linux/fs/ubifs/orphan.c
<<
>>
Prefs
   1// SPDX-License-Identifier: GPL-2.0-only
   2/*
   3 * This file is part of UBIFS.
   4 *
   5 * Copyright (C) 2006-2008 Nokia Corporation.
   6 *
   7 * Author: Adrian Hunter
   8 */
   9
  10#include "ubifs.h"
  11
  12/*
  13 * An orphan is an inode number whose inode node has been committed to the index
  14 * with a link count of zero. That happens when an open file is deleted
  15 * (unlinked) and then a commit is run. In the normal course of events the inode
  16 * would be deleted when the file is closed. However in the case of an unclean
  17 * unmount, orphans need to be accounted for. After an unclean unmount, the
  18 * orphans' inodes must be deleted which means either scanning the entire index
  19 * looking for them, or keeping a list on flash somewhere. This unit implements
  20 * the latter approach.
  21 *
  22 * The orphan area is a fixed number of LEBs situated between the LPT area and
  23 * the main area. The number of orphan area LEBs is specified when the file
  24 * system is created. The minimum number is 1. The size of the orphan area
  25 * should be so that it can hold the maximum number of orphans that are expected
  26 * to ever exist at one time.
  27 *
  28 * The number of orphans that can fit in a LEB is:
  29 *
  30 *         (c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64)
  31 *
  32 * For example: a 15872 byte LEB can fit 1980 orphans so 1 LEB may be enough.
  33 *
  34 * Orphans are accumulated in a rb-tree. When an inode's link count drops to
  35 * zero, the inode number is added to the rb-tree. It is removed from the tree
  36 * when the inode is deleted.  Any new orphans that are in the orphan tree when
  37 * the commit is run, are written to the orphan area in 1 or more orphan nodes.
  38 * If the orphan area is full, it is consolidated to make space.  There is
  39 * always enough space because validation prevents the user from creating more
  40 * than the maximum number of orphans allowed.
  41 */
  42
  43static int dbg_check_orphans(struct ubifs_info *c);
  44
  45static struct ubifs_orphan *orphan_add(struct ubifs_info *c, ino_t inum,
  46                                       struct ubifs_orphan *parent_orphan)
  47{
  48        struct ubifs_orphan *orphan, *o;
  49        struct rb_node **p, *parent = NULL;
  50
  51        orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_NOFS);
  52        if (!orphan)
  53                return ERR_PTR(-ENOMEM);
  54        orphan->inum = inum;
  55        orphan->new = 1;
  56        INIT_LIST_HEAD(&orphan->child_list);
  57
  58        spin_lock(&c->orphan_lock);
  59        if (c->tot_orphans >= c->max_orphans) {
  60                spin_unlock(&c->orphan_lock);
  61                kfree(orphan);
  62                return ERR_PTR(-ENFILE);
  63        }
  64        p = &c->orph_tree.rb_node;
  65        while (*p) {
  66                parent = *p;
  67                o = rb_entry(parent, struct ubifs_orphan, rb);
  68                if (inum < o->inum)
  69                        p = &(*p)->rb_left;
  70                else if (inum > o->inum)
  71                        p = &(*p)->rb_right;
  72                else {
  73                        ubifs_err(c, "orphaned twice");
  74                        spin_unlock(&c->orphan_lock);
  75                        kfree(orphan);
  76                        return ERR_PTR(-EINVAL);
  77                }
  78        }
  79        c->tot_orphans += 1;
  80        c->new_orphans += 1;
  81        rb_link_node(&orphan->rb, parent, p);
  82        rb_insert_color(&orphan->rb, &c->orph_tree);
  83        list_add_tail(&orphan->list, &c->orph_list);
  84        list_add_tail(&orphan->new_list, &c->orph_new);
  85
  86        if (parent_orphan) {
  87                list_add_tail(&orphan->child_list,
  88                              &parent_orphan->child_list);
  89        }
  90
  91        spin_unlock(&c->orphan_lock);
  92        dbg_gen("ino %lu", (unsigned long)inum);
  93        return orphan;
  94}
  95
  96static struct ubifs_orphan *lookup_orphan(struct ubifs_info *c, ino_t inum)
  97{
  98        struct ubifs_orphan *o;
  99        struct rb_node *p;
 100
 101        p = c->orph_tree.rb_node;
 102        while (p) {
 103                o = rb_entry(p, struct ubifs_orphan, rb);
 104                if (inum < o->inum)
 105                        p = p->rb_left;
 106                else if (inum > o->inum)
 107                        p = p->rb_right;
 108                else {
 109                        return o;
 110                }
 111        }
 112        return NULL;
 113}
 114
 115static void __orphan_drop(struct ubifs_info *c, struct ubifs_orphan *o)
 116{
 117        rb_erase(&o->rb, &c->orph_tree);
 118        list_del(&o->list);
 119        c->tot_orphans -= 1;
 120
 121        if (o->new) {
 122                list_del(&o->new_list);
 123                c->new_orphans -= 1;
 124        }
 125
 126        kfree(o);
 127}
 128
 129static void orphan_delete(struct ubifs_info *c, struct ubifs_orphan *orph)
 130{
 131        if (orph->del) {
 132                dbg_gen("deleted twice ino %lu", orph->inum);
 133                return;
 134        }
 135
 136        if (orph->cmt) {
 137                orph->del = 1;
 138                orph->dnext = c->orph_dnext;
 139                c->orph_dnext = orph;
 140                dbg_gen("delete later ino %lu", orph->inum);
 141                return;
 142        }
 143
 144        __orphan_drop(c, orph);
 145}
 146
 147/**
 148 * ubifs_add_orphan - add an orphan.
