diff options
Diffstat (limited to 'fs/xfs')
111 files changed, 12309 insertions, 2192 deletions
diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index fc593c869493..26ef1958b65b 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -52,8 +52,11 @@ xfs-y += $(addprefix libxfs/, \ xfs_inode_fork.o \ xfs_inode_buf.o \ xfs_log_rlimit.o \ + xfs_ag_resv.o \ xfs_rmap.o \ xfs_rmap_btree.o \ + xfs_refcount.o \ + xfs_refcount_btree.o \ xfs_sb.o \ xfs_symlink_remote.o \ xfs_trans_resv.o \ @@ -87,6 +90,7 @@ xfs-y += xfs_aops.o \ xfs_message.o \ xfs_mount.o \ xfs_mru_cache.o \ + xfs_reflink.o \ xfs_stats.o \ xfs_super.o \ xfs_symlink.o \ @@ -99,16 +103,20 @@ xfs-y += xfs_aops.o \ # low-level transaction/log code xfs-y += xfs_log.o \ xfs_log_cil.o \ + xfs_bmap_item.o \ xfs_buf_item.o \ xfs_extfree_item.o \ xfs_icreate_item.o \ xfs_inode_item.o \ + xfs_refcount_item.o \ xfs_rmap_item.o \ xfs_log_recover.o \ xfs_trans_ail.o \ + xfs_trans_bmap.o \ xfs_trans_buf.o \ xfs_trans_extfree.o \ xfs_trans_inode.o \ + xfs_trans_refcount.o \ xfs_trans_rmap.o \ # optional features diff --git a/fs/xfs/libxfs/xfs_ag_resv.c b/fs/xfs/libxfs/xfs_ag_resv.c new file mode 100644 index 000000000000..d346d42c54d1 --- /dev/null +++ b/fs/xfs/libxfs/xfs_ag_resv.c @@ -0,0 +1,341 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_alloc.h" +#include "xfs_error.h" +#include "xfs_trace.h" +#include "xfs_cksum.h" +#include "xfs_trans.h" +#include "xfs_bit.h" +#include "xfs_bmap.h" +#include "xfs_bmap_btree.h" +#include "xfs_ag_resv.h" +#include "xfs_trans_space.h" +#include "xfs_rmap_btree.h" +#include "xfs_btree.h" +#include "xfs_refcount_btree.h" + +/* + * Per-AG Block Reservations + * + * For some kinds of allocation group metadata structures, it is advantageous + * to reserve a small number of blocks in each AG so that future expansions of + * that data structure do not encounter ENOSPC because errors during a btree + * split cause the filesystem to go offline. + * + * Prior to the introduction of reflink, this wasn't an issue because the free + * space btrees maintain a reserve of space (the AGFL) to handle any expansion + * that may be necessary; and allocations of other metadata (inodes, BMBT, + * dir/attr) aren't restricted to a single AG. However, with reflink it is + * possible to allocate all the space in an AG, have subsequent reflink/CoW + * activity expand the refcount btree, and discover that there's no space left + * to handle that expansion. Since we can calculate the maximum size of the + * refcount btree, we can reserve space for it and avoid ENOSPC. + * + * Handling per-AG reservations consists of three changes to the allocator's + * behavior: First, because these reservations are always needed, we decrease + * the ag_max_usable counter to reflect the size of the AG after the reserved + * blocks are taken. Second, the reservations must be reflected in the + * fdblocks count to maintain proper accounting. Third, each AG must maintain + * its own reserved block counter so that we can calculate the amount of space + * that must remain free to maintain the reservations. Fourth, the "remaining + * reserved blocks" count must be used when calculating the length of the + * longest free extent in an AG and to clamp maxlen in the per-AG allocation + * functions. In other words, we maintain a virtual allocation via in-core + * accounting tricks so that we don't have to clean up after a crash. :) + * + * Reserved blocks can be managed by passing one of the enum xfs_ag_resv_type + * values via struct xfs_alloc_arg or directly to the xfs_free_extent + * function. It might seem a little funny to maintain a reservoir of blocks + * to feed another reservoir, but the AGFL only holds enough blocks to get + * through the next transaction. The per-AG reservation is to ensure (we + * hope) that each AG never runs out of blocks. Each data structure wanting + * to use the reservation system should update ask/used in xfs_ag_resv_init. + */ + +/* + * Are we critically low on blocks? For now we'll define that as the number + * of blocks we can get our hands on being less than 10% of what we reserved + * or less than some arbitrary number (maximum btree height). + */ +bool +xfs_ag_resv_critical( + struct xfs_perag *pag, + enum xfs_ag_resv_type type) +{ + xfs_extlen_t avail; + xfs_extlen_t orig; + + switch (type) { + case XFS_AG_RESV_METADATA: + avail = pag->pagf_freeblks - pag->pag_agfl_resv.ar_reserved; + orig = pag->pag_meta_resv.ar_asked; + break; + case XFS_AG_RESV_AGFL: + avail = pag->pagf_freeblks + pag->pagf_flcount - + pag->pag_meta_resv.ar_reserved; + orig = pag->pag_agfl_resv.ar_asked; + break; + default: + ASSERT(0); + return false; + } + + trace_xfs_ag_resv_critical(pag, type, avail); + + /* Critically low if less than 10% or max btree height remains. */ + return XFS_TEST_ERROR(avail < orig / 10 || avail < XFS_BTREE_MAXLEVELS, + pag->pag_mount, XFS_ERRTAG_AG_RESV_CRITICAL, + XFS_RANDOM_AG_RESV_CRITICAL); +} + +/* + * How many blocks are reserved but not used, and therefore must not be + * allocated away? + */ +xfs_extlen_t +xfs_ag_resv_needed( + struct xfs_perag *pag, + enum xfs_ag_resv_type type) +{ + xfs_extlen_t len; + + len = pag->pag_meta_resv.ar_reserved + pag->pag_agfl_resv.ar_reserved; + switch (type) { + case XFS_AG_RESV_METADATA: + case XFS_AG_RESV_AGFL: + len -= xfs_perag_resv(pag, type)->ar_reserved; + break; + case XFS_AG_RESV_NONE: + /* empty */ + break; + default: + ASSERT(0); + } + + trace_xfs_ag_resv_needed(pag, type, len); + + return len; +} + +/* Clean out a reservation */ +static int +__xfs_ag_resv_free( + struct xfs_perag *pag, + enum xfs_ag_resv_type type) +{ + struct xfs_ag_resv *resv; + xfs_extlen_t oldresv; + int error; + + trace_xfs_ag_resv_free(pag, type, 0); + + resv = xfs_perag_resv(pag, type); + pag->pag_mount->m_ag_max_usable += resv->ar_asked; + /* + * AGFL blocks are always considered "free", so whatever + * was reserved at mount time must be given back at umount. + */ + if (type == XFS_AG_RESV_AGFL) + oldresv = resv->ar_orig_reserved; + else + oldresv = resv->ar_reserved; + error = xfs_mod_fdblocks(pag->pag_mount, oldresv, true); + resv->ar_reserved = 0; + resv->ar_asked = 0; + + if (error) + trace_xfs_ag_resv_free_error(pag->pag_mount, pag->pag_agno, + error, _RET_IP_); + return error; +} + +/* Free a per-AG reservation. */ +int +xfs_ag_resv_free( + struct xfs_perag *pag) +{ + int error; + int err2; + + error = __xfs_ag_resv_free(pag, XFS_AG_RESV_AGFL); + err2 = __xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA); + if (err2 && !error) + error = err2; + return error; +} + +static int +__xfs_ag_resv_init( + struct xfs_perag *pag, + enum xfs_ag_resv_type type, + xfs_extlen_t ask, + xfs_extlen_t used) +{ + struct xfs_mount *mp = pag->pag_mount; + struct xfs_ag_resv *resv; + int error; + + resv = xfs_perag_resv(pag, type); + if (used > ask) + ask = used; + resv->ar_asked = ask; + resv->ar_reserved = resv->ar_orig_reserved = ask - used; + mp->m_ag_max_usable -= ask; + + trace_xfs_ag_resv_init(pag, type, ask); + + error = xfs_mod_fdblocks(mp, -(int64_t)resv->ar_reserved, true); + if (error) + trace_xfs_ag_resv_init_error(pag->pag_mount, pag->pag_agno, + error, _RET_IP_); + + return error; +} + +/* Create a per-AG block reservation. */ +int +xfs_ag_resv_init( + struct xfs_perag *pag) +{ + xfs_extlen_t ask; + xfs_extlen_t used; + int error = 0; + + /* Create the metadata reservation. */ + if (pag->pag_meta_resv.ar_asked == 0) { + ask = used = 0; + + error = xfs_refcountbt_calc_reserves(pag->pag_mount, + pag->pag_agno, &ask, &used); + if (error) + goto out; + + error = __xfs_ag_resv_init(pag, XFS_AG_RESV_METADATA, + ask, used); + if (error) + goto out; + } + + /* Create the AGFL metadata reservation */ + if (pag->pag_agfl_resv.ar_asked == 0) { + ask = used = 0; + + error = xfs_rmapbt_calc_reserves(pag->pag_mount, pag->pag_agno, + &ask, &used); + if (error) + goto out; + + error = __xfs_ag_resv_init(pag, XFS_AG_RESV_AGFL, ask, used); + if (error) + goto out; + } + + ASSERT(xfs_perag_resv(pag, XFS_AG_RESV_METADATA)->ar_reserved + + xfs_perag_resv(pag, XFS_AG_RESV_AGFL)->ar_reserved <= + pag->pagf_freeblks + pag->pagf_flcount); +out: + return error; +} + +/* Allocate a block from the reservation. */ +void +xfs_ag_resv_alloc_extent( + struct xfs_perag *pag, + enum xfs_ag_resv_type type, + struct xfs_alloc_arg *args) +{ + struct xfs_ag_resv *resv; + xfs_extlen_t len; + uint field; + + trace_xfs_ag_resv_alloc_extent(pag, type, args->len); + + switch (type) { + case XFS_AG_RESV_METADATA: + case XFS_AG_RESV_AGFL: + resv = xfs_perag_resv(pag, type); + break; + default: + ASSERT(0); + /* fall through */ + case XFS_AG_RESV_NONE: + field = args->wasdel ? XFS_TRANS_SB_RES_FDBLOCKS : + XFS_TRANS_SB_FDBLOCKS; + xfs_trans_mod_sb(args->tp, field, -(int64_t)args->len); + return; + } + + len = min_t(xfs_extlen_t, args->len, resv->ar_reserved); + resv->ar_reserved -= len; + if (type == XFS_AG_RESV_AGFL) + return; + /* Allocations of reserved blocks only need on-disk sb updates... */ + xfs_trans_mod_sb(args->tp, XFS_TRANS_SB_RES_FDBLOCKS, -(int64_t)len); + /* ...but non-reserved blocks need in-core and on-disk updates. */ + if (args->len > len) + xfs_trans_mod_sb(args->tp, XFS_TRANS_SB_FDBLOCKS, + -((int64_t)args->len - len)); +} + +/* Free a block to the reservation. */ +void +xfs_ag_resv_free_extent( + struct xfs_perag *pag, + enum xfs_ag_resv_type type, + struct xfs_trans *tp, + xfs_extlen_t len) +{ + xfs_extlen_t leftover; + struct xfs_ag_resv *resv; + + trace_xfs_ag_resv_free_extent(pag, type, len); + + switch (type) { + case XFS_AG_RESV_METADATA: + case XFS_AG_RESV_AGFL: + resv = xfs_perag_resv(pag, type); + break; + default: + ASSERT(0); + /* fall through */ + case XFS_AG_RESV_NONE: + xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, (int64_t)len); + return; + } + + leftover = min_t(xfs_extlen_t, len, resv->ar_asked - resv->ar_reserved); + resv->ar_reserved += leftover; + if (type == XFS_AG_RESV_AGFL) + return; + /* Freeing into the reserved pool only requires on-disk update... */ + xfs_trans_mod_sb(tp, XFS_TRANS_SB_RES_FDBLOCKS, len); + /* ...but freeing beyond that requires in-core and on-disk update. */ + if (len > leftover) + xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, len - leftover); +} diff --git a/fs/xfs/libxfs/xfs_ag_resv.h b/fs/xfs/libxfs/xfs_ag_resv.h new file mode 100644 index 000000000000..8d6c687deef3 --- /dev/null +++ b/fs/xfs/libxfs/xfs_ag_resv.h @@ -0,0 +1,35 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_AG_RESV_H__ +#define __XFS_AG_RESV_H__ + +int xfs_ag_resv_free(struct xfs_perag *pag); +int xfs_ag_resv_init(struct xfs_perag *pag); + +bool xfs_ag_resv_critical(struct xfs_perag *pag, enum xfs_ag_resv_type type); +xfs_extlen_t xfs_ag_resv_needed(struct xfs_perag *pag, + enum xfs_ag_resv_type type); + +void xfs_ag_resv_alloc_extent(struct xfs_perag *pag, enum xfs_ag_resv_type type, + struct xfs_alloc_arg *args); +void xfs_ag_resv_free_extent(struct xfs_perag *pag, enum xfs_ag_resv_type type, + struct xfs_trans *tp, xfs_extlen_t len); + +#endif /* __XFS_AG_RESV_H__ */ diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 05b5243d89f6..5050056a0b06 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -37,6 +37,7 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" +#include "xfs_ag_resv.h" struct workqueue_struct *xfs_alloc_wq; @@ -51,10 +52,23 @@ STATIC int xfs_alloc_ag_vextent_size(xfs_alloc_arg_t *); STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *, xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *); +unsigned int +xfs_refc_block( + struct xfs_mount *mp) +{ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + return XFS_RMAP_BLOCK(mp) + 1; + if (xfs_sb_version_hasfinobt(&mp->m_sb)) + return XFS_FIBT_BLOCK(mp) + 1; + return XFS_IBT_BLOCK(mp) + 1; +} + xfs_extlen_t xfs_prealloc_blocks( struct xfs_mount *mp) { + if (xfs_sb_version_hasreflink(&mp->m_sb)) + return xfs_refc_block(mp) + 1; if (xfs_sb_version_hasrmapbt(&mp->m_sb)) return XFS_RMAP_BLOCK(mp) + 1; if (xfs_sb_version_hasfinobt(&mp->m_sb)) @@ -74,14 +88,8 @@ xfs_prealloc_blocks( * extents need to be actually allocated. To get around this, we explicitly set * aside a few blocks which will not be reserved in delayed allocation. * - * When rmap is disabled, we need to reserve 4 fsbs _per AG_ for the freelist - * and 4 more to handle a potential split of the file's bmap btree. - * - * When rmap is enabled, we must also be able to handle two rmap btree inserts - * to record both the file data extent and a new bmbt block. The bmbt block - * might not be in the same AG as the file data extent. In the worst case - * the bmap btree splits multiple levels and all the new blocks come from - * different AGs, so set aside enough to handle rmap btree splits in all AGs. + * We need to reserve 4 fsbs _per AG_ for the freelist and 4 more to handle a + * potential split of the file's bmap btree. */ unsigned int xfs_alloc_set_aside( @@ -90,8 +98,6 @@ xfs_alloc_set_aside( unsigned int blocks; blocks = 4 + (mp->m_sb.sb_agcount * XFS_ALLOC_AGFL_RESERVE); - if (xfs_sb_version_hasrmapbt(&mp->m_sb)) - blocks += mp->m_sb.sb_agcount * mp->m_rmap_maxlevels; return blocks; } @@ -122,6 +128,8 @@ xfs_alloc_ag_max_usable( blocks++; /* finobt root block */ if (xfs_sb_version_hasrmapbt(&mp->m_sb)) blocks++; /* rmap root block */ + if (xfs_sb_version_hasreflink(&mp->m_sb)) + blocks++; /* refcount root block */ return mp->m_sb.sb_agblocks - blocks; } @@ -265,7 +273,7 @@ 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? */ + int datatype, /* 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 */ @@ -276,6 +284,7 @@ xfs_alloc_compute_diff( xfs_extlen_t newlen1=0; /* length with newbno1 */ xfs_extlen_t newlen2=0; /* length with newbno2 */ xfs_agblock_t wantend; /* end of target extent */ + bool userdata = xfs_alloc_is_userdata(datatype); ASSERT(freelen >= wantlen); freeend = freebno + freelen; @@ -680,12 +689,29 @@ xfs_alloc_ag_vextent( xfs_alloc_arg_t *args) /* argument structure for allocation */ { int error=0; + xfs_extlen_t reservation; + xfs_extlen_t oldmax; ASSERT(args->minlen > 0); ASSERT(args->maxlen > 0); ASSERT(args->minlen <= args->maxlen); ASSERT(args->mod < args->prod); ASSERT(args->alignment > 0); + + /* + * Clamp maxlen to the amount of free space minus any reservations + * that have been made. + */ + oldmax = args->maxlen; + reservation = xfs_ag_resv_needed(args->pag, args->resv); + if (args->maxlen > args->pag->pagf_freeblks - reservation) + args->maxlen = args->pag->pagf_freeblks - reservation; + if (args->maxlen == 0) { + args->agbno = NULLAGBLOCK; + args->maxlen = oldmax; + return 0; + } + /* * Branch to correct routine based on the type. */ @@ -705,12 +731,14 @@ xfs_alloc_ag_vextent( /* NOTREACHED */ } + args->maxlen = oldmax; + if (error || args->agbno == NULLAGBLOCK) return error; ASSERT(args->len >= args->minlen); ASSERT(args->len <= args->maxlen); - ASSERT(!args->wasfromfl || !args->isfl); + ASSERT(!args->wasfromfl || args->resv != XFS_AG_RESV_AGFL); ASSERT(args->agbno % args->alignment == 0); /* if not file data, insert new block into the reverse map btree */ @@ -732,12 +760,7 @@ xfs_alloc_ag_vextent( args->agbno, args->len)); } - if (!args->isfl) { - xfs_trans_mod_sb(args->tp, args->wasdel ? - XFS_TRANS_SB_RES_FDBLOCKS : - XFS_TRANS_SB_FDBLOCKS, - -((long)(args->len))); - } + xfs_ag_resv_alloc_extent(args->pag, args->resv, args); XFS_STATS_INC(args->mp, xs_allocx); XFS_STATS_ADD(args->mp, xs_allocb, args->len); @@ -917,7 +940,7 @@ xfs_alloc_find_best_extent( sdiff = xfs_alloc_compute_diff(args->agbno, args->len, args->alignment, - args->userdata, *sbnoa, + args->datatype, *sbnoa, *slena, &new); /* @@ -1101,7 +1124,7 @@ restart: if (args->len < blen) continue; ltdiff = xfs_alloc_compute_diff(args->agbno, args->len, - args->alignment, args->userdata, ltbnoa, + args->alignment, args->datatype, ltbnoa, ltlena, <new); if (ltnew != NULLAGBLOCK && (args->len > blen || ltdiff < bdiff)) { @@ -1254,7 +1277,7 @@ 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, args->userdata, ltbnoa, + args->alignment, args->datatype, ltbnoa, ltlena, <new); error = xfs_alloc_find_best_extent(args, @@ -1271,7 +1294,7 @@ 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, args->userdata, gtbnoa, + args->alignment, args->datatype, gtbnoa, gtlena, >new); error = xfs_alloc_find_best_extent(args, @@ -1331,7 +1354,7 @@ restart: } rlen = args->len; (void)xfs_alloc_compute_diff(args->agbno, rlen, args->alignment, - args->userdata, ltbnoa, ltlena, <new); + args->datatype, 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)); @@ -1583,6 +1606,7 @@ xfs_alloc_ag_vextent_small( int *stat) /* status: 0-freelist, 1-normal/none */ { struct xfs_owner_info oinfo; + struct xfs_perag *pag; int error; xfs_agblock_t fbno; xfs_extlen_t flen; @@ -1600,7 +1624,8 @@ xfs_alloc_ag_vextent_small( * to respect minleft even when pulling from the * freelist. */ - else if (args->minlen == 1 && args->alignment == 1 && !args->isfl && + else if (args->minlen == 1 && args->alignment == 1 && + args->resv != XFS_AG_RESV_AGFL && (be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_flcount) > args->minleft)) { error = xfs_alloc_get_freelist(args->tp, args->agbp, &fbno, 0); @@ -1608,9 +1633,9 @@ xfs_alloc_ag_vextent_small( goto error0; if (fbno != NULLAGBLOCK) { xfs_extent_busy_reuse(args->mp, args->agno, fbno, 1, - args->userdata); + xfs_alloc_allow_busy_reuse(args->datatype)); - if (args->userdata) { + if (xfs_alloc_is_userdata(args->datatype)) { xfs_buf_t *bp; bp = xfs_btree_get_bufs(args->mp, args->tp, @@ -1629,13 +1654,18 @@ xfs_alloc_ag_vextent_small( /* * If we're feeding an AGFL block to something that * doesn't live in the free space, we need to clear - * out the OWN_AG rmap. + * out the OWN_AG rmap and add the block back to + * the AGFL per-AG reservation. */ xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_AG); error = xfs_rmap_free(args->tp, args->agbp, args->agno, fbno, 1, &oinfo); if (error) goto error0; + pag = xfs_perag_get(args->mp, args->agno); + xfs_ag_resv_free_extent(pag, XFS_AG_RESV_AGFL, + args->tp, 1); + xfs_perag_put(pag); *stat = 0; return 0; @@ -1683,7 +1713,7 @@ xfs_free_ag_extent( xfs_agblock_t bno, xfs_extlen_t len, struct xfs_owner_info *oinfo, - int isfl) + enum xfs_ag_resv_type type) { xfs_btree_cur_t *bno_cur; /* cursor for by-block btree */ xfs_btree_cur_t *cnt_cur; /* cursor for by-size btree */ @@ -1911,21 +1941,22 @@ xfs_free_ag_extent( */ pag = xfs_perag_get(mp, agno); error = xfs_alloc_update_counters(tp, pag, agbp, len); + xfs_ag_resv_free_extent(pag, type, tp, len); xfs_perag_put(pag); if (error) goto error0; - if (!isfl) - xfs_trans_mod_sb(tp, XFS_TRANS_SB_FDBLOCKS, (long)len); XFS_STATS_INC(mp, xs_freex); XFS_STATS_ADD(mp, xs_freeb, len); - trace_xfs_free_extent(mp, agno, bno, len, isfl, haveleft, haveright); + trace_xfs_free_extent(mp, agno, bno, len, type == XFS_AG_RESV_AGFL, + haveleft, haveright); return 0; error0: - trace_xfs_free_extent(mp, agno, bno, len, isfl, -1, -1); + trace_xfs_free_extent(mp, agno, bno, len, type == XFS_AG_RESV_AGFL, + -1, -1); if (bno_cur) xfs_btree_del_cursor(bno_cur, XFS_BTREE_ERROR); if (cnt_cur) @@ -1950,21 +1981,43 @@ xfs_alloc_compute_maxlevels( } /* - * Find the length of the longest extent in an AG. + * Find the length of the longest extent in an AG. The 'need' parameter + * specifies how much space we're going to need for the AGFL and the + * 'reserved' parameter tells us how many blocks in this AG are reserved for + * other callers. */ xfs_extlen_t xfs_alloc_longest_free_extent( struct xfs_mount *mp, struct xfs_perag *pag, - xfs_extlen_t need) + xfs_extlen_t need, + xfs_extlen_t reserved) { xfs_extlen_t delta = 0; + /* + * If the AGFL needs a recharge, we'll have to subtract that from the + * longest extent. + */ if (need > pag->pagf_flcount) delta = need - pag->pagf_flcount; + /* + * If we cannot maintain others' reservations with space from the + * not-longest freesp extents, we'll have to subtract /that/ from + * the longest extent too. + */ + if (pag->pagf_freeblks - pag->pagf_longest < reserved) + delta += reserved - (pag->pagf_freeblks - pag->pagf_longest); + + /* + * If the longest extent is long enough to satisfy all the + * reservations and AGFL rules in place, we can return this extent. + */ if (pag->pagf_longest > delta) return pag->pagf_longest - delta; + + /* Otherwise, let the caller try for 1 block if there's space. */ return pag->pagf_flcount > 0 || pag->pagf_longest > 0; } @@ -2004,20 +2057,24 @@ xfs_alloc_space_available( { struct xfs_perag *pag = args->pag; xfs_extlen_t longest; + xfs_extlen_t reservation; /* blocks that are still reserved */ int available; if (flags & XFS_ALLOC_FLAG_FREEING) return true; + reservation = xfs_ag_resv_needed(pag, args->resv); + /* do we have enough contiguous free space for the allocation? */ - longest = xfs_alloc_longest_free_extent(args->mp, pag, min_free); + longest = xfs_alloc_longest_free_extent(args->mp, pag, min_free, + reservation); if ((args->minlen + args->alignment + args->minalignslop - 1) > longest) return false; - /* do have enough free space remaining for the allocation? */ + /* do we have enough free space remaining for the allocation? */ available = (int)(pag->pagf_freeblks + pag->pagf_flcount - - min_free - args->total); - if (available < (int)args->minleft) + reservation - min_free - args->total); + if (available < (int)args->minleft || available <= 0) return false; return true; @@ -2058,7 +2115,7 @@ xfs_alloc_fix_freelist( * somewhere else if we are not being asked to try harder at this * point */ - if (pag->pagf_metadata && args->userdata && + if (pag->pagf_metadata && xfs_alloc_is_userdata(args->datatype) && (flags & XFS_ALLOC_FLAG_TRYLOCK)) { ASSERT(!(flags & XFS_ALLOC_FLAG_FREEING)); goto out_agbp_relse; @@ -2124,7 +2181,7 @@ xfs_alloc_fix_freelist( if (error) goto out_agbp_relse; error = xfs_free_ag_extent(tp, agbp, args->agno, bno, 1, - &targs.oinfo, 1); + &targs.oinfo, XFS_AG_RESV_AGFL); if (error) goto out_agbp_relse; bp = xfs_btree_get_bufs(mp, tp, args->agno, bno, 0); @@ -2135,7 +2192,7 @@ xfs_alloc_fix_freelist( targs.mp = mp; targs.agbp = agbp; targs.agno = args->agno; - targs.alignment = targs.minlen = targs.prod = targs.isfl = 1; + targs.alignment = targs.minlen = targs.prod = 1; targs.type = XFS_ALLOCTYPE_THIS_AG; targs.pag = pag; error = xfs_alloc_read_agfl(mp, tp, targs.agno, &agflbp); @@ -2146,6 +2203,7 @@ xfs_alloc_fix_freelist( while (pag->pagf_flcount < need) { targs.agbno = 0; targs.maxlen = need - pag->pagf_flcount; + targs.resv = XFS_AG_RESV_AGFL; /* Allocate as many blocks as possible at once. */ error = xfs_alloc_ag_vextent(&targs); @@ -2278,6 +2336,9 @@ xfs_alloc_log_agf( offsetof(xfs_agf_t, agf_btreeblks), offsetof(xfs_agf_t, agf_uuid), offsetof(xfs_agf_t, agf_rmap_blocks), + offsetof(xfs_agf_t, agf_refcount_blocks), + offsetof(xfs_agf_t, agf_refcount_root), + offsetof(xfs_agf_t, agf_refcount_level), /* needed so that we don't log the whole rest of the structure: */ offsetof(xfs_agf_t, agf_spare64), sizeof(xfs_agf_t) @@ -2394,12 +2455,15 @@ xfs_agf_verify( be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp))) return false; - if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS || + if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 || + be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 || + be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS || be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) > XFS_BTREE_MAXLEVELS) return false; if (xfs_sb_version_hasrmapbt(&mp->m_sb) && - be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS) + (be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) < 1 || + be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS)) return false; /* @@ -2415,6 +2479,11 @@ xfs_agf_verify( be32_to_cpu(agf->agf_btreeblks) > be32_to_cpu(agf->agf_length)) return false; + if (xfs_sb_version_hasreflink(&mp->m_sb) && + (be32_to_cpu(agf->agf_refcount_level) < 1 || + be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS)) + return false; + return true;; } @@ -2535,6 +2604,7 @@ xfs_alloc_read_agf( be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]); pag->pagf_levels[XFS_BTNUM_RMAPi] = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]); + pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level); spin_lock_init(&pag->pagb_lock); pag->pagb_count = 0; pag->pagb_tree = RB_ROOT; @@ -2633,7 +2703,7 @@ xfs_alloc_vextent( * Try near allocation first, then anywhere-in-ag after * the first a.g. fails. */ - if ((args->userdata & XFS_ALLOC_INITIAL_USER_DATA) && + if ((args->datatype & XFS_ALLOC_INITIAL_USER_DATA) && (mp->m_flags & XFS_MOUNT_32BITINODES)) { args->fsbno = XFS_AGB_TO_FSB(mp, ((mp->m_agfrotor / rotorstep) % @@ -2766,7 +2836,7 @@ xfs_alloc_vextent( #endif /* Zero the extent if we were asked to do so */ - if (args->userdata & XFS_ALLOC_USERDATA_ZERO) { + if (args->datatype & XFS_ALLOC_USERDATA_ZERO) { error = xfs_zero_extent(args->ip, args->fsbno, args->len); if (error) goto error0; @@ -2825,7 +2895,8 @@ xfs_free_extent( struct xfs_trans *tp, /* transaction pointer */ xfs_fsblock_t bno, /* starting block number of extent */ xfs_extlen_t len, /* length of extent */ - struct xfs_owner_info *oinfo) /* extent owner */ + struct xfs_owner_info *oinfo, /* extent owner */ + enum xfs_ag_resv_type type) /* block reservation type */ { struct xfs_mount *mp = tp->t_mountp; struct xfs_buf *agbp; @@ -2834,6 +2905,7 @@ xfs_free_extent( int error; ASSERT(len != 0); + ASSERT(type != XFS_AG_RESV_AGFL); if (XFS_TEST_ERROR(false, mp, XFS_ERRTAG_FREE_EXTENT, @@ -2851,7 +2923,7 @@ xfs_free_extent( agbno + len <= be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_length), err); - error = xfs_free_ag_extent(tp, agbp, agno, agbno, len, oinfo, 0); + error = xfs_free_ag_extent(tp, agbp, agno, agbno, len, oinfo, type); if (error) goto err; diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h index 6fe2d6b7cfe9..7c404a6b0ae3 100644 --- a/fs/xfs/libxfs/xfs_alloc.h +++ b/fs/xfs/libxfs/xfs_alloc.h @@ -85,20 +85,33 @@ typedef struct xfs_alloc_arg { xfs_extlen_t len; /* output: actual size of extent */ xfs_alloctype_t type; /* allocation type XFS_ALLOCTYPE_... */ xfs_alloctype_t otype; /* original allocation type */ + int datatype; /* mask defining data type treatment */ char wasdel; /* set if allocation was prev delayed */ char wasfromfl; /* set if allocation is from freelist */ - char isfl; /* set if is freelist blocks - !acctg */ - char userdata; /* mask defining userdata treatment */ xfs_fsblock_t firstblock; /* io first block allocated */ struct xfs_owner_info oinfo; /* owner of blocks being allocated */ + enum xfs_ag_resv_type resv; /* block reservation to use */ } xfs_alloc_arg_t; /* - * Defines for userdata + * Defines for datatype */ #define XFS_ALLOC_USERDATA (1 << 0)/* allocation is for user data*/ #define XFS_ALLOC_INITIAL_USER_DATA (1 << 1)/* special case start of file */ #define XFS_ALLOC_USERDATA_ZERO (1 << 2)/* zero extent on allocation */ +#define XFS_ALLOC_NOBUSY (1 << 3)/* Busy extents not allowed */ + +static inline bool +xfs_alloc_is_userdata(int datatype) +{ + return (datatype & ~XFS_ALLOC_NOBUSY) != 0; +} + +static inline bool +xfs_alloc_allow_busy_reuse(int datatype) +{ + return (datatype & XFS_ALLOC_NOBUSY) == 0; +} /* freespace limit calculations */ #define XFS_ALLOC_AGFL_RESERVE 4 @@ -106,7 +119,8 @@ unsigned int xfs_alloc_set_aside(struct xfs_mount *mp); unsigned int xfs_alloc_ag_max_usable(struct xfs_mount *mp); xfs_extlen_t xfs_alloc_longest_free_extent(struct xfs_mount *mp, - struct xfs_perag *pag, xfs_extlen_t need); + struct xfs_perag *pag, xfs_extlen_t need, + xfs_extlen_t reserved); unsigned int xfs_alloc_min_freelist(struct xfs_mount *mp, struct xfs_perag *pag); @@ -184,7 +198,8 @@ xfs_free_extent( struct xfs_trans *tp, /* transaction pointer */ xfs_fsblock_t bno, /* starting block number of extent */ xfs_extlen_t len, /* length of extent */ - struct xfs_owner_info *oinfo);/* extent owner */ + struct xfs_owner_info *oinfo, /* extent owner */ + enum xfs_ag_resv_type type); /* block reservation type */ int /* error */ xfs_alloc_lookup_ge( diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c index 5ba2dac5e67c..efb467b10a71 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.c +++ b/fs/xfs/libxfs/xfs_alloc_btree.c @@ -421,13 +421,17 @@ xfs_allocbt_init_cursor( ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT); - cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP); + cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); cur->bc_tp = tp; cur->bc_mp = mp; cur->bc_btnum = btnum; cur->bc_blocklog = mp->m_sb.sb_blocklog; cur->bc_ops = &xfs_allocbt_ops; + if (btnum == XFS_BTNUM_BNO) + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2); + else + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2); if (btnum == XFS_BTNUM_CNT) { cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]); diff --git a/fs/xfs/libxfs/xfs_attr_leaf.c b/fs/xfs/libxfs/xfs_attr_leaf.c index 8ea91f363093..2852521fc8ec 100644 --- a/fs/xfs/libxfs/xfs_attr_leaf.c +++ b/fs/xfs/libxfs/xfs_attr_leaf.c @@ -253,6 +253,7 @@ xfs_attr3_leaf_verify( { struct xfs_mount *mp = bp->b_target->bt_mount; struct xfs_attr_leafblock *leaf = bp->b_addr; + struct xfs_perag *pag = bp->b_pag; struct xfs_attr3_icleaf_hdr ichdr; xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &ichdr, leaf); @@ -273,7 +274,12 @@ xfs_attr3_leaf_verify( if (ichdr.magic != XFS_ATTR_LEAF_MAGIC) return false; } - if (ichdr.count == 0) + /* + * In recovery there is a transient state where count == 0 is valid + * because we may have transitioned an empty shortform attr to a leaf + * if the attr didn't fit in shortform. + */ + if (pag && pag->pagf_init && ichdr.count == 0) return false; /* XXX: need to range check rest of attr header values */ diff --git a/fs/xfs/libxfs/xfs_attr_leaf.h b/fs/xfs/libxfs/xfs_attr_leaf.h index 4f2aed04f827..f7dda0c237b0 100644 --- a/fs/xfs/libxfs/xfs_attr_leaf.h +++ b/fs/xfs/libxfs/xfs_attr_leaf.h @@ -51,7 +51,7 @@ int xfs_attr_shortform_getvalue(struct xfs_da_args *args); int xfs_attr_shortform_to_leaf(struct xfs_da_args *args); int xfs_attr_shortform_remove(struct xfs_da_args *args); int xfs_attr_shortform_allfit(struct xfs_buf *bp, struct xfs_inode *dp); -int xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes); +int xfs_attr_shortform_bytesfit(struct xfs_inode *dp, int bytes); void xfs_attr_fork_remove(struct xfs_inode *ip, struct xfs_trans *tp); /* @@ -77,7 +77,7 @@ int xfs_attr3_leaf_add(struct xfs_buf *leaf_buffer, struct xfs_da_args *args); int xfs_attr3_leaf_remove(struct xfs_buf *leaf_buffer, struct xfs_da_args *args); -int xfs_attr3_leaf_list_int(struct xfs_buf *bp, +void xfs_attr3_leaf_list_int(struct xfs_buf *bp, struct xfs_attr_list_context *context); /* diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index b060bca93402..2760bc3b2536 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -47,6 +47,10 @@ #include "xfs_attr_leaf.h" #include "xfs_filestream.h" #include "xfs_rmap.h" +#include "xfs_ag_resv.h" +#include "xfs_refcount.h" +#include "xfs_rmap_btree.h" +#include "xfs_icache.h" kmem_zone_t *xfs_bmap_free_item_zone; @@ -139,7 +143,8 @@ xfs_bmbt_lookup_ge( */ static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork) { - return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS && + return whichfork != XFS_COW_FORK && + XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS && XFS_IFORK_NEXTENTS(ip, whichfork) > XFS_IFORK_MAXEXT(ip, whichfork); } @@ -149,7 +154,8 @@ static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork) */ static inline bool xfs_bmap_wants_extents(struct xfs_inode *ip, int whichfork) { - return XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && + return whichfork != XFS_COW_FORK && + XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE && XFS_IFORK_NEXTENTS(ip, whichfork) <= XFS_IFORK_MAXEXT(ip, whichfork); } @@ -186,8 +192,12 @@ xfs_bmap_worst_indlen( int maxrecs; /* maximum record count at this level */ xfs_mount_t *mp; /* mount structure */ xfs_filblks_t rval; /* return value */ + xfs_filblks_t orig_len; mp = ip->i_mount; + + /* Calculate the worst-case size of the bmbt. */ + orig_len = len; maxrecs = mp->m_bmap_dmxr[0]; for (level = 0, rval = 0; level < XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK); @@ -195,12 +205,20 @@ xfs_bmap_worst_indlen( len += maxrecs - 1; do_div(len, maxrecs); rval += len; - if (len == 1) - return rval + XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - + if (len == 1) { + rval += XFS_BM_MAXLEVELS(mp, XFS_DATA_FORK) - level - 1; + break; + } if (level == 0) maxrecs = mp->m_bmap_dmxr[1]; } + + /* Calculate the worst-case size of the rmapbt. */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + rval += 1 + xfs_rmapbt_calc_size(mp, orig_len) + + mp->m_rmap_maxlevels; + return rval; } @@ -500,7 +518,7 @@ void xfs_bmap_trace_exlist( xfs_inode_t *ip, /* incore inode pointer */ xfs_extnum_t cnt, /* count of entries in the list */ - int whichfork, /* data or attr fork */ + int whichfork, /* data or attr or cow fork */ unsigned long caller_ip) { xfs_extnum_t idx; /* extent record index */ @@ -509,11 +527,13 @@ xfs_bmap_trace_exlist( if (whichfork == XFS_ATTR_FORK) state |= BMAP_ATTRFORK; + else if (whichfork == XFS_COW_FORK) + state |= BMAP_COWFORK; ifp = XFS_IFORK_PTR(ip, whichfork); - ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))); + ASSERT(cnt == xfs_iext_count(ifp)); for (idx = 0; idx < cnt; idx++) - trace_xfs_extlist(ip, idx, whichfork, caller_ip); + trace_xfs_extlist(ip, idx, state, caller_ip); } /* @@ -639,6 +659,7 @@ xfs_bmap_btree_to_extents( mp = ip->i_mount; ifp = XFS_IFORK_PTR(ip, whichfork); + ASSERT(whichfork != XFS_COW_FORK); ASSERT(ifp->if_flags & XFS_IFEXTENTS); ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE); rblock = ifp->if_broot; @@ -705,6 +726,7 @@ xfs_bmap_extents_to_btree( xfs_bmbt_ptr_t *pp; /* root block address pointer */ mp = ip->i_mount; + ASSERT(whichfork != XFS_COW_FORK); ifp = XFS_IFORK_PTR(ip, whichfork); ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS); @@ -747,6 +769,7 @@ xfs_bmap_extents_to_btree( args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = XFS_INO_TO_FSB(mp, ip->i_ino); } else if (dfops->dop_low) { +try_another_ag: args.type = XFS_ALLOCTYPE_START_BNO; args.fsbno = *firstblock; } else { @@ -761,6 +784,21 @@ xfs_bmap_extents_to_btree( xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); return error; } + + /* + * During a CoW operation, the allocation and bmbt updates occur in + * different transactions. The mapping code tries to put new bmbt + * blocks near extents being mapped, but the only way to guarantee this + * is if the alloc and the mapping happen in a single transaction that + * has a block reservation. That isn't the case here, so if we run out + * of space we'll try again with another AG. + */ + if (xfs_sb_version_hasreflink(&cur->bc_mp->m_sb) && + args.fsbno == NULLFSBLOCK && + args.type == XFS_ALLOCTYPE_NEAR_BNO) { + dfops->dop_low = true; + goto try_another_ag; + } /* * Allocation can't fail, the space was reserved. */ @@ -789,7 +827,7 @@ xfs_bmap_extents_to_btree( XFS_BTREE_LONG_PTRS); arp = XFS_BMBT_REC_ADDR(mp, ablock, 1); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); for (cnt = i = 0; i < nextents; i++) { ep = xfs_iext_get_ext(ifp, i); if (!isnullstartblock(xfs_bmbt_get_startblock(ep))) { @@ -836,6 +874,7 @@ xfs_bmap_local_to_extents_empty( { struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); + ASSERT(whichfork != XFS_COW_FORK); ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL); ASSERT(ifp->if_bytes == 0); ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0); @@ -895,6 +934,7 @@ xfs_bmap_local_to_extents( * file currently fits in an inode. */ if (*firstblock == NULLFSBLOCK) { +try_another_ag: args.fsbno = XFS_INO_TO_FSB(args.mp, ip->i_ino); args.type = XFS_ALLOCTYPE_START_BNO; } else { @@ -907,6 +947,19 @@ xfs_bmap_local_to_extents( if (error) goto done; + /* + * During a CoW operation, the allocation and bmbt updates occur in + * different transactions. The mapping code tries to put new bmbt + * blocks near extents being mapped, but the only way to guarantee this + * is if the alloc and the mapping happen in a single transaction that + * has a block reservation. That isn't the case here, so if we run out + * of space we'll try again with another AG. + */ + if (xfs_sb_version_hasreflink(&ip->i_mount->m_sb) && + args.fsbno == NULLFSBLOCK && + args.type == XFS_ALLOCTYPE_NEAR_BNO) { + goto try_another_ag; + } /* Can't fail, the space was reserved. */ ASSERT(args.fsbno != NULLFSBLOCK); ASSERT(args.len == 1); @@ -1100,6 +1153,10 @@ xfs_bmap_add_attrfork( goto trans_cancel; if (XFS_IFORK_Q(ip)) goto trans_cancel; + if (ip->i_d.di_anextents != 0) { + error = -EFSCORRUPTED; + goto trans_cancel; + } if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) { /* * For inodes coming from pre-6.2 filesystems. @@ -1107,7 +1164,6 @@ xfs_bmap_add_attrfork( ASSERT(ip->i_d.di_aformat == 0); ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS; } - ASSERT(ip->i_d.di_anextents == 0); xfs_trans_ijoin(tp, ip, 0); xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); @@ -1259,7 +1315,7 @@ xfs_bmap_read_extents( /* * Here with bp and block set to the leftmost leaf node in the tree. */ - room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + room = xfs_iext_count(ifp); i = 0; /* * Loop over all leaf nodes. Copy information to the extent records. @@ -1324,8 +1380,9 @@ xfs_bmap_read_extents( return error; block = XFS_BUF_TO_BLOCK(bp); } - ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))); - ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork)); + if (i != XFS_IFORK_NEXTENTS(ip, whichfork)) + return -EFSCORRUPTED; + ASSERT(i == xfs_iext_count(ifp)); XFS_BMAP_TRACE_EXLIST(ip, i, whichfork); return 0; error0: @@ -1333,97 +1390,6 @@ error0: return -EFSCORRUPTED; } - -/* - * Search the extent records for the entry containing block bno. - * If bno lies in a hole, point to the next entry. If bno lies - * past eof, *eofp will be set, and *prevp will contain the last - * entry (null if none). Else, *lastxp will be set to the index - * of the found entry; *gotp will contain the entry. - */ -STATIC xfs_bmbt_rec_host_t * /* pointer to found extent entry */ -xfs_bmap_search_multi_extents( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_fileoff_t bno, /* block number searched for */ - int *eofp, /* out: end of file found */ - xfs_extnum_t *lastxp, /* out: last extent index */ - xfs_bmbt_irec_t *gotp, /* out: extent entry found */ - xfs_bmbt_irec_t *prevp) /* out: previous extent entry found */ -{ - xfs_bmbt_rec_host_t *ep; /* extent record pointer */ - xfs_extnum_t lastx; /* last extent index */ - - /* - * Initialize the extent entry structure to catch access to - * uninitialized br_startblock field. - */ - gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL; - gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL; - gotp->br_state = XFS_EXT_INVALID; - gotp->br_startblock = 0xffffa5a5a5a5a5a5LL; - prevp->br_startoff = NULLFILEOFF; - - ep = xfs_iext_bno_to_ext(ifp, bno, &lastx); - if (lastx > 0) { - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp); - } - if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) { - xfs_bmbt_get_all(ep, gotp); - *eofp = 0; - } else { - if (lastx > 0) { - *gotp = *prevp; - } - *eofp = 1; - ep = NULL; - } - *lastxp = lastx; - return ep; -} - -/* - * Search the extents list for the inode, for the extent containing bno. - * If bno lies in a hole, point to the next entry. If bno lies past eof, - * *eofp will be set, and *prevp will contain the last entry (null if none). - * Else, *lastxp will be set to the index of the found - * entry; *gotp will contain the entry. - */ -STATIC xfs_bmbt_rec_host_t * /* pointer to found extent entry */ -xfs_bmap_search_extents( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_fileoff_t bno, /* block number searched for */ - int fork, /* data or attr fork */ - int *eofp, /* out: end of file found */ - xfs_extnum_t *lastxp, /* out: last extent index */ - xfs_bmbt_irec_t *gotp, /* out: extent entry found */ - xfs_bmbt_irec_t *prevp) /* out: previous extent entry found */ -{ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_bmbt_rec_host_t *ep; /* extent record pointer */ - - XFS_STATS_INC(ip->i_mount, xs_look_exlist); - ifp = XFS_IFORK_PTR(ip, fork); - - ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp); - - if (unlikely(!(gotp->br_startblock) && (*lastxp != NULLEXTNUM) && - !(XFS_IS_REALTIME_INODE(ip) && fork == XFS_DATA_FORK))) { - xfs_alert_tag(ip->i_mount, XFS_PTAG_FSBLOCK_ZERO, - "Access to block zero in inode %llu " - "start_block: %llx start_off: %llx " - "blkcnt: %llx extent-state: %x lastx: %x", - (unsigned long long)ip->i_ino, - (unsigned long long)gotp->br_startblock, - (unsigned long long)gotp->br_startoff, - (unsigned long long)gotp->br_blockcount, - gotp->br_state, *lastxp); - *lastxp = NULLEXTNUM; - *eofp = 1; - return NULL; - } - return ep; -} - /* * Returns the file-relative block number of the first unused block(s) * in the file with at least "len" logically contiguous blocks free. @@ -1460,7 +1426,7 @@ xfs_bmap_first_unused( (error = xfs_iread_extents(tp, ip, whichfork))) return error; lowest = *first_unused; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) { xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, idx); off = xfs_bmbt_get_startoff(ep); @@ -1486,44 +1452,44 @@ xfs_bmap_first_unused( */ int /* error */ xfs_bmap_last_before( - xfs_trans_t *tp, /* transaction pointer */ - xfs_inode_t *ip, /* incore inode */ - xfs_fileoff_t *last_block, /* last block */ - int whichfork) /* data or attr fork */ + struct xfs_trans *tp, /* transaction pointer */ + struct xfs_inode *ip, /* incore inode */ + xfs_fileoff_t *last_block, /* last block */ + int whichfork) /* data or attr fork */ { - xfs_fileoff_t bno; /* input file offset */ - int eof; /* hit end of file */ - xfs_bmbt_rec_host_t *ep; /* pointer to last extent */ - int error; /* error return value */ - xfs_bmbt_irec_t got; /* current extent value */ - xfs_ifork_t *ifp; /* inode fork pointer */ - xfs_extnum_t lastx; /* last extent used */ - xfs_bmbt_irec_t prev; /* previous extent value */ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); + struct xfs_bmbt_irec got; + xfs_extnum_t idx; + int error; - if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE && - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL) - return -EIO; - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { + switch (XFS_IFORK_FORMAT(ip, whichfork)) { + case XFS_DINODE_FMT_LOCAL: *last_block = 0; return 0; + case XFS_DINODE_FMT_BTREE: + case XFS_DINODE_FMT_EXTENTS: + break; + default: + return -EIO; } - ifp = XFS_IFORK_PTR(ip, whichfork); - if (!(ifp->if_flags & XFS_IFEXTENTS) && - (error = xfs_iread_extents(tp, ip, whichfork))) - return error; - bno = *last_block - 1; - ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, - &prev); - if (eof || xfs_bmbt_get_startoff(ep) > bno) { - if (prev.br_startoff == NULLFILEOFF) - *last_block = 0; - else - *last_block = prev.br_startoff + prev.br_blockcount; + + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(tp, ip, whichfork); + if (error) + return error; } - /* - * Otherwise *last_block is already the right answer. - */ + + if (xfs_iext_lookup_extent(ip, ifp, *last_block - 1, &idx, &got)) { + if (got.br_startoff <= *last_block - 1) + return 0; + } + + if (xfs_iext_get_extent(ifp, idx - 1, &got)) { + *last_block = got.br_startoff + got.br_blockcount; + return 0; + } + + *last_block = 0; return 0; } @@ -1545,7 +1511,7 @@ xfs_bmap_last_extent( return error; } - nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); if (nextents == 0) { *is_empty = 1; return 0; @@ -1669,7 +1635,8 @@ xfs_bmap_one_block( */ STATIC int /* error */ xfs_bmap_add_extent_delay_real( - struct xfs_bmalloca *bma) + struct xfs_bmalloca *bma, + int whichfork) { struct xfs_bmbt_irec *new = &bma->got; int diff; /* temp value */ @@ -1687,14 +1654,17 @@ xfs_bmap_add_extent_delay_real( xfs_filblks_t temp=0; /* value for da_new calculations */ xfs_filblks_t temp2=0;/* value for da_new calculations */ int tmp_rval; /* partial logging flags */ - int whichfork = XFS_DATA_FORK; struct xfs_mount *mp; + xfs_extnum_t *nextents; mp = bma->ip->i_mount; ifp = XFS_IFORK_PTR(bma->ip, whichfork); + ASSERT(whichfork != XFS_ATTR_FORK); + nextents = (whichfork == XFS_COW_FORK ? &bma->ip->i_cnextents : + &bma->ip->i_d.di_nextents); ASSERT(bma->idx >= 0); - ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(bma->idx <= xfs_iext_count(ifp)); ASSERT(!isnullstartblock(new->br_startblock)); ASSERT(!bma->cur || (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); @@ -1705,6 +1675,9 @@ xfs_bmap_add_extent_delay_real( #define RIGHT r[1] #define PREV r[2] + if (whichfork == XFS_COW_FORK) + state |= BMAP_COWFORK; + /* * Set up a bunch of variables to make the tests simpler. */ @@ -1750,7 +1723,7 @@ xfs_bmap_add_extent_delay_real( * Don't set contiguous if the combined extent would be too large. * Also check for all-three-contiguous being too large. */ - if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) { + if (bma->idx < xfs_iext_count(ifp) - 1) { state |= BMAP_RIGHT_VALID; xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx + 1), &RIGHT); @@ -1791,7 +1764,7 @@ xfs_bmap_add_extent_delay_real( trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_iext_remove(bma->ip, bma->idx + 1, 2, state); - bma->ip->i_d.di_nextents--; + (*nextents)--; if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { @@ -1893,7 +1866,7 @@ xfs_bmap_add_extent_delay_real( xfs_bmbt_set_startblock(ep, new->br_startblock); trace_xfs_bmap_post_update(bma->ip, bma->idx, state, _THIS_IP_); - bma->ip->i_d.di_nextents++; + (*nextents)++; if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { @@ -1963,7 +1936,7 @@ xfs_bmap_add_extent_delay_real( temp = PREV.br_blockcount - new->br_blockcount; xfs_bmbt_set_blockcount(ep, temp); xfs_iext_insert(bma->ip, bma->idx, 1, new, state); - bma->ip->i_d.di_nextents++; + (*nextents)++; if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { @@ -2047,7 +2020,7 @@ xfs_bmap_add_extent_delay_real( trace_xfs_bmap_pre_update(bma->ip, bma->idx, state, _THIS_IP_); xfs_bmbt_set_blockcount(ep, temp); xfs_iext_insert(bma->ip, bma->idx + 1, 1, new, state); - bma->ip->i_d.di_nextents++; + (*nextents)++; if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { @@ -2116,7 +2089,7 @@ xfs_bmap_add_extent_delay_real( RIGHT.br_blockcount = temp2; /* insert LEFT (r[0]) and RIGHT (r[1]) at the same time */ xfs_iext_insert(bma->ip, bma->idx + 1, 2, &LEFT, state); - bma->ip->i_d.di_nextents++; + (*nextents)++; if (bma->cur == NULL) rval = XFS_ILOG_CORE | XFS_ILOG_DEXT; else { @@ -2214,7 +2187,8 @@ xfs_bmap_add_extent_delay_real( xfs_bmap_check_leaf_extents(bma->cur, bma->ip, whichfork); done: - bma->logflags |= rval; + if (whichfork != XFS_COW_FORK) + bma->logflags |= rval; return error; #undef LEFT #undef RIGHT @@ -2255,7 +2229,7 @@ xfs_bmap_add_extent_unwritten_real( ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); ASSERT(*idx >= 0); - ASSERT(*idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(*idx <= xfs_iext_count(ifp)); ASSERT(!isnullstartblock(new->br_startblock)); XFS_STATS_INC(mp, xs_add_exlist); @@ -2311,7 +2285,7 @@ xfs_bmap_add_extent_unwritten_real( * Don't set contiguous if the combined extent would be too large. * Also check for all-three-contiguous being too large. */ - if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1) { + if (*idx < xfs_iext_count(&ip->i_df) - 1) { state |= BMAP_RIGHT_VALID; xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx + 1), &RIGHT); if (isnullstartblock(RIGHT.br_startblock)) @@ -2758,6 +2732,7 @@ done: STATIC void xfs_bmap_add_extent_hole_delay( xfs_inode_t *ip, /* incore inode pointer */ + int whichfork, xfs_extnum_t *idx, /* extent number to update/insert */ xfs_bmbt_irec_t *new) /* new data to add to file extents */ { @@ -2769,8 +2744,10 @@ xfs_bmap_add_extent_hole_delay( int state; /* state bits, accessed thru macros */ xfs_filblks_t temp=0; /* temp for indirect calculations */ - ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + ifp = XFS_IFORK_PTR(ip, whichfork); state = 0; + if (whichfork == XFS_COW_FORK) + state |= BMAP_COWFORK; ASSERT(isnullstartblock(new->br_startblock)); /* @@ -2788,7 +2765,7 @@ xfs_bmap_add_extent_hole_delay( * Check and set flags if the current (right) segment exists. * If it doesn't exist, we're converting the hole at end-of-file. */ - if (*idx < ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) { + if (*idx < xfs_iext_count(ifp)) { state |= BMAP_RIGHT_VALID; xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *idx), &right); @@ -2918,10 +2895,11 @@ xfs_bmap_add_extent_hole_real( ifp = XFS_IFORK_PTR(bma->ip, whichfork); ASSERT(bma->idx >= 0); - ASSERT(bma->idx <= ifp->if_bytes / sizeof(struct xfs_bmbt_rec)); + ASSERT(bma->idx <= xfs_iext_count(ifp)); ASSERT(!isnullstartblock(new->br_startblock)); ASSERT(!bma->cur || !(bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL)); + ASSERT(whichfork != XFS_COW_FORK); XFS_STATS_INC(mp, xs_add_exlist); @@ -2943,7 +2921,7 @@ xfs_bmap_add_extent_hole_real( * Check and set flags if this segment has a current value. * Not true if we're inserting into the "hole" at eof. */ - if (bma->idx < ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)) { + if (bma->idx < xfs_iext_count(ifp)) { state |= BMAP_RIGHT_VALID; xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx), &right); if (isnullstartblock(right.br_startblock)) @@ -3347,7 +3325,8 @@ xfs_bmap_adjacent( mp = ap->ip->i_mount; nullfb = *ap->firstblock == NULLFSBLOCK; - rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata; + rt = XFS_IS_REALTIME_INODE(ap->ip) && + xfs_alloc_is_userdata(ap->datatype); fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); /* * If allocating at eof, and there's a previous real block, @@ -3501,7 +3480,8 @@ xfs_bmap_longest_free_extent( } longest = xfs_alloc_longest_free_extent(mp, pag, - xfs_alloc_min_freelist(mp, pag)); + xfs_alloc_min_freelist(mp, pag), + xfs_ag_resv_needed(pag, XFS_AG_RESV_NONE)); if (*blen < longest) *blen = longest; @@ -3622,7 +3602,7 @@ xfs_bmap_btalloc( { xfs_mount_t *mp; /* mount point structure */ xfs_alloctype_t atype = 0; /* type for allocation routines */ - xfs_extlen_t align; /* minimum allocation alignment */ + xfs_extlen_t align = 0; /* minimum allocation alignment */ xfs_agnumber_t fb_agno; /* ag number of ap->firstblock */ xfs_agnumber_t ag; xfs_alloc_arg_t args; @@ -3645,7 +3625,10 @@ xfs_bmap_btalloc( else if (mp->m_dalign) stripe_align = mp->m_dalign; - align = ap->userdata ? xfs_get_extsz_hint(ap->ip) : 0; + if (ap->flags & XFS_BMAPI_COWFORK) + align = xfs_get_cowextsz_hint(ap->ip); + else if (xfs_alloc_is_userdata(ap->datatype)) + align = xfs_get_extsz_hint(ap->ip); if (unlikely(align)) { error = xfs_bmap_extsize_align(mp, &ap->got, &ap->prev, align, 0, ap->eof, 0, ap->conv, @@ -3658,7 +3641,8 @@ xfs_bmap_btalloc( nullfb = *ap->firstblock == NULLFSBLOCK; fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, *ap->firstblock); if (nullfb) { - if (ap->userdata && xfs_inode_is_filestream(ap->ip)) { + if (xfs_alloc_is_userdata(ap->datatype) && + xfs_inode_is_filestream(ap->ip)) { ag = xfs_filestream_lookup_ag(ap->ip); ag = (ag != NULLAGNUMBER) ? ag : 0; ap->blkno = XFS_AGB_TO_FSB(mp, ag, 0); @@ -3698,7 +3682,8 @@ xfs_bmap_btalloc( * enough for the request. If one isn't found, then adjust * the minimum allocation size to the largest space found. */ - if (ap->userdata && xfs_inode_is_filestream(ap->ip)) + if (xfs_alloc_is_userdata(ap->datatype) && + xfs_inode_is_filestream(ap->ip)) error = xfs_bmap_btalloc_filestreams(ap, &args, &blen); else error = xfs_bmap_btalloc_nullfb(ap, &args, &blen); @@ -3781,9 +3766,9 @@ xfs_bmap_btalloc( } args.minleft = ap->minleft; args.wasdel = ap->wasdel; - args.isfl = 0; - args.userdata = ap->userdata; - if (ap->userdata & XFS_ALLOC_USERDATA_ZERO) + args.resv = XFS_AG_RESV_NONE; + args.datatype = ap->datatype; + if (ap->datatype & XFS_ALLOC_USERDATA_ZERO) args.ip = ap->ip; error = xfs_alloc_vextent(&args); @@ -3850,7 +3835,8 @@ xfs_bmap_btalloc( ASSERT(nullfb || fb_agno == args.agno || (ap->dfops->dop_low && fb_agno < args.agno)); ap->length = args.len; - ap->ip->i_d.di_nblocks += args.len; + if (!(ap->flags & XFS_BMAPI_COWFORK)) + ap->ip->i_d.di_nblocks += args.len; xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); if (ap->wasdel) ap->ip->i_delayed_blks -= args.len; @@ -3870,6 +3856,60 @@ xfs_bmap_btalloc( } /* + * For a remap operation, just "allocate" an extent at the address that the + * caller passed in, and ensure that the AGFL is the right size. The caller + * will then map the "allocated" extent into the file somewhere. + */ +STATIC int +xfs_bmap_remap_alloc( + struct xfs_bmalloca *ap) +{ + struct xfs_trans *tp = ap->tp; + struct xfs_mount *mp = tp->t_mountp; + xfs_agblock_t bno; + struct xfs_alloc_arg args; + int error; + + /* + * validate that the block number is legal - the enables us to detect + * and handle a silent filesystem corruption rather than crashing. + */ + memset(&args, 0, sizeof(struct xfs_alloc_arg)); + args.tp = ap->tp; + args.mp = ap->tp->t_mountp; + bno = *ap->firstblock; + args.agno = XFS_FSB_TO_AGNO(mp, bno); + args.agbno = XFS_FSB_TO_AGBNO(mp, bno); + if (args.agno >= mp->m_sb.sb_agcount || + args.agbno >= mp->m_sb.sb_agblocks) + return -EFSCORRUPTED; + + /* "Allocate" the extent from the range we passed in. */ + trace_xfs_bmap_remap_alloc(ap->ip, *ap->firstblock, ap->length); + ap->blkno = bno; + ap->ip->i_d.di_nblocks += ap->length; + xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE); + + /* Fix the freelist, like a real allocator does. */ + args.datatype = ap->datatype; + args.pag = xfs_perag_get(args.mp, args.agno); + ASSERT(args.pag); + + /* + * The freelist fixing code will decline the allocation if + * the size and shape of the free space doesn't allow for + * allocating the extent and updating all the metadata that + * happens during an allocation. We're remapping, not + * allocating, so skip that check by pretending to be freeing. + */ + error = xfs_alloc_fix_freelist(&args, XFS_ALLOC_FLAG_FREEING); + xfs_perag_put(args.pag); + if (error) + trace_xfs_bmap_remap_alloc_error(ap->ip, error, _RET_IP_); + return error; +} + +/* * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file. * It figures out where to ask the underlying allocator to put the new extent. */ @@ -3877,11 +3917,47 @@ STATIC int xfs_bmap_alloc( struct xfs_bmalloca *ap) /* bmap alloc argument struct */ { - if (XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata) + if (ap->flags & XFS_BMAPI_REMAP) + return xfs_bmap_remap_alloc(ap); + if (XFS_IS_REALTIME_INODE(ap->ip) && + xfs_alloc_is_userdata(ap->datatype)) return xfs_bmap_rtalloc(ap); return xfs_bmap_btalloc(ap); } +/* Trim extent to fit a logical block range. */ +void +xfs_trim_extent( + struct xfs_bmbt_irec *irec, + xfs_fileoff_t bno, + xfs_filblks_t len) +{ + xfs_fileoff_t distance; + xfs_fileoff_t end = bno + len; + + if (irec->br_startoff + irec->br_blockcount <= bno || + irec->br_startoff >= end) { + irec->br_blockcount = 0; + return; + } + + if (irec->br_startoff < bno) { + distance = bno - irec->br_startoff; + if (isnullstartblock(irec->br_startblock)) + irec->br_startblock = DELAYSTARTBLOCK; + if (irec->br_startblock != DELAYSTARTBLOCK && + irec->br_startblock != HOLESTARTBLOCK) + irec->br_startblock += distance; + irec->br_startoff += distance; + irec->br_blockcount -= distance; + } + + if (end < irec->br_startoff + irec->br_blockcount) { + distance = irec->br_startoff + irec->br_blockcount - end; + irec->br_blockcount -= distance; + } +} + /* * Trim the returned map to the required bounds */ @@ -3998,19 +4074,17 @@ xfs_bmapi_read( struct xfs_mount *mp = ip->i_mount; struct xfs_ifork *ifp; struct xfs_bmbt_irec got; - struct xfs_bmbt_irec prev; xfs_fileoff_t obno; xfs_fileoff_t end; - xfs_extnum_t lastx; + xfs_extnum_t idx; int error; - int eof; + bool eof = false; int n = 0; - int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? - XFS_ATTR_FORK : XFS_DATA_FORK; + int whichfork = xfs_bmapi_whichfork(flags); ASSERT(*nmap >= 1); ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE| - XFS_BMAPI_IGSTATE))); + XFS_BMAPI_IGSTATE|XFS_BMAPI_COWFORK))); ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)); if (unlikely(XFS_TEST_ERROR( @@ -4028,13 +4102,24 @@ xfs_bmapi_read( ifp = XFS_IFORK_PTR(ip, whichfork); + /* No CoW fork? Return a hole. */ + if (whichfork == XFS_COW_FORK && !ifp) { + mval->br_startoff = bno; + mval->br_startblock = HOLESTARTBLOCK; + mval->br_blockcount = len; + mval->br_state = XFS_EXT_NORM; + *nmap = 1; + return 0; + } + if (!