linux/fs/xfs/xfs_alloc_btree.c
<<
>>
Prefs
   1/*
   2 * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
   3 * All Rights Reserved.
   4 *
   5 * This program is free software; you can redistribute it and/or
   6 * modify it under the terms of the GNU General Public License as
   7 * published by the Free Software Foundation.
   8 *
   9 * This program is distributed in the hope that it would be useful,
  10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 * GNU General Public License for more details.
  13 *
  14 * You should have received a copy of the GNU General Public License
  15 * along with this program; if not, write the Free Software Foundation,
  16 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
  17 */
  18#include "xfs.h"
  19#include "xfs_fs.h"
  20#include "xfs_types.h"
  21#include "xfs_log.h"
  22#include "xfs_trans.h"
  23#include "xfs_sb.h"
  24#include "xfs_ag.h"
  25#include "xfs_mount.h"
  26#include "xfs_bmap_btree.h"
  27#include "xfs_alloc_btree.h"
  28#include "xfs_ialloc_btree.h"
  29#include "xfs_dinode.h"
  30#include "xfs_inode.h"
  31#include "xfs_btree.h"
  32#include "xfs_alloc.h"
  33#include "xfs_extent_busy.h"
  34#include "xfs_error.h"
  35#include "xfs_trace.h"
  36#include "xfs_cksum.h"
  37
  38
  39STATIC struct xfs_btree_cur *
  40xfs_allocbt_dup_cursor(
  41        struct xfs_btree_cur    *cur)
  42{
  43        return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
  44                        cur->bc_private.a.agbp, cur->bc_private.a.agno,
  45                        cur->bc_btnum);
  46}
  47
  48STATIC void
  49xfs_allocbt_set_root(
  50        struct xfs_btree_cur    *cur,
  51        union xfs_btree_ptr     *ptr,
  52        int                     inc)
  53{
  54        struct xfs_buf          *agbp = cur->bc_private.a.agbp;
  55        struct xfs_agf          *agf = XFS_BUF_TO_AGF(agbp);
  56        xfs_agnumber_t          seqno = be32_to_cpu(agf->agf_seqno);
  57        int                     btnum = cur->bc_btnum;
  58        struct xfs_perag        *pag = xfs_perag_get(cur->bc_mp, seqno);
  59
  60        ASSERT(ptr->s != 0);
  61
  62        agf->agf_roots[btnum] = ptr->s;
  63        be32_add_cpu(&agf->agf_levels[btnum], inc);
  64        pag->pagf_levels[btnum] += inc;
  65        xfs_perag_put(pag);
  66
  67        xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
  68}
  69
  70STATIC int
  71xfs_allocbt_alloc_block(
  72        struct xfs_btree_cur    *cur,
  73        union xfs_btree_ptr     *start,
  74        union xfs_btree_ptr     *new,
  75        int                     length,
  76        int                     *stat)
  77{
  78        int                     error;
  79        xfs_agblock_t           bno;
  80
  81        XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
  82
  83        /* Allocate the new block from the freelist. If we can't, give up.  */
  84        error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
  85                                       &bno, 1);
  86        if (error) {
  87                XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
  88                return error;
  89        }
  90
  91        if (bno == NULLAGBLOCK) {
  92                XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
  93                *stat = 0;
  94                return 0;
  95        }
  96
  97        xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1, false);
  98