 149 * @c: UBIFS file-system description object
 150 * @inum: orphan inode number
 151 *
 152 * Add an orphan. This function is called when an inodes link count drops to
 153 * zero.
 154 */
 155int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
 156{
 157        int err = 0;
 158        ino_t xattr_inum;
 159        union ubifs_key key;
 160        struct ubifs_dent_node *xent;
 161        struct fscrypt_name nm = {0};
 162        struct ubifs_orphan *xattr_orphan;
 163        struct ubifs_orphan *orphan;
 164
 165        orphan = orphan_add(c, inum, NULL);
 166        if (IS_ERR(orphan))
 167                return PTR_ERR(orphan);
 168
 169        lowest_xent_key(c, &key, inum);
 170        while (1) {
 171                xent = ubifs_tnc_next_ent(c, &key, &nm);
 172                if (IS_ERR(xent)) {
 173                        err = PTR_ERR(xent);
 174                        if (err == -ENOENT)
 175                                break;
 176                        return err;
 177                }
 178
 179                fname_name(&nm) = xent->name;
 180                fname_len(&nm) = le16_to_cpu(xent->nlen);
 181                xattr_inum = le64_to_cpu(xent->inum);
 182
 183                xattr_orphan = orphan_add(c, xattr_inum, orphan);
 184                if (IS_ERR(xattr_orphan))
 185                        return PTR_ERR(xattr_orphan);
 186
 187                key_read(c, &xent->key, &key);
 188        }
 189
 190        return 0;
 191}
 192
 193/**
 194 * ubifs_delete_orphan - delete an orphan.
 195 * @c: UBIFS file-system description object
 196 * @inum: orphan inode number
 197 *
 198 * Delete an orphan. This function is called when an inode is deleted.
 199 */
 200void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
 201{
 202        struct ubifs_orphan *orph, *child_orph, *tmp_o;
 203
 204        spin_lock(&c->orphan_lock);
 205
 206        orph = lookup_orphan(c, inum);
 207        if (!orph) {
 208                spin_unlock(&c->orphan_lock);
 209                ubifs_err(c, "missing orphan ino %lu", (unsigned long)inum);
 210                dump_stack();
 211
 212                return;
 213        }
 214
 215        list_for_each_entry_safe(child_orph, tmp_o, &orph->child_list, child_list) {
 216                list_del(&child_orph->child_list);
 217                orphan_delete(c, child_orph);
 218        }
 219        
 220        orphan_delete(c, orph);
 221
 222        spin_unlock(&c->orphan_lock);
 223}
 224
 225/**
 226 * ubifs_orphan_start_commit - start commit of orphans.
 227 * @c: UBIFS file-system description object
 228 *
 229 * Start commit of orphans.
