreq);
5562                         s = ceph_get_mds_session(s);
5563                         mutex_unlock(&mdsc->mutex);
5564
5565                         /* send flush mdlog request to MDS */
5566                         if (last_session != s) {
5567                                 send_flush_mdlog(s);
5568                                 ceph_put_mds_session(last_session);
5569                                 last_session = s;
5570                         } else {
5571                                 ceph_put_mds_session(s);
5572                         }
5573                         doutc(cl, "wait on %llu (want %llu)\n",
5574                               req->r_tid, want_tid);
5575                         wait_for_completion(&req->r_safe_completion);
5576
5577                         mutex_lock(&mdsc->mutex);
5578                         ceph_mdsc_put_request(req);
5579                         if (!nextreq)
5580                                 break;  /* next dne before, so we're done! */
5581                         if (RB_EMPTY_NODE(&nextreq->r_node)) {
5582                                 /* next request was removed from tree */
5583                                 ceph_mdsc_put_request(nextreq);
5584                                 goto restart;
5585                         }
5586                         ceph_mdsc_put_request(nextreq);  /* won't go away */
5587                 }
5588                 req = nextreq;
5589         }
5590         mutex_unlock(&mdsc->mutex);
5591         ceph_put_mds_session(last_session);
5592         doutc(cl, "done\n");
5593 }
5594
5595 void ceph_mdsc_sync(struct ceph_mds_client *mdsc)
5596 {
5597         struct ceph_client *cl = mdsc->fsc->client;
5598         u64 want_tid, want_flush;
5599
5600         if (READ_ONCE(mdsc->fsc->mount_state) >= CEPH_MOUNT_SHUTDOWN)
5601                 return;
5602
5603         doutc(cl, "sync\n");
5604         mutex_lock(&mdsc->mutex);
5605         want_tid = mdsc->last_tid;
5606         mutex_unlock(&mdsc->mutex);
5607
5608         ceph_flush_dirty_caps(mdsc);
5609         spin_lock(&mdsc->cap_dirty_lock);
5610         want_flush = mdsc->last_cap_flush_tid;
5611         if (!list_empty(&mdsc->cap_flush_list)) {
5612                 struct ceph_cap_flush *cf =
5613                         list_last_entry(&mdsc->cap_flush_list,
5614                                         struct ceph_cap_flush, g_list);
5615                 cf->wake = true;
5616         }
5617         spin_unlock(&mdsc->cap_dirty_lock);
5618
5619         doutc(cl, "sync want tid %lld flush_seq %lld\n", want_tid, want_flush);
5620
5621         flush_mdlog_and_wait_mdsc_unsafe_requests(mdsc, want_tid);
5622         wait_caps_flush(mdsc, want_flush);
5623 }
5624
5625 /*
5626  * true if all sessions are closed, or we force unmount
5627  */
5628 static bool done_closing_sessions(struct ceph_mds_client *mdsc, int skipped)
5629 {
5630         if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN)
5631                 return true;
5632         return atomic_read(&mdsc->num_sessions) <= skipped;
5633 }
5634
5635 /*
5636  * called after sb is ro or when metadata corrupted.
