diff options
Diffstat (limited to 'fs/xfs/xfs_alloc.c')
| -rw-r--r-- | fs/xfs/xfs_alloc.c | 347 |
1 files changed, 233 insertions, 114 deletions
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index 4f33c32affe..d43813267a8 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -17,23 +17,25 @@ */ #include "xfs.h" #include "xfs_fs.h" -#include "xfs_types.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_shared.h" +#include "xfs_trans_resv.h" #include "xfs_bit.h" -#include "xfs_log.h" -#include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_ag.h" #include "xfs_mount.h" -#include "xfs_bmap_btree.h" -#include "xfs_alloc_btree.h" -#include "xfs_ialloc_btree.h" -#include "xfs_dinode.h" #include "xfs_inode.h" #include "xfs_btree.h" +#include "xfs_alloc_btree.h" #include "xfs_alloc.h" #include "xfs_extent_busy.h" #include "xfs_error.h" +#include "xfs_cksum.h" #include "xfs_trace.h" +#include "xfs_trans.h" +#include "xfs_buf_item.h" +#include "xfs_log.h" struct workqueue_struct *xfs_alloc_wq; @@ -173,6 +175,7 @@ xfs_alloc_compute_diff( xfs_agblock_t wantbno, /* target starting block */ xfs_extlen_t wantlen, /* target length */ xfs_extlen_t alignment, /* target alignment */ + char userdata, /* are we allocating data? */ xfs_agblock_t freebno, /* freespace's starting block */ xfs_extlen_t freelen, /* freespace's length */ xfs_agblock_t *newbnop) /* result: best start block from free */ @@ -187,7 +190,14 @@ xfs_alloc_compute_diff( ASSERT(freelen >= wantlen); freeend = freebno + freelen; wantend = wantbno + wantlen; - if (freebno >= wantbno) { + /* + * We want to allocate from the start of a free extent if it is past + * the desired block or if we are allocating user data and the free + * extent is before desired block. The second case is there to allow + * for contiguous allocation from the remaining free space if the file + * grows in the short term. + */ + if (freebno >= wantbno || (userdata && freeend < wantend)) { if ((newbno1 = roundup(freebno, alignment)) >= freeend) newbno1 = NULLAGBLOCK; } else if (freeend >= wantend && alignment > 1) { @@ -247,16 +257,14 @@ xfs_alloc_fix_len( k = rlen % args->prod; if (k == args->mod) return; - if (k > args->mod) { - if ((int)(rlen = rlen - k - args->mod) < (int)args->minlen) - return; - } else { - if ((int)(rlen = rlen - args->prod - (args->mod - k)) < - (int)args->minlen) - return; - } - ASSERT(rlen >= args->minlen); - ASSERT(rlen <= args->maxlen); + if (k > args->mod) + rlen = rlen - (k - args->mod); + else + rlen = rlen - args->prod + (args->mod - k); + if ((int)rlen < (int)args->minlen) + return; + ASSERT(rlen >= args->minlen && rlen <= args->maxlen); + ASSERT(rlen % args->prod == args->mod); args->len = rlen; } @@ -430,6 +438,87 @@ xfs_alloc_fixup_trees( return 0; } +static bool +xfs_agfl_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + struct xfs_agfl *agfl = XFS_BUF_TO_AGFL(bp); + int i; + + if (!uuid_equal(&agfl->agfl_uuid, &mp->m_sb.sb_uuid)) + return false; + if (be32_to_cpu(agfl->agfl_magicnum) != XFS_AGFL_MAGIC) + return false; + /* + * during growfs operations, the perag is not fully initialised, + * so we can't use it for any useful checking. growfs ensures we can't + * use it by using uncached buffers that don't have the perag attached + * so