(ifp->if_flags & XFS_IFEXTENTS)) { error = xfs_iread_extents(NULL, ip, whichfork); if (error) return error; } - xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, &prev); + if (!xfs_iext_lookup_extent(ip, ifp, bno, &idx, &got)) + eof = true; end = bno + len; obno = bno; @@ -4065,41 +4150,55 @@ xfs_bmapi_read( break; /* Else go on to the next record. */ - if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); - else - eof = 1; + if (!xfs_iext_get_extent(ifp, ++idx, &got)) + eof = true; } *nmap = n; return 0; } -STATIC int +int xfs_bmapi_reserve_delalloc( struct xfs_inode *ip, - xfs_fileoff_t aoff, + int whichfork, + xfs_fileoff_t off, xfs_filblks_t len, + xfs_filblks_t prealloc, struct xfs_bmbt_irec *got, - struct xfs_bmbt_irec *prev, xfs_extnum_t *lastx, int eof) { struct xfs_mount *mp = ip->i_mount; - struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); xfs_extlen_t alen; xfs_extlen_t indlen; char rt = XFS_IS_REALTIME_INODE(ip); xfs_extlen_t extsz; int error; + xfs_fileoff_t aoff = off; - alen = XFS_FILBLKS_MIN(len, MAXEXTLEN); + /* + * Cap the alloc length. Keep track of prealloc so we know whether to + * tag the inode before we return. + */ + alen = XFS_FILBLKS_MIN(len + prealloc, MAXEXTLEN); if (!eof) alen = XFS_FILBLKS_MIN(alen, got->br_startoff - aoff); + if (prealloc && alen >= len) + prealloc = alen - len; /* Figure out the extent size, adjust alen */ - extsz = xfs_get_extsz_hint(ip); + if (whichfork == XFS_COW_FORK) + extsz = xfs_get_cowextsz_hint(ip); + else + extsz = xfs_get_extsz_hint(ip); if (extsz) { - error = xfs_bmap_extsize_align(mp, got, prev, extsz, rt, eof, + struct xfs_bmbt_irec prev; + + if (!xfs_iext_get_extent(ifp, *lastx - 1, &prev)) + prev.br_startoff = NULLFILEOFF; + + error = xfs_bmap_extsize_align(mp, got, &prev, extsz, rt, eof, 1, 0, &aoff, &alen); ASSERT(!error); } @@ -4144,7 +4243,7 @@ xfs_bmapi_reserve_delalloc( got->br_startblock = nullstartblock(indlen); got->br_blockcount = alen; got->br_state = XFS_EXT_NORM; - xfs_bmap_add_extent_hole_delay(ip, lastx, got); + xfs_bmap_add_extent_hole_delay(ip, whichfork, lastx, got); /* * Update our extent pointer, given that xfs_bmap_add_extent_hole_delay @@ -4152,6 +4251,16 @@ xfs_bmapi_reserve_delalloc( */ xfs_bmbt_get_all(xfs_iext_get_ext(ifp, *lastx), got); + /* + * Tag the inode if blocks were preallocated. Note that COW fork + * preallocation can occur at the start or end of the extent, even when + * prealloc == 0, so we must also check the aligned offset and length. + */ + if (whichfork == XFS_DATA_FORK && prealloc) + xfs_inode_set_eofblocks_tag(ip); + if (whichfork == XFS_COW_FORK && (prealloc || aoff < off || alen > len)) + xfs_inode_set_cowblocks_tag(ip); + ASSERT(got->br_startoff <= aoff); ASSERT(got->br_startoff + got->br_blockcount >= aoff + alen); ASSERT(isnullstartblock(got->br_startblock)); @@ -4170,98 +4279,12 @@ out_unreserve_quota: return error; } -/* - * Map file blocks to filesystem blocks, adding delayed allocations as needed. - */ -int -xfs_bmapi_delay( - struct xfs_inode *ip, /* incore inode */ - xfs_fileoff_t bno, /* starting file offs. mapped */ - xfs_filblks_t len, /* length to map in file */ - struct xfs_bmbt_irec *mval, /* output: map values */ - int *nmap, /* i/o: mval size/count */ - int flags) /* XFS_BMAPI_... */ -{ - struct xfs_mount *mp = ip->i_mount; - struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); - struct xfs_bmbt_irec got; /* current file extent record */ - struct xfs_bmbt_irec prev; /* previous file extent record */ - xfs_fileoff_t obno; /* old block number (offset) */ - xfs_fileoff_t end; /* end of mapped file region */ - xfs_extnum_t lastx; /* last useful extent number */ - int eof; /* we've hit the end of extents */ - int n = 0; /* current extent index */ - int error = 0; - - ASSERT(*nmap >= 1); - ASSERT(*nmap <= XFS_BMAP_MAX_NMAP); - ASSERT(!(flags & ~XFS_BMAPI_ENTIRE)); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); - - if (unlikely(XFS_TEST_ERROR( - (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS && - XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_BTREE), - mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { - XFS_ERROR_REPORT("xfs_bmapi_delay", XFS_ERRLEVEL_LOW, mp); - return -EFSCORRUPTED; - } - - if (XFS_FORCED_SHUTDOWN(mp)) - return -EIO; - - XFS_STATS_INC(mp, xs_blk_mapw); - - if (!(ifp->if_flags & XFS_IFEXTENTS)) { - error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK); - if (error) - return error; - } - - xfs_bmap_search_extents(ip, bno, XFS_DATA_FORK, &eof, &lastx, &got, &prev); - end = bno + len; - obno = bno; - - while (bno < end && n < *nmap) { - if (eof || got.br_startoff > bno) { - error = xfs_bmapi_reserve_delalloc(ip, bno, len, &got, - &prev, &lastx, eof); - if (error) { - if (n == 0) { - *nmap = 0; - return error; - } - break; - } - } - - /* set up the extent map to return. */ - xfs_bmapi_trim_map(mval, &got, &bno, len, obno, end, n, flags); - xfs_bmapi_update_map(&mval, &bno, &len, obno, end, &n, flags); - - /* If we're done, stop now. */ - if (bno >= end || n >= *nmap) - break; - - /* Else go on to the next record. */ - prev = got; - if (++lastx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx), &got); - else - eof = 1; - } - - *nmap = n; - return 0; -} - - static int xfs_bmapi_allocate( struct xfs_bmalloca *bma) { struct xfs_mount *mp = bma->ip->i_mount; - int whichfork = (bma->flags & XFS_BMAPI_ATTRFORK) ? - XFS_ATTR_FORK : XFS_DATA_FORK; + int whichfork = xfs_bmapi_whichfork(bma->flags); struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); int tmp_logflags = 0; int error; @@ -4275,7 +4298,7 @@ xfs_bmapi_allocate( if (bma->wasdel) { bma->length = (xfs_extlen_t)bma->got.br_blockcount; bma->offset = bma->got.br_startoff; - if (bma->idx != NULLEXTNUM && bma->idx) { + if (bma->idx) { xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma->idx - 1), &bma->prev); } @@ -4287,15 +4310,21 @@ xfs_bmapi_allocate( } /* - * Indicate if this is the first user data in the file, or just any - * user data. And if it is userdata, indicate whether it needs to - * be initialised to zero during allocation. + * Set the data type being allocated. For the data fork, the first data + * in the file is treated differently to all other allocations. For the + * attribute fork, we only need to ensure the allocated range is not on + * the busy list. */ if (!(bma->flags & XFS_BMAPI_METADATA)) { - bma->userdata = (bma->offset == 0) ? - XFS_ALLOC_INITIAL_USER_DATA : XFS_ALLOC_USERDATA; + bma->datatype = XFS_ALLOC_NOBUSY; + if (whichfork == XFS_DATA_FORK) { + if (bma->offset == 0) + bma->datatype |= XFS_ALLOC_INITIAL_USER_DATA; + else + bma->datatype |= XFS_ALLOC_USERDATA; + } if (bma->flags & XFS_BMAPI_ZERO) - bma->userdata |= XFS_ALLOC_USERDATA_ZERO; + bma->datatype |= XFS_ALLOC_USERDATA_ZERO; } bma->minlen = (bma->flags & XFS_BMAPI_CONTIG) ? bma->length : 1; @@ -4350,7 +4379,7 @@ xfs_bmapi_allocate( bma->got.br_state = XFS_EXT_UNWRITTEN; if (bma->wasdel) - error = xfs_bmap_add_extent_delay_real(bma); + error = xfs_bmap_add_extent_delay_real(bma, whichfork); else error = xfs_bmap_add_extent_hole_real(bma, whichfork); @@ -4380,8 +4409,7 @@ xfs_bmapi_convert_unwritten( xfs_filblks_t len, int flags) { - int whichfork = (flags & XFS_BMAPI_ATTRFORK) ? - XFS_ATTR_FORK : XFS_DATA_FORK; + int whichfork = xfs_bmapi_whichfork(flags); struct xfs_ifork *ifp = XFS_IFORK_PTR(bma->ip, whichfork); int tmp_logflags = 0; int error; @@ -4397,6 +4425,8 @@ xfs_bmapi_convert_unwritten( (XFS_BMAPI_PREALLOC | XFS_BMAPI_CONVERT)) return 0; + ASSERT(whichfork != XFS_COW_FORK); + /* * Modify (by adding) the state flag, if writing. */ @@ -4482,7 +4512,7 @@ xfs_bmapi_write( struct xfs_ifork *ifp; struct xfs_bmalloca bma = { NULL }; /* args for xfs_bmap_alloc */ xfs_fileoff_t end; /* end of mapped file region */ - int eof; /* after the end of extents */ + bool eof = false; /* after the end of extents */ int error; /* error return */ int n; /* current extent index */ xfs_fileoff_t obno; /* old block number (offset) */ @@ -4503,8 +4533,7 @@ xfs_bmapi_write( orig_mval = mval; orig_nmap = *nmap; #endif - whichfork = (flags & XFS_BMAPI_ATTRFORK) ? - XFS_ATTR_FORK : XFS_DATA_FORK; + whichfork = xfs_bmapi_whichfork(flags); ASSERT(*nmap >= 1); ASSERT(*nmap <= XFS_BMAP_MAX_NMAP); @@ -4513,6 +4542,11 @@ xfs_bmapi_write( ASSERT(len > 0); ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL); ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + ASSERT(!(flags & XFS_BMAPI_REMAP) || whichfork == XFS_DATA_FORK); + ASSERT(!(flags & XFS_BMAPI_PREALLOC) || !(flags & XFS_BMAPI_REMAP)); + ASSERT(!(flags & XFS_BMAPI_CONVERT) || !(flags & XFS_BMAPI_REMAP)); + ASSERT(!(flags & XFS_BMAPI_PREALLOC) || whichfork != XFS_COW_FORK); + ASSERT(!(flags & XFS_BMAPI_CONVERT) || whichfork != XFS_COW_FORK); /* zeroing is for currently only for data extents, not metadata */ ASSERT((flags & (XFS_BMAPI_METADATA | XFS_BMAPI_ZERO)) != @@ -4556,16 +4590,18 @@ xfs_bmapi_write( goto error0; } - xfs_bmap_search_extents(ip, bno, whichfork, &eof, &bma.idx, &bma.got, - &bma.prev); n = 0; end = bno + len; obno = bno; + if (!xfs_iext_lookup_extent(ip, ifp, bno, &bma.idx, &bma.got)) + eof = true; + if (!xfs_iext_get_extent(ifp, bma.idx - 1, &bma.prev)) + bma.prev.br_startoff = NULLFILEOFF; bma.tp = tp; bma.ip = ip; bma.total = total; - bma.userdata = 0; + bma.datatype = 0; bma.dfops = dfops; bma.firstblock = firstblock; @@ -4574,6 +4610,14 @@ xfs_bmapi_write( wasdelay = !inhole && isnullstartblock(bma.got.br_startblock); /* + * Make sure we only reflink into a hole. + */ + if (flags & XFS_BMAPI_REMAP) + ASSERT(inhole); + if (flags & XFS_BMAPI_COWFORK) + ASSERT(!inhole); + + /* * First, deal with the hole before the allocated space * that we found, if any. */ @@ -4603,6 +4647,17 @@ xfs_bmapi_write( goto error0; if (bma.blkno == NULLFSBLOCK) break; + + /* + * If this is a CoW allocation, record the data in + * the refcount btree for orphan recovery. + */ + if (whichfork == XFS_COW_FORK) { + error = xfs_refcount_alloc_cow_extent(mp, dfops, + bma.blkno, bma.length); + if (error) + goto error0; + } } /* Deal with the allocated space we found. */ @@ -4629,11 +4684,8 @@ xfs_bmapi_write( /* Else go on to the next record. */ bma.prev = bma.got; - if (++bma.idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)) { - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, bma.idx), - &bma.got); - } else - eof = 1; + if (!xfs_iext_get_extent(ifp, ++bma.idx, &bma.got)) + eof = true; } *nmap = n; @@ -4755,6 +4807,222 @@ xfs_bmap_split_indlen( return stolen; } +int +xfs_bmap_del_extent_delay( + struct xfs_inode *ip, + int whichfork, + xfs_extnum_t *idx, + struct xfs_bmbt_irec *got, + struct xfs_bmbt_irec *del) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); + struct xfs_bmbt_irec new; + int64_t da_old, da_new, da_diff = 0; + xfs_fileoff_t del_endoff, got_endoff; + xfs_filblks_t got_indlen, new_indlen, stolen; + int error = 0, state = 0; + bool isrt; + + XFS_STATS_INC(mp, xs_del_exlist); + + isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip); + del_endoff = del->br_startoff + del->br_blockcount; + got_endoff = got->br_startoff + got->br_blockcount; + da_old = startblockval(got->br_startblock); + da_new = 0; + + ASSERT(*idx >= 0); + ASSERT(*idx <= xfs_iext_count(ifp)); + ASSERT(del->br_blockcount > 0); + ASSERT(got->br_startoff <= del->br_startoff); + ASSERT(got_endoff >= del_endoff); + + if (isrt) { + int64_t rtexts = XFS_FSB_TO_B(mp, del->br_blockcount); + + do_div(rtexts, mp->m_sb.sb_rextsize); + xfs_mod_frextents(mp, rtexts); + } + + /* + * Update the inode delalloc counter now and wait to update the + * sb counters as we might have to borrow some blocks for the + * indirect block accounting. + */ + error = xfs_trans_reserve_quota_nblks(NULL, ip, + -((long)del->br_blockcount), 0, + isrt ? XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS); + if (error) + return error; + ip->i_delayed_blks -= del->br_blockcount; + + if (whichfork == XFS_COW_FORK) + state |= BMAP_COWFORK; + + if (got->br_startoff == del->br_startoff) + state |= BMAP_LEFT_CONTIG; + if (got_endoff == del_endoff) + state |= BMAP_RIGHT_CONTIG; + + switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) { + case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG: + /* + * Matches the whole extent. Delete the entry. + */ + xfs_iext_remove(ip, *idx, 1, state); + --*idx; + break; + case BMAP_LEFT_CONTIG: + /* + * Deleting the first part of the extent. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + got->br_startoff = del_endoff; + got->br_blockcount -= del->br_blockcount; + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, + got->br_blockcount), da_old); + got->br_startblock = nullstartblock((int)da_new); + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + break; + case BMAP_RIGHT_CONTIG: + /* + * Deleting the last part of the extent. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + got->br_blockcount = got->br_blockcount - del->br_blockcount; + da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, + got->br_blockcount), da_old); + got->br_startblock = nullstartblock((int)da_new); + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + break; + case 0: + /* + * Deleting the middle of the extent. + * + * Distribute the original indlen reservation across the two new + * extents. Steal blocks from the deleted extent if necessary. + * Stealing blocks simply fudges the fdblocks accounting below. + * Warn if either of the new indlen reservations is zero as this + * can lead to delalloc problems. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + + got->br_blockcount = del->br_startoff - got->br_startoff; + got_indlen = xfs_bmap_worst_indlen(ip, got->br_blockcount); + + new.br_blockcount = got_endoff - del_endoff; + new_indlen = xfs_bmap_worst_indlen(ip, new.br_blockcount); + + WARN_ON_ONCE(!got_indlen || !new_indlen); + stolen = xfs_bmap_split_indlen(da_old, &got_indlen, &new_indlen, + del->br_blockcount); + + got->br_startblock = nullstartblock((int)got_indlen); + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, 0, _THIS_IP_); + + new.br_startoff = del_endoff; + new.br_state = got->br_state; + new.br_startblock = nullstartblock((int)new_indlen); + + ++*idx; + xfs_iext_insert(ip, *idx, 1, &new, state); + + da_new = got_indlen + new_indlen - stolen; + del->br_blockcount -= stolen; + break; + } + + ASSERT(da_old >= da_new); + da_diff = da_old - da_new; + if (!isrt) + da_diff += del->br_blockcount; + if (da_diff) + xfs_mod_fdblocks(mp, da_diff, false); + return error; +} + +void +xfs_bmap_del_extent_cow( + struct xfs_inode *ip, + xfs_extnum_t *idx, + struct xfs_bmbt_irec *got, + struct xfs_bmbt_irec *del) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + struct xfs_bmbt_irec new; + xfs_fileoff_t del_endoff, got_endoff; + int state = BMAP_COWFORK; + + XFS_STATS_INC(mp, xs_del_exlist); + + del_endoff = del->br_startoff + del->br_blockcount; + got_endoff = got->br_startoff + got->br_blockcount; + + ASSERT(*idx >= 0); + ASSERT(*idx <= xfs_iext_count(ifp)); + ASSERT(del->br_blockcount > 0); + ASSERT(got->br_startoff <= del->br_startoff); + ASSERT(got_endoff >= del_endoff); + ASSERT(!isnullstartblock(got->br_startblock)); + + if (got->br_startoff == del->br_startoff) + state |= BMAP_LEFT_CONTIG; + if (got_endoff == del_endoff) + state |= BMAP_RIGHT_CONTIG; + + switch (state & (BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG)) { + case BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG: + /* + * Matches the whole extent. Delete the entry. + */ + xfs_iext_remove(ip, *idx, 1, state); + --*idx; + break; + case BMAP_LEFT_CONTIG: + /* + * Deleting the first part of the extent. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + got->br_startoff = del_endoff; + got->br_blockcount -= del->br_blockcount; + got->br_startblock = del->br_startblock + del->br_blockcount; + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + break; + case BMAP_RIGHT_CONTIG: + /* + * Deleting the last part of the extent. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + got->br_blockcount -= del->br_blockcount; + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + break; + case 0: + /* + * Deleting the middle of the extent. + */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); + got->br_blockcount = del->br_startoff - got->br_startoff; + xfs_bmbt_set_all(xfs_iext_get_ext(ifp, *idx), got); + trace_xfs_bmap_post_update(ip, *idx, state, _THIS_IP_); + + new.br_startoff = del_endoff; + new.br_blockcount = got_endoff - del_endoff; + new.br_state = got->br_state; + new.br_startblock = del->br_startblock + del->br_blockcount; + + ++*idx; + xfs_iext_insert(ip, *idx, 1, &new, state); + break; + } +} + /* * Called by xfs_bmapi to update file extent records and the btree * after removing space (or undoing a delayed allocation). @@ -4768,7 +5036,8 @@ xfs_bmap_del_extent( xfs_btree_cur_t *cur, /* if null, not a btree */ xfs_bmbt_irec_t *del, /* data to remove from extents */ int *logflagsp, /* inode logging flags */ - int whichfork) /* data or attr fork */ + int whichfork, /* data or attr fork */ + int bflags) /* bmapi flags */ { xfs_filblks_t da_new; /* new delay-alloc indirect blocks */ xfs_filblks_t da_old; /* old delay-alloc indirect blocks */ @@ -4797,10 +5066,11 @@ xfs_bmap_del_extent( if (whichfork == XFS_ATTR_FORK) state |= BMAP_ATTRFORK; + else if (whichfork == XFS_COW_FORK) + state |= BMAP_COWFORK; ifp = XFS_IFORK_PTR(ip, whichfork); - ASSERT((*idx >= 0) && (*idx < ifp->if_bytes / - (uint)sizeof(xfs_bmbt_rec_t))); + ASSERT((*idx >= 0) && (*idx < xfs_iext_count(ifp))); ASSERT(del->br_blockcount > 0); ep = xfs_iext_get_ext(ifp, *idx); xfs_bmbt_get_all(ep, &got); @@ -4877,6 +5147,7 @@ xfs_bmap_del_extent( /* * Matches the whole extent. Delete the entry. */ + trace_xfs_bmap_pre_update(ip, *idx, state, _THIS_IP_); xfs_iext_remove(ip, *idx, 1, whichfork == XFS_ATTR_FORK ? BMAP_ATTRFORK : 0); --*idx; @@ -5060,9 +5331,16 @@ xfs_bmap_del_extent( /* * If we need to, add to list of extents to delete. */ - if (do_fx) - xfs_bmap_add_free(mp, dfops, del->br_startblock, - del->br_blockcount, NULL); + if (do_fx && !(bflags & XFS_BMAPI_REMAP)) { + if (xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK) { + error = xfs_refcount_decrease_extent(mp, dfops, del); + if (error) + goto done; + } else + xfs_bmap_add_free(mp, dfops, del->br_startblock, + del->br_blockcount, NULL); + } + /* * Adjust inode # blocks in the file. */ @@ -5071,7 +5349,7 @@ xfs_bmap_del_extent( /* * Adjust quota data. */ - if (qfield) + if (qfield && !(bflags & XFS_BMAPI_REMAP)) xfs_trans_mod_dquot_byino(tp, ip, qfield, (long)-nblks); /* @@ -5093,22 +5371,19 @@ done: * *done is set. */ int /* error */ -xfs_bunmapi( +__xfs_bunmapi( xfs_trans_t *tp, /* transaction pointer */ struct xfs_inode *ip, /* incore inode */ xfs_fileoff_t bno, /* starting offset to unmap */ - xfs_filblks_t len, /* length to unmap in file */ + xfs_filblks_t *rlen, /* i/o: amount remaining */ int flags, /* misc flags */ xfs_extnum_t nexts, /* number of extents max */ xfs_fsblock_t *firstblock, /* first allocated block controls a.g. for allocs */ - struct xfs_defer_ops *dfops, /* i/o: list extents to free */ - int *done) /* set if not done yet */ + struct xfs_defer_ops *dfops) /* i/o: deferred updates */ { xfs_btree_cur_t *cur; /* bmap btree cursor */ xfs_bmbt_irec_t del; /* extent being deleted */ - int eof; /* is deleting at eof */ - xfs_bmbt_rec_host_t *ep; /* extent record pointer */ int error; /* error return value */ xfs_extnum_t extno; /* extent number in list */ xfs_bmbt_irec_t got; /* current extent record */ @@ -5118,18 +5393,17 @@ xfs_bunmapi( int logflags; /* transaction logging flags */ xfs_extlen_t mod; /* rt extent offset */ xfs_mount_t *mp; /* mount structure */ - xfs_extnum_t nextents; /* number of file extents */ - xfs_bmbt_irec_t prev; /* previous extent record */ xfs_fileoff_t start; /* first file offset deleted */ int tmp_logflags; /* partial logging flags */ int wasdel; /* was a delayed alloc extent */ int whichfork; /* data or attribute fork */ xfs_fsblock_t sum; + xfs_filblks_t len = *rlen; /* length to unmap in file */ trace_xfs_bunmap(ip, bno, len, flags, _RET_IP_); - whichfork = (flags & XFS_BMAPI_ATTRFORK) ? - XFS_ATTR_FORK : XFS_DATA_FORK; + whichfork = xfs_bmapi_whichfork(flags); + ASSERT(whichfork != XFS_COW_FORK); ifp = XFS_IFORK_PTR(ip, whichfork); if (unlikely( XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS && @@ -5149,27 +5423,25 @@ xfs_bunmapi( if (!(ifp->if_flags & XFS_IFEXTENTS) && (error = xfs_iread_extents(tp, ip, whichfork))) return error; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - if (nextents == 0) { - *done = 1; + if (xfs_iext_count(ifp) == 0) { + *rlen = 0; return 0; } XFS_STATS_INC(mp, xs_blk_unmap); isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip); start = bno; bno = start + len - 1; - ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got, - &prev); /* * Check to see if the given block number is past the end of the * file, back up to the last block if so... */ - if (eof) { - ep = xfs_iext_get_ext(ifp, --lastx); - xfs_bmbt_get_all(ep, &got); + if (!xfs_iext_lookup_extent(ip, ifp, bno, &lastx, &got)) { + ASSERT(lastx > 0); + xfs_iext_get_extent(ifp, --lastx, &got); bno = got.br_startoff + got.br_blockcount - 1; } + logflags = 0; if (ifp->if_flags & XFS_IFBROOT) { ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE); @@ -5200,8 +5472,7 @@ xfs_bunmapi( if (got.br_startoff > bno) { if (--lastx < 0) break; - ep = xfs_iext_get_ext(ifp, lastx); - xfs_bmbt_get_all(ep, &got); + xfs_iext_get_extent(ifp, lastx, &got); } /* * Is the last block of this extent before the range @@ -5215,7 +5486,6 @@ xfs_bunmapi( * Then deal with the (possibly delayed) allocated space * we found. */ - ASSERT(ep != NULL); del = got; wasdel = isnullstartblock(del.br_startblock); if (got.br_startoff < start) { @@ -5296,15 +5566,12 @@ xfs_bunmapi( */ ASSERT(bno >= del.br_blockcount); bno -= del.br_blockcount; - if (got.br_startoff > bno) { - if (--lastx >= 0) { - ep = xfs_iext_get_ext(ifp, - lastx); - xfs_bmbt_get_all(ep, &got); - } - } + if (got.br_startoff > bno && --lastx >= 0) + xfs_iext_get_extent(ifp, lastx, &got); continue; } else if (del.br_state == XFS_EXT_UNWRITTEN) { + struct xfs_bmbt_irec prev; + /* * This one is already unwritten. * It must have a written left neighbor. @@ -5312,8 +5579,7 @@ xfs_bunmapi( * try again. */ ASSERT(lastx > 0); - xfs_bmbt_get_all(xfs_iext_get_ext(ifp, - lastx - 1), &prev); + xfs_iext_get_extent(ifp, lastx - 1, &prev); ASSERT(prev.br_state == XFS_EXT_NORM); ASSERT(!isnullstartblock(prev.br_startblock)); ASSERT(del.br_startblock == @@ -5396,7 +5662,7 @@ xfs_bunmapi( cur->bc_private.b.flags &= ~XFS_BTCUR_BPRV_WASDEL; error = xfs_bmap_del_extent(ip, tp, &lastx, dfops, cur, &del, - &tmp_logflags, whichfork); + &tmp_logflags, whichfork, flags); logflags |= tmp_logflags; if (error) goto error0; @@ -5411,18 +5677,17 @@ nodelete: */ if (bno != (xfs_fileoff_t)-1 && bno >= start) { if (lastx >= 0) { - ep = xfs_iext_get_ext(ifp, lastx); - if (xfs_bmbt_get_startoff(ep) > bno) { - if (--lastx >= 0) - ep = xfs_iext_get_ext(ifp, - lastx); - } - xfs_bmbt_get_all(ep, &got); + xfs_iext_get_extent(ifp, lastx, &got); + if (got.br_startoff > bno && --lastx >= 0) + xfs_iext_get_extent(ifp, lastx, &got); } extno++; } } - *done = bno == (xfs_fileoff_t)-1 || bno < start || lastx < 0; + if (bno == (xfs_fileoff_t)-1 || bno < start || lastx < 0) + *rlen = 0; + else + *rlen = bno - start + 1; /* * Convert to a btree if necessary. @@ -5478,6 +5743,27 @@ error0: return error; } +/* Unmap a range of a file. */ +int +xfs_bunmapi( + xfs_trans_t *tp, + struct xfs_inode *ip, + xfs_fileoff_t bno, + xfs_filblks_t len, + int flags, + xfs_extnum_t nexts, + xfs_fsblock_t *firstblock, + struct xfs_defer_ops *dfops, + int *done) +{ + int error; + + error = __xfs_bunmapi(tp, ip, bno, &len, flags, nexts, firstblock, + dfops); + *done = (len == 0); + return error; +} + /* * Determine whether an extent shift can be accomplished by a merge with the * extent that precedes the target hole of the shift. @@ -5611,7 +5897,7 @@ xfs_bmse_shift_one( mp = ip->i_mount; ifp = XFS_IFORK_PTR(ip, whichfork); - total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t); + total_extents = xfs_iext_count(ifp); xfs_bmbt_get_all(gotp, &got); @@ -5788,7 +6074,7 @@ xfs_bmap_shift_extents( * are collapsing out, so we cannot use the count of real extents here. * Instead we have to calculate it from the incore fork. */ - total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t); + total_extents = xfs_iext_count(ifp); if (total_extents == 0) { *done = 1; goto del_cursor; @@ -5848,7 +6134,7 @@ xfs_bmap_shift_extents( * count can change. Update the total and grade the next record. */ if (direction == SHIFT_LEFT) { - total_extents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t); + total_extents = xfs_iext_count(ifp); stop_extent = total_extents; } @@ -6057,3 +6343,146 @@ out: xfs_trans_cancel(tp); return error; } + +/* Deferred mapping is only for real extents in the data fork. */ +static bool +xfs_bmap_is_update_needed( + struct xfs_bmbt_irec *bmap) +{ + return bmap->br_startblock != HOLESTARTBLOCK && + bmap->br_startblock != DELAYSTARTBLOCK; +} + +/* Record a bmap intent. */ +static int +__xfs_bmap_add( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + enum xfs_bmap_intent_type type, + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *bmap) +{ + int error; + struct xfs_bmap_intent *bi; + + trace_xfs_bmap_defer(mp, + XFS_FSB_TO_AGNO(mp, bmap->br_startblock), + type, + XFS_FSB_TO_AGBNO(mp, bmap->br_startblock), + ip->i_ino, whichfork, + bmap->br_startoff, + bmap->br_blockcount, + bmap->br_state); + + bi = kmem_alloc(sizeof(struct xfs_bmap_intent), KM_SLEEP | KM_NOFS); + INIT_LIST_HEAD(&bi->bi_list); + bi->bi_type = type; + bi->bi_owner = ip; + bi->bi_whichfork = whichfork; + bi->bi_bmap = *bmap; + + error = xfs_defer_join(dfops, bi->bi_owner); + if (error) { + kmem_free(bi); + return error; + } + + xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_BMAP, &bi->bi_list); + return 0; +} + +/* Map an extent into a file. */ +int +xfs_bmap_map_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_bmap_is_update_needed(PREV)) + return 0; + + return __xfs_bmap_add(mp, dfops, XFS_BMAP_MAP, ip, + XFS_DATA_FORK, PREV); +} + +/* Unmap an extent out of a file. */ +int +xfs_bmap_unmap_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_bmap_is_update_needed(PREV)) + return 0; + + return __xfs_bmap_add(mp, dfops, XFS_BMAP_UNMAP, ip, + XFS_DATA_FORK, PREV); +} + +/* + * Process one of the deferred bmap operations. We pass back the + * btree cursor to maintain our lock on the bmapbt between calls. + */ +int +xfs_bmap_finish_one( + struct xfs_trans *tp, + struct xfs_defer_ops *dfops, + struct xfs_inode *ip, + enum xfs_bmap_intent_type type, + int whichfork, + xfs_fileoff_t startoff, + xfs_fsblock_t startblock, + xfs_filblks_t blockcount, + xfs_exntst_t state) +{ + struct xfs_bmbt_irec bmap; + int nimaps = 1; + xfs_fsblock_t firstfsb; + int flags = XFS_BMAPI_REMAP; + int done; + int error = 0; + + bmap.br_startblock = startblock; + bmap.br_startoff = startoff; + bmap.br_blockcount = blockcount; + bmap.br_state = state; + + trace_xfs_bmap_deferred(tp->t_mountp, + XFS_FSB_TO_AGNO(tp->t_mountp, startblock), type, + XFS_FSB_TO_AGBNO(tp->t_mountp, startblock), + ip->i_ino, whichfork, startoff, blockcount, state); + + if (whichfork != XFS_DATA_FORK && whichfork != XFS_ATTR_FORK) + return -EFSCORRUPTED; + if (whichfork == XFS_ATTR_FORK) + flags |= XFS_BMAPI_ATTRFORK; + + if (XFS_TEST_ERROR(false, tp->t_mountp, + XFS_ERRTAG_BMAP_FINISH_ONE, + XFS_RANDOM_BMAP_FINISH_ONE)) + return -EIO; + + switch (type) { + case XFS_BMAP_MAP: + firstfsb = bmap.br_startblock; + error = xfs_bmapi_write(tp, ip, bmap.br_startoff, + bmap.br_blockcount, flags, &firstfsb, + bmap.br_blockcount, &bmap, &nimaps, + dfops); + break; + case XFS_BMAP_UNMAP: + error = xfs_bunmapi(tp, ip, bmap.br_startoff, + bmap.br_blockcount, flags, 1, &firstfsb, + dfops, &done); + ASSERT(done); + break; + default: + ASSERT(0); + error = -EFSCORRUPTED; + } + + return error; +} diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index 254034f96941..cecd094404cc 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -54,7 +54,7 @@ struct xfs_bmalloca { bool wasdel; /* replacing a delayed allocation */ bool aeof; /* allocated space at eof */ bool conv; /* overwriting unwritten extents */ - char userdata;/* userdata mask */ + int datatype;/* data type being allocated */ int flags; }; @@ -97,6 +97,19 @@ struct xfs_extent_free_item */ #define XFS_BMAPI_ZERO 0x080 +/* + * Map the inode offset to the block given in ap->firstblock. Primarily + * used for reflink. The range must be in a hole, and this flag cannot be + * turned on with PREALLOC or CONVERT, and cannot be used on the attr fork. + * + * For bunmapi, this flag unmaps the range without adjusting quota, reducing + * refcount, or freeing the blocks. + */ +#define XFS_BMAPI_REMAP 0x100 + +/* Map something in the CoW fork. */ +#define XFS_BMAPI_COWFORK 0x200 + #define XFS_BMAPI_FLAGS \ { XFS_BMAPI_ENTIRE, "ENTIRE" }, \ { XFS_BMAPI_METADATA, "METADATA" }, \ @@ -105,12 +118,24 @@ struct xfs_extent_free_item { XFS_BMAPI_IGSTATE, "IGSTATE" }, \ { XFS_BMAPI_CONTIG, "CONTIG" }, \ { XFS_BMAPI_CONVERT, "CONVERT" }, \ - { XFS_BMAPI_ZERO, "ZERO" } + { XFS_BMAPI_ZERO, "ZERO" }, \ + { XFS_BMAPI_REMAP, "REMAP" }, \ + { XFS_BMAPI_COWFORK, "COWFORK" } static inline int xfs_bmapi_aflag(int w) { - return (w == XFS_ATTR_FORK ? XFS_BMAPI_ATTRFORK : 0); + return (w == XFS_ATTR_FORK ? XFS_BMAPI_ATTRFORK : + (w == XFS_COW_FORK ? XFS_BMAPI_COWFORK : 0)); +} + +static inline int xfs_bmapi_whichfork(int bmapi_flags) +{ + if (bmapi_flags & XFS_BMAPI_COWFORK) + return XFS_COW_FORK; + else if (bmapi_flags & XFS_BMAPI_ATTRFORK) + return XFS_ATTR_FORK; + return XFS_DATA_FORK; } /* @@ -131,13 +156,15 @@ static inline int xfs_bmapi_aflag(int w) #define BMAP_LEFT_VALID (1 << 6) #define BMAP_RIGHT_VALID (1 << 7) #define BMAP_ATTRFORK (1 << 8) +#define BMAP_COWFORK (1 << 9) #define XFS_BMAP_EXT_FLAGS \ { BMAP_LEFT_CONTIG, "LC" }, \ { BMAP_RIGHT_CONTIG, "RC" }, \ { BMAP_LEFT_FILLING, "LF" }, \ { BMAP_RIGHT_FILLING, "RF" }, \ - { BMAP_ATTRFORK, "ATTR" } + { BMAP_ATTRFORK, "ATTR" }, \ + { BMAP_COWFORK, "COW" } /* @@ -163,6 +190,8 @@ void xfs_bmap_trace_exlist(struct xfs_inode *ip, xfs_extnum_t cnt, #define XFS_BMAP_TRACE_EXLIST(ip,c,w) #endif +void xfs_trim_extent(struct xfs_bmbt_irec *irec, xfs_fileoff_t bno, + xfs_filblks_t len); int xfs_bmap_add_attrfork(struct xfs_inode *ip, int size, int rsvd); void xfs_bmap_local_to_extents_empty(struct xfs_inode *ip, int whichfork); void xfs_bmap_add_free(struct xfs_mount *mp, struct xfs_defer_ops *dfops, @@ -181,18 +210,24 @@ int xfs_bmap_read_extents(struct xfs_trans *tp, struct xfs_inode *ip, int xfs_bmapi_read(struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, struct xfs_bmbt_irec *mval, int *nmap, int flags); -int xfs_bmapi_delay(struct xfs_inode *ip, xfs_fileoff_t bno, - xfs_filblks_t len, struct xfs_bmbt_irec *mval, - int *nmap, int flags); int xfs_bmapi_write(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_fsblock_t *firstblock, xfs_extlen_t total, struct xfs_bmbt_irec *mval, int *nmap, struct xfs_defer_ops *dfops); +int __xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_fileoff_t bno, xfs_filblks_t *rlen, int flags, + xfs_extnum_t nexts, xfs_fsblock_t *firstblock, + struct xfs_defer_ops *dfops); int xfs_bunmapi(struct xfs_trans *tp, struct xfs_inode *ip, xfs_fileoff_t bno, xfs_filblks_t len, int flags, xfs_extnum_t nexts, xfs_fsblock_t *firstblock, struct xfs_defer_ops *dfops, int *done); +int xfs_bmap_del_extent_delay(struct xfs_inode *ip, int whichfork, + xfs_extnum_t *idx, struct xfs_bmbt_irec *got, + struct xfs_bmbt_irec *del); +void xfs_bmap_del_extent_cow(struct xfs_inode *ip, xfs_extnum_t *idx, + struct xfs_bmbt_irec *got, struct xfs_bmbt_irec *del); int xfs_check_nostate_extents(struct xfs_ifork *ifp, xfs_extnum_t idx, xfs_extnum_t num); uint xfs_default_attroffset(struct xfs_inode *ip); @@ -202,5 +237,30 @@ int xfs_bmap_shift_extents(struct xfs_trans *tp, struct xfs_inode *ip, struct xfs_defer_ops *dfops, enum shift_direction direction, int num_exts); int xfs_bmap_split_extent(struct xfs_inode *ip, xfs_fileoff_t split_offset); +int xfs_bmapi_reserve_delalloc(struct xfs_inode *ip, int whichfork, + xfs_fileoff_t off, xfs_filblks_t len, xfs_filblks_t prealloc, + struct xfs_bmbt_irec *got, xfs_extnum_t *lastx, int eof); + +enum xfs_bmap_intent_type { + XFS_BMAP_MAP = 1, + XFS_BMAP_UNMAP, +}; + +struct xfs_bmap_intent { + struct list_head bi_list; + enum xfs_bmap_intent_type bi_type; + struct xfs_inode *bi_owner; + int bi_whichfork; + struct xfs_bmbt_irec bi_bmap; +}; + +int xfs_bmap_finish_one(struct xfs_trans *tp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, enum xfs_bmap_intent_type type, + int whichfork, xfs_fileoff_t startoff, xfs_fsblock_t startblock, + xfs_filblks_t blockcount, xfs_exntst_t state); +int xfs_bmap_map_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, struct xfs_bmbt_irec *imap); +int xfs_bmap_unmap_extent(struct xfs_mount *mp, struct xfs_defer_ops *dfops, + struct xfs_inode *ip, struct xfs_bmbt_irec *imap); #endif /* __XFS_BMAP_H__ */ diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index cd85274e810c..d6330c297ca0 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -453,6 +453,7 @@ xfs_bmbt_alloc_block( if (args.fsbno == NULLFSBLOCK) { args.fsbno = be64_to_cpu(start->l); +try_another_ag: args.type = XFS_ALLOCTYPE_START_BNO; /* * Make sure there is sufficient room left in the AG to @@ -482,6 +483,22 @@ xfs_bmbt_alloc_block( if (error) goto error0; + /* + * During a CoW operation, the allocation and bmbt updates occur in + * different transactions. The mapping code tries to put new bmbt + * blocks near extents being mapped, but the only way to guarantee this + * is if the alloc and the mapping happen in a single transaction that + * has a block reservation. That isn't the case here, so if we run out + * of space we'll try again with another AG. + */ + if (xfs_sb_version_hasreflink(&cur->bc_mp->m_sb) && + args.fsbno == NULLFSBLOCK && + args.type == XFS_ALLOCTYPE_NEAR_BNO) { + cur->bc_private.b.dfops->dop_low = true; + args.fsbno = cur->bc_private.b.firstblock; + goto try_another_ag; + } + if (args.fsbno == NULLFSBLOCK && args.minleft) { /* * Could not find an AG with enough free space to satisfy @@ -777,14 +794,16 @@ xfs_bmbt_init_cursor( { struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork); struct xfs_btree_cur *cur; + ASSERT(whichfork != XFS_COW_FORK); - cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP); + cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); cur->bc_tp = tp; cur->bc_mp = mp; cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1; cur->bc_btnum = XFS_BTNUM_BMAP; cur->bc_blocklog = mp->m_sb.sb_blocklog; + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2); cur->bc_ops = &xfs_bmbt_ops; cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE; diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index 08569792fe20..21e6a6ab6b9a 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -45,9 +45,10 @@ kmem_zone_t *xfs_btree_cur_zone; */ static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = { { XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, 0, XFS_BMAP_MAGIC, XFS_IBT_MAGIC, - XFS_FIBT_MAGIC }, + XFS_FIBT_MAGIC, 0 }, { XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, XFS_RMAP_CRC_MAGIC, - XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC } + XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC, + XFS_REFC_CRC_MAGIC } }; #define xfs_btree_magic(cur) \ xfs_magics[!!((cur)->bc_flags & XFS_BTREE_CRC_BLOCKS)][cur->bc_btnum] @@ -1216,6 +1217,9 @@ xfs_btree_set_refs( case XFS_BTNUM_RMAP: xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF); break; + case XFS_BTNUM_REFC: + xfs_buf_set_ref(bp, XFS_REFC_BTREE_REF); + break; default: ASSERT(0); } @@ -1765,8 +1769,28 @@ xfs_btree_lookup_get_block( if (error) return error; + /* Check the inode owner since the verifiers don't. */ + if (xfs_sb_version_hascrc(&cur->bc_mp->m_sb) && + (cur->bc_flags & XFS_BTREE_LONG_PTRS) && + be64_to_cpu((*blkp)->bb_u.l.bb_owner) != + cur->bc_private.b.ip->i_ino) + goto out_bad; + + /* Did we get the level we were looking for? */ + if (be16_to_cpu((*blkp)->bb_level) != level) + goto out_bad; + + /* Check that internal nodes have at least one record. */ + if (level != 0 && be16_to_cpu((*blkp)->bb_numrecs) == 0) + goto out_bad; + xfs_btree_setbuf(cur, level, bp); return 0; + +out_bad: + *blkp = NULL; + xfs_trans_brelse(cur->bc_tp, bp); + return -EFSCORRUPTED; } /* @@ -2070,7 +2094,7 @@ __xfs_btree_updkeys( struct xfs_buf *bp0, bool force_all) { - union xfs_btree_bigkey key; /* keys from current level */ + union xfs_btree_key key; /* keys from current level */ union xfs_btree_key *lkey; /* keys from the next level up */ union xfs_btree_key *hkey; union xfs_btree_key *nlkey; /* keys from the next level up */ @@ -2086,7 +2110,7 @@ __xfs_btree_updkeys( trace_xfs_btree_updkeys(cur, level, bp0); - lkey = (union xfs_btree_key *)&key; + lkey = &key; hkey = xfs_btree_high_key_from_key(cur, lkey); xfs_btree_get_keys(cur, block, lkey); for (level++; level < cur->bc_nlevels; level++) { @@ -3226,7 +3250,7 @@ xfs_btree_insrec( struct xfs_buf *bp; /* buffer for block */ union xfs_btree_ptr nptr; /* new block ptr */ struct xfs_btree_cur *ncur; /* new btree cursor */ - union xfs_btree_bigkey nkey; /* new block key */ + union xfs_btree_key nkey; /* new block key */ union xfs_btree_key *lkey; int optr; /* old key/record index */ int ptr; /* key/record index */ @@ -3241,7 +3265,7 @@ xfs_btree_insrec( XFS_BTREE_TRACE_ARGIPR(cur, level, *ptrp, &rec); ncur = NULL; - lkey = (union xfs_btree_key *)&nkey; + lkey = &nkey; /* * If we have an external root pointer, and we've made it to the @@ -3444,14 +3468,14 @@ xfs_btree_insert( union xfs_btree_ptr nptr; /* new block number (split result) */ struct xfs_btree_cur *ncur; /* new cursor (split result) */ struct xfs_btree_cur *pcur; /* previous level's cursor */ - union xfs_btree_bigkey bkey; /* key of block to insert */ + union xfs_btree_key bkey; /* key of block to insert */ union xfs_btree_key *key; union xfs_btree_rec rec; /* record to insert */ level = 0; ncur = NULL; pcur = cur; - key = (union xfs_btree_key *)&bkey; + key = &bkey; xfs_btree_set_ptr_null(cur, &nptr); @@ -4797,3 +4821,50 @@ xfs_btree_query_range( return xfs_btree_overlapped_query_range(cur, &low_key, &high_key, fn, priv); } + +/* + * Calculate the number of blocks needed to store a given number of records + * in a short-format (per-AG metadata) btree. + */ +xfs_extlen_t +xfs_btree_calc_size( + struct xfs_mount *mp, + uint *limits, + unsigned long long len) +{ + int level; + int maxrecs; + xfs_extlen_t rval; + + maxrecs = limits[0]; + for (level = 0, rval = 0; len > 1; level++) { + len += maxrecs - 1; + do_div(len, maxrecs); + maxrecs = limits[1]; + rval += len; + } + return rval; +} + +static int +xfs_btree_count_blocks_helper( + struct xfs_btree_cur *cur, + int level, + void *data) +{ + xfs_extlen_t *blocks = data; + (*blocks)++; + + return 0; +} + +/* Count the blocks in a btree and return the result in *blocks. */ +int +xfs_btree_count_blocks( + struct xfs_btree_cur *cur, + xfs_extlen_t *blocks) +{ + *blocks = 0; + return xfs_btree_visit_blocks(cur, xfs_btree_count_blocks_helper, + blocks); +} diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index 04d0865e5e6d..b69b947c4c1b 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -37,30 +37,19 @@ union xfs_btree_ptr { __be64 l; /* long form ptr */ }; -union xfs_btree_key { - struct xfs_bmbt_key bmbt; - xfs_bmdr_key_t bmbr; /* bmbt root block */ - xfs_alloc_key_t alloc; - struct xfs_inobt_key inobt; - struct xfs_rmap_key rmap; -}; - /* - * In-core key that holds both low and high keys for overlapped btrees. - * The two keys are packed next to each other on disk, so do the same - * in memory. Preserve the existing xfs_btree_key as a single key to - * avoid the mental model breakage that would happen if we passed a - * bigkey into a function that operates on a single key. + * The in-core btree key. Overlapping btrees actually store two keys + * per pointer, so we reserve enough memory to hold both. The __*bigkey + * items should never be accessed directly. */ -union xfs_btree_bigkey { +union xfs_btree_key { struct xfs_bmbt_key bmbt; xfs_bmdr_key_t bmbr; /* bmbt root block */ xfs_alloc_key_t alloc; struct xfs_inobt_key inobt; - struct { - struct xfs_rmap_key rmap; - struct xfs_rmap_key rmap_hi; - }; + struct xfs_rmap_key rmap; + struct xfs_rmap_key __rmap_bigkey[2]; + struct xfs_refcount_key refc; }; union xfs_btree_rec { @@ -69,6 +58,7 @@ union xfs_btree_rec { struct xfs_alloc_rec alloc; struct xfs_inobt_rec inobt; struct xfs_rmap_rec rmap; + struct xfs_refcount_rec refc; }; /* @@ -84,6 +74,7 @@ union xfs_btree_rec { #define XFS_BTNUM_INO ((xfs_btnum_t)XFS_BTNUM_INOi) #define XFS_BTNUM_FINO ((xfs_btnum_t)XFS_BTNUM_FINOi) #define XFS_BTNUM_RMAP ((xfs_btnum_t)XFS_BTNUM_RMAPi) +#define XFS_BTNUM_REFC ((xfs_btnum_t)XFS_BTNUM_REFCi) /* * For logging record fields. @@ -105,43 +96,10 @@ union xfs_btree_rec { /* * Generic stats interface */ -#define __XFS_BTREE_STATS_INC(mp, type, stat) \ - XFS_STATS_INC(mp, xs_ ## type ## _2_ ## stat) #define XFS_BTREE_STATS_INC(cur, stat) \ -do { \ - struct xfs_mount *__mp = cur->bc_mp; \ - switch (cur->bc_btnum) { \ - case XFS_BTNUM_BNO: __XFS_BTREE_STATS_INC(__mp, abtb, stat); break; \ - case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(__mp, abtc, stat); break; \ - case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(__mp, bmbt, stat); break; \ - case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \ - case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \ - case XFS_BTNUM_RMAP: __XFS_BTREE_STATS_INC(__mp, rmap, stat); break; \ - case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ - } \ -} while (0) - -#define __XFS_BTREE_STATS_ADD(mp, type, stat, val) \ - XFS_STATS_ADD(mp, xs_ ## type ## _2_ ## stat, val) -#define XFS_BTREE_STATS_ADD(cur, stat, val) \ -do { \ - struct xfs_mount *__mp = cur->bc_mp; \ - switch (cur->bc_btnum) { \ - case XFS_BTNUM_BNO: \ - __XFS_BTREE_STATS_ADD(__mp, abtb, stat, val); break; \ - case XFS_BTNUM_CNT: \ - __XFS_BTREE_STATS_ADD(__mp, abtc, stat, val); break; \ - case XFS_BTNUM_BMAP: \ - __XFS_BTREE_STATS_ADD(__mp, bmbt, stat, val); break; \ - case XFS_BTNUM_INO: \ - __XFS_BTREE_STATS_ADD(__mp, ibt, stat, val); break; \ - case XFS_BTNUM_FINO: \ - __XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \ - case XFS_BTNUM_RMAP: \ - __XFS_BTREE_STATS_ADD(__mp, rmap, stat, val); break; \ - case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \ - } \ -} while (0) + XFS_STATS_INC_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat) +#define XFS_BTREE_STATS_ADD(cur, stat, val) \ + XFS_STATS_ADD_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat, val) #define XFS_BTREE_MAXLEVELS 9 /* max of all btrees */ @@ -229,6 +187,15 @@ union xfs_btree_irec { struct xfs_bmbt_irec b; struct xfs_inobt_rec_incore i; struct xfs_rmap_irec r; + struct xfs_refcount_irec rc; +}; + +/* Per-AG btree private information. */ +union xfs_btree_cur_private { + struct { + unsigned long nr_ops; /* # record updates */ + int shape_changes; /* # of extent splits */ + } refc; }; /* @@ -250,11 +217,13 @@ typedef struct xfs_btree_cur __uint8_t bc_nlevels; /* number of levels in the tree */ __uint8_t bc_blocklog; /* log2(blocksize) of btree blocks */ xfs_btnum_t bc_btnum; /* identifies which btree type */ + int bc_statoff; /* offset of btre stats array */ union { struct { /* needed for BNO, CNT, INO */ struct xfs_buf *agbp; /* agf/agi buffer pointer */ struct xfs_defer_ops *dfops; /* deferred updates */ xfs_agnumber_t agno; /* ag number */ + union xfs_btree_cur_private priv; } a; struct { /* needed for BMAP */ struct xfs_inode *ip; /* pointer to our inode */ @@ -513,6 +482,8 @@ bool xfs_btree_sblock_v5hdr_verify(struct xfs_buf *bp); bool xfs_btree_sblock_verify(struct xfs_buf *bp, unsigned int max_recs); uint xfs_btree_compute_maxlevels(struct xfs_mount *mp, uint *limits, unsigned long len); +xfs_extlen_t xfs_btree_calc_size(struct xfs_mount *mp, uint *limits, + unsigned long long len); /* return codes */ #define XFS_BTREE_QUERY_RANGE_CONTINUE 0 /* keep iterating */ @@ -529,4 +500,6 @@ typedef int (*xfs_btree_visit_blocks_fn)(struct xfs_btree_cur *cur, int level, int xfs_btree_visit_blocks(struct xfs_btree_cur *cur, xfs_btree_visit_blocks_fn fn, void *data); +int xfs_btree_count_blocks(struct xfs_btree_cur *cur, xfs_extlen_t *blocks); + #endif /* __XFS_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_cksum.h b/fs/xfs/libxfs/xfs_cksum.h index fad1676ad8cd..a416c7cb23ea 100644 --- a/fs/xfs/libxfs/xfs_cksum.h +++ b/fs/xfs/libxfs/xfs_cksum.h @@ -6,10 +6,11 @@ /* * Calculate the intermediate checksum for a buffer that has the CRC field * inside it. The offset of the 32bit crc fields is passed as the - * cksum_offset parameter. + * cksum_offset parameter. We do not modify the buffer during verification, + * hence we have to split the CRC calculation across the cksum_offset. */ static inline __uint32_t -xfs_start_cksum(char *buffer, size_t length, unsigned long cksum_offset) +xfs_start_cksum_safe(char *buffer, size_t length, unsigned long cksum_offset) { __uint32_t zero = 0; __uint32_t crc; @@ -26,6 +27,20 @@ xfs_start_cksum(char *buffer, size_t length, unsigned long cksum_offset) } /* + * Fast CRC method where the buffer is modified. Callers must have exclusive + * access to the buffer while the calculation takes place. + */ +static inline __uint32_t +xfs_start_cksum_update(char *buffer, size_t length, unsigned long cksum_offset) +{ + /* zero the CRC field */ + *(__le32 *)(buffer + cksum_offset) = 0; + + /* single pass CRC calculation for the entire buffer */ + return crc32c(XFS_CRC_SEED, buffer, length); +} + +/* * Convert the intermediate checksum to the final ondisk format. * * The CRC32c calculation uses LE format even on BE machines, but returns the @@ -40,11 +55,14 @@ xfs_end_cksum(__uint32_t crc) /* * Helper to generate the checksum for a buffer. + * + * This modifies the buffer temporarily - callers must have exclusive + * access to the buffer while the calculation takes place. */ static inline void xfs_update_cksum(char *buffer, size_t length, unsigned long cksum_offset) { - __uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset); + __uint32_t crc = xfs_start_cksum_update(buffer, length, cksum_offset); *(__le32 *)(buffer + cksum_offset) = xfs_end_cksum(crc); } @@ -55,7 +73,7 @@ xfs_update_cksum(char *buffer, size_t length, unsigned long cksum_offset) static inline int xfs_verify_cksum(char *buffer, size_t length, unsigned long cksum_offset) { - __uint32_t crc = xfs_start_cksum(buffer, length, cksum_offset); + __uint32_t crc = xfs_start_cksum_safe(buffer, length, cksum_offset); return *(__le32 *)(buffer + cksum_offset) == xfs_end_cksum(crc); } diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c index c221d0ecd52e..5c2929f94bd3 100644 --- a/fs/xfs/libxfs/xfs_defer.c +++ b/fs/xfs/libxfs/xfs_defer.c @@ -81,6 +81,10 @@ * - For each work item attached to the log intent item, * * Perform the described action. * * Attach the work item to the log done item. + * * If the result of doing the work was -EAGAIN, ->finish work + * wants a new transaction. See the "Requesting a Fresh + * Transaction while Finishing Deferred Work" section below for + * details. * * The key here is that we must log an intent item for all pending * work items every time we roll the transaction, and that we must log @@ -88,6 +92,34 @@ * we can perform complex remapping operations, chaining intent items * as needed. * + * Requesting a Fresh Transaction while Finishing Deferred Work + * + * If ->finish_item decides that it needs a fresh transaction to + * finish the work, it must ask its caller (xfs_defer_finish) for a + * continuation. The most likely cause of this circumstance are the + * refcount adjust functions deciding that they've logged enough items + * to be at risk of exceeding the transaction reservation. + * + * To get a fresh transaction, we want to log the existing log done + * item to prevent the log intent item from replaying, immediately log + * a new log intent item with the unfinished work items, roll the + * transaction, and re-call ->finish_item wherever it left off. The + * log done item and the new log intent item must be in the same + * transaction or atomicity cannot be guaranteed; defer_finish ensures + * that this happens. + * + * This requires some coordination between ->finish_item and + * defer_finish. Upon deciding to request a new transaction, + * ->finish_item should update the current work item to reflect the + * unfinished work. Next, it should reset the log done item's list + * count to the number of items finished, and return -EAGAIN. + * defer_finish sees the -EAGAIN, logs the new log intent item + * with the remaining work items, and leaves the xfs_defer_pending + * item at the head of the dop_work queue. Then it rolls the + * transaction and picks up processing where it left off. It is + * required that ->finish_item must be careful to leave enough + * transaction reservation to fit the new log intent item. + * * This is an example of remapping the extent (E, E+B) into file X at * offset A and dealing with the extent (C, C+B) already being mapped * there: @@ -104,21 +136,26 @@ * | Intent to add rmap (X, E, A, B) | * +-------------------------------------------------+ * | Reduce refcount for extent (C, B) | t2 - * | Done reducing refcount for extent (C, B) | + * | Done reducing refcount for extent (C, 9) | + * | Intent to reduce refcount for extent (C+9, B-9) | + * | (ran out of space after 9 refcount updates) | + * +-------------------------------------------------+ + * | Reduce refcount for extent (C+9, B+9) | t3 + * | Done reducing refcount for extent (C+9, B-9) | * | Increase refcount for extent (E, B) | * | Done increasing refcount for extent (E, B) | * | Intent to free extent (C, B) | * | Intent to free extent (F, 1) (refcountbt block) | * | Intent to remove rmap (F, 1, REFC) | * +-------------------------------------------------+ - * | Remove rmap (X, C, A, B) | t3 + * | Remove rmap (X, C, A, B) | t4 * | Done removing rmap (X, C, A, B) | * | Add rmap (X, E, A, B) | * | Done adding rmap (X, E, A, B) | * | Remove rmap (F, 1, REFC) | * | Done removing rmap (F, 1, REFC) | * +-------------------------------------------------+ - * | Free extent (C, B) | t4 + * | Free extent (C, B) | t5 * | Done freeing extent (C, B) | * | Free extent (D, 1) | * | Done freeing extent (D, 1) | @@ -141,6 +178,9 @@ * - Intent to free extent (C, B) * - Intent to free extent (F, 1) (refcountbt block) * - Intent to remove rmap (F, 1, REFC) + * + * Note that the continuation requested between t2 and t3 is likely to + * reoccur. */ static const struct xfs_defer_op_type *defer_op_types[XFS_DEFER_OPS_TYPE_MAX]; @@ -159,9 +199,9 @@ xfs_defer_intake_work( struct xfs_defer_pending *dfp; list_for_each_entry(dfp, &dop->dop_intake, dfp_list) { - trace_xfs_defer_intake_work(tp->t_mountp, dfp); dfp->dfp_intent = dfp->dfp_type->create_intent(tp, dfp->dfp_count); + trace_xfs_defer_intake_work(tp->t_mountp, dfp); list_sort(tp->t_mountp, &dfp->dfp_work, dfp->dfp_type->diff_items); list_for_each(li, &dfp->dfp_work) @@ -181,21 +221,14 @@ xfs_defer_trans_abort( struct xfs_defer_pending *dfp; trace_xfs_defer_trans_abort(tp->t_mountp, dop); - /* - * If the transaction was committed, drop the intent reference - * since we're bailing out of here. The other reference is - * dropped when the intent hits the AIL. If the transaction - * was not committed, the intent is freed by the intent item - * unlock handler on abort. - */ - if (!dop->dop_committed) - return; - /* Abort intent items. */ + /* Abort intent items that don't have a done item. */ list_for_each_entry(dfp, &dop->dop_pending, dfp_list) { trace_xfs_defer_pending_abort(tp->t_mountp, dfp); - if (!dfp->dfp_done) + if (dfp->dfp_intent && !dfp->dfp_done) { dfp->dfp_type->abort_intent(dfp->dfp_intent); + dfp->dfp_intent = NULL; + } } /* Shut down FS. */ @@ -323,7 +356,16 @@ xfs_defer_finish( dfp->dfp_count--; error = dfp->dfp_type->finish_item(*tp, dop, li, dfp->dfp_done, &state); - if (error) { + if (error == -EAGAIN) { + /* + * Caller wants a fresh transaction; + * put the work item back on the list + * and jump out. + */ + list_add(li, &dfp->dfp_work); + dfp->dfp_count++; + break; + } else if (error) { /* * Clean up after ourselves and jump out. * xfs_defer_cancel will take care of freeing @@ -335,9 +377,25 @@ xfs_defer_finish( goto out; } } - /* Done with the dfp, free it. */ - list_del(&dfp->dfp_list); - kmem_free(dfp); + if (error == -EAGAIN) { + /* + * Caller wants a fresh transaction, so log a + * new log intent item to replace the old one + * and roll the transaction. See "Requesting + * a Fresh Transaction while Finishing + * Deferred Work" above. + */ + dfp->dfp_intent = dfp->dfp_type->create_intent(*tp, + dfp->dfp_count); + dfp->dfp_done = NULL; + list_for_each(li, &dfp->dfp_work) + dfp->dfp_type->log_item(*tp, dfp->dfp_intent, + li); + } else { + /* Done with the dfp, free it. */ + list_del(&dfp->dfp_list); + kmem_free(dfp); + } if (cleanup_fn) cleanup_fn(*tp, state, error); diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h index e96533d178cf..f6e93ef0bffe 100644 --- a/fs/xfs/libxfs/xfs_defer.h +++ b/fs/xfs/libxfs/xfs_defer.h @@ -51,6 +51,8 @@ struct xfs_defer_pending { * find all the space it needs. */ enum xfs_defer_ops_type { + XFS_DEFER_OPS_TYPE_BMAP, + XFS_DEFER_OPS_TYPE_REFCOUNT, XFS_DEFER_OPS_TYPE_RMAP, XFS_DEFER_OPS_TYPE_FREE, XFS_DEFER_OPS_TYPE_MAX, diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c index 20a96dd5af7e..c58d72c220f5 100644 --- a/fs/xfs/libxfs/xfs_dir2.c +++ b/fs/xfs/libxfs/xfs_dir2.c @@ -93,7 +93,7 @@ xfs_ascii_ci_compname( return result; } -static struct xfs_nameops xfs_ascii_ci_nameops = { +static const struct xfs_nameops xfs_ascii_ci_nameops = { .hashname = xfs_ascii_ci_hashname, .compname = xfs_ascii_ci_compname, }; diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h index becc926c3e3d..0197590fa7d7 100644 --- a/fs/xfs/libxfs/xfs_dir2.h +++ b/fs/xfs/libxfs/xfs_dir2.h @@ -157,6 +157,9 @@ extern int xfs_dir2_isleaf(struct xfs_da_args *args, int *r); extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db, struct xfs_buf *bp); +extern void xfs_dir2_data_freescan_int(struct xfs_da_geometry *geo, + const struct xfs_dir_ops *ops, + struct xfs_dir2_data_hdr *hdr, int *loghead); extern void xfs_dir2_data_freescan(struct xfs_inode *dp, struct xfs_dir2_data_hdr *hdr, int *loghead); extern void xfs_dir2_data_log_entry(struct xfs_da_args *args, @@ -177,6 +180,8 @@ extern struct xfs_dir2_data_free *xfs_dir2_data_freefind( struct xfs_dir2_data_hdr *hdr, struct xfs_dir2_data_free *bf, struct xfs_dir2_data_unused *dup); +extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino); + extern const struct xfs_buf_ops xfs_dir3_block_buf_ops; extern const struct xfs_buf_ops xfs_dir3_leafn_buf_ops; extern const struct xfs_buf_ops xfs_dir3_leaf1_buf_ops; diff --git a/fs/xfs/libxfs/xfs_dir2_data.c b/fs/xfs/libxfs/xfs_dir2_data.c index 725fc7841fde..d478065b9544 100644 --- a/fs/xfs/libxfs/xfs_dir2_data.c +++ b/fs/xfs/libxfs/xfs_dir2_data.c @@ -329,7 +329,7 @@ xfs_dir3_data_read( err = xfs_da_read_buf(tp, dp, bno, mapped_bno, bpp, XFS_DATA_FORK, &xfs_dir3_data_buf_ops); - if (!err && tp) + if (!err && tp && *bpp) xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_DATA_BUF); return err; } @@ -505,8 +505,9 @@ xfs_dir2_data_freeremove( * Given a data block, reconstruct its bestfree map. */ void -xfs_dir2_data_freescan( - struct xfs_inode *dp, +xfs_dir2_data_freescan_int( + struct xfs_da_geometry *geo, + const struct xfs_dir_ops *ops, struct xfs_dir2_data_hdr *hdr, int *loghead) { @@ -516,7 +517,6 @@ xfs_dir2_data_freescan( struct xfs_dir2_data_free *bf; char *endp; /* end of block's data */ char *p; /* current entry pointer */ - struct xfs_da_geometry *geo = dp->i_mount->m_dir_geo; ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) || hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) || @@ -526,13 +526,13 @@ xfs_dir2_data_freescan( /* * Start by clearing the table. */ - bf = dp->d_ops->data_bestfree_p(hdr); + bf = ops->data_bestfree_p(hdr); memset(bf, 0, sizeof(*bf) * XFS_DIR2_DATA_FD_COUNT); *loghead = 1; /* * Set up pointers. */ - p = (char *)dp->d_ops->data_entry_p(hdr); + p = (char *)ops->data_entry_p(hdr); if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) || hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) { btp = xfs_dir2_block_tail_p(geo, hdr); @@ -559,12 +559,22 @@ xfs_dir2_data_freescan( else { dep = (xfs_dir2_data_entry_t *)p; ASSERT((char *)dep - (char *)hdr == - be16_to_cpu(*dp->d_ops->data_entry_tag_p(dep))); - p += dp->d_ops->data_entsize(dep->namelen); + be16_to_cpu(*ops->data_entry_tag_p(dep))); + p += ops->data_entsize(dep->namelen); } } } +void +xfs_dir2_data_freescan( + struct xfs_inode *dp, + struct xfs_dir2_data_hdr *hdr, + int *loghead) +{ + return xfs_dir2_data_freescan_int(dp->i_mount->m_dir_geo, dp->d_ops, + hdr, loghead); +} + /* * Initialize a data block at the given block number in the directory. * Give back the buffer for the created block. diff --git a/fs/xfs/libxfs/xfs_dir2_priv.h b/fs/xfs/libxfs/xfs_dir2_priv.h index ef9f6ead96a4..d04547fcf274 100644 --- a/fs/xfs/libxfs/xfs_dir2_priv.h +++ b/fs/xfs/libxfs/xfs_dir2_priv.h @@ -21,7 +21,6 @@ struct dir_context; /* xfs_dir2.c */ -extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino); extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space, xfs_dir2_db_t *dbp); extern int xfs_dir_cilookup_result(struct xfs_da_args *args, diff --git a/fs/xfs/libxfs/xfs_dquot_buf.c b/fs/xfs/libxfs/xfs_dquot_buf.c index 3cc3cf767474..ac9a003dd29a 100644 --- a/fs/xfs/libxfs/xfs_dquot_buf.c +++ b/fs/xfs/libxfs/xfs_dquot_buf.c @@ -191,8 +191,7 @@ xfs_dquot_buf_verify_crc( if (mp->m_quotainfo) ndquots = mp->m_quotainfo->qi_dqperchunk; else - ndquots = xfs_calc_dquots_per_chunk( - XFS_BB_TO_FSB(mp, bp->b_length)); + ndquots = xfs_calc_dquots_per_chunk(bp->b_length); for (i = 0; i < ndquots; i++, d++) { if (!xfs_verify_cksum((char *)d, sizeof(struct xfs_dqblk), diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index 270fb5cf4fa1..6b7579e7b60a 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -456,9 +456,11 @@ xfs_sb_has_compat_feature( #define XFS_SB_FEAT_RO_COMPAT_FINOBT (1 << 0) /* free inode btree */ #define XFS_SB_FEAT_RO_COMPAT_RMAPBT (1 << 1) /* reverse map btree */ +#define XFS_SB_FEAT_RO_COMPAT_REFLINK (1 << 2) /* reflinked files */ #define XFS_SB_FEAT_RO_COMPAT_ALL \ (XFS_SB_FEAT_RO_COMPAT_FINOBT | \ - XFS_SB_FEAT_RO_COMPAT_RMAPBT) + XFS_SB_FEAT_RO_COMPAT_RMAPBT | \ + XFS_SB_FEAT_RO_COMPAT_REFLINK) #define XFS_SB_FEAT_RO_COMPAT_UNKNOWN ~XFS_SB_FEAT_RO_COMPAT_ALL static inline bool xfs_sb_has_ro_compat_feature( @@ -546,6 +548,12 @@ static inline bool xfs_sb_version_hasrmapbt(struct xfs_sb *sbp) (sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_RMAPBT); } +static inline bool xfs_sb_version_hasreflink(struct xfs_sb *sbp) +{ + return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 && + (sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_REFLINK); +} + /* * end of superblock version macros */ @@ -641,14 +649,17 @@ typedef struct xfs_agf { uuid_t agf_uuid; /* uuid of filesystem */ __be32 agf_rmap_blocks; /* rmapbt blocks used */ - __be32 agf_padding; /* padding */ + __be32 agf_refcount_blocks; /* refcountbt blocks used */ + + __be32 agf_refcount_root; /* refcount tree root block */ + __be32 agf_refcount_level; /* refcount btree levels */ /* * reserve some contiguous space for future logged fields before we add * the unlogged fields. This makes the range logging via flags and * structure offsets much simpler. */ - __be64 agf_spare64[15]; + __be64 agf_spare64[14]; /* unlogged fields, written during buffer writeback. */ __be64 agf_lsn; /* last write sequence */ @@ -674,8 +685,11 @@ typedef struct xfs_agf { #define XFS_AGF_BTREEBLKS 0x00000800 #define XFS_AGF_UUID 0x00001000 #define XFS_AGF_RMAP_BLOCKS 0x00002000 -#define XFS_AGF_SPARE64 0x00004000 -#define XFS_AGF_NUM_BITS 15 +#define XFS_AGF_REFCOUNT_BLOCKS 0x00004000 +#define XFS_AGF_REFCOUNT_ROOT 0x00008000 +#define XFS_AGF_REFCOUNT_LEVEL 0x00010000 +#define XFS_AGF_SPARE64 0x00020000 +#define XFS_AGF_NUM_BITS 18 #define XFS_AGF_ALL_BITS ((1 << XFS_AGF_NUM_BITS) - 1) #define XFS_AGF_FLAGS \ @@ -693,6 +707,9 @@ typedef struct xfs_agf { { XFS_AGF_BTREEBLKS, "BTREEBLKS" }, \ { XFS_AGF_UUID, "UUID" }, \ { XFS_AGF_RMAP_BLOCKS, "RMAP_BLOCKS" }, \ + { XFS_AGF_REFCOUNT_BLOCKS, "REFCOUNT_BLOCKS" }, \ + { XFS_AGF_REFCOUNT_ROOT, "REFCOUNT_ROOT" }, \ + { XFS_AGF_REFCOUNT_LEVEL, "REFCOUNT_LEVEL" }, \ { XFS_AGF_SPARE64, "SPARE64" } /* disk block (xfs_daddr_t) in the AG */ @@ -848,7 +865,6 @@ typedef struct xfs_timestamp { * padding field for v3 inodes. */ #define XFS_DINODE_MAGIC 0x494e /* 'IN' */ -#define XFS_DINODE_GOOD_VERSION(v) ((v) >= 1 && (v) <= 3) typedef struct xfs_dinode { __be16 di_magic; /* inode magic # = XFS_DINODE_MAGIC */ __be16 di_mode; /* mode and type of file */ @@ -885,7 +901,8 @@ typedef struct xfs_dinode { __be64 di_changecount; /* number of attribute changes */ __be64 di_lsn; /* flush sequence */ __be64 di_flags2; /* more random flags */ - __u8 di_pad2[16]; /* more padding for future expansion */ + __be32 di_cowextsize; /* basic cow extent size for file */ + __u8 di_pad2[12]; /* more padding for future expansion */ /* fields only written to during inode creation */ xfs_timestamp_t di_crtime; /* time created */ @@ -1041,9 +1058,14 @@ static inline void xfs_dinode_put_rdev(struct xfs_dinode *dip, xfs_dev_t rdev) * 16 bits of the XFS_XFLAG_s range. */ #define XFS_DIFLAG2_DAX_BIT 0 /* use DAX for this inode */ +#define XFS_DIFLAG2_REFLINK_BIT 1 /* file's blocks may be shared */ +#define XFS_DIFLAG2_COWEXTSIZE_BIT 2 /* copy on write extent size hint */ #define XFS_DIFLAG2_DAX (1 << XFS_DIFLAG2_DAX_BIT) +#define XFS_DIFLAG2_REFLINK (1 << XFS_DIFLAG2_REFLINK_BIT) +#define XFS_DIFLAG2_COWEXTSIZE (1 << XFS_DIFLAG2_COWEXTSIZE_BIT) -#define XFS_DIFLAG2_ANY (XFS_DIFLAG2_DAX) +#define XFS_DIFLAG2_ANY \ + (XFS_DIFLAG2_DAX | XFS_DIFLAG2_REFLINK | XFS_DIFLAG2_COWEXTSIZE) /* * Inode number format: @@ -1353,7 +1375,9 @@ struct xfs_owner_info { #define XFS_RMAP_OWN_AG (-5ULL) /* AG freespace btree blocks */ #define XFS_RMAP_OWN_INOBT (-6ULL) /* Inode btree blocks */ #define XFS_RMAP_OWN_INODES (-7ULL) /* Inode chunk */ -#define XFS_RMAP_OWN_MIN (-8ULL) /* guard */ +#define XFS_RMAP_OWN_REFC (-8ULL) /* refcount tree */ +#define XFS_RMAP_OWN_COW (-9ULL) /* cow allocations */ +#define XFS_RMAP_OWN_MIN (-10ULL) /* guard */ #define XFS_RMAP_NON_INODE_OWNER(owner) (!!