5637  */
5638 void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc)
5639 {
5640         struct ceph_options *opts = mdsc->fsc->client->options;
5641         struct ceph_client *cl = mdsc->fsc->client;
5642         struct ceph_mds_session *session;
5643         int i;
5644         int skipped = 0;
5645
5646         doutc(cl, "begin\n");
5647
5648         /* close sessions */
5649         mutex_lock(&mdsc->mutex);
5650         for (i = 0; i < mdsc->max_sessions; i++) {
5651                 session = __ceph_lookup_mds_session(mdsc, i);
5652                 if (!session)
5653                         continue;
5654                 mutex_unlock(&mdsc->mutex);
5655                 mutex_lock(&session->s_mutex);
5656                 if (__close_session(mdsc, session) <= 0)
5657                         skipped++;
5658                 mutex_unlock(&session->s_mutex);
5659                 ceph_put_mds_session(session);
5660                 mutex_lock(&mdsc->mutex);
5661         }
5662         mutex_unlock(&mdsc->mutex);
5663
5664         doutc(cl, "waiting for sessions to close\n");
5665         wait_event_timeout(mdsc->session_close_wq,
5666                            done_closing_sessions(mdsc, skipped),
5667                            ceph_timeout_jiffies(opts->mount_timeout));
5668
5669         /* tear down remaining sessions */
5670         mutex_lock(&mdsc->mutex);
5671         for (i = 0; i < mdsc->max_sessions; i++) {
5672                 if (mdsc->sessions[i]) {
5673                         session = ceph_get_mds_session(mdsc->sessions[i]);
5674                         __unregister_session(mdsc, session);
5675                         mutex_unlock(&mdsc->mutex);
5676                         mutex_lock(&session->s_mutex);
5677                         remove_session_caps(session);
5678                         mutex_unlock(&session->s_mutex);
5679                         ceph_put_mds_session(session);
5680                         mutex_lock(&mdsc->mutex);
5681                 }
5682         }
5683         WARN_ON(!list_empty(&mdsc->cap_delay_list));
5684         mutex_unlock(&mdsc->mutex);
5685
5686         ceph_cleanup_snapid_map(mdsc);
5687         ceph_cleanup_global_and_empty_realms(mdsc);
5688
5689         cancel_work_sync(&mdsc->cap_reclaim_work);
5690         cancel_work_sync(&mdsc->cap_unlink_work);
5691         cancel_delayed_work_sync(&mdsc->delayed_work); /* cancel timer */
5692
5693         doutc(cl, "done\n");
5694 }
5695
5696 void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc)
5697 {
5698         struct ceph_mds_session *session;
5699         int mds;
5700
5701         doutc(mdsc->fsc->client, "force umount\n");
5702
5703         mutex_lock(&mdsc->mutex);
5704         for (mds = 0; mds < mdsc->max_sessions; mds++) {
5705                 session = __ceph_lookup_mds_session(mdsc, mds);
5706                 if (!session)
5707                         continue;
5708
5709                 if (session->s_state == CEPH_MDS_SESSION_REJECTED)
5710                         __unregister_session(mdsc, session);
5711                 __wake_requests(mdsc, &session->s_waiting);
5712                 mutex_unlock(&mdsc->mutex);
5713
5714                 mutex_lock(&session->s_mutex);
5715                 __close_session(mdsc, session);
5716                 if (session->s_state == CEPH_MDS_SESSION_CLOSING) {
5717                         cleanup_session_requests(mdsc, session);
5718                         remove_session_caps(session);
5719                 }
5720                 mutex_unlock(&session->s_mutex);
5721                 ceph_put_mds_session(session);
5722
5723                 mutex_lock(&mdsc->mutex);
5724                 kick_requests(mdsc, mds);
5725         }
5726         __wake_requests(mdsc, &mdsc->waiting_for_map);
5727         mutex_unlock(&mdsc->mutex);
5728 }
5729
5730 static void ceph_mdsc_stop(struct ceph_mds_client *mdsc)
5731 {
5732         doutc(mdsc->fsc->client, "stop\n");
5733         /*
5734          * Make sure the delayed work stopped before releasing
5735          * the resources.
5736          *
5737          * Because the cancel_delayed_work_sync() will only
5738          * guarantee that the work finishes executing. But the
5739          * delayed work will re-arm itself again after that.