 230 */
 231int ubifs_orphan_start_commit(struct ubifs_info *c)
 232{
 233        struct ubifs_orphan *orphan, **last;
 234
 235        spin_lock(&c->orphan_lock);
 236        last = &c->orph_cnext;
 237        list_for_each_entry(orphan, &c->orph_new, new_list) {
 238                ubifs_assert(c, orphan->new);
 239                ubifs_assert(c, !orphan->cmt);
 240                orphan->new = 0;
 241                orphan->cmt = 1;
 242                *last = orphan;
 243                last = &orphan->cnext;
 244        }
 245        *last = NULL;
 246        c->cmt_orphans = c->new_orphans;
 247        c->new_orphans = 0;
 248        dbg_cmt("%d orphans to commit", c->cmt_orphans);
 249        INIT_LIST_HEAD(&c->orph_new);
 250        if (c->tot_orphans == 0)
 251                c->no_orphs = 1;
 252        else
 253                c->no_orphs = 0;
 254        spin_unlock(&c->orphan_lock);
 255        return 0;
 256}
 257
 258/**
 259 * avail_orphs - calculate available space.
 260 * @c: UBIFS file-system description object
 261 *
 262 * This function returns the number of orphans that can be written in the
 263 * available space.
 264 */
 265static int avail_orphs(struct ubifs_info *c)
 266{
 267        int avail_lebs, avail, gap;
 268
 269        avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
 270        avail = avail_lebs *
 271               ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
 272        gap = c->leb_size - c->ohead_offs;
 273        if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
 274                avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
 275        return avail;
 276}
 277
 278/**
 279 * tot_avail_orphs - calculate total space.
 280 * @c: UBIFS file-system description object
 281 *
 282 * This function returns the number of orphans that can be written in half
 283 * the total space. That leaves half the space for adding new orphans.
 284 */
 285static int tot_avail_orphs(struct ubifs_info *c)
 286{
 287        int avail_lebs, avail;
 288
 289        avail_lebs = c->orph_lebs;
 290        avail = avail_lebs *
 291               ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
 292        return avail / 2;
 293}
 294
 295/**
 296 * do_write_orph_node - write a node to the orphan head.
 297 * @c: UBIFS file-system description object
 298 * @len: length of node
 299 * @atomic: write atomically
 300 *
 301 * This function writes a node to the orphan head from the orphan buffer. If
 302 * %atomic is not zero, then the write is done atomically. On success, %0 is
 303 * returned, otherwise a negative error code is returned.
 304 */
 305static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
 306{
 307        int err = 0;
 308
 309        if (atomic) {
 310                ubifs_assert(c, c->ohead_offs == 0);
 311                ubifs_prepare_node(c, c->orph_buf, len, 1);
 312                len = ALIGN(len, c->min_io_size);
 313                err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
 314        } else {
 315                if (c->ohead_offs == 0) {
 316                        /* Ensure LEB has been unmapped */
 317                        err = ubifs_leb_unmap(c, c->ohead_lnum);
 318                        if (err)
 319                                return err;
 320                }
 321                err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
 322                                       c->ohead_offs);
 323        }
 324        return err;
 325}
 326
 327/**
 328 * write_orph_node - write an orphan node.
 329 * @c: UBIFS file-system description object
 330 * @atomic: write atomically
 331 *
 332 * This function builds an orphan node from the cnext list and writes it to the
 333 * orphan head. On success, %0 is returned, otherwise a negative error code
 334 * is returned.
 335 */
 336static int write_orph_node(struct ubifs_info *c, int atomic)
 337{
 338        struct ubifs_orphan *orphan, *cnext;
 339        struct ubifs_orph_node *orph;
 340        int gap, err, len, cnt, i;
 341
 342        ubifs_assert(c, c->cmt_orphans > 0);
 343        gap = c->leb_size - c->ohead_offs;
 344        if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
 345                c->ohead_lnum += 1;
 346                c->ohead_offs = 0;
 347                gap = c->leb_size;
 348                if (c->ohead_lnum > c->orph_last) {
 349                        /*
 350                         * We limit the number of orphans so that this should
 351                         * never happen.