((owner) & (1ULL << 63))) @@ -1434,6 +1458,62 @@ typedef __be32 xfs_rmap_ptr_t; XFS_IBT_BLOCK(mp) + 1) /* + * Reference Count Btree format definitions + * + */ +#define XFS_REFC_CRC_MAGIC 0x52334643 /* 'R3FC' */ + +unsigned int xfs_refc_block(struct xfs_mount *mp); + +/* + * Data record/key structure + * + * Each record associates a range of physical blocks (starting at + * rc_startblock and ending rc_blockcount blocks later) with a reference + * count (rc_refcount). Extents that are being used to stage a copy on + * write (CoW) operation are recorded in the refcount btree with a + * refcount of 1. All other records must have a refcount > 1 and must + * track an extent mapped only by file data forks. + * + * Extents with a single owner (attributes, metadata, non-shared file + * data) are not tracked here. Free space is also not tracked here. + * This is consistent with pre-reflink XFS. + */ + +/* + * Extents that are being used to stage a copy on write are stored + * in the refcount btree with a refcount of 1 and the upper bit set + * on the startblock. This speeds up mount time deletion of stale + * staging extents because they're all at the right side of the tree. + */ +#define XFS_REFC_COW_START ((xfs_agblock_t)(1U << 31)) +#define REFCNTBT_COWFLAG_BITLEN 1 +#define REFCNTBT_AGBLOCK_BITLEN 31 + +struct xfs_refcount_rec { + __be32 rc_startblock; /* starting block number */ + __be32 rc_blockcount; /* count of blocks */ + __be32 rc_refcount; /* number of inodes linked here */ +}; + +struct xfs_refcount_key { + __be32 rc_startblock; /* starting block number */ +}; + +struct xfs_refcount_irec { + xfs_agblock_t rc_startblock; /* starting block number */ + xfs_extlen_t rc_blockcount; /* count of free blocks */ + xfs_nlink_t rc_refcount; /* number of inodes linked here */ +}; + +#define MAXREFCOUNT ((xfs_nlink_t)~0U) +#define MAXREFCEXTLEN ((xfs_extlen_t)~0U) + +/* btree pointer type */ +typedef __be32 xfs_refcount_ptr_t; + + +/* * BMAP Btree format definitions * * This includes both the root block definition that sits inside an inode fork diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h index 79455058b752..b72dc821d78b 100644 --- a/fs/xfs/libxfs/xfs_fs.h +++ b/fs/xfs/libxfs/xfs_fs.h @@ -81,14 +81,16 @@ struct getbmapx { #define BMV_IF_PREALLOC 0x4 /* rtn status BMV_OF_PREALLOC if req */ #define BMV_IF_DELALLOC 0x8 /* rtn status BMV_OF_DELALLOC if req */ #define BMV_IF_NO_HOLES 0x10 /* Do not return holes */ +#define BMV_IF_COWFORK 0x20 /* return CoW fork rather than data */ #define BMV_IF_VALID \ (BMV_IF_ATTRFORK|BMV_IF_NO_DMAPI_READ|BMV_IF_PREALLOC| \ - BMV_IF_DELALLOC|BMV_IF_NO_HOLES) + BMV_IF_DELALLOC|BMV_IF_NO_HOLES|BMV_IF_COWFORK) /* bmv_oflags values - returned for each non-header segment */ #define BMV_OF_PREALLOC 0x1 /* segment = unwritten pre-allocation */ #define BMV_OF_DELALLOC 0x2 /* segment = delayed allocation */ #define BMV_OF_LAST 0x4 /* segment is the last in the file */ +#define BMV_OF_SHARED 0x8 /* segment shared with another file */ /* * Structure for XFS_IOC_FSSETDM. @@ -206,7 +208,8 @@ typedef struct xfs_fsop_resblks { #define XFS_FSOP_GEOM_FLAGS_FTYPE 0x10000 /* inode directory types */ #define XFS_FSOP_GEOM_FLAGS_FINOBT 0x20000 /* free inode btree */ #define XFS_FSOP_GEOM_FLAGS_SPINODES 0x40000 /* sparse inode chunks */ -#define XFS_FSOP_GEOM_FLAGS_RMAPBT 0x80000 /* Reverse mapping btree */ +#define XFS_FSOP_GEOM_FLAGS_RMAPBT 0x80000 /* reverse mapping btree */ +#define XFS_FSOP_GEOM_FLAGS_REFLINK 0x100000 /* files can share blocks */ /* * Minimum and maximum sizes need for growth checks. @@ -275,7 +278,8 @@ typedef struct xfs_bstat { #define bs_projid bs_projid_lo /* (previously just bs_projid) */ __u16 bs_forkoff; /* inode fork offset in bytes */ __u16 bs_projid_hi; /* higher part of project id */ - unsigned char bs_pad[10]; /* pad space, unused */ + unsigned char bs_pad[6]; /* pad space, unused */ + __u32 bs_cowextsize; /* cow extent size */ __u32 bs_dmevmask; /* DMIG event mask */ __u16 bs_dmstate; /* DMIG state info */ __u16 bs_aextents; /* attribute number of extents */ diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index 51b4e0de1fdc..f272abff11e1 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -2344,7 +2344,8 @@ xfs_imap( imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, agbno); imap->im_len = XFS_FSB_TO_BB(mp, 1); - imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog); + imap->im_boffset = (unsigned short)(offset << + mp->m_sb.sb_inodelog); return 0; } @@ -2372,7 +2373,7 @@ out_map: imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, cluster_agbno); imap->im_len = XFS_FSB_TO_BB(mp, blks_per_cluster); - imap->im_boffset = (ushort)(offset << mp->m_sb.sb_inodelog); + imap->im_boffset = (unsigned short)(offset << mp->m_sb.sb_inodelog); /* * If the inode number maps to a block outside the bounds @@ -2450,8 +2451,6 @@ xfs_ialloc_log_agi( ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC)); #endif - xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGI_BUF); - /* * Compute byte offsets for the first and last fields in the first * region and log the agi buffer. This only logs up through @@ -2512,8 +2511,15 @@ xfs_agi_verify( if (!XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum))) return false; - if (be32_to_cpu(agi->agi_level) > XFS_BTREE_MAXLEVELS) + if (be32_to_cpu(agi->agi_level) < 1 || + be32_to_cpu(agi->agi_level) > XFS_BTREE_MAXLEVELS) + return false; + + if (xfs_sb_version_hasfinobt(&mp->m_sb) && + (be32_to_cpu(agi->agi_free_level) < 1 || + be32_to_cpu(agi->agi_free_level) > XFS_BTREE_MAXLEVELS)) 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 @@ -2592,6 +2598,8 @@ xfs_read_agi( XFS_FSS_TO_BB(mp, 1), 0, bpp, &xfs_agi_buf_ops); if (error) return error; + if (tp) + xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_AGI_BUF); xfs_buf_set_ref(*bpp, XFS_AGI_REF); return 0; diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index 31ca2208c03d..0fd086d03d41 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -132,7 +132,7 @@ xfs_inobt_free_block( xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INOBT); return xfs_free_extent(cur->bc_tp, XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp)), 1, - &oinfo); + &oinfo, XFS_AG_RESV_NONE); } STATIC int @@ -357,7 +357,7 @@ xfs_inobt_init_cursor( struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp); struct xfs_btree_cur *cur; - cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP); + cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); cur->bc_tp = tp; cur->bc_mp = mp; @@ -365,9 +365,11 @@ xfs_inobt_init_cursor( if (btnum == XFS_BTNUM_INO) { cur->bc_nlevels = be32_to_cpu(agi->agi_level); cur->bc_ops = &xfs_inobt_ops; + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2); } else { cur->bc_nlevels = be32_to_cpu(agi->agi_free_level); cur->bc_ops = &xfs_finobt_ops; + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2); } cur->bc_blocklog = mp->m_sb.sb_blocklog; diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c index 4b9769e23c83..dd483e2767f7 100644 --- a/fs/xfs/libxfs/xfs_inode_buf.c +++ b/fs/xfs/libxfs/xfs_inode_buf.c @@ -57,6 +57,17 @@ xfs_inobp_check( } #endif +bool +xfs_dinode_good_version( + struct xfs_mount *mp, + __u8 version) +{ + if (xfs_sb_version_hascrc(&mp->m_sb)) + return version == 3; + + return version == 1 || version == 2; +} + /* * If we are doing readahead on an inode buffer, we might be in log recovery * reading an inode allocation buffer that hasn't yet been replayed, and hence @@ -91,7 +102,7 @@ xfs_inode_buf_verify( dip = xfs_buf_offset(bp, (i << mp->m_sb.sb_inodelog)); di_ok = dip->di_magic == cpu_to_be16(XFS_DINODE_MAGIC) && - XFS_DINODE_GOOD_VERSION(dip->di_version); + xfs_dinode_good_version(mp, dip->di_version); if (unlikely(XFS_TEST_ERROR(!di_ok, mp, XFS_ERRTAG_ITOBP_INOTOBP, XFS_RANDOM_ITOBP_INOTOBP))) { @@ -256,6 +267,7 @@ xfs_inode_from_disk( to->di_crtime.t_sec = be32_to_cpu(from->di_crtime.t_sec); to->di_crtime.t_nsec = be32_to_cpu(from->di_crtime.t_nsec); to->di_flags2 = be64_to_cpu(from->di_flags2); + to->di_cowextsize = be32_to_cpu(from->di_cowextsize); } } @@ -305,7 +317,7 @@ xfs_inode_to_disk( to->di_crtime.t_sec = cpu_to_be32(from->di_crtime.t_sec); to->di_crtime.t_nsec = cpu_to_be32(from->di_crtime.t_nsec); to->di_flags2 = cpu_to_be64(from->di_flags2); - + to->di_cowextsize = cpu_to_be32(from->di_cowextsize); to->di_ino = cpu_to_be64(ip->i_ino); to->di_lsn = cpu_to_be64(lsn); memset(to->di_pad2, 0, sizeof(to->di_pad2)); @@ -357,6 +369,7 @@ xfs_log_dinode_to_disk( to->di_crtime.t_sec = cpu_to_be32(from->di_crtime.t_sec); to->di_crtime.t_nsec = cpu_to_be32(from->di_crtime.t_nsec); to->di_flags2 = cpu_to_be64(from->di_flags2); + to->di_cowextsize = cpu_to_be32(from->di_cowextsize); to->di_ino = cpu_to_be64(from->di_ino); to->di_lsn = cpu_to_be64(from->di_lsn); memcpy(to->di_pad2, from->di_pad2, sizeof(to->di_pad2)); @@ -370,12 +383,23 @@ xfs_log_dinode_to_disk( static bool xfs_dinode_verify( struct xfs_mount *mp, - struct xfs_inode *ip, + xfs_ino_t ino, struct xfs_dinode *dip) { + uint16_t flags; + uint64_t flags2; + if (dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC)) return false; + /* don't allow invalid i_size */ + if (be64_to_cpu(dip->di_size) & (1ULL << 63)) + return false; + + /* No zero-length symlinks. */ + if (S_ISLNK(be16_to_cpu(dip->di_mode)) && dip->di_size == 0) + return false; + /* only version 3 or greater inodes are extensively verified here */ if (dip->di_version < 3) return true; @@ -385,10 +409,27 @@ xfs_dinode_verify( if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize, XFS_DINODE_CRC_OFF)) return false; - if (be64_to_cpu(dip->di_ino) != ip->i_ino) + if (be64_to_cpu(dip->di_ino) != ino) return false; if (!uuid_equal(&dip->di_uuid, &mp->m_sb.sb_meta_uuid)) return false; + + flags = be16_to_cpu(dip->di_flags); + flags2 = be64_to_cpu(dip->di_flags2); + + /* don't allow reflink/cowextsize if we don't have reflink */ + if ((flags2 & (XFS_DIFLAG2_REFLINK | XFS_DIFLAG2_COWEXTSIZE)) && + !xfs_sb_version_hasreflink(&mp->m_sb)) + return false; + + /* don't let reflink and realtime mix */ + if ((flags2 & XFS_DIFLAG2_REFLINK) && (flags & XFS_DIFLAG_REALTIME)) + return false; + + /* don't let reflink and dax mix */ + if ((flags2 & XFS_DIFLAG2_REFLINK) && (flags2 & XFS_DIFLAG2_DAX)) + return false; + return true; } @@ -403,7 +444,7 @@ xfs_dinode_calc_crc( return; ASSERT(xfs_sb_version_hascrc(&mp->m_sb)); - crc = xfs_start_cksum((char *)dip, mp->m_sb.sb_inodesize, + crc = xfs_start_cksum_update((char *)dip, mp->m_sb.sb_inodesize, XFS_DINODE_CRC_OFF); dip->di_crc = xfs_end_cksum(crc); } @@ -460,7 +501,7 @@ xfs_iread( return error; /* even unallocated inodes are verified */ - if (!xfs_dinode_verify(mp, ip, dip)) { + if (!xfs_dinode_verify(mp, ip->i_ino, dip)) { xfs_alert(mp, "%s: validation failed for inode %lld failed", __func__, ip->i_ino); diff --git a/fs/xfs/libxfs/xfs_inode_buf.h b/fs/xfs/libxfs/xfs_inode_buf.h index 7c4dd321b215..6848a0afbce7 100644 --- a/fs/xfs/libxfs/xfs_inode_buf.h +++ b/fs/xfs/libxfs/xfs_inode_buf.h @@ -47,6 +47,7 @@ struct xfs_icdinode { __uint16_t di_flags; /* random flags, XFS_DIFLAG_... */ __uint64_t di_flags2; /* more random flags */ + __uint32_t di_cowextsize; /* basic cow extent size for file */ xfs_ictimestamp_t di_crtime; /* time created */ }; @@ -57,8 +58,8 @@ struct xfs_icdinode { */ struct xfs_imap { xfs_daddr_t im_blkno; /* starting BB of inode chunk */ - ushort im_len; /* length in BBs of inode chunk */ - ushort im_boffset; /* inode offset in block in bytes */ + unsigned short im_len; /* length in BBs of inode chunk */ + unsigned short im_boffset; /* inode offset in block in bytes */ }; int xfs_imap_to_bp(struct xfs_mount *, struct xfs_trans *, @@ -73,6 +74,8 @@ void xfs_inode_from_disk(struct xfs_inode *ip, struct xfs_dinode *from); void xfs_log_dinode_to_disk(struct xfs_log_dinode *from, struct xfs_dinode *to); +bool xfs_dinode_good_version(struct xfs_mount *mp, __u8 version); + #if defined(DEBUG) void xfs_inobp_check(struct xfs_mount *, struct xfs_buf *); #else diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index bbcc8c7a44b3..222e103356c6 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -121,6 +121,26 @@ xfs_iformat_fork( return -EFSCORRUPTED; } + if (unlikely(xfs_is_reflink_inode(ip) && + (VFS_I(ip)->i_mode & S_IFMT) != S_IFREG)) { + xfs_warn(ip->i_mount, + "corrupt dinode %llu, wrong file type for reflink.", + ip->i_ino); + XFS_CORRUPTION_ERROR("xfs_iformat(reflink)", + XFS_ERRLEVEL_LOW, ip->i_mount, dip); + return -EFSCORRUPTED; + } + + if (unlikely(xfs_is_reflink_inode(ip) && + (ip->i_d.di_flags & XFS_DIFLAG_REALTIME))) { + xfs_warn(ip->i_mount, + "corrupt dinode %llu, has reflink+realtime flag set.", + ip->i_ino); + XFS_CORRUPTION_ERROR("xfs_iformat(reflink)", + XFS_ERRLEVEL_LOW, ip->i_mount, dip); + return -EFSCORRUPTED; + } + switch (VFS_I(ip)->i_mode & S_IFMT) { case S_IFIFO: case S_IFCHR: @@ -186,9 +206,14 @@ xfs_iformat_fork( XFS_ERROR_REPORT("xfs_iformat(7)", XFS_ERRLEVEL_LOW, ip->i_mount); return -EFSCORRUPTED; } - if (error) { + if (error) return error; + + if (xfs_is_reflink_inode(ip)) { + ASSERT(ip->i_cowfp == NULL); + xfs_ifork_init_cow(ip); } + if (!XFS_DFORK_Q(dip)) return 0; @@ -208,7 +233,8 @@ xfs_iformat_fork( XFS_CORRUPTION_ERROR("xfs_iformat(8)", XFS_ERRLEVEL_LOW, ip->i_mount, dip); - return -EFSCORRUPTED; + error = -EFSCORRUPTED; + break; } error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK, size); @@ -226,6 +252,9 @@ xfs_iformat_fork( if (error) { kmem_zone_free(xfs_ifork_zone, ip->i_afp); ip->i_afp = NULL; + if (ip->i_cowfp) + kmem_zone_free(xfs_ifork_zone, ip->i_cowfp); + ip->i_cowfp = NULL; xfs_idestroy_fork(ip, XFS_DATA_FORK); } return error; @@ -740,9 +769,19 @@ xfs_idestroy_fork( if (whichfork == XFS_ATTR_FORK) { kmem_zone_free(xfs_ifork_zone, ip->i_afp); ip->i_afp = NULL; + } else if (whichfork == XFS_COW_FORK) { + kmem_zone_free(xfs_ifork_zone, ip->i_cowfp); + ip->i_cowfp = NULL; } } +/* Count number of incore extents based on if_bytes */ +xfs_extnum_t +xfs_iext_count(struct xfs_ifork *ifp) +{ + return ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); +} + /* * Convert in-core extents to on-disk form * @@ -771,7 +810,7 @@ xfs_iextents_copy( ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); ASSERT(ifp->if_bytes > 0); - nrecs = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nrecs = xfs_iext_count(ifp); XFS_BMAP_TRACE_EXLIST(ip, nrecs, whichfork); ASSERT(nrecs > 0); @@ -909,7 +948,7 @@ xfs_iext_get_ext( xfs_extnum_t idx) /* index of target extent */ { ASSERT(idx >= 0); - ASSERT(idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t)); + ASSERT(idx < xfs_iext_count(ifp)); if ((ifp->if_flags & XFS_IFEXTIREC) && (idx == 0)) { return ifp->if_u1.if_ext_irec->er_extbuf; @@ -927,6 +966,19 @@ xfs_iext_get_ext( } } +/* Convert bmap state flags to an inode fork. */ +struct xfs_ifork * +xfs_iext_state_to_fork( + struct xfs_inode *ip, + int state) +{ + if (state & BMAP_COWFORK) + return ip->i_cowfp; + else if (state & BMAP_ATTRFORK) + return ip->i_afp; + return &ip->i_df; +} + /* * Insert new item(s) into the extent records for incore inode * fork 'ifp'. 'count' new items are inserted at index 'idx'. @@ -939,7 +991,7 @@ xfs_iext_insert( xfs_bmbt_irec_t *new, /* items to insert */ int state) /* type of extent conversion */ { - xfs_ifork_t *ifp = (state & BMAP_ATTRFORK) ? ip->i_afp : &ip->i_df; + xfs_ifork_t *ifp = xfs_iext_state_to_fork(ip, state); xfs_extnum_t i; /* extent record index */ trace_xfs_iext_insert(ip, idx, new, state, _RET_IP_); @@ -972,7 +1024,7 @@ xfs_iext_add( int new_size; /* size of extents after adding */ xfs_extnum_t nextents; /* number of extents in file */ - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); ASSERT((idx >= 0) && (idx <= nextents)); byte_diff = ext_diff * sizeof(xfs_bmbt_rec_t); new_size = ifp->if_bytes + byte_diff; @@ -1189,14 +1241,14 @@ xfs_iext_remove( int ext_diff, /* number of extents to remove */ int state) /* type of extent conversion */ { - xfs_ifork_t *ifp = (state & BMAP_ATTRFORK) ? ip->i_afp : &ip->i_df; + xfs_ifork_t *ifp = xfs_iext_state_to_fork(ip, state); xfs_extnum_t nextents; /* number of extents in file */ int new_size; /* size of extents after removal */ trace_xfs_iext_remove(ip, idx, state, _RET_IP_); ASSERT(ext_diff > 0); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); new_size = (nextents - ext_diff) * sizeof(xfs_bmbt_rec_t); if (new_size == 0) { @@ -1225,7 +1277,7 @@ xfs_iext_remove_inline( ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); ASSERT(idx < XFS_INLINE_EXTS); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); ASSERT(((nextents - ext_diff) > 0) && (nextents - ext_diff) < XFS_INLINE_EXTS); @@ -1264,7 +1316,7 @@ xfs_iext_remove_direct( ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); new_size = ifp->if_bytes - (ext_diff * sizeof(xfs_bmbt_rec_t)); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); if (new_size == 0) { xfs_iext_destroy(ifp); @@ -1501,7 +1553,7 @@ xfs_iext_indirect_to_direct( int size; /* size of file extents */ ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); ASSERT(nextents <= XFS_LINEAR_EXTS); size = nextents * sizeof(xfs_bmbt_rec_t); @@ -1575,7 +1627,7 @@ xfs_iext_bno_to_ext( xfs_extnum_t nextents; /* number of file extents */ xfs_fileoff_t startoff = 0; /* start offset of extent */ - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); if (nextents == 0) { *idxp = 0; return NULL; @@ -1688,8 +1740,8 @@ xfs_iext_idx_to_irec( ASSERT(ifp->if_flags & XFS_IFEXTIREC); ASSERT(page_idx >= 0); - ASSERT(page_idx <= ifp->if_bytes / sizeof(xfs_bmbt_rec_t)); - ASSERT(page_idx < ifp->if_bytes / sizeof(xfs_bmbt_rec_t) || realloc); + ASSERT(page_idx <= xfs_iext_count(ifp)); + ASSERT(page_idx < xfs_iext_count(ifp) || realloc); nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; erp_idx = 0; @@ -1737,7 +1789,7 @@ xfs_iext_irec_init( xfs_extnum_t nextents; /* number of extents in file */ ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); ASSERT(nextents <= XFS_LINEAR_EXTS); erp = kmem_alloc(sizeof(xfs_ext_irec_t), KM_NOFS); @@ -1861,7 +1913,7 @@ xfs_iext_irec_compact( ASSERT(ifp->if_flags & XFS_IFEXTIREC); nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); if (nextents == 0) { xfs_iext_destroy(ifp); @@ -1934,3 +1986,66 @@ xfs_iext_irec_update_extoffs( ifp->if_u1.if_ext_irec[i].er_extoff += ext_diff; } } + +/* + * Initialize an inode's copy-on-write fork. + */ +void +xfs_ifork_init_cow( + struct xfs_inode *ip) +{ + if (ip->i_cowfp) + return; + + ip->i_cowfp = kmem_zone_zalloc(xfs_ifork_zone, + KM_SLEEP | KM_NOFS); + ip->i_cowfp->if_flags = XFS_IFEXTENTS; + ip->i_cformat = XFS_DINODE_FMT_EXTENTS; + ip->i_cnextents = 0; +} + +/* + * Lookup the extent covering bno. + * + * If there is an extent covering bno return the extent index, and store the + * expanded extent structure in *gotp, and the extent index in *idx. + * If there is no extent covering bno, but there is an extent after it (e.g. + * it lies in a hole) return that extent in *gotp and its index in *idx + * instead. + * If bno is beyond the last extent return false, and return the index after + * the last valid index in *idxp. + */ +bool +xfs_iext_lookup_extent( + struct xfs_inode *ip, + struct xfs_ifork *ifp, + xfs_fileoff_t bno, + xfs_extnum_t *idxp, + struct xfs_bmbt_irec *gotp) +{ + struct xfs_bmbt_rec_host *ep; + + XFS_STATS_INC(ip->i_mount, xs_look_exlist); + + ep = xfs_iext_bno_to_ext(ifp, bno, idxp); + if (!ep) + return false; + xfs_bmbt_get_all(ep, gotp); + return true; +} + +/* + * Return true if there is an extent at index idx, and return the expanded + * extent structure at idx in that case. Else return false. + */ +bool +xfs_iext_get_extent( + struct xfs_ifork *ifp, + xfs_extnum_t idx, + struct xfs_bmbt_irec *gotp) +{ + if (idx < 0 || idx >= xfs_iext_count(ifp)) + return false; + xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), gotp); + return true; +} diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h index f95e072ae646..7fb8365326d1 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.h +++ b/fs/xfs/libxfs/xfs_inode_fork.h @@ -92,7 +92,9 @@ typedef struct xfs_ifork { #define XFS_IFORK_PTR(ip,w) \ ((w) == XFS_DATA_FORK ? \ &(ip)->i_df : \ - (ip)->i_afp) + ((w) == XFS_ATTR_FORK ? \ + (ip)->i_afp : \ + (ip)->i_cowfp)) #define XFS_IFORK_DSIZE(ip) \ (XFS_IFORK_Q(ip) ? \ XFS_IFORK_BOFF(ip) : \ @@ -105,26 +107,38 @@ typedef struct xfs_ifork { #define XFS_IFORK_SIZE(ip,w) \ ((w) == XFS_DATA_FORK ? \ XFS_IFORK_DSIZE(ip) : \ - XFS_IFORK_ASIZE(ip)) + ((w) == XFS_ATTR_FORK ? \ + XFS_IFORK_ASIZE(ip) : \ + 0)) #define XFS_IFORK_FORMAT(ip,w) \ ((w) == XFS_DATA_FORK ? \ (ip)->i_d.di_format : \ - (ip)->i_d.di_aformat) + ((w) == XFS_ATTR_FORK ? \ + (ip)->i_d.di_aformat : \ + (ip)->i_cformat)) #define XFS_IFORK_FMT_SET(ip,w,n) \ ((w) == XFS_DATA_FORK ? \ ((ip)->i_d.di_format = (n)) : \ - ((ip)->i_d.di_aformat = (n))) + ((w) == XFS_ATTR_FORK ? \ + ((ip)->i_d.di_aformat = (n)) : \ + ((ip)->i_cformat = (n)))) #define XFS_IFORK_NEXTENTS(ip,w) \ ((w) == XFS_DATA_FORK ? \ (ip)->i_d.di_nextents : \ - (ip)->i_d.di_anextents) + ((w) == XFS_ATTR_FORK ? \ + (ip)->i_d.di_anextents : \ + (ip)->i_cnextents)) #define XFS_IFORK_NEXT_SET(ip,w,n) \ ((w) == XFS_DATA_FORK ? \ ((ip)->i_d.di_nextents = (n)) : \ - ((ip)->i_d.di_anextents = (n))) + ((w) == XFS_ATTR_FORK ? \ + ((ip)->i_d.di_anextents = (n)) : \ + ((ip)->i_cnextents = (n)))) #define XFS_IFORK_MAXEXT(ip, w) \ (XFS_IFORK_SIZE(ip, w) / sizeof(xfs_bmbt_rec_t)) +struct xfs_ifork *xfs_iext_state_to_fork(struct xfs_inode *ip, int state); + int xfs_iformat_fork(struct xfs_inode *, struct xfs_dinode *); void xfs_iflush_fork(struct xfs_inode *, struct xfs_dinode *, struct xfs_inode_log_item *, int); @@ -138,6 +152,7 @@ void xfs_init_local_fork(struct xfs_inode *, int, const void *, int); struct xfs_bmbt_rec_host * xfs_iext_get_ext(struct xfs_ifork *, xfs_extnum_t); +xfs_extnum_t xfs_iext_count(struct xfs_ifork *); void xfs_iext_insert(struct xfs_inode *, xfs_extnum_t, xfs_extnum_t, struct xfs_bmbt_irec *, int); void xfs_iext_add(struct xfs_ifork *, xfs_extnum_t, int); @@ -167,6 +182,14 @@ void xfs_iext_irec_compact_pages(struct xfs_ifork *); void xfs_iext_irec_compact_full(struct xfs_ifork *); void xfs_iext_irec_update_extoffs(struct xfs_ifork *, int, int); +bool xfs_iext_lookup_extent(struct xfs_inode *ip, + struct xfs_ifork *ifp, xfs_fileoff_t bno, + xfs_extnum_t *idxp, struct xfs_bmbt_irec *gotp); +bool xfs_iext_get_extent(struct xfs_ifork *ifp, xfs_extnum_t idx, + struct xfs_bmbt_irec *gotp); + extern struct kmem_zone *xfs_ifork_zone; +extern void xfs_ifork_init_cow(struct xfs_inode *ip); + #endif /* __XFS_INODE_FORK_H__ */ diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h index a6eed43fa7cd..7ae571f8e34a 100644 --- a/fs/xfs/libxfs/xfs_log_format.h +++ b/fs/xfs/libxfs/xfs_log_format.h @@ -112,7 +112,11 @@ static inline uint xlog_get_cycle(char *ptr) #define XLOG_REG_TYPE_ICREATE 20 #define XLOG_REG_TYPE_RUI_FORMAT 21 #define XLOG_REG_TYPE_RUD_FORMAT 22 -#define XLOG_REG_TYPE_MAX 22 +#define XLOG_REG_TYPE_CUI_FORMAT 23 +#define XLOG_REG_TYPE_CUD_FORMAT 24 +#define XLOG_REG_TYPE_BUI_FORMAT 25 +#define XLOG_REG_TYPE_BUD_FORMAT 26 +#define XLOG_REG_TYPE_MAX 26 /* * Flags to log operation header @@ -231,6 +235,10 @@ typedef struct xfs_trans_header { #define XFS_LI_ICREATE 0x123f #define XFS_LI_RUI 0x1240 /* rmap update intent */ #define XFS_LI_RUD 0x1241 +#define XFS_LI_CUI 0x1242 /* refcount update intent */ +#define XFS_LI_CUD 0x1243 +#define XFS_LI_BUI 0x1244 /* bmbt update intent */ +#define XFS_LI_BUD 0x1245 #define XFS_LI_TYPE_DESC \ { XFS_LI_EFI, "XFS_LI_EFI" }, \ @@ -242,7 +250,11 @@ typedef struct xfs_trans_header { { XFS_LI_QUOTAOFF, "XFS_LI_QUOTAOFF" }, \ { XFS_LI_ICREATE, "XFS_LI_ICREATE" }, \ { XFS_LI_RUI, "XFS_LI_RUI" }, \ - { XFS_LI_RUD, "XFS_LI_RUD" } + { XFS_LI_RUD, "XFS_LI_RUD" }, \ + { XFS_LI_CUI, "XFS_LI_CUI" }, \ + { XFS_LI_CUD, "XFS_LI_CUD" }, \ + { XFS_LI_BUI, "XFS_LI_BUI" }, \ + { XFS_LI_BUD, "XFS_LI_BUD" } /* * Inode Log Item Format definitions. @@ -411,7 +423,8 @@ struct xfs_log_dinode { __uint64_t di_changecount; /* number of attribute changes */ xfs_lsn_t di_lsn; /* flush sequence */ __uint64_t di_flags2; /* more random flags */ - __uint8_t di_pad2[16]; /* more padding for future expansion */ + __uint32_t di_cowextsize; /* basic cow extent size for file */ + __uint8_t di_pad2[12]; /* more padding for future expansion */ /* fields only written to during inode creation */ xfs_ictimestamp_t di_crtime; /* time created */ @@ -468,8 +481,8 @@ static inline uint xfs_log_dinode_size(int version) typedef struct xfs_buf_log_format { unsigned short blf_type; /* buf log item type indicator */ unsigned short blf_size; /* size of this item */ - ushort blf_flags; /* misc state */ - ushort blf_len; /* number of blocks in this buf */ + unsigned short blf_flags; /* misc state */ + unsigned short blf_len; /* number of blocks in this buf */ __int64_t blf_blkno; /* starting blkno of this buf */ unsigned int blf_map_size; /* used size of data bitmap in words */ unsigned int blf_data_map[XFS_BLF_DATAMAP_SIZE]; /* dirty bitmap */ @@ -622,8 +635,11 @@ struct xfs_map_extent { /* rmap me_flags: upper bits are flags, lower byte is type code */ #define XFS_RMAP_EXTENT_MAP 1 +#define XFS_RMAP_EXTENT_MAP_SHARED 2 #define XFS_RMAP_EXTENT_UNMAP 3 +#define XFS_RMAP_EXTENT_UNMAP_SHARED 4 #define XFS_RMAP_EXTENT_CONVERT 5 +#define XFS_RMAP_EXTENT_CONVERT_SHARED 6 #define XFS_RMAP_EXTENT_ALLOC 7 #define XFS_RMAP_EXTENT_FREE 8 #define XFS_RMAP_EXTENT_TYPE_MASK 0xFF @@ -647,9 +663,17 @@ struct xfs_rui_log_format { __uint16_t rui_size; /* size of this item */ __uint32_t rui_nextents; /* # extents to free */ __uint64_t rui_id; /* rui identifier */ - struct xfs_map_extent rui_extents[1]; /* array of extents to rmap */ + struct xfs_map_extent rui_extents[]; /* array of extents to rmap */ }; +static inline size_t +xfs_rui_log_format_sizeof( + unsigned int nr) +{ + return sizeof(struct xfs_rui_log_format) + + nr * sizeof(struct xfs_map_extent); +} + /* * This is the structure used to lay out an rud log item in the * log. The rud_extents array is a variable size array whose @@ -663,6 +687,102 @@ struct xfs_rud_log_format { }; /* + * CUI/CUD (refcount update) log format definitions + */ +struct xfs_phys_extent { + __uint64_t pe_startblock; + __uint32_t pe_len; + __uint32_t pe_flags; +}; + +/* refcount pe_flags: upper bits are flags, lower byte is type code */ +/* Type codes are taken directly from enum xfs_refcount_intent_type. */ +#define XFS_REFCOUNT_EXTENT_TYPE_MASK 0xFF + +#define XFS_REFCOUNT_EXTENT_FLAGS (XFS_REFCOUNT_EXTENT_TYPE_MASK) + +/* + * This is the structure used to lay out a cui log item in the + * log. The cui_extents field is a variable size array whose + * size is given by cui_nextents. + */ +struct xfs_cui_log_format { + __uint16_t cui_type; /* cui log item type */ + __uint16_t cui_size; /* size of this item */ + __uint32_t cui_nextents; /* # extents to free */ + __uint64_t cui_id; /* cui identifier */ + struct xfs_phys_extent cui_extents[]; /* array of extents */ +}; + +static inline size_t +xfs_cui_log_format_sizeof( + unsigned int nr) +{ + return sizeof(struct xfs_cui_log_format) + + nr * sizeof(struct xfs_phys_extent); +} + +/* + * This is the structure used to lay out a cud log item in the + * log. The cud_extents array is a variable size array whose + * size is given by cud_nextents; + */ +struct xfs_cud_log_format { + __uint16_t cud_type; /* cud log item type */ + __uint16_t cud_size; /* size of this item */ + __uint32_t __pad; + __uint64_t cud_cui_id; /* id of corresponding cui */ +}; + +/* + * BUI/BUD (inode block mapping) log format definitions + */ + +/* bmbt me_flags: upper bits are flags, lower byte is type code */ +/* Type codes are taken directly from enum xfs_bmap_intent_type. */ +#define XFS_BMAP_EXTENT_TYPE_MASK 0xFF + +#define XFS_BMAP_EXTENT_ATTR_FORK (1U << 31) +#define XFS_BMAP_EXTENT_UNWRITTEN (1U << 30) + +#define XFS_BMAP_EXTENT_FLAGS (XFS_BMAP_EXTENT_TYPE_MASK | \ + XFS_BMAP_EXTENT_ATTR_FORK | \ + XFS_BMAP_EXTENT_UNWRITTEN) + +/* + * This is the structure used to lay out an bui log item in the + * log. The bui_extents field is a variable size array whose + * size is given by bui_nextents. + */ +struct xfs_bui_log_format { + __uint16_t bui_type; /* bui log item type */ + __uint16_t bui_size; /* size of this item */ + __uint32_t bui_nextents; /* # extents to free */ + __uint64_t bui_id; /* bui identifier */ + struct xfs_map_extent bui_extents[]; /* array of extents to bmap */ +}; + +static inline size_t +xfs_bui_log_format_sizeof( + unsigned int nr) +{ + return sizeof(struct xfs_bui_log_format) + + nr * sizeof(struct xfs_map_extent); +} + +/* + * This is the structure used to lay out an bud log item in the + * log. The bud_extents array is a variable size array whose + * size is given by bud_nextents; + */ +struct xfs_bud_log_format { + __uint16_t bud_type; /* bud log item type */ + __uint16_t bud_size; /* size of this item */ + __uint32_t __pad; + __uint64_t bud_bui_id; /* id of corresponding bui */ +}; + +/* * Dquot Log format definitions. * * The first two fields must be the type and size fitting into diff --git a/fs/xfs/libxfs/xfs_log_recover.h b/fs/xfs/libxfs/xfs_log_recover.h index 8e385f91d660..d9f65e2d5cc8 100644 --- a/fs/xfs/libxfs/xfs_log_recover.h +++ b/fs/xfs/libxfs/xfs_log_recover.h @@ -52,7 +52,7 @@ typedef struct xlog_recover { struct list_head r_itemq; /* q for items */ } xlog_recover_t; -#define ITEM_TYPE(i) (*(ushort *)(i)->ri_buf[0].i_addr) +#define ITEM_TYPE(i) (*(unsigned short *)(i)->ri_buf[0].i_addr) /* * This is the number of entries in the l_buf_cancel_table used during diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c new file mode 100644 index 000000000000..b177ef33cd4c --- /dev/null +++ b/fs/xfs/libxfs/xfs_refcount.c @@ -0,0 +1,1698 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_bmap.h" +#include "xfs_refcount_btree.h" +#include "xfs_alloc.h" +#include "xfs_error.h" +#include "xfs_trace.h" +#include "xfs_cksum.h" +#include "xfs_trans.h" +#include "xfs_bit.h" +#include "xfs_refcount.h" +#include "xfs_rmap.h" + +/* Allowable refcount adjustment amounts. */ +enum xfs_refc_adjust_op { + XFS_REFCOUNT_ADJUST_INCREASE = 1, + XFS_REFCOUNT_ADJUST_DECREASE = -1, + XFS_REFCOUNT_ADJUST_COW_ALLOC = 0, + XFS_REFCOUNT_ADJUST_COW_FREE = -1, +}; + +STATIC int __xfs_refcount_cow_alloc(struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, xfs_extlen_t aglen, + struct xfs_defer_ops *dfops); +STATIC int __xfs_refcount_cow_free(struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, xfs_extlen_t aglen, + struct xfs_defer_ops *dfops); + +/* + * Look up the first record less than or equal to [bno, len] in the btree + * given by cur. + */ +int +xfs_refcount_lookup_le( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + int *stat) +{ + trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno, + XFS_LOOKUP_LE); + cur->bc_rec.rc.rc_startblock = bno; + cur->bc_rec.rc.rc_blockcount = 0; + return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat); +} + +/* + * Look up the first record greater than or equal to [bno, len] in the btree + * given by cur. + */ +int +xfs_refcount_lookup_ge( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + int *stat) +{ + trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno, + XFS_LOOKUP_GE); + cur->bc_rec.rc.rc_startblock = bno; + cur->bc_rec.rc.rc_blockcount = 0; + return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat); +} + +/* Convert on-disk record to in-core format. */ +static inline void +xfs_refcount_btrec_to_irec( + union xfs_btree_rec *rec, + struct xfs_refcount_irec *irec) +{ + irec->rc_startblock = be32_to_cpu(rec->refc.rc_startblock); + irec->rc_blockcount = be32_to_cpu(rec->refc.rc_blockcount); + irec->rc_refcount = be32_to_cpu(rec->refc.rc_refcount); +} + +/* + * Get the data from the pointed-to record. + */ +int +xfs_refcount_get_rec( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *irec, + int *stat) +{ + union xfs_btree_rec *rec; + int error; + + error = xfs_btree_get_rec(cur, &rec, stat); + if (!error && *stat == 1) { + xfs_refcount_btrec_to_irec(rec, irec); + trace_xfs_refcount_get(cur->bc_mp, cur->bc_private.a.agno, + irec); + } + return error; +} + +/* + * Update the record referred to by cur to the value given + * by [bno, len, refcount]. + * This either works (return 0) or gets an EFSCORRUPTED error. + */ +STATIC int +xfs_refcount_update( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *irec) +{ + union xfs_btree_rec rec; + int error; + + trace_xfs_refcount_update(cur->bc_mp, cur->bc_private.a.agno, irec); + rec.refc.rc_startblock = cpu_to_be32(irec->rc_startblock); + rec.refc.rc_blockcount = cpu_to_be32(irec->rc_blockcount); + rec.refc.rc_refcount = cpu_to_be32(irec->rc_refcount); + error = xfs_btree_update(cur, &rec); + if (error) + trace_xfs_refcount_update_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Insert the record referred to by cur to the value given + * by [bno, len, refcount]. + * This either works (return 0) or gets an EFSCORRUPTED error. + */ +STATIC int +xfs_refcount_insert( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *irec, + int *i) +{ + int error; + + trace_xfs_refcount_insert(cur->bc_mp, cur->bc_private.a.agno, irec); + cur->bc_rec.rc.rc_startblock = irec->rc_startblock; + cur->bc_rec.rc.rc_blockcount = irec->rc_blockcount; + cur->bc_rec.rc.rc_refcount = irec->rc_refcount; + error = xfs_btree_insert(cur, i); + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, *i == 1, out_error); +out_error: + if (error) + trace_xfs_refcount_insert_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Remove the record referred to by cur, then set the pointer to the spot + * where the record could be re-inserted, in case we want to increment or + * decrement the cursor. + * This either works (return 0) or gets an EFSCORRUPTED error. + */ +STATIC int +xfs_refcount_delete( + struct xfs_btree_cur *cur, + int *i) +{ + struct xfs_refcount_irec irec; + int found_rec; + int error; + + error = xfs_refcount_get_rec(cur, &irec, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + trace_xfs_refcount_delete(cur->bc_mp, cur->bc_private.a.agno, &irec); + error = xfs_btree_delete(cur, i); + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, *i == 1, out_error); + if (error) + goto out_error; + error = xfs_refcount_lookup_ge(cur, irec.rc_startblock, &found_rec); +out_error: + if (error) + trace_xfs_refcount_delete_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Adjusting the Reference Count + * + * As stated elsewhere, the reference count btree (refcbt) stores + * >1 reference counts for extents of physical blocks. In this + * operation, we're either raising or lowering the reference count of + * some subrange stored in the tree: + * + * <------ adjustment range ------> + * ----+ +---+-----+ +--+--------+--------- + * 2 | | 3 | 4 | |17| 55 | 10 + * ----+ +---+-----+ +--+--------+--------- + * X axis is physical blocks number; + * reference counts are the numbers inside the rectangles + * + * The first thing we need to do is to ensure that there are no + * refcount extents crossing either boundary of the range to be + * adjusted. For any extent that does cross a boundary, split it into + * two extents so that we can increment the refcount of one of the + * pieces later: + * + * <------ adjustment range ------> + * ----+ +---+-----+ +--+--------+----+---- + * 2 | | 3 | 2 | |17| 55 | 10 | 10 + * ----+ +---+-----+ +--+--------+----+---- + * + * For this next step, let's assume that all the physical blocks in + * the adjustment range are mapped to a file and are therefore in use + * at least once. Therefore, we can infer that any gap in the + * refcount tree within the adjustment range represents a physical + * extent with refcount == 1: + * + * <------ adjustment range ------> + * ----+---+---+-----+-+--+--------+----+---- + * 2 |"1"| 3 | 2 |1|17| 55 | 10 | 10 + * ----+---+---+-----+-+--+--------+----+---- + * ^ + * + * For each extent that falls within the interval range, figure out + * which extent is to the left or the right of that extent. Now we + * have a left, current, and right extent. If the new reference count + * of the center extent enables us to merge left, center, and right + * into one record covering all three, do so. If the center extent is + * at the left end of the range, abuts the left extent, and its new + * reference count matches the left extent's record, then merge them. + * If the center extent is at the right end of the range, abuts the + * right extent, and the reference counts match, merge those. In the + * example, we can left merge (assuming an increment operation): + * + * <------ adjustment range ------> + * --------+---+-----+-+--+--------+----+---- + * 2 | 3 | 2 |1|17| 55 | 10 | 10 + * --------+---+-----+-+--+--------+----+---- + * ^ + * + * For all other extents within the range, adjust the reference count + * or delete it if the refcount falls below 2. If we were + * incrementing, the end result looks like this: + * + * <------ adjustment range ------> + * --------+---+-----+-+--+--------+----+---- + * 2 | 4 | 3 |2|18| 56 | 11 | 10 + * --------+---+-----+-+--+--------+----+---- + * + * The result of a decrement operation looks as such: + * + * <------ adjustment range ------> + * ----+ +---+ +--+--------+----+---- + * 2 | | 2 | |16| 54 | 9 | 10 + * ----+ +---+ +--+--------+----+---- + * DDDD 111111DD + * + * The blocks marked "D" are freed; the blocks marked "1" are only + * referenced once and therefore the record is removed from the + * refcount btree. + */ + +/* Next block after this extent. */ +static inline xfs_agblock_t +xfs_refc_next( + struct xfs_refcount_irec *rc) +{ + return rc->rc_startblock + rc->rc_blockcount; +} + +/* + * Split a refcount extent that crosses agbno. + */ +STATIC int +xfs_refcount_split_extent( + struct xfs_btree_cur *cur, + xfs_agblock_t agbno, + bool *shape_changed) +{ + struct xfs_refcount_irec rcext, tmp; + int found_rec; + int error; + + *shape_changed = false; + error = xfs_refcount_lookup_le(cur, agbno, &found_rec); + if (error) + goto out_error; + if (!found_rec) + return 0; + + error = xfs_refcount_get_rec(cur, &rcext, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + if (rcext.rc_startblock == agbno || xfs_refc_next(&rcext) <= agbno) + return 0; + + *shape_changed = true; + trace_xfs_refcount_split_extent(cur->bc_mp, cur->bc_private.a.agno, + &rcext, agbno); + + /* Establish the right extent. */ + tmp = rcext; + tmp.rc_startblock = agbno; + tmp.rc_blockcount -= (agbno - rcext.rc_startblock); + error = xfs_refcount_update(cur, &tmp); + if (error) + goto out_error; + + /* Insert the left extent. */ + tmp = rcext; + tmp.rc_blockcount = agbno - rcext.rc_startblock; + error = xfs_refcount_insert(cur, &tmp, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + return error; + +out_error: + trace_xfs_refcount_split_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Merge the left, center, and right extents. + */ +STATIC int +xfs_refcount_merge_center_extents( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *left, + struct xfs_refcount_irec *center, + struct xfs_refcount_irec *right, + unsigned long long extlen, + xfs_agblock_t *agbno, + xfs_extlen_t *aglen) +{ + int error; + int found_rec; + + trace_xfs_refcount_merge_center_extents(cur->bc_mp, + cur->bc_private.a.agno, left, center, right); + + /* + * Make sure the center and right extents are not in the btree. + * If the center extent was synthesized, the first delete call + * removes the right extent and we skip the second deletion. + * If center and right were in the btree, then the first delete + * call removes the center and the second one removes the right + * extent. + */ + error = xfs_refcount_lookup_ge(cur, center->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + if (center->rc_refcount > 1) { + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + } + + /* Enlarge the left extent. */ + error = xfs_refcount_lookup_le(cur, left->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + left->rc_blockcount = extlen; + error = xfs_refcount_update(cur, left); + if (error) + goto out_error; + + *aglen = 0; + return error; + +out_error: + trace_xfs_refcount_merge_center_extents_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Merge with the left extent. + */ +STATIC int +xfs_refcount_merge_left_extent( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *left, + struct xfs_refcount_irec *cleft, + xfs_agblock_t *agbno, + xfs_extlen_t *aglen) +{ + int error; + int found_rec; + + trace_xfs_refcount_merge_left_extent(cur->bc_mp, + cur->bc_private.a.agno, left, cleft); + + /* If the extent at agbno (cleft) wasn't synthesized, remove it. */ + if (cleft->rc_refcount > 1) { + error = xfs_refcount_lookup_le(cur, cleft->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + } + + /* Enlarge the left extent. */ + error = xfs_refcount_lookup_le(cur, left->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + left->rc_blockcount += cleft->rc_blockcount; + error = xfs_refcount_update(cur, left); + if (error) + goto out_error; + + *agbno += cleft->rc_blockcount; + *aglen -= cleft->rc_blockcount; + return error; + +out_error: + trace_xfs_refcount_merge_left_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Merge with the right extent. + */ +STATIC int +xfs_refcount_merge_right_extent( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *right, + struct xfs_refcount_irec *cright, + xfs_agblock_t *agbno, + xfs_extlen_t *aglen) +{ + int error; + int found_rec; + + trace_xfs_refcount_merge_right_extent(cur->bc_mp, + cur->bc_private.a.agno, cright, right); + + /* + * If the extent ending at agbno+aglen (cright) wasn't synthesized, + * remove it. + */ + if (cright->rc_refcount > 1) { + error = xfs_refcount_lookup_le(cur, cright->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + } + + /* Enlarge the right extent. */ + error = xfs_refcount_lookup_le(cur, right->rc_startblock, + &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + right->rc_startblock -= cright->rc_blockcount; + right->rc_blockcount += cright->rc_blockcount; + error = xfs_refcount_update(cur, right); + if (error) + goto out_error; + + *aglen -= cright->rc_blockcount; + return error; + +out_error: + trace_xfs_refcount_merge_right_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +#define XFS_FIND_RCEXT_SHARED 1 +#define XFS_FIND_RCEXT_COW 2 +/* + * Find the left extent and the one after it (cleft). This function assumes + * that we've already split any extent crossing agbno. + */ +STATIC int +xfs_refcount_find_left_extents( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *left, + struct xfs_refcount_irec *cleft, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + int flags) +{ + struct xfs_refcount_irec tmp; + int error; + int found_rec; + + left->rc_startblock = cleft->rc_startblock = NULLAGBLOCK; + error = xfs_refcount_lookup_le(cur, agbno - 1, &found_rec); + if (error) + goto out_error; + if (!found_rec) + return 0; + + error = xfs_refcount_get_rec(cur, &tmp, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + if (xfs_refc_next(&tmp) != agbno) + return 0; + if ((flags & XFS_FIND_RCEXT_SHARED) && tmp.rc_refcount < 2) + return 0; + if ((flags & XFS_FIND_RCEXT_COW) && tmp.rc_refcount > 1) + return 0; + /* We have a left extent; retrieve (or invent) the next right one */ + *left = tmp; + + error = xfs_btree_increment(cur, 0, &found_rec); + if (error) + goto out_error; + if (found_rec) { + error = xfs_refcount_get_rec(cur, &tmp, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + + /* if tmp starts at the end of our range, just use that */ + if (tmp.rc_startblock == agbno) + *cleft = tmp; + else { + /* + * There's a gap in the refcntbt at the start of the + * range we're interested in (refcount == 1) so + * synthesize the implied extent and pass it back. + * We assume here that the agbno/aglen range was + * passed in from a data fork extent mapping and + * therefore is allocated to exactly one owner. + */ + cleft->rc_startblock = agbno; + cleft->rc_blockcount = min(aglen, + tmp.rc_startblock - agbno); + cleft->rc_refcount = 1; + } + } else { + /* + * No extents, so pretend that there's one covering the whole + * range. + */ + cleft->rc_startblock = agbno; + cleft->rc_blockcount = aglen; + cleft->rc_refcount = 1; + } + trace_xfs_refcount_find_left_extent(cur->bc_mp, cur->bc_private.a.agno, + left, cleft, agbno); + return error; + +out_error: + trace_xfs_refcount_find_left_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Find the right extent and the one before it (cright). This function + * assumes that we've already split any extents crossing agbno + aglen. + */ +STATIC int +xfs_refcount_find_right_extents( + struct xfs_btree_cur *cur, + struct xfs_refcount_irec *right, + struct xfs_refcount_irec *cright, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + int flags) +{ + struct xfs_refcount_irec tmp; + int error; + int found_rec; + + right->rc_startblock = cright->rc_startblock = NULLAGBLOCK; + error = xfs_refcount_lookup_ge(cur, agbno + aglen, &found_rec); + if (error) + goto out_error; + if (!found_rec) + return 0; + + error = xfs_refcount_get_rec(cur, &tmp, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, out_error); + + if (tmp.rc_startblock != agbno + aglen) + return 0; + if ((flags & XFS_FIND_RCEXT_SHARED) && tmp.rc_refcount < 2) + return 0; + if ((flags & XFS_FIND_RCEXT_COW) && tmp.rc_refcount > 1) + return 0; + /* We have a right extent; retrieve (or invent) the next left one */ + *right = tmp; + + error = xfs_btree_decrement(cur, 0, &found_rec); + if (error) + goto out_error; + if (found_rec) { + error = xfs_refcount_get_rec(cur, &tmp, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, found_rec == 1, + out_error); + + /* if tmp ends at the end of our range, just use that */ + if (xfs_refc_next(&tmp) == agbno + aglen) + *cright = tmp; + else { + /* + * There's a gap in the refcntbt at the end of the + * range we're interested in (refcount == 1) so + * create the implied extent and pass it back. + * We assume here that the agbno/aglen range was + * passed in from a data fork extent mapping and + * therefore is allocated to exactly one owner. + */ + cright->rc_startblock = max(agbno, xfs_refc_next(&tmp)); + cright->rc_blockcount = right->rc_startblock - + cright->rc_startblock; + cright->rc_refcount = 1; + } + } else { + /* + * No extents, so pretend that there's one covering the whole + * range. + */ + cright->rc_startblock = agbno; + cright->rc_blockcount = aglen; + cright->rc_refcount = 1; + } + trace_xfs_refcount_find_right_extent(cur->bc_mp, cur->bc_private.a.agno, + cright, right, agbno + aglen); + return error; + +out_error: + trace_xfs_refcount_find_right_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* Is this extent valid? */ +static inline bool +xfs_refc_valid( + struct xfs_refcount_irec *rc) +{ + return rc->rc_startblock != NULLAGBLOCK; +} + +/* + * Try to merge with any extents on the boundaries of the adjustment range. + */ +STATIC int +xfs_refcount_merge_extents( + struct xfs_btree_cur *cur, + xfs_agblock_t *agbno, + xfs_extlen_t *aglen, + enum xfs_refc_adjust_op adjust, + int flags, + bool *shape_changed) +{ + struct xfs_refcount_irec left = {0}, cleft = {0}; + struct xfs_refcount_irec cright = {0}, right = {0}; + int error; + unsigned long long ulen; + bool cequal; + + *shape_changed = false; + /* + * Find the extent just below agbno [left], just above agbno [cleft], + * just below (agbno + aglen) [cright], and just above (agbno + aglen) + * [right]. + */ + error = xfs_refcount_find_left_extents(cur, &left, &cleft, *agbno, + *aglen, flags); + if (error) + return error; + error = xfs_refcount_find_right_extents(cur, &right, &cright, *agbno, + *aglen, flags); + if (error) + return error; + + /* No left or right extent to merge; exit. */ + if (!xfs_refc_valid(&left) && !xfs_refc_valid(&right)) + return 0; + + cequal = (cleft.rc_startblock == cright.rc_startblock) && + (cleft.rc_blockcount == cright.rc_blockcount); + + /* Try to merge left, cleft, and right. cleft must == cright. */ + ulen = (unsigned long long)left.rc_blockcount + cleft.rc_blockcount + + right.rc_blockcount; + if (xfs_refc_valid(&left) && xfs_refc_valid(&right) && + xfs_refc_valid(&cleft) && xfs_refc_valid(&cright) && cequal && + left.rc_refcount == cleft.rc_refcount + adjust && + right.rc_refcount == cleft.rc_refcount + adjust && + ulen < MAXREFCEXTLEN) { + *shape_changed = true; + return xfs_refcount_merge_center_extents(cur, &left, &cleft, + &right, ulen, agbno, aglen); + } + + /* Try to merge left and cleft. */ + ulen = (unsigned long long)left.rc_blockcount + cleft.rc_blockcount; + if (xfs_refc_valid(&left) && xfs_refc_valid(&cleft) && + left.rc_refcount == cleft.rc_refcount + adjust && + ulen < MAXREFCEXTLEN) { + *shape_changed = true; + error = xfs_refcount_merge_left_extent(cur, &left, &cleft, + agbno, aglen); + if (error) + return error; + + /* + * If we just merged left + cleft and cleft == cright, + * we no longer have a cright to merge with right. We're done. + */ + if (cequal) + return 0; + } + + /* Try to merge cright and right. */ + ulen = (unsigned long long)right.rc_blockcount + cright.rc_blockcount; + if (xfs_refc_valid(&right) && xfs_refc_valid(&cright) && + right.rc_refcount == cright.rc_refcount + adjust && + ulen < MAXREFCEXTLEN) { + *shape_changed = true; + return xfs_refcount_merge_right_extent(cur, &right, &cright, + agbno, aglen); + } + + return error; +} + +/* + * While we're adjusting the refcounts records of an extent, we have + * to keep an eye on the number of extents we're dirtying -- run too + * many in a single transaction and we'll exceed the transaction's + * reservation and crash the fs. Each record adds 12 bytes to the + * log (plus any key updates) so we'll conservatively assume 24 bytes + * per record. We must also leave space for btree splits on both ends + * of the range and space for the CUD and a new CUI. + * + * XXX: This is a pretty hand-wavy estimate. The penalty for guessing + * true incorrectly is a shutdown FS; the penalty for guessing false + * incorrectly is more transaction rolls than might be necessary. + * Be conservative here. + */ +static bool +xfs_refcount_still_have_space( + struct xfs_btree_cur *cur) +{ + unsigned long overhead; + + overhead = cur->bc_private.a.priv.refc.shape_changes * + xfs_allocfree_log_count(cur->bc_mp, 1); + overhead *= cur->bc_mp->m_sb.sb_blocksize; + + /* + * Only allow 2 refcount extent updates per transaction if the + * refcount continue update "error" has been injected. + */ + if (cur->bc_private.a.priv.refc.nr_ops > 2 && + XFS_TEST_ERROR(false, cur->bc_mp, + XFS_ERRTAG_REFCOUNT_CONTINUE_UPDATE, + XFS_RANDOM_REFCOUNT_CONTINUE_UPDATE)) + return false; + + if (cur->bc_private.a.priv.refc.nr_ops == 0) + return true; + else if (overhead > cur->bc_tp->t_log_res) + return false; + return cur->bc_tp->t_log_res - overhead > + cur->bc_private.a.priv.refc.nr_ops * 32; +} + +/* + * Adjust the refcounts of middle extents. At this point we should have + * split extents that crossed the adjustment range; merged with adjacent + * extents; and updated agbno/aglen to reflect the merges. Therefore, + * all we have to do is update the extents inside [agbno, agbno + aglen]. + */ +STATIC int +xfs_refcount_adjust_extents( + struct xfs_btree_cur *cur, + xfs_agblock_t *agbno, + xfs_extlen_t *aglen, + enum xfs_refc_adjust_op adj, + struct xfs_defer_ops *dfops, + struct xfs_owner_info *oinfo) +{ + struct xfs_refcount_irec ext, tmp; + int error; + int found_rec, found_tmp; + xfs_fsblock_t fsbno; + + /* Merging did all the work already. */ + if (*aglen == 0) + return 0; + + error = xfs_refcount_lookup_ge(cur, *agbno, &found_rec); + if (error) + goto out_error; + + while (*aglen > 0 && xfs_refcount_still_have_space(cur)) { + error = xfs_refcount_get_rec(cur, &ext, &found_rec); + if (error) + goto out_error; + if (!found_rec) { + ext.rc_startblock = cur->bc_mp->m_sb.sb_agblocks; + ext.rc_blockcount = 0; + ext.rc_refcount = 0; + } + + /* + * Deal with a hole in the refcount tree; if a file maps to + * these blocks and there's no refcountbt record, pretend that + * there is one with refcount == 1. + */ + if (ext.rc_startblock != *agbno) { + tmp.rc_startblock = *agbno; + tmp.rc_blockcount = min(*aglen, + ext.rc_startblock - *agbno); + tmp.rc_refcount = 1 + adj; + trace_xfs_refcount_modify_extent(cur->bc_mp, + cur->bc_private.a.agno, &tmp); + + /* + * Either cover the hole (increment) or + * delete the range (decrement). + */ + if (tmp.rc_refcount) { + error = xfs_refcount_insert(cur, &tmp, + &found_tmp); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + found_tmp == 1, out_error); + cur->bc_private.a.priv.refc.nr_ops++; + } else { + fsbno = XFS_AGB_TO_FSB(cur->bc_mp, + cur->bc_private.a.agno, + tmp.rc_startblock); + xfs_bmap_add_free(cur->bc_mp, dfops, fsbno, + tmp.rc_blockcount, oinfo); + } + + (*agbno) += tmp.rc_blockcount; + (*aglen) -= tmp.rc_blockcount; + + error = xfs_refcount_lookup_ge(cur, *agbno, + &found_rec); + if (error) + goto out_error; + } + + /* Stop if there's nothing left to modify */ + if (*aglen == 0 || !xfs_refcount_still_have_space(cur)) + break; + + /* + * Adjust the reference count and either update the tree + * (incr) or free the blocks (decr). + */ + if (ext.rc_refcount == MAXREFCOUNT) + goto skip; + ext.rc_refcount += adj; + trace_xfs_refcount_modify_extent(cur->bc_mp, + cur->bc_private.a.agno, &ext); + if (ext.rc_refcount > 1) { + error = xfs_refcount_update(cur, &ext); + if (error) + goto out_error; + cur->bc_private.a.priv.refc.nr_ops++; + } else if (ext.rc_refcount == 1) { + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + found_rec == 1, out_error); + cur->bc_private.a.priv.refc.nr_ops++; + goto advloop; + } else { + fsbno = XFS_AGB_TO_FSB(cur->bc_mp, + cur->bc_private.a.agno, + ext.rc_startblock); + xfs_bmap_add_free(cur->bc_mp, dfops, fsbno, + ext.rc_blockcount, oinfo); + } + +skip: + error = xfs_btree_increment(cur, 0, &found_rec); + if (error) + goto out_error; + +advloop: + (*agbno) += ext.rc_blockcount; + (*aglen) -= ext.rc_blockcount; + } + + return error; +out_error: + trace_xfs_refcount_modify_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* Adjust the reference count of a range of AG blocks. */ +STATIC int +xfs_refcount_adjust( + struct