5740          */
5741         flush_delayed_work(&mdsc->delayed_work);
5742
5743         if (mdsc->mdsmap)
5744                 ceph_mdsmap_destroy(mdsc->mdsmap);
5745         kfree(mdsc->sessions);
5746         ceph_caps_finalize(mdsc);
5747         ceph_pool_perm_destroy(mdsc);
5748 }
5749
5750 void ceph_mdsc_destroy(struct ceph_fs_client *fsc)
5751 {
5752         struct ceph_mds_client *mdsc = fsc->mdsc;
5753         doutc(fsc->client, "%p\n", mdsc);
5754
5755         if (!mdsc)
5756                 return;
5757
5758         /* flush out any connection work with references to us */
5759         ceph_msgr_flush();
5760
5761         ceph_mdsc_stop(mdsc);
5762
5763         ceph_metric_destroy(&mdsc->metric);
5764
5765         fsc->mdsc = NULL;
5766         kfree(mdsc);
5767         doutc(fsc->client, "%p done\n", mdsc);
5768 }
5769
5770 void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
5771 {
5772         struct ceph_fs_client *fsc = mdsc->fsc;
5773         struct ceph_client *cl = fsc->client;
5774         const char *mds_namespace = fsc->mount_options->mds_namespace;
5775         void *p = msg->front.iov_base;
5776         void *end = p + msg->front.iov_len;
5777         u32 epoch;
5778         u32 num_fs;
5779         u32 mount_fscid = (u32)-1;
5780         int err = -EINVAL;
5781
5782         ceph_decode_need(&p, end, sizeof(u32), bad);
5783         epoch = ceph_decode_32(&p);
5784
5785         doutc(cl, "epoch %u\n", epoch);
5786
5787         /* struct_v, struct_cv, map_len, epoch, legacy_client_fscid */
5788         ceph_decode_skip_n(&p, end, 2 + sizeof(u32) * 3, bad);
5789
5790         ceph_decode_32_safe(&p, end, num_fs, bad);
5791         while (num_fs-- > 0) {
5792                 void *info_p, *info_end;
5793                 u32 info_len;
5794                 u32 fscid, namelen;
5795
5796                 ceph_decode_need(&p, end, 2 + sizeof(u32), bad);
5797                 p += 2;         // info_v, info_cv
5798                 info_len = ceph_decode_32(&p);
5799                 ceph_decode_need(&p, end, info_len, bad);
5800                 info_p = p;
5801                 info_end = p + info_len;
5802                 p = info_end;
5803
5804                 ceph_decode_need(&info_p, info_end, sizeof(u32) * 2, bad);
5805                 fscid = ceph_decode_32(&info_p);
5806                 namelen = ceph_decode_32(&info_p);
5807                 ceph_decode_need(&info_p, info_end, namelen, bad);
5808
5809                 if (mds_namespace &&
5810                     strlen(mds_namespace) == namelen &&
5811                     !strncmp(mds_namespace, (char *)info_p, namelen)) {
5812                         mount_fscid = fscid;
5813                         break;
5814                 }
5815         }
5816
5817         ceph_monc_got_map(&fsc->client->monc, CEPH_SUB_FSMAP, epoch);
5818         if (mount_fscid != (u32)-1) {
5819                 fsc->client->monc.fs_cluster_id = mount_fscid;
5820                 ceph_monc_want_map(&fsc->client->monc, CEPH_SUB_MDSMAP,
5821                                    0, true);
5822                 ceph_monc_renew_subs(&fsc->client->monc);
5823         } else {
5824                 err = -ENOENT;
5825                 goto err_out;
5826         }
5827         return;
5828
5829 bad:
5830         pr_err_client(cl, "error decoding fsmap %d. Shutting down mount.\n",
5831                       err);
5832         ceph_umount_begin(mdsc->fsc->sb);
5833         ceph_msg_dump(msg);
5834 err_out:
5835         mutex_lock(&mdsc->mutex);
5836         mdsc->mdsmap_err = err;
5837         __wake_requests(mdsc, &mdsc->waiting_for_map);
5838         mutex_unlock(&mdsc->mutex);
5839 }
5840
5841 /*
5842  * handle mds map update.