 352                         */
 353                        ubifs_err(c, "out of space in orphan area");
 354                        return -EINVAL;
 355                }
 356        }
 357        cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
 358        if (cnt > c->cmt_orphans)
 359                cnt = c->cmt_orphans;
 360        len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
 361        ubifs_assert(c, c->orph_buf);
 362        orph = c->orph_buf;
 363        orph->ch.node_type = UBIFS_ORPH_NODE;
 364        spin_lock(&c->orphan_lock);
 365        cnext = c->orph_cnext;
 366        for (i = 0; i < cnt; i++) {
 367                orphan = cnext;
 368                ubifs_assert(c, orphan->cmt);
 369                orph->inos[i] = cpu_to_le64(orphan->inum);
 370                orphan->cmt = 0;
 371                cnext = orphan->cnext;
 372                orphan->cnext = NULL;
 373        }
 374        c->orph_cnext = cnext;
 375        c->cmt_orphans -= cnt;
 376        spin_unlock(&c->orphan_lock);
 377        if (c->cmt_orphans)
 378                orph->cmt_no = cpu_to_le64(c->cmt_no);
 379        else
 380                /* Mark the last node of the commit */
 381                orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
 382        ubifs_assert(c, c->ohead_offs + len <= c->leb_size);
 383        ubifs_assert(c, c->ohead_lnum >= c->orph_first);
 384        ubifs_assert(c, c->ohead_lnum <= c->orph_last);
 385        err = do_write_orph_node(c, len, atomic);
 386        c->ohead_offs += ALIGN(len, c->min_io_size);
 387        c->ohead_offs = ALIGN(c->ohead_offs, 8);
 388        return err;
 389}
 390
 391/**
 392 * write_orph_nodes - write orphan nodes until there are no more to commit.
 393 * @c: UBIFS file-system description object
 394 * @atomic: write atomically
 395 *
 396 * This function writes orphan nodes for all the orphans to commit. On success,
 397 * %0 is returned, otherwise a negative error code is returned.
 398 */
 399static int write_orph_nodes(struct ubifs_info *c, int atomic)
 400{
 401        int err;
 402
 403        while (c->cmt_orphans > 0) {
 404                err = write_orph_node(c, atomic);
 405                if (err)
 406                        return err;
 407        }
 408        if (atomic) {
 409                int lnum;
 410
 411                /* Unmap any unused LEBs after consolidation */
 412                for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
 413                        err = ubifs_leb_unmap(c, lnum);
 414                        if (err)
 415                                return err;
 416                }
 417        }
 418        return 0;
 419}
 420
 421/**
 422 * consolidate - consolidate the orphan area.
 423 * @c: UBIFS file-system description object
 424 *
 425 * This function enables consolidation by putting all the orphans into the list
 426 * to commit. The list is in the order that the orphans were added, and the
 427 * LEBs are written atomically in order, so at no time can orphans be lost by
 428 * an unclean unmount.
 429 *
 430 * This function returns %0 on success and a negative error code on failure.
 431 */
 432static int consolidate(struct ubifs_info *c)
 433{
 434        int tot_avail = tot_avail_orphs(c), err = 0;
 435
 436        spin_lock(&c->orphan_lock);
 437        dbg_cmt("there is space for %d orphans and there are %d",
 438                tot_avail, c->tot_orphans);
 439        if (c->tot_orphans - c->new_orphans <= tot_avail) {
 440                struct ubifs_orphan *orphan, **last;
 441                int cnt = 0;
 442
 443                /* Change the cnext list to include all non-new orphans */
 444                last = &c->orph_cnext;
 445                list_for_each_entry(orphan, &c->orph_list, list) {
 446                        if (orphan->new)
 447                                continue;
 448                        orphan->cmt = 1;
 449                        *last = orphan;
 450                        last = &orphan->cnext;
 451                        cnt += 1;
 452                }
 453                *last = NULL;
 454                ubifs_assert(c, cnt == c->tot_orphans - c->new_orphans);
 455                c->cmt_orphans = cnt;
 456                c->ohead_lnum = c->orph_first;
 457                c->ohead_offs = 0;
 458        } else {
 459                /*
 460                 * We limit the number of orphans so that this should
 461                 * never happen.
 462                 */
 463                ubifs_err(c, "out of space in orphan area");
 464                err = -EINVAL;
 465        }
 466        spin_unlock(&c->orphan_lock);
 467        return err;
 468}
 469
 470/**
 471 * commit_orphans - commit orphans.
 472 * @c: UBIFS file-system description object
 473 *
 474 * This function commits orphans to flash. On success, %0 is returned,
 475 * otherwise a negative error code is returned.
 476 */
 477static int commit_orphans(struct ubifs_info *c)
 478{
 479        int avail, atomic = 0, err;
 480
 481        ubifs_assert(c, c->cmt_orphans > 0);
 482        avail = avail_orphs(c);
 483        if (avail < c->cmt_orphans) {
 484                /* Not enough space to write new orphans, so consolidate */
 485                err = consolidate(c);
 486                if (err)
 487                        return err;
 488                atomic = 1;
 489        }
 490        err = write_orph_nodes(c, atomic);
 491        return err;
 492}
 493
 494/**
 495 * erase_deleted - erase the orphans marked for deletion.