xfs_btree_cur *cur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + xfs_agblock_t *new_agbno, + xfs_extlen_t *new_aglen, + enum xfs_refc_adjust_op adj, + struct xfs_defer_ops *dfops, + struct xfs_owner_info *oinfo) +{ + bool shape_changed; + int shape_changes = 0; + int error; + + *new_agbno = agbno; + *new_aglen = aglen; + if (adj == XFS_REFCOUNT_ADJUST_INCREASE) + trace_xfs_refcount_increase(cur->bc_mp, cur->bc_private.a.agno, + agbno, aglen); + else + trace_xfs_refcount_decrease(cur->bc_mp, cur->bc_private.a.agno, + agbno, aglen); + + /* + * Ensure that no rcextents cross the boundary of the adjustment range. + */ + error = xfs_refcount_split_extent(cur, agbno, &shape_changed); + if (error) + goto out_error; + if (shape_changed) + shape_changes++; + + error = xfs_refcount_split_extent(cur, agbno + aglen, &shape_changed); + if (error) + goto out_error; + if (shape_changed) + shape_changes++; + + /* + * Try to merge with the left or right extents of the range. + */ + error = xfs_refcount_merge_extents(cur, new_agbno, new_aglen, adj, + XFS_FIND_RCEXT_SHARED, &shape_changed); + if (error) + goto out_error; + if (shape_changed) + shape_changes++; + if (shape_changes) + cur->bc_private.a.priv.refc.shape_changes++; + + /* Now that we've taken care of the ends, adjust the middle extents */ + error = xfs_refcount_adjust_extents(cur, new_agbno, new_aglen, + adj, dfops, oinfo); + if (error) + goto out_error; + + return 0; + +out_error: + trace_xfs_refcount_adjust_error(cur->bc_mp, cur->bc_private.a.agno, + error, _RET_IP_); + return error; +} + +/* Clean up after calling xfs_refcount_finish_one. */ +void +xfs_refcount_finish_one_cleanup( + struct xfs_trans *tp, + struct xfs_btree_cur *rcur, + int error) +{ + struct xfs_buf *agbp; + + if (rcur == NULL) + return; + agbp = rcur->bc_private.a.agbp; + xfs_btree_del_cursor(rcur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); + if (error) + xfs_trans_brelse(tp, agbp); +} + +/* + * Process one of the deferred refcount operations. We pass back the + * btree cursor to maintain our lock on the btree between calls. + * This saves time and eliminates a buffer deadlock between the + * superblock and the AGF because we'll always grab them in the same + * order. + */ +int +xfs_refcount_finish_one( + struct xfs_trans *tp, + struct xfs_defer_ops *dfops, + enum xfs_refcount_intent_type type, + xfs_fsblock_t startblock, + xfs_extlen_t blockcount, + xfs_fsblock_t *new_fsb, + xfs_extlen_t *new_len, + struct xfs_btree_cur **pcur) +{ + struct xfs_mount *mp = tp->t_mountp; + struct xfs_btree_cur *rcur; + struct xfs_buf *agbp = NULL; + int error = 0; + xfs_agnumber_t agno; + xfs_agblock_t bno; + xfs_agblock_t new_agbno; + unsigned long nr_ops = 0; + int shape_changes = 0; + + agno = XFS_FSB_TO_AGNO(mp, startblock); + ASSERT(agno != NULLAGNUMBER); + bno = XFS_FSB_TO_AGBNO(mp, startblock); + + trace_xfs_refcount_deferred(mp, XFS_FSB_TO_AGNO(mp, startblock), + type, XFS_FSB_TO_AGBNO(mp, startblock), + blockcount); + + if (XFS_TEST_ERROR(false, mp, + XFS_ERRTAG_REFCOUNT_FINISH_ONE, + XFS_RANDOM_REFCOUNT_FINISH_ONE)) + return -EIO; + + /* + * If we haven't gotten a cursor or the cursor AG doesn't match + * the startblock, get one now. + */ + rcur = *pcur; + if (rcur != NULL && rcur->bc_private.a.agno != agno) { + nr_ops = rcur->bc_private.a.priv.refc.nr_ops; + shape_changes = rcur->bc_private.a.priv.refc.shape_changes; + xfs_refcount_finish_one_cleanup(tp, rcur, 0); + rcur = NULL; + *pcur = NULL; + } + if (rcur == NULL) { + error = xfs_alloc_read_agf(tp->t_mountp, tp, agno, + XFS_ALLOC_FLAG_FREEING, &agbp); + if (error) + return error; + if (!agbp) + return -EFSCORRUPTED; + + rcur = xfs_refcountbt_init_cursor(mp, tp, agbp, agno, dfops); + if (!rcur) { + error = -ENOMEM; + goto out_cur; + } + rcur->bc_private.a.priv.refc.nr_ops = nr_ops; + rcur->bc_private.a.priv.refc.shape_changes = shape_changes; + } + *pcur = rcur; + + switch (type) { + case XFS_REFCOUNT_INCREASE: + error = xfs_refcount_adjust(rcur, bno, blockcount, &new_agbno, + new_len, XFS_REFCOUNT_ADJUST_INCREASE, dfops, NULL); + *new_fsb = XFS_AGB_TO_FSB(mp, agno, new_agbno); + break; + case XFS_REFCOUNT_DECREASE: + error = xfs_refcount_adjust(rcur, bno, blockcount, &new_agbno, + new_len, XFS_REFCOUNT_ADJUST_DECREASE, dfops, NULL); + *new_fsb = XFS_AGB_TO_FSB(mp, agno, new_agbno); + break; + case XFS_REFCOUNT_ALLOC_COW: + *new_fsb = startblock + blockcount; + *new_len = 0; + error = __xfs_refcount_cow_alloc(rcur, bno, blockcount, dfops); + break; + case XFS_REFCOUNT_FREE_COW: + *new_fsb = startblock + blockcount; + *new_len = 0; + error = __xfs_refcount_cow_free(rcur, bno, blockcount, dfops); + break; + default: + ASSERT(0); + error = -EFSCORRUPTED; + } + if (!error && *new_len > 0) + trace_xfs_refcount_finish_one_leftover(mp, agno, type, + bno, blockcount, new_agbno, *new_len); + return error; + +out_cur: + xfs_trans_brelse(tp, agbp); + + return error; +} + +/* + * Record a refcount intent for later processing. + */ +static int +__xfs_refcount_add( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + enum xfs_refcount_intent_type type, + xfs_fsblock_t startblock, + xfs_extlen_t blockcount) +{ + struct xfs_refcount_intent *ri; + + trace_xfs_refcount_defer(mp, XFS_FSB_TO_AGNO(mp, startblock), + type, XFS_FSB_TO_AGBNO(mp, startblock), + blockcount); + + ri = kmem_alloc(sizeof(struct xfs_refcount_intent), + KM_SLEEP | KM_NOFS); + INIT_LIST_HEAD(&ri->ri_list); + ri->ri_type = type; + ri->ri_startblock = startblock; + ri->ri_blockcount = blockcount; + + xfs_defer_add(dfops, XFS_DEFER_OPS_TYPE_REFCOUNT, &ri->ri_list); + return 0; +} + +/* + * Increase the reference count of the blocks backing a file's extent. + */ +int +xfs_refcount_increase_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_INCREASE, + PREV->br_startblock, PREV->br_blockcount); +} + +/* + * Decrease the reference count of the blocks backing a file's extent. + */ +int +xfs_refcount_decrease_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + struct xfs_bmbt_irec *PREV) +{ + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_DECREASE, + PREV->br_startblock, PREV->br_blockcount); +} + +/* + * Given an AG extent, find the lowest-numbered run of shared blocks + * within that range and return the range in fbno/flen. If + * find_end_of_shared is set, return the longest contiguous extent of + * shared blocks; if not, just return the first extent we find. If no + * shared blocks are found, fbno and flen will be set to NULLAGBLOCK + * and 0, respectively. + */ +int +xfs_refcount_find_shared( + struct xfs_btree_cur *cur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + xfs_agblock_t *fbno, + xfs_extlen_t *flen, + bool find_end_of_shared) +{ + struct xfs_refcount_irec tmp; + int i; + int have; + int error; + + trace_xfs_refcount_find_shared(cur->bc_mp, cur->bc_private.a.agno, + agbno, aglen); + + /* By default, skip the whole range */ + *fbno = NULLAGBLOCK; + *flen = 0; + + /* Try to find a refcount extent that crosses the start */ + error = xfs_refcount_lookup_le(cur, agbno, &have); + if (error) + goto out_error; + if (!have) { + /* No left extent, look at the next one */ + error = xfs_btree_increment(cur, 0, &have); + if (error) + goto out_error; + if (!have) + goto done; + } + error = xfs_refcount_get_rec(cur, &tmp, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, out_error); + + /* If the extent ends before the start, look at the next one */ + if (tmp.rc_startblock + tmp.rc_blockcount <= agbno) { + error = xfs_btree_increment(cur, 0, &have); + if (error) + goto out_error; + if (!have) + goto done; + error = xfs_refcount_get_rec(cur, &tmp, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, out_error); + } + + /* If the extent starts after the range we want, bail out */ + if (tmp.rc_startblock >= agbno + aglen) + goto done; + + /* We found the start of a shared extent! */ + if (tmp.rc_startblock < agbno) { + tmp.rc_blockcount -= (agbno - tmp.rc_startblock); + tmp.rc_startblock = agbno; + } + + *fbno = tmp.rc_startblock; + *flen = min(tmp.rc_blockcount, agbno + aglen - *fbno); + if (!find_end_of_shared) + goto done; + + /* Otherwise, find the end of this shared extent */ + while (*fbno + *flen < agbno + aglen) { + error = xfs_btree_increment(cur, 0, &have); + if (error) + goto out_error; + if (!have) + break; + error = xfs_refcount_get_rec(cur, &tmp, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, i == 1, out_error); + if (tmp.rc_startblock >= agbno + aglen || + tmp.rc_startblock != *fbno + *flen) + break; + *flen = min(*flen + tmp.rc_blockcount, agbno + aglen - *fbno); + } + +done: + trace_xfs_refcount_find_shared_result(cur->bc_mp, + cur->bc_private.a.agno, *fbno, *flen); + +out_error: + if (error) + trace_xfs_refcount_find_shared_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Recovering CoW Blocks After a Crash + * + * Due to the way that the copy on write mechanism works, there's a window of + * opportunity in which we can lose track of allocated blocks during a crash. + * Because CoW uses delayed allocation in the in-core CoW fork, writeback + * causes blocks to be allocated and stored in the CoW fork. The blocks are + * no longer in the free space btree but are not otherwise recorded anywhere + * until the write completes and the blocks are mapped into the file. A crash + * in between allocation and remapping results in the replacement blocks being + * lost. This situation is exacerbated by the CoW extent size hint because + * allocations can hang around for long time. + * + * However, there is a place where we can record these allocations before they + * become mappings -- the reference count btree. The btree does not record + * extents with refcount == 1, so we can record allocations with a refcount of + * 1. Blocks being used for CoW writeout cannot be shared, so there should be + * no conflict with shared block records. These mappings should be created + * when we allocate blocks to the CoW fork and deleted when they're removed + * from the CoW fork. + * + * Minor nit: records for in-progress CoW allocations and records for shared + * extents must never be merged, to preserve the property that (except for CoW + * allocations) there are no refcount btree entries with refcount == 1. The + * only time this could potentially happen is when unsharing a block that's + * adjacent to CoW allocations, so we must be careful to avoid this. + * + * At mount time we recover lost CoW allocations by searching the refcount + * btree for these refcount == 1 mappings. These represent CoW allocations + * that were in progress at the time the filesystem went down, so we can free + * them to get the space back. + * + * This mechanism is superior to creating EFIs for unmapped CoW extents for + * several reasons -- first, EFIs pin the tail of the log and would have to be + * periodically relogged to avoid filling up the log. Second, CoW completions + * will have to file an EFD and create new EFIs for whatever remains in the + * CoW fork; this partially takes care of (1) but extent-size reservations + * will have to periodically relog even if there's no writeout in progress. + * This can happen if the CoW extent size hint is set, which you really want. + * Third, EFIs cannot currently be automatically relogged into newer + * transactions to advance the log tail. Fourth, stuffing the log full of + * EFIs places an upper bound on the number of CoW allocations that can be + * held filesystem-wide at any given time. Recording them in the refcount + * btree doesn't require us to maintain any state in memory and doesn't pin + * the log. + */ +/* + * Adjust the refcounts of CoW allocations. These allocations are "magic" + * in that they're not referenced anywhere else in the filesystem, so we + * stash them in the refcount btree with a refcount of 1 until either file + * remapping (or CoW cancellation) happens. + */ +STATIC int +xfs_refcount_adjust_cow_extents( + struct xfs_btree_cur *cur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + enum xfs_refc_adjust_op adj, + struct xfs_defer_ops *dfops, + struct xfs_owner_info *oinfo) +{ + struct xfs_refcount_irec ext, tmp; + int error; + int found_rec, found_tmp; + + if (aglen == 0) + return 0; + + /* Find any overlapping refcount records */ + error = xfs_refcount_lookup_ge(cur, agbno, &found_rec); + if (error) + goto out_error; + error = xfs_refcount_get_rec(cur, &ext, &found_rec); + if (error) + goto out_error; + if (!found_rec) { + ext.rc_startblock = cur->bc_mp->m_sb.sb_agblocks + + XFS_REFC_COW_START; + ext.rc_blockcount = 0; + ext.rc_refcount = 0; + } + + switch (adj) { + case XFS_REFCOUNT_ADJUST_COW_ALLOC: + /* Adding a CoW reservation, there should be nothing here. */ + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + ext.rc_startblock >= agbno + aglen, out_error); + + tmp.rc_startblock = agbno; + tmp.rc_blockcount = aglen; + tmp.rc_refcount = 1; + trace_xfs_refcount_modify_extent(cur->bc_mp, + cur->bc_private.a.agno, &tmp); + + error = xfs_refcount_insert(cur, &tmp, + &found_tmp); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + found_tmp == 1, out_error); + break; + case XFS_REFCOUNT_ADJUST_COW_FREE: + /* Removing a CoW reservation, there should be one extent. */ + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + ext.rc_startblock == agbno, out_error); + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + ext.rc_blockcount == aglen, out_error); + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + ext.rc_refcount == 1, out_error); + + ext.rc_refcount = 0; + trace_xfs_refcount_modify_extent(cur->bc_mp, + cur->bc_private.a.agno, &ext); + error = xfs_refcount_delete(cur, &found_rec); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(cur->bc_mp, + found_rec == 1, out_error); + break; + default: + ASSERT(0); + } + + return error; +out_error: + trace_xfs_refcount_modify_extent_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Add or remove refcount btree entries for CoW reservations. + */ +STATIC int +xfs_refcount_adjust_cow( + struct xfs_btree_cur *cur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + enum xfs_refc_adjust_op adj, + struct xfs_defer_ops *dfops) +{ + bool shape_changed; + int error; + + agbno += XFS_REFC_COW_START; + + /* + * Ensure that no rcextents cross the boundary of the adjustment range. + */ + error = xfs_refcount_split_extent(cur, agbno, &shape_changed); + if (error) + goto out_error; + + error = xfs_refcount_split_extent(cur, agbno + aglen, &shape_changed); + if (error) + goto out_error; + + /* + * Try to merge with the left or right extents of the range. + */ + error = xfs_refcount_merge_extents(cur, &agbno, &aglen, adj, + XFS_FIND_RCEXT_COW, &shape_changed); + if (error) + goto out_error; + + /* Now that we've taken care of the ends, adjust the middle extents */ + error = xfs_refcount_adjust_cow_extents(cur, agbno, aglen, adj, + dfops, NULL); + if (error) + goto out_error; + + return 0; + +out_error: + trace_xfs_refcount_adjust_cow_error(cur->bc_mp, cur->bc_private.a.agno, + error, _RET_IP_); + return error; +} + +/* + * Record a CoW allocation in the refcount btree. + */ +STATIC int +__xfs_refcount_cow_alloc( + struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + struct xfs_defer_ops *dfops) +{ + int error; + + trace_xfs_refcount_cow_increase(rcur->bc_mp, rcur->bc_private.a.agno, + agbno, aglen); + + /* Add refcount btree reservation */ + error = xfs_refcount_adjust_cow(rcur, agbno, aglen, + XFS_REFCOUNT_ADJUST_COW_ALLOC, dfops); + if (error) + return error; + + /* Add rmap entry */ + if (xfs_sb_version_hasrmapbt(&rcur->bc_mp->m_sb)) { + error = xfs_rmap_alloc_extent(rcur->bc_mp, dfops, + rcur->bc_private.a.agno, + agbno, aglen, XFS_RMAP_OWN_COW); + if (error) + return error; + } + + return error; +} + +/* + * Remove a CoW allocation from the refcount btree. + */ +STATIC int +__xfs_refcount_cow_free( + struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + struct xfs_defer_ops *dfops) +{ + int error; + + trace_xfs_refcount_cow_decrease(rcur->bc_mp, rcur->bc_private.a.agno, + agbno, aglen); + + /* Remove refcount btree reservation */ + error = xfs_refcount_adjust_cow(rcur, agbno, aglen, + XFS_REFCOUNT_ADJUST_COW_FREE, dfops); + if (error) + return error; + + /* Remove rmap entry */ + if (xfs_sb_version_hasrmapbt(&rcur->bc_mp->m_sb)) { + error = xfs_rmap_free_extent(rcur->bc_mp, dfops, + rcur->bc_private.a.agno, + agbno, aglen, XFS_RMAP_OWN_COW); + if (error) + return error; + } + + return error; +} + +/* Record a CoW staging extent in the refcount btree. */ +int +xfs_refcount_alloc_cow_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + xfs_fsblock_t fsb, + xfs_extlen_t len) +{ + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_ALLOC_COW, + fsb, len); +} + +/* Forget a CoW staging event in the refcount btree. */ +int +xfs_refcount_free_cow_extent( + struct xfs_mount *mp, + struct xfs_defer_ops *dfops, + xfs_fsblock_t fsb, + xfs_extlen_t len) +{ + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_FREE_COW, + fsb, len); +} + +struct xfs_refcount_recovery { + struct list_head rr_list; + struct xfs_refcount_irec rr_rrec; +}; + +/* Stuff an extent on the recovery list. */ +STATIC int +xfs_refcount_recover_extent( + struct xfs_btree_cur *cur, + union xfs_btree_rec *rec, + void *priv) +{ + struct list_head *debris = priv; + struct xfs_refcount_recovery *rr; + + if (be32_to_cpu(rec->refc.rc_refcount) != 1) + return -EFSCORRUPTED; + + rr = kmem_alloc(sizeof(struct xfs_refcount_recovery), KM_SLEEP); + xfs_refcount_btrec_to_irec(rec, &rr->rr_rrec); + list_add_tail(&rr->rr_list, debris); + + return 0; +} + +/* Find and remove leftover CoW reservations. */ +int +xfs_refcount_recover_cow_leftovers( + struct xfs_mount *mp, + xfs_agnumber_t agno) +{ + struct xfs_trans *tp; + struct xfs_btree_cur *cur; + struct xfs_buf *agbp; + struct xfs_refcount_recovery *rr, *n; + struct list_head debris; + union xfs_btree_irec low; + union xfs_btree_irec high; + struct xfs_defer_ops dfops; + xfs_fsblock_t fsb; + xfs_agblock_t agbno; + int error; + + if (mp->m_sb.sb_agblocks >= XFS_REFC_COW_START) + return -EOPNOTSUPP; + + error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp); + if (error) + return error; + cur = xfs_refcountbt_init_cursor(mp, NULL, agbp, agno, NULL); + + /* Find all the leftover CoW staging extents. */ + INIT_LIST_HEAD(&debris); + memset(&low, 0, sizeof(low)); + memset(&high, 0, sizeof(high)); + low.rc.rc_startblock = XFS_REFC_COW_START; + high.rc.rc_startblock = -1U; + error = xfs_btree_query_range(cur, &low, &high, + xfs_refcount_recover_extent, &debris); + if (error) + goto out_cursor; + xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); + xfs_buf_relse(agbp); + + /* Now iterate the list to free the leftovers */ + list_for_each_entry(rr, &debris, rr_list) { + /* Set up transaction. */ + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0, &tp); + if (error) + goto out_free; + + trace_xfs_refcount_recover_extent(mp, agno, &rr->rr_rrec); + + /* Free the orphan record */ + xfs_defer_init(&dfops, &fsb); + agbno = rr->rr_rrec.rc_startblock - XFS_REFC_COW_START; + fsb = XFS_AGB_TO_FSB(mp, agno, agbno); + error = xfs_refcount_free_cow_extent(mp, &dfops, fsb, + rr->rr_rrec.rc_blockcount); + if (error) + goto out_defer; + + /* Free the block. */ + xfs_bmap_add_free(mp, &dfops, fsb, + rr->rr_rrec.rc_blockcount, NULL); + + error = xfs_defer_finish(&tp, &dfops, NULL); + if (error) + goto out_defer; + + error = xfs_trans_commit(tp); + if (error) + goto out_free; + } + +out_free: + /* Free the leftover list */ + list_for_each_entry_safe(rr, n, &debris, rr_list) { + list_del(&rr->rr_list); + kmem_free(rr); + } + return error; + +out_cursor: + xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); + xfs_buf_relse(agbp); + goto out_free; + +out_defer: + xfs_defer_cancel(&dfops); + xfs_trans_cancel(tp); + goto out_free; +} diff --git a/fs/xfs/libxfs/xfs_refcount.h b/fs/xfs/libxfs/xfs_refcount.h new file mode 100644 index 000000000000..098dc668ab2c --- /dev/null +++ b/fs/xfs/libxfs/xfs_refcount.h @@ -0,0 +1,70 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_REFCOUNT_H__ +#define __XFS_REFCOUNT_H__ + +extern int xfs_refcount_lookup_le(struct xfs_btree_cur *cur, + xfs_agblock_t bno, int *stat); +extern int xfs_refcount_lookup_ge(struct xfs_btree_cur *cur, + xfs_agblock_t bno, int *stat); +extern int xfs_refcount_get_rec(struct xfs_btree_cur *cur, + struct xfs_refcount_irec *irec, int *stat); + +enum xfs_refcount_intent_type { + XFS_REFCOUNT_INCREASE = 1, + XFS_REFCOUNT_DECREASE, + XFS_REFCOUNT_ALLOC_COW, + XFS_REFCOUNT_FREE_COW, +}; + +struct xfs_refcount_intent { + struct list_head ri_list; + enum xfs_refcount_intent_type ri_type; + xfs_fsblock_t ri_startblock; + xfs_extlen_t ri_blockcount; +}; + +extern int xfs_refcount_increase_extent(struct xfs_mount *mp, + struct xfs_defer_ops *dfops, struct xfs_bmbt_irec *irec); +extern int xfs_refcount_decrease_extent(struct xfs_mount *mp, + struct xfs_defer_ops *dfops, struct xfs_bmbt_irec *irec); + +extern void xfs_refcount_finish_one_cleanup(struct xfs_trans *tp, + struct xfs_btree_cur *rcur, int error); +extern int xfs_refcount_finish_one(struct xfs_trans *tp, + struct xfs_defer_ops *dfops, enum xfs_refcount_intent_type type, + xfs_fsblock_t startblock, xfs_extlen_t blockcount, + xfs_fsblock_t *new_fsb, xfs_extlen_t *new_len, + struct xfs_btree_cur **pcur); + +extern int xfs_refcount_find_shared(struct xfs_btree_cur *cur, + xfs_agblock_t agbno, xfs_extlen_t aglen, xfs_agblock_t *fbno, + xfs_extlen_t *flen, bool find_end_of_shared); + +extern int xfs_refcount_alloc_cow_extent(struct xfs_mount *mp, + struct xfs_defer_ops *dfops, xfs_fsblock_t fsb, + xfs_extlen_t len); +extern int xfs_refcount_free_cow_extent(struct xfs_mount *mp, + struct xfs_defer_ops *dfops, xfs_fsblock_t fsb, + xfs_extlen_t len); +extern int xfs_refcount_recover_cow_leftovers(struct xfs_mount *mp, + xfs_agnumber_t agno); + +#endif /* __XFS_REFCOUNT_H__ */ diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c new file mode 100644 index 000000000000..50add5272807 --- /dev/null +++ b/fs/xfs/libxfs/xfs_refcount_btree.c @@ -0,0 +1,455 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_sb.h" +#include "xfs_mount.h" +#include "xfs_btree.h" +#include "xfs_bmap.h" +#include "xfs_refcount_btree.h" +#include "xfs_alloc.h" +#include "xfs_error.h" +#include "xfs_trace.h" +#include "xfs_cksum.h" +#include "xfs_trans.h" +#include "xfs_bit.h" +#include "xfs_rmap.h" + +static struct xfs_btree_cur * +xfs_refcountbt_dup_cursor( + struct xfs_btree_cur *cur) +{ + return xfs_refcountbt_init_cursor(cur->bc_mp, cur->bc_tp, + cur->bc_private.a.agbp, cur->bc_private.a.agno, + cur->bc_private.a.dfops); +} + +STATIC void +xfs_refcountbt_set_root( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr, + int inc) +{ + struct xfs_buf *agbp = cur->bc_private.a.agbp; + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + xfs_agnumber_t seqno = be32_to_cpu(agf->agf_seqno); + struct xfs_perag *pag = xfs_perag_get(cur->bc_mp, seqno); + + ASSERT(ptr->s != 0); + + agf->agf_refcount_root = ptr->s; + be32_add_cpu(&agf->agf_refcount_level, inc); + pag->pagf_refcount_level += inc; + xfs_perag_put(pag); + + xfs_alloc_log_agf(cur->bc_tp, agbp, + XFS_AGF_REFCOUNT_ROOT | XFS_AGF_REFCOUNT_LEVEL); +} + +STATIC int +xfs_refcountbt_alloc_block( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *start, + union xfs_btree_ptr *new, + int *stat) +{ + struct xfs_buf *agbp = cur->bc_private.a.agbp; + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + struct xfs_alloc_arg args; /* block allocation args */ + int error; /* error return value */ + + XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY); + + memset(&args, 0, sizeof(args)); + args.tp = cur->bc_tp; + args.mp = cur->bc_mp; + args.type = XFS_ALLOCTYPE_NEAR_BNO; + args.fsbno = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno, + xfs_refc_block(args.mp)); + args.firstblock = args.fsbno; + xfs_rmap_ag_owner(&args.oinfo, XFS_RMAP_OWN_REFC); + args.minlen = args.maxlen = args.prod = 1; + args.resv = XFS_AG_RESV_METADATA; + + error = xfs_alloc_vextent(&args); + if (error) + goto out_error; + trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_private.a.agno, + args.agbno, 1); + if (args.fsbno == NULLFSBLOCK) { + XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); + *stat = 0; + return 0; + } + ASSERT(args.agno == cur->bc_private.a.agno); + ASSERT(args.len == 1); + + new->s = cpu_to_be32(args.agbno); + be32_add_cpu(&agf->agf_refcount_blocks, 1); + xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS); + + XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT); + *stat = 1; + return 0; + +out_error: + XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR); + return error; +} + +STATIC int +xfs_refcountbt_free_block( + struct xfs_btree_cur *cur, + struct xfs_buf *bp) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_buf *agbp = cur->bc_private.a.agbp; + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + xfs_fsblock_t fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp)); + struct xfs_owner_info oinfo; + int error; + + trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_private.a.agno, + XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1); + xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_REFC); + be32_add_cpu(&agf->agf_refcount_blocks, -1); + xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS); + error = xfs_free_extent(cur->bc_tp, fsbno, 1, &oinfo, + XFS_AG_RESV_METADATA); + if (error) + return error; + + return error; +} + +STATIC int +xfs_refcountbt_get_minrecs( + struct xfs_btree_cur *cur, + int level) +{ + return cur->bc_mp->m_refc_mnr[level != 0]; +} + +STATIC int +xfs_refcountbt_get_maxrecs( + struct xfs_btree_cur *cur, + int level) +{ + return cur->bc_mp->m_refc_mxr[level != 0]; +} + +STATIC void +xfs_refcountbt_init_key_from_rec( + union xfs_btree_key *key, + union xfs_btree_rec *rec) +{ + key->refc.rc_startblock = rec->refc.rc_startblock; +} + +STATIC void +xfs_refcountbt_init_high_key_from_rec( + union xfs_btree_key *key, + union xfs_btree_rec *rec) +{ + __u32 x; + + x = be32_to_cpu(rec->refc.rc_startblock); + x += be32_to_cpu(rec->refc.rc_blockcount) - 1; + key->refc.rc_startblock = cpu_to_be32(x); +} + +STATIC void +xfs_refcountbt_init_rec_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_rec *rec) +{ + rec->refc.rc_startblock = cpu_to_be32(cur->bc_rec.rc.rc_startblock); + rec->refc.rc_blockcount = cpu_to_be32(cur->bc_rec.rc.rc_blockcount); + rec->refc.rc_refcount = cpu_to_be32(cur->bc_rec.rc.rc_refcount); +} + +STATIC void +xfs_refcountbt_init_ptr_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr) +{ + struct xfs_agf *agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp); + + ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno)); + ASSERT(agf->agf_refcount_root != 0); + + ptr->s = agf->agf_refcount_root; +} + +STATIC __int64_t +xfs_refcountbt_key_diff( + struct xfs_btree_cur *cur, + union xfs_btree_key *key) +{ + struct xfs_refcount_irec *rec = &cur->bc_rec.rc; + struct xfs_refcount_key *kp = &key->refc; + + return (__int64_t)be32_to_cpu(kp->rc_startblock) - rec->rc_startblock; +} + +STATIC __int64_t +xfs_refcountbt_diff_two_keys( + struct xfs_btree_cur *cur, + union xfs_btree_key *k1, + union xfs_btree_key *k2) +{ + return (__int64_t)be32_to_cpu(k1->refc.rc_startblock) - + be32_to_cpu(k2->refc.rc_startblock); +} + +STATIC bool +xfs_refcountbt_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_target->bt_mount; + struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); + struct xfs_perag *pag = bp->b_pag; + unsigned int level; + + if (block->bb_magic != cpu_to_be32(XFS_REFC_CRC_MAGIC)) + return false; + + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return false; + if (!xfs_btree_sblock_v5hdr_verify(bp)) + return false; + + level = be16_to_cpu(block->bb_level); + if (pag && pag->pagf_init) { + if (level >= pag->pagf_refcount_level) + return false; + } else if (level >= mp->m_refc_maxlevels) + return false; + + return xfs_btree_sblock_verify(bp, mp->m_refc_mxr[level != 0]); +} + +STATIC void +xfs_refcountbt_read_verify( + struct xfs_buf *bp) +{ + if (!xfs_btree_sblock_verify_crc(bp)) + xfs_buf_ioerror(bp, -EFSBADCRC); + else if (!xfs_refcountbt_verify(bp)) + xfs_buf_ioerror(bp, -EFSCORRUPTED); + + if (bp->b_error) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_verifier_error(bp); + } +} + +STATIC void +xfs_refcountbt_write_verify( + struct xfs_buf *bp) +{ + if (!xfs_refcountbt_verify(bp)) { + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_buf_ioerror(bp, -EFSCORRUPTED); + xfs_verifier_error(bp); + return; + } + xfs_btree_sblock_calc_crc(bp); + +} + +const struct xfs_buf_ops xfs_refcountbt_buf_ops = { + .name = "xfs_refcountbt", + .verify_read = xfs_refcountbt_read_verify, + .verify_write = xfs_refcountbt_write_verify, +}; + +#if defined(DEBUG) || defined(XFS_WARN) +STATIC int +xfs_refcountbt_keys_inorder( + struct xfs_btree_cur *cur, + union xfs_btree_key *k1, + union xfs_btree_key *k2) +{ + return be32_to_cpu(k1->refc.rc_startblock) < + be32_to_cpu(k2->refc.rc_startblock); +} + +STATIC int +xfs_refcountbt_recs_inorder( + struct xfs_btree_cur *cur, + union xfs_btree_rec *r1, + union xfs_btree_rec *r2) +{ + return be32_to_cpu(r1->refc.rc_startblock) + + be32_to_cpu(r1->refc.rc_blockcount) <= + be32_to_cpu(r2->refc.rc_startblock); +} +#endif + +static const struct xfs_btree_ops xfs_refcountbt_ops = { + .rec_len = sizeof(struct xfs_refcount_rec), + .key_len = sizeof(struct xfs_refcount_key), + + .dup_cursor = xfs_refcountbt_dup_cursor, + .set_root = xfs_refcountbt_set_root, + .alloc_block = xfs_refcountbt_alloc_block, + .free_block = xfs_refcountbt_free_block, + .get_minrecs = xfs_refcountbt_get_minrecs, + .get_maxrecs = xfs_refcountbt_get_maxrecs, + .init_key_from_rec = xfs_refcountbt_init_key_from_rec, + .init_high_key_from_rec = xfs_refcountbt_init_high_key_from_rec, + .init_rec_from_cur = xfs_refcountbt_init_rec_from_cur, + .init_ptr_from_cur = xfs_refcountbt_init_ptr_from_cur, + .key_diff = xfs_refcountbt_key_diff, + .buf_ops = &xfs_refcountbt_buf_ops, + .diff_two_keys = xfs_refcountbt_diff_two_keys, +#if defined(DEBUG) || defined(XFS_WARN) + .keys_inorder = xfs_refcountbt_keys_inorder, + .recs_inorder = xfs_refcountbt_recs_inorder, +#endif +}; + +/* + * Allocate a new refcount btree cursor. + */ +struct xfs_btree_cur * +xfs_refcountbt_init_cursor( + struct xfs_mount *mp, + struct xfs_trans *tp, + struct xfs_buf *agbp, + xfs_agnumber_t agno, + struct xfs_defer_ops *dfops) +{ + struct xfs_agf *agf = XFS_BUF_TO_AGF(agbp); + struct xfs_btree_cur *cur; + + ASSERT(agno != NULLAGNUMBER); + ASSERT(agno < mp->m_sb.sb_agcount); + cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); + + cur->bc_tp = tp; + cur->bc_mp = mp; + cur->bc_btnum = XFS_BTNUM_REFC; + cur->bc_blocklog = mp->m_sb.sb_blocklog; + cur->bc_ops = &xfs_refcountbt_ops; + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2); + + cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level); + + cur->bc_private.a.agbp = agbp; + cur->bc_private.a.agno = agno; + cur->bc_private.a.dfops = dfops; + cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; + + cur->bc_private.a.priv.refc.nr_ops = 0; + cur->bc_private.a.priv.refc.shape_changes = 0; + + return cur; +} + +/* + * Calculate the number of records in a refcount btree block. + */ +int +xfs_refcountbt_maxrecs( + struct xfs_mount *mp, + int blocklen, + bool leaf) +{ + blocklen -= XFS_REFCOUNT_BLOCK_LEN; + + if (leaf) + return blocklen / sizeof(struct xfs_refcount_rec); + return blocklen / (sizeof(struct xfs_refcount_key) + + sizeof(xfs_refcount_ptr_t)); +} + +/* Compute the maximum height of a refcount btree. */ +void +xfs_refcountbt_compute_maxlevels( + struct xfs_mount *mp) +{ + mp->m_refc_maxlevels = xfs_btree_compute_maxlevels(mp, + mp->m_refc_mnr, mp->m_sb.sb_agblocks); +} + +/* Calculate the refcount btree size for some records. */ +xfs_extlen_t +xfs_refcountbt_calc_size( + struct xfs_mount *mp, + unsigned long long len) +{ + return xfs_btree_calc_size(mp, mp->m_refc_mnr, len); +} + +/* + * Calculate the maximum refcount btree size. + */ +xfs_extlen_t +xfs_refcountbt_max_size( + struct xfs_mount *mp, + xfs_agblock_t agblocks) +{ + /* Bail out if we're uninitialized, which can happen in mkfs. */ + if (mp->m_refc_mxr[0] == 0) + return 0; + + return xfs_refcountbt_calc_size(mp, agblocks); +} + +/* + * Figure out how many blocks to reserve and how many are used by this btree. + */ +int +xfs_refcountbt_calc_reserves( + struct xfs_mount *mp, + xfs_agnumber_t agno, + xfs_extlen_t *ask, + xfs_extlen_t *used) +{ + struct xfs_buf *agbp; + struct xfs_agf *agf; + xfs_agblock_t agblocks; + xfs_extlen_t tree_len; + int error; + + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + + error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp); + if (error) + return error; + + agf = XFS_BUF_TO_AGF(agbp); + agblocks = be32_to_cpu(agf->agf_length); + tree_len = be32_to_cpu(agf->agf_refcount_blocks); + xfs_buf_relse(agbp); + + *ask += xfs_refcountbt_max_size(mp, agblocks); + *used += tree_len; + + return error; +} diff --git a/fs/xfs/libxfs/xfs_refcount_btree.h b/fs/xfs/libxfs/xfs_refcount_btree.h new file mode 100644 index 000000000000..9db008b955b7 --- /dev/null +++ b/fs/xfs/libxfs/xfs_refcount_btree.h @@ -0,0 +1,75 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_REFCOUNT_BTREE_H__ +#define __XFS_REFCOUNT_BTREE_H__ + +/* + * Reference Count Btree on-disk structures + */ + +struct xfs_buf; +struct xfs_btree_cur; +struct xfs_mount; + +/* + * Btree block header size + */ +#define XFS_REFCOUNT_BLOCK_LEN XFS_BTREE_SBLOCK_CRC_LEN + +/* + * Record, key, and pointer address macros for btree blocks. + * + * (note that some of these may appear unused, but they are used in userspace) + */ +#define XFS_REFCOUNT_REC_ADDR(block, index) \ + ((struct xfs_refcount_rec *) \ + ((char *)(block) + \ + XFS_REFCOUNT_BLOCK_LEN + \ + (((index) - 1) * sizeof(struct xfs_refcount_rec)))) + +#define XFS_REFCOUNT_KEY_ADDR(block, index) \ + ((struct xfs_refcount_key *) \ + ((char *)(block) + \ + XFS_REFCOUNT_BLOCK_LEN + \ + ((index) - 1) * sizeof(struct xfs_refcount_key))) + +#define XFS_REFCOUNT_PTR_ADDR(block, index, maxrecs) \ + ((xfs_refcount_ptr_t *) \ + ((char *)(block) + \ + XFS_REFCOUNT_BLOCK_LEN + \ + (maxrecs) * sizeof(struct xfs_refcount_key) + \ + ((index) - 1) * sizeof(xfs_refcount_ptr_t))) + +extern struct xfs_btree_cur *xfs_refcountbt_init_cursor(struct xfs_mount *mp, + struct xfs_trans *tp, struct xfs_buf *agbp, xfs_agnumber_t agno, + struct xfs_defer_ops *dfops); +extern int xfs_refcountbt_maxrecs(struct xfs_mount *mp, int blocklen, + bool leaf); +extern void xfs_refcountbt_compute_maxlevels(struct xfs_mount *mp); + +extern xfs_extlen_t xfs_refcountbt_calc_size(struct xfs_mount *mp, + unsigned long long len); +extern xfs_extlen_t xfs_refcountbt_max_size(struct xfs_mount *mp, + xfs_agblock_t agblocks); + +extern int xfs_refcountbt_calc_reserves(struct xfs_mount *mp, + xfs_agnumber_t agno, xfs_extlen_t *ask, xfs_extlen_t *used); + +#endif /* __XFS_REFCOUNT_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 73d05407d663..3a8cc7139912 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -148,6 +148,37 @@ done: return error; } +STATIC int +xfs_rmap_delete( + struct xfs_btree_cur *rcur, + xfs_agblock_t agbno, + xfs_extlen_t len, + uint64_t owner, + uint64_t offset, + unsigned int flags) +{ + int i; + int error; + + trace_xfs_rmap_delete(rcur->bc_mp, rcur->bc_private.a.agno, agbno, + len, owner, offset, flags); + + error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(rcur->bc_mp, i == 1, done); + + error = xfs_btree_delete(rcur, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(rcur->bc_mp, i == 1, done); +done: + if (error) + trace_xfs_rmap_delete_error(rcur->bc_mp, + rcur->bc_private.a.agno, error, _RET_IP_); + return error; +} + static int xfs_rmap_btrec_to_irec( union xfs_btree_rec *rec, @@ -180,6 +211,160 @@ xfs_rmap_get_rec( return xfs_rmap_btrec_to_irec(rec, irec); } +struct xfs_find_left_neighbor_info { + struct xfs_rmap_irec high; + struct xfs_rmap_irec *irec; + int *stat; +}; + +/* For each rmap given, figure out if it matches the key we want. */ +STATIC int +xfs_rmap_find_left_neighbor_helper( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *rec, + void *priv) +{ + struct xfs_find_left_neighbor_info *info = priv; + + trace_xfs_rmap_find_left_neighbor_candidate(cur->bc_mp, + cur->bc_private.a.agno, rec->rm_startblock, + rec->rm_blockcount, rec->rm_owner, rec->rm_offset, + rec->rm_flags); + + if (rec->rm_owner != info->high.rm_owner) + return XFS_BTREE_QUERY_RANGE_CONTINUE; + if (!XFS_RMAP_NON_INODE_OWNER(rec->rm_owner) && + !(rec->rm_flags & XFS_RMAP_BMBT_BLOCK) && + rec->rm_offset + rec->rm_blockcount - 1 != info->high.rm_offset) + return XFS_BTREE_QUERY_RANGE_CONTINUE; + + *info->irec = *rec; + *info->stat = 1; + return XFS_BTREE_QUERY_RANGE_ABORT; +} + +/* + * Find the record to the left of the given extent, being careful only to + * return a match with the same owner and adjacent physical and logical + * block ranges. + */ +int +xfs_rmap_find_left_neighbor( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + uint64_t owner, + uint64_t offset, + unsigned int flags, + struct xfs_rmap_irec *irec, + int *stat) +{ + struct xfs_find_left_neighbor_info info; + int error; + + *stat = 0; + if (bno == 0) + return 0; + info.high.rm_startblock = bno - 1; + info.high.rm_owner = owner; + if (!XFS_RMAP_NON_INODE_OWNER(owner) && + !(flags & XFS_RMAP_BMBT_BLOCK)) { + if (offset == 0) + return 0; + info.high.rm_offset = offset - 1; + } else + info.high.rm_offset = 0; + info.high.rm_flags = flags; + info.high.rm_blockcount = 0; + info.irec = irec; + info.stat = stat; + + trace_xfs_rmap_find_left_neighbor_query(cur->bc_mp, + cur->bc_private.a.agno, bno, 0, owner, offset, flags); + + error = xfs_rmap_query_range(cur, &info.high, &info.high, + xfs_rmap_find_left_neighbor_helper, &info); + if (error == XFS_BTREE_QUERY_RANGE_ABORT) + error = 0; + if (*stat) + trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, irec->rm_startblock, + irec->rm_blockcount, irec->rm_owner, + irec->rm_offset, irec->rm_flags); + return error; +} + +/* For each rmap given, figure out if it matches the key we want. */ +STATIC int +xfs_rmap_lookup_le_range_helper( + struct xfs_btree_cur *cur, + struct xfs_rmap_irec *rec, + void *priv) +{ + struct xfs_find_left_neighbor_info *info = priv; + + trace_xfs_rmap_lookup_le_range_candidate(cur->bc_mp, + cur->bc_private.a.agno, rec->rm_startblock, + rec->rm_blockcount, rec->rm_owner, rec->rm_offset, + rec->rm_flags); + + if (rec->rm_owner != info->high.rm_owner) + return XFS_BTREE_QUERY_RANGE_CONTINUE; + if (!XFS_RMAP_NON_INODE_OWNER(rec->rm_owner) && + !(rec->rm_flags & XFS_RMAP_BMBT_BLOCK) && + (rec->rm_offset > info->high.rm_offset || + rec->rm_offset + rec->rm_blockcount <= info->high.rm_offset)) + return XFS_BTREE_QUERY_RANGE_CONTINUE; + + *info->irec = *rec; + *info->stat = 1; + return XFS_BTREE_QUERY_RANGE_ABORT; +} + +/* + * Find the record to the left of the given extent, being careful only to + * return a match with the same owner and overlapping physical and logical + * block ranges. This is the overlapping-interval version of + * xfs_rmap_lookup_le. + */ +int +xfs_rmap_lookup_le_range( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + uint64_t owner, + uint64_t offset, + unsigned int flags, + struct xfs_rmap_irec *irec, + int *stat) +{ + struct xfs_find_left_neighbor_info info; + int error; + + info.high.rm_startblock = bno; + info.high.rm_owner = owner; + if (!XFS_RMAP_NON_INODE_OWNER(owner) && !(flags & XFS_RMAP_BMBT_BLOCK)) + info.high.rm_offset = offset; + else + info.high.rm_offset = 0; + info.high.rm_flags = flags; + info.high.rm_blockcount = 0; + *stat = 0; + info.irec = irec; + info.stat = stat; + + trace_xfs_rmap_lookup_le_range(cur->bc_mp, + cur->bc_private.a.agno, bno, 0, owner, offset, flags); + error = xfs_rmap_query_range(cur, &info.high, &info.high, + xfs_rmap_lookup_le_range_helper, &info); + if (error == XFS_BTREE_QUERY_RANGE_ABORT) + error = 0; + if (*stat) + trace_xfs_rmap_lookup_le_range_result(cur->bc_mp, + cur->bc_private.a.agno, irec->rm_startblock, + irec->rm_blockcount, irec->rm_owner, + irec->rm_offset, irec->rm_flags); + return error; +} + /* * Find the extent in the rmap btree and remove it. * @@ -1093,11 +1278,704 @@ done: return error; } +/* + * Convert an unwritten extent to a real extent or vice versa. If there is no + * possibility of overlapping extents, delegate to the simpler convert + * function. + */ +STATIC int +xfs_rmap_convert_shared( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec r[4]; /* neighbor extent entries */ + /* left is 0, right is 1, prev is 2 */ + /* new is 3 */ + uint64_t owner; + uint64_t offset; + uint64_t new_endoff; + unsigned int oldext; + unsigned int newext; + unsigned int flags = 0; + int i; + int state = 0; + int error; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + ASSERT(!(XFS_RMAP_NON_INODE_OWNER(owner) || + (flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK)))); + oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0; + new_endoff = offset + len; + trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* + * For the initial lookup, look for and exact match or the left-adjacent + * record for our insertion point. This will also give us the record for + * start block contiguity tests. + */ + error = xfs_rmap_lookup_le_range(cur, bno, owner, offset, flags, + &PREV, &i); + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + + ASSERT(PREV.rm_offset <= offset); + ASSERT(PREV.rm_offset + PREV.rm_blockcount >= new_endoff); + ASSERT((PREV.rm_flags & XFS_RMAP_UNWRITTEN) == oldext); + newext = ~oldext & XFS_RMAP_UNWRITTEN; + + /* + * Set flags determining what part of the previous oldext allocation + * extent is being replaced by a newext allocation. + */ + if (PREV.rm_offset == offset) + state |= RMAP_LEFT_FILLING; + if (PREV.rm_offset + PREV.rm_blockcount == new_endoff) + state |= RMAP_RIGHT_FILLING; + + /* Is there a left record that abuts our range? */ + error = xfs_rmap_find_left_neighbor(cur, bno, owner, offset, newext, + &LEFT, &i); + if (error) + goto done; + if (i) { + state |= RMAP_LEFT_VALID; + XFS_WANT_CORRUPTED_GOTO(mp, + LEFT.rm_startblock + LEFT.rm_blockcount <= bno, + done); + if (xfs_rmap_is_mergeable(&LEFT, owner, newext)) + state |= RMAP_LEFT_CONTIG; + } + + /* Is there a right record that abuts our range? */ + error = xfs_rmap_lookup_eq(cur, bno + len, len, owner, offset + len, + newext, &i); + if (error) + goto done; + if (i) { + state |= RMAP_RIGHT_VALID; + error = xfs_rmap_get_rec(cur, &RIGHT, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= RIGHT.rm_startblock, + done); + trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, RIGHT.rm_startblock, + RIGHT.rm_blockcount, RIGHT.rm_owner, + RIGHT.rm_offset, RIGHT.rm_flags); + if (xfs_rmap_is_mergeable(&RIGHT, owner, newext)) + state |= RMAP_RIGHT_CONTIG; + } + + /* check that left + prev + right is not too long */ + if ((state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) == + (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG) && + (unsigned long)LEFT.rm_blockcount + len + + RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX) + state &= ~RMAP_RIGHT_CONTIG; + + trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state, + _RET_IP_); + /* + * Switch out based on the FILLING and CONTIG state bits. + */ + switch (state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) { + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | + RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The left and right neighbors are both contiguous with new. + */ + error = xfs_rmap_delete(cur, RIGHT.rm_startblock, + RIGHT.rm_blockcount, RIGHT.rm_owner, + RIGHT.rm_offset, RIGHT.rm_flags); + if (error) + goto done; + error = xfs_rmap_delete(cur, PREV.rm_startblock, + PREV.rm_blockcount, PREV.rm_owner, + PREV.rm_offset, PREV.rm_flags); + if (error) + goto done; + NEW = LEFT; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount += PREV.rm_blockcount + RIGHT.rm_blockcount; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The left neighbor is contiguous, the right is not. + */ + error = xfs_rmap_delete(cur, PREV.rm_startblock, + PREV.rm_blockcount, PREV.rm_owner, + PREV.rm_offset, PREV.rm_flags); + if (error) + goto done; + NEW = LEFT; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount += PREV.rm_blockcount; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting all of a previous oldext extent to newext. + * The right neighbor is contiguous, the left is not. + */ + error = xfs_rmap_delete(cur, RIGHT.rm_startblock, + RIGHT.rm_blockcount, RIGHT.rm_owner, + RIGHT.rm_offset, RIGHT.rm_flags); + if (error) + goto done; + NEW = PREV; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount += RIGHT.rm_blockcount; + NEW.rm_flags = RIGHT.rm_flags; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING: + /* + * Setting all of a previous oldext extent to newext. + * Neither the left nor right neighbors are contiguous with + * the new one. + */ + NEW = PREV; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_flags = newext; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG: + /* + * Setting the first part of a previous oldext extent to newext. + * The left neighbor is contiguous. + */ + NEW = PREV; + error = xfs_rmap_delete(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + NEW.rm_offset += len; + NEW.rm_startblock += len; + NEW.rm_blockcount -= len; + error = xfs_rmap_insert(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + NEW = LEFT; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount += len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING: + /* + * Setting the first part of a previous oldext extent to newext. + * The left neighbor is not contiguous. + */ + NEW = PREV; + error = xfs_rmap_delete(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + NEW.rm_offset += len; + NEW.rm_startblock += len; + NEW.rm_blockcount -= len; + error = xfs_rmap_insert(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + error = xfs_rmap_insert(cur, bno, len, owner, offset, newext); + if (error) + goto done; + break; + + case RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG: + /* + * Setting the last part of a previous oldext extent to newext. + * The right neighbor is contiguous with the new allocation. + */ + NEW = PREV; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount = offset - NEW.rm_offset; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + NEW = RIGHT; + error = xfs_rmap_delete(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + NEW.rm_offset = offset; + NEW.rm_startblock = bno; + NEW.rm_blockcount += len; + error = xfs_rmap_insert(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags); + if (error) + goto done; + break; + + case RMAP_RIGHT_FILLING: + /* + * Setting the last part of a previous oldext extent to newext. + * The right neighbor is not contiguous. + */ + NEW = PREV; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount -= len; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + error = xfs_rmap_insert(cur, bno, len, owner, offset, newext); + if (error) + goto done; + break; + + case 0: + /* + * Setting the middle part of a previous oldext extent to + * newext. Contiguity is impossible here. + * One extent becomes three extents. + */ + /* new right extent - oldext */ + NEW.rm_startblock = bno + len; + NEW.rm_owner = owner; + NEW.rm_offset = new_endoff; + NEW.rm_blockcount = PREV.rm_offset + PREV.rm_blockcount - + new_endoff; + NEW.rm_flags = PREV.rm_flags; + error = xfs_rmap_insert(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, NEW.rm_offset, + NEW.rm_flags); + if (error) + goto done; + /* new left extent - oldext */ + NEW = PREV; + error = xfs_rmap_lookup_eq(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, + NEW.rm_offset, NEW.rm_flags, &i); + if (error) + goto done; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done); + NEW.rm_blockcount = offset - NEW.rm_offset; + error = xfs_rmap_update(cur, &NEW); + if (error) + goto done; + /* new middle extent - newext */ + NEW.rm_startblock = bno; + NEW.rm_blockcount = len; + NEW.rm_owner = owner; + NEW.rm_offset = offset; + NEW.rm_flags = newext; + error = xfs_rmap_insert(cur, NEW.rm_startblock, + NEW.rm_blockcount, NEW.rm_owner, NEW.rm_offset, + NEW.rm_flags); + if (error) + goto done; + break; + + case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_LEFT_FILLING | RMAP_RIGHT_CONTIG: + case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG: + case RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG: + case RMAP_LEFT_CONTIG: + case RMAP_RIGHT_CONTIG: + /* + * These cases are all impossible. + */ + ASSERT(0); + } + + trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +done: + if (error) + trace_xfs_rmap_convert_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + #undef NEW #undef LEFT #undef RIGHT #undef PREV +/* + * Find an extent in the rmap btree and unmap it. For rmap extent types that + * can overlap (data fork rmaps on reflink filesystems) we must be careful + * that the prev/next records in the btree might belong to another owner. + * Therefore we must use delete+insert to alter any of the key fields. + * + * For every other situation there can only be one owner for a given extent, + * so we can call the regular _free function. + */ +STATIC int +xfs_rmap_unmap_shared( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec ltrec; + uint64_t ltoff; + int error = 0; + int i; + uint64_t owner; + uint64_t offset; + unsigned int flags; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + if (unwritten) + flags |= XFS_RMAP_UNWRITTEN; + trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* + * We should always have a left record because there's a static record + * for the AG headers at rm_startblock == 0 created by mkfs/growfs that + * will not ever be removed from the tree. + */ + error = xfs_rmap_lookup_le_range(cur, bno, owner, offset, flags, + <rec, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + ltoff = ltrec.rm_offset; + + /* Make sure the extent we found covers the entire freeing range. */ + XFS_WANT_CORRUPTED_GOTO(mp, ltrec.rm_startblock <= bno && + ltrec.rm_startblock + ltrec.rm_blockcount >= + bno + len, out_error); + + /* Make sure the owner matches what we expect to find in the tree. */ + XFS_WANT_CORRUPTED_GOTO(mp, owner == ltrec.rm_owner, out_error); + + /* Make sure the unwritten flag matches. */ + XFS_WANT_CORRUPTED_GOTO(mp, (flags & XFS_RMAP_UNWRITTEN) == + (ltrec.rm_flags & XFS_RMAP_UNWRITTEN), out_error); + + /* Check the offset. */ + XFS_WANT_CORRUPTED_GOTO(mp, ltrec.rm_offset <= offset, out_error); + XFS_WANT_CORRUPTED_GOTO(mp, offset <= ltoff + ltrec.rm_blockcount, + out_error); + + if (ltrec.rm_startblock == bno && ltrec.rm_blockcount == len) { + /* Exact match, simply remove the record from rmap tree. */ + error = xfs_rmap_delete(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags); + if (error) + goto out_error; + } else if (ltrec.rm_startblock == bno) { + /* + * Overlap left hand side of extent: move the start, trim the + * length and update the current record. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrrrrrrr| + * bno len + */ + + /* Delete prev rmap. */ + error = xfs_rmap_delete(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags); + if (error) + goto out_error; + + /* Add an rmap at the new offset. */ + ltrec.rm_startblock += len; + ltrec.rm_blockcount -= len; + ltrec.rm_offset += len; + error = xfs_rmap_insert(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags); + if (error) + goto out_error; + } else if (ltrec.rm_startblock + ltrec.rm_blockcount == bno + len) { + /* + * Overlap right hand side of extent: trim the length and + * update the current record. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrrrrrrr| + * bno len + */ + error = xfs_rmap_lookup_eq(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + ltrec.rm_blockcount -= len; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + } else { + /* + * Overlap middle of extent: trim the length of the existing + * record to the length of the new left-extent size, increment + * the insertion position so we can insert a new record + * containing the remaining right-extent space. + * + * ltbno ltlen + * Orig: |oooooooooooooooooooo| + * Freeing: |fffffffff| + * Result: |rrrrr| |rrrr| + * bno len + */ + xfs_extlen_t orig_len = ltrec.rm_blockcount; + + /* Shrink the left side of the rmap */ + error = xfs_rmap_lookup_eq(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + ltrec.rm_blockcount = bno - ltrec.rm_startblock; + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + + /* Add an rmap at the new offset */ + error = xfs_rmap_insert(cur, bno + len, + orig_len - len - ltrec.rm_blockcount, + ltrec.rm_owner, offset + len, + ltrec.rm_flags); + if (error) + goto out_error; + } + + trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +out_error: + if (error) + trace_xfs_rmap_unmap_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + +/* + * Find an extent in the rmap btree and map it. For rmap extent types that + * can overlap (data fork rmaps on reflink filesystems) we must be careful + * that the prev/next records in the btree might belong to another owner. + * Therefore we must use delete+insert to alter any of the key fields. + * + * For every other situation there can only be one owner for a given extent, + * so we can call the regular _alloc function. + */ +STATIC int +xfs_rmap_map_shared( + struct xfs_btree_cur *cur, + xfs_agblock_t bno, + xfs_extlen_t len, + bool unwritten, + struct xfs_owner_info *oinfo) +{ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_rmap_irec ltrec; + struct xfs_rmap_irec gtrec; + int have_gt; + int have_lt; + int error = 0; + int i; + uint64_t owner; + uint64_t offset; + unsigned int flags = 0; + + xfs_owner_info_unpack(oinfo, &owner, &offset, &flags); + if (unwritten) + flags |= XFS_RMAP_UNWRITTEN; + trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); + + /* Is there a left record that abuts our range? */ + error = xfs_rmap_find_left_neighbor(cur, bno, owner, offset, flags, + <rec, &have_lt); + if (error) + goto out_error; + if (have_lt && + !xfs_rmap_is_mergeable(<rec, owner, flags)) + have_lt = 0; + + /* Is there a right record that abuts our range? */ + error = xfs_rmap_lookup_eq(cur, bno + len, len, owner, offset + len, + flags, &have_gt); + if (error) + goto out_error; + if (have_gt) { + error = xfs_rmap_get_rec(cur, >rec, &have_gt); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, have_gt == 1, out_error); + trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp, + cur->bc_private.a.agno, gtrec.rm_startblock, + gtrec.rm_blockcount, gtrec.rm_owner, + gtrec.rm_offset, gtrec.rm_flags); + + if (!xfs_rmap_is_mergeable(>rec, owner, flags)) + have_gt = 0; + } + + if (have_lt && + ltrec.rm_startblock + ltrec.rm_blockcount == bno && + ltrec.rm_offset + ltrec.rm_blockcount == offset) { + /* + * Left edge contiguous, merge into left record. + * + * ltbno ltlen + * orig: |ooooooooo| + * adding: |aaaaaaaaa| + * result: |rrrrrrrrrrrrrrrrrrr| + * bno len + */ + ltrec.rm_blockcount += len; + if (have_gt && + bno + len == gtrec.rm_startblock && + offset + len == gtrec.rm_offset) { + /* + * Right edge also contiguous, delete right record + * and merge into left record. + * + * ltbno ltlen gtbno gtlen + * orig: |ooooooooo| |ooooooooo| + * adding: |aaaaaaaaa| + * result: |rrrrrrrrrrrrrrrrrrrrrrrrrrrrr| + */ + ltrec.rm_blockcount += gtrec.rm_blockcount; + error = xfs_rmap_delete(cur, gtrec.rm_startblock, + gtrec.rm_blockcount, gtrec.rm_owner, + gtrec.rm_offset, gtrec.rm_flags); + if (error) + goto out_error; + } + + /* Point the cursor back to the left record and update. */ + error = xfs_rmap_lookup_eq(cur, ltrec.rm_startblock, + ltrec.rm_blockcount, ltrec.rm_owner, + ltrec.rm_offset, ltrec.rm_flags, &i); + if (error) + goto out_error; + XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error); + + error = xfs_rmap_update(cur, <rec); + if (error) + goto out_error; + } else if (have_gt && + bno + len == gtrec.rm_startblock && + offset + len == gtrec.rm_offset) { + /* + * Right edge contiguous, merge into right record. + * + * gtbno gtlen + * Orig: |ooooooooo| + * adding: |aaaaaaaaa| + * Result: |rrrrrrrrrrrrrrrrrrr| + * bno len + */ + /* Delete the old record. */ + error = xfs_rmap_delete(cur, gtrec.rm_startblock, + gtrec.rm_blockcount, gtrec.rm_owner, + gtrec.rm_offset, gtrec.rm_flags); + if (error) + goto out_error; + + /* Move the start and re-add it. */ + gtrec.rm_startblock = bno; + gtrec.rm_blockcount += len; + gtrec.rm_offset = offset; + error = xfs_rmap_insert(cur, gtrec.rm_startblock, + gtrec.rm_blockcount, gtrec.rm_owner, + gtrec.rm_offset, gtrec.rm_flags); + if (error) + goto out_error; + } else { + /* + * No contiguous edge with identical owner, insert + * new record at current cursor position. + */ + error = xfs_rmap_insert(cur, bno, len, owner, offset, flags); + if (error) + goto out_error; + } + + trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len, + unwritten, oinfo); +out_error: + if (error) + trace_xfs_rmap_map_error(cur->bc_mp, + cur->bc_private.a.agno, error, _RET_IP_); + return error; +} + struct xfs_rmap_query_range_info { xfs_rmap_query_range_fn fn; void *priv; @@ -1237,15 +2115,27 @@ xfs_rmap_finish_one( case XFS_RMAP_MAP: error = xfs_rmap_map(rcur, bno, blockcount, unwritten, &oinfo); break; + case XFS_RMAP_MAP_SHARED: + error = xfs_rmap_map_shared(rcur, bno, blockcount, unwritten, + &oinfo); + break; case XFS_RMAP_FREE: case XFS_RMAP_UNMAP: error = xfs_rmap_unmap(rcur, bno, blockcount, unwritten, &oinfo); break; + case XFS_RMAP_UNMAP_SHARED: + error = xfs_rmap_unmap_shared(rcur, bno, blockcount, unwritten, + &oinfo); + break; case XFS_RMAP_CONVERT: error = xfs_rmap_convert(rcur, bno, blockcount, !unwritten, &oinfo); break; + case XFS_RMAP_CONVERT_SHARED: + error = xfs_rmap_convert_shared(rcur, bno, blockcount, + !unwritten, &oinfo); + break; default: ASSERT(0); error = -EFSCORRUPTED; @@ -1263,9 +2153,10 @@ out_cur: */ static bool xfs_rmap_update_is_needed( - struct xfs_mount *mp) + struct xfs_mount *mp, + int whichfork) { - return xfs_sb_version_hasrmapbt(&mp->m_sb); + return xfs_sb_version_hasrmapbt(&mp->m_sb) && whichfork != XFS_COW_FORK; } /* @@ -1311,10 +2202,11 @@ xfs_rmap_map_extent( int whichfork, struct xfs_bmbt_irec *PREV) { - if (!xfs_rmap_update_is_needed(mp)) + if (!xfs_rmap_update_is_needed(mp, whichfork)) return 0; - return __xfs_rmap_add(mp, dfops, XFS_RMAP_MAP, ip->i_ino, + return __xfs_rmap_add(mp, dfops, xfs_is_reflink_inode(ip) ? + XFS_RMAP_MAP_SHARED : XFS_RMAP_MAP, ip->i_ino, whichfork, PREV); } @@ -1327,10 +2219,11 @@ xfs_rmap_unmap_extent( int whichfork, struct xfs_bmbt_irec *PREV) { - if (!xfs_rmap_update_is_needed(mp)) + if (!xfs_rmap_update_is_needed(mp, whichfork)) return 0; - return __xfs_rmap_add(mp, dfops, XFS_RMAP_UNMAP, ip->i_ino, + return __xfs_rmap_add(mp, dfops, xfs_is_reflink_inode(ip) ? + XFS_RMAP_UNMAP_SHARED : XFS_RMAP_UNMAP, ip->i_ino, whichfork, PREV); } @@ -1343,10 +2236,11 @@ xfs_rmap_convert_extent( int whichfork, struct xfs_bmbt_irec *PREV) { - if (!xfs_rmap_update_is_needed(mp)) + if (!xfs_rmap_update_is_needed(mp, whichfork)) return 0; - return __xfs_rmap_add(mp, dfops, XFS_RMAP_CONVERT, ip->i_ino, + return __xfs_rmap_add(mp, dfops, xfs_is_reflink_inode(ip) ? + XFS_RMAP_CONVERT_SHARED : XFS_RMAP_CONVERT, ip->i_ino, whichfork, PREV); } @@ -1362,7 +2256,7 @@ xfs_rmap_alloc_extent( { struct xfs_bmbt_irec bmap; - if (!xfs_rmap_update_is_needed(mp)) + if (!xfs_rmap_update_is_needed(mp, XFS_DATA_FORK)) return 0; bmap.br_startblock = XFS_AGB_TO_FSB(mp, agno, bno); @@ -1386,7 +2280,7 @@ xfs_rmap_free_extent( { struct xfs_bmbt_irec bmap; - if (!xfs_rmap_update_is_needed(mp)) + if (!xfs_rmap_update_is_needed(mp, XFS_DATA_FORK)) return 0; bmap.br_startblock = XFS_AGB_TO_FSB(mp, agno, bno); diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index 71cf99a4acba..789930599339 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -206,4 +206,11 @@ int xfs_rmap_finish_one(struct xfs_trans *tp, enum xfs_rmap_intent_type type, xfs_fsblock_t startblock, xfs_filblks_t blockcount, xfs_exntst_t state, struct xfs_btree_cur **pcur); +int xfs_rmap_find_left_neighbor(struct xfs_btree_cur *cur, xfs_agblock_t bno, + uint64_t owner, uint64_t offset, unsigned int flags, + struct xfs_rmap_irec *irec, int *stat); +int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_agblock_t bno, + uint64_t owner, uint64_t offset, unsigned int flags, + struct xfs_rmap_irec *irec, int *stat); + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c index 17b8eeb34ac8..74e5a54bc428 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.c +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -35,6 +35,7 @@ #include "xfs_cksum.h" #include "xfs_error.h" #include "xfs_extent_busy.h" +#include "xfs_ag_resv.h" /* * Reverse map btree. @@ -483,6 +484,7 @@ xfs_rmapbt_init_cursor( cur->bc_blocklog = mp->m_sb.sb_blocklog; cur->bc_ops = &xfs_rmapbt_ops; cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]); + cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2); cur->bc_private.a.agbp = agbp; cur->bc_private.a.agno = agno; @@ -512,6 +514,83 @@ void xfs_rmapbt_compute_maxlevels( struct xfs_mount *mp) { - mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(mp, - mp->m_rmap_mnr, mp->m_sb.sb_agblocks); + /* + * On a non-reflink filesystem, the maximum number of rmap + * records is the number of blocks in the AG, hence the max + * rmapbt height is log_$maxrecs($agblocks). However, with + * reflink each AG block can have up to 2^32 (per the refcount + * record format) owners, which means that theoretically we + * could face up to 2^64 rmap records. + * + * That effectively means that the max rmapbt height must be + * XFS_BTREE_MAXLEVELS. "Fortunately" we'll run out of AG + * blocks to feed the rmapbt long before the rmapbt reaches + * maximum height. The reflink code uses ag_resv_critical to + * disallow reflinking when less than 10% of the per-AG metadata + * block reservation since the fallback is a regular file copy. + */ + if (xfs_sb_version_hasreflink(&mp->m_sb)) + mp->m_rmap_maxlevels = XFS_BTREE_MAXLEVELS; + else + mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(mp, + mp->m_rmap_mnr, mp->m_sb.sb_agblocks); +} + +/* Calculate the refcount btree size for some records. */ +xfs_extlen_t +xfs_rmapbt_calc_size( + struct xfs_mount *mp, + unsigned long long len) +{ + return xfs_btree_calc_size(mp, mp->m_rmap_mnr, len); +} + +/* + * Calculate the maximum refcount btree size. + */ +xfs_extlen_t +xfs_rmapbt_max_size( + struct xfs_mount *mp, + xfs_agblock_t agblocks) +{ + /* Bail out if we're uninitialized, which can happen in mkfs. */ + if (mp->m_rmap_mxr[0] == 0) + return 0; + + return xfs_rmapbt_calc_size(mp, agblocks); +} + +/* + * Figure out how many blocks to reserve and how many are used by this btree. + */ +int +xfs_rmapbt_calc_reserves( + struct xfs_mount *mp, + xfs_agnumber_t agno, + xfs_extlen_t *ask, + xfs_extlen_t *used) +{ + struct xfs_buf *agbp; + struct xfs_agf *agf; + xfs_agblock_t agblocks; + xfs_extlen_t tree_len; + int error; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return 0; + + error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp); + if (error) + return error; + + agf = XFS_BUF_TO_AGF(agbp); + agblocks = be32_to_cpu(agf->agf_length); + tree_len = be32_to_cpu(agf->agf_rmap_blocks); + xfs_buf_relse(agbp); + + /* Reserve 1% of the AG or enough for 1 block per record. */ + *ask += max(agblocks / 100, xfs_rmapbt_max_size(mp, agblocks)); + *used += tree_len; + + return error; } diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h index e73a55357dab..19c08e933049 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.h +++ b/fs/xfs/libxfs/xfs_rmap_btree.h @@ -58,4 +58,12 @@ struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp, int xfs_rmapbt_maxrecs(struct xfs_mount *mp, int blocklen, int leaf); extern void xfs_rmapbt_compute_maxlevels(struct xfs_mount *mp); +extern xfs_extlen_t xfs_rmapbt_calc_size(struct xfs_mount *mp, + unsigned long long len); +extern xfs_extlen_t xfs_rmapbt_max_size(struct xfs_mount *mp, + xfs_agblock_t agblocks); + +extern int xfs_rmapbt_calc_reserves(struct xfs_mount *mp, + xfs_agnumber_t agno, xfs_extlen_t *ask, xfs_extlen_t *used); + #endif /* __XFS_RMAP_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c index e2e1106c9fad..ea45584a9913 100644 --- a/fs/xfs/libxfs/xfs_rtbitmap.c +++ b/fs/xfs/libxfs/xfs_rtbitmap.c @@ -1016,4 +1016,3 @@ xfs_rtfree_extent( } return 0; } - diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 4aecc5fefe96..2580262e4ea0 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -38,6 +38,8 @@ #include "xfs_ialloc_btree.h" #include "xfs_log.h" #include "xfs_rmap_btree.h" +#include "xfs_bmap.h" +#include "xfs_refcount_btree.h" /* * Physical superblock buffer manipulations. Shared with libxfs in userspace. @@ -260,6 +262,12 @@ xfs_mount_validate_sb( return -EFSCORRUPTED; } + if (xfs_sb_version_hascrc(&mp->m_sb) && + sbp->sb_blocksize < XFS_MIN_CRC_BLOCKSIZE) { + xfs_notice(mp, "v5 SB sanity check failed"); + return -EFSCORRUPTED; + } + /* * Until this is fixed only page-sized or smaller data blocks work. */ @@ -336,13 +344,16 @@ xfs_sb_quota_from_disk(struct xfs_sb *sbp) XFS_PQUOTA_CHKD : XFS_GQUOTA_CHKD; sbp->sb_qflags &= ~(XFS_OQUOTA_ENFD | XFS_OQUOTA_CHKD); - if (sbp->sb_qflags & XFS_PQUOTA_ACCT) { + if (sbp->sb_qflags & XFS_PQUOTA_ACCT && + sbp->sb_gquotino != NULLFSINO) { /* * In older version of superblock, on-disk superblock only * has sb_gquotino, and in-core superblock has both sb_gquotino * and sb_pquotino. But, only one of them is supported at any * point of time. So, if PQUOTA is set in disk superblock, - * copy over sb_gquotino to sb_pquotino. + * copy over sb_gquotino to sb_pquotino. The NULLFSINO test + * above is to make sure we don't do this twice and wipe them + * both out! */ sbp->sb_pquotino = sbp->sb_gquotino; sbp->sb_gquotino = NULLFSINO; @@ -737,6 +748,13 @@ xfs_sb_mount_common( mp->m_rmap_mnr[0] = mp->m_rmap_mxr[0] / 2; mp->m_rmap_mnr[1] = mp->m_rmap_mxr[1] / 2; + mp->m_refc_mxr[0] = xfs_refcountbt_maxrecs(mp, sbp->sb_blocksize, + true); + mp->m_refc_mxr[1] = xfs_refcountbt_maxrecs(mp, sbp->sb_blocksize, + false); + mp->m_refc_mnr[0] = mp->m_refc_mxr[0] / 2; + mp->m_refc_mnr[1] = mp->m_refc_mxr[1] / 2; + mp->m_bsize = XFS_FSB_TO_BB(mp, 1); mp->m_ialloc_inos = (int)MAX((__uint16_t)XFS_INODES_PER_CHUNK, sbp->sb_inopblock); diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h index 0c5b30bd884c..c6f4eb46fe26 100644 --- a/fs/xfs/libxfs/xfs_shared.h +++ b/fs/xfs/libxfs/xfs_shared.h @@ -39,6 +39,7 @@ extern const struct xfs_buf_ops xfs_agf_buf_ops; extern const struct xfs_buf_ops xfs_agfl_buf_ops; extern const struct xfs_buf_ops xfs_allocbt_buf_ops; extern const struct xfs_buf_ops xfs_rmapbt_buf_ops; +extern const struct xfs_buf_ops xfs_refcountbt_buf_ops; extern const struct xfs_buf_ops xfs_attr3_leaf_buf_ops; extern const struct xfs_buf_ops xfs_attr3_rmt_buf_ops; extern const struct xfs_buf_ops xfs_bmbt_buf_ops; @@ -122,6 +123,7 @@ int xfs_log_calc_minimum_size(struct xfs_mount *); #define XFS_INO_REF 2 #define XFS_ATTR_BTREE_REF 1 #define XFS_DQUOT_REF 1 +#define XFS_REFC_BTREE_REF 1 /* * Flags for xfs_trans_ichgtime(). diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c index 301ef2f4dbd6..b456cca1bfb2 100644 --- a/fs/xfs/libxfs/xfs_trans_resv.c +++ b/fs/xfs/libxfs/xfs_trans_resv.c @@ -67,13 +67,14 @@ xfs_calc_buf_res( * Per-extent log reservation for the btree changes involved in freeing or * allocating an extent. In classic XFS there were two trees that will be * modified (bnobt + cntbt). With rmap enabled, there are three trees - * (rmapbt). The number of blocks reserved is based on the formula: + * (rmapbt). With reflink, there are four trees (refcountbt). The number of + * blocks reserved is based on the formula: * * num trees * ((2 blocks/level * max depth) - 1) * * Keep in mind that max depth is calculated separately for each type of tree. */ -static uint +uint xfs_allocfree_log_count( struct xfs_mount *mp, uint num_ops) @@ -83,6 +84,8 @@ xfs_allocfree_log_count( blocks = num_ops * 2 * (2 * mp->m_ag_maxlevels - 1); if (xfs_sb_version_hasrmapbt(&mp->m_sb)) blocks += num_ops * (2 * mp->m_rmap_maxlevels - 1); + if (xfs_sb_version_hasreflink(&mp->m_sb)) + blocks += num_ops * (2 * mp->m_refc_maxlevels - 1); return blocks; } @@ -809,11 +812,18 @@ xfs_trans_resv_calc( * require a permanent reservation on space. */ resp->tr_write.tr_logres = xfs_calc_write_reservation(mp); - resp->tr_write.tr_logcount = XFS_WRITE_LOG_COUNT; + if (xfs_sb_version_hasreflink(&mp->m_sb)) + resp->tr_write.tr_logcount = XFS_WRITE_LOG_COUNT_REFLINK; + else + resp->tr_write.tr_logcount = XFS_WRITE_LOG_COUNT; resp->tr_write.tr_logflags |= XFS_TRANS_PERM_LOG_RES; resp->tr_itruncate.tr_logres = xfs_calc_itruncate_reservation(mp); - resp->tr_itruncate.tr_logcount = XFS_ITRUNCATE_LOG_COUNT; + if (xfs_sb_version_hasreflink(&mp->m_sb)) + resp->tr_itruncate.tr_logcount = + XFS_ITRUNCATE_LOG_COUNT_REFLINK; + else + resp->tr_itruncate.tr_logcount = XFS_ITRUNCATE_LOG_COUNT; resp->tr_itruncate.tr_logflags |= XFS_TRANS_PERM_LOG_RES; resp->tr_rename.tr_logres = xfs_calc_rename_reservation(mp); @@ -870,7 +880,10 @@ xfs_trans_resv_calc( resp->tr_growrtalloc.tr_logflags |= XFS_TRANS_PERM_LOG_RES; resp->tr_qm_dqalloc.tr_logres = xfs_calc_qm_dqalloc_reservation(mp); - resp->tr_qm_dqalloc.tr_logcount = XFS_WRITE_LOG_COUNT; + if (xfs_sb_version_hasreflink(&mp->m_sb)) + resp->tr_qm_dqalloc.tr_logcount = XFS_WRITE_LOG_COUNT_REFLINK; + else + resp->tr_qm_dqalloc.tr_logcount = XFS_WRITE_LOG_COUNT; resp->tr_qm_dqalloc.tr_logflags |= XFS_TRANS_PERM_LOG_RES; /* diff --git a/fs/xfs/libxfs/xfs_trans_resv.h b/fs/xfs/libxfs/xfs_trans_resv.h index 0eb46ed6d404..b7e5357d060a 100644 --- a/fs/xfs/libxfs/xfs_trans_resv.h +++ b/fs/xfs/libxfs/xfs_trans_resv.h @@ -87,6 +87,7 @@ struct xfs_trans_resv { #define XFS_DEFAULT_LOG_COUNT 1 #define XFS_DEFAULT_PERM_LOG_COUNT 2 #define XFS_ITRUNCATE_LOG_COUNT 2 +#define XFS_ITRUNCATE_LOG_COUNT_REFLINK 8 #define XFS_INACTIVE_LOG_COUNT 2 #define XFS_CREATE_LOG_COUNT 2 #define XFS_CREATE_TMPFILE_LOG_COUNT 2 @@ -96,11 +97,13 @@ struct xfs_trans_resv { #define XFS_LINK_LOG_COUNT 2 #define XFS_RENAME_LOG_COUNT 2 #define XFS_WRITE_LOG_COUNT 2 +#define XFS_WRITE_LOG_COUNT_REFLINK 8 #define XFS_ADDAFORK_LOG_COUNT 2 #define XFS_ATTRINVAL_LOG_COUNT 1 #define XFS_ATTRSET_LOG_COUNT 3 #define XFS_ATTRRM_LOG_COUNT 3 void xfs_trans_resv_calc(struct xfs_mount *mp, struct xfs_trans_resv *resp); +uint xfs_allocfree_log_count(struct xfs_mount *mp, uint num_ops); #endif /* __XFS_TRANS_RESV_H__ */ diff --git a/fs/xfs/libxfs/xfs_trans_space.h b/fs/xfs/libxfs/xfs_trans_space.h index 41e0428d8175..7917f6e44286 100644 --- a/fs/xfs/libxfs/xfs_trans_space.h +++ b/fs/xfs/libxfs/xfs_trans_space.h @@ -21,6 +21,8 @@ /* * Components of space reservations. */ +#define XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp) \ + (((mp)->m_rmap_mxr[0]) - ((mp)->m_rmap_mnr[0])) #define XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp) \ (((mp)->m_alloc_mxr[0]) - ((mp)->m_alloc_mnr[0])) #define XFS_EXTENTADD_SPACE_RES(mp,w) (XFS_BM_MAXLEVELS(mp,w) - 1) @@ -28,6 +30,13 @@ (((b + XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp) - 1) / \ XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp)) * \ XFS_EXTENTADD_SPACE_RES(mp,w)) +#define XFS_SWAP_RMAP_SPACE_RES(mp,b,w)\ + (((b + XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp) - 1) / \ + XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp)) * \ + XFS_EXTENTADD_SPACE_RES(mp,w) + \ + ((b + XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp) - 1) / \ + XFS_MAX_CONTIG_RMAPS_PER_BLOCK(mp)) * \ + (mp)->m_rmap_maxlevels) #define XFS_DAENTER_1B(mp,w) \ ((w) == XFS_DATA_FORK ? (mp)->m_dir_geo->fsbcount : 1) #define XFS_DAENTER_DBS(mp,w) \ diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h index 3d503647f26b..717909f2f7b7 100644 --- a/fs/xfs/libxfs/xfs_types.h +++ b/fs/xfs/libxfs/xfs_types.h @@ -57,7 +57,6 @@ typedef __int64_t xfs_sfiloff_t; /* signed block number in a file */ #define NULLAGBLOCK ((xfs_agblock_t)-1) #define NULLAGNUMBER ((xfs_agnumber_t)-1) -#define NULLEXTNUM ((xfs_extnum_t)-1) #define NULLCOMMITLSN ((xfs_lsn_t)-1) @@ -75,11 +74,14 @@ typedef __int64_t xfs_sfiloff_t; /* signed block number in a file */ * Minimum and maximum blocksize and sectorsize. * The blocksize upper limit is pretty much arbitrary. * The sectorsize upper limit is due to sizeof(sb_sectsize). + * CRC enable filesystems use 512 byte inodes, meaning 512 byte block sizes + * cannot be used. */ #define XFS_MIN_BLOCKSIZE_LOG 9 /* i.e. 512 bytes */ #define XFS_MAX_BLOCKSIZE_LOG 16 /* i.e. 65536 bytes */ #define XFS_MIN_BLOCKSIZE (1 << XFS_MIN_BLOCKSIZE_LOG) #define XFS_MAX_BLOCKSIZE (1 << XFS_MAX_BLOCKSIZE_LOG) +#define XFS_MIN_CRC_BLOCKSIZE (1 << (XFS_MIN_BLOCKSIZE_LOG + 1)) #define XFS_MIN_SECTORSIZE_LOG 9 /* i.e. 512 bytes */ #define XFS_MAX_SECTORSIZE_LOG 15 /* i.e. 32768 bytes */ #define XFS_MIN_SECTORSIZE (1 << XFS_MIN_SECTORSIZE_LOG) @@ -90,6 +92,7 @@ typedef __int64_t xfs_sfiloff_t; /* signed block number in a file */ */ #define XFS_DATA_FORK 0 #define XFS_ATTR_FORK 1 +#define XFS_COW_FORK 2 /* * Min numbers of data/attr fork btree root pointers. @@ -109,7 +112,7 @@ typedef enum { typedef enum { XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi, - XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_MAX + XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_MAX } xfs_btnum_t; struct xfs_name { diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index b6e527b8eccb..b468e041f207 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -236,7 +236,7 @@ xfs_set_mode(struct inode *inode, umode_t mode) iattr.ia_valid = ATTR_MODE | ATTR_CTIME; iattr.ia_mode = mode; - iattr.ia_ctime = current_fs_time(inode->i_sb); + iattr.ia_ctime = current_time(inode); error = xfs_setattr_nonsize(XFS_I(inode), &iattr, XFS_ATTR_NOACL); } @@ -257,16 +257,11 @@ xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type) return error; if (type == ACL_TYPE_ACCESS) { - umode_t mode = inode->i_mode; - error = posix_acl_equiv_mode(acl, &mode); - - if (error <= 0) { - acl = NULL; - - if (error < 0) - return error; - } + umode_t mode; + error = posix_acl_update_mode(inode, &mode, &acl); + if (error) + return error; error = xfs_set_mode(inode, mode); if (error) return error; diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c index 7575cfc3ad15..0f56fcd3a5d5 100644 --- a/fs/xfs/xfs_aops.c +++ b/fs/xfs/xfs_aops.c @@ -31,15 +31,12 @@ #include "xfs_bmap.h" #include "xfs_bmap_util.h" #include "xfs_bmap_btree.h" +#include "xfs_reflink.h" #include <linux/gfp.h> #include <linux/mpage.h> #include <linux/pagevec.h> #include <linux/writeback.h> -/* flags for direct write completions */ -#define XFS_DIO_FLAG_UNWRITTEN (1 << 0) -#define XFS_DIO_FLAG_APPEND (1 << 1) - /* * structure owned by writepages passed to individual writepage calls */ @@ -200,7 +197,7 @@ xfs_setfilesize_trans_alloc( * Update on-disk file size now that data has been written to disk. */ STATIC int -xfs_setfilesize( +__xfs_setfilesize( struct xfs_inode *ip, struct xfs_trans *tp, xfs_off_t offset, @@ -225,6 +222,23 @@ xfs_setfilesize( return xfs_trans_commit(tp); } +int +xfs_setfilesize( + struct xfs_inode *ip, + xfs_off_t offset, + size_t size) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + int error; + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, &tp); + if (error) + return error; + + return __xfs_setfilesize(ip, tp, offset, size); +} + STATIC int xfs_setfilesize_ioend( struct xfs_ioend *ioend, @@ -247,7 +261,7 @@ xfs_setfilesize_ioend( return error; } - return xfs_setfilesize(ip, tp, ioend->io_offset, ioend->io_size); + return __xfs_setfilesize(ip, tp, ioend->io_offset, ioend->io_size); } /* @@ -270,6 +284,25 @@ xfs_end_io( error = -EIO; /* + * For a CoW extent, we need to move the mapping from the CoW fork + * to the data fork. If instead an error happened, just dump the + * new blocks. + */ + if (ioend->io_type == XFS_IO_COW) { + if (error) + goto done; + if (ioend->io_bio->bi_error) { + error = xfs_reflink_cancel_cow_range(ip, + ioend->io_offset, ioend->io_size); + goto done; + } + error = xfs_reflink_end_cow(ip, ioend->io_offset, + ioend->io_size); + if (error) + goto done; + } + + /* * For unwritten extents we need to issue transactions to convert a * range to normal written extens after the data I/O has finished. * Detecting and handling completion IO errors is done individually @@ -284,7 +317,8 @@ xfs_end_io( } else if (ioend->io_append_trans) { error = xfs_setfilesize_ioend(ioend, error); } else { - ASSERT(!xfs_ioend_is_append(ioend)); + ASSERT(!xfs_ioend_is_append(ioend) || + ioend->io_type == XFS_IO_COW); } done: @@ -298,7 +332,7 @@ xfs_end_bio( struct xfs_ioend *ioend = bio->bi_private; struct xfs_mount *mp = XFS_I(ioend->io_inode)->i_mount; - if (ioend->io_type == XFS_IO_UNWRITTEN) + if (ioend->io_type == XFS_IO_UNWRITTEN || ioend->io_type == XFS_IO_COW) queue_work(mp->m_unwritten_workqueue, &ioend->io_work); else if (ioend->io_append_trans) queue_work(mp->m_data_workqueue, &ioend->io_work); @@ -324,6 +358,7 @@ xfs_map_blocks( if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; + ASSERT(type != XFS_IO_COW); if (type == XFS_IO_UNWRITTEN) bmapi_flags |= XFS_BMAPI_IGSTATE; @@ -338,6 +373,13 @@ xfs_map_blocks( offset_fsb = XFS_B_TO_FSBT(mp, offset); error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, imap, &nimaps, bmapi_flags); + /* + * Truncate an overwrite extent if there's a pending CoW + * reservation before the end of this extent. This forces us + * to come back to writepage to take care of the CoW. + */ + if (nimaps && type == XFS_IO_OVERWRITE) + xfs_reflink_trim_irec_to_next_cow(ip, offset_fsb, imap); xfs_iunlock(ip, XFS_ILOCK_SHARED); if (error) @@ -345,7 +387,8 @@ xfs_map_blocks( if (type == XFS_IO_DELALLOC && (!nimaps || isnullstartblock(imap->br_startblock))) { - error = xfs_iomap_write_allocate(ip, offset, imap); + error = xfs_iomap_write_allocate(ip, XFS_DATA_FORK, offset, + imap); if (!error) trace_xfs_map_blocks_alloc(ip, offset, count, type, imap); return error; @@ -447,8 +490,8 @@ xfs_submit_ioend( ioend->io_bio->bi_private = ioend; ioend->io_bio->bi_end_io = xfs_end_bio; - bio_set_op_attrs(ioend->io_bio, REQ_OP_WRITE, - (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : 0); + ioend->io_bio->bi_opf = REQ_OP_WRITE | wbc_to_write_flags(wbc); + /* * If we are failing the IO now, just mark the ioend with an * error and finish it. This will run IO completion immediately @@ -519,8 +562,7 @@ xfs_chain_bio( bio_chain(ioend->io_bio, new); bio_get(ioend->io_bio); /* for xfs_destroy_ioend */ - bio_set_op_attrs(ioend->io_bio, REQ_OP_WRITE, - (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : 0); + ioend->io_bio->bi_opf = REQ_OP_WRITE | wbc_to_write_flags(wbc); submit_bio(ioend->io_bio); ioend->io_bio = new; } @@ -720,6 +762,56 @@ out_invalidate: return; } +static int +xfs_map_cow( + struct xfs_writepage_ctx *wpc, + struct inode *inode, + loff_t offset, + unsigned int *new_type) +{ + struct xfs_inode *ip = XFS_I(inode); + struct xfs_bmbt_irec imap; + bool is_cow = false; + int error; + + /* + * If we already have a valid COW mapping keep using it. + */ + if (wpc->io_type == XFS_IO_COW) { + wpc->imap_valid = xfs_imap_valid(inode, &wpc->imap, offset); + if (wpc->imap_valid) { + *new_type = XFS_IO_COW; + return 0; + } + } + + /* + * Else we need to check if there is a COW mapping at this offset. + */ + xfs_ilock(ip, XFS_ILOCK_SHARED); + is_cow = xfs_reflink_find_cow_mapping(ip, offset, &imap); + xfs_iunlock(ip, XFS_ILOCK_SHARED); + + if (!is_cow) + return 0; + + /* + * And if the COW mapping has a delayed extent here we need to + * allocate real space for it now. + */ + if (isnullstartblock(imap.br_startblock)) { + error = xfs_iomap_write_allocate(ip, XFS_COW_FORK, offset, + &imap); + if (error) + return error; + } + + wpc->io_type = *new_type = XFS_IO_COW; + wpc->imap_valid = true; + wpc->imap = imap; + return 0; +} + /* * We implement an immediate ioend submission policy here to avoid needing to * chain multiple ioends and hence nest mempool allocations which can violate @@ -752,6 +844,7 @@ xfs_writepage_map( int error = 0; int count = 0; int uptodate = 1; + unsigned int new_type; bh = head = page_buffers(page); offset = page_offset(page); @@ -772,22 +865,13 @@ xfs_writepage_map( continue; } - if (buffer_unwritten(bh)) { - if (wpc->io_type != XFS_IO_UNWRITTEN) { - wpc->io_type = XFS_IO_UNWRITTEN; - wpc->imap_valid = false; - } - } else if (buffer_delay(bh)) { - if (wpc->io_type != XFS_IO_DELALLOC) { - wpc->io_type = XFS_IO_DELALLOC; - wpc->imap_valid = false; - } - } else if (buffer_uptodate(bh)) { - if (wpc->io_type != XFS_IO_OVERWRITE) { - wpc->io_type = XFS_IO_OVERWRITE; - wpc->imap_valid = false; - } - } else { + if (buffer_unwritten(bh)) + new_type = XFS_IO_UNWRITTEN; + else if (buffer_delay(bh)) + new_type = XFS_IO_DELALLOC; + else if (buffer_uptodate(bh)) + new_type = XFS_IO_OVERWRITE; + else { if (PageUptodate(page)) ASSERT(buffer_mapped(bh)); /* @@ -800,6 +884,17 @@ xfs_writepage_map( continue; } + if (xfs_is_reflink_inode(XFS_I(inode))) { + error = xfs_map_cow(wpc, inode, offset, &new_type); + if (error) + goto out; + } + + if (wpc->io_type != new_type) { + wpc->io_type = new_type; + wpc->imap_valid = false; + } + if (wpc->imap_valid) wpc->imap_valid = xfs_imap_valid(inode, &wpc->imap, offset); @@ -1075,39 +1170,6 @@ xfs_vm_releasepage( } /* - * When we map a DIO buffer, we may need to pass flags to - * xfs_end_io_direct_write to tell it what kind of write IO we are doing. - * - * Note that for DIO, an IO to the highest supported file block offset (i.e. - * 2^63 - 1FSB bytes) will result in the offset + count overflowing a signed 64 - * bit variable. Hence if we see this overflow, we have to assume that the IO is - * extending the file size. We won't know for sure until IO completion is run - * and the actual max write offset is communicated to the IO completion - * routine. - */ -static void -xfs_map_direct( - struct inode *inode, - struct buffer_head *bh_result, - struct xfs_bmbt_irec *imap, - xfs_off_t offset) -{ - uintptr_t *flags = (uintptr_t *)&bh_result->b_private; - xfs_off_t size = bh_result->b_size; - - trace_xfs_get_blocks_map_direct(XFS_I(inode), offset, size, - ISUNWRITTEN(imap) ? XFS_IO_UNWRITTEN : XFS_IO_OVERWRITE, imap); - - if (ISUNWRITTEN(imap)) { - *flags |= XFS_DIO_FLAG_UNWRITTEN; - set_buffer_defer_completion(bh_result); - } else if (offset + size > i_size_read(inode) || offset + size < 0) { - *flags |= XFS_DIO_FLAG_APPEND; - set_buffer_defer_completion(bh_result); - } -} - -/* * If this is O_DIRECT or the mpage code calling tell them how large the mapping * is, so that we can avoid repeated get_blocks calls. * @@ -1147,14 +1209,12 @@ xfs_map_trim_size( bh_result->b_size = mapping_size; } -STATIC int -__xfs_get_blocks( +static int +xfs_get_blocks( struct inode *inode, sector_t iblock, struct buffer_head *bh_result, - int create, - bool direct, - bool dax_fault) + int create) { struct xfs_inode *ip = XFS_I(inode); struct xfs_mount *mp = ip->i_mount; @@ -1165,9 +1225,8 @@ __xfs_get_blocks( int nimaps = 1; xfs_off_t offset; ssize_t size; - int new = 0; - BUG_ON(create && !direct); + BUG_ON(create); if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; @@ -1176,7 +1235,7 @@ __xfs_get_blocks( ASSERT(bh_result->b_size >= (1 << inode->i_blkbits)); size = bh_result->b_size; - if (!create && offset >= i_size_read(inode)) + if (offset >= i_size_read(inode)) return 0; /* @@ -1196,29 +1255,7 @@ __xfs_get_blocks( if (error) goto out_unlock; - /* for DAX, we convert unwritten extents directly */ - if (create && - (!nimaps || - (imap.br_startblock == HOLESTARTBLOCK || - imap.br_startblock == DELAYSTARTBLOCK) || - (IS_DAX(inode) && ISUNWRITTEN(&imap)))) { - /* - * xfs_iomap_write_direct() expects the shared lock. It - * is unlocked on return. - */ - if (lockmode == XFS_ILOCK_EXCL) - xfs_ilock_demote(ip, lockmode); - - error = xfs_iomap_write_direct(ip, offset, size, - &imap, nimaps); - if (error) - return error; - new = 1; - - trace_xfs_get_blocks_alloc(ip, offset, size, - ISUNWRITTEN(&imap) ? XFS_IO_UNWRITTEN - : XFS_IO_DELALLOC, &imap); - } else if (nimaps) { + if (nimaps) { trace_xfs_get_blocks_found(ip, offset, size, ISUNWRITTEN(&imap) ? XFS_IO_UNWRITTEN : XFS_IO_OVERWRITE, &imap); @@ -1228,12 +1265,6 @@ __xfs_get_blocks( goto out_unlock; } - if (IS_DAX(inode) && create) { - ASSERT(!ISUNWRITTEN(&imap)); - /* zeroing is not needed at a higher layer */ - new = 0; - } - /* trim mapping down to size requested */ xfs_map_trim_size(inode, iblock, bh_result, &imap, offset, size); @@ -1243,42 +1274,14 @@ __xfs_get_blocks( */ if (imap.br_startblock != HOLESTARTBLOCK && imap.br_startblock != DELAYSTARTBLOCK && - (create || !ISUNWRITTEN(&imap))) { + !ISUNWRITTEN(&imap)) xfs_map_buffer(inode, bh_result, &imap, offset); - if (ISUNWRITTEN(&imap)) - set_buffer_unwritten(bh_result); - /* direct IO needs special help */ - if (create) { - if (dax_fault) - ASSERT(!ISUNWRITTEN(&imap)); - else - xfs_map_direct(inode, bh_result, &imap, offset); - } - } /* * If this is a realtime file, data may be on a different device. * to that pointed to from the buffer_head b_bdev currently. */ bh_result->b_bdev = xfs_find_bdev_for_inode(inode); - - /* - * If we previously allocated a block out beyond eof and we are now - * coming back to use it then we will need to flag it as new even if it - * has a disk address. - * - * With sub-block writes into unwritten extents we also need to mark - * the buffer as new so that the unwritten parts of the buffer gets - * correctly zeroed. - */ - if (create && - ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) || - (offset >= i_size_read(inode)) || - (new || ISUNWRITTEN(&imap)))) - set_buffer_new(bh_result); - - BUG_ON(direct && imap.br_startblock == DELAYSTARTBLOCK); - return 0; out_unlock: @@ -1286,113 +1289,6 @@ out_unlock: return error; } -int -xfs_get_blocks( - struct inode *inode, - sector_t iblock, - struct buffer_head *bh_result, - int create) -{ - return __xfs_get_blocks(inode, iblock, bh_result, create, false, false); -} - -int -xfs_get_blocks_direct( - struct inode *inode, - sector_t iblock, - struct buffer_head *bh_result, - int create) -{ - return __xfs_get_blocks(inode, iblock, bh_result, create, true, false); -} - -int -xfs_get_blocks_dax_fault( - struct inode *inode, - sector_t iblock, - struct buffer_head *bh_result, - int create) -{ - return __xfs_get_blocks(inode, iblock, bh_result, create, true, true); -} - -/* - * Complete a direct I/O write request. - * - * xfs_map_direct passes us some flags in the private data to tell us what to - * do. If no flags are set, then the write IO is an overwrite wholly within - * the existing allocated file size and so there is nothing for us to do. - * - * Note that in this case the completion can be called in interrupt context, - * whereas if we have flags set we will always be called in task context - * (i.e. from a workqueue). - */ -int -xfs_end_io_direct_write( - struct kiocb *iocb, - loff_t offset, - ssize_t size, - void *private) -{ - struct inode *inode = file_inode(iocb->ki_filp); - struct xfs_inode *ip = XFS_I(inode); - struct xfs_mount *mp = ip->i_mount; - uintptr_t flags = (uintptr_t)private; - int error = 0; - - trace_xfs_end_io_direct_write(ip, offset, size); - - if (XFS_FORCED_SHUTDOWN(mp)) - return -EIO; - - if (size <= 0) - return size; - - /* - * The flags tell us whether we are doing unwritten extent conversions - * or an append transaction that updates the on-disk file size. These - * cases are the only cases where we should *potentially* be needing - * to update the VFS inode size. - */ - if (flags == 0) { - ASSERT(offset + size <= i_size_read(inode)); - return 0; - } - - /* - * We need to update the in-core inode size here so that we don't end up - * with the on-disk inode size being outside the in-core inode size. We - * have no other method of updating EOF for AIO, so always do it here - * if necessary. - * - * We need to lock the test/set EOF update as we can be racing with - * other IO completions here to update the EOF. Failing to serialise - * here can result in EOF moving backwards and Bad Things Happen when - * that occurs. - */ - spin_lock(&ip->i_flags_lock); - if (offset + size > i_size_read(inode)) - i_size_write(inode, offset + size); - spin_unlock(&ip->i_flags_lock); - - if (flags & XFS_DIO_FLAG_UNWRITTEN) { - trace_xfs_end_io_direct_write_unwritten(ip, offset, size); - - error = xfs_iomap_write_unwritten(ip, offset, size); - } else if (flags & XFS_DIO_FLAG_APPEND) { - struct xfs_trans *tp; - - trace_xfs_end_io_direct_write_append(ip, offset, size); - - error = xfs_trans_alloc(mp, &M_RES(mp)->tr_fsyncts, 0, 0, 0, - &tp); - if (!error) - error = xfs_setfilesize(ip, tp, offset, size); - } - - return error; -} - STATIC ssize_t xfs_vm_direct_IO( struct kiocb *iocb, @@ -1413,9 +1309,17 @@ xfs_vm_bmap( struct xfs_inode *ip = XFS_I(inode); trace_xfs_vm_bmap(XFS_I(inode)); - xfs_ilock(ip, XFS_IOLOCK_SHARED); + + /* + * The swap code (ab-)uses ->bmap to get a block mapping and then + * bypasseѕ the file system for actual I/O. We really can't allow + * that on reflinks inodes, so we have to skip out here. And yes, + * 0 is the magic code for a bmap error.. + */ + if (xfs_is_reflink_inode(ip)) + return 0; + filemap_write_and_wait(mapping); - xfs_iunlock(ip, XFS_IOLOCK_SHARED); return generic_block_bmap(mapping, block, xfs_get_blocks); } diff --git a/fs/xfs/xfs_aops.h b/fs/xfs/xfs_aops.h index bf2d9a141a73..cc174ec6c2fd 100644 --- a/fs/xfs/xfs_aops.h +++ b/fs/xfs/xfs_aops.h @@ -28,13 +28,15 @@ enum { XFS_IO_DELALLOC, /* covers delalloc region */ XFS_IO_UNWRITTEN, /* covers allocated but uninitialized data */ XFS_IO_OVERWRITE, /* covers already allocated extent */ + XFS_IO_COW, /* covers copy-on-write extent */ }; #define XFS_IO_TYPES \ { XFS_IO_INVALID, "invalid" }, \ { XFS_IO_DELALLOC, "delalloc" }, \ { XFS_IO_UNWRITTEN, "unwritten" }, \ - { XFS_IO_OVERWRITE, "overwrite" } + { XFS_IO_OVERWRITE, "overwrite" }, \ + { XFS_IO_COW, "CoW" } /* * Structure for buffered I/O completions. @@ -53,15 +55,7 @@ struct xfs_ioend { extern const struct address_space_operations xfs_address_space_operations; -int xfs_get_blocks(struct inode *inode, sector_t offset, - struct buffer_head *map_bh, int create); -int xfs_get_blocks_direct(struct inode *inode, sector_t offset, - struct buffer_head *map_bh, int create); -int xfs_get_blocks_dax_fault(struct inode *inode, sector_t offset, - struct buffer_head *map_bh, int create); - -int xfs_end_io_direct_write(struct kiocb *iocb, loff_t offset, - ssize_t size, void *private); +int xfs_setfilesize(struct xfs_inode *ip, xfs_off_t offset, size_t size); extern void xfs_count_page_state(struct page *, int *, int *); extern struct block_device *xfs_find_bdev_for_inode(struct inode *); diff --git a/fs/xfs/xfs_attr.h b/fs/xfs/xfs_attr.h index e3da5d448bcf..d14691aa02b4 100644 --- a/fs/xfs/xfs_attr.h +++ b/fs/xfs/xfs_attr.h @@ -112,8 +112,8 @@ typedef struct attrlist_cursor_kern { *========================================================================*/ -/* Return 0 on success, or -errno; other state communicated via *context */ -typedef int (*put_listent_func_t)(struct xfs_attr_list_context *, int, +/* void; state communicated via *context */ +typedef void (*put_listent_func_t)(struct xfs_attr_list_context *, int, unsigned char *, int, int); typedef struct xfs_attr_list_context { diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index 25e76cd6c053..97c45b6eb91e 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -74,7 +74,6 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) xfs_attr_sf_entry_t *sfe; xfs_inode_t *dp; int sbsize, nsbuf, count, i; - int error; ASSERT(context != NULL); dp = context->dp; @@ -102,13 +101,11 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) (XFS_ISRESET_CURSOR(cursor) && (dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize)) { for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) { - error = context->put_listent(context, - sfe->flags, - sfe->nameval, - (int)sfe->namelen, - (int)sfe->valuelen); - if (error) - return error; + context->put_listent(context, + sfe->flags, + sfe->nameval, + (int)sfe->namelen, + (int)sfe->valuelen); /* * Either search callback finished early or * didn't fit it all in the buffer after all. @@ -193,15 +190,11 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context) cursor->hashval = sbp->hash; cursor->offset = 0; } - error = context->put_listent(context, - sbp->flags, - sbp->name, - sbp->namelen, - sbp->valuelen); - if (error) { - kmem_free(sbuf); - return error; - } + context->put_listent(context, + sbp->flags, + sbp->name, + sbp->namelen, + sbp->valuelen); if (context->seen_enough) break; cursor->offset++; @@ -335,11 +328,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context) */ for (;;) { leaf = bp->b_addr; - error = xfs_attr3_leaf_list_int(bp, context); - if (error) { - xfs_trans_brelse(NULL, bp); - return error; - } + xfs_attr3_leaf_list_int(bp, context); xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf); if (context->seen_enough || leafhdr.forw == 0) break; @@ -356,7 +345,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context) /* * Copy out attribute list entries for attr_list(), for leaf attribute lists. */ -int +void xfs_attr3_leaf_list_int( struct xfs_buf *bp, struct xfs_attr_list_context *context) @@ -366,7 +355,6 @@ xfs_attr3_leaf_list_int( struct xfs_attr3_icleaf_hdr ichdr; struct xfs_attr_leaf_entry *entries; struct xfs_attr_leaf_entry *entry; - int retval; int i; struct xfs_mount *mp = context->dp->i_mount; @@ -399,7 +387,7 @@ xfs_attr3_leaf_list_int( } if (i == ichdr.count) { trace_xfs_attr_list_notfound(context); - return 0; + return; } } else { entry = &entries[0]; @@ -410,7 +398,6 @@ xfs_attr3_leaf_list_int( /* * We have found our place, start copying out the new attributes. */ - retval = 0; for (; i < ichdr.count; entry++, i++) { char *name; int namelen, valuelen; @@ -439,16 +426,14 @@ xfs_attr3_leaf_list_int( valuelen = be32_to_cpu(name_rmt->valuelen); } - retval = context->put_listent(context, entry->flags, + context->put_listent(context, entry->flags, name, namelen, valuelen); - if (retval) - break; if (context->seen_enough) break; cursor->offset++; } trace_xfs_attr_list_leaf_end(context); - return retval; + return; } /* @@ -467,9 +452,9 @@ xfs_attr_leaf_list(xfs_attr_list_context_t *context) if (error) return error; - error = xfs_attr3_leaf_list_int(bp, context); + xfs_attr3_leaf_list_int(bp, context); xfs_trans_brelse(NULL, bp); - return error; + return 0; } int @@ -513,7 +498,7 @@ xfs_attr_list_int( * Take care to check values and protect against them changing later, * we may be reading them directly out of a user buffer. */ -STATIC int +STATIC void xfs_attr_put_listent( xfs_attr_list_context_t *context, int flags, @@ -536,10 +521,10 @@ xfs_attr_put_listent( */ if (((context->flags & ATTR_SECURE) == 0) != ((flags & XFS_ATTR_SECURE) == 0)) - return 0; + return; if (((context->flags & ATTR_ROOT) == 0) != ((flags & XFS_ATTR_ROOT) == 0)) - return 0; + return; arraytop = sizeof(*alist) + context->count * sizeof(alist->al_offset[0]); @@ -548,7 +533,7 @@ xfs_attr_put_listent( trace_xfs_attr_list_full(context); alist->al_more = 1; context->seen_enough = 1; - return 0; + return; } aep = (attrlist_ent_t *)&context->alist[context->firstu]; @@ -558,7 +543,7 @@ xfs_attr_put_listent( alist->al_offset[context->count++] = context->firstu; alist->al_count = context->count; trace_xfs_attr_list_add(context); - return 0; + return; } /* diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c new file mode 100644 index 000000000000..9bf57c76623b --- /dev/null +++ b/fs/xfs/xfs_bmap_item.c @@ -0,0 +1,508 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_bit.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_inode.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_buf_item.h" +#include "xfs_bmap_item.h" +#include "xfs_log.h" +#include "xfs_bmap.h" +#include "xfs_icache.h" +#include "xfs_trace.h" + + +kmem_zone_t *xfs_bui_zone; +kmem_zone_t *xfs_bud_zone; + +static inline struct xfs_bui_log_item *BUI_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_bui_log_item, bui_item); +} + +void +xfs_bui_item_free( + struct xfs_bui_log_item *buip) +{ + kmem_zone_free(xfs_bui_zone, buip); +} + +STATIC void +xfs_bui_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + struct xfs_bui_log_item *buip = BUI_ITEM(lip); + + *nvecs += 1; + *nbytes += xfs_bui_log_format_sizeof(buip->bui_format.bui_nextents); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given bui log item. We use only 1 iovec, and we point that + * at the bui_log_format structure embedded in the bui item. + * It is at this point that we assert that all of the extent + * slots in the bui item have been filled. + */ +STATIC void +xfs_bui_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_bui_log_item *buip = BUI_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + ASSERT(atomic_read(&buip->bui_next_extent) == + buip->bui_format.bui_nextents); + + buip->bui_format.bui_type = XFS_LI_BUI; + buip->bui_format.bui_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_BUI_FORMAT, &buip->bui_format, + xfs_bui_log_format_sizeof(buip->bui_format.bui_nextents)); +} + +/* + * Pinning has no meaning for an bui item, so just return. + */ +STATIC void +xfs_bui_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * The unpin operation is the last place an BUI is manipulated in the log. It is + * either inserted in the AIL or aborted in the event of a log I/O error. In + * either case, the BUI transaction has been successfully committed to make it + * this far. Therefore, we expect whoever committed the BUI to either construct + * and commit the BUD or drop the BUD's reference in the event of error. Simply + * drop the log's BUI reference now that the log is done with it. + */ +STATIC void +xfs_bui_item_unpin( + struct xfs_log_item *lip, + int remove) +{ + struct xfs_bui_log_item *buip = BUI_ITEM(lip); + + xfs_bui_release(buip); +} + +/* + * BUI items have no locking or pushing. However, since BUIs are pulled from + * the AIL when their corresponding BUDs are committed to disk, their situation + * is very similar to being pinned. Return XFS_ITEM_PINNED so that the caller + * will eventually flush the log. This should help in getting the BUI out of + * the AIL. + */ +STATIC uint +xfs_bui_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The BUI has been either committed or aborted if the transaction has been + * cancelled. If the transaction was cancelled, an BUD isn't going to be + * constructed and thus we free the BUI here directly. + */ +STATIC void +xfs_bui_item_unlock( + struct xfs_log_item *lip) +{ + if (lip->li_flags & XFS_LI_ABORTED) + xfs_bui_item_free(BUI_ITEM(lip)); +} + +/* + * The BUI is logged only once and cannot be moved in the log, so simply return + * the lsn at which it's been logged. + */ +STATIC xfs_lsn_t +xfs_bui_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + return lsn; +} + +/* + * The BUI dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_bui_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all bui log items. + */ +static const struct xfs_item_ops xfs_bui_item_ops = { + .iop_size = xfs_bui_item_size, + .iop_format = xfs_bui_item_format, + .iop_pin = xfs_bui_item_pin, + .iop_unpin = xfs_bui_item_unpin, + .iop_unlock = xfs_bui_item_unlock, + .iop_committed = xfs_bui_item_committed, + .iop_push = xfs_bui_item_push, + .iop_committing = xfs_bui_item_committing, +}; + +/* + * Allocate and initialize an bui item with the given number of extents. + */ +struct xfs_bui_log_item * +xfs_bui_init( + struct xfs_mount *mp) + +{ + struct xfs_bui_log_item *buip; + + buip = kmem_zone_zalloc(xfs_bui_zone, KM_SLEEP); + + xfs_log_item_init(mp, &buip->bui_item, XFS_LI_BUI, &xfs_bui_item_ops); + buip->bui_format.bui_nextents = XFS_BUI_MAX_FAST_EXTENTS; + buip->bui_format.bui_id = (uintptr_t)(void *)buip; + atomic_set(&buip->bui_next_extent, 0); + atomic_set(&buip->bui_refcount, 2); + + return buip; +} + +/* + * Freeing the BUI requires that we remove it from the AIL if it has already + * been placed there. However, the BUI may not yet have been placed in the AIL + * when called by xfs_bui_release() from BUD processing due to the ordering of + * committed vs unpin operations in bulk insert operations. Hence the reference + * count to ensure only the last caller frees the BUI. + */ +void +xfs_bui_release( + struct xfs_bui_log_item *buip) +{ + if (atomic_dec_and_test(&buip->bui_refcount)) { + xfs_trans_ail_remove(&buip->bui_item, SHUTDOWN_LOG_IO_ERROR); + xfs_bui_item_free(buip); + } +} + +static inline struct xfs_bud_log_item *BUD_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_bud_log_item, bud_item); +} + +STATIC void +xfs_bud_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + *nvecs += 1; + *nbytes += sizeof(struct xfs_bud_log_format); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given bud log item. We use only 1 iovec, and we point that + * at the bud_log_format structure embedded in the bud item. + * It is at this point that we assert that all of the extent + * slots in the bud item have been filled. + */ +STATIC void +xfs_bud_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_bud_log_item *budp = BUD_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + budp->bud_format.bud_type = XFS_LI_BUD; + budp->bud_format.bud_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_BUD_FORMAT, &budp->bud_format, + sizeof(struct xfs_bud_log_format)); +} + +/* + * Pinning has no meaning for an bud item, so just return. + */ +STATIC void +xfs_bud_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * Since pinning has no meaning for an bud item, unpinning does + * not either. + */ +STATIC void +xfs_bud_item_unpin( + struct xfs_log_item *lip, + int remove) +{ +} + +/* + * There isn't much you can do to push on an bud item. It is simply stuck + * waiting for the log to be flushed to disk. + */ +STATIC uint +xfs_bud_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The BUD is either committed or aborted if the transaction is cancelled. If + * the transaction is cancelled, drop our reference to the BUI and free the + * BUD. + */ +STATIC void +xfs_bud_item_unlock( + struct xfs_log_item *lip) +{ + struct xfs_bud_log_item *budp = BUD_ITEM(lip); + + if (lip->li_flags & XFS_LI_ABORTED) { + xfs_bui_release(budp->bud_buip); + kmem_zone_free(xfs_bud_zone, budp); + } +} + +/* + * When the bud item is committed to disk, all we need to do is delete our + * reference to our partner bui item and then free ourselves. Since we're + * freeing ourselves we must return -1 to keep the transaction code from + * further referencing this item. + */ +STATIC xfs_lsn_t +xfs_bud_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + struct xfs_bud_log_item *budp = BUD_ITEM(lip); + + /* + * Drop the BUI reference regardless of whether the BUD has been + * aborted. Once the BUD transaction is constructed, it is the sole + * responsibility of the BUD to release the BUI (even if the BUI is + * aborted due to log I/O error). + */ + xfs_bui_release(budp->bud_buip); + kmem_zone_free(xfs_bud_zone, budp); + + return (xfs_lsn_t)-1; +} + +/* + * The BUD dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_bud_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all bud log items. + */ +static const struct xfs_item_ops xfs_bud_item_ops = { + .iop_size = xfs_bud_item_size, + .iop_format = xfs_bud_item_format, + .iop_pin = xfs_bud_item_pin, + .iop_unpin = xfs_bud_item_unpin, + .iop_unlock = xfs_bud_item_unlock, + .iop_committed = xfs_bud_item_committed, + .iop_push = xfs_bud_item_push, + .iop_committing = xfs_bud_item_committing, +}; + +/* + * Allocate and initialize an bud item with the given number of extents. + */ +struct xfs_bud_log_item * +xfs_bud_init( + struct xfs_mount *mp, + struct xfs_bui_log_item *buip) + +{ + struct xfs_bud_log_item *budp; + + budp = kmem_zone_zalloc(xfs_bud_zone, KM_SLEEP); + xfs_log_item_init(mp, &budp->bud_item, XFS_LI_BUD, &xfs_bud_item_ops); + budp->bud_buip = buip; + budp->bud_format.bud_bui_id = buip->bui_format.bui_id; + + return budp; +} + +/* + * Process a bmap update intent item that was recovered from the log. + * We need to update some inode's bmbt. + */ +int +xfs_bui_recover( + struct xfs_mount *mp, + struct xfs_bui_log_item *buip) +{ + int error = 0; + unsigned int bui_type; + struct xfs_map_extent *bmap; + xfs_fsblock_t startblock_fsb; + xfs_fsblock_t inode_fsb; + bool op_ok; + struct xfs_bud_log_item *budp; + enum xfs_bmap_intent_type type; + int whichfork; + xfs_exntst_t state; + struct xfs_trans *tp; + struct xfs_inode *ip = NULL; + struct xfs_defer_ops dfops; + xfs_fsblock_t firstfsb; + + ASSERT(!test_bit(XFS_BUI_RECOVERED, &buip->bui_flags)); + + /* Only one mapping operation per BUI... */ + if (buip->bui_format.bui_nextents != XFS_BUI_MAX_FAST_EXTENTS) { + set_bit(XFS_BUI_RECOVERED, &buip->bui_flags); + xfs_bui_release(buip); + return -EIO; + } + + /* + * First check the validity of the extent described by the + * BUI. If anything is bad, then toss the BUI. + */ + bmap = &buip->bui_format.bui_extents[0]; + startblock_fsb = XFS_BB_TO_FSB(mp, + XFS_FSB_TO_DADDR(mp, bmap->me_startblock)); + inode_fsb = XFS_BB_TO_FSB(mp, XFS_FSB_TO_DADDR(mp, + XFS_INO_TO_FSB(mp, bmap->me_owner))); + switch (bmap->me_flags & XFS_BMAP_EXTENT_TYPE_MASK) { + case XFS_BMAP_MAP: + case XFS_BMAP_UNMAP: + op_ok = true; + break; + default: + op_ok = false; + break; + } + if (!op_ok || startblock_fsb == 0 || + bmap->me_len == 0 || + inode_fsb == 0 || + startblock_fsb >= mp->m_sb.sb_dblocks || + bmap->me_len >= mp->m_sb.sb_agblocks || + inode_fsb >= mp->m_sb.sb_dblocks || + (bmap->me_flags & ~XFS_BMAP_EXTENT_FLAGS)) { + /* + * This will pull the BUI from the AIL and + * free the memory associated with it. + */ + set_bit(XFS_BUI_RECOVERED, &buip->bui_flags); + xfs_bui_release(buip); + return -EIO; + } + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); + if (error) + return error; + budp = xfs_trans_get_bud(tp, buip); + + /* Grab the inode. */ + error = xfs_iget(mp, tp, bmap->me_owner, 0, XFS_ILOCK_EXCL, &ip); + if (error) + goto err_inode; + + if (VFS_I(ip)->i_nlink == 0) + xfs_iflags_set(ip, XFS_IRECOVERY); + xfs_defer_init(&dfops, &firstfsb); + + /* Process deferred bmap item. */ + state = (bmap->me_flags & XFS_BMAP_EXTENT_UNWRITTEN) ? + XFS_EXT_UNWRITTEN : XFS_EXT_NORM; + whichfork = (bmap->me_flags & XFS_BMAP_EXTENT_ATTR_FORK) ? + XFS_ATTR_FORK : XFS_DATA_FORK; + bui_type = bmap->me_flags & XFS_BMAP_EXTENT_TYPE_MASK; + switch (bui_type) { + case XFS_BMAP_MAP: + case XFS_BMAP_UNMAP: + type = bui_type; + break; + default: + error = -EFSCORRUPTED; + goto err_dfops; + } + xfs_trans_ijoin(tp, ip, 0); + + error = xfs_trans_log_finish_bmap_update(tp, budp, &dfops, type, + ip, whichfork, bmap->me_startoff, + bmap->me_startblock, bmap->me_len, + state); + if (error) + goto err_dfops; + + /* Finish transaction, free inodes. */ + error = xfs_defer_finish(&tp, &dfops, NULL); + if (error) + goto err_dfops; + + set_bit(XFS_BUI_RECOVERED, &buip->bui_flags); + error = xfs_trans_commit(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + IRELE(ip); + + return error; + +err_dfops: + xfs_defer_cancel(&dfops); +err_inode: + xfs_trans_cancel(tp); + if (ip) { + xfs_iunlock(ip, XFS_ILOCK_EXCL); + IRELE(ip); + } + return error; +} diff --git a/fs/xfs/xfs_bmap_item.h b/fs/xfs/xfs_bmap_item.h new file mode 100644 index 000000000000..c867daae4a3c --- /dev/null +++ b/fs/xfs/xfs_bmap_item.h @@ -0,0 +1,98 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_BMAP_ITEM_H__ +#define __XFS_BMAP_ITEM_H__ + +/* + * There are (currently) two pairs of bmap btree redo item types: map & unmap. + * The common abbreviations for these are BUI (bmap update intent) and BUD + * (bmap update done). The redo item type is encoded in the flags field of + * each xfs_map_extent. + * + * *I items should be recorded in the *first* of a series of rolled + * transactions, and the *D items should be recorded in the same transaction + * that records the associated bmbt updates. + * + * Should the system crash after the commit of the first transaction but + * before the commit of the final transaction in a series, log recovery will + * use the redo information recorded by the intent items to replay the + * bmbt metadata updates in the non-first transaction. + */ + +/* kernel only BUI/BUD definitions */ + +struct xfs_mount; +struct kmem_zone; + +/* + * Max number of extents in fast allocation path. + */ +#define XFS_BUI_MAX_FAST_EXTENTS 1 + +/* + * Define BUI flag bits. Manipulated by set/clear/test_bit operators. + */ +#define XFS_BUI_RECOVERED 1 + +/* + * This is the "bmap update intent" log item. It is used to log the fact that + * some reverse mappings need to change. It is used in conjunction with the + * "bmap update done" log item described below. + * + * These log items follow the same rules as struct xfs_efi_log_item; see the + * comments about that structure (in xfs_extfree_item.h) for more details. + */ +struct xfs_bui_log_item { + struct xfs_log_item bui_item; + atomic_t bui_refcount; + atomic_t bui_next_extent; + unsigned long bui_flags; /* misc flags */ + struct xfs_bui_log_format bui_format; +}; + +static inline size_t +xfs_bui_log_item_sizeof( + unsigned int nr) +{ + return offsetof(struct xfs_bui_log_item, bui_format) + + xfs_bui_log_format_sizeof(nr); +} + +/* + * This is the "bmap update done" log item. It is used to log the fact that + * some bmbt updates mentioned in an earlier bui item have been performed. + */ +struct xfs_bud_log_item { + struct xfs_log_item bud_item; + struct xfs_bui_log_item *bud_buip; + struct xfs_bud_log_format bud_format; +}; + +extern struct kmem_zone *xfs_bui_zone; +extern struct kmem_zone *xfs_bud_zone; + +struct xfs_bui_log_item *xfs_bui_init(struct xfs_mount *); +struct xfs_bud_log_item *xfs_bud_init(struct xfs_mount *, + struct xfs_bui_log_item *); +void xfs_bui_item_free(struct xfs_bui_log_item *); +void xfs_bui_release(struct xfs_bui_log_item *); +int xfs_bui_recover(struct xfs_mount *mp, struct xfs_bui_log_item *buip); + +#endif /* __XFS_BMAP_ITEM_H__ */ diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index 4ece4f2ffc72..b9abce524c33 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -42,6 +42,9 @@ #include "xfs_icache.h" #include "xfs_log.h" #include "xfs_rmap_btree.h" +#include "xfs_iomap.h" +#include "xfs_reflink.h" +#include "xfs_refcount.h" /* Kernel only BMAP related definitions and functions */ @@ -182,7 +185,7 @@ xfs_bmap_rtalloc( XFS_TRANS_DQ_RTBCOUNT, (long) ralen); /* Zero the extent if we were asked to do so */ - if (ap->userdata & XFS_ALLOC_USERDATA_ZERO) { + if (ap->datatype & XFS_ALLOC_USERDATA_ZERO) { error = xfs_zero_extent(ap->ip, ap->blkno, ap->length); if (error) return error; @@ -356,9 +359,7 @@ xfs_bmap_count_blocks( mp = ip->i_mount; ifp = XFS_IFORK_PTR(ip, whichfork); if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) { - xfs_bmap_count_leaves(ifp, 0, - ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t), - count); + xfs_bmap_count_leaves(ifp, 0, xfs_iext_count(ifp), count); return 0; } @@ -389,11 +390,13 @@ xfs_bmap_count_blocks( STATIC int xfs_getbmapx_fix_eof_hole( xfs_inode_t *ip, /* xfs incore inode pointer */ + int whichfork, struct getbmapx *out, /* output structure */ int prealloced, /* this is a file with * preallocated data space */ __int64_t end, /* last block requested */ - xfs_fsblock_t startblock) + xfs_fsblock_t startblock, + bool moretocome) { __int64_t fixlen; xfs_mount_t *mp; /* file system mount point */ @@ -418,15 +421,91 @@ xfs_getbmapx_fix_eof_hole( else out->bmv_block = xfs_fsb_to_db(ip, startblock); fileblock = XFS_BB_TO_FSB(ip->i_mount, out->bmv_offset); - ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); - if (xfs_iext_bno_to_ext(ifp, fileblock, &lastx) && - (lastx == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))-1)) + ifp = XFS_IFORK_PTR(ip, whichfork); + if (!moretocome && + xfs_iext_bno_to_ext(ifp, fileblock, &lastx) && + (lastx == xfs_iext_count(ifp) - 1)) out->bmv_oflags |= BMV_OF_LAST; } return 1; } +/* Adjust the reported bmap around shared/unshared extent transitions. */ +STATIC int +xfs_getbmap_adjust_shared( + struct xfs_inode *ip, + int whichfork, + struct xfs_bmbt_irec *map, + struct getbmapx *out, + struct xfs_bmbt_irec *next_map) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_agnumber_t agno; + xfs_agblock_t agbno; + xfs_agblock_t ebno; + xfs_extlen_t elen; + xfs_extlen_t nlen; + int error; + + next_map->br_startblock = NULLFSBLOCK; + next_map->br_startoff = NULLFILEOFF; + next_map->br_blockcount = 0; + + /* Only written data blocks can be shared. */ + if (!xfs_is_reflink_inode(ip) || whichfork != XFS_DATA_FORK || + map->br_startblock == DELAYSTARTBLOCK || + map->br_startblock == HOLESTARTBLOCK || + ISUNWRITTEN(map)) + return 0; + + agno = XFS_FSB_TO_AGNO(mp, map->br_startblock); + agbno = XFS_FSB_TO_AGBNO(mp, map->br_startblock); + error = xfs_reflink_find_shared(mp, agno, agbno, map->br_blockcount, + &ebno, &elen, true); + if (error) + return error; + + if (ebno == NULLAGBLOCK) { + /* No shared blocks at all. */ + return 0; + } else if (agbno == ebno) { + /* + * Shared extent at (agbno, elen). Shrink the reported + * extent length and prepare to move the start of map[i] + * to agbno+elen, with the aim of (re)formatting the new + * map[i] the next time through the inner loop. + */ + out->bmv_length = XFS_FSB_TO_BB(mp, elen); + out->bmv_oflags |= BMV_OF_SHARED; + if (elen != map->br_blockcount) { + *next_map = *map; + next_map->br_startblock += elen; + next_map->br_startoff += elen; + next_map->br_blockcount -= elen; + } + map->br_blockcount -= elen; + } else { + /* + * There's an unshared extent (agbno, ebno - agbno) + * followed by shared extent at (ebno, elen). Shrink + * the reported extent length to cover only the unshared + * extent and prepare to move up the start of map[i] to + * ebno, with the aim of (re)formatting the new map[i] + * the next time through the inner loop. + */ + *next_map = *map; + nlen = ebno - agbno; + out->bmv_length = XFS_FSB_TO_BB(mp, nlen); + next_map->br_startblock += nlen; + next_map->br_startoff += nlen; + next_map->br_blockcount -= nlen; + map->br_blockcount -= nlen; + } + + return 0; +} + /* * Get inode's extents as described in bmv, and format for output. * Calls formatter to fill the user's buffer until all extents @@ -459,12 +538,28 @@ xfs_getbmap( int iflags; /* interface flags */ int bmapi_flags; /* flags for xfs_bmapi */ int cur_ext = 0; + struct xfs_bmbt_irec inject_map; mp = ip->i_mount; iflags = bmv->bmv_iflags; - whichfork = iflags & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK; - if (whichfork == XFS_ATTR_FORK) { +#ifndef DEBUG + /* Only allow CoW fork queries if we're debugging. */ + if (iflags & BMV_IF_COWFORK) + return -EINVAL; +#endif + if ((iflags & BMV_IF_ATTRFORK) && (iflags & BMV_IF_COWFORK)) + return -EINVAL; + + if (iflags & BMV_IF_ATTRFORK) + whichfork = XFS_ATTR_FORK; + else if (iflags & BMV_IF_COWFORK) + whichfork = XFS_COW_FORK; + else + whichfork = XFS_DATA_FORK; + + switch (whichfork) { + case XFS_ATTR_FORK: if (XFS_IFORK_Q(ip)) { if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS && ip->i_d.di_aformat != XFS_DINODE_FMT_BTREE && @@ -480,7 +575,20 @@ xfs_getbmap( prealloced = 0; fixlen = 1LL << 32; - } else { + break; + case XFS_COW_FORK: + if (ip->i_cformat != XFS_DINODE_FMT_EXTENTS) + return -EINVAL; + + if (xfs_get_cowextsz_hint(ip)) { + prealloced = 1; + fixlen = mp->m_super->s_maxbytes; + } else { + prealloced = 0; + fixlen = XFS_ISIZE(ip); + } + break; + default: if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS && ip->i_d.di_format != XFS_DINODE_FMT_BTREE && ip->i_d.di_format != XFS_DINODE_FMT_LOCAL) @@ -494,6 +602,7 @@ xfs_getbmap( prealloced = 0; fixlen = XFS_ISIZE(ip); } + break; } if (bmv->bmv_length == -1) { @@ -520,7 +629,8 @@ xfs_getbmap( return -ENOMEM; xfs_ilock(ip, XFS_IOLOCK_SHARED); - if (whichfork == XFS_DATA_FORK) { + switch (whichfork) { + case XFS_DATA_FORK: if (!