  99        xfs_trans_agbtree_delta(cur->bc_tp, 1);
 100        new->s = cpu_to_be32(bno);
 101
 102        XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 103        *stat = 1;
 104        return 0;
 105}
 106
 107STATIC int
 108xfs_allocbt_free_block(
 109        struct xfs_btree_cur    *cur,
 110        struct xfs_buf          *bp)
 111{
 112        struct xfs_buf          *agbp = cur->bc_private.a.agbp;
 113        struct xfs_agf          *agf = XFS_BUF_TO_AGF(agbp);
 114        xfs_agblock_t           bno;
 115        int                     error;
 116
 117        bno = xfs_daddr_to_agbno(cur->bc_mp, XFS_BUF_ADDR(bp));
 118        error = xfs_alloc_put_freelist(cur->bc_tp, agbp, NULL, bno, 1);
 119        if (error)
 120                return error;
 121
 122        xfs_extent_busy_insert(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1,
 123                              XFS_EXTENT_BUSY_SKIP_DISCARD);
 124        xfs_trans_agbtree_delta(cur->bc_tp, -1);
 125
 126        xfs_trans_binval(cur->bc_tp, bp);
 127        return 0;
 128}
 129
 130/*
 131 * Update the longest extent in the AGF
 132 */
 133STATIC void
 134xfs_allocbt_update_lastrec(
 135        struct xfs_btree_cur    *cur,
 136        struct xfs_btree_block  *block,
 137        union xfs_btree_rec     *rec,
 138        int                     ptr,
 139        int                     reason)
 140{
 141        struct xfs_agf          *agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
 142        xfs_agnumber_t          seqno = be32_to_cpu(agf->agf_seqno);
 143        struct xfs_perag        *pag;
 144        __be32                  len;
 145        int                     numrecs;
 146
 147        ASSERT(cur->bc_btnum == XFS_BTNUM_CNT);
 148
 149        switch (reason) {
 150        case LASTREC_UPDATE:
 151                /*
 152                 * If this is the last leaf block and it's the last record,
 153                 * then update the size of the longest extent in the AG.
 154                 */
 155                if (ptr != xfs_btree_get_numrecs(block))
 156                        return;
 157                len = rec->alloc.ar_blockcount;
 158                break;
 159        case LASTREC_INSREC:
 160                if (be32_to_cpu(rec->alloc.ar_blockcount) <=
 161                    be32_to_cpu(agf->agf_longest))
 162                        return;
 163                len = rec->alloc.ar_blockcount;
 164                break;
 165        case LASTREC_DELREC:
 166                numrecs = xfs_btree_get_numrecs(block);
 167                if (ptr <= numrecs)
 168                        return;
 169                ASSERT(ptr == numrecs + 1);
 170
 171                if (numrecs) {
 172                        xfs_alloc_rec_t *rrp;
 173
 174                        rrp = XFS_ALLOC_REC_ADDR(cur->bc_mp, block, numrecs);
 175                        len = rrp->ar_blockcount;
 176                } else {
 177                        len = 0;
 178                }
 179
 180                break;
 181        default:
 182                ASSERT(0);
 183                return;
 184        }
 185
 186        agf->agf_longest = len;
 187        pag = xfs_perag_get(cur->bc_mp, seqno);
 188        pag->pagf_longest = be32_to_cpu(len);
 189        xfs_perag_put(pag);
 190        xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST);
 191}
 192
 193STATIC int
 194xfs_allocbt_get_minrecs(
 195        struct xfs_btree_cur    *cur,