 496 * @c: UBIFS file-system description object
 497 *
 498 * During commit, the orphans being committed cannot be deleted, so they are
 499 * marked for deletion and deleted by this function. Also, the recovery
 500 * adds killed orphans to the deletion list, and therefore they are deleted
 501 * here too.
 502 */
 503static void erase_deleted(struct ubifs_info *c)
 504{
 505        struct ubifs_orphan *orphan, *dnext;
 506
 507        spin_lock(&c->orphan_lock);
 508        dnext = c->orph_dnext;
 509        while (dnext) {
 510                orphan = dnext;
 511                dnext = orphan->dnext;
 512                ubifs_assert(c, !orphan->new);
 513                ubifs_assert(c, orphan->del);
 514                rb_erase(&orphan->rb, &c->orph_tree);
 515                list_del(&orphan->list);
 516                c->tot_orphans -= 1;
 517                dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
 518                kfree(orphan);
 519        }
 520        c->orph_dnext = NULL;
 521        spin_unlock(&c->orphan_lock);
 522}
 523
 524/**
 525 * ubifs_orphan_end_commit - end commit of orphans.
 526 * @c: UBIFS file-system description object
 527 *
 528 * End commit of orphans.
 529 */
 530int ubifs_orphan_end_commit(struct ubifs_info *c)
 531{
 532        int err;
 533
 534        if (c->cmt_orphans != 0) {
 535                err = commit_orphans(c);
 536                if (err)
 537                        return err;
 538        }
 539        erase_deleted(c);
 540        err = dbg_check_orphans(c);
 541        return err;
 542}
 543
 544/**
 545 * ubifs_clear_orphans - erase all LEBs used for orphans.
 546 * @c: UBIFS file-system description object
 547 *
 548 * If recovery is not required, then the orphans from the previous session
 549 * are not needed. This function locates the LEBs used to record
 550 * orphans, and un-maps them.
 551 */
 552int ubifs_clear_orphans(struct ubifs_info *c)
 553{
 554        int lnum, err;
 555
 556        for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
 557                err = ubifs_leb_unmap(c, lnum);
 558                if (err)
 559                        return err;
 560        }
 561        c->ohead_lnum = c->orph_first;
 562        c->ohead_offs = 0;
 563        return 0;
 564}
 565
 566/**
 567 * insert_dead_orphan - insert an orphan.
 568 * @c: UBIFS file-system description object
 569 * @inum: orphan inode number
 570 *
 571 * This function is a helper to the 'do_kill_orphans()' function. The orphan
 572 * must be kept until the next commit, so it is added to the rb-tree and the
 573 * deletion list.
 574 */
 575static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
 576{
 577        struct ubifs_orphan *orphan, *o;
 578        struct rb_node **p, *parent = NULL;
 579
 580        orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
 581        if (!orphan)
 582                return -ENOMEM;
 583        orphan->inum = inum;
 584
 585        p = &c->orph_tree.rb_node;
 586        while (*p) {
 587                parent = *p;
 588                o = rb_entry(parent, struct ubifs_orphan, rb);
 589                if (inum < o->inum)
 590                        p = &(*p)->rb_left;
 591                else if (inum > o->inum)
 592                        p = &(*p)->rb_right;
 593                else {
 594                        /* Already added - no problem */
 595                        kfree(orphan);
 596                        return 0;
 597                }
 598        }
 599        c->tot_orphans += 1;
 600        rb_link_node(&orphan->rb, parent, p);
 601        rb_insert_color(&orphan->rb, &c->orph_tree);
 602        list_add_tail(&orphan->list, &c->orph_list);
 603        orphan->del = 1;
 604        orphan->dnext = c->orph_dnext;
 605        c->orph_dnext = orphan;
 606        dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
 607                c->new_orphans, c->tot_orphans);
 608        return 0;
 609}
 610
 611/**
 612 * do_kill_orphans - remove orphan inodes from the index.
 613 * @c: UBIFS file-system description object
 614 * @sleb: scanned LEB
 615 * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
 616 * @outofdate: whether the LEB is out of date is returned here
 617 * @last_flagged: whether the end orphan node is encountered
 618 *
 619 * This function is a helper to the 'kill_orphans()' function. It goes through
 620 * every orphan node in a LEB and for every inode number recorded, removes
 621 * all keys for that inode from the TNC.