(iflags & BMV_IF_DELALLOC) && (ip->i_delayed_blks || XFS_ISIZE(ip) > ip->i_d.di_size)) { error = filemap_write_and_wait(VFS_I(ip)->i_mapping); @@ -538,8 +648,14 @@ xfs_getbmap( } lock = xfs_ilock_data_map_shared(ip); - } else { + break; + case XFS_COW_FORK: + lock = XFS_ILOCK_SHARED; + xfs_ilock(ip, lock); + break; + case XFS_ATTR_FORK: lock = xfs_ilock_attr_map_shared(ip); + break; } /* @@ -581,7 +697,8 @@ xfs_getbmap( goto out_free_map; ASSERT(nmap <= subnex); - for (i = 0; i < nmap && nexleft && bmv->bmv_length; i++) { + for (i = 0; i < nmap && nexleft && bmv->bmv_length && + cur_ext < bmv->bmv_count; i++) { out[cur_ext].bmv_oflags = 0; if (map[i].br_state == XFS_EXT_UNWRITTEN) out[cur_ext].bmv_oflags |= BMV_OF_PREALLOC; @@ -614,9 +731,16 @@ xfs_getbmap( goto out_free_map; } - if (!xfs_getbmapx_fix_eof_hole(ip, &out[cur_ext], - prealloced, bmvend, - map[i].br_startblock)) + /* Is this a shared block? */ + error = xfs_getbmap_adjust_shared(ip, whichfork, + &map[i], &out[cur_ext], &inject_map); + if (error) + goto out_free_map; + + if (!xfs_getbmapx_fix_eof_hole(ip, whichfork, + &out[cur_ext], prealloced, bmvend, + map[i].br_startblock, + inject_map.br_startblock != NULLFSBLOCK)) goto out_free_map; bmv->bmv_offset = @@ -636,11 +760,16 @@ xfs_getbmap( continue; } - nexleft--; + if (inject_map.br_startblock != NULLFSBLOCK) { + map[i] = inject_map; + i--; + } else + nexleft--; bmv->bmv_entries++; cur_ext++; } - } while (nmap && nexleft && bmv->bmv_length); + } while (nmap && nexleft && bmv->bmv_length && + cur_ext < bmv->bmv_count); out_free_map: kmem_free(map); @@ -1433,8 +1562,8 @@ xfs_insert_file_space( */ static int xfs_swap_extents_check_format( - xfs_inode_t *ip, /* target inode */ - xfs_inode_t *tip) /* tmp inode */ + struct xfs_inode *ip, /* target inode */ + struct xfs_inode *tip) /* tmp inode */ { /* Should never get a local format */ @@ -1450,6 +1579,13 @@ xfs_swap_extents_check_format( return -EINVAL; /* + * If we have to use the (expensive) rmap swap method, we can + * handle any number of extents and any format. + */ + if (xfs_sb_version_hasrmapbt(&ip->i_mount->m_sb)) + return 0; + + /* * if the target inode is in extent form and the temp inode is in btree * form then we will end up with the target inode in the wrong format * as we already know there are less extents in the temp inode. @@ -1518,125 +1654,162 @@ xfs_swap_extent_flush( return 0; } -int -xfs_swap_extents( - xfs_inode_t *ip, /* target inode */ - xfs_inode_t *tip, /* tmp inode */ - xfs_swapext_t *sxp) +/* + * Move extents from one file to another, when rmap is enabled. + */ +STATIC int +xfs_swap_extent_rmap( + struct xfs_trans **tpp, + struct xfs_inode *ip, + struct xfs_inode *tip) { - xfs_mount_t *mp = ip->i_mount; - xfs_trans_t *tp; - xfs_bstat_t *sbp = &sxp->sx_stat; - xfs_ifork_t *tempifp, *ifp, *tifp; - int src_log_flags, target_log_flags; - int error = 0; - int aforkblks = 0; - int taforkblks = 0; - __uint64_t tmp; - int lock_flags; - - /* XXX: we can't do this with rmap, will fix later */ - if (xfs_sb_version_hasrmapbt(&mp->m_sb)) - return -EOPNOTSUPP; - - tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL); - if (!tempifp) { - error = -ENOMEM; - goto out; - } + struct xfs_bmbt_irec irec; + struct xfs_bmbt_irec uirec; + struct xfs_bmbt_irec tirec; + xfs_fileoff_t offset_fsb; + xfs_fileoff_t end_fsb; + xfs_filblks_t count_fsb; + xfs_fsblock_t firstfsb; + struct xfs_defer_ops dfops; + int error; + xfs_filblks_t ilen; + xfs_filblks_t rlen; + int nimaps; + __uint64_t tip_flags2; /* - * Lock the inodes against other IO, page faults and truncate to - * begin with. Then we can ensure the inodes are flushed and have no - * page cache safely. Once we have done this we can take the ilocks and - * do the rest of the checks. + * If the source file has shared blocks, we must flag the donor + * file as having shared blocks so that we get the shared-block + * rmap functions when we go to fix up the rmaps. The flags + * will be switch for reals later. */ - lock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL; - xfs_lock_two_inodes(ip, tip, XFS_IOLOCK_EXCL); - xfs_lock_two_inodes(ip, tip, XFS_MMAPLOCK_EXCL); - - /* Verify that both files have the same format */ - if ((VFS_I(ip)->i_mode & S_IFMT) != (VFS_I(tip)->i_mode & S_IFMT)) { - error = -EINVAL; - goto out_unlock; - } + tip_flags2 = tip->i_d.di_flags2; + if (ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK) + tip->i_d.di_flags2 |= XFS_DIFLAG2_REFLINK; + + offset_fsb = 0; + end_fsb = XFS_B_TO_FSB(ip->i_mount, i_size_read(VFS_I(ip))); + count_fsb = (xfs_filblks_t)(end_fsb - offset_fsb); + + while (count_fsb) { + /* Read extent from the donor file */ + nimaps = 1; + error = xfs_bmapi_read(tip, offset_fsb, count_fsb, &tirec, + &nimaps, 0); + if (error) + goto out; + ASSERT(nimaps == 1); + ASSERT(tirec.br_startblock != DELAYSTARTBLOCK); + + trace_xfs_swap_extent_rmap_remap(tip, &tirec); + ilen = tirec.br_blockcount; + + /* Unmap the old blocks in the source file. */ + while (tirec.br_blockcount) { + xfs_defer_init(&dfops, &firstfsb); + trace_xfs_swap_extent_rmap_remap_piece(tip, &tirec); + + /* Read extent from the source file */ + nimaps = 1; + error = xfs_bmapi_read(ip, tirec.br_startoff, + tirec.br_blockcount, &irec, + &nimaps, 0); + if (error) + goto out_defer; + ASSERT(nimaps == 1); + ASSERT(tirec.br_startoff == irec.br_startoff); + trace_xfs_swap_extent_rmap_remap_piece(ip, &irec); + + /* Trim the extent. */ + uirec = tirec; + uirec.br_blockcount = rlen = min_t(xfs_filblks_t, + tirec.br_blockcount, + irec.br_blockcount); + trace_xfs_swap_extent_rmap_remap_piece(tip, &uirec); + + /* Remove the mapping from the donor file. */ + error = xfs_bmap_unmap_extent((*tpp)->t_mountp, &dfops, + tip, &uirec); + if (error) + goto out_defer; - /* Verify both files are either real-time or non-realtime */ - if (XFS_IS_REALTIME_INODE(ip) != XFS_IS_REALTIME_INODE(tip)) { - error = -EINVAL; - goto out_unlock; - } + /* Remove the mapping from the source file. */ + error = xfs_bmap_unmap_extent((*tpp)->t_mountp, &dfops, + ip, &irec); + if (error) + goto out_defer; - error = xfs_swap_extent_flush(ip); - if (error) - goto out_unlock; - error = xfs_swap_extent_flush(tip); - if (error) - goto out_unlock; + /* Map the donor file's blocks into the source file. */ + error = xfs_bmap_map_extent((*tpp)->t_mountp, &dfops, + ip, &uirec); + if (error) + goto out_defer; - error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); - if (error) - goto out_unlock; + /* Map the source file's blocks into the donor file. */ + error = xfs_bmap_map_extent((*tpp)->t_mountp, &dfops, + tip, &irec); + if (error) + goto out_defer; - /* - * Lock and join the inodes to the tansaction so that transaction commit - * or cancel will unlock the inodes from this point onwards. - */ - xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL); - lock_flags |= XFS_ILOCK_EXCL; - xfs_trans_ijoin(tp, ip, lock_flags); - xfs_trans_ijoin(tp, tip, lock_flags); + error = xfs_defer_finish(tpp, &dfops, ip); + if (error) + goto out_defer; + tirec.br_startoff += rlen; + if (tirec.br_startblock != HOLESTARTBLOCK && + tirec.br_startblock != DELAYSTARTBLOCK) + tirec.br_startblock += rlen; + tirec.br_blockcount -= rlen; + } - /* Verify all data are being swapped */ - if (sxp->sx_offset != 0 || - sxp->sx_length != ip->i_d.di_size || - sxp->sx_length != tip->i_d.di_size) { - error = -EFAULT; - goto out_trans_cancel; + /* Roll on... */ + count_fsb -= ilen; + offset_fsb += ilen; } - trace_xfs_swap_extent_before(ip, 0); - trace_xfs_swap_extent_before(tip, 1); + tip->i_d.di_flags2 = tip_flags2; + return 0; - /* check inode formats now that data is flushed */ - error = xfs_swap_extents_check_format(ip, tip); - if (error) { - xfs_notice(mp, - "%s: inode 0x%llx format is incompatible for exchanging.", - __func__, ip->i_ino); - goto out_trans_cancel; - } +out_defer: + xfs_defer_cancel(&dfops); +out: + trace_xfs_swap_extent_rmap_error(ip, error, _RET_IP_); + tip->i_d.di_flags2 = tip_flags2; + return error; +} + +/* Swap the extents of two files by swapping data forks. */ +STATIC int +xfs_swap_extent_forks( + struct xfs_trans *tp, + struct xfs_inode *ip, + struct xfs_inode *tip, + int *src_log_flags, + int *target_log_flags) +{ + struct xfs_ifork tempifp, *ifp, *tifp; + int aforkblks = 0; + int taforkblks = 0; + xfs_extnum_t nextents; + __uint64_t tmp; + int error; - /* - * Compare the current change & modify times with that - * passed in. If they differ, we abort this swap. - * This is the mechanism used to ensure the calling - * process that the file was not changed out from - * under it. - */ - if ((sbp->bs_ctime.tv_sec != VFS_I(ip)->i_ctime.tv_sec) || - (sbp->bs_ctime.tv_nsec != VFS_I(ip)->i_ctime.tv_nsec) || - (sbp->bs_mtime.tv_sec != VFS_I(ip)->i_mtime.tv_sec) || - (sbp->bs_mtime.tv_nsec != VFS_I(ip)->i_mtime.tv_nsec)) { - error = -EBUSY; - goto out_trans_cancel; - } /* * Count the number of extended attribute blocks */ if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) && (ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) { - error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks); + error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, + &aforkblks); if (error) - goto out_trans_cancel; + return error; } if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) && (tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) { error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK, - &taforkblks); + &taforkblks); if (error) - goto out_trans_cancel; + return error; } /* @@ -1645,31 +1818,23 @@ xfs_swap_extents( * buffers, and so the validation done on read will expect the owner * field to be correctly set. Once we change the owners, we can swap the * inode forks. - * - * Note the trickiness in setting the log flags - we set the owner log - * flag on the opposite inode (i.e. the inode we are setting the new - * owner to be) because once we swap the forks and log that, log - * recovery is going to see the fork as owned by the swapped inode, - * not the pre-swapped inodes. */ - src_log_flags = XFS_ILOG_CORE; - target_log_flags = XFS_ILOG_CORE; if (ip->i_d.di_version == 3 && ip->i_d.di_format == XFS_DINODE_FMT_BTREE) { - target_log_flags |= XFS_ILOG_DOWNER; + (*target_log_flags) |= XFS_ILOG_DOWNER; error = xfs_bmbt_change_owner(tp, ip, XFS_DATA_FORK, tip->i_ino, NULL); if (error) - goto out_trans_cancel; + return error; } if (tip->i_d.di_version == 3 && tip->i_d.di_format == XFS_DINODE_FMT_BTREE) { - src_log_flags |= XFS_ILOG_DOWNER; + (*src_log_flags) |= XFS_ILOG_DOWNER; error = xfs_bmbt_change_owner(tp, tip, XFS_DATA_FORK, ip->i_ino, NULL); if (error) - goto out_trans_cancel; + return error; } /* @@ -1677,9 +1842,9 @@ xfs_swap_extents( */ ifp = &ip->i_df; tifp = &tip->i_df; - *tempifp = *ifp; /* struct copy */ + tempifp = *ifp; /* struct copy */ *ifp = *tifp; /* struct copy */ - *tifp = *tempifp; /* struct copy */ + *tifp = tempifp; /* struct copy */ /* * Fix the on-disk inode values @@ -1711,42 +1876,191 @@ xfs_swap_extents( switch (ip->i_d.di_format) { case XFS_DINODE_FMT_EXTENTS: - /* If the extents fit in the inode, fix the - * pointer. Otherwise it's already NULL or - * pointing to the extent. + /* + * If the extents fit in the inode, fix the pointer. Otherwise + * it's already NULL or pointing to the extent. */ - if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) { - ifp->if_u1.if_extents = - ifp->if_u2.if_inline_ext; - } - src_log_flags |= XFS_ILOG_DEXT; + nextents = xfs_iext_count(&ip->i_df); + if (nextents <= XFS_INLINE_EXTS) + ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext; + (*src_log_flags) |= XFS_ILOG_DEXT; break; case XFS_DINODE_FMT_BTREE: ASSERT(ip->i_d.di_version < 3 || - (src_log_flags & XFS_ILOG_DOWNER)); - src_log_flags |= XFS_ILOG_DBROOT; + (*src_log_flags & XFS_ILOG_DOWNER)); + (*src_log_flags) |= XFS_ILOG_DBROOT; break; } switch (tip->i_d.di_format) { case XFS_DINODE_FMT_EXTENTS: - /* If the extents fit in the inode, fix the - * pointer. Otherwise it's already NULL or - * pointing to the extent. + /* + * If the extents fit in the inode, fix the pointer. Otherwise + * it's already NULL or pointing to the extent. */ - if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) { - tifp->if_u1.if_extents = - tifp->if_u2.if_inline_ext; - } - target_log_flags |= XFS_ILOG_DEXT; + nextents = xfs_iext_count(&tip->i_df); + if (nextents <= XFS_INLINE_EXTS) + tifp->if_u1.if_extents = tifp->if_u2.if_inline_ext; + (*target_log_flags) |= XFS_ILOG_DEXT; break; case XFS_DINODE_FMT_BTREE: - target_log_flags |= XFS_ILOG_DBROOT; + (*target_log_flags) |= XFS_ILOG_DBROOT; ASSERT(tip->i_d.di_version < 3 || - (target_log_flags & XFS_ILOG_DOWNER)); + (*target_log_flags & XFS_ILOG_DOWNER)); break; } + return 0; +} + +int +xfs_swap_extents( + struct xfs_inode *ip, /* target inode */ + struct xfs_inode *tip, /* tmp inode */ + struct xfs_swapext *sxp) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + struct xfs_bstat *sbp = &sxp->sx_stat; + int src_log_flags, target_log_flags; + int error = 0; + int lock_flags; + struct xfs_ifork *cowfp; + __uint64_t f; + int resblks; + + /* + * Lock the inodes against other IO, page faults and truncate to + * begin with. Then we can ensure the inodes are flushed and have no + * page cache safely. Once we have done this we can take the ilocks and + * do the rest of the checks. + */ + lock_two_nondirectories(VFS_I(ip), VFS_I(tip)); + lock_flags = XFS_MMAPLOCK_EXCL; + xfs_lock_two_inodes(ip, tip, XFS_MMAPLOCK_EXCL); + + /* Verify that both files have the same format */ + if ((VFS_I(ip)->i_mode & S_IFMT) != (VFS_I(tip)->i_mode & S_IFMT)) { + error = -EINVAL; + goto out_unlock; + } + + /* Verify both files are either real-time or non-realtime */ + if (XFS_IS_REALTIME_INODE(ip) != XFS_IS_REALTIME_INODE(tip)) { + error = -EINVAL; + goto out_unlock; + } + + error = xfs_swap_extent_flush(ip); + if (error) + goto out_unlock; + error = xfs_swap_extent_flush(tip); + if (error) + goto out_unlock; + + /* + * Extent "swapping" with rmap requires a permanent reservation and + * a block reservation because it's really just a remap operation + * performed with log redo items! + */ + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { + /* + * Conceptually this shouldn't affect the shape of either + * bmbt, but since we atomically move extents one by one, + * we reserve enough space to rebuild both trees. + */ + resblks = XFS_SWAP_RMAP_SPACE_RES(mp, + XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK), + XFS_DATA_FORK) + + XFS_SWAP_RMAP_SPACE_RES(mp, + XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK), + XFS_DATA_FORK); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, + 0, 0, &tp); + } else + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, + 0, 0, &tp); + if (error) + goto out_unlock; + + /* + * Lock and join the inodes to the tansaction so that transaction commit + * or cancel will unlock the inodes from this point onwards. + */ + xfs_lock_two_inodes(ip, tip, XFS_ILOCK_EXCL); + lock_flags |= XFS_ILOCK_EXCL; + xfs_trans_ijoin(tp, ip, 0); + xfs_trans_ijoin(tp, tip, 0); + + + /* Verify all data are being swapped */ + if (sxp->sx_offset != 0 || + sxp->sx_length != ip->i_d.di_size || + sxp->sx_length != tip->i_d.di_size) { + error = -EFAULT; + goto out_trans_cancel; + } + + trace_xfs_swap_extent_before(ip, 0); + trace_xfs_swap_extent_before(tip, 1); + + /* check inode formats now that data is flushed */ + error = xfs_swap_extents_check_format(ip, tip); + if (error) { + xfs_notice(mp, + "%s: inode 0x%llx format is incompatible for exchanging.", + __func__, ip->i_ino); + goto out_trans_cancel; + } + + /* + * Compare the current change & modify times with that + * passed in. If they differ, we abort this swap. + * This is the mechanism used to ensure the calling + * process that the file was not changed out from + * under it. + */ + if ((sbp->bs_ctime.tv_sec != VFS_I(ip)->i_ctime.tv_sec) || + (sbp->bs_ctime.tv_nsec != VFS_I(ip)->i_ctime.tv_nsec) || + (sbp->bs_mtime.tv_sec != VFS_I(ip)->i_mtime.tv_sec) || + (sbp->bs_mtime.tv_nsec != VFS_I(ip)->i_mtime.tv_nsec)) { + error = -EBUSY; + goto out_trans_cancel; + } + + /* + * Note the trickiness in setting the log flags - we set the owner log + * flag on the opposite inode (i.e. the inode we are setting the new + * owner to be) because once we swap the forks and log that, log + * recovery is going to see the fork as owned by the swapped inode, + * not the pre-swapped inodes. + */ + src_log_flags = XFS_ILOG_CORE; + target_log_flags = XFS_ILOG_CORE; + + if (xfs_sb_version_hasrmapbt(&mp->m_sb)) + error = xfs_swap_extent_rmap(&tp, ip, tip); + else + error = xfs_swap_extent_forks(tp, ip, tip, &src_log_flags, + &target_log_flags); + if (error) + goto out_trans_cancel; + + /* Do we have to swap reflink flags? */ + if ((ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK) ^ + (tip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK)) { + f = ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK; + ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + ip->i_d.di_flags2 |= tip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK; + tip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + tip->i_d.di_flags2 |= f & XFS_DIFLAG2_REFLINK; + cowfp = ip->i_cowfp; + ip->i_cowfp = tip->i_cowfp; + tip->i_cowfp = cowfp; + xfs_inode_set_cowblocks_tag(ip); + xfs_inode_set_cowblocks_tag(tip); + } + xfs_trans_log_inode(tp, ip, src_log_flags); xfs_trans_log_inode(tp, tip, target_log_flags); @@ -1761,16 +2075,14 @@ xfs_swap_extents( trace_xfs_swap_extent_after(ip, 0); trace_xfs_swap_extent_after(tip, 1); -out: - kmem_free(tempifp); - return error; out_unlock: xfs_iunlock(ip, lock_flags); xfs_iunlock(tip, lock_flags); - goto out; + unlock_two_nondirectories(VFS_I(ip), VFS_I(tip)); + return error; out_trans_cancel: xfs_trans_cancel(tp); - goto out; + goto out_unlock; } diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index b5b9bffe3520..7f0a01f7b592 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -219,7 +219,6 @@ _xfs_buf_alloc( init_completion(&bp->b_iowait); INIT_LIST_HEAD(&bp->b_lru); INIT_LIST_HEAD(&bp->b_list); - RB_CLEAR_NODE(&bp->b_rbnode); sema_init(&bp->b_sema, 0); /* held, no waiters */ spin_lock_init(&bp->b_lock); XB_SET_OWNER(bp); @@ -473,6 +472,62 @@ _xfs_buf_map_pages( /* * Finding and Reading Buffers */ +static int +_xfs_buf_obj_cmp( + struct rhashtable_compare_arg *arg, + const void *obj) +{ + const struct xfs_buf_map *map = arg->key; + const struct xfs_buf *bp = obj; + + /* + * The key hashing in the lookup path depends on the key being the + * first element of the compare_arg, make sure to assert this. + */ + BUILD_BUG_ON(offsetof(struct xfs_buf_map, bm_bn) != 0); + + if (bp->b_bn != map->bm_bn) + return 1; + + if (unlikely(bp->b_length != map->bm_len)) { + /* + * found a block number match. If the range doesn't + * match, the only way this is allowed is if the buffer + * in the cache is stale and the transaction that made + * it stale has not yet committed. i.e. we are + * reallocating a busy extent. Skip this buffer and + * continue searching for an exact match. + */ + ASSERT(bp->b_flags & XBF_STALE); + return 1; + } + return 0; +} + +static const struct rhashtable_params xfs_buf_hash_params = { + .min_size = 32, /* empty AGs have minimal footprint */ + .nelem_hint = 16, + .key_len = sizeof(xfs_daddr_t), + .key_offset = offsetof(struct xfs_buf, b_bn), + .head_offset = offsetof(struct xfs_buf, b_rhash_head), + .automatic_shrinking = true, + .obj_cmpfn = _xfs_buf_obj_cmp, +}; + +int +xfs_buf_hash_init( + struct xfs_perag *pag) +{ + spin_lock_init(&pag->pag_buf_lock); + return rhashtable_init(&pag->pag_buf_hash, &xfs_buf_hash_params); +} + +void +xfs_buf_hash_destroy( + struct xfs_perag *pag) +{ + rhashtable_destroy(&pag->pag_buf_hash); +} /* * Look up, and creates if absent, a lockable buffer for @@ -488,27 +543,24 @@ _xfs_buf_find( xfs_buf_t *new_bp) { struct xfs_perag *pag; - struct rb_node **rbp; - struct rb_node *parent; xfs_buf_t *bp; - xfs_daddr_t blkno = map[0].bm_bn; + struct xfs_buf_map cmap = { .bm_bn = map[0].bm_bn }; xfs_daddr_t eofs; - int numblks = 0; int i; for (i = 0; i < nmaps; i++) - numblks += map[i].bm_len; + cmap.bm_len += map[i].bm_len; /* Check for IOs smaller than the sector size / not sector aligned */ - ASSERT(!(BBTOB(numblks) < btp->bt_meta_sectorsize)); - ASSERT(!(BBTOB(blkno) & (xfs_off_t)btp->bt_meta_sectormask)); + ASSERT(!(BBTOB(cmap.bm_len) < btp->bt_meta_sectorsize)); + ASSERT(!(BBTOB(cmap.bm_bn) & (xfs_off_t)btp->bt_meta_sectormask)); /* * Corrupted block numbers can get through to here, unfortunately, so we * have to check that the buffer falls within the filesystem bounds. */ eofs = XFS_FSB_TO_BB(btp->bt_mount, btp->bt_mount->m_sb.sb_dblocks); - if (blkno < 0 || blkno >= eofs) { + if (cmap.bm_bn < 0 || cmap.bm_bn >= eofs) { /* * XXX (dgc): we should really be returning -EFSCORRUPTED here, * but none of the higher level infrastructure supports @@ -516,53 +568,29 @@ _xfs_buf_find( */ xfs_alert(btp->bt_mount, "%s: Block out of range: block 0x%llx, EOFS 0x%llx ", - __func__, blkno, eofs); + __func__, cmap.bm_bn, eofs); WARN_ON(1); return NULL; } - /* get tree root */ pag = xfs_perag_get(btp->bt_mount, - xfs_daddr_to_agno(btp->bt_mount, blkno)); + xfs_daddr_to_agno(btp->bt_mount, cmap.bm_bn)); - /* walk tree */ spin_lock(&pag->pag_buf_lock); - rbp = &pag->pag_buf_tree.rb_node; - parent = NULL; - bp = NULL; - while (*rbp) { - parent = *rbp; - bp = rb_entry(parent, struct xfs_buf, b_rbnode); - - if (blkno < bp->b_bn) - rbp = &(*rbp)->rb_left; - else if (blkno > bp->b_bn) - rbp = &(*rbp)->rb_right; - else { - /* - * found a block number match. If the range doesn't - * match, the only way this is allowed is if the buffer - * in the cache is stale and the transaction that made - * it stale has not yet committed. i.e. we are - * reallocating a busy extent. Skip this buffer and - * continue searching to the right for an exact match. - */ - if (bp->b_length != numblks) { - ASSERT(bp->b_flags & XBF_STALE); - rbp = &(*rbp)->rb_right; - continue; - } - atomic_inc(&bp->b_hold); - goto found; - } + bp = rhashtable_lookup_fast(&pag->pag_buf_hash, &cmap, + xfs_buf_hash_params); + if (bp) { + atomic_inc(&bp->b_hold); + goto found; } /* No match found */ if (new_bp) { - rb_link_node(&new_bp->b_rbnode, parent, rbp); - rb_insert_color(&new_bp->b_rbnode, &pag->pag_buf_tree); /* the buffer keeps the perag reference until it is freed */ new_bp->b_pag = pag; + rhashtable_insert_fast(&pag->pag_buf_hash, + &new_bp->b_rhash_head, + xfs_buf_hash_params); spin_unlock(&pag->pag_buf_lock); } else { XFS_STATS_INC(btp->bt_mount, xb_miss_locked); @@ -930,7 +958,6 @@ xfs_buf_rele( if (!pag) { ASSERT(list_empty(&bp->b_lru)); - ASSERT(RB_EMPTY_NODE(&bp->b_rbnode)); if (atomic_dec_and_test(&bp->b_hold)) { xfs_buf_ioacct_dec(bp); xfs_buf_free(bp); @@ -938,8 +965,6 @@ xfs_buf_rele( return; } - ASSERT(!RB_EMPTY_NODE(&bp->b_rbnode)); - ASSERT(atomic_read(&bp->b_hold) > 0); release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); @@ -983,7 +1008,8 @@ xfs_buf_rele( } ASSERT(!(bp->b_flags & _XBF_DELWRI_Q)); - rb_erase(&bp->b_rbnode, &pag->pag_buf_tree); + rhashtable_remove_fast(&pag->pag_buf_hash, &bp->b_rhash_head, + xfs_buf_hash_params); spin_unlock(&pag->pag_buf_lock); xfs_perag_put(pag); freebuf = true; @@ -1304,7 +1330,7 @@ _xfs_buf_ioapply( if (bp->b_flags & XBF_WRITE) { op = REQ_OP_WRITE; if (bp->b_flags & XBF_SYNCIO) - op_flags = WRITE_SYNC; + op_flags = REQ_SYNC; if (bp->b_flags & XBF_FUA) op_flags |= REQ_FUA; if (bp->b_flags & XBF_FLUSH) @@ -1711,8 +1737,7 @@ xfs_free_buftarg( percpu_counter_destroy(&btp->bt_io_count); list_lru_destroy(&btp->bt_lru); - if (mp->m_flags & XFS_MOUNT_BARRIER) - xfs_blkdev_issue_flush(btp); + xfs_blkdev_issue_flush(btp); kmem_free(btp); } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 1c2e52b2d926..8a9d3a9599f0 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -71,6 +71,7 @@ typedef unsigned int xfs_buf_flags_t; { XBF_READ, "READ" }, \ { XBF_WRITE, "WRITE" }, \ { XBF_READ_AHEAD, "READ_AHEAD" }, \ + { XBF_NO_IOACCT, "NO_IOACCT" }, \ { XBF_ASYNC, "ASYNC" }, \ { XBF_DONE, "DONE" }, \ { XBF_STALE, "STALE" }, \ @@ -150,7 +151,7 @@ typedef struct xfs_buf { * which is the only bit that is touched if we hit the semaphore * fast-path on locking. */ - struct rb_node b_rbnode; /* rbtree node */ + struct rhash_head b_rhash_head; /* pag buffer hash node */ xfs_daddr_t b_bn; /* block number of buffer */ int b_length; /* size of buffer in BBs */ atomic_t b_hold; /* reference count */ diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index e455f9098d49..2975cb2319f4 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -865,7 +865,7 @@ xfs_buf_item_log_segment( */ if (bit) { end_bit = MIN(bit + bits_to_set, (uint)NBWORD); - mask = ((1 << (end_bit - bit)) - 1) << bit; + mask = ((1U << (end_bit - bit)) - 1) << bit; *wordp |= mask; wordp++; bits_set = end_bit - bit; @@ -888,7 +888,7 @@ xfs_buf_item_log_segment( */ end_bit = bits_to_set - bits_set; if (end_bit) { - mask = (1 << end_bit) - 1; + mask = (1U << end_bit) - 1; *wordp |= mask; } } @@ -1095,7 +1095,8 @@ xfs_buf_iodone_callback_error( bp->b_last_error != bp->b_error) { bp->b_flags |= (XBF_WRITE | XBF_DONE | XBF_WRITE_FAIL); bp->b_last_error = bp->b_error; - if (cfg->retry_timeout && !bp->b_first_retry_time) + if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER && + !bp->b_first_retry_time) bp->b_first_retry_time = jiffies; xfs_buf_ioerror(bp, 0); @@ -1111,7 +1112,7 @@ xfs_buf_iodone_callback_error( if (cfg->max_retries != XFS_ERR_RETRY_FOREVER && ++bp->b_retries > cfg->max_retries) goto permanent_error; - if (cfg->retry_timeout && + if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER && time_after(jiffies, cfg->retry_timeout + bp->b_first_retry_time)) goto permanent_error; diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c index f44f79996978..003a99b83bd8 100644 --- a/fs/xfs/xfs_dir2_readdir.c +++ b/fs/xfs/xfs_dir2_readdir.c @@ -84,7 +84,8 @@ xfs_dir2_sf_getdents( sfp = (xfs_dir2_sf_hdr_t *)dp->i_df.if_u1.if_data; - ASSERT(dp->i_d.di_size >= xfs_dir2_sf_hdr_size(sfp->i8count)); + if (dp->i_d.di_size < xfs_dir2_sf_hdr_size(sfp->i8count)) + return -EFSCORRUPTED; /* * If the block number in the offset is out of range, we're done. @@ -676,7 +677,6 @@ xfs_readdir( args.dp = dp; args.geo = dp->i_mount->m_dir_geo; - xfs_ilock(dp, XFS_IOLOCK_SHARED); if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) rval = xfs_dir2_sf_getdents(&args, ctx); else if ((rval = xfs_dir2_isblock(&args, &v))) @@ -685,7 +685,6 @@ xfs_readdir( rval = xfs_dir2_block_getdents(&args, ctx); else rval = xfs_dir2_leaf_getdents(&args, ctx, bufsize); - xfs_iunlock(dp, XFS_IOLOCK_SHARED); return rval; } diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h index 3d224702fbc0..05f8666733a0 100644 --- a/fs/xfs/xfs_error.h +++ b/fs/xfs/xfs_error.h @@ -92,7 +92,11 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_ERRTAG_BMAPIFORMAT 21 #define XFS_ERRTAG_FREE_EXTENT 22 #define XFS_ERRTAG_RMAP_FINISH_ONE 23 -#define XFS_ERRTAG_MAX 24 +#define XFS_ERRTAG_REFCOUNT_CONTINUE_UPDATE 24 +#define XFS_ERRTAG_REFCOUNT_FINISH_ONE 25 +#define XFS_ERRTAG_BMAP_FINISH_ONE 26 +#define XFS_ERRTAG_AG_RESV_CRITICAL 27 +#define XFS_ERRTAG_MAX 28 /* * Random factors for above tags, 1 means always, 2 means 1/2 time, etc. @@ -121,6 +125,10 @@ extern void xfs_verifier_error(struct xfs_buf *bp); #define XFS_RANDOM_BMAPIFORMAT XFS_RANDOM_DEFAULT #define XFS_RANDOM_FREE_EXTENT 1 #define XFS_RANDOM_RMAP_FINISH_ONE 1 +#define XFS_RANDOM_REFCOUNT_CONTINUE_UPDATE 1 +#define XFS_RANDOM_REFCOUNT_FINISH_ONE 1 +#define XFS_RANDOM_BMAP_FINISH_ONE 1 +#define XFS_RANDOM_AG_RESV_CRITICAL 4 #ifdef DEBUG extern int xfs_error_test_active; diff --git a/fs/xfs/xfs_extent_busy.c b/fs/xfs/xfs_extent_busy.c index c263e079273e..162dc186cf04 100644 --- a/fs/xfs/xfs_extent_busy.c +++ b/fs/xfs/xfs_extent_busy.c @@ -384,7 +384,7 @@ restart: * If this is a metadata allocation, try to reuse the busy * extent instead of trimming the allocation. */ - if (!args->userdata && + if (!xfs_alloc_is_userdata(args->datatype) && !(busyp->flags & XFS_EXTENT_BUSY_DISCARDED)) { if (!xfs_extent_busy_update_extent(args->mp, args->pag, busyp, fbno, flen, diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index e612a0233710..bbb9eb6811b2 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -38,6 +38,7 @@ #include "xfs_icache.h" #include "xfs_pnfs.h" #include "xfs_iomap.h" +#include "xfs_reflink.h" #include <linux/dcache.h> #include <linux/falloc.h> @@ -47,40 +48,6 @@ static const struct vm_operations_struct xfs_file_vm_ops; /* - * Locking primitives for read and write IO paths to ensure we consistently use - * and order the inode->i_mutex, ip->i_lock and ip->i_iolock. - */ -static inline void -xfs_rw_ilock( - struct xfs_inode *ip, - int type) -{ - if (type & XFS_IOLOCK_EXCL) - inode_lock(VFS_I(ip)); - xfs_ilock(ip, type); -} - -static inline void -xfs_rw_iunlock( - struct xfs_inode *ip, - int type) -{ - xfs_iunlock(ip, type); - if (type & XFS_IOLOCK_EXCL) - inode_unlock(VFS_I(ip)); -} - -static inline void -xfs_rw_ilock_demote( - struct xfs_inode *ip, - int type) -{ - xfs_ilock_demote(ip, type); - if (type & XFS_IOLOCK_EXCL) - inode_unlock(VFS_I(ip)); -} - -/* * Clear the specified ranges to zero through either the pagecache or DAX. * Holes and unwritten extents will be left as-is as they already are zeroed. */ @@ -182,19 +149,16 @@ xfs_file_fsync( xfs_iflags_clear(ip, XFS_ITRUNCATED); - if (mp->m_flags & XFS_MOUNT_BARRIER) { - /* - * If we have an RT and/or log subvolume we need to make sure - * to flush the write cache the device used for file data - * first. This is to ensure newly written file data make - * it to disk before logging the new inode size in case of - * an extending write. - */ - if (XFS_IS_REALTIME_INODE(ip)) - xfs_blkdev_issue_flush(mp->m_rtdev_targp); - else if (mp->m_logdev_targp != mp->m_ddev_targp) - xfs_blkdev_issue_flush(mp->m_ddev_targp); - } + /* + * If we have an RT and/or log subvolume we need to make sure to flush + * the write cache the device used for file data first. This is to + * ensure newly written file data make it to disk before logging the new + * inode size in case of an extending write. + */ + if (XFS_IS_REALTIME_INODE(ip)) + xfs_blkdev_issue_flush(mp->m_rtdev_targp); + else if (mp->m_logdev_targp != mp->m_ddev_targp) + xfs_blkdev_issue_flush(mp->m_ddev_targp); /* * All metadata updates are logged, which means that we just have to @@ -229,10 +193,8 @@ xfs_file_fsync( * an already allocated file and thus do not have any metadata to * commit. */ - if ((mp->m_flags & XFS_MOUNT_BARRIER) && - mp->m_logdev_targp == mp->m_ddev_targp && - !XFS_IS_REALTIME_INODE(ip) && - !log_flushed) + if (!log_flushed && !XFS_IS_REALTIME_INODE(ip) && + mp->m_logdev_targp == mp->m_ddev_targp) xfs_blkdev_issue_flush(mp->m_ddev_targp); return error; @@ -243,87 +205,21 @@ xfs_file_dio_aio_read( struct kiocb *iocb, struct iov_iter *to) { - struct address_space *mapping = iocb->ki_filp->f_mapping; - struct inode *inode = mapping->host; - struct xfs_inode *ip = XFS_I(inode); - loff_t isize = i_size_read(inode); + struct xfs_inode *ip = XFS_I(file_inode(iocb->ki_filp)); size_t count = iov_iter_count(to); - struct iov_iter data; - struct xfs_buftarg *target; - ssize_t ret = 0; + ssize_t ret; trace_xfs_file_direct_read(ip, count, iocb->ki_pos); if (!count) return 0; /* skip atime */ - if (XFS_IS_REALTIME_INODE(ip)) - target = ip->i_mount->m_rtdev_targp; - else - target = ip->i_mount->m_ddev_targp; - - /* DIO must be aligned to device logical sector size */ - if ((iocb->ki_pos | count) & target->bt_logical_sectormask) { - if (iocb->ki_pos == isize) - return 0; - return -EINVAL; - } - - /* - * Locking is a bit tricky here. If we take an exclusive lock for direct - * IO, we effectively serialise all new concurrent read IO to this file - * and block it behind IO that is currently in progress because IO in - * progress holds the IO lock shared. We only need to hold the lock - * exclusive to blow away the page cache, so only take lock exclusively - * if the page cache needs invalidation. This allows the normal direct - * IO case of no page cache pages to proceeed concurrently without - * serialisation. - */ - xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); - if (mapping->nrpages) { - xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); - xfs_rw_ilock(ip, XFS_IOLOCK_EXCL); - - /* - * The generic dio code only flushes the range of the particular - * I/O. Because we take an exclusive lock here, this whole - * sequence is considerably more expensive for us. This has a - * noticeable performance impact for any file with cached pages, - * even when outside of the range of the particular I/O. - * - * Hence, amortize the cost of the lock against a full file - * flush and reduce the chances of repeated iolock cycles going - * forward. - */ - if (mapping->nrpages) { - ret = filemap_write_and_wait(mapping); - if (ret) { - xfs_rw_iunlock(ip, XFS_IOLOCK_EXCL); - return ret; - } + file_accessed(iocb->ki_filp); - /* - * Invalidate whole pages. This can return an error if - * we fail to invalidate a page, but this should never - * happen on XFS. Warn if it does fail. - */ - ret = invalidate_inode_pages2(mapping); - WARN_ON_ONCE(ret); - ret = 0; - } - xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL); - } + xfs_ilock(ip, XFS_IOLOCK_SHARED); + ret = iomap_dio_rw(iocb, to, &xfs_iomap_ops, NULL); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); - data = *to; - ret = __blockdev_direct_IO(iocb, inode, target->bt_bdev, &data, - xfs_get_blocks_direct, NULL, NULL, 0); - if (ret > 0) { - iocb->ki_pos += ret; - iov_iter_advance(to, ret); - } - xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); - - file_accessed(iocb->ki_filp); return ret; } @@ -332,10 +228,7 @@ xfs_file_dax_read( struct kiocb *iocb, struct iov_iter *to) { - struct address_space *mapping = iocb->ki_filp->f_mapping; - struct inode *inode = mapping->host; - struct xfs_inode *ip = XFS_I(inode); - struct iov_iter data = *to; + struct xfs_inode *ip = XFS_I(iocb->ki_filp->f_mapping->host); size_t count = iov_iter_count(to); ssize_t ret = 0; @@ -344,13 +237,9 @@ xfs_file_dax_read( if (!count) return 0; /* skip atime */ - xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); - ret = dax_do_io(iocb, inode, &data, xfs_get_blocks_direct, NULL, 0); - if (ret > 0) { - iocb->ki_pos += ret; - iov_iter_advance(to, ret); - } - xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); + xfs_ilock(ip, XFS_IOLOCK_SHARED); + ret = dax_iomap_rw(iocb, to, &xfs_iomap_ops); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); file_accessed(iocb->ki_filp); return ret; @@ -366,9 +255,9 @@ xfs_file_buffered_aio_read( trace_xfs_file_buffered_read(ip, iov_iter_count(to), iocb->ki_pos); - xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); + xfs_ilock(ip, XFS_IOLOCK_SHARED); ret = generic_file_read_iter(iocb, to); - xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); return ret; } @@ -399,45 +288,6 @@ xfs_file_read_iter( return ret; } -STATIC ssize_t -xfs_file_splice_read( - struct file *infilp, - loff_t *ppos, - struct pipe_inode_info *pipe, - size_t count, - unsigned int flags) -{ - struct xfs_inode *ip = XFS_I(infilp->f_mapping->host); - ssize_t ret; - - XFS_STATS_INC(ip->i_mount, xs_read_calls); - - if (XFS_FORCED_SHUTDOWN(ip->i_mount)) - return -EIO; - - trace_xfs_file_splice_read(ip, count, *ppos); - - /* - * DAX inodes cannot ues the page cache for splice, so we have to push - * them through the VFS IO path. This means it goes through - * ->read_iter, which for us takes the XFS_IOLOCK_SHARED. Hence we - * cannot lock the splice operation at this level for DAX inodes. - */ - if (IS_DAX(VFS_I(ip))) { - ret = default_file_splice_read(infilp, ppos, pipe, count, - flags); - goto out; - } - - xfs_rw_ilock(ip, XFS_IOLOCK_SHARED); - ret = generic_file_splice_read(infilp, ppos, pipe, count, flags); - xfs_rw_iunlock(ip, XFS_IOLOCK_SHARED); -out: - if (ret > 0) - XFS_STATS_ADD(ip->i_mount, xs_read_bytes, ret); - return ret; -} - /* * Zero any on disk space between the current EOF and the new, larger EOF. * @@ -488,15 +338,18 @@ restart: if (error <= 0) return error; - error = xfs_break_layouts(inode, iolock, true); + error = xfs_break_layouts(inode, iolock); if (error) return error; - /* For changing security info in file_remove_privs() we need i_mutex */ + /* + * For changing security info in file_remove_privs() we need i_rwsem + * exclusively. + */ if (*iolock == XFS_IOLOCK_SHARED && !IS_NOSEC(inode)) { - xfs_rw_iunlock(ip, *iolock); + xfs_iunlock(ip, *iolock); *iolock = XFS_IOLOCK_EXCL; - xfs_rw_ilock(ip, *iolock); + xfs_ilock(ip, *iolock); goto restart; } /* @@ -521,9 +374,9 @@ restart: spin_unlock(&ip->i_flags_lock); if (!drained_dio) { if (*iolock == XFS_IOLOCK_SHARED) { - xfs_rw_iunlock(ip, *iolock); + xfs_iunlock(ip, *iolock); *iolock = XFS_IOLOCK_EXCL; - xfs_rw_ilock(ip, *iolock); + xfs_ilock(ip, *iolock); iov_iter_reexpand(from, count); } /* @@ -566,6 +419,58 @@ restart: return 0; } +static int +xfs_dio_write_end_io( + struct kiocb *iocb, + ssize_t size, + unsigned flags) +{ + struct inode *inode = file_inode(iocb->ki_filp); + struct xfs_inode *ip = XFS_I(inode); + loff_t offset = iocb->ki_pos; + bool update_size = false; + int error = 0; + + trace_xfs_end_io_direct_write(ip, offset, size); + + if (XFS_FORCED_SHUTDOWN(ip->i_mount)) + return -EIO; + + if (size <= 0) + return size; + + /* + * We need to update the in-core inode size here so that we don't end up + * with the on-disk inode size being outside the in-core inode size. We + * have no other method of updating EOF for AIO, so always do it here + * if necessary. + * + * We need to lock the test/set EOF update as we can be racing with + * other IO completions here to update the EOF. Failing to serialise + * here can result in EOF moving backwards and Bad Things Happen when + * that occurs. + */ + spin_lock(&ip->i_flags_lock); + if (offset + size > i_size_read(inode)) { + i_size_write(inode, offset + size); + update_size = true; + } + spin_unlock(&ip->i_flags_lock); + + if (flags & IOMAP_DIO_COW) { + error = xfs_reflink_end_cow(ip, offset, size); + if (error) + return error; + } + + if (flags & IOMAP_DIO_UNWRITTEN) + error = xfs_iomap_write_unwritten(ip, offset, size); + else if (update_size) + error = xfs_setfilesize(ip, offset, size); + + return error; +} + /* * xfs_file_dio_aio_write - handle direct IO writes * @@ -605,98 +510,59 @@ xfs_file_dio_aio_write( int unaligned_io = 0; int iolock; size_t count = iov_iter_count(from); - loff_t end; - struct iov_iter data; - struct xfs_buftarg *target = XFS_IS_REALTIME_INODE(ip) ? + struct xfs_buftarg *target = XFS_IS_REALTIME_INODE(ip) ? mp->m_rtdev_targp : mp->m_ddev_targp; /* DIO must be aligned to device logical sector size */ if ((iocb->ki_pos | count) & target->bt_logical_sectormask) return -EINVAL; - /* "unaligned" here means not aligned to a filesystem block */ - if ((iocb->ki_pos & mp->m_blockmask) || - ((iocb->ki_pos + count) & mp->m_blockmask)) - unaligned_io = 1; - /* - * We don't need to take an exclusive lock unless there page cache needs - * to be invalidated or unaligned IO is being executed. We don't need to - * consider the EOF extension case here because - * xfs_file_aio_write_checks() will relock the inode as necessary for - * EOF zeroing cases and fill out the new inode size as appropriate. + * Don't take the exclusive iolock here unless the I/O is unaligned to + * the file system block size. We don't need to consider the EOF + * extension case here because xfs_file_aio_write_checks() will relock + * the inode as necessary for EOF zeroing cases and fill out the new + * inode size as appropriate. */ - if (unaligned_io || mapping->nrpages) + if ((iocb->ki_pos & mp->m_blockmask) || + ((iocb->ki_pos + count) & mp->m_blockmask)) { + unaligned_io = 1; iolock = XFS_IOLOCK_EXCL; - else + } else { iolock = XFS_IOLOCK_SHARED; - xfs_rw_ilock(ip, iolock); - - /* - * Recheck if there are cached pages that need invalidate after we got - * the iolock to protect against other threads adding new pages while - * we were waiting for the iolock. - */ - if (mapping->nrpages && iolock == XFS_IOLOCK_SHARED) { - xfs_rw_iunlock(ip, iolock); - iolock = XFS_IOLOCK_EXCL; - xfs_rw_ilock(ip, iolock); } + xfs_ilock(ip, iolock); + ret = xfs_file_aio_write_checks(iocb, from, &iolock); if (ret) goto out; count = iov_iter_count(from); - end = iocb->ki_pos + count - 1; - - /* - * See xfs_file_dio_aio_read() for why we do a full-file flush here. - */ - if (mapping->nrpages) { - ret = filemap_write_and_wait(VFS_I(ip)->i_mapping); - if (ret) - goto out; - /* - * Invalidate whole pages. This can return an error if we fail - * to invalidate a page, but this should never happen on XFS. - * Warn if it does fail. - */ - ret = invalidate_inode_pages2(VFS_I(ip)->i_mapping); - WARN_ON_ONCE(ret); - ret = 0; - } /* * If we are doing unaligned IO, wait for all other IO to drain, - * otherwise demote the lock if we had to flush cached pages + * otherwise demote the lock if we had to take the exclusive lock + * for other reasons in xfs_file_aio_write_checks. */ if (unaligned_io) inode_dio_wait(inode); else if (iolock == XFS_IOLOCK_EXCL) { - xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL); + xfs_ilock_demote(ip, XFS_IOLOCK_EXCL); iolock = XFS_IOLOCK_SHARED; } trace_xfs_file_direct_write(ip, count, iocb->ki_pos); - data = *from; - ret = __blockdev_direct_IO(iocb, inode, target->bt_bdev, &data, - xfs_get_blocks_direct, xfs_end_io_direct_write, - NULL, DIO_ASYNC_EXTEND); - - /* see generic_file_direct_write() for why this is necessary */ - if (mapping->nrpages) { - invalidate_inode_pages2_range(mapping, - iocb->ki_pos >> PAGE_SHIFT, - end >> PAGE_SHIFT); + /* If this is a block-aligned directio CoW, remap immediately. */ + if (xfs_is_reflink_inode(ip) && !unaligned_io) { + ret = xfs_reflink_allocate_cow_range(ip, iocb->ki_pos, count); + if (ret) + goto out; } - if (ret > 0) { - iocb->ki_pos += ret; - iov_iter_advance(from, ret); - } + ret = iomap_dio_rw(iocb, from, &xfs_iomap_ops, xfs_dio_write_end_io); out: - xfs_rw_iunlock(ip, iolock); + xfs_iunlock(ip, iolock); /* * No fallback to buffered IO on errors for XFS, direct IO will either @@ -711,70 +577,30 @@ xfs_file_dax_write( struct kiocb *iocb, struct iov_iter *from) { - struct address_space *mapping = iocb->ki_filp->f_mapping; - struct inode *inode = mapping->host; + struct inode *inode = iocb->ki_filp->f_mapping->host; struct xfs_inode *ip = XFS_I(inode); - struct xfs_mount *mp = ip->i_mount; - ssize_t ret = 0; - int unaligned_io = 0; - int iolock; - struct iov_iter data; - - /* "unaligned" here means not aligned to a filesystem block */ - if ((iocb->ki_pos & mp->m_blockmask) || - ((iocb->ki_pos + iov_iter_count(from)) & mp->m_blockmask)) { - unaligned_io = 1; - iolock = XFS_IOLOCK_EXCL; - } else if (mapping->nrpages) { - iolock = XFS_IOLOCK_EXCL; - } else { - iolock = XFS_IOLOCK_SHARED; - } - xfs_rw_ilock(ip, iolock); + int iolock = XFS_IOLOCK_EXCL; + ssize_t ret, error = 0; + size_t count; + loff_t pos; + xfs_ilock(ip, iolock); ret = xfs_file_aio_write_checks(iocb, from, &iolock); if (ret) goto out; - /* - * Yes, even DAX files can have page cache attached to them: A zeroed - * page is inserted into the pagecache when we have to serve a write - * fault on a hole. It should never be dirtied and can simply be - * dropped from the pagecache once we get real data for the page. - * - * XXX: This is racy against mmap, and there's nothing we can do about - * it. dax_do_io() should really do this invalidation internally as - * it will know if we've allocated over a holei for this specific IO and - * if so it needs to update the mapping tree and invalidate existing - * PTEs over the newly allocated range. Remove this invalidation when - * dax_do_io() is fixed up. - */ - if (mapping->nrpages) { - loff_t end = iocb->ki_pos + iov_iter_count(from) - 1; - - ret = invalidate_inode_pages2_range(mapping, - iocb->ki_pos >> PAGE_SHIFT, - end >> PAGE_SHIFT); - WARN_ON_ONCE(ret); - } - - if (iolock == XFS_IOLOCK_EXCL && !unaligned_io) { - xfs_rw_ilock_demote(ip, XFS_IOLOCK_EXCL); - iolock = XFS_IOLOCK_SHARED; - } - - trace_xfs_file_dax_write(ip, iov_iter_count(from), iocb->ki_pos); + pos = iocb->ki_pos; + count = iov_iter_count(from); - data = *from; - ret = dax_do_io(iocb, inode, &data, xfs_get_blocks_direct, - xfs_end_io_direct_write, 0); - if (ret > 0) { - iocb->ki_pos += ret; - iov_iter_advance(from, ret); + trace_xfs_file_dax_write(ip, count, pos); + ret = dax_iomap_rw(iocb, from, &xfs_iomap_ops); + if (ret > 0 && iocb->ki_pos > i_size_read(inode)) { + i_size_write(inode, iocb->ki_pos); + error = xfs_setfilesize(ip, pos, ret); } out: - xfs_rw_iunlock(ip, iolock); - return ret; + xfs_iunlock(ip, iolock); + return error ? error : ret; } STATIC ssize_t @@ -790,7 +616,7 @@ xfs_file_buffered_aio_write( int enospc = 0; int iolock = XFS_IOLOCK_EXCL; - xfs_rw_ilock(ip, iolock); + xfs_ilock(ip, iolock); ret = xfs_file_aio_write_checks(iocb, from, &iolock); if (ret) @@ -818,6 +644,9 @@ write_retry: enospc = xfs_inode_free_quota_eofblocks(ip); if (enospc) goto write_retry; + enospc = xfs_inode_free_quota_cowblocks(ip); + if (enospc) + goto write_retry; } else if (ret == -ENOSPC && !enospc) { struct xfs_eofblocks eofb = {0}; @@ -831,7 +660,7 @@ write_retry: current->backing_dev_info = NULL; out: - xfs_rw_iunlock(ip, iolock); + xfs_iunlock(ip, iolock); return ret; } @@ -857,10 +686,20 @@ xfs_file_write_iter( if (IS_DAX(inode)) ret = xfs_file_dax_write(iocb, from); - else if (iocb->ki_flags & IOCB_DIRECT) + else if (iocb->ki_flags & IOCB_DIRECT) { + /* + * Allow a directio write to fall back to a buffered + * write *only* in the case that we're doing a reflink + * CoW. In all other directio scenarios we do not + * allow an operation to fall back to buffered mode. + */ ret = xfs_file_dio_aio_write(iocb, from); - else + if (ret == -EREMCHG) + goto buffered; + } else { +buffered: ret = xfs_file_buffered_aio_write(iocb, from); + } if (ret > 0) { XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret); @@ -874,7 +713,7 @@ xfs_file_write_iter( #define XFS_FALLOC_FL_SUPPORTED \ (FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | \ FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE | \ - FALLOC_FL_INSERT_RANGE) + FALLOC_FL_INSERT_RANGE | FALLOC_FL_UNSHARE_RANGE) STATIC long xfs_file_fallocate( @@ -897,7 +736,7 @@ xfs_file_fallocate( return -EOPNOTSUPP; xfs_ilock(ip, iolock); - error = xfs_break_layouts(inode, &iolock, false); + error = xfs_break_layouts(inode, &iolock); if (error) goto out_unlock; @@ -964,9 +803,15 @@ xfs_file_fallocate( if (mode & FALLOC_FL_ZERO_RANGE) error = xfs_zero_file_space(ip, offset, len); - else + else { + if (mode & FALLOC_FL_UNSHARE_RANGE) { + error = xfs_reflink_unshare(ip, offset, len); + if (error) + goto out_unlock; + } error = xfs_alloc_file_space(ip, offset, len, XFS_BMAPI_PREALLOC); + } if (error) goto out_unlock; } @@ -984,7 +829,7 @@ xfs_file_fallocate( iattr.ia_valid = ATTR_SIZE; iattr.ia_size = new_size; - error = xfs_setattr_size(ip, &iattr); + error = xfs_vn_setattr_size(file_dentry(file), &iattr); if (error) goto out_unlock; } @@ -1003,6 +848,34 @@ out_unlock: return error; } +STATIC int +xfs_file_clone_range( + struct file *file_in, + loff_t pos_in, + struct file *file_out, + loff_t pos_out, + u64 len) +{ + return xfs_reflink_remap_range(file_in, pos_in, file_out, pos_out, + len, false); +} + +STATIC ssize_t +xfs_file_dedupe_range( + struct file *src_file, + u64 loff, + u64 len, + struct file *dst_file, + u64 dst_loff) +{ + int error; + + error = xfs_reflink_remap_range(src_file, loff, dst_file, dst_loff, + len, true); + if (error) + return error; + return len; +} STATIC int xfs_file_open( @@ -1513,7 +1386,7 @@ xfs_filemap_page_mkwrite( xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED); if (IS_DAX(inode)) { - ret = dax_mkwrite(vma, vmf, xfs_get_blocks_dax_fault); + ret = dax_iomap_fault(vma, vmf, &xfs_iomap_ops); } else { ret = iomap_page_mkwrite(vma, vmf, &xfs_iomap_ops); ret = block_page_mkwrite_return(ret); @@ -1540,15 +1413,9 @@ xfs_filemap_fault( return xfs_filemap_page_mkwrite(vma, vmf); xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED); - if (IS_DAX(inode)) { - /* - * we do not want to trigger unwritten extent conversion on read - * faults - that is unnecessary overhead and would also require - * changes to xfs_get_blocks_direct() to map unwritten extent - * ioend for conversion on read-only mappings. - */ - ret = dax_fault(vma, vmf, xfs_get_blocks_dax_fault); - } else + if (IS_DAX(inode)) + ret = dax_iomap_fault(vma, vmf, &xfs_iomap_ops); + else ret = filemap_fault(vma, vmf); xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED); @@ -1584,7 +1451,7 @@ xfs_filemap_pmd_fault( } xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED); - ret = dax_pmd_fault(vma, addr, pmd, flags, xfs_get_blocks_dax_fault); + ret = dax_iomap_pmd_fault(vma, addr, pmd, flags, &xfs_iomap_ops); xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED); if (flags & FAULT_FLAG_WRITE) @@ -1652,7 +1519,7 @@ const struct file_operations xfs_file_operations = { .llseek = xfs_file_llseek, .read_iter = xfs_file_read_iter, .write_iter = xfs_file_write_iter, - .splice_read = xfs_file_splice_read, + .splice_read = generic_file_splice_read, .splice_write = iter_file_splice_write, .unlocked_ioctl = xfs_file_ioctl, #ifdef CONFIG_COMPAT @@ -1662,7 +1529,10 @@ const struct file_operations xfs_file_operations = { .open = xfs_file_open, .release = xfs_file_release, .fsync = xfs_file_fsync, + .get_unmapped_area = thp_get_unmapped_area, .fallocate = xfs_file_fallocate, + .clone_file_range = xfs_file_clone_range, + .dedupe_file_range = xfs_file_dedupe_range, }; const struct file_operations xfs_dir_file_operations = { diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 4a33a3304369..043ca3808ea2 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -30,6 +30,7 @@ #include "xfs_mru_cache.h" #include "xfs_filestream.h" #include "xfs_trace.h" +#include "xfs_ag_resv.h" struct xfs_fstrm_item { struct xfs_mru_cache_elem mru; @@ -198,7 +199,8 @@ xfs_filestream_pick_ag( } longest = xfs_alloc_longest_free_extent(mp, pag, - xfs_alloc_min_freelist(mp, pag)); + xfs_alloc_min_freelist(mp, pag), + xfs_ag_resv_needed(pag, XFS_AG_RESV_NONE)); if (((minlen && longest >= minlen) || (!minlen && pag->pagf_freeblks >= minfree)) && (!pag->pagf_metadata || !(flags & XFS_PICK_USERDATA) || @@ -369,7 +371,8 @@ xfs_filestream_new_ag( struct xfs_mount *mp = ip->i_mount; xfs_extlen_t minlen = ap->length; xfs_agnumber_t startag = 0; - int flags, err = 0; + int flags = 0; + int err = 0; struct xfs_mru_cache_elem *mru; *agp = NULLAGNUMBER; @@ -385,8 +388,10 @@ xfs_filestream_new_ag( startag = (item->ag + 1) % mp->m_sb.sb_agcount; } - flags = (ap->userdata ? XFS_PICK_USERDATA : 0) | - (ap->dfops->dop_low ? XFS_PICK_LOWSPACE : 0); + if (xfs_alloc_is_userdata(ap->datatype)) + flags |= XFS_PICK_USERDATA; + if (ap->dfops->dop_low) + flags |= XFS_PICK_LOWSPACE; err = xfs_filestream_pick_ag(pip, startag, agp, flags, minlen); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index 0b7f986745c1..242e8091296d 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -43,6 +43,7 @@ #include "xfs_log.h" #include "xfs_filestream.h" #include "xfs_rmap.h" +#include "xfs_ag_resv.h" /* * File system operations @@ -108,7 +109,9 @@ xfs_fs_geometry( (xfs_sb_version_hassparseinodes(&mp->m_sb) ? XFS_FSOP_GEOM_FLAGS_SPINODES : 0) | (xfs_sb_version_hasrmapbt(&mp->m_sb) ? - XFS_FSOP_GEOM_FLAGS_RMAPBT : 0); + XFS_FSOP_GEOM_FLAGS_RMAPBT : 0) | + (xfs_sb_version_hasreflink(&mp->m_sb) ? + XFS_FSOP_GEOM_FLAGS_REFLINK : 0); geo->logsectsize = xfs_sb_version_hassector(&mp->m_sb) ? mp->m_sb.sb_logsectsize : BBSIZE; geo->rtsectsize = mp->m_sb.sb_blocksize; @@ -259,6 +262,12 @@ xfs_growfs_data_private( agf->agf_longest = cpu_to_be32(tmpsize); if (xfs_sb_version_hascrc(&mp->m_sb)) uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid); + if (xfs_sb_version_hasreflink(&mp->m_sb)) { + agf->agf_refcount_root = cpu_to_be32( + xfs_refc_block(mp)); + agf->agf_refcount_level = cpu_to_be32(1); + agf->agf_refcount_blocks = cpu_to_be32(1); + } error = xfs_bwrite(bp); xfs_buf_relse(bp); @@ -450,6 +459,17 @@ xfs_growfs_data_private( rrec->rm_offset = 0; be16_add_cpu(&block->bb_numrecs, 1); + /* account for refc btree root */ + if (xfs_sb_version_hasreflink(&mp->m_sb)) { + rrec = XFS_RMAP_REC_ADDR(block, 5); + rrec->rm_startblock = cpu_to_be32( + xfs_refc_block(mp)); + rrec->rm_blockcount = cpu_to_be32(1); + rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_REFC); + rrec->rm_offset = 0; + be16_add_cpu(&block->bb_numrecs, 1); + } + error = xfs_bwrite(bp); xfs_buf_relse(bp); if (error) @@ -507,6 +527,28 @@ xfs_growfs_data_private( goto error0; } + /* + * refcount btree root block + */ + if (xfs_sb_version_hasreflink(&mp->m_sb)) { + bp = xfs_growfs_get_hdr_buf(mp, + XFS_AGB_TO_DADDR(mp, agno, xfs_refc_block(mp)), + BTOBB(mp->m_sb.sb_blocksize), 0, + &xfs_refcountbt_buf_ops); + if (!bp) { + error = -ENOMEM; + goto error0; + } + + xfs_btree_init_block(mp, bp, XFS_REFC_CRC_MAGIC, + 0, 0, agno, + XFS_BTREE_CRC_BLOCKS); + + error = xfs_bwrite(bp); + xfs_buf_relse(bp); + if (error) + goto error0; + } } xfs_trans_agblocks_delta(tp, nfree); /* @@ -553,7 +595,7 @@ xfs_growfs_data_private( error = xfs_free_extent(tp, XFS_AGB_TO_FSB(mp, agno, be32_to_cpu(agf->agf_length) - new), - new, &oinfo); + new, &oinfo, XFS_AG_RESV_NONE); if (error) goto error0; } @@ -589,6 +631,25 @@ xfs_growfs_data_private( xfs_set_low_space_thresholds(mp); mp->m_alloc_set_aside = xfs_alloc_set_aside(mp); + /* + * If we expanded the last AG, free the per-AG reservation + * so we can reinitialize it with the new size. + */ + if (new) { + struct xfs_perag *pag; + + pag = xfs_perag_get(mp, agno); + error = xfs_ag_resv_free(pag); + xfs_perag_put(pag); + if (error) + goto out; + } + + /* Reserve AG metadata blocks. */ + error = xfs_fs_reserve_ag_blocks(mp); + if (error && error != -ENOSPC) + goto out; + /* update secondary superblocks. */ for (agno = 1; agno < nagcount; agno++) { error = 0; @@ -639,6 +700,8 @@ xfs_growfs_data_private( continue; } } + + out: return saved_error ? saved_error : error; error0: @@ -948,3 +1011,59 @@ xfs_do_force_shutdown( "Please umount the filesystem and rectify the problem(s)"); } } + +/* + * Reserve free space for per-AG metadata. + */ +int +xfs_fs_reserve_ag_blocks( + struct xfs_mount *mp) +{ + xfs_agnumber_t agno; + struct xfs_perag *pag; + int error = 0; + int err2; + + for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { + pag = xfs_perag_get(mp, agno); + err2 = xfs_ag_resv_init(pag); + xfs_perag_put(pag); + if (err2 && !error) + error = err2; + } + + if (error && error != -ENOSPC) { + xfs_warn(mp, + "Error %d reserving per-AG metadata reserve pool.", error); + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + } + + return error; +} + +/* + * Free space reserved for per-AG metadata. + */ +int +xfs_fs_unreserve_ag_blocks( + struct xfs_mount *mp) +{ + xfs_agnumber_t agno; + struct xfs_perag *pag; + int error = 0; + int err2; + + for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { + pag = xfs_perag_get(mp, agno); + err2 = xfs_ag_resv_free(pag); + xfs_perag_put(pag); + if (err2 && !error) + error = err2; + } + + if (error) + xfs_warn(mp, + "Error %d freeing per-AG metadata reserve pool.", error); + + return error; +} diff --git a/fs/xfs/xfs_fsops.h b/fs/xfs/xfs_fsops.h index f32713f14f9a..f34915898fea 100644 --- a/fs/xfs/xfs_fsops.h +++ b/fs/xfs/xfs_fsops.h @@ -26,4 +26,7 @@ extern int xfs_reserve_blocks(xfs_mount_t *mp, __uint64_t *inval, xfs_fsop_resblks_t *outval); extern int xfs_fs_goingdown(xfs_mount_t *mp, __uint32_t inflags); +extern int xfs_fs_reserve_ag_blocks(struct xfs_mount *mp); +extern int xfs_fs_unreserve_ag_blocks(struct xfs_mount *mp); + #endif /* __XFS_FSOPS_H__ */ diff --git a/fs/xfs/xfs_globals.c b/fs/xfs/xfs_globals.c index 4d41b241298f..687a4b01fc53 100644 --- a/fs/xfs/xfs_globals.c +++ b/fs/xfs/xfs_globals.c @@ -21,8 +21,8 @@ /* * Tunable XFS parameters. xfs_params is required even when CONFIG_SYSCTL=n, * other XFS code uses these values. Times are measured in centisecs (i.e. - * 100ths of a second) with the exception of eofb_timer, which is measured in - * seconds. + * 100ths of a second) with the exception of eofb_timer and cowb_timer, which + * are measured in seconds. */ xfs_param_t xfs_params = { /* MIN DFLT MAX */ @@ -42,6 +42,7 @@ xfs_param_t xfs_params = { .inherit_nodfrg = { 0, 1, 1 }, .fstrm_timer = { 1, 30*100, 3600*100}, .eofb_timer = { 1, 300, 3600*24}, + .cowb_timer = { 1, 1800, 3600*24}, }; struct xfs_globals xfs_globals = { diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index fb39a66914dd..70ca4f608321 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -33,6 +33,7 @@ #include "xfs_bmap_util.h" #include "xfs_dquot_item.h" #include "xfs_dquot.h" +#include "xfs_reflink.h" #include <linux/kthread.h> #include <linux/freezer.h> @@ -69,13 +70,14 @@ xfs_inode_alloc( ASSERT(!xfs_isiflocked(ip)); ASSERT(ip->i_ino == 0); - mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino); - /* initialise the xfs inode */ ip->i_ino = ino; ip->i_mount = mp; memset(&ip->i_imap, 0, sizeof(struct xfs_imap)); ip->i_afp = NULL; + ip->i_cowfp = NULL; + ip->i_cnextents = 0; + ip->i_cformat = XFS_DINODE_FMT_EXTENTS; memset(&ip->i_df, 0, sizeof(xfs_ifork_t)); ip->i_flags = 0; ip->i_delayed_blks = 0; @@ -101,6 +103,8 @@ xfs_inode_free_callback( if (ip->i_afp) xfs_idestroy_fork(ip, XFS_ATTR_FORK); + if (ip->i_cowfp) + xfs_idestroy_fork(ip, XFS_COW_FORK); if (ip->i_itemp) { ASSERT(!