 196        int                     level)
 197{
 198        return cur->bc_mp->m_alloc_mnr[level != 0];
 199}
 200
 201STATIC int
 202xfs_allocbt_get_maxrecs(
 203        struct xfs_btree_cur    *cur,
 204        int                     level)
 205{
 206        return cur->bc_mp->m_alloc_mxr[level != 0];
 207}
 208
 209STATIC void
 210xfs_allocbt_init_key_from_rec(
 211        union xfs_btree_key     *key,
 212        union xfs_btree_rec     *rec)
 213{
 214        ASSERT(rec->alloc.ar_startblock != 0);
 215
 216        key->alloc.ar_startblock = rec->alloc.ar_startblock;
 217        key->alloc.ar_blockcount = rec->alloc.ar_blockcount;
 218}
 219
 220STATIC void
 221xfs_allocbt_init_rec_from_key(
 222        union xfs_btree_key     *key,
 223        union xfs_btree_rec     *rec)
 224{
 225        ASSERT(key->alloc.ar_startblock != 0);
 226
 227        rec->alloc.ar_startblock = key->alloc.ar_startblock;
 228        rec->alloc.ar_blockcount = key->alloc.ar_blockcount;
 229}
 230
 231STATIC void
 232xfs_allocbt_init_rec_from_cur(
 233        struct xfs_btree_cur    *cur,
 234        union xfs_btree_rec     *rec)
 235{
 236        ASSERT(cur->bc_rec.a.ar_startblock != 0);
 237
 238        rec->alloc.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
 239        rec->alloc.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
 240}
 241
 242STATIC void
 243xfs_allocbt_init_ptr_from_cur(
 244        struct xfs_btree_cur    *cur,
 245        union xfs_btree_ptr     *ptr)
 246{
 247        struct xfs_agf          *agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
 248
 249        ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
 250        ASSERT(agf->agf_roots[cur->bc_btnum] != 0);
 251
 252        ptr->s = agf->agf_roots[cur->bc_btnum];
 253}
 254
 255STATIC __int64_t
 256xfs_allocbt_key_diff(
 257        struct xfs_btree_cur    *cur,
 258        union xfs_btree_key     *key)
 259{
 260        xfs_alloc_rec_incore_t  *rec = &cur->bc_rec.a;
 261        xfs_alloc_key_t         *kp = &key->alloc;
 262        __int64_t               diff;
 263
 264        if (cur->bc_btnum == XFS_BTNUM_BNO) {
 265                return (__int64_t)be32_to_cpu(kp->ar_startblock) -
 266                                rec->ar_startblock;
 267        }
 268
 269        diff = (__int64_t)be32_to_cpu(kp->ar_blockcount) - rec->ar_blockcount;
 270        if (diff)
 271                return diff;
 272
 273        return (__int64_t)be32_to_cpu(kp->ar_startblock) - rec->ar_startblock;
 274}
 275
 276static bool
 277xfs_allocbt_verify(
 278        struct xfs_buf          *bp)
 279{
 280        struct xfs_mount        *mp = bp->b_target->bt_mount;
 281        struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);
 282        struct xfs_perag        *pag = bp->b_pag;
 283        unsigned int            level;
 284
 285        /*
 286         * magic number and level verification
 287         *
 288         * During growfs operations, we can't verify the exact level or owner as
 289         * the perag is not fully initialised and hence not attached to the
 290         * buffer.  In this case, check against the maximum tree depth.
 291         *
 292         * Similarly, during log recovery we will have a perag structure
 293         * attached, but the agf information will not yet have been initialised
 294         * from the on disk AGF. Again, we can only check against maximum limits
 295         * in this case.