5843  */
5844 void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc, struct ceph_msg *msg)
5845 {
5846         struct ceph_client *cl = mdsc->fsc->client;
5847         u32 epoch;
5848         u32 maplen;
5849         void *p = msg->front.iov_base;
5850         void *end = p + msg->front.iov_len;
5851         struct ceph_mdsmap *newmap, *oldmap;
5852         struct ceph_fsid fsid;
5853         int err = -EINVAL;
5854
5855         ceph_decode_need(&p, end, sizeof(fsid)+2*sizeof(u32), bad);
5856         ceph_decode_copy(&p, &fsid, sizeof(fsid));
5857         if (ceph_check_fsid(mdsc->fsc->client, &fsid) < 0)
5858                 return;
5859         epoch = ceph_decode_32(&p);
5860         maplen = ceph_decode_32(&p);
5861         doutc(cl, "epoch %u len %d\n", epoch, (int)maplen);
5862
5863         /* do we need it? */
5864         mutex_lock(&mdsc->mutex);
5865         if (mdsc->mdsmap && epoch <= mdsc->mdsmap->m_epoch) {
5866                 doutc(cl, "epoch %u <= our %u\n", epoch, mdsc->mdsmap->m_epoch);
5867                 mutex_unlock(&mdsc->mutex);
5868                 return;
5869         }
5870
5871         newmap = ceph_mdsmap_decode(mdsc, &p, end, ceph_msgr2(mdsc->fsc->client));
5872         if (IS_ERR(newmap)) {
5873                 err = PTR_ERR(newmap);
5874                 goto bad_unlock;
5875         }
5876
5877         /* swap into place */
5878         if (mdsc->mdsmap) {
5879                 oldmap = mdsc->mdsmap;
5880                 mdsc->mdsmap = newmap;
5881                 check_new_map(mdsc, newmap, oldmap);
5882                 ceph_mdsmap_destroy(oldmap);
5883         } else {
5884                 mdsc->mdsmap = newmap;  /* first mds map */
5885         }
5886         mdsc->fsc->max_file_size = min((loff_t)mdsc->mdsmap->m_max_file_size,
5887                                         MAX_LFS_FILESIZE);
5888
5889         __wake_requests(mdsc, &mdsc->waiting_for_map);
5890         ceph_monc_got_map(&mdsc->fsc->client->monc, CEPH_SUB_MDSMAP,
5891                           mdsc->mdsmap->m_epoch);
5892
5893         mutex_unlock(&mdsc->mutex);
5894         schedule_delayed(mdsc, 0);
5895         return;
5896
5897 bad_unlock:
5898         mutex_unlock(&mdsc->mutex);
5899 bad:
5900         pr_err_client(cl, "error decoding mdsmap %d. Shutting down mount.\n",
5901                       err);
5902         ceph_umount_begin(mdsc->fsc->sb);
5903         ceph_msg_dump(msg);
5904         return;
5905 }
5906
5907 static struct ceph_connection *mds_get_con(struct ceph_connection *con)
5908 {
5909         struct ceph_mds_session *s = con->private;
5910
5911         if (ceph_get_mds_session(s))
5912                 return con;
5913         return NULL;
5914 }
5915
5916 static void mds_put_con(struct ceph_connection *con)
5917 {
5918         struct ceph_mds_session *s = con->private;
5919
5920         ceph_put_mds_session(s);
5921 }
5922
5923 /*
5924  * if the client is unresponsive for long enough, the mds will kill
5925  * the session entirely.
5926  */
5927 static void mds_peer_reset(struct ceph_connection *con)
5928 {
5929         struct ceph_mds_session *s = con->private;
5930         struct ceph_mds_client *mdsc = s->s_mdsc;
5931
5932         pr_warn_client(mdsc->fsc->client, "mds%d closed our session\n",
5933                        s->s_mds);
5934         if (READ_ONCE(mdsc->fsc->mount_state) != CEPH_MOUNT_FENCE_IO &&
5935             ceph_mdsmap_get_state(mdsc->mdsmap, s->s_mds) >= CEPH_MDS_STATE_RECONNECT)
5936                 send_mds_reconnect(mdsc, s);
5937 }
5938
5939 static void mds_dispatch(struct ceph_connection *con, struct ceph_msg *msg)
5940 {
5941         struct ceph_mds_session *s = con->private;
5942         struct ceph_mds_client *mdsc = s->s_mdsc;
5943         struct ceph_client *cl = mdsc->fsc->client;
5944         int type = le16_to_cpu(msg->hdr.type);
5945
5946         mutex_lock(&mdsc->mutex);
5947         if (__verify_registered_session(mdsc, s) < 0) {
5948                 mutex_unlock(&mdsc->mutex);
5949                 goto out;
5950         }
5951         mutex_unlock(&mdsc->mutex);
5952
5953         switch (type) {
5954         case CEPH_MSG_MDS_MAP:
5955                 ceph_mdsc_handle_mdsmap(mdsc, msg);
5956                 break;
5957         case CEPH_MSG_FS_MAP_USER:
5958                 ceph_mdsc_handle_fsmap(mdsc, msg);
5959                 break;
5960         case CEPH_MSG_CLIENT_SESSION:
5961                 handle_session(s, msg);
5962                 break;
5963         case CEPH_MSG_CLIENT_REPLY:
5964                 handle_reply(s, msg);
5965                 break;
5966         case CEPH_MSG_CLIENT_REQUEST_FORWARD:
5967                 handle_forward(mdsc, s, msg);
5968                 break;
5969         case CEPH_MSG_CLIENT_CAPS:
5970                 ceph_handle_caps(s, msg);
5971                 break;
5972         case CEPH_MSG_CLIENT_SNAP:
5973                 ceph_handle_snap(mdsc, s, msg);
5974                 break;
5975         case CEPH_MSG_CLIENT_LEASE:
5976                 handle_lease(mdsc, s, msg);
5977                 break;
5978         case CEPH_MSG_CLIENT_QUOTA:
5979                 ceph_handle_quota(mdsc, s, msg);
5980                 break;
5981
5982         default:
5983                 pr_err_client(cl, "received unknown message type %d %s\n",
5984                               type, ceph_msg_type_name(type));
5985         }
5986 out:
5987         ceph_msg_put(msg);
5988 }
5989
5990 /*
5991  * authentication
5992  */
5993
5994 /*
5995  * Note: returned pointer is the address of a structure that's
5996  * managed separately.  Caller must *not* attempt to free it.