we can detect and avoid this problem. + */ + if (bp->b_pag && be32_to_cpu(agfl->agfl_seqno) != bp->b_pag->pag_agno) + return false; + + for (i = 0; i < XFS_AGFL_SIZE(mp); i++) { + if (be32_to_cpu(agfl->agfl_bno[i]) != NULLAGBLOCK && + be32_to_cpu(agfl->agfl_bno[i]) >= mp->m_sb.sb_agblocks) + return false; + } + return true; +} + +static void +xfs_agfl_read_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + + /* + * There is no verification of non-crc AGFLs because mkfs does not + * initialise the AGFL to zero or NULL. Hence the only valid part of the + * AGFL is what the AGF says is active. We can't get to the AGF, so we + * can't verify just those entries are valid. + */ + if (!xfs_sb_version_hascrc(&mp->m_sb)) + return; + + if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (!xfs_agfl_verify(bp)) + xfs_buf_ioerror(bp, EFSCORRUPTED); + + if (bp->b_error) + xfs_verifier_error(bp); +} + +static void +xfs_agfl_write_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + struct xfs_buf_log_item *bip = bp->b_fspriv; + + /* no verification of non-crc AGFLs */ + if (!xfs_sb_version_hascrc(&mp->m_sb)) + return; + + if (!xfs_agfl_verify(bp)) { + xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); + return; + } + + if (bip) + XFS_BUF_TO_AGFL(bp)->agfl_lsn = cpu_to_be64(bip->bli_item.li_lsn); + + xfs_buf_update_cksum(bp, XFS_AGFL_CRC_OFF); +} + +const struct xfs_buf_ops xfs_agfl_buf_ops = { + .verify_read = xfs_agfl_read_verify, + .verify_write = xfs_agfl_write_verify, +}; + /* * Read in the allocation group free block array. */ @@ -447,10 +536,9 @@ xfs_alloc_read_agfl( error = xfs_trans_read_buf( mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGFL_DADDR(mp)), - XFS_FSS_TO_BB(mp, 1), 0, &bp); + XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_agfl_buf_ops); if (error) return error; - ASSERT(!xfs_buf_geterror(bp)); xfs_buf_set_ref(bp, XFS_AGFL_REF); *bpp = bp; return 0; @@ -718,7 +806,8 @@ xfs_alloc_find_best_extent( xfs_alloc_fix_len(args); sdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, *sbnoa, + args->alignment, + args->userdata, *sbnoa, *slena, &new); /* @@ -782,13 +871,13 @@ xfs_alloc_ag_vextent_near( xfs_agblock_t ltnew; /* useful start bno of left side */ xfs_extlen_t rlen; /* length of returned extent */ int forced = 0; -#if defined(DEBUG) && defined(__KERNEL__) +#ifdef DEBUG /* * Randomly don't execute the first algorithm. */ int dofirst; /* set to do first algorithm */ - dofirst = random32() & 1; + dofirst = prandom_u32() & 1; #endif restart: @@ -842,8 +931,8 @@ restart: xfs_extlen_t blen=0; xfs_agblock_t bnew=0; -#if defined(DEBUG) && defined(__KERNEL__) - if (!dofirst) +#ifdef DEBUG + if (dofirst) break; #endif /* @@ -889,7 +978,8 @@ restart: if (args->len < blen) continue; ltdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, ltbnoa, ltlena, <new); + args->alignment, args->userdata, ltbnoa, + ltlena, <new); if (ltnew != NULLAGBLOCK && (args->len > blen || ltdiff < bdiff)) { bdiff = ltdiff; @@ -1041,7 +1131,8 @@ restart: args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen); xfs_alloc_fix_len(args); ltdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, ltbnoa, ltlena, <new); + args->alignment, args->userdata, ltbnoa, + ltlena, <new); error = xfs_alloc_find_best_extent(args, &bno_cur_lt, &bno_cur_gt, @@ -1057,7 +1148,8 @@ restart: args->len = XFS_EXTLEN_MIN(gtlena, args->maxlen); xfs_alloc_fix_len(args); gtdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, gtbnoa, gtlena, >new); + args->alignment, args->userdata, gtbnoa, + gtlena, >new); error = xfs_alloc_find_best_extent(args, &bno_cur_gt, &bno_cur_lt, @@ -1116,7 +1208,7 @@ restart: } rlen = args->len; (void)xfs_alloc_compute_diff(args->agbno, rlen, args->alignment, - ltbnoa, ltlena, <new); + args->userdata, ltbnoa, ltlena, <new); ASSERT(ltnew >= ltbno); ASSERT(ltnew + rlen <= ltbnoa + ltlena); ASSERT(ltnew + rlen <= be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length)); @@ -1866,12 +1958,11 @@ xfs_alloc_fix_freelist( /* * Initialize the args structure. */ + memset(&targs, 0, sizeof(targs)); targs.tp = tp; targs.mp = mp; targs.agbp = agbp; targs.agno = args->agno; - targs.mod = targs.minleft = targs.wasdel = targs.userdata = - targs.minalignslop = 0; targs.alignment = targs.minlen = targs.prod = targs.isfl = 1; targs.type = XFS_ALLOCTYPE_THIS_AG; targs.pag = pag; @@ -1929,18 +2020,18 @@ xfs_alloc_get_freelist( int btreeblk) /* destination is a AGF btree */ { xfs_agf_t *agf; /* a.g. freespace structure */ - xfs_agfl_t *agfl; /* a.g. freelist structure */ xfs_buf_t *agflbp;/* buffer for a.g. freelist structure */ xfs_agblock_t bno; /* block number returned */ + __be32 *agfl_bno; int error; int logflags; - xfs_mount_t *mp; /* mount structure */ + xfs_mount_t *mp = tp->t_mountp; xfs_perag_t *pag; /* per allocation group data */ - agf = XFS_BUF_TO_AGF(agbp); /* * Freelist is empty, give up. */ + agf = XFS_BUF_TO_AGF(agbp); if (!agf->agf_flcount) { *bnop = NULLAGBLOCK; return 0; @@ -1948,15 +2039,17 @@ xfs_alloc_get_freelist( /* * Read the array of free blocks. */ - mp = tp->t_mountp; - if ((error = xfs_alloc_read_agfl(mp, tp, - be32_to_cpu(agf->agf_seqno), &agflbp))) + error = xfs_alloc_read_agfl(mp, tp, be32_to_cpu(agf->agf_seqno), + &agflbp); + if (error) return error; - agfl = XFS_BUF_TO_AGFL(agflbp); + + /* * Get the block number and update the data structures. */ - bno = be32_to_cpu(agfl->agfl_bno[be32_to_cpu(agf->agf_flfirst)]); + agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp); + bno = be32_to_cpu(agfl_bno[be32_to_cpu(agf->agf_flfirst)]); be32_add_cpu(&agf->agf_flfirst, 1); xfs_trans_brelse(tp, agflbp); if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp)) @@ -2005,11 +2098,14 @@ xfs_alloc_log_agf( offsetof(xfs_agf_t, agf_freeblks), offsetof(xfs_agf_t, agf_longest), offsetof(xfs_agf_t, agf_btreeblks), + offsetof(xfs_agf_t, agf_uuid), sizeof(xfs_agf_t) }; trace_xfs_agf(tp->t_mountp, XFS_BUF_TO_AGF(bp), fields, _RET_IP_); + xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGF_BUF); + xfs_btree_offsets(fields, offsets, XFS_AGF_NUM_BITS, &first, &last); xfs_trans_log_buf(tp, bp, (uint)first, (uint)last); } @@ -2046,12 +2142,13 @@ xfs_alloc_put_freelist( int btreeblk) /* block came from a AGF btree */ { xfs_agf_t *agf; /* a.g. freespace structure */ - xfs_agfl_t *agfl; /* a.g. free block array */ __be32 *blockp;/* pointer to array entry */ int error; int logflags; xfs_mount_t *mp; /* mount structure */ xfs_perag_t *pag; /* per allocation group data */ + __be32 *agfl_bno; + int startoff; agf = XFS_BUF_TO_AGF(agbp); mp = tp->t_mountp; @@ -2059,7 +2156,6 @@ xfs_alloc_put_freelist( if (!agflbp && (error = xfs_alloc_read_agfl(mp, tp, be32_to_cpu(agf->agf_seqno), &agflbp))) return error; - agfl = XFS_BUF_TO_AGFL(agflbp); be32_add_cpu(&agf->agf_fllast, 1); if (be32_to_cpu(agf->agf_fllast) == XFS_AGFL_SIZE(mp)) agf->agf_fllast = 0; @@ -2080,16 +2176,101 @@ xfs_alloc_put_freelist( xfs_alloc_log_agf(tp, agbp, logflags); ASSERT(be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp)); - blockp = &agfl->agfl_bno[be32_to_cpu(agf->agf_fllast)]; + + agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp); + blockp = &agfl_bno[be32_to_cpu(agf->agf_fllast)]; *blockp = cpu_to_be32(bno); + startoff = (char *)blockp - (char *)agflbp->b_addr; + xfs_alloc_log_agf(tp, agbp, logflags); - xfs_trans_log_buf(tp, agflbp, - (int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl), - (int)((xfs_caddr_t)blockp - (xfs_caddr_t)agfl + - sizeof(xfs_agblock_t) - 1)); + + xfs_trans_buf_set_type(tp, agflbp, XFS_BLFT_AGFL_BUF); + xfs_trans_log_buf(tp, agflbp, startoff, + startoff + sizeof(xfs_agblock_t) - 1); return 0; } +static bool +xfs_agf_verify( + struct xfs_mount *mp, + struct xfs_buf *bp) + { + struct xfs_agf *agf = XFS_BUF_TO_AGF(bp); + + if (xfs_sb_version_hascrc(&mp->m_sb) && + !uuid_equal(&agf->agf_uuid, &mp->m_sb.sb_uuid)) + return false; + + if (!(agf->agf_magicnum == cpu_to_be32(XFS_AGF_MAGIC) && + XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) && + be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) && + be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) && + be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) && + be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp))) + return false; + + /* + * during growfs operations, the perag is not fully initialised, + * so we can't use it for any useful checking. growfs ensures we can't + * use it by using uncached buffers that don't have the perag attached + * so we can detect and avoid this problem. + */ + if (bp->b_pag && be32_to_cpu(agf->agf_seqno) != bp->b_pag->pag_agno) + return false; + + if (xfs_sb_version_haslazysbcount(&mp->m_sb) && + be32_to_cpu(agf->agf_btreeblks) > be32_to_cpu(agf->agf_length)) + return false; + + return true;; + +} + +static void +xfs_agf_read_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + + if (xfs_sb_version_hascrc(&mp->m_sb) && + !xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF)) + xfs_buf_ioerror(bp, EFSBADCRC); + else if (XFS_TEST_ERROR(!xfs_agf_verify(mp, bp), mp, + XFS_ERRTAG_ALLOC_READ_AGF, + XFS_RANDOM_ALLOC_READ_AGF)) + xfs_buf_ioerror(bp, EFSCORRUPTED); + + if (bp->b_error) + xfs_verifier_error(bp); +} + +static void +xfs_agf_write_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + struct xfs_buf_log_item *bip = bp->b_fspriv; + + if (!xfs_agf_verify(mp, bp)) { + xfs_buf_ioerror(bp, EFSCORRUPTED); + xfs_verifier_error(bp); + return; + } + + if (!xfs_sb_version_hascrc(&mp->m_sb)) + return; + + if (bip) + XFS_BUF_TO_AGF(bp)->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn); + + xfs_buf_update_cksum(bp, XFS_AGF_CRC_OFF); +} + +const struct xfs_buf_ops xfs_agf_buf_ops = { + .verify_read = xfs_agf_read_verify, + .verify_write = xfs_agf_write_verify, +}; + /* * Read in the