 622 */
 623static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
 624                           unsigned long long *last_cmt_no, int *outofdate,
 625                           int *last_flagged)
 626{
 627        struct ubifs_scan_node *snod;
 628        struct ubifs_orph_node *orph;
 629        struct ubifs_ino_node *ino = NULL;
 630        unsigned long long cmt_no;
 631        ino_t inum;
 632        int i, n, err, first = 1;
 633
 634        list_for_each_entry(snod, &sleb->nodes, list) {
 635                if (snod->type != UBIFS_ORPH_NODE) {
 636                        ubifs_err(c, "invalid node type %d in orphan area at %d:%d",
 637                                  snod->type, sleb->lnum, snod->offs);
 638                        ubifs_dump_node(c, snod->node);
 639                        return -EINVAL;
 640                }
 641
 642                orph = snod->node;
 643
 644                /* Check commit number */
 645                cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
 646                /*
 647                 * The commit number on the master node may be less, because
 648                 * of a failed commit. If there are several failed commits in a
 649                 * row, the commit number written on orphan nodes will continue
 650                 * to increase (because the commit number is adjusted here) even
 651                 * though the commit number on the master node stays the same
 652                 * because the master node has not been re-written.
 653                 */
 654                if (cmt_no > c->cmt_no)
 655                        c->cmt_no = cmt_no;
 656                if (cmt_no < *last_cmt_no && *last_flagged) {
 657                        /*
 658                         * The last orphan node had a higher commit number and
 659                         * was flagged as the last written for that commit
 660                         * number. That makes this orphan node, out of date.
 661                         */
 662                        if (!first) {
 663                                ubifs_err(c, "out of order commit number %llu in orphan node at %d:%d",
 664                                          cmt_no, sleb->lnum, snod->offs);
 665                                ubifs_dump_node(c, snod->node);
 666                                return -EINVAL;
 667                        }
 668                        dbg_rcvry("out of date LEB %d", sleb->lnum);
 669                        *outofdate = 1;
 670                        return 0;
 671                }
 672
 673                if (first)
 674                        first = 0;
 675
 676                ino = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
 677                if (!ino)
 678                        return -ENOMEM;
 679
 680                n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
 681                for (i = 0; i < n; i++) {
 682                        union ubifs_key key1, key2;
 683
 684                        inum = le64_to_cpu(orph->inos[i]);
 685
 686                        ino_key_init(c, &key1, inum);
 687                        err = ubifs_tnc_lookup(c, &key1, ino);
 688                        if (err)
 689                                goto out_free;
 690
 691                        /*
 692                         * Check whether an inode can really get deleted.
 693                         * linkat() with O_TMPFILE allows rebirth of an inode.
 694                         */
 695                        if (ino->nlink == 0) {
 696                                dbg_rcvry("deleting orphaned inode %lu",
 697                                          (unsigned long)inum);
 698
 699                                lowest_ino_key(c, &key1, inum);
 700                                highest_ino_key(c, &key2, inum);
 701
 702                                err = ubifs_tnc_remove_range(c, &key1, &key2);
 703                                if (err)
 704                                        goto out_ro;
 705                        }
 706
 707                        err = insert_dead_orphan(c, inum);
 708                        if (err)
 709                                goto out_free;
 710                }
 711
 712                *last_cmt_no = cmt_no;
 713                if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
 714                        dbg_rcvry("last orph node for commit %llu at %d:%d",
 715                                  cmt_no, sleb->lnum, snod->offs);
 716                        *last_flagged = 1;
 717                } else
 718                        *last_flagged = 0;
 719        }
 720
 721        err = 0;
 722out_free:
 723        kfree(ino);
 724        return err;
 725
 726out_ro:
 727        ubifs_ro_mode(c, err);
 728        kfree(ino);
 729        return err;
 730}
 731
 732/**
 733 * kill_orphans - remove all orphan inodes from the index.
 734 * @c: UBIFS file-system description object
 735 *
 736 * If recovery is required, then orphan inodes recorded during the previous
 737 * session (which ended with an unclean unmount) must be deleted from the index.
 738 * This is done by updating the TNC, but since the index is not updated until
 739 * the next commit, the LEBs where the orphan information is recorded are not
 740 * erased until the next commit.