5997  */
5998 static struct ceph_auth_handshake *
5999 mds_get_authorizer(struct ceph_connection *con, int *proto, int force_new)
6000 {
6001         struct ceph_mds_session *s = con->private;
6002         struct ceph_mds_client *mdsc = s->s_mdsc;
6003         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
6004         struct ceph_auth_handshake *auth = &s->s_auth;
6005         int ret;
6006
6007         ret = __ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_MDS,
6008                                          force_new, proto, NULL, NULL);
6009         if (ret)
6010                 return ERR_PTR(ret);
6011
6012         return auth;
6013 }
6014
6015 static int mds_add_authorizer_challenge(struct ceph_connection *con,
6016                                     void *challenge_buf, int challenge_buf_len)
6017 {
6018         struct ceph_mds_session *s = con->private;
6019         struct ceph_mds_client *mdsc = s->s_mdsc;
6020         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
6021
6022         return ceph_auth_add_authorizer_challenge(ac, s->s_auth.authorizer,
6023                                             challenge_buf, challenge_buf_len);
6024 }
6025
6026 static int mds_verify_authorizer_reply(struct ceph_connection *con)
6027 {
6028         struct ceph_mds_session *s = con->private;
6029         struct ceph_mds_client *mdsc = s->s_mdsc;
6030         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
6031         struct ceph_auth_handshake *auth = &s->s_auth;
6032
6033         return ceph_auth_verify_authorizer_reply(ac, auth->authorizer,
6034                 auth->authorizer_reply_buf, auth->authorizer_reply_buf_len,
6035                 NULL, NULL, NULL, NULL);
6036 }
6037
6038 static int mds_invalidate_authorizer(struct ceph_connection *con)
6039 {
6040         struct ceph_mds_session *s = con->private;
6041         struct ceph_mds_client *mdsc = s->s_mdsc;
6042         struct ceph_auth_client *ac = mdsc->fsc->client->monc.auth;
6043
6044         ceph_auth_invalidate_authorizer(ac, CEPH_ENTITY_TYPE_MDS);
6045
6046         return ceph_monc_validate_auth(&mdsc->fsc->client->monc);
6047 }
6048
6049 static int mds_get_auth_request(struct ceph_connection *con,
6050                                 void *buf, int *buf_len,
6051                                 void **authorizer, int *authorizer_len)
6052 {
6053         struct ceph_mds_session *s = con->private;
6054         struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth;
6055         struct ceph_auth_handshake *auth = &s->s_auth;
6056         int ret;
6057
6058         ret = ceph_auth_get_authorizer(ac, auth, CEPH_ENTITY_TYPE_MDS,
6059                                        buf, buf_len);
6060         if (ret)
6061                 return ret;
6062
6063         *authorizer = auth->authorizer_buf;
6064         *authorizer_len = auth->authorizer_buf_len;
6065         return 0;
6066 }
6067
6068 static int mds_handle_auth_reply_more(struct ceph_connection *con,
6069                                       void *reply, int reply_len,
6070                                       void *buf, int *buf_len,
6071                                       void **authorizer, int *authorizer_len)
6072 {
6073         struct ceph_mds_session *s = con->private;
6074         struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth;
6075         struct ceph_auth_handshake *auth = &s->s_auth;
6076         int ret;
6077
6078         ret = ceph_auth_handle_svc_reply_more(ac, auth, reply, reply_len,
6079                                               buf, buf_len);
6080         if (ret)
6081                 return ret;
6082
6083         *authorizer = auth->authorizer_buf;
6084         *authorizer_len = auth->authorizer_buf_len;
6085         return 0;
6086 }
6087
6088 static int mds_handle_auth_done(struct ceph_connection *con,