allocation group header (free/alloc section). */ @@ -2101,44 +2282,21 @@ xfs_read_agf( int flags, /* XFS_BUF_ */ struct xfs_buf **bpp) /* buffer for the ag freelist header */ { - struct xfs_agf *agf; /* ag freelist header */ - int agf_ok; /* set if agf is consistent */ int error; + trace_xfs_read_agf(mp, agno); + ASSERT(agno != NULLAGNUMBER); error = xfs_trans_read_buf( mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGF_DADDR(mp)), - XFS_FSS_TO_BB(mp, 1), flags, bpp); + XFS_FSS_TO_BB(mp, 1), flags, bpp, &xfs_agf_buf_ops); if (error) return error; if (!*bpp) return 0; ASSERT(!(*bpp)->b_error); - agf = XFS_BUF_TO_AGF(*bpp); - - /* - * Validate the magic number of the agf block. - */ - agf_ok = - agf->agf_magicnum == cpu_to_be32(XFS_AGF_MAGIC) && - XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) && - be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) && - be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) && - be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) && - be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp) && - be32_to_cpu(agf->agf_seqno) == agno; - if (xfs_sb_version_haslazysbcount(&mp->m_sb)) - agf_ok = agf_ok && be32_to_cpu(agf->agf_btreeblks) <= - be32_to_cpu(agf->agf_length); - if (unlikely(XFS_TEST_ERROR(!agf_ok, mp, XFS_ERRTAG_ALLOC_READ_AGF, - XFS_RANDOM_ALLOC_READ_AGF))) { - XFS_CORRUPTION_ERROR("xfs_alloc_read_agf", - XFS_ERRLEVEL_LOW, mp, agf); - xfs_trans_brelse(tp, *bpp); - return XFS_ERROR(EFSCORRUPTED); - } xfs_buf_set_ref(*bpp, XFS_AGF_REF); return 0; } @@ -2158,8 +2316,9 @@ xfs_alloc_read_agf( struct xfs_perag *pag; /* per allocation group data */ int error; - ASSERT(agno != NULLAGNUMBER); + trace_xfs_alloc_read_agf(mp, agno); + ASSERT(agno != NULLAGNUMBER); error = xfs_read_agf(mp, tp, agno, (flags & XFS_ALLOC_FLAG_TRYLOCK) ? XBF_TRYLOCK : 0, bpp); @@ -2207,7 +2366,7 @@ xfs_alloc_read_agf( * group or loop over the allocation groups to find the result. */ int /* error */ -__xfs_alloc_vextent( +xfs_alloc_vextent( xfs_alloc_arg_t *args) /* allocation argument structure */ { xfs_agblock_t agsize; /* allocation group size */ @@ -2417,46 +2576,6 @@ error0: return error; } -static void -xfs_alloc_vextent_worker( - struct work_struct *work) -{ - struct xfs_alloc_arg *args = container_of(work, - struct xfs_alloc_arg, work); - unsigned long pflags; - - /* we are in a transaction context here */ - current_set_flags_nested(&pflags, PF_FSTRANS); - - args->result = __xfs_alloc_vextent(args); - complete(args->done); - - current_restore_flags_nested(&pflags, PF_FSTRANS); -} - -/* - * Data allocation requests often come in with little stack to work on. Push - * them off to a worker thread so there is lots of stack to use. Metadata - * requests, OTOH, are generally from low stack usage paths, so avoid the - * context switch overhead here. - */ -int -xfs_alloc_vextent( - struct xfs_alloc_arg *args) -{ - DECLARE_COMPLETION_ONSTACK(done); - - if (!args->userdata) - return __xfs_alloc_vextent(args); - - - args->done = &done; - INIT_WORK_ONSTACK(&args->work, xfs_alloc_vextent_worker); - queue_work(xfs_alloc_wq, &args->work); - wait_for_completion(&done); - return args->result; -} - /* * Free an extent. * Just break up the extent address and hand off to xfs_free_ag_extent |