 741 */
 742static int kill_orphans(struct ubifs_info *c)
 743{
 744        unsigned long long last_cmt_no = 0;
 745        int lnum, err = 0, outofdate = 0, last_flagged = 0;
 746
 747        c->ohead_lnum = c->orph_first;
 748        c->ohead_offs = 0;
 749        /* Check no-orphans flag and skip this if no orphans */
 750        if (c->no_orphs) {
 751                dbg_rcvry("no orphans");
 752                return 0;
 753        }
 754        /*
 755         * Orph nodes always start at c->orph_first and are written to each
 756         * successive LEB in turn. Generally unused LEBs will have been unmapped
 757         * but may contain out of date orphan nodes if the unmap didn't go
 758         * through. In addition, the last orphan node written for each commit is
 759         * marked (top bit of orph->cmt_no is set to 1). It is possible that
 760         * there are orphan nodes from the next commit (i.e. the commit did not
 761         * complete successfully). In that case, no orphans will have been lost
 762         * due to the way that orphans are written, and any orphans added will
 763         * be valid orphans anyway and so can be deleted.
 764         */
 765        for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
 766                struct ubifs_scan_leb *sleb;
 767
 768                dbg_rcvry("LEB %d", lnum);
 769                sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
 770                if (IS_ERR(sleb)) {
 771                        if (PTR_ERR(sleb) == -EUCLEAN)
 772                                sleb = ubifs_recover_leb(c, lnum, 0,
 773                                                         c->sbuf, -1);
 774                        if (IS_ERR(sleb)) {
 775                                err = PTR_ERR(sleb);
 776                                break;
 777                        }
 778                }
 779                err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
 780                                      &last_flagged);
 781                if (err || outofdate) {
 782                        ubifs_scan_destroy(sleb);
 783                        break;
 784                }
 785                if (sleb->endpt) {
 786                        c->ohead_lnum = lnum;
 787                        c->ohead_offs = sleb->endpt;
 788                }
 789                ubifs_scan_destroy(sleb);
 790        }
 791        return err;
 792}
 793
 794/**
 795 * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
 796 * @c: UBIFS file-system description object
 797 * @unclean: indicates recovery from unclean unmount
 798 * @read_only: indicates read only mount
 799 *
 800 * This function is called when mounting to erase orphans from the previous
 801 * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
 802 * orphans are deleted.
 803 */
 804int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
 805{
 806        int err = 0;
 807
 808        c->max_orphans = tot_avail_orphs(c);
 809
 810        if (!read_only) {
 811                c->orph_buf = vmalloc(c->leb_size);
 812                if (!c->orph_buf)
 813                        return -ENOMEM;
 814        }
 815
 816        if (unclean)
 817                err = kill_orphans(c);
 818        else if (!read_only)
 819                err = ubifs_clear_orphans(c);
 820
 821        return err;
 822}
 823
 824/*
 825 * Everything below is related to debugging.
 826 */
 827
 828struct check_orphan {
 829        struct rb_node rb;
 830        ino_t inum;
 831};
 832
 833struct check_info {
 834        unsigned long last_ino;
 835        unsigned long tot_inos;
 836        unsigned long missing;
 837        unsigned long long leaf_cnt;
 838        struct ubifs_ino_node *node;
 839        struct rb_root root;
 840};
 841
 842static bool dbg_find_orphan(struct ubifs_info *c, ino_t inum)
 843{
 844        bool found = false;
 845
 846        spin_lock(&c->orphan_lock);
 847        found = !!lookup_orphan(c, inum);
 848        spin_unlock(&c->orphan_lock);
 849
 850        return found;
 851}
 852
 853static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
 854{
 855        struct check_orphan *orphan, *o;
 856        struct rb_node **p, *parent = NULL;
 857
 858        orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
 859        if (!orphan)
 860                return -ENOMEM;
 861        orphan->inum = inum;
 862
 863        p = &root->rb_node;
 864        while (*p) {
 865                parent = *p;
 866                o = rb_entry(parent, struct check_orphan, rb);
 867                if (inum < o->inum)
 868                        p = &(*p)->rb_left;
 869                else if (inum > o->inum)
 870                        p = &(*p)->rb_right;
 871                else {
 872                        kfree(orphan);
 873                        return 0;
 874                }
 875        }
 876        rb_link_node(&orphan->rb, parent, p);
 877        rb_insert_color(&orphan->rb, root);
 878        return 0;
 879}
 880
 881static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
 882{
 883        struct check_orphan *o;
 884        struct rb_node *p;
 885
 886        p = root->rb_node;
 887        while (p) {
 888                o = rb_entry(p, struct check_orphan, rb);