6089                                 u64 global_id, void *reply, int reply_len,
6090                                 u8 *session_key, int *session_key_len,
6091                                 u8 *con_secret, int *con_secret_len)
6092 {
6093         struct ceph_mds_session *s = con->private;
6094         struct ceph_auth_client *ac = s->s_mdsc->fsc->client->monc.auth;
6095         struct ceph_auth_handshake *auth = &s->s_auth;
6096
6097         return ceph_auth_handle_svc_reply_done(ac, auth, reply, reply_len,
6098                                                session_key, session_key_len,
6099                                                con_secret, con_secret_len);
6100 }
6101
6102 static int mds_handle_auth_bad_method(struct ceph_connection *con,
6103                                       int used_proto, int result,
6104                                       const int *allowed_protos, int proto_cnt,
6105                                       const int *allowed_modes, int mode_cnt)
6106 {
6107         struct ceph_mds_session *s = con->private;
6108         struct ceph_mon_client *monc = &s->s_mdsc->fsc->client->monc;
6109         int ret;
6110
6111         if (ceph_auth_handle_bad_authorizer(monc->auth, CEPH_ENTITY_TYPE_MDS,
6112                                             used_proto, result,
6113                                             allowed_protos, proto_cnt,
6114                                             allowed_modes, mode_cnt)) {
6115                 ret = ceph_monc_validate_auth(monc);
6116                 if (ret)
6117                         return ret;
6118         }
6119
6120         return -EACCES;
6121 }
6122
6123 static struct ceph_msg *mds_alloc_msg(struct ceph_connection *con,
6124                                 struct ceph_msg_header *hdr, int *skip)
6125 {
6126         struct ceph_msg *msg;
6127         int type = (int) le16_to_cpu(hdr->type);
6128         int front_len = (int) le32_to_cpu(hdr->front_len);
6129
6130         if (con->in_msg)
6131                 return con->in_msg;
6132
6133         *skip = 0;
6134         msg = ceph_msg_new(type, front_len, GFP_NOFS, false);
6135         if (!msg) {
6136                 pr_err("unable to allocate msg type %d len %d\n",
6137                        type, front_len);
6138                 return NULL;
6139         }
6140
6141         return msg;
6142 }
6143
6144 static int mds_sign_message(struct ceph_msg *msg)
6145 {
6146        struct ceph_mds_session *s = msg->con->private;
6147        struct ceph_auth_handshake *auth = &s->s_auth;
6148
6149        return ceph_auth_sign_message(auth, msg);
6150 }
6151
6152 static int mds_check_message_signature(struct ceph_msg *msg)
6153 {
6154        struct ceph_mds_session *s = msg->con->private;
6155        struct ceph_auth_handshake *auth = &s->s_auth;
6156
6157        return ceph_auth_check_message_signature(auth, msg);
6158 }
6159
6160 static const struct ceph_connection_operations mds_con_ops = {
6161         .get = mds_get_con,
6162         .put = mds_put_con,
6163         .alloc_msg = mds_alloc_msg,
6164         .dispatch = mds_dispatch,
6165         .peer_reset = mds_peer_reset,
6166         .get_authorizer = mds_get_authorizer,
6167         .add_authorizer_challenge = mds_add_authorizer_challenge,
6168         .verify_authorizer_reply = mds_verify_authorizer_reply,
6169         .invalidate_authorizer = mds_invalidate_authorizer,
6170         .sign_message = mds_sign_message,
6171         .check_message_signature = mds_check_message_signature,
6172         .get_auth_request = mds_get_auth_request,
6173         .handle_auth_reply_more = mds_handle_auth_reply_more,
6174         .handle_auth_done = mds_handle_auth_done,
6175         .handle_auth_bad_method = mds_handle_auth_bad_method,
6176 };
6177
6178 /* eof */