(ip->i_itemp->ili_item.li_flags & XFS_LI_IN_AIL)); @@ -117,7 +121,6 @@ __xfs_inode_free( { /* asserts to verify all state is correct here */ ASSERT(atomic_read(&ip->i_pincount) == 0); - ASSERT(!xfs_isiflocked(ip)); XFS_STATS_DEC(ip->i_mount, vn_active); call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback); @@ -127,6 +130,8 @@ void xfs_inode_free( struct xfs_inode *ip) { + ASSERT(!xfs_isiflocked(ip)); + /* * Because we use RCU freeing we need to ensure the inode always * appears to be reclaimed with an invalid inode number when in the @@ -387,8 +392,8 @@ xfs_iget_cache_hit( xfs_inode_clear_reclaim_tag(pag, ip->i_ino); inode->i_state = I_NEW; - ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock)); - mrlock_init(&ip->i_iolock, MRLOCK_BARRIER, "xfsio", ip->i_ino); + ASSERT(!rwsem_is_locked(&inode->i_rwsem)); + init_rwsem(&inode->i_rwsem); spin_unlock(&ip->i_flags_lock); spin_unlock(&pag->pag_ici_lock); @@ -787,6 +792,33 @@ xfs_eofblocks_worker( xfs_queue_eofblocks(mp); } +/* + * Background scanning to trim preallocated CoW space. This is queued + * based on the 'speculative_cow_prealloc_lifetime' tunable (5m by default). + * (We'll just piggyback on the post-EOF prealloc space workqueue.) + */ +STATIC void +xfs_queue_cowblocks( + struct xfs_mount *mp) +{ + rcu_read_lock(); + if (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_COWBLOCKS_TAG)) + queue_delayed_work(mp->m_eofblocks_workqueue, + &mp->m_cowblocks_work, + msecs_to_jiffies(xfs_cowb_secs * 1000)); + rcu_read_unlock(); +} + +void +xfs_cowblocks_worker( + struct work_struct *work) +{ + struct xfs_mount *mp = container_of(to_delayed_work(work), + struct xfs_mount, m_cowblocks_work); + xfs_icache_free_cowblocks(mp, NULL); + xfs_queue_cowblocks(mp); +} + int xfs_inode_ag_iterator( struct xfs_mount *mp, @@ -948,6 +980,7 @@ restart: if (XFS_FORCED_SHUTDOWN(ip->i_mount)) { xfs_iunpin_wait(ip); + /* xfs_iflush_abort() drops the flush lock */ xfs_iflush_abort(ip, false); goto reclaim; } @@ -956,10 +989,10 @@ restart: goto out_ifunlock; xfs_iunpin_wait(ip); } - if (xfs_iflags_test(ip, XFS_ISTALE)) - goto reclaim; - if (xfs_inode_clean(ip)) + if (xfs_iflags_test(ip, XFS_ISTALE) || xfs_inode_clean(ip)) { + xfs_ifunlock(ip); goto reclaim; + } /* * Never flush out dirty data during non-blocking reclaim, as it would @@ -997,25 +1030,24 @@ restart: xfs_buf_relse(bp); } - xfs_iflock(ip); reclaim: + ASSERT(!xfs_isiflocked(ip)); + /* * Because we use RCU freeing we need to ensure the inode always appears * to be reclaimed with an invalid inode number when in the free state. - * We do this as early as possible under the ILOCK and flush lock so - * that xfs_iflush_cluster() can be guaranteed to detect races with us - * here. By doing this, we guarantee that once xfs_iflush_cluster has - * locked both the XFS_ILOCK and the flush lock that it will see either - * a valid, flushable inode that will serialise correctly against the - * locks below, or it will see a clean (and invalid) inode that it can - * skip. + * We do this as early as possible under the ILOCK so that + * xfs_iflush_cluster() can be guaranteed to detect races with us here. + * By doing this, we guarantee that once xfs_iflush_cluster has locked + * XFS_ILOCK that it will see either a valid, flushable inode that will + * serialise correctly, or it will see a clean (and invalid) inode that + * it can skip. */ spin_lock(&ip->i_flags_lock); ip->i_flags = XFS_IRECLAIM; ip->i_ino = 0; spin_unlock(&ip->i_flags_lock); - xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); XFS_STATS_INC(ip->i_mount, xs_ig_reclaims); @@ -1343,18 +1375,30 @@ xfs_inode_free_eofblocks( return ret; } -int -xfs_icache_free_eofblocks( +static int +__xfs_icache_free_eofblocks( struct xfs_mount *mp, - struct xfs_eofblocks *eofb) + struct xfs_eofblocks *eofb, + int (*execute)(struct xfs_inode *ip, int flags, + void *args), + int tag) { int flags = SYNC_TRYLOCK; if (eofb && (eofb->eof_flags & XFS_EOF_FLAGS_SYNC)) flags = SYNC_WAIT; - return xfs_inode_ag_iterator_tag(mp, xfs_inode_free_eofblocks, flags, - eofb, XFS_ICI_EOFBLOCKS_TAG); + return xfs_inode_ag_iterator_tag(mp, execute, flags, + eofb, tag); +} + +int +xfs_icache_free_eofblocks( + struct xfs_mount *mp, + struct xfs_eofblocks *eofb) +{ + return __xfs_icache_free_eofblocks(mp, eofb, xfs_inode_free_eofblocks, + XFS_ICI_EOFBLOCKS_TAG); } /* @@ -1363,9 +1407,11 @@ xfs_icache_free_eofblocks( * failure. We make a best effort by including each quota under low free space * conditions (less than 1% free space) in the scan. */ -int -xfs_inode_free_quota_eofblocks( - struct xfs_inode *ip) +static int +__xfs_inode_free_quota_eofblocks( + struct xfs_inode *ip, + int (*execute)(struct xfs_mount *mp, + struct xfs_eofblocks *eofb)) { int scan = 0; struct xfs_eofblocks eofb = {0}; @@ -1401,41 +1447,58 @@ xfs_inode_free_quota_eofblocks( } if (scan) - xfs_icache_free_eofblocks(ip->i_mount, &eofb); + execute(ip->i_mount, &eofb); return scan; } -void -xfs_inode_set_eofblocks_tag( - xfs_inode_t *ip) +int +xfs_inode_free_quota_eofblocks( + struct xfs_inode *ip) +{ + return __xfs_inode_free_quota_eofblocks(ip, xfs_icache_free_eofblocks); +} + +static void +__xfs_inode_set_eofblocks_tag( + xfs_inode_t *ip, + void (*execute)(struct xfs_mount *mp), + void (*set_tp)(struct xfs_mount *mp, xfs_agnumber_t agno, + int error, unsigned long caller_ip), + int tag) { struct xfs_mount *mp = ip->i_mount; struct xfs_perag *pag; int tagged; + /* + * Don't bother locking the AG and looking up in the radix trees + * if we already know that we have the tag set. + */ + if (ip->i_flags & XFS_IEOFBLOCKS) + return; + spin_lock(&ip->i_flags_lock); + ip->i_flags |= XFS_IEOFBLOCKS; + spin_unlock(&ip->i_flags_lock); + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); spin_lock(&pag->pag_ici_lock); - trace_xfs_inode_set_eofblocks_tag(ip); - tagged = radix_tree_tagged(&pag->pag_ici_root, - XFS_ICI_EOFBLOCKS_TAG); + tagged = radix_tree_tagged(&pag->pag_ici_root, tag); radix_tree_tag_set(&pag->pag_ici_root, - XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino), - XFS_ICI_EOFBLOCKS_TAG); + XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino), tag); if (!tagged) { /* propagate the eofblocks tag up into the perag radix tree */ spin_lock(&ip->i_mount->m_perag_lock); radix_tree_tag_set(&ip->i_mount->m_perag_tree, XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), - XFS_ICI_EOFBLOCKS_TAG); + tag); spin_unlock(&ip->i_mount->m_perag_lock); /* kick off background trimming */ - xfs_queue_eofblocks(ip->i_mount); + execute(ip->i_mount); - trace_xfs_perag_set_eofblocks(ip->i_mount, pag->pag_agno, - -1, _RET_IP_); + set_tp(ip->i_mount, pag->pag_agno, -1, _RET_IP_); } spin_unlock(&pag->pag_ici_lock); @@ -1443,31 +1506,172 @@ xfs_inode_set_eofblocks_tag( } void -xfs_inode_clear_eofblocks_tag( +xfs_inode_set_eofblocks_tag( xfs_inode_t *ip) { + trace_xfs_inode_set_eofblocks_tag(ip); + return __xfs_inode_set_eofblocks_tag(ip, xfs_queue_eofblocks, + trace_xfs_perag_set_eofblocks, + XFS_ICI_EOFBLOCKS_TAG); +} + +static void +__xfs_inode_clear_eofblocks_tag( + xfs_inode_t *ip, + void (*clear_tp)(struct xfs_mount *mp, xfs_agnumber_t agno, + int error, unsigned long caller_ip), + int tag) +{ struct xfs_mount *mp = ip->i_mount; struct xfs_perag *pag; + spin_lock(&ip->i_flags_lock); + ip->i_flags &= ~XFS_IEOFBLOCKS; + spin_unlock(&ip->i_flags_lock); + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); spin_lock(&pag->pag_ici_lock); - trace_xfs_inode_clear_eofblocks_tag(ip); radix_tree_tag_clear(&pag->pag_ici_root, - XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino), - XFS_ICI_EOFBLOCKS_TAG); - if (!radix_tree_tagged(&pag->pag_ici_root, XFS_ICI_EOFBLOCKS_TAG)) { + XFS_INO_TO_AGINO(ip->i_mount, ip->i_ino), tag); + if (!radix_tree_tagged(&pag->pag_ici_root, tag)) { /* clear the eofblocks tag from the perag radix tree */ spin_lock(&ip->i_mount->m_perag_lock); radix_tree_tag_clear(&ip->i_mount->m_perag_tree, XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), - XFS_ICI_EOFBLOCKS_TAG); + tag); spin_unlock(&ip->i_mount->m_perag_lock); - trace_xfs_perag_clear_eofblocks(ip->i_mount, pag->pag_agno, - -1, _RET_IP_); + clear_tp(ip->i_mount, pag->pag_agno, -1, _RET_IP_); } spin_unlock(&pag->pag_ici_lock); xfs_perag_put(pag); } +void +xfs_inode_clear_eofblocks_tag( + xfs_inode_t *ip) +{ + trace_xfs_inode_clear_eofblocks_tag(ip); + return __xfs_inode_clear_eofblocks_tag(ip, + trace_xfs_perag_clear_eofblocks, XFS_ICI_EOFBLOCKS_TAG); +} + +/* + * Automatic CoW Reservation Freeing + * + * These functions automatically garbage collect leftover CoW reservations + * that were made on behalf of a cowextsize hint when we start to run out + * of quota or when the reservations sit around for too long. If the file + * has dirty pages or is undergoing writeback, its CoW reservations will + * be retained. + * + * The actual garbage collection piggybacks off the same code that runs + * the speculative EOF preallocation garbage collector. + */ +STATIC int +xfs_inode_free_cowblocks( + struct xfs_inode *ip, + int flags, + void *args) +{ + int ret; + struct xfs_eofblocks *eofb = args; + bool need_iolock = true; + int match; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + + ASSERT(!eofb || (eofb && eofb->eof_scan_owner != 0)); + + /* + * Just clear the tag if we have an empty cow fork or none at all. It's + * possible the inode was fully unshared since it was originally tagged. + */ + if (!xfs_is_reflink_inode(ip) || !ifp->if_bytes) { + trace_xfs_inode_free_cowblocks_invalid(ip); + xfs_inode_clear_cowblocks_tag(ip); + return 0; + } + + /* + * If the mapping is dirty or under writeback we cannot touch the + * CoW fork. Leave it alone if we're in the midst of a directio. + */ + if ((VFS_I(ip)->i_state & I_DIRTY_PAGES) || + mapping_tagged(VFS_I(ip)->i_mapping, PAGECACHE_TAG_DIRTY) || + mapping_tagged(VFS_I(ip)->i_mapping, PAGECACHE_TAG_WRITEBACK) || + atomic_read(&VFS_I(ip)->i_dio_count)) + return 0; + + if (eofb) { + if (eofb->eof_flags & XFS_EOF_FLAGS_UNION) + match = xfs_inode_match_id_union(ip, eofb); + else + match = xfs_inode_match_id(ip, eofb); + if (!match) + return 0; + + /* skip the inode if the file size is too small */ + if (eofb->eof_flags & XFS_EOF_FLAGS_MINFILESIZE && + XFS_ISIZE(ip) < eofb->eof_min_file_size) + return 0; + + /* + * A scan owner implies we already hold the iolock. Skip it in + * xfs_free_eofblocks() to avoid deadlock. This also eliminates + * the possibility of EAGAIN being returned. + */ + if (eofb->eof_scan_owner == ip->i_ino) + need_iolock = false; + } + + /* Free the CoW blocks */ + if (need_iolock) { + xfs_ilock(ip, XFS_IOLOCK_EXCL); + xfs_ilock(ip, XFS_MMAPLOCK_EXCL); + } + + ret = xfs_reflink_cancel_cow_range(ip, 0, NULLFILEOFF); + + if (need_iolock) { + xfs_iunlock(ip, XFS_MMAPLOCK_EXCL); + xfs_iunlock(ip, XFS_IOLOCK_EXCL); + } + + return ret; +} + +int +xfs_icache_free_cowblocks( + struct xfs_mount *mp, + struct xfs_eofblocks *eofb) +{ + return __xfs_icache_free_eofblocks(mp, eofb, xfs_inode_free_cowblocks, + XFS_ICI_COWBLOCKS_TAG); +} + +int +xfs_inode_free_quota_cowblocks( + struct xfs_inode *ip) +{ + return __xfs_inode_free_quota_eofblocks(ip, xfs_icache_free_cowblocks); +} + +void +xfs_inode_set_cowblocks_tag( + xfs_inode_t *ip) +{ + trace_xfs_inode_set_cowblocks_tag(ip); + return __xfs_inode_set_eofblocks_tag(ip, xfs_queue_cowblocks, + trace_xfs_perag_set_cowblocks, + XFS_ICI_COWBLOCKS_TAG); +} + +void +xfs_inode_clear_cowblocks_tag( + xfs_inode_t *ip) +{ + trace_xfs_inode_clear_cowblocks_tag(ip); + return __xfs_inode_clear_eofblocks_tag(ip, + trace_xfs_perag_clear_cowblocks, XFS_ICI_COWBLOCKS_TAG); +} diff --git a/fs/xfs/xfs_icache.h b/fs/xfs/xfs_icache.h index 05bac99bef75..a1e02f4708ab 100644 --- a/fs/xfs/xfs_icache.h +++ b/fs/xfs/xfs_icache.h @@ -40,6 +40,7 @@ struct xfs_eofblocks { in xfs_inode_ag_iterator */ #define XFS_ICI_RECLAIM_TAG 0 /* inode is to be reclaimed */ #define XFS_ICI_EOFBLOCKS_TAG 1 /* inode has blocks beyond EOF */ +#define XFS_ICI_COWBLOCKS_TAG 2 /* inode can have cow blocks to gc */ /* * Flags for xfs_iget() @@ -70,6 +71,12 @@ int xfs_inode_free_quota_eofblocks(struct xfs_inode *ip); void xfs_eofblocks_worker(struct work_struct *); void xfs_queue_eofblocks(struct xfs_mount *); +void xfs_inode_set_cowblocks_tag(struct xfs_inode *ip); +void xfs_inode_clear_cowblocks_tag(struct xfs_inode *ip); +int xfs_icache_free_cowblocks(struct xfs_mount *, struct xfs_eofblocks *); +int xfs_inode_free_quota_cowblocks(struct xfs_inode *ip); +void xfs_cowblocks_worker(struct work_struct *); + int xfs_inode_ag_iterator(struct xfs_mount *mp, int (*execute)(struct xfs_inode *ip, int flags, void *args), int flags, void *args); diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c index d45ca72af6fb..865ad1373e5e 100644 --- a/fs/xfs/xfs_icreate_item.c +++ b/fs/xfs/xfs_icreate_item.c @@ -133,7 +133,7 @@ xfs_icreate_item_committing( /* * This is the ops vector shared by all buf log items. */ -static struct xfs_item_ops xfs_icreate_item_ops = { +static const struct xfs_item_ops xfs_icreate_item_ops = { .iop_size = xfs_icreate_item_size, .iop_format = xfs_icreate_item_format, .iop_pin = xfs_icreate_item_pin, diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index e08eaea6327b..b9557795eb74 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -49,6 +49,7 @@ #include "xfs_trans_priv.h" #include "xfs_log.h" #include "xfs_bmap_btree.h" +#include "xfs_reflink.h" kmem_zone_t *xfs_inode_zone; @@ -77,6 +78,29 @@ xfs_get_extsz_hint( } /* + * Helper function to extract CoW extent size hint from inode. + * Between the extent size hint and the CoW extent size hint, we + * return the greater of the two. If the value is zero (automatic), + * use the default size. + */ +xfs_extlen_t +xfs_get_cowextsz_hint( + struct xfs_inode *ip) +{ + xfs_extlen_t a, b; + + a = 0; + if (ip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) + a = ip->i_d.di_cowextsize; + b = xfs_get_extsz_hint(ip); + + a = max(a, b); + if (a == 0) + return XFS_DEFAULT_COWEXTSZ_HINT; + return a; +} + +/* * These two are wrapper routines around the xfs_ilock() routine used to * centralize some grungy code. They are used in places that wish to lock the * inode solely for reading the extents. The reason these places can't just @@ -118,31 +142,31 @@ xfs_ilock_attr_map_shared( } /* - * The xfs inode contains 3 multi-reader locks: the i_iolock the i_mmap_lock and - * the i_lock. This routine allows various combinations of the locks to be - * obtained. + * In addition to i_rwsem in the VFS inode, the xfs inode contains 2 + * multi-reader locks: i_mmap_lock and the i_lock. This routine allows + * various combinations of the locks to be obtained. * * The 3 locks should always be ordered so that the IO lock is obtained first, * the mmap lock second and the ilock last in order to prevent deadlock. * * Basic locking order: * - * i_iolock -> i_mmap_lock -> page_lock -> i_ilock + * i_rwsem -> i_mmap_lock -> page_lock -> i_ilock * * mmap_sem locking order: * - * i_iolock -> page lock -> mmap_sem + * i_rwsem -> page lock -> mmap_sem * mmap_sem -> i_mmap_lock -> page_lock * * The difference in mmap_sem locking order mean that we cannot hold the * i_mmap_lock over syscall based read(2)/write(2) based IO. These IO paths can * fault in pages during copy in/out (for buffered IO) or require the mmap_sem * in get_user_pages() to map the user pages into the kernel address space for - * direct IO. Similarly the i_iolock cannot be taken inside a page fault because + * direct IO. Similarly the i_rwsem cannot be taken inside a page fault because * page faults already hold the mmap_sem. * * Hence to serialise fully against both syscall and mmap based IO, we need to - * take both the i_iolock and the i_mmap_lock. These locks should *only* be both + * take both the i_rwsem and the i_mmap_lock. These locks should *only* be both * taken in places where we need to invalidate the page cache in a race * free manner (e.g. truncate, hole punch and other extent manipulation * functions). @@ -167,10 +191,13 @@ xfs_ilock( (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0); - if (lock_flags & XFS_IOLOCK_EXCL) - mrupdate_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags)); - else if (lock_flags & XFS_IOLOCK_SHARED) - mraccess_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags)); + if (lock_flags & XFS_IOLOCK_EXCL) { + down_write_nested(&VFS_I(ip)->i_rwsem, + XFS_IOLOCK_DEP(lock_flags)); + } else if (lock_flags & XFS_IOLOCK_SHARED) { + down_read_nested(&VFS_I(ip)->i_rwsem, + XFS_IOLOCK_DEP(lock_flags)); + } if (lock_flags & XFS_MMAPLOCK_EXCL) mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags)); @@ -216,10 +243,10 @@ xfs_ilock_nowait( ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0); if (lock_flags & XFS_IOLOCK_EXCL) { - if (!mrtryupdate(&ip->i_iolock)) + if (!down_write_trylock(&VFS_I(ip)->i_rwsem)) goto out; } else if (lock_flags & XFS_IOLOCK_SHARED) { - if (!mrtryaccess(&ip->i_iolock)) + if (!down_read_trylock(&VFS_I(ip)->i_rwsem)) goto out; } @@ -247,9 +274,9 @@ out_undo_mmaplock: mrunlock_shared(&ip->i_mmaplock); out_undo_iolock: if (lock_flags & XFS_IOLOCK_EXCL) - mrunlock_excl(&ip->i_iolock); + up_write(&VFS_I(ip)->i_rwsem); else if (lock_flags & XFS_IOLOCK_SHARED) - mrunlock_shared(&ip->i_iolock); + up_read(&VFS_I(ip)->i_rwsem); out: return 0; } @@ -286,9 +313,9 @@ xfs_iunlock( ASSERT(lock_flags != 0); if (lock_flags & XFS_IOLOCK_EXCL) - mrunlock_excl(&ip->i_iolock); + up_write(&VFS_I(ip)->i_rwsem); else if (lock_flags & XFS_IOLOCK_SHARED) - mrunlock_shared(&ip->i_iolock); + up_read(&VFS_I(ip)->i_rwsem); if (lock_flags & XFS_MMAPLOCK_EXCL) mrunlock_excl(&ip->i_mmaplock); @@ -321,7 +348,7 @@ xfs_ilock_demote( if (lock_flags & XFS_MMAPLOCK_EXCL) mrdemote(&ip->i_mmaplock); if (lock_flags & XFS_IOLOCK_EXCL) - mrdemote(&ip->i_iolock); + downgrade_write(&VFS_I(ip)->i_rwsem); trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_); } @@ -346,8 +373,9 @@ xfs_isilocked( if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) { if (!(lock_flags & XFS_IOLOCK_SHARED)) - return !!ip->i_iolock.mr_writer; - return rwsem_is_locked(&ip->i_iolock.mr_lock); + return !debug_locks || + lockdep_is_held_type(&VFS_I(ip)->i_rwsem, 0); + return rwsem_is_locked(&VFS_I(ip)->i_rwsem); } ASSERT(0); @@ -397,11 +425,7 @@ xfs_lock_inumorder(int lock_mode, int subclass) if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) { ASSERT(subclass <= XFS_IOLOCK_MAX_SUBCLASS); - ASSERT(xfs_lockdep_subclass_ok(subclass + - XFS_IOLOCK_PARENT_VAL)); class += subclass << XFS_IOLOCK_SHIFT; - if (lock_mode & XFS_IOLOCK_PARENT) - class += XFS_IOLOCK_PARENT_VAL << XFS_IOLOCK_SHIFT; } if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) { @@ -453,8 +477,6 @@ xfs_lock_inodes( XFS_ILOCK_EXCL)); ASSERT(!(lock_mode & (XFS_IOLOCK_SHARED | XFS_MMAPLOCK_SHARED | XFS_ILOCK_SHARED))); - ASSERT(!(lock_mode & XFS_IOLOCK_EXCL) || - inodes <= XFS_IOLOCK_MAX_SUBCLASS + 1); ASSERT(!(lock_mode & XFS_MMAPLOCK_EXCL) || inodes <= XFS_MMAPLOCK_MAX_SUBCLASS + 1); ASSERT(!(lock_mode & XFS_ILOCK_EXCL) || @@ -557,10 +579,8 @@ xfs_lock_two_inodes( int attempts = 0; xfs_log_item_t *lp; - if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) { - ASSERT(!(lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL))); - ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))); - } else if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) + ASSERT(!(lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL))); + if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) ASSERT(!(lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL))); ASSERT(ip0->i_ino != ip1->i_ino); @@ -651,6 +671,8 @@ _xfs_dic2xflags( if (di_flags2 & XFS_DIFLAG2_ANY) { if (di_flags2 & XFS_DIFLAG2_DAX) flags |= FS_XFLAG_DAX; + if (di_flags2 & XFS_DIFLAG2_COWEXTSIZE) + flags |= FS_XFLAG_COWEXTSIZE; } if (has_attr) @@ -689,7 +711,6 @@ xfs_lookup( if (XFS_FORCED_SHUTDOWN(dp->i_mount)) return -EIO; - xfs_ilock(dp, XFS_IOLOCK_SHARED); error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name); if (error) goto out_unlock; @@ -698,14 +719,12 @@ xfs_lookup( if (error) goto out_free_name; - xfs_iunlock(dp, XFS_IOLOCK_SHARED); return 0; out_free_name: if (ci_name) kmem_free(ci_name->name); out_unlock: - xfs_iunlock(dp, XFS_IOLOCK_SHARED); *ipp = NULL; return error; } @@ -821,7 +840,7 @@ xfs_ialloc( ip->i_d.di_nextents = 0; ASSERT(ip->i_d.di_nblocks == 0); - tv = current_fs_time(mp->m_super); + tv = current_time(inode); inode->i_mtime = tv; inode->i_atime = tv; inode->i_ctime = tv; @@ -834,6 +853,7 @@ xfs_ialloc( if (ip->i_d.di_version == 3) { inode->i_version = 1; ip->i_d.di_flags2 = 0; + ip->i_d.di_cowextsize = 0; ip->i_d.di_crtime.t_sec = (__int32_t)tv.tv_sec; ip->i_d.di_crtime.t_nsec = (__int32_t)tv.tv_nsec; } @@ -896,6 +916,15 @@ xfs_ialloc( ip->i_d.di_flags |= di_flags; ip->i_d.di_flags2 |= di_flags2; } + if (pip && + (pip->i_d.di_flags2 & XFS_DIFLAG2_ANY) && + pip->i_d.di_version == 3 && + ip->i_d.di_version == 3) { + if (pip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) { + ip->i_d.di_flags2 |= XFS_DIFLAG2_COWEXTSIZE; + ip->i_d.di_cowextsize = pip->i_d.di_cowextsize; + } + } /* FALLTHROUGH */ case S_IFLNK: ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS; @@ -1179,8 +1208,7 @@ xfs_create( if (error) goto out_release_inode; - xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL | - XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); + xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT); unlock_dp_on_error = true; xfs_defer_init(&dfops, &first_block); @@ -1216,7 +1244,7 @@ xfs_create( * the transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = false; error = xfs_dir_createname(tp, dp, name, ip->i_ino, @@ -1289,7 +1317,7 @@ xfs_create( xfs_qm_dqrele(pdqp); if (unlock_dp_on_error) - xfs_iunlock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_iunlock(dp, XFS_ILOCK_EXCL); return error; } @@ -1430,11 +1458,10 @@ xfs_link( if (error) goto std_return; - xfs_ilock(tdp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, tdp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL); /* * If we are using project inheritance, we only allow hard link @@ -1586,6 +1613,20 @@ xfs_itruncate_extents( goto out; } + /* Remove all pending CoW reservations. */ + error = xfs_reflink_cancel_cow_blocks(ip, &tp, first_unmap_block, + last_block); + if (error) + goto out; + + /* + * Clear the reflink flag if we truncated everything. + */ + if (ip->i_d.di_nblocks == 0 && xfs_is_reflink_inode(ip)) { + ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + xfs_inode_clear_cowblocks_tag(ip); + } + /* * Always re-log the inode so that our permanent transaction can keep * on rolling it forward in the log. @@ -1710,7 +1751,7 @@ xfs_inactive_truncate( /* * Log the inode size first to prevent stale data exposure in the event * of a system crash before the truncate completes. See the related - * comment in xfs_setattr_size() for details. + * comment in xfs_vn_setattr_size() for details. */ ip->i_d.di_size = 0; xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); @@ -1850,6 +1891,7 @@ xfs_inactive( } mp = ip->i_mount; + ASSERT(!xfs_iflags_test(ip, XFS_IRECOVERY)); /* If this is a read-only mount, don't do this (would generate I/O) */ if (mp->m_flags & XFS_MOUNT_RDONLY) @@ -1990,7 +2032,6 @@ xfs_iunlink( agi->agi_unlinked[bucket_index] = cpu_to_be32(agino); offset = offsetof(xfs_agi_t, agi_unlinked) + (sizeof(xfs_agino_t) * bucket_index); - xfs_trans_buf_set_type(tp, agibp, XFS_BLFT_AGI_BUF); xfs_trans_log_buf(tp, agibp, offset, (offset + sizeof(xfs_agino_t) - 1)); return 0; @@ -2082,7 +2123,6 @@ xfs_iunlink_remove( agi->agi_unlinked[bucket_index] = cpu_to_be32(next_agino); offset = offsetof(xfs_agi_t, agi_unlinked) + (sizeof(xfs_agino_t) * bucket_index); - xfs_trans_buf_set_type(tp, agibp, XFS_BLFT_AGI_BUF); xfs_trans_log_buf(tp, agibp, offset, (offset + sizeof(xfs_agino_t) - 1)); } else { @@ -2528,10 +2568,9 @@ xfs_remove( goto std_return; } - xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL); - xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); /* @@ -2912,12 +2951,6 @@ xfs_rename( * whether the target directory is the same as the source * directory, we can lock from 2 to 4 inodes. */ - if (!new_parent) - xfs_ilock(src_dp, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); - else - xfs_lock_two_inodes(src_dp, target_dp, - XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT); - xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL); /* @@ -2925,9 +2958,9 @@ xfs_rename( * we can rely on either trans_commit or trans_cancel to unlock * them. */ - xfs_trans_ijoin(tp, src_dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL); if (new_parent) - xfs_trans_ijoin(tp, target_dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL); if (target_ip) xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index e1a411e08f00..10dcf27b4c85 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -47,6 +47,7 @@ typedef struct xfs_inode { /* Extent information. */ xfs_ifork_t *i_afp; /* attribute fork pointer */ + xfs_ifork_t *i_cowfp; /* copy on write extents */ xfs_ifork_t i_df; /* data fork */ /* operations vectors */ @@ -55,7 +56,6 @@ typedef struct xfs_inode { /* Transaction and locking information. */ struct xfs_inode_log_item *i_itemp; /* logging information */ mrlock_t i_lock; /* inode lock */ - mrlock_t i_iolock; /* inode IO lock */ mrlock_t i_mmaplock; /* inode mmap IO lock */ atomic_t i_pincount; /* inode pin count */ spinlock_t i_flags_lock; /* inode i_flags lock */ @@ -65,6 +65,9 @@ typedef struct xfs_inode { struct xfs_icdinode i_d; /* most of ondisk inode */ + xfs_extnum_t i_cnextents; /* # of extents in cow fork */ + unsigned int i_cformat; /* format of cow fork */ + /* VFS inode */ struct inode i_vnode; /* embedded VFS inode */ } xfs_inode_t; @@ -202,6 +205,11 @@ xfs_get_initial_prid(struct xfs_inode *dp) return XFS_PROJID_DEFAULT; } +static inline bool xfs_is_reflink_inode(struct xfs_inode *ip) +{ + return ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK; +} + /* * In-core inode flags. */ @@ -216,6 +224,13 @@ xfs_get_initial_prid(struct xfs_inode *dp) #define __XFS_IPINNED_BIT 8 /* wakeup key for zero pin count */ #define XFS_IPINNED (1 << __XFS_IPINNED_BIT) #define XFS_IDONTCACHE (1 << 9) /* don't cache the inode long term */ +#define XFS_IEOFBLOCKS (1 << 10)/* has the preallocblocks tag set */ +/* + * If this unlinked inode is in the middle of recovery, don't let drop_inode + * truncate and free the inode. This can happen if we iget the inode during + * log recovery to replay a bmap operation on the inode. + */ +#define XFS_IRECOVERY (1 << 11) /* * Per-lifetime flags need to be reset when re-using a reclaimable inode during @@ -230,6 +245,11 @@ xfs_get_initial_prid(struct xfs_inode *dp) * Synchronize processes attempting to flush the in-core inode back to disk. */ +static inline int xfs_isiflocked(struct xfs_inode *ip) +{ + return xfs_iflags_test(ip, XFS_IFLOCK); +} + extern void __xfs_iflock(struct xfs_inode *ip); static inline int xfs_iflock_nowait(struct xfs_inode *ip) @@ -245,16 +265,12 @@ static inline void xfs_iflock(struct xfs_inode *ip) static inline void xfs_ifunlock(struct xfs_inode *ip) { + ASSERT(xfs_isiflocked(ip)); xfs_iflags_clear(ip, XFS_IFLOCK); smp_mb(); wake_up_bit(&ip->i_flags, __XFS_IFLOCK_BIT); } -static inline int xfs_isiflocked(struct xfs_inode *ip) -{ - return xfs_iflags_test(ip, XFS_IFLOCK); -} - /* * Flags for inode locking. * Bit ranges: 1<<1 - 1<<16-1 -- iolock/ilock modes (bitfield) @@ -316,7 +332,7 @@ static inline int xfs_isiflocked(struct xfs_inode *ip) * IOLOCK values * * 0-3 subclass value - * 4-7 PARENT subclass values + * 4-7 unused * * MMAPLOCK values * @@ -331,10 +347,8 @@ static inline int xfs_isiflocked(struct xfs_inode *ip) * */ #define XFS_IOLOCK_SHIFT 16 -#define XFS_IOLOCK_PARENT_VAL 4 -#define XFS_IOLOCK_MAX_SUBCLASS (XFS_IOLOCK_PARENT_VAL - 1) +#define XFS_IOLOCK_MAX_SUBCLASS 3 #define XFS_IOLOCK_DEP_MASK 0x000f0000 -#define XFS_IOLOCK_PARENT (XFS_IOLOCK_PARENT_VAL << XFS_IOLOCK_SHIFT) #define XFS_MMAPLOCK_SHIFT 20 #define XFS_MMAPLOCK_NUMORDER 0 @@ -410,6 +424,7 @@ int xfs_iflush(struct xfs_inode *, struct xfs_buf **); void xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint); xfs_extlen_t xfs_get_extsz_hint(struct xfs_inode *ip); +xfs_extlen_t xfs_get_cowextsz_hint(struct xfs_inode *ip); int xfs_dir_ialloc(struct xfs_trans **, struct xfs_inode *, umode_t, xfs_nlink_t, xfs_dev_t, prid_t, int, @@ -473,4 +488,7 @@ do { \ extern struct kmem_zone *xfs_inode_zone; +/* The default CoW extent size hint. */ +#define XFS_DEFAULT_COWEXTSZ_HINT 32 + #endif /* __XFS_INODE_H__ */ diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 892c2aced207..d90e7811ccdd 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -164,7 +164,7 @@ xfs_inode_item_format_data_fork( struct xfs_bmbt_rec *p; ASSERT(ip->i_df.if_u1.if_extents != NULL); - ASSERT(ip->i_df.if_bytes / sizeof(xfs_bmbt_rec_t) > 0); + ASSERT(xfs_iext_count(&ip->i_df) > 0); p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IEXT); data_bytes = xfs_iextents_copy(ip, p, XFS_DATA_FORK); @@ -261,7 +261,7 @@ xfs_inode_item_format_attr_fork( ip->i_afp->if_bytes > 0) { struct xfs_bmbt_rec *p; - ASSERT(ip->i_afp->if_bytes / sizeof(xfs_bmbt_rec_t) == + ASSERT(xfs_iext_count(ip->i_afp) == ip->i_d.di_anextents); ASSERT(ip->i_afp->if_u1.if_extents != NULL); @@ -368,7 +368,7 @@ xfs_inode_to_log_dinode( to->di_crtime.t_sec = from->di_crtime.t_sec; to->di_crtime.t_nsec = from->di_crtime.t_nsec; to->di_flags2 = from->di_flags2; - + to->di_cowextsize = from->di_cowextsize; to->di_ino = ip->i_ino; to->di_lsn = lsn; memset(to->di_pad2, 0, sizeof(to->di_pad2)); diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index 96a70fd1f5d6..c67cfb451fd3 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -287,7 +287,7 @@ xfs_readlink_by_handle( return PTR_ERR(dentry); /* Restrict this handle operation to symlinks only. */ - if (!d_inode(dentry)->i_op->readlink) { + if (!d_is_symlink(dentry)) { error = -EINVAL; goto out_dput; } @@ -297,7 +297,7 @@ xfs_readlink_by_handle( goto out_dput; } - error = d_inode(dentry)->i_op->readlink(dentry, hreq->ohandle, olen); + error = vfs_readlink(dentry, hreq->ohandle, olen); out_dput: dput(dentry); @@ -639,7 +639,7 @@ xfs_ioc_space( return error; xfs_ilock(ip, iolock); - error = xfs_break_layouts(inode, &iolock, false); + error = xfs_break_layouts(inode, &iolock); if (error) goto out_unlock; @@ -720,7 +720,7 @@ xfs_ioc_space( iattr.ia_valid = ATTR_SIZE; iattr.ia_size = bf->l_start; - error = xfs_setattr_size(ip, &iattr); + error = xfs_vn_setattr_size(file_dentry(filp), &iattr); break; default: ASSERT(0); @@ -903,21 +903,21 @@ xfs_ioc_fsgetxattr( xfs_ilock(ip, XFS_ILOCK_SHARED); fa.fsx_xflags = xfs_ip2xflags(ip); fa.fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog; + fa.fsx_cowextsize = ip->i_d.di_cowextsize << + ip->i_mount->m_sb.sb_blocklog; fa.fsx_projid = xfs_get_projid(ip); if (attr) { if (ip->i_afp) { if (ip->i_afp->if_flags & XFS_IFEXTENTS) - fa.fsx_nextents = ip->i_afp->if_bytes / - sizeof(xfs_bmbt_rec_t); + fa.fsx_nextents = xfs_iext_count(ip->i_afp); else fa.fsx_nextents = ip->i_d.di_anextents; } else fa.fsx_nextents = 0; } else { if (ip->i_df.if_flags & XFS_IFEXTENTS) - fa.fsx_nextents = ip->i_df.if_bytes / - sizeof(xfs_bmbt_rec_t); + fa.fsx_nextents = xfs_iext_count(&ip->i_df); else fa.fsx_nextents = ip->i_d.di_nextents; } @@ -973,12 +973,13 @@ xfs_set_diflags( if (ip->i_d.di_version < 3) return; - di_flags2 = 0; + di_flags2 = (ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK); if (xflags & FS_XFLAG_DAX) di_flags2 |= XFS_DIFLAG2_DAX; + if (xflags & FS_XFLAG_COWEXTSIZE) + di_flags2 |= XFS_DIFLAG2_COWEXTSIZE; ip->i_d.di_flags2 = di_flags2; - } STATIC void @@ -1031,6 +1032,14 @@ xfs_ioctl_setattr_xflags( return -EINVAL; } + /* Clear reflink if we are actually able to set the rt flag. */ + if ((fa->fsx_xflags & FS_XFLAG_REALTIME) && xfs_is_reflink_inode(ip)) + ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + + /* Don't allow us to set DAX mode for a reflinked file for now. */ + if ((fa->fsx_xflags & FS_XFLAG_DAX) && xfs_is_reflink_inode(ip)) + return -EINVAL; + /* * Can't modify an immutable/append-only file unless * we have appropriate permission. @@ -1219,6 +1228,56 @@ xfs_ioctl_setattr_check_extsize( return 0; } +/* + * CoW extent size hint validation rules are: + * + * 1. CoW extent size hint can only be set if reflink is enabled on the fs. + * The inode does not have to have any shared blocks, but it must be a v3. + * 2. FS_XFLAG_COWEXTSIZE is only valid for directories and regular files; + * for a directory, the hint is propagated to new files. + * 3. Can be changed on files & directories at any time. + * 4. CoW extsize hint of 0 turns off hints, clears inode flags. + * 5. Extent size must be a multiple of the appropriate block size. + * 6. The extent size hint must be limited to half the AG size to avoid + * alignment extending the extent beyond the limits of the AG. + */ +static int +xfs_ioctl_setattr_check_cowextsize( + struct xfs_inode *ip, + struct fsxattr *fa) +{ + struct xfs_mount *mp = ip->i_mount; + + if (!(fa->fsx_xflags & FS_XFLAG_COWEXTSIZE)) + return 0; + + if (!xfs_sb_version_hasreflink(&ip->i_mount->m_sb) || + ip->i_d.di_version != 3) + return -EINVAL; + + if (!S_ISREG(VFS_I(ip)->i_mode) && !S_ISDIR(VFS_I(ip)->i_mode)) + return -EINVAL; + + if (fa->fsx_cowextsize != 0) { + xfs_extlen_t size; + xfs_fsblock_t cowextsize_fsb; + + cowextsize_fsb = XFS_B_TO_FSB(mp, fa->fsx_cowextsize); + if (cowextsize_fsb > MAXEXTLEN) + return -EINVAL; + + size = mp->m_sb.sb_blocksize; + if (cowextsize_fsb > mp->m_sb.sb_agblocks / 2) + return -EINVAL; + + if (fa->fsx_cowextsize % size) + return -EINVAL; + } else + fa->fsx_xflags &= ~FS_XFLAG_COWEXTSIZE; + + return 0; +} + static int xfs_ioctl_setattr_check_projid( struct xfs_inode *ip, @@ -1311,6 +1370,10 @@ xfs_ioctl_setattr( if (code) goto error_trans_cancel; + code = xfs_ioctl_setattr_check_cowextsize(ip, fa); + if (code) + goto error_trans_cancel; + code = xfs_ioctl_setattr_xflags(tp, ip, fa); if (code) goto error_trans_cancel; @@ -1346,6 +1409,12 @@ xfs_ioctl_setattr( ip->i_d.di_extsize = fa->fsx_extsize >> mp->m_sb.sb_blocklog; else ip->i_d.di_extsize = 0; + if (ip->i_d.di_version == 3 && + (ip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE)) + ip->i_d.di_cowextsize = fa->fsx_cowextsize >> + mp->m_sb.sb_blocklog; + else + ip->i_d.di_cowextsize = 0; code = xfs_trans_commit(tp); diff --git a/fs/xfs/xfs_ioctl32.c b/fs/xfs/xfs_ioctl32.c index 321f57721b92..7c49938c5aed 100644 --- a/fs/xfs/xfs_ioctl32.c +++ b/fs/xfs/xfs_ioctl32.c @@ -19,7 +19,7 @@ #include <linux/ioctl.h> #include <linux/mount.h> #include <linux/slab.h> -#include <asm/uaccess.h> +#include <linux/uaccess.h> #include "xfs.h" #include "xfs_fs.h" #include "xfs_format.h" diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 2af0dda1c978..0d147428971e 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -1,5 +1,6 @@ /* * Copyright (c) 2000-2006 Silicon Graphics, Inc. + * Copyright (c) 2016 Christoph Hellwig. * All Rights Reserved. * * This program is free software; you can redistribute it and/or @@ -38,21 +39,45 @@ #include "xfs_quota.h" #include "xfs_dquot_item.h" #include "xfs_dquot.h" +#include "xfs_reflink.h" #define XFS_WRITEIO_ALIGN(mp,off) (((off) >> mp->m_writeio_log) \ << mp->m_writeio_log) -#define XFS_WRITE_IMAPS XFS_BMAP_MAX_NMAP -STATIC int -xfs_iomap_eof_align_last_fsb( - xfs_mount_t *mp, - xfs_inode_t *ip, - xfs_extlen_t extsize, - xfs_fileoff_t *last_fsb) +void +xfs_bmbt_to_iomap( + struct xfs_inode *ip, + struct iomap *iomap, + struct xfs_bmbt_irec *imap) { - xfs_extlen_t align = 0; - int eof, error; + struct xfs_mount *mp = ip->i_mount; + + if (imap->br_startblock == HOLESTARTBLOCK) { + iomap->blkno = IOMAP_NULL_BLOCK; + iomap->type = IOMAP_HOLE; + } else if (imap->br_startblock == DELAYSTARTBLOCK) { + iomap->blkno = IOMAP_NULL_BLOCK; + iomap->type = IOMAP_DELALLOC; + } else { + iomap->blkno = xfs_fsb_to_db(ip, imap->br_startblock); + if (imap->br_state == XFS_EXT_UNWRITTEN) + iomap->type = IOMAP_UNWRITTEN; + else + iomap->type = IOMAP_MAPPED; + } + iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff); + iomap->length = XFS_FSB_TO_B(mp, imap->br_blockcount); + iomap->bdev = xfs_find_bdev_for_inode(VFS_I(ip)); +} + +xfs_extlen_t +xfs_eof_alignment( + struct xfs_inode *ip, + xfs_extlen_t extsize) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_extlen_t align = 0; if (!XFS_IS_REALTIME_INODE(ip)) { /* @@ -83,8 +108,21 @@ xfs_iomap_eof_align_last_fsb( align = extsize; } + return align; +} + +STATIC int +xfs_iomap_eof_align_last_fsb( + struct xfs_inode *ip, + xfs_extlen_t extsize, + xfs_fileoff_t *last_fsb) +{ + xfs_extlen_t align = xfs_eof_alignment(ip, extsize); + if (align) { xfs_fileoff_t new_last_fsb = roundup_64(*last_fsb, align); + int eof, error; + error = xfs_bmap_eof(ip, new_last_fsb, XFS_DATA_FORK, &eof); if (error) return error; @@ -154,7 +192,7 @@ xfs_iomap_write_direct( */ ASSERT(XFS_IFORK_PTR(ip, XFS_DATA_FORK)->if_flags & XFS_IFEXTENTS); - error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb); + error = xfs_iomap_eof_align_last_fsb(ip, extsz, &last_fsb); if (error) goto out_unlock; } else { @@ -274,130 +312,6 @@ out_trans_cancel: goto out_unlock; } -/* - * If the caller is doing a write at the end of the file, then extend the - * allocation out to the file system's write iosize. We clean up any extra - * space left over when the file is closed in xfs_inactive(). - * - * If we find we already have delalloc preallocation beyond EOF, don't do more - * preallocation as it it not needed. - */ -STATIC int -xfs_iomap_eof_want_preallocate( - xfs_mount_t *mp, - xfs_inode_t *ip, - xfs_off_t offset, - size_t count, - xfs_bmbt_irec_t *imap, - int nimaps, - int *prealloc) -{ - xfs_fileoff_t start_fsb; - xfs_filblks_t count_fsb; - int n, error, imaps; - int found_delalloc = 0; - - *prealloc = 0; - if (offset + count <= XFS_ISIZE(ip)) - return 0; - - /* - * If the file is smaller than the minimum prealloc and we are using - * dynamic preallocation, don't do any preallocation at all as it is - * likely this is the only write to the file that is going to be done. - */ - if (!(mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) && - XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_writeio_blocks)) - return 0; - - /* - * If there are any real blocks past eof, then don't - * do any speculative allocation. - */ - start_fsb = XFS_B_TO_FSBT(mp, ((xfs_ufsize_t)(offset + count - 1))); - count_fsb = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes); - while (count_fsb > 0) { - imaps = nimaps; - error = xfs_bmapi_read(ip, start_fsb, count_fsb, imap, &imaps, - 0); - if (error) - return error; - for (n = 0; n < imaps; n++) { - if ((imap[n].br_startblock != HOLESTARTBLOCK) && - (imap[n].br_startblock != DELAYSTARTBLOCK)) - return 0; - start_fsb += imap[n].br_blockcount; - count_fsb -= imap[n].br_blockcount; - - if (imap[n].br_startblock == DELAYSTARTBLOCK) - found_delalloc = 1; - } - } - if (!found_delalloc) - *prealloc = 1; - return 0; -} - -/* - * Determine the initial size of the preallocation. We are beyond the current - * EOF here, but we need to take into account whether this is a sparse write or - * an extending write when determining the preallocation size. Hence we need to - * look up the extent that ends at the current write offset and use the result - * to determine the preallocation size. - * - * If the extent is a hole, then preallocation is essentially disabled. - * Otherwise we take the size of the preceeding data extent as the basis for the - * preallocation size. If the size of the extent is greater than half the - * maximum extent length, then use the current offset as the basis. This ensures - * that for large files the preallocation size always extends to MAXEXTLEN - * rather than falling short due to things like stripe unit/width alignment of - * real extents. - */ -STATIC xfs_fsblock_t -xfs_iomap_eof_prealloc_initial_size( - struct xfs_mount *mp, - struct xfs_inode *ip, - xfs_off_t offset, - xfs_bmbt_irec_t *imap, - int nimaps) -{ - xfs_fileoff_t start_fsb; - int imaps = 1; - int error; - - ASSERT(nimaps >= imaps); - - /* if we are using a specific prealloc size, return now */ - if (mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) - return 0; - - /* If the file is small, then use the minimum prealloc */ - if (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign)) - return 0; - - /* - * As we write multiple pages, the offset will always align to the - * start of a page and hence point to a hole at EOF. i.e. if the size is - * 4096 bytes, we only have one block at FSB 0, but XFS_B_TO_FSB(4096) - * will return FSB 1. Hence if there are blocks in the file, we want to - * point to the block prior to the EOF block and not the hole that maps - * directly at @offset. - */ - start_fsb = XFS_B_TO_FSB(mp, offset); - if (start_fsb) - start_fsb--; - error = xfs_bmapi_read(ip, start_fsb, 1, imap, &imaps, XFS_BMAPI_ENTIRE); - if (error) - return 0; - - ASSERT(imaps == 1); - if (imap[0].br_startblock == HOLESTARTBLOCK) - return 0; - if (imap[0].br_blockcount <= (MAXEXTLEN >> 1)) - return imap[0].br_blockcount << 1; - return XFS_B_TO_FSB(mp, offset); -} - STATIC bool xfs_quota_need_throttle( struct xfs_inode *ip, @@ -459,27 +373,77 @@ xfs_quota_calc_throttle( } /* + * If we are doing a write at the end of the file and there are no allocations + * past this one, then extend the allocation out to the file system's write + * iosize. + * * If we don't have a user specified preallocation size, dynamically increase - * the preallocation size as the size of the file grows. Cap the maximum size + * the preallocation size as the size of the file grows. Cap the maximum size * at a single extent or less if the filesystem is near full. The closer the * filesystem is to full, the smaller the maximum prealocation. + * + * As an exception we don't do any preallocation at all if the file is smaller + * than the minimum preallocation and we are using the default dynamic + * preallocation scheme, as it is likely this is the only write to the file that + * is going to be done. + * + * We clean up any extra space left over when the file is closed in + * xfs_inactive(). */ STATIC xfs_fsblock_t xfs_iomap_prealloc_size( - struct xfs_mount *mp, struct xfs_inode *ip, - xfs_off_t offset, - struct xfs_bmbt_irec *imap, - int nimaps) + loff_t offset, + loff_t count, + xfs_extnum_t idx) { - xfs_fsblock_t alloc_blocks = 0; + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + xfs_fileoff_t offset_fsb = XFS_B_TO_FSBT(mp, offset); + struct xfs_bmbt_irec prev; int shift = 0; int64_t freesp; xfs_fsblock_t qblocks; int qshift = 0; + xfs_fsblock_t alloc_blocks = 0; + + if (offset + count <= XFS_ISIZE(ip)) + return 0; + + if (!(mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) && + (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_writeio_blocks))) + return 0; - alloc_blocks = xfs_iomap_eof_prealloc_initial_size(mp, ip, offset, - imap, nimaps); + /* + * If an explicit allocsize is set, the file is small, or we + * are writing behind a hole, then use the minimum prealloc: + */ + if ((mp->m_flags & XFS_MOUNT_DFLT_IOSIZE) || + XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign) || + !xfs_iext_get_extent(ifp, idx - 1, &prev) || + prev.br_startoff + prev.br_blockcount < offset_fsb) + return mp->m_writeio_blocks; + + /* + * Determine the initial size of the preallocation. We are beyond the + * current EOF here, but we need to take into account whether this is + * a sparse write or an extending write when determining the + * preallocation size. Hence we need to look up the extent that ends + * at the current write offset and use the result to determine the + * preallocation size. + * + * If the extent is a hole, then preallocation is essentially disabled. + * Otherwise we take the size of the preceding data extent as the basis + * for the preallocation size. If the size of the extent is greater than + * half the maximum extent length, then use the current offset as the + * basis. This ensures that for large files the preallocation size + * always extends to MAXEXTLEN rather than falling short due to things + * like stripe unit/width alignment of real extents. + */ + if (prev.br_blockcount <= (MAXEXTLEN >> 1)) + alloc_blocks = prev.br_blockcount << 1; + else + alloc_blocks = XFS_B_TO_FSB(mp, offset); if (!alloc_blocks) goto check_writeio; qblocks = alloc_blocks; @@ -550,120 +514,145 @@ xfs_iomap_prealloc_size( */ while (alloc_blocks && alloc_blocks >= freesp) alloc_blocks >>= 4; - check_writeio: if (alloc_blocks < mp->m_writeio_blocks) alloc_blocks = mp->m_writeio_blocks; - trace_xfs_iomap_prealloc_size(ip, alloc_blocks, shift, mp->m_writeio_blocks); - return alloc_blocks; } -int -xfs_iomap_write_delay( - xfs_inode_t *ip, - xfs_off_t offset, - size_t count, - xfs_bmbt_irec_t *ret_imap) +static int +xfs_file_iomap_begin_delay( + struct inode *inode, + loff_t offset, + loff_t count, + unsigned flags, + struct iomap *iomap) { - xfs_mount_t *mp = ip->i_mount; - xfs_fileoff_t offset_fsb; - xfs_fileoff_t last_fsb; - xfs_off_t aligned_offset; - xfs_fileoff_t ioalign; - xfs_extlen_t extsz; - int nimaps; - xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS]; - int prealloc; - int error; + struct xfs_inode *ip = XFS_I(inode); + struct xfs_mount *mp = ip->i_mount; + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK); + xfs_fileoff_t offset_fsb = XFS_B_TO_FSBT(mp, offset); + xfs_fileoff_t maxbytes_fsb = + XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes); + xfs_fileoff_t end_fsb; + int error = 0, eof = 0; + struct xfs_bmbt_irec got; + xfs_extnum_t idx; + xfs_fsblock_t prealloc_blocks = 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + ASSERT(!XFS_IS_REALTIME_INODE(ip)); + ASSERT(!xfs_get_extsz_hint(ip)); - /* - * Make sure that the dquots are there. This doesn't hold - * the ilock across a disk read. - */ - error = xfs_qm_dqattach_locked(ip, 0); - if (error) - return error; + xfs_ilock(ip, XFS_ILOCK_EXCL); - extsz = xfs_get_extsz_hint(ip); - offset_fsb = XFS_B_TO_FSBT(mp, offset); + if (unlikely(XFS_TEST_ERROR( + (XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_EXTENTS && + XFS_IFORK_FORMAT(ip, XFS_DATA_FORK) != XFS_DINODE_FMT_BTREE), + mp, XFS_ERRTAG_BMAPIFORMAT, XFS_RANDOM_BMAPIFORMAT))) { + XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, mp); + error = -EFSCORRUPTED; + goto out_unlock; + } - error = xfs_iomap_eof_want_preallocate(mp, ip, offset, count, - imap, XFS_WRITE_IMAPS, &prealloc); - if (error) - return error; + XFS_STATS_INC(mp, xs_blk_mapw); -retry: - if (prealloc) { - xfs_fsblock_t alloc_blocks; + if (!(ifp->if_flags & XFS_IFEXTENTS)) { + error = xfs_iread_extents(NULL, ip, XFS_DATA_FORK); + if (error) + goto out_unlock; + } - alloc_blocks = xfs_iomap_prealloc_size(mp, ip, offset, imap, - XFS_WRITE_IMAPS); + eof = !xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got); + if (!eof && got.br_startoff <= offset_fsb) { + if (xfs_is_reflink_inode(ip)) { + bool shared; - aligned_offset = XFS_WRITEIO_ALIGN(mp, (offset + count - 1)); - ioalign = XFS_B_TO_FSBT(mp, aligned_offset); - last_fsb = ioalign + alloc_blocks; - } else { - last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count))); - } + end_fsb = min(XFS_B_TO_FSB(mp, offset + count), + maxbytes_fsb); + xfs_trim_extent(&got, offset_fsb, end_fsb - offset_fsb); + error = xfs_reflink_reserve_cow(ip, &got, &shared); + if (error) + goto out_unlock; + } - if (prealloc || extsz) { - error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb); - if (error) - return error; + trace_xfs_iomap_found(ip, offset, count, 0, &got); + goto done; } + error = xfs_qm_dqattach_locked(ip, 0); + if (error) + goto out_unlock; + /* - * Make sure preallocation does not create extents beyond the range we - * actually support in this filesystem. + * We cap the maximum length we map here to MAX_WRITEBACK_PAGES pages + * to keep the chunks of work done where somewhat symmetric with the + * work writeback does. This is a completely arbitrary number pulled + * out of thin air as a best guess for initial testing. + * + * Note that the values needs to be less than 32-bits wide until + * the lower level functions are updated. */ - if (last_fsb > XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes)) - last_fsb = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes); - - ASSERT(last_fsb > offset_fsb); + count = min_t(loff_t, count, 1024 * PAGE_SIZE); + end_fsb = min(XFS_B_TO_FSB(mp, offset + count), maxbytes_fsb); + + if (eof) { + prealloc_blocks = xfs_iomap_prealloc_size(ip, offset, count, idx); + if (prealloc_blocks) { + xfs_extlen_t align; + xfs_off_t end_offset; + xfs_fileoff_t p_end_fsb; + + end_offset = XFS_WRITEIO_ALIGN(mp, offset + count - 1); + p_end_fsb = XFS_B_TO_FSBT(mp, end_offset) + + prealloc_blocks; + + align = xfs_eof_alignment(ip, 0); + if (align) + p_end_fsb = roundup_64(p_end_fsb, align); + + p_end_fsb = min(p_end_fsb, maxbytes_fsb); + ASSERT(p_end_fsb > offset_fsb); + prealloc_blocks = p_end_fsb - end_fsb; + } + } - nimaps = XFS_WRITE_IMAPS; - error = xfs_bmapi_delay(ip, offset_fsb, last_fsb - offset_fsb, - imap, &nimaps, XFS_BMAPI_ENTIRE); +retry: + error = xfs_bmapi_reserve_delalloc(ip, XFS_DATA_FORK, offset_fsb, + end_fsb - offset_fsb, prealloc_blocks, &got, &idx, eof); switch (error) { case 0: + break; case -ENOSPC: case -EDQUOT: - break; - default: - return error; - } - - /* - * If bmapi returned us nothing, we got either ENOSPC or EDQUOT. Retry - * without EOF preallocation. - */ - if (nimaps == 0) { + /* retry without any preallocation */ trace_xfs_delalloc_enospc(ip, offset, count); - if (prealloc) { - prealloc = 0; - error = 0; + if (prealloc_blocks) { + prealloc_blocks = 0; goto retry; } - return error ? error : -ENOSPC; + /*FALLTHRU*/ + default: + goto out_unlock; } - if (!(imap[0].br_startblock || XFS_IS_REALTIME_INODE(ip))) - return xfs_alert_fsblock_zero(ip, &imap[0]); + trace_xfs_iomap_alloc(ip, offset, count, 0, &got); +done: + if (isnullstartblock(got.br_startblock)) + got.br_startblock = DELAYSTARTBLOCK; - /* - * Tag the inode as speculatively preallocated so we can reclaim this - * space on demand, if necessary. - */ - if (prealloc) - xfs_inode_set_eofblocks_tag(ip); + if (!got.br_startblock) { + error = xfs_alert_fsblock_zero(ip, &got); + if (error) + goto out_unlock; + } - *ret_imap = imap[0]; - return 0; + xfs_bmbt_to_iomap(ip, iomap, &got); + +out_unlock: + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return error; } /* @@ -679,6 +668,7 @@ retry: int xfs_iomap_write_allocate( xfs_inode_t *ip, + int whichfork, xfs_off_t offset, xfs_bmbt_irec_t *imap) { @@ -691,8 +681,12 @@ xfs_iomap_write_allocate( xfs_trans_t *tp; int nimaps; int error = 0; + int flags = 0; int nres; + if (whichfork == XFS_COW_FORK) + flags |= XFS_BMAPI_COWFORK; + /* * Make sure that the dquots are there. */ @@ -786,7 +780,7 @@ xfs_iomap_write_allocate( * pointer that the caller gave to us. */ error = xfs_bmapi_write(tp, ip, map_start_fsb, - count_fsb, 0, &first_block, + count_fsb, flags, &first_block, nres, imap, &nimaps, &dfops); if (error) @@ -947,37 +941,26 @@ error_on_bmapi_transaction: return error; } -void -xfs_bmbt_to_iomap( - struct xfs_inode *ip, - struct iomap *iomap, - struct xfs_bmbt_irec *imap) +static inline bool imap_needs_alloc(struct inode *inode, + struct xfs_bmbt_irec *imap, int nimaps) { - struct xfs_mount *mp = ip->i_mount; - - if (imap->br_startblock == HOLESTARTBLOCK) { - iomap->blkno = IOMAP_NULL_BLOCK; - iomap->type = IOMAP_HOLE; - } else if (imap->br_startblock == DELAYSTARTBLOCK) { - iomap->blkno = IOMAP_NULL_BLOCK; - iomap->type = IOMAP_DELALLOC; - } else { - iomap->blkno = xfs_fsb_to_db(ip, imap->br_startblock); - if (imap->br_state == XFS_EXT_UNWRITTEN) - iomap->type = IOMAP_UNWRITTEN; - else - iomap->type = IOMAP_MAPPED; - } - iomap->offset = XFS_FSB_TO_B(mp, imap->br_startoff); - iomap->length = XFS_FSB_TO_B(mp, imap->br_blockcount); - iomap->bdev = xfs_find_bdev_for_inode(VFS_I(ip)); + return !nimaps || + imap->br_startblock == HOLESTARTBLOCK || + imap->br_startblock == DELAYSTARTBLOCK || + (IS_DAX(inode) && ISUNWRITTEN(imap)); } -static inline bool imap_needs_alloc(struct xfs_bmbt_irec *imap, int nimaps) +static inline bool need_excl_ilock(struct xfs_inode *ip, unsigned flags) { - return !nimaps || - imap->br_startblock == HOLESTARTBLOCK || - imap->br_startblock == DELAYSTARTBLOCK; + /* + * COW writes will allocate delalloc space, so we need to make sure + * to take the lock exclusively here. + */ + if (xfs_is_reflink_inode(ip) && (flags & (IOMAP_WRITE | IOMAP_ZERO))) + return true; + if ((flags & IOMAP_DIRECT) && (flags & IOMAP_WRITE)) + return true; + return false; } static int @@ -993,11 +976,25 @@ xfs_file_iomap_begin( struct xfs_bmbt_irec imap; xfs_fileoff_t offset_fsb, end_fsb; int nimaps = 1, error = 0; + bool shared = false, trimmed = false; + unsigned lockmode; if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; - xfs_ilock(ip, XFS_ILOCK_EXCL); + if (((flags & (IOMAP_WRITE | IOMAP_DIRECT)) == IOMAP_WRITE) && + !IS_DAX(inode) && !xfs_get_extsz_hint(ip)) { + /* Reserve delalloc blocks for regular writeback. */ + return xfs_file_iomap_begin_delay(inode, offset, length, flags, + iomap); + } + + if (need_excl_ilock(ip, flags)) { + lockmode = XFS_ILOCK_EXCL; + xfs_ilock(ip, XFS_ILOCK_EXCL); + } else { + lockmode = xfs_ilock_data_map_shared(ip); + } ASSERT(offset <= mp->m_super->s_maxbytes); if ((xfs_fsize_t)offset + length > mp->m_super->s_maxbytes) @@ -1005,14 +1002,53 @@ xfs_file_iomap_begin( offset_fsb = XFS_B_TO_FSBT(mp, offset); end_fsb = XFS_B_TO_FSB(mp, offset + length); + if (xfs_is_reflink_inode(ip) && + (flags & IOMAP_WRITE) && (flags & IOMAP_DIRECT)) { + shared = xfs_reflink_find_cow_mapping(ip, offset, &imap); + if (shared) { + xfs_iunlock(ip, lockmode); + goto alloc_done; + } + ASSERT(!isnullstartblock(imap.br_startblock)); + } + error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, &imap, - &nimaps, XFS_BMAPI_ENTIRE); - if (error) { - xfs_iunlock(ip, XFS_ILOCK_EXCL); - return error; + &nimaps, 0); + if (error) + goto out_unlock; + + if ((flags & IOMAP_REPORT) || + (xfs_is_reflink_inode(ip) && + (flags & IOMAP_WRITE) && (flags & IOMAP_DIRECT))) { + /* Trim the mapping to the nearest shared extent boundary. */ + error = xfs_reflink_trim_around_shared(ip, &imap, &shared, + &trimmed); + if (error) + goto out_unlock; + + /* + * We're here because we're trying to do a directio write to a + * region that isn't aligned to a filesystem block. If the + * extent is shared, fall back to buffered mode to handle the + * RMW. + */ + if (!(flags & IOMAP_REPORT) && shared) { + trace_xfs_reflink_bounce_dio_write(ip, &imap); + error = -EREMCHG; + goto out_unlock; + } } - if ((flags & IOMAP_WRITE) && imap_needs_alloc(&imap, nimaps)) { + if ((flags & (IOMAP_WRITE | IOMAP_ZERO)) && xfs_is_reflink_inode(ip)) { + error = xfs_reflink_reserve_cow(ip, &imap, &shared); + if (error) + goto out_unlock; + + end_fsb = imap.br_startoff + imap.br_blockcount; + length = XFS_FSB_TO_B(mp, end_fsb) - offset; + } + + if ((flags & IOMAP_WRITE) && imap_needs_alloc(inode, &imap, nimaps)) { /* * We cap the maximum length we map here to MAX_WRITEBACK_PAGES * pages to keep the chunks of work done where somewhat symmetric @@ -1024,32 +1060,34 @@ xfs_file_iomap_begin( * the lower level functions are updated. */ length = min_t(loff_t, length, 1024 * PAGE_SIZE); - if (xfs_get_extsz_hint(ip)) { - /* - * xfs_iomap_write_direct() expects the shared lock. It - * is unlocked on return. - */ - xfs_ilock_demote(ip, XFS_ILOCK_EXCL); - error = xfs_iomap_write_direct(ip, offset, length, &imap, - nimaps); - } else { - error = xfs_iomap_write_delay(ip, offset, length, &imap); - xfs_iunlock(ip, XFS_ILOCK_EXCL); - } - + /* + * xfs_iomap_write_direct() expects the shared lock. It + * is unlocked on return. + */ + if (lockmode == XFS_ILOCK_EXCL) + xfs_ilock_demote(ip, lockmode); + error = xfs_iomap_write_direct(ip, offset, length, &imap, + nimaps); if (error) return error; +alloc_done: + iomap->flags = IOMAP_F_NEW; trace_xfs_iomap_alloc(ip, offset, length, 0, &imap); } else { ASSERT(nimaps); - xfs_iunlock(ip, XFS_ILOCK_EXCL); + xfs_iunlock(ip, lockmode); trace_xfs_iomap_found(ip, offset, length, 0, &imap); } xfs_bmbt_to_iomap(ip, iomap, &imap); + if (shared) + iomap->flags |= IOMAP_F_SHARED; return 0; +out_unlock: + xfs_iunlock(ip, lockmode); + return error; } static int diff --git a/fs/xfs/xfs_iomap.h b/fs/xfs/xfs_iomap.h index fb8aca3d69ab..6d45cf01fcff 100644 --- a/fs/xfs/xfs_iomap.h +++ b/fs/xfs/xfs_iomap.h @@ -25,14 +25,13 @@ struct xfs_bmbt_irec; int xfs_iomap_write_direct(struct xfs_inode *, xfs_off_t, size_t, struct xfs_bmbt_irec *, int); -int xfs_iomap_write_delay(struct xfs_inode *, xfs_off_t, size_t, - struct xfs_bmbt_irec *); -int xfs_iomap_write_allocate(struct xfs_inode *, xfs_off_t, +int xfs_iomap_write_allocate(struct xfs_inode *, int, xfs_off_t, struct xfs_bmbt_irec *); int xfs_iomap_write_unwritten(struct xfs_inode *, xfs_off_t, xfs_off_t); void xfs_bmbt_to_iomap(struct xfs_inode *, struct iomap *, struct xfs_bmbt_irec *); +xfs_extlen_t xfs_eof_alignment(struct xfs_inode *ip, xfs_extlen_t extsize); extern struct iomap_ops xfs_iomap_ops; extern struct iomap_ops xfs_xattr_iomap_ops; diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index b24c3102fa93..308bebb6dfd2 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -542,6 +542,28 @@ xfs_setattr_time( inode->i_mtime = iattr->ia_mtime; } +static int +xfs_vn_change_ok( + struct dentry *dentry, + struct iattr *iattr) +{ + struct xfs_mount *mp = XFS_I(d_inode(dentry))->i_mount; + + if (mp->m_flags & XFS_MOUNT_RDONLY) + return -EROFS; + + if (XFS_FORCED_SHUTDOWN(mp)) + return -EIO; + + return setattr_prepare(dentry, iattr); +} + +/* + * Set non-size attributes of an inode. + * + * Caution: The caller of this function is responsible for calling + * setattr_prepare() or otherwise verifying the change is fine. + */ int xfs_setattr_nonsize( struct xfs_inode *ip, @@ -558,21 +580,6 @@ xfs_setattr_nonsize( struct xfs_dquot *udqp = NULL, *gdqp = NULL; struct xfs_dquot *olddquot1 = NULL, *olddquot2 = NULL; - trace_xfs_setattr(ip); - - /* If acls are being inherited, we already have this checked */ - if (!