 889                if (inum < o->inum)
 890                        p = p->rb_left;
 891                else if (inum > o->inum)
 892                        p = p->rb_right;
 893                else
 894                        return 1;
 895        }
 896        return 0;
 897}
 898
 899static void dbg_free_check_tree(struct rb_root *root)
 900{
 901        struct check_orphan *o, *n;
 902
 903        rbtree_postorder_for_each_entry_safe(o, n, root, rb)
 904                kfree(o);
 905}
 906
 907static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
 908                            void *priv)
 909{
 910        struct check_info *ci = priv;
 911        ino_t inum;
 912        int err;
 913
 914        inum = key_inum(c, &zbr->key);
 915        if (inum != ci->last_ino) {
 916                /* Lowest node type is the inode node, so it comes first */
 917                if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
 918                        ubifs_err(c, "found orphan node ino %lu, type %d",
 919                                  (unsigned long)inum, key_type(c, &zbr->key));
 920                ci->last_ino = inum;
 921                ci->tot_inos += 1;
 922                err = ubifs_tnc_read_node(c, zbr, ci->node);
 923                if (err) {
 924                        ubifs_err(c, "node read failed, error %d", err);
 925                        return err;
 926                }
 927                if (ci->node->nlink == 0)
 928                        /* Must be recorded as an orphan */
 929                        if (!dbg_find_check_orphan(&ci->root, inum) &&
 930                            !dbg_find_orphan(c, inum)) {
 931                                ubifs_err(c, "missing orphan, ino %lu",
 932                                          (unsigned long)inum);
 933                                ci->missing += 1;
 934                        }
 935        }
 936        ci->leaf_cnt += 1;
 937        return 0;
 938}
 939
 940static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
 941{
 942        struct ubifs_scan_node *snod;
 943        struct ubifs_orph_node *orph;
 944        ino_t inum;
 945        int i, n, err;
 946
 947        list_for_each_entry(snod, &sleb->nodes, list) {
 948                cond_resched();
 949                if (snod->type != UBIFS_ORPH_NODE)
 950                        continue;
 951                orph = snod->node;
 952                n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
 953                for (i = 0; i < n; i++) {
 954                        inum = le64_to_cpu(orph->inos[i]);
 955                        err = dbg_ins_check_orphan(&ci->root, inum);
 956                        if (err)
 957                                return err;
 958                }
 959        }
 960        return 0;
 961}
 962
 963static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
 964{
 965        int lnum, err = 0;
 966        void *buf;
 967
 968        /* Check no-orphans flag and skip this if no orphans */
 969        if (c->no_orphs)
 970                return 0;
 971
 972        buf = __vmalloc(c->leb_size, GFP_NOFS, PAGE_KERNEL);
 973        if (!buf) {
 974                ubifs_err(c, "cannot allocate memory to check orphans");
 975                return 0;
 976        }
 977
 978        for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
 979                struct ubifs_scan_leb *sleb;
 980
 981                sleb = ubifs_scan(c, lnum, 0, buf, 0);
 982                if (IS_ERR(sleb)) {
 983                        err = PTR_ERR(sleb);
 984                        break;
 985                }
 986
 987                err = dbg_read_orphans(ci, sleb);
 988                ubifs_scan_destroy(sleb);
 989                if (err)
 990                        break;
 991        }
 992
 993        vfree(buf);
 994        return err;
 995}
 996
 997static int dbg_check_orphans(struct ubifs_info *c)
 998{
 999        struct check_info ci;
1000        int err;
1001
1002        if (!dbg_is_chk_orph(c))
1003                return 0;
1004
1005        ci.last_ino = 0;
1006        ci.tot_inos = 0;
1007        ci.missing  = 0;
1008        ci.leaf_cnt = 0;
1009        ci.root = RB_ROOT;
1010        ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
1011        if (!ci.node) {
1012                ubifs_err(c, "out of memory");
1013                return -ENOMEM;
1014        }
1015
1016        err = dbg_scan_orphans(c, &ci);
1017        if (err)
1018                goto out;
1019
1020        err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
1021        if (err) {
1022                ubifs_err(c, "cannot scan TNC, error %d", err);
1023                goto out;
1024        }
1025
1026        if (ci.missing) {
1027                ubifs_err(c, "%lu missing orphan(s)", ci.missing);
1028                err = -EINVAL;
1029                goto out;
1030        }
1031
1032        dbg_cmt("last inode number is %lu", ci.last_ino);
1033        dbg_cmt("total number of inodes is %lu", ci.tot_inos);
1034        dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
1035
1036out:
1037        dbg_free_check_tree(&ci.root);
1038        kfree(ci.node);
1039        return err;
1040}
1041