 296         */
 297        level = be16_to_cpu(block->bb_level);
 298        switch (block->bb_magic) {
 299        case cpu_to_be32(XFS_ABTB_CRC_MAGIC):
 300                if (!xfs_sb_version_hascrc(&mp->m_sb))
 301                        return false;
 302                if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
 303                        return false;
 304                if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
 305                        return false;
 306                if (pag &&
 307                    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
 308                        return false;
 309                /* fall through */
 310        case cpu_to_be32(XFS_ABTB_MAGIC):
 311                if (pag && pag->pagf_init) {
 312                        if (level >= pag->pagf_levels[XFS_BTNUM_BNOi])
 313                                return false;
 314                } else if (level >= mp->m_ag_maxlevels)
 315                        return false;
 316                break;
 317        case cpu_to_be32(XFS_ABTC_CRC_MAGIC):
 318                if (!xfs_sb_version_hascrc(&mp->m_sb))
 319                        return false;
 320                if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
 321                        return false;
 322                if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
 323                        return false;
 324                if (pag &&
 325                    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
 326                        return false;
 327                /* fall through */
 328        case cpu_to_be32(XFS_ABTC_MAGIC):
 329                if (pag && pag->pagf_init) {
 330                        if (level >= pag->pagf_levels[XFS_BTNUM_CNTi])
 331                                return false;
 332                } else if (level >= mp->m_ag_maxlevels)
 333                        return false;
 334                break;
 335        default:
 336                return false;
 337        }
 338
 339        /* numrecs verification */
 340        if (be16_to_cpu(block->bb_numrecs) > mp->m_alloc_mxr[level != 0])
 341                return false;
 342
 343        /* sibling pointer verification */
 344        if (!block->bb_u.s.bb_leftsib ||
 345            (be32_to_cpu(block->bb_u.s.bb_leftsib) >= mp->m_sb.sb_agblocks &&
 346             block->bb_u.s.bb_leftsib != cpu_to_be32(NULLAGBLOCK)))
 347                return false;
 348        if (!block->bb_u.s.bb_rightsib ||
 349            (be32_to_cpu(block->bb_u.s.bb_rightsib) >= mp->m_sb.sb_agblocks &&
 350             block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK)))
 351                return false;
 352
 353        return true;
 354}
 355
 356static void
 357xfs_allocbt_read_verify(
 358        struct xfs_buf  *bp)
 359{
 360        if (!(xfs_btree_sblock_verify_crc(bp) &&
 361              xfs_allocbt_verify(bp))) {
 362                trace_xfs_btree_corrupt(bp, _RET_IP_);
 363                XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW,
 364                                     bp->b_target->bt_mount, bp->b_addr);
 365                xfs_buf_ioerror(bp, EFSCORRUPTED);
 366        }
 367}
 368
 369static void
 370xfs_allocbt_write_verify(
 371        struct xfs_buf  *bp)
 372{
 373        if (!xfs_allocbt_verify(bp)) {
 374                trace_xfs_btree_corrupt(bp, _RET_IP_);
 375                XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW,
 376                                     bp->b_target->bt_mount, bp->b_addr);
 377                xfs_buf_ioerror(bp, EFSCORRUPTED);
 378        }
 379        xfs_btree_sblock_calc_crc(bp);
 380
 381}
 382
 383const struct xfs_buf_ops xfs_allocbt_buf_ops = {
 384        .verify_read = xfs_allocbt_read_verify,
 385        .verify_write = xfs_allocbt_write_verify,
 386};
 387
 388
 389#if defined(DEBUG) || defined(XFS_WARN)
 390STATIC int
 391xfs_allocbt_keys_inorder(
 392        struct xfs_btree_cur    *cur,
 393        union xfs_btree_key     *k1,
 394        union xfs_btree_key     *k2)
 395{
 396        if (cur->bc_btnum == XFS_BTNUM_BNO) {
 397                return be32_to_cpu(k1->alloc.ar_startblock) <
 398                       be32_to_cpu(k2->alloc.ar_startblock);
 399        } else {
 400                return be32_to_cpu(k1->alloc.ar_blockcount) <
 401                        be32_to_cpu(k2->alloc.ar_blockcount) ||
 402                        (k1->alloc.ar_blockcount == k2->alloc.ar_blockcount &&
 403                         be32_to_cpu(k1->alloc.ar_startblock) <
 404                         be32_to_cpu(k2->alloc.ar_startblock));
 405        }
 406}
 407
 408STATIC int
 409xfs_allocbt_recs_inorder(
 410        struct xfs_btree_cur    *cur,
 411        union xfs_btree_rec     *r1,
 412        union xfs_btree_rec     *r2)
 413{
 414        if (cur->bc_btnum == XFS_BTNUM_BNO) {
 415                return be32_to_cpu(r1->alloc.ar_startblock) +
 416                        be32_to_cpu(r1->alloc.ar_blockcount) <=
 417                        be32_to_cpu(r2->alloc.ar_startblock);
 418        } else {
 419                return be32_to_cpu(r1->alloc.ar_blockcount) <
 420                        be32_to_cpu(r2->alloc.ar_blockcount) ||
 421                        (r1->alloc.ar_blockcount == r2->alloc.ar_blockcount &&
 422                         be32_to_cpu(r1->alloc.ar_startblock) <
 423                         be32_to_cpu(r2->alloc.ar_startblock));
 424        }
 425}
 426#endif  /* DEBUG */
 427
 428static const struct xfs_btree_ops xfs_allocbt_ops = {
 429        .rec_len                = sizeof(xfs_alloc_rec_t),
 430        .key_len                = sizeof(xfs_alloc_key_t),
 431
 432        .dup_cursor             = xfs_allocbt_dup_cursor,
 433        .set_root               = xfs_allocbt_set_root,
 434        .alloc_block            = xfs_allocbt_alloc_block,
 435        .free_block             = xfs_allocbt_free_block,
 436        .update_lastrec         = xfs_allocbt_update_lastrec,
 437        .get_minrecs            = xfs_allocbt_get_minrecs,
 438        .get_maxrecs            = xfs_allocbt_get_maxrecs,
 439        .init_key_from_rec      = xfs_allocbt_init_key_from_rec,
 440        .init_rec_from_key      = xfs_allocbt_init_rec_from_key,
 441        .init_rec_from_cur      = xfs_allocbt_init_rec_from_cur,
 442        .init_ptr_from_cur      = xfs_allocbt_init_ptr_from_cur,
 443        .key_diff               = xfs_allocbt_key_diff,
 444        .buf_ops                = &xfs_allocbt_buf_ops,
 445#if defined(DEBUG) || defined(XFS_WARN)
 446        .keys_inorder           = xfs_allocbt_keys_inorder,
 447        .recs_inorder           = xfs_allocbt_recs_inorder,
 448#endif
 449};
 450
 451/*
 452 * Allocate a new allocation btree cursor.
 453 */
 454struct xfs_btree_cur *                  /* new alloc btree cursor */
 455xfs_allocbt_init_cursor(
 456        struct xfs_mount        *mp,            /* file system mount point */
 457        struct xfs_trans        *tp,            /* transaction pointer */
 458        struct xfs_buf          *agbp,          /* buffer for agf structure */
 459        xfs_agnumber_t          agno,           /* allocation group number */
 460        xfs_btnum_t             btnum)          /* btree identifier */
 461{
 462        struct xfs_agf          *agf = XFS_BUF_TO_AGF(agbp);
 463        struct xfs_btree_cur    *cur;
 464
 465        ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
 466
 467        cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
 468
 469        cur->bc_tp = tp;
 470        cur->bc_mp = mp;
 471        cur->bc_btnum = btnum;
 472        cur->bc_blocklog = mp->m_sb.sb_blocklog;
 473        cur->bc_ops = &xfs_allocbt_ops;
 474
 475        if (btnum == XFS_BTNUM_CNT) {
 476                cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
 477                cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
 478        } else {
 479                cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
 480        }
 481
 482        cur->bc_private.a.agbp = agbp;
 483        cur->bc_private.a.agno = agno;
 484
 485        if (xfs_sb_version_hascrc(&mp->m_sb))
 486                cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 487
 488        return cur;
 489}
 490
 491/*
 492 * Calculate number of records in an alloc btree block.
 493 */
 494int
 495xfs_allocbt_maxrecs(
 496        struct xfs_mount        *mp,
 497        int                     blocklen,
 498        int                     leaf)
 499{
 500        blocklen -= XFS_ALLOC_BLOCK_LEN(mp);
 501
 502        if (leaf)
 503                return blocklen / sizeof(xfs_alloc_rec_t);
 504        return blocklen / (sizeof(xfs_alloc_key_t) + sizeof(xfs_alloc_ptr_t));
 505}
 506