(flags & XFS_ATTR_NOACL)) { - if (mp->m_flags & XFS_MOUNT_RDONLY) - return -EROFS; - - if (XFS_FORCED_SHUTDOWN(mp)) - return -EIO; - - error = inode_change_ok(inode, iattr); - if (error) - return error; - } - ASSERT((mask & ATTR_SIZE) == 0); /* @@ -743,8 +750,27 @@ out_dqrele: return error; } +int +xfs_vn_setattr_nonsize( + struct dentry *dentry, + struct iattr *iattr) +{ + struct xfs_inode *ip = XFS_I(d_inode(dentry)); + int error; + + trace_xfs_setattr(ip); + + error = xfs_vn_change_ok(dentry, iattr); + if (error) + return error; + return xfs_setattr_nonsize(ip, iattr, 0); +} + /* * Truncate file. Must have write permission and not be a directory. + * + * Caution: The caller of this function is responsible for calling + * setattr_prepare() or otherwise verifying the change is fine. */ int xfs_setattr_size( @@ -759,18 +785,6 @@ xfs_setattr_size( uint lock_flags = 0; bool did_zeroing = false; - trace_xfs_setattr(ip); - - if (mp->m_flags & XFS_MOUNT_RDONLY) - return -EROFS; - - if (XFS_FORCED_SHUTDOWN(mp)) - return -EIO; - - error = inode_change_ok(inode, iattr); - if (error) - return error; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL)); ASSERT(S_ISREG(inode->i_mode)); @@ -882,7 +896,7 @@ xfs_setattr_size( if (newsize != oldsize && !(iattr->ia_valid & (ATTR_CTIME | ATTR_MTIME))) { iattr->ia_ctime = iattr->ia_mtime = - current_fs_time(inode->i_sb); + current_time(inode); iattr->ia_valid |= ATTR_CTIME | ATTR_MTIME; } @@ -942,28 +956,42 @@ out_trans_cancel: goto out_unlock; } +int +xfs_vn_setattr_size( + struct dentry *dentry, + struct iattr *iattr) +{ + struct xfs_inode *ip = XFS_I(d_inode(dentry)); + int error; + + trace_xfs_setattr(ip); + + error = xfs_vn_change_ok(dentry, iattr); + if (error) + return error; + return xfs_setattr_size(ip, iattr); +} + STATIC int xfs_vn_setattr( struct dentry *dentry, struct iattr *iattr) { - struct xfs_inode *ip = XFS_I(d_inode(dentry)); int error; if (iattr->ia_valid & ATTR_SIZE) { - uint iolock = XFS_IOLOCK_EXCL; + struct xfs_inode *ip = XFS_I(d_inode(dentry)); + uint iolock = XFS_IOLOCK_EXCL; - xfs_ilock(ip, iolock); - error = xfs_break_layouts(d_inode(dentry), &iolock, true); - if (!error) { - xfs_ilock(ip, XFS_MMAPLOCK_EXCL); - iolock |= XFS_MMAPLOCK_EXCL; + error = xfs_break_layouts(d_inode(dentry), &iolock); + if (error) + return error; - error = xfs_setattr_size(ip, iattr); - } - xfs_iunlock(ip, iolock); + xfs_ilock(ip, XFS_MMAPLOCK_EXCL); + error = xfs_vn_setattr_size(dentry, iattr); + xfs_iunlock(ip, XFS_MMAPLOCK_EXCL); } else { - error = xfs_setattr_nonsize(ip, iattr, 0); + error = xfs_vn_setattr_nonsize(dentry, iattr); } return error; @@ -1036,9 +1064,6 @@ static const struct inode_operations xfs_inode_operations = { .set_acl = xfs_set_acl, .getattr = xfs_vn_getattr, .setattr = xfs_vn_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .fiemap = xfs_vn_fiemap, .update_time = xfs_vn_update_time, @@ -1059,14 +1084,11 @@ static const struct inode_operations xfs_dir_inode_operations = { */ .rmdir = xfs_vn_unlink, .mknod = xfs_vn_mknod, - .rename2 = xfs_vn_rename, + .rename = xfs_vn_rename, .get_acl = xfs_get_acl, .set_acl = xfs_set_acl, .getattr = xfs_vn_getattr, .setattr = xfs_vn_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, .tmpfile = xfs_vn_tmpfile, @@ -1087,39 +1109,28 @@ static const struct inode_operations xfs_dir_ci_inode_operations = { */ .rmdir = xfs_vn_unlink, .mknod = xfs_vn_mknod, - .rename2 = xfs_vn_rename, + .rename = xfs_vn_rename, .get_acl = xfs_get_acl, .set_acl = xfs_set_acl, .getattr = xfs_vn_getattr, .setattr = xfs_vn_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, .tmpfile = xfs_vn_tmpfile, }; static const struct inode_operations xfs_symlink_inode_operations = { - .readlink = generic_readlink, .get_link = xfs_vn_get_link, .getattr = xfs_vn_getattr, .setattr = xfs_vn_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, }; static const struct inode_operations xfs_inline_symlink_inode_operations = { - .readlink = generic_readlink, .get_link = xfs_vn_get_link_inline, .getattr = xfs_vn_getattr, .setattr = xfs_vn_setattr, - .setxattr = generic_setxattr, - .getxattr = generic_getxattr, - .removexattr = generic_removexattr, .listxattr = xfs_vn_listxattr, .update_time = xfs_vn_update_time, }; @@ -1144,6 +1155,7 @@ xfs_diflags_to_iflags( inode->i_flags |= S_NOATIME; if (S_ISREG(inode->i_mode) && ip->i_mount->m_sb.sb_blocksize == PAGE_SIZE && + !xfs_is_reflink_inode(ip) && (ip->i_mount->m_flags & XFS_MOUNT_DAX || ip->i_d.di_flags2 & XFS_DIFLAG2_DAX)) inode->i_flags |= S_DAX; diff --git a/fs/xfs/xfs_iops.h b/fs/xfs/xfs_iops.h index a0f84abb0d09..0259a383721a 100644 --- a/fs/xfs/xfs_iops.h +++ b/fs/xfs/xfs_iops.h @@ -33,6 +33,7 @@ extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size); extern void xfs_setattr_time(struct xfs_inode *ip, struct iattr *iattr); extern int xfs_setattr_nonsize(struct xfs_inode *ip, struct iattr *vap, int flags); -extern int xfs_setattr_size(struct xfs_inode *ip, struct iattr *vap); +extern int xfs_vn_setattr_nonsize(struct dentry *dentry, struct iattr *vap); +extern int xfs_vn_setattr_size(struct dentry *dentry, struct iattr *vap); #endif /* __XFS_IOPS_H__ */ diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index ce73eb34620d..66e881790c17 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -66,7 +66,7 @@ xfs_bulkstat_one_int( if (!buffer || xfs_internal_inum(mp, ino)) return -EINVAL; - buf = kmem_alloc(sizeof(*buf), KM_SLEEP | KM_MAYFAIL); + buf = kmem_zalloc(sizeof(*buf), KM_SLEEP | KM_MAYFAIL); if (!buf) return -ENOMEM; @@ -111,6 +111,12 @@ xfs_bulkstat_one_int( buf->bs_aextents = dic->di_anextents; buf->bs_forkoff = XFS_IFORK_BOFF(ip); + if (dic->di_version == 3) { + if (dic->di_flags2 & XFS_DIFLAG2_COWEXTSIZE) + buf->bs_cowextsize = dic->di_cowextsize << + mp->m_sb.sb_blocklog; + } + switch (dic->di_format) { case XFS_DINODE_FMT_DEV: buf->bs_rdev = ip->i_df.if_u2.if_rdev; diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h index b8d64d520e12..e467218c0098 100644 --- a/fs/xfs/xfs_linux.h +++ b/fs/xfs/xfs_linux.h @@ -78,11 +78,12 @@ typedef __u32 xfs_nlink_t; #include <linux/freezer.h> #include <linux/list_sort.h> #include <linux/ratelimit.h> +#include <linux/rhashtable.h> #include <asm/page.h> #include <asm/div64.h> #include <asm/param.h> -#include <asm/uaccess.h> +#include <linux/uaccess.h> #include <asm/byteorder.h> #include <asm/unaligned.h> @@ -116,6 +117,7 @@ typedef __u32 xfs_nlink_t; #define xfs_inherit_nodefrag xfs_params.inherit_nodfrg.val #define xfs_fstrm_centisecs xfs_params.fstrm_timer.val #define xfs_eofb_secs xfs_params.eofb_timer.val +#define xfs_cowb_secs xfs_params.cowb_timer.val #define current_cpu() (raw_smp_processor_id()) #define current_pid() (current->pid) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 3b74fa011bb1..c39ac14ff540 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -1668,7 +1668,7 @@ xlog_cksum( __uint32_t crc; /* first generate the crc for the record header ... */ - crc = xfs_start_cksum((char *)rhead, + crc = xfs_start_cksum_update((char *)rhead, sizeof(struct xlog_rec_header), offsetof(struct xlog_rec_header, h_crc)); @@ -1862,26 +1862,21 @@ xlog_sync( bp->b_io_length = BTOBB(count); bp->b_fspriv = iclog; - bp->b_flags &= ~(XBF_FUA | XBF_FLUSH); - bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE); + bp->b_flags &= ~XBF_FLUSH; + bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA); - if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) { - bp->b_flags |= XBF_FUA; - - /* - * Flush the data device before flushing the log to make - * sure all meta data written back from the AIL actually made - * it to disk before stamping the new log tail LSN into the - * log buffer. For an external log we need to issue the - * flush explicitly, and unfortunately synchronously here; - * for an internal log we can simply use the block layer - * state machine for preflushes. - */ - if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp) - xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp); - else - bp->b_flags |= XBF_FLUSH; - } + /* + * Flush the data device before flushing the log to make sure all meta + * data written back from the AIL actually made it to disk before + * stamping the new log tail LSN into the log buffer. For an external + * log we need to issue the flush explicitly, and unfortunately + * synchronously here; for an internal log we can simply use the block + * layer state machine for preflushes. + */ + if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp) + xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp); + else + bp->b_flags |= XBF_FLUSH; ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1); ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize); @@ -1906,10 +1901,8 @@ xlog_sync( xfs_buf_associate_memory(bp, (char *)&iclog->ic_header + count, split); bp->b_fspriv = iclog; - bp->b_flags &= ~(XBF_FUA | XBF_FLUSH); - bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE); - if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) - bp->b_flags |= XBF_FUA; + bp->b_flags &= ~XBF_FLUSH; + bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA); ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1); ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize); diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 765f084759b5..2b6eec52178e 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -413,7 +413,8 @@ struct xlog { /* log record crc error injection factor */ uint32_t l_badcrc_factor; #endif - + /* log recovery lsn tracking (for buffer submission */ + xfs_lsn_t l_recovery_lsn; }; #define XLOG_BUF_CANCEL_BUCKET(log, blkno) \ diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index e8638fd2c0c3..4a98762ec8b4 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -44,6 +44,9 @@ #include "xfs_error.h" #include "xfs_dir2.h" #include "xfs_rmap_item.h" +#include "xfs_buf_item.h" +#include "xfs_refcount_item.h" +#include "xfs_bmap_item.h" #define BLK_AVG(blk1, blk2) ((blk1+blk2) >> 1) @@ -381,6 +384,15 @@ xlog_recover_iodone( SHUTDOWN_META_IO_ERROR); } } + + /* + * On v5 supers, a bli could be attached to update the metadata LSN. + * Clean it up. + */ + if (bp->b_fspriv) + xfs_buf_item_relse(bp); + ASSERT(bp->b_fspriv == NULL); + bp->b_iodone = NULL; xfs_buf_ioend(bp); } @@ -1914,6 +1926,10 @@ xlog_recover_reorder_trans( case XFS_LI_EFI: case XFS_LI_RUI: case XFS_LI_RUD: + case XFS_LI_CUI: + case XFS_LI_CUD: + case XFS_LI_BUI: + case XFS_LI_BUD: trace_xfs_log_recover_item_reorder_tail(log, trans, item, pass); list_move_tail(&item->ri_list, &inode_list); @@ -2009,7 +2025,7 @@ xlog_peek_buffer_cancelled( struct xlog *log, xfs_daddr_t blkno, uint len, - ushort flags) + unsigned short flags) { struct list_head *bucket; struct xfs_buf_cancel *bcp; @@ -2049,7 +2065,7 @@ xlog_check_buffer_cancelled( struct xlog *log, xfs_daddr_t blkno, uint len, - ushort flags) + unsigned short flags) { struct xfs_buf_cancel *bcp; @@ -2232,6 +2248,7 @@ xlog_recover_get_buf_lsn( case XFS_ABTB_MAGIC: case XFS_ABTC_MAGIC: case XFS_RMAP_CRC_MAGIC: + case XFS_REFC_CRC_MAGIC: case XFS_IBT_CRC_MAGIC: case XFS_IBT_MAGIC: { struct xfs_btree_block *btb = blk; @@ -2360,12 +2377,14 @@ static void xlog_recover_validate_buf_type( struct xfs_mount *mp, struct xfs_buf *bp, - xfs_buf_log_format_t *buf_f) + xfs_buf_log_format_t *buf_f, + xfs_lsn_t current_lsn) { struct xfs_da_blkinfo *info = bp->b_addr; __uint32_t magic32; __uint16_t magic16; __uint16_t magicda; + char *warnmsg = NULL; /* * We can only do post recovery validation on items on CRC enabled @@ -2403,32 +2422,31 @@ xlog_recover_validate_buf_type( case XFS_RMAP_CRC_MAGIC: bp->b_ops = &xfs_rmapbt_buf_ops; break; + case XFS_REFC_CRC_MAGIC: + bp->b_ops = &xfs_refcountbt_buf_ops; + break; default: - xfs_warn(mp, "Bad btree block magic!"); - ASSERT(0); + warnmsg = "Bad btree block magic!"; break; } break; case XFS_BLFT_AGF_BUF: if (magic32 != XFS_AGF_MAGIC) { - xfs_warn(mp, "Bad AGF block magic!"); - ASSERT(0); + warnmsg = "Bad AGF block magic!"; break; } bp->b_ops = &xfs_agf_buf_ops; break; case XFS_BLFT_AGFL_BUF: if (magic32 != XFS_AGFL_MAGIC) { - xfs_warn(mp, "Bad AGFL block magic!"); - ASSERT(0); + warnmsg = "Bad AGFL block magic!"; break; } bp->b_ops = &xfs_agfl_buf_ops; break; case XFS_BLFT_AGI_BUF: if (magic32 != XFS_AGI_MAGIC) { - xfs_warn(mp, "Bad AGI block magic!"); - ASSERT(0); + warnmsg = "Bad AGI block magic!"; break; } bp->b_ops = &xfs_agi_buf_ops; @@ -2438,8 +2456,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_GDQUOT_BUF: #ifdef CONFIG_XFS_QUOTA if (magic16 != XFS_DQUOT_MAGIC) { - xfs_warn(mp, "Bad DQUOT block magic!"); - ASSERT(0); + warnmsg = "Bad DQUOT block magic!"; break; } bp->b_ops = &xfs_dquot_buf_ops; @@ -2451,16 +2468,14 @@ xlog_recover_validate_buf_type( break; case XFS_BLFT_DINO_BUF: if (magic16 != XFS_DINODE_MAGIC) { - xfs_warn(mp, "Bad INODE block magic!"); - ASSERT(0); + warnmsg = "Bad INODE block magic!"; break; } bp->b_ops = &xfs_inode_buf_ops; break; case XFS_BLFT_SYMLINK_BUF: if (magic32 != XFS_SYMLINK_MAGIC) { - xfs_warn(mp, "Bad symlink block magic!"); - ASSERT(0); + warnmsg = "Bad symlink block magic!"; break; } bp->b_ops = &xfs_symlink_buf_ops; @@ -2468,8 +2483,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DIR_BLOCK_BUF: if (magic32 != XFS_DIR2_BLOCK_MAGIC && magic32 != XFS_DIR3_BLOCK_MAGIC) { - xfs_warn(mp, "Bad dir block magic!"); - ASSERT(0); + warnmsg = "Bad dir block magic!"; break; } bp->b_ops = &xfs_dir3_block_buf_ops; @@ -2477,8 +2491,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DIR_DATA_BUF: if (magic32 != XFS_DIR2_DATA_MAGIC && magic32 != XFS_DIR3_DATA_MAGIC) { - xfs_warn(mp, "Bad dir data magic!"); - ASSERT(0); + warnmsg = "Bad dir data magic!"; break; } bp->b_ops = &xfs_dir3_data_buf_ops; @@ -2486,8 +2499,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DIR_FREE_BUF: if (magic32 != XFS_DIR2_FREE_MAGIC && magic32 != XFS_DIR3_FREE_MAGIC) { - xfs_warn(mp, "Bad dir3 free magic!"); - ASSERT(0); + warnmsg = "Bad dir3 free magic!"; break; } bp->b_ops = &xfs_dir3_free_buf_ops; @@ -2495,8 +2507,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DIR_LEAF1_BUF: if (magicda != XFS_DIR2_LEAF1_MAGIC && magicda != XFS_DIR3_LEAF1_MAGIC) { - xfs_warn(mp, "Bad dir leaf1 magic!"); - ASSERT(0); + warnmsg = "Bad dir leaf1 magic!"; break; } bp->b_ops = &xfs_dir3_leaf1_buf_ops; @@ -2504,8 +2515,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DIR_LEAFN_BUF: if (magicda != XFS_DIR2_LEAFN_MAGIC && magicda != XFS_DIR3_LEAFN_MAGIC) { - xfs_warn(mp, "Bad dir leafn magic!"); - ASSERT(0); + warnmsg = "Bad dir leafn magic!"; break; } bp->b_ops = &xfs_dir3_leafn_buf_ops; @@ -2513,8 +2523,7 @@ xlog_recover_validate_buf_type( case XFS_BLFT_DA_NODE_BUF: if (magicda != XFS_DA_NODE_MAGIC && magicda != XFS_DA3_NODE_MAGIC) { - xfs_warn(mp, "Bad da node magic!"); - ASSERT(0); + warnmsg = "Bad da node magic!"; break; } bp->b_ops = &xfs_da3_node_buf_ops; @@ -2522,24 +2531,21 @@ xlog_recover_validate_buf_type( case XFS_BLFT_ATTR_LEAF_BUF: if (magicda != XFS_ATTR_LEAF_MAGIC && magicda != XFS_ATTR3_LEAF_MAGIC) { - xfs_warn(mp, "Bad attr leaf magic!"); - ASSERT(0); + warnmsg = "Bad attr leaf magic!"; break; } bp->b_ops = &xfs_attr3_leaf_buf_ops; break; case XFS_BLFT_ATTR_RMT_BUF: if (magic32 != XFS_ATTR3_RMT_MAGIC) { - xfs_warn(mp, "Bad attr remote magic!"); - ASSERT(0); + warnmsg = "Bad attr remote magic!"; break; } bp->b_ops = &xfs_attr3_rmt_buf_ops; break; case XFS_BLFT_SB_BUF: if (magic32 != XFS_SB_MAGIC) { - xfs_warn(mp, "Bad SB block magic!"); - ASSERT(0); + warnmsg = "Bad SB block magic!"; break; } bp->b_ops = &xfs_sb_buf_ops; @@ -2556,6 +2562,40 @@ xlog_recover_validate_buf_type( xfs_blft_from_flags(buf_f)); break; } + + /* + * Nothing else to do in the case of a NULL current LSN as this means + * the buffer is more recent than the change in the log and will be + * skipped. + */ + if (current_lsn == NULLCOMMITLSN) + return; + + if (warnmsg) { + xfs_warn(mp, warnmsg); + ASSERT(0); + } + + /* + * We must update the metadata LSN of the buffer as it is written out to + * ensure that older transactions never replay over this one and corrupt + * the buffer. This can occur if log recovery is interrupted at some + * point after the current transaction completes, at which point a + * subsequent mount starts recovery from the beginning. + * + * Write verifiers update the metadata LSN from log items attached to + * the buffer. Therefore, initialize a bli purely to carry the LSN to + * the verifier. We'll clean it up in our ->iodone() callback. + */ + if (bp->b_ops) { + struct xfs_buf_log_item *bip; + + ASSERT(!bp->b_iodone || bp->b_iodone == xlog_recover_iodone); + bp->b_iodone = xlog_recover_iodone; + xfs_buf_item_init(bp, mp); + bip = bp->b_fspriv; + bip->bli_item.li_lsn = current_lsn; + } } /* @@ -2569,7 +2609,8 @@ xlog_recover_do_reg_buffer( struct xfs_mount *mp, xlog_recover_item_t *item, struct xfs_buf *bp, - xfs_buf_log_format_t *buf_f) + xfs_buf_log_format_t *buf_f, + xfs_lsn_t current_lsn) { int i; int bit; @@ -2642,7 +2683,7 @@ xlog_recover_do_reg_buffer( /* Shouldn't be any more regions */ ASSERT(i == item->ri_total); - xlog_recover_validate_buf_type(mp, bp, buf_f); + xlog_recover_validate_buf_type(mp, bp, buf_f, current_lsn); } /* @@ -2685,7 +2726,7 @@ xlog_recover_do_dquot_buffer( if (log->l_quotaoffs_flag & type) return false; - xlog_recover_do_reg_buffer(mp, item, bp, buf_f); + xlog_recover_do_reg_buffer(mp, item, bp, buf_f, NULLCOMMITLSN); return true; } @@ -2773,7 +2814,8 @@ xlog_recover_buffer_pass2( */ lsn = xlog_recover_get_buf_lsn(mp, bp); if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) { - xlog_recover_validate_buf_type(mp, bp, buf_f); + trace_xfs_log_recover_buf_skip(log, buf_f); + xlog_recover_validate_buf_type(mp, bp, buf_f, NULLCOMMITLSN); goto out_release; } @@ -2789,7 +2831,7 @@ xlog_recover_buffer_pass2( if (!dirty) goto out_release; } else { - xlog_recover_do_reg_buffer(mp, item, bp, buf_f); + xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn); } /* @@ -3515,6 +3557,242 @@ xlog_recover_rud_pass2( } /* + * Copy an CUI format buffer from the given buf, and into the destination + * CUI format structure. The CUI/CUD items were designed not to need any + * special alignment handling. + */ +static int +xfs_cui_copy_format( + struct xfs_log_iovec *buf, + struct xfs_cui_log_format *dst_cui_fmt) +{ + struct xfs_cui_log_format *src_cui_fmt; + uint len; + + src_cui_fmt = buf->i_addr; + len = xfs_cui_log_format_sizeof(src_cui_fmt->cui_nextents); + + if (buf->i_len == len) { + memcpy(dst_cui_fmt, src_cui_fmt, len); + return 0; + } + return -EFSCORRUPTED; +} + +/* + * This routine is called to create an in-core extent refcount update + * item from the cui format structure which was logged on disk. + * It allocates an in-core cui, copies the extents from the format + * structure into it, and adds the cui to the AIL with the given + * LSN. + */ +STATIC int +xlog_recover_cui_pass2( + struct xlog *log, + struct xlog_recover_item *item, + xfs_lsn_t lsn) +{ + int error; + struct xfs_mount *mp = log->l_mp; + struct xfs_cui_log_item *cuip; + struct xfs_cui_log_format *cui_formatp; + + cui_formatp = item->ri_buf[0].i_addr; + + cuip = xfs_cui_init(mp, cui_formatp->cui_nextents); + error = xfs_cui_copy_format(&item->ri_buf[0], &cuip->cui_format); + if (error) { + xfs_cui_item_free(cuip); + return error; + } + atomic_set(&cuip->cui_next_extent, cui_formatp->cui_nextents); + + spin_lock(&log->l_ailp->xa_lock); + /* + * The CUI has two references. One for the CUD and one for CUI to ensure + * it makes it into the AIL. Insert the CUI into the AIL directly and + * drop the CUI reference. Note that xfs_trans_ail_update() drops the + * AIL lock. + */ + xfs_trans_ail_update(log->l_ailp, &cuip->cui_item, lsn); + xfs_cui_release(cuip); + return 0; +} + + +/* + * This routine is called when an CUD format structure is found in a committed + * transaction in the log. Its purpose is to cancel the corresponding CUI if it + * was still in the log. To do this it searches the AIL for the CUI with an id + * equal to that in the CUD format structure. If we find it we drop the CUD + * reference, which removes the CUI from the AIL and frees it. + */ +STATIC int +xlog_recover_cud_pass2( + struct xlog *log, + struct xlog_recover_item *item) +{ + struct xfs_cud_log_format *cud_formatp; + struct xfs_cui_log_item *cuip = NULL; + struct xfs_log_item *lip; + __uint64_t cui_id; + struct xfs_ail_cursor cur; + struct xfs_ail *ailp = log->l_ailp; + + cud_formatp = item->ri_buf[0].i_addr; + if (item->ri_buf[0].i_len != sizeof(struct xfs_cud_log_format)) + return -EFSCORRUPTED; + cui_id = cud_formatp->cud_cui_id; + + /* + * Search for the CUI with the id in the CUD format structure in the + * AIL. + */ + spin_lock(&ailp->xa_lock); + lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); + while (lip != NULL) { + if (lip->li_type == XFS_LI_CUI) { + cuip = (struct xfs_cui_log_item *)lip; + if (cuip->cui_format.cui_id == cui_id) { + /* + * Drop the CUD reference to the CUI. This + * removes the CUI from the AIL and frees it. + */ + spin_unlock(&ailp->xa_lock); + xfs_cui_release(cuip); + spin_lock(&ailp->xa_lock); + break; + } + } + lip = xfs_trans_ail_cursor_next(ailp, &cur); + } + + xfs_trans_ail_cursor_done(&cur); + spin_unlock(&ailp->xa_lock); + + return 0; +} + +/* + * Copy an BUI format buffer from the given buf, and into the destination + * BUI format structure. The BUI/BUD items were designed not to need any + * special alignment handling. + */ +static int +xfs_bui_copy_format( + struct xfs_log_iovec *buf, + struct xfs_bui_log_format *dst_bui_fmt) +{ + struct xfs_bui_log_format *src_bui_fmt; + uint len; + + src_bui_fmt = buf->i_addr; + len = xfs_bui_log_format_sizeof(src_bui_fmt->bui_nextents); + + if (buf->i_len == len) { + memcpy(dst_bui_fmt, src_bui_fmt, len); + return 0; + } + return -EFSCORRUPTED; +} + +/* + * This routine is called to create an in-core extent bmap update + * item from the bui format structure which was logged on disk. + * It allocates an in-core bui, copies the extents from the format + * structure into it, and adds the bui to the AIL with the given + * LSN. + */ +STATIC int +xlog_recover_bui_pass2( + struct xlog *log, + struct xlog_recover_item *item, + xfs_lsn_t lsn) +{ + int error; + struct xfs_mount *mp = log->l_mp; + struct xfs_bui_log_item *buip; + struct xfs_bui_log_format *bui_formatp; + + bui_formatp = item->ri_buf[0].i_addr; + + if (bui_formatp->bui_nextents != XFS_BUI_MAX_FAST_EXTENTS) + return -EFSCORRUPTED; + buip = xfs_bui_init(mp); + error = xfs_bui_copy_format(&item->ri_buf[0], &buip->bui_format); + if (error) { + xfs_bui_item_free(buip); + return error; + } + atomic_set(&buip->bui_next_extent, bui_formatp->bui_nextents); + + spin_lock(&log->l_ailp->xa_lock); + /* + * The RUI has two references. One for the RUD and one for RUI to ensure + * it makes it into the AIL. Insert the RUI into the AIL directly and + * drop the RUI reference. Note that xfs_trans_ail_update() drops the + * AIL lock. + */ + xfs_trans_ail_update(log->l_ailp, &buip->bui_item, lsn); + xfs_bui_release(buip); + return 0; +} + + +/* + * This routine is called when an BUD format structure is found in a committed + * transaction in the log. Its purpose is to cancel the corresponding BUI if it + * was still in the log. To do this it searches the AIL for the BUI with an id + * equal to that in the BUD format structure. If we find it we drop the BUD + * reference, which removes the BUI from the AIL and frees it. + */ +STATIC int +xlog_recover_bud_pass2( + struct xlog *log, + struct xlog_recover_item *item) +{ + struct xfs_bud_log_format *bud_formatp; + struct xfs_bui_log_item *buip = NULL; + struct xfs_log_item *lip; + __uint64_t bui_id; + struct xfs_ail_cursor cur; + struct xfs_ail *ailp = log->l_ailp; + + bud_formatp = item->ri_buf[0].i_addr; + if (item->ri_buf[0].i_len != sizeof(struct xfs_bud_log_format)) + return -EFSCORRUPTED; + bui_id = bud_formatp->bud_bui_id; + + /* + * Search for the BUI with the id in the BUD format structure in the + * AIL. + */ + spin_lock(&ailp->xa_lock); + lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); + while (lip != NULL) { + if (lip->li_type == XFS_LI_BUI) { + buip = (struct xfs_bui_log_item *)lip; + if (buip->bui_format.bui_id == bui_id) { + /* + * Drop the BUD reference to the BUI. This + * removes the BUI from the AIL and frees it. + */ + spin_unlock(&ailp->xa_lock); + xfs_bui_release(buip); + spin_lock(&ailp->xa_lock); + break; + } + } + lip = xfs_trans_ail_cursor_next(ailp, &cur); + } + + xfs_trans_ail_cursor_done(&cur); + spin_unlock(&ailp->xa_lock); + + return 0; +} + +/* * This routine is called when an inode create format structure is found in a * committed transaction in the log. It's purpose is to initialise the inodes * being allocated on disk. This requires us to get inode cluster buffers that @@ -3741,6 +4019,10 @@ xlog_recover_ra_pass2( case XFS_LI_QUOTAOFF: case XFS_LI_RUI: case XFS_LI_RUD: + case XFS_LI_CUI: + case XFS_LI_CUD: + case XFS_LI_BUI: + case XFS_LI_BUD: default: break; } @@ -3766,6 +4048,10 @@ xlog_recover_commit_pass1( case XFS_LI_ICREATE: case XFS_LI_RUI: case XFS_LI_RUD: + case XFS_LI_CUI: + case XFS_LI_CUD: + case XFS_LI_BUI: + case XFS_LI_BUD: /* nothing to do in pass 1 */ return 0; default: @@ -3800,6 +4086,14 @@ xlog_recover_commit_pass2( return xlog_recover_rui_pass2(log, item, trans->r_lsn); case XFS_LI_RUD: return xlog_recover_rud_pass2(log, item); + case XFS_LI_CUI: + return xlog_recover_cui_pass2(log, item, trans->r_lsn); + case XFS_LI_CUD: + return xlog_recover_cud_pass2(log, item); + case XFS_LI_BUI: + return xlog_recover_bui_pass2(log, item, trans->r_lsn); + case XFS_LI_BUD: + return xlog_recover_bud_pass2(log, item); case XFS_LI_DQUOT: return xlog_recover_dquot_pass2(log, buffer_list, item, trans->r_lsn); @@ -3846,14 +4140,13 @@ STATIC int xlog_recover_commit_trans( struct xlog *log, struct xlog_recover *trans, - int pass) + int pass, + struct list_head *buffer_list) { int error = 0; - int error2; int items_queued = 0; struct xlog_recover_item *item; struct xlog_recover_item *next; - LIST_HEAD (buffer_list); LIST_HEAD (ra_list); LIST_HEAD (done_list); @@ -3876,7 +4169,7 @@ xlog_recover_commit_trans( items_queued++; if (items_queued >= XLOG_RECOVER_COMMIT_QUEUE_MAX) { error = xlog_recover_items_pass2(log, trans, - &buffer_list, &ra_list); + buffer_list, &ra_list); list_splice_tail_init(&ra_list, &done_list); items_queued = 0; } @@ -3894,15 +4187,14 @@ out: if (!list_empty(&ra_list)) { if (!error) error = xlog_recover_items_pass2(log, trans, - &buffer_list, &ra_list); + buffer_list, &ra_list); list_splice_tail_init(&ra_list, &done_list); } if (!list_empty(&done_list)) list_splice_init(&done_list, &trans->r_itemq); - error2 = xfs_buf_delwri_submit(&buffer_list); - return error ? error : error2; + return error; } STATIC void @@ -4085,7 +4377,8 @@ xlog_recovery_process_trans( char *dp, unsigned int len, unsigned int flags, - int pass) + int pass, + struct list_head *buffer_list) { int error = 0; bool freeit = false; @@ -4109,7 +4402,8 @@ xlog_recovery_process_trans( error = xlog_recover_add_to_cont_trans(log, trans, dp, len); break; case XLOG_COMMIT_TRANS: - error = xlog_recover_commit_trans(log, trans, pass); + error = xlog_recover_commit_trans(log, trans, pass, + buffer_list); /* success or fail, we are now done with this transaction. */ freeit = true; break; @@ -4191,10 +4485,12 @@ xlog_recover_process_ophdr( struct xlog_op_header *ohead, char *dp, char *end, - int pass) + int pass, + struct list_head *buffer_list) { struct xlog_recover *trans; unsigned int len; + int error; /* Do we understand who wrote this op? */ if (ohead->oh_clientid != XFS_TRANSACTION && @@ -4221,8 +4517,39 @@ xlog_recover_process_ophdr( return 0; } + /* + * The recovered buffer queue is drained only once we know that all + * recovery items for the current LSN have been processed. This is + * required because: + * + * - Buffer write submission updates the metadata LSN of the buffer. + * - Log recovery skips items with a metadata LSN >= the current LSN of + * the recovery item. + * - Separate recovery items against the same metadata buffer can share + * a current LSN. I.e., consider that the LSN of a recovery item is + * defined as the starting LSN of the first record in which its + * transaction appears, that a record can hold multiple transactions, + * and/or that a transaction can span multiple records. + * + * In other words, we are allowed to submit a buffer from log recovery + * once per current LSN. Otherwise, we may incorrectly skip recovery + * items and cause corruption. + * + * We don't know up front whether buffers are updated multiple times per + * LSN. Therefore, track the current LSN of each commit log record as it + * is processed and drain the queue when it changes. Use commit records + * because they are ordered correctly by the logging code. + */ + if (log->l_recovery_lsn != trans->r_lsn && + ohead->oh_flags & XLOG_COMMIT_TRANS) { + error = xfs_buf_delwri_submit(buffer_list); + if (error) + return error; + log->l_recovery_lsn = trans->r_lsn; + } + return xlog_recovery_process_trans(log, trans, dp, len, - ohead->oh_flags, pass); + ohead->oh_flags, pass, buffer_list); } /* @@ -4240,7 +4567,8 @@ xlog_recover_process_data( struct hlist_head rhash[], struct xlog_rec_header *rhead, char *dp, - int pass) + int pass, + struct list_head *buffer_list) { struct xlog_op_header *ohead; char *end; @@ -4254,6 +4582,7 @@ xlog_recover_process_data( if (xlog_header_check_recover(log->l_mp, rhead)) return -EIO; + trace_xfs_log_recover_record(log, rhead, pass); while ((dp < end) && num_logops) { ohead = (struct xlog_op_header *)dp; @@ -4262,7 +4591,7 @@ xlog_recover_process_data( /* errors will abort recovery */ error = xlog_recover_process_ophdr(log, rhash, rhead, ohead, - dp, end, pass); + dp, end, pass, buffer_list); if (error) return error; @@ -4352,12 +4681,94 @@ xlog_recover_cancel_rui( spin_lock(&ailp->xa_lock); } +/* Recover the CUI if necessary. */ +STATIC int +xlog_recover_process_cui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_cui_log_item *cuip; + int error; + + /* + * Skip CUIs that we've already processed. + */ + cuip = container_of(lip, struct xfs_cui_log_item, cui_item); + if (test_bit(XFS_CUI_RECOVERED, &cuip->cui_flags)) + return 0; + + spin_unlock(&ailp->xa_lock); + error = xfs_cui_recover(mp, cuip); + spin_lock(&ailp->xa_lock); + + return error; +} + +/* Release the CUI since we're cancelling everything. */ +STATIC void +xlog_recover_cancel_cui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_cui_log_item *cuip; + + cuip = container_of(lip, struct xfs_cui_log_item, cui_item); + + spin_unlock(&ailp->xa_lock); + xfs_cui_release(cuip); + spin_lock(&ailp->xa_lock); +} + +/* Recover the BUI if necessary. */ +STATIC int +xlog_recover_process_bui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_bui_log_item *buip; + int error; + + /* + * Skip BUIs that we've already processed. + */ + buip = container_of(lip, struct xfs_bui_log_item, bui_item); + if (test_bit(XFS_BUI_RECOVERED, &buip->bui_flags)) + return 0; + + spin_unlock(&ailp->xa_lock); + error = xfs_bui_recover(mp, buip); + spin_lock(&ailp->xa_lock); + + return error; +} + +/* Release the BUI since we're cancelling everything. */ +STATIC void +xlog_recover_cancel_bui( + struct xfs_mount *mp, + struct xfs_ail *ailp, + struct xfs_log_item *lip) +{ + struct xfs_bui_log_item *buip; + + buip = container_of(lip, struct xfs_bui_log_item, bui_item); + + spin_unlock(&ailp->xa_lock); + xfs_bui_release(buip); + spin_lock(&ailp->xa_lock); +} + /* Is this log item a deferred action intent? */ static inline bool xlog_item_is_intent(struct xfs_log_item *lip) { switch (lip->li_type) { case XFS_LI_EFI: case XFS_LI_RUI: + case XFS_LI_CUI: + case XFS_LI_BUI: return true; default: return false; @@ -4421,6 +4832,12 @@ xlog_recover_process_intents( case XFS_LI_RUI: error = xlog_recover_process_rui(log->l_mp, ailp, lip); break; + case XFS_LI_CUI: + error = xlog_recover_process_cui(log->l_mp, ailp, lip); + break; + case XFS_LI_BUI: + error = xlog_recover_process_bui(log->l_mp, ailp, lip); + break; } if (error) goto out; @@ -4468,6 +4885,12 @@ xlog_recover_cancel_intents( case XFS_LI_RUI: xlog_recover_cancel_rui(log->l_mp, ailp, lip); break; + case XFS_LI_CUI: + xlog_recover_cancel_cui(log->l_mp, ailp, lip); + break; + case XFS_LI_BUI: + xlog_recover_cancel_bui(log->l_mp, ailp, lip); + break; } lip = xfs_trans_ail_cursor_next(ailp, &cur); @@ -4546,6 +4969,7 @@ xlog_recover_process_one_iunlink( if (error) goto fail_iput; + xfs_iflags_clear(ip, XFS_IRECOVERY); ASSERT(VFS_I(ip)->i_nlink == 0); ASSERT(VFS_I(ip)->i_mode != 0); @@ -4685,22 +5109,25 @@ xlog_recover_process( struct hlist_head rhash[], struct xlog_rec_header *rhead, char *dp, - int pass) + int pass, + struct list_head *buffer_list) { int error; + __le32 old_crc = rhead->h_crc; __le32 crc; + crc = xlog_cksum(log, rhead, dp, be32_to_cpu(rhead->h_len)); /* * Nothing else to do if this is a CRC verification pass. Just return * if this a record with a non-zero crc. Unfortunately, mkfs always - * sets h_crc to 0 so we must consider this valid even on v5 supers. + * sets old_crc to 0 so we must consider this valid even on v5 supers. * Otherwise, return EFSBADCRC on failure so the callers up the stack * know precisely what failed. */ if (pass == XLOG_RECOVER_CRCPASS) { - if (rhead->h_crc && crc != rhead->h_crc) + if (old_crc && crc != old_crc) return -EFSBADCRC; return 0; } @@ -4711,11 +5138,11 @@ xlog_recover_process( * zero CRC check prevents warnings from being emitted when upgrading * the kernel from one that does not add CRCs by default. */ - if (crc != rhead->h_crc) { - if (rhead->h_crc || xfs_sb_version_hascrc(&log->l_mp->m_sb)) { + if (crc != old_crc) { + if (old_crc || xfs_sb_version_hascrc(&log->l_mp->m_sb)) { xfs_alert(log->l_mp, "log record CRC mismatch: found 0x%x, expected 0x%x.", - le32_to_cpu(rhead->h_crc), + le32_to_cpu(old_crc), le32_to_cpu(crc)); xfs_hex_dump(dp, 32); } @@ -4732,7 +5159,8 @@ xlog_recover_process( if (error) return error; - return xlog_recover_process_data(log, rhash, rhead, dp, pass); + return xlog_recover_process_data(log, rhash, rhead, dp, pass, + buffer_list); } STATIC int @@ -4793,9 +5221,11 @@ xlog_do_recovery_pass( char *offset; xfs_buf_t *hbp, *dbp; int error = 0, h_size, h_len; + int error2 = 0; int bblks, split_bblks; int hblks, split_hblks, wrapped_hblks; struct hlist_head rhash[XLOG_RHASH_SIZE]; + LIST_HEAD (buffer_list); ASSERT(head_blk != tail_blk); rhead_blk = 0; @@ -4981,7 +5411,7 @@ xlog_do_recovery_pass( } error = xlog_recover_process(log, rhash, rhead, offset, - pass); + pass, &buffer_list); if (error) goto bread_err2; @@ -5012,7 +5442,8 @@ xlog_do_recovery_pass( if (error) goto bread_err2; - error = xlog_recover_process(log, rhash, rhead, offset, pass); + error = xlog_recover_process(log, rhash, rhead, offset, pass, + &buffer_list); if (error) goto bread_err2; @@ -5025,10 +5456,17 @@ xlog_do_recovery_pass( bread_err1: xlog_put_bp(hbp); + /* + * Submit buffers that have been added from the last record processed, + * regardless of error status. + */ + if (!list_empty(&buffer_list)) + error2 = xfs_buf_delwri_submit(&buffer_list); + if (error && first_bad) *first_bad = rhead_blk; - return error; + return error ? error : error2; } /* diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index faeead671f9f..9b9540db17a6 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -43,6 +43,8 @@ #include "xfs_icache.h" #include "xfs_sysfs.h" #include "xfs_rmap_btree.h" +#include "xfs_refcount_btree.h" +#include "xfs_reflink.h" static DEFINE_MUTEX(xfs_uuid_table_mutex); @@ -155,6 +157,7 @@ xfs_free_perag( spin_unlock(&mp->m_perag_lock); ASSERT(pag); ASSERT(atomic_read(&pag->pag_ref) == 0); + xfs_buf_hash_destroy(pag); call_rcu(&pag->rcu_head, __xfs_free_perag); } } @@ -210,8 +213,8 @@ xfs_initialize_perag( spin_lock_init(&pag->pag_ici_lock); mutex_init(&pag->pag_ici_reclaim_lock); INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC); - spin_lock_init(&pag->pag_buf_lock); - pag->pag_buf_tree = RB_ROOT; + if (xfs_buf_hash_init(pag)) + goto out_unwind; if (radix_tree_preload(GFP_NOFS)) goto out_unwind; @@ -237,9 +240,11 @@ xfs_initialize_perag( return 0; out_unwind: + xfs_buf_hash_destroy(pag); kmem_free(pag); for (; index > first_initialised; index--) { pag = radix_tree_delete(&mp->m_perag_tree, index); + xfs_buf_hash_destroy(pag); kmem_free(pag); } return error; @@ -684,6 +689,7 @@ xfs_mountfs( xfs_bmap_compute_maxlevels(mp, XFS_ATTR_FORK); xfs_ialloc_compute_maxlevels(mp); xfs_rmapbt_compute_maxlevels(mp); + xfs_refcountbt_compute_maxlevels(mp); xfs_set_maxicount(mp); @@ -923,6 +929,15 @@ xfs_mountfs( } /* + * During the second phase of log recovery, we need iget and + * iput to behave like they do for an active filesystem. + * xfs_fs_drop_inode needs to be able to prevent the deletion + * of inodes before we're done replaying log items on those + * inodes. + */ + mp->m_super->s_flags |= MS_ACTIVE; + + /* * Finish recovering the file system. This part needed to be delayed * until after the root and real-time bitmap inodes were consistently * read in. @@ -934,6 +949,20 @@ xfs_mountfs( } /* + * Now the log is fully replayed, we can transition to full read-only + * mode for read-only mounts. This will sync all the metadata and clean + * the log so that the recovery we just performed does not have to be + * replayed again on the next mount. + * + * We use the same quiesce mechanism as the rw->ro remount, as they are + * semantically identical operations. + */ + if ((mp->m_flags & (XFS_MOUNT_RDONLY|XFS_MOUNT_NORECOVERY)) == + XFS_MOUNT_RDONLY) { + xfs_quiesce_attr(mp); + } + + /* * Complete the quota initialisation, post-log-replay component. */ if (quotamount) { @@ -960,11 +989,30 @@ xfs_mountfs( if (error) xfs_warn(mp, "Unable to allocate reserve blocks. Continuing without reserve pool."); + + /* Recover any CoW blocks that never got remapped. */ + error = xfs_reflink_recover_cow(mp); + if (error) { + xfs_err(mp, + "Error %d recovering leftover CoW allocations.", error); + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + goto out_quota; + } + + /* Reserve AG blocks for future btree expansion. */ + error = xfs_fs_reserve_ag_blocks(mp); + if (error && error != -ENOSPC) + goto out_agresv; } return 0; + out_agresv: + xfs_fs_unreserve_ag_blocks(mp); + out_quota: + xfs_qm_unmount_quotas(mp); out_rtunmount: + mp->m_super->s_flags &= ~MS_ACTIVE; xfs_rtunmount_inodes(mp); out_rele_rip: IRELE(rip); @@ -1005,7 +1053,9 @@ xfs_unmountfs( int error; cancel_delayed_work_sync(&mp->m_eofblocks_work); + cancel_delayed_work_sync(&mp->m_cowblocks_work); + xfs_fs_unreserve_ag_blocks(mp); xfs_qm_unmount_quotas(mp); xfs_rtunmount_inodes(mp); IRELE(mp->m_rootip); diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index b36676cde103..84f785218907 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -57,10 +57,16 @@ enum { #define XFS_ERR_RETRY_FOREVER -1 +/* + * Although retry_timeout is in jiffies which is normally an unsigned long, + * we limit the retry timeout to 86400 seconds, or one day. So even a + * signed 32-bit long is sufficient for a HZ value up to 24855. Making it + * signed lets us store the special "-1" value, meaning retry forever. + */ struct xfs_error_cfg { struct xfs_kobj kobj; int max_retries; - unsigned long retry_timeout; /* in jiffies, 0 = no timeout */ + long retry_timeout; /* in jiffies, -1 = infinite */ }; typedef struct xfs_mount { @@ -118,10 +124,13 @@ typedef struct xfs_mount { uint m_inobt_mnr[2]; /* min inobt btree records */ uint m_rmap_mxr[2]; /* max rmap btree records */ uint m_rmap_mnr[2]; /* min rmap btree records */ + uint m_refc_mxr[2]; /* max refc btree records */ + uint m_refc_mnr[2]; /* min refc btree records */ uint m_ag_maxlevels; /* XFS_AG_MAXLEVELS */ uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */ uint m_in_maxlevels; /* max inobt btree levels. */ uint m_rmap_maxlevels; /* max rmap btree levels */ + uint m_refc_maxlevels; /* max refcount btree level */ xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */ uint m_alloc_set_aside; /* space we can't use */ uint m_ag_max_usable; /* max space per AG */ @@ -155,6 +164,8 @@ typedef struct xfs_mount { struct delayed_work m_reclaim_work; /* background inode reclaim */ struct delayed_work m_eofblocks_work; /* background eof blocks trimming */ + struct delayed_work m_cowblocks_work; /* background cow blocks + trimming */ bool m_update_sb; /* sb needs update in mount */ int64_t m_low_space[XFS_LOWSP_MAX]; /* low free space thresholds */ @@ -325,6 +336,22 @@ xfs_mp_fail_writes(struct xfs_mount *mp) } #endif +/* per-AG block reservation data structures*/ +enum xfs_ag_resv_type { + XFS_AG_RESV_NONE = 0, + XFS_AG_RESV_METADATA, + XFS_AG_RESV_AGFL, +}; + +struct xfs_ag_resv { + /* number of blocks originally reserved here */ + xfs_extlen_t ar_orig_reserved; + /* number of blocks reserved here */ + xfs_extlen_t ar_reserved; + /* number of blocks originally asked for */ + xfs_extlen_t ar_asked; +}; + /* * Per-ag incore structure, copies of information in agf and agi, to improve the * performance of allocation group selection. @@ -366,14 +393,40 @@ typedef struct xfs_perag { unsigned long pag_ici_reclaim_cursor; /* reclaim restart point */ /* buffer cache index */ - spinlock_t pag_buf_lock; /* lock for pag_buf_tree */ - struct rb_root pag_buf_tree; /* ordered tree of active buffers */ + spinlock_t pag_buf_lock; /* lock for pag_buf_hash */ + struct rhashtable pag_buf_hash; /* for rcu-safe freeing */ struct rcu_head rcu_head; int pagb_count; /* pagb slots in use */ + + /* Blocks reserved for all kinds of metadata. */ + struct xfs_ag_resv pag_meta_resv; + /* Blocks reserved for just AGFL-based metadata. */ + struct xfs_ag_resv pag_agfl_resv; + + /* reference count */ + __uint8_t pagf_refcount_level; } xfs_perag_t; +static inline struct xfs_ag_resv * +xfs_perag_resv( + struct xfs_perag *pag, + enum xfs_ag_resv_type type) +{ + switch (type) { + case XFS_AG_RESV_METADATA: + return &pag->pag_meta_resv; + case XFS_AG_RESV_AGFL: + return &pag->pag_agfl_resv; + default: + return NULL; + } +} + +int xfs_buf_hash_init(xfs_perag_t *pag); +void xfs_buf_hash_destroy(xfs_perag_t *pag); + extern void xfs_uuid_table_free(void); extern int xfs_log_sbcount(xfs_mount_t *); extern __uint64_t xfs_default_resblks(xfs_mount_t *mp); diff --git a/fs/xfs/xfs_ondisk.h b/fs/xfs/xfs_ondisk.h index 69e2986a3776..0c381d71b242 100644 --- a/fs/xfs/xfs_ondisk.h +++ b/fs/xfs/xfs_ondisk.h @@ -49,6 +49,8 @@ xfs_check_ondisk_structs(void) XFS_CHECK_STRUCT_SIZE(struct xfs_dsymlink_hdr, 56); XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_key, 4); XFS_CHECK_STRUCT_SIZE(struct xfs_inobt_rec, 16); + XFS_CHECK_STRUCT_SIZE(struct xfs_refcount_key, 4); + XFS_CHECK_STRUCT_SIZE(struct xfs_refcount_rec, 12); XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_key, 20); XFS_CHECK_STRUCT_SIZE(struct xfs_rmap_rec, 24); XFS_CHECK_STRUCT_SIZE(struct xfs_timestamp, 8); @@ -56,6 +58,7 @@ xfs_check_ondisk_structs(void) XFS_CHECK_STRUCT_SIZE(xfs_alloc_ptr_t, 4); XFS_CHECK_STRUCT_SIZE(xfs_alloc_rec_t, 8); XFS_CHECK_STRUCT_SIZE(xfs_inobt_ptr_t, 4); + XFS_CHECK_STRUCT_SIZE(xfs_refcount_ptr_t, 4); XFS_CHECK_STRUCT_SIZE(xfs_rmap_ptr_t, 4); /* dir/attr trees */ diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c index 0f14b2e4bf6c..2f2dc3c09ad0 100644 --- a/fs/xfs/xfs_pnfs.c +++ b/fs/xfs/xfs_pnfs.c @@ -32,8 +32,7 @@ int xfs_break_layouts( struct inode *inode, - uint *iolock, - bool with_imutex) + uint *iolock) { struct xfs_inode *ip = XFS_I(inode); int error; @@ -42,12 +41,8 @@ xfs_break_layouts( while ((error = break_layout(inode, false) == -EWOULDBLOCK)) { xfs_iunlock(ip, *iolock); - if (with_imutex && (*iolock & XFS_IOLOCK_EXCL)) - inode_unlock(inode); error = break_layout(inode, true); *iolock = XFS_IOLOCK_EXCL; - if (with_imutex) - inode_lock(inode); xfs_ilock(ip, *iolock); } @@ -114,6 +109,13 @@ xfs_fs_map_blocks( return -ENXIO; /* + * The pNFS block layout spec actually supports reflink like + * functionality, but the Linux pNFS server doesn't implement it yet. + */ + if (xfs_is_reflink_inode(ip)) + return -ENXIO; + + /* * Lock out any other I/O before we flush and invalidate the pagecache, * and then hand out a layout to the remote system. This is very * similar to direct I/O, except that the synchronization is much more diff --git a/fs/xfs/xfs_pnfs.h b/fs/xfs/xfs_pnfs.h index e8339f74966b..b587cb99b2b7 100644 --- a/fs/xfs/xfs_pnfs.h +++ b/fs/xfs/xfs_pnfs.h @@ -8,10 +8,10 @@ int xfs_fs_map_blocks(struct inode *inode, loff_t offset, u64 length, int xfs_fs_commit_blocks(struct inode *inode, struct iomap *maps, int nr_maps, struct iattr *iattr); -int xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex); +int xfs_break_layouts(struct inode *inode, uint *iolock); #else static inline int -xfs_break_layouts(struct inode *inode, uint *iolock, bool with_imutex) +xfs_break_layouts(struct inode *inode, uint *iolock) { return 0; } diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index a60d9e2739d1..45e50ea90769 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -1135,7 +1135,7 @@ xfs_qm_get_rtblks( return error; } rtblks = 0; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); + nextents = xfs_iext_count(ifp); for (idx = 0; idx < nextents; idx++) rtblks += xfs_bmbt_get_blockcount(xfs_iext_get_ext(ifp, idx)); *O_rtblks = (xfs_qcnt_t)rtblks; diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c new file mode 100644 index 000000000000..6e4c7446c3d4 --- /dev/null +++ b/fs/xfs/xfs_refcount_item.c @@ -0,0 +1,540 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_bit.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_buf_item.h" +#include "xfs_refcount_item.h" +#include "xfs_log.h" +#include "xfs_refcount.h" + + +kmem_zone_t *xfs_cui_zone; +kmem_zone_t *xfs_cud_zone; + +static inline struct xfs_cui_log_item *CUI_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_cui_log_item, cui_item); +} + +void +xfs_cui_item_free( + struct xfs_cui_log_item *cuip) +{ + if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS) + kmem_free(cuip); + else + kmem_zone_free(xfs_cui_zone, cuip); +} + +STATIC void +xfs_cui_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + struct xfs_cui_log_item *cuip = CUI_ITEM(lip); + + *nvecs += 1; + *nbytes += xfs_cui_log_format_sizeof(cuip->cui_format.cui_nextents); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given cui log item. We use only 1 iovec, and we point that + * at the cui_log_format structure embedded in the cui item. + * It is at this point that we assert that all of the extent + * slots in the cui item have been filled. + */ +STATIC void +xfs_cui_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_cui_log_item *cuip = CUI_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + ASSERT(atomic_read(&cuip->cui_next_extent) == + cuip->cui_format.cui_nextents); + + cuip->cui_format.cui_type = XFS_LI_CUI; + cuip->cui_format.cui_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUI_FORMAT, &cuip->cui_format, + xfs_cui_log_format_sizeof(cuip->cui_format.cui_nextents)); +} + +/* + * Pinning has no meaning for an cui item, so just return. + */ +STATIC void +xfs_cui_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * The unpin operation is the last place an CUI is manipulated in the log. It is + * either inserted in the AIL or aborted in the event of a log I/O error. In + * either case, the CUI transaction has been successfully committed to make it + * this far. Therefore, we expect whoever committed the CUI to either construct + * and commit the CUD or drop the CUD's reference in the event of error. Simply + * drop the log's CUI reference now that the log is done with it. + */ +STATIC void +xfs_cui_item_unpin( + struct xfs_log_item *lip, + int remove) +{ + struct xfs_cui_log_item *cuip = CUI_ITEM(lip); + + xfs_cui_release(cuip); +} + +/* + * CUI items have no locking or pushing. However, since CUIs are pulled from + * the AIL when their corresponding CUDs are committed to disk, their situation + * is very similar to being pinned. Return XFS_ITEM_PINNED so that the caller + * will eventually flush the log. This should help in getting the CUI out of + * the AIL. + */ +STATIC uint +xfs_cui_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The CUI has been either committed or aborted if the transaction has been + * cancelled. If the transaction was cancelled, an CUD isn't going to be + * constructed and thus we free the CUI here directly. + */ +STATIC void +xfs_cui_item_unlock( + struct xfs_log_item *lip) +{ + if (lip->li_flags & XFS_LI_ABORTED) + xfs_cui_item_free(CUI_ITEM(lip)); +} + +/* + * The CUI is logged only once and cannot be moved in the log, so simply return + * the lsn at which it's been logged. + */ +STATIC xfs_lsn_t +xfs_cui_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + return lsn; +} + +/* + * The CUI dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_cui_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all cui log items. + */ +static const struct xfs_item_ops xfs_cui_item_ops = { + .iop_size = xfs_cui_item_size, + .iop_format = xfs_cui_item_format, + .iop_pin = xfs_cui_item_pin, + .iop_unpin = xfs_cui_item_unpin, + .iop_unlock = xfs_cui_item_unlock, + .iop_committed = xfs_cui_item_committed, + .iop_push = xfs_cui_item_push, + .iop_committing = xfs_cui_item_committing, +}; + +/* + * Allocate and initialize an cui item with the given number of extents. + */ +struct xfs_cui_log_item * +xfs_cui_init( + struct xfs_mount *mp, + uint nextents) + +{ + struct xfs_cui_log_item *cuip; + + ASSERT(nextents > 0); + if (nextents > XFS_CUI_MAX_FAST_EXTENTS) + cuip = kmem_zalloc(xfs_cui_log_item_sizeof(nextents), + KM_SLEEP); + else + cuip = kmem_zone_zalloc(xfs_cui_zone, KM_SLEEP); + + xfs_log_item_init(mp, &cuip->cui_item, XFS_LI_CUI, &xfs_cui_item_ops); + cuip->cui_format.cui_nextents = nextents; + cuip->cui_format.cui_id = (uintptr_t)(void *)cuip; + atomic_set(&cuip->cui_next_extent, 0); + atomic_set(&cuip->cui_refcount, 2); + + return cuip; +} + +/* + * Freeing the CUI requires that we remove it from the AIL if it has already + * been placed there. However, the CUI may not yet have been placed in the AIL + * when called by xfs_cui_release() from CUD processing due to the ordering of + * committed vs unpin operations in bulk insert operations. Hence the reference + * count to ensure only the last caller frees the CUI. + */ +void +xfs_cui_release( + struct xfs_cui_log_item *cuip) +{ + if (atomic_dec_and_test(&cuip->cui_refcount)) { + xfs_trans_ail_remove(&cuip->cui_item, SHUTDOWN_LOG_IO_ERROR); + xfs_cui_item_free(cuip); + } +} + +static inline struct xfs_cud_log_item *CUD_ITEM(struct xfs_log_item *lip) +{ + return container_of(lip, struct xfs_cud_log_item, cud_item); +} + +STATIC void +xfs_cud_item_size( + struct xfs_log_item *lip, + int *nvecs, + int *nbytes) +{ + *nvecs += 1; + *nbytes += sizeof(struct xfs_cud_log_format); +} + +/* + * This is called to fill in the vector of log iovecs for the + * given cud log item. We use only 1 iovec, and we point that + * at the cud_log_format structure embedded in the cud item. + * It is at this point that we assert that all of the extent + * slots in the cud item have been filled. + */ +STATIC void +xfs_cud_item_format( + struct xfs_log_item *lip, + struct xfs_log_vec *lv) +{ + struct xfs_cud_log_item *cudp = CUD_ITEM(lip); + struct xfs_log_iovec *vecp = NULL; + + cudp->cud_format.cud_type = XFS_LI_CUD; + cudp->cud_format.cud_size = 1; + + xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_CUD_FORMAT, &cudp->cud_format, + sizeof(struct xfs_cud_log_format)); +} + +/* + * Pinning has no meaning for an cud item, so just return. + */ +STATIC void +xfs_cud_item_pin( + struct xfs_log_item *lip) +{ +} + +/* + * Since pinning has no meaning for an cud item, unpinning does + * not either. + */ +STATIC void +xfs_cud_item_unpin( + struct xfs_log_item *lip, + int remove) +{ +} + +/* + * There isn't much you can do to push on an cud item. It is simply stuck + * waiting for the log to be flushed to disk. + */ +STATIC uint +xfs_cud_item_push( + struct xfs_log_item *lip, + struct list_head *buffer_list) +{ + return XFS_ITEM_PINNED; +} + +/* + * The CUD is either committed or aborted if the transaction is cancelled. If + * the transaction is cancelled, drop our reference to the CUI and free the + * CUD. + */ +STATIC void +xfs_cud_item_unlock( + struct xfs_log_item *lip) +{ + struct xfs_cud_log_item *cudp = CUD_ITEM(lip); + + if (lip->li_flags & XFS_LI_ABORTED) { + xfs_cui_release(cudp->cud_cuip); + kmem_zone_free(xfs_cud_zone, cudp); + } +} + +/* + * When the cud item is committed to disk, all we need to do is delete our + * reference to our partner cui item and then free ourselves. Since we're + * freeing ourselves we must return -1 to keep the transaction code from + * further referencing this item. + */ +STATIC xfs_lsn_t +xfs_cud_item_committed( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ + struct xfs_cud_log_item *cudp = CUD_ITEM(lip); + + /* + * Drop the CUI reference regardless of whether the CUD has been + * aborted. Once the CUD transaction is constructed, it is the sole + * responsibility of the CUD to release the CUI (even if the CUI is + * aborted due to log I/O error). + */ + xfs_cui_release(cudp->cud_cuip); + kmem_zone_free(xfs_cud_zone, cudp); + + return (xfs_lsn_t)-1; +} + +/* + * The CUD dependency tracking op doesn't do squat. It can't because + * it doesn't know where the free extent is coming from. The dependency + * tracking has to be handled by the "enclosing" metadata object. For + * example, for inodes, the inode is locked throughout the extent freeing + * so the dependency should be recorded there. + */ +STATIC void +xfs_cud_item_committing( + struct xfs_log_item *lip, + xfs_lsn_t lsn) +{ +} + +/* + * This is the ops vector shared by all cud log items. + */ +static const struct xfs_item_ops xfs_cud_item_ops = { + .iop_size = xfs_cud_item_size, + .iop_format = xfs_cud_item_format, + .iop_pin = xfs_cud_item_pin, + .iop_unpin = xfs_cud_item_unpin, + .iop_unlock = xfs_cud_item_unlock, + .iop_committed = xfs_cud_item_committed, + .iop_push = xfs_cud_item_push, + .iop_committing = xfs_cud_item_committing, +}; + +/* + * Allocate and initialize an cud item with the given number of extents. + */ +struct xfs_cud_log_item * +xfs_cud_init( + struct xfs_mount *mp, + struct xfs_cui_log_item *cuip) + +{ + struct xfs_cud_log_item *cudp; + + cudp = kmem_zone_zalloc(xfs_cud_zone, KM_SLEEP); + xfs_log_item_init(mp, &cudp->cud_item, XFS_LI_CUD, &xfs_cud_item_ops); + cudp->cud_cuip = cuip; + cudp->cud_format.cud_cui_id = cuip->cui_format.cui_id; + + return cudp; +} + +/* + * Process a refcount update intent item that was recovered from the log. + * We need to update the refcountbt. + */ +int +xfs_cui_recover( + struct xfs_mount *mp, + struct xfs_cui_log_item *cuip) +{ + int i; + int error = 0; + unsigned int refc_type; + struct xfs_phys_extent *refc; + xfs_fsblock_t startblock_fsb; + bool op_ok; + struct xfs_cud_log_item *cudp; + struct xfs_trans *tp; + struct xfs_btree_cur *rcur = NULL; + enum xfs_refcount_intent_type type; + xfs_fsblock_t firstfsb; + xfs_fsblock_t new_fsb; + xfs_extlen_t new_len; + struct xfs_bmbt_irec irec; + struct xfs_defer_ops dfops; + bool requeue_only = false; + + ASSERT(!test_bit(XFS_CUI_RECOVERED, &cuip->cui_flags)); + + /* + * First check the validity of the extents described by the + * CUI. If any are bad, then assume that all are bad and + * just toss the CUI. + */ + for (i = 0; i < cuip->cui_format.cui_nextents; i++) { + refc = &cuip->cui_format.cui_extents[i]; + startblock_fsb = XFS_BB_TO_FSB(mp, + XFS_FSB_TO_DADDR(mp, refc->pe_startblock)); + switch (refc->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK) { + case XFS_REFCOUNT_INCREASE: + case XFS_REFCOUNT_DECREASE: + case XFS_REFCOUNT_ALLOC_COW: + case XFS_REFCOUNT_FREE_COW: + op_ok = true; + break; + default: + op_ok = false; + break; + } + if (!op_ok || startblock_fsb == 0 || + refc->pe_len == 0 || + startblock_fsb >= mp->m_sb.sb_dblocks || + refc->pe_len >= mp->m_sb.sb_agblocks || + (refc->pe_flags & ~XFS_REFCOUNT_EXTENT_FLAGS)) { + /* + * This will pull the CUI from the AIL and + * free the memory associated with it. + */ + set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags); + xfs_cui_release(cuip); + return -EIO; + } + } + + /* + * Under normal operation, refcount updates are deferred, so we + * wouldn't be adding them directly to a transaction. All + * refcount updates manage reservation usage internally and + * dynamically by deferring work that won't fit in the + * transaction. Normally, any work that needs to be deferred + * gets attached to the same defer_ops that scheduled the + * refcount update. However, we're in log recovery here, so we + * we create our own defer_ops and use that to finish up any + * work that doesn't fit. + */ + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp); + if (error) + return error; + cudp = xfs_trans_get_cud(tp, cuip); + + xfs_defer_init(&dfops, &firstfsb); + for (i = 0; i < cuip->cui_format.cui_nextents; i++) { + refc = &cuip->cui_format.cui_extents[i]; + refc_type = refc->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK; + switch (refc_type) { + case XFS_REFCOUNT_INCREASE: + case XFS_REFCOUNT_DECREASE: + case XFS_REFCOUNT_ALLOC_COW: + case XFS_REFCOUNT_FREE_COW: + type = refc_type; + break; + default: + error = -EFSCORRUPTED; + goto abort_error; + } + if (requeue_only) { + new_fsb = refc->pe_startblock; + new_len = refc->pe_len; + } else + error = xfs_trans_log_finish_refcount_update(tp, cudp, + &dfops, type, refc->pe_startblock, refc->pe_len, + &new_fsb, &new_len, &rcur); + if (error) + goto abort_error; + + /* Requeue what we didn't finish. */ + if (new_len > 0) { + irec.br_startblock = new_fsb; + irec.br_blockcount = new_len; + switch (type) { + case XFS_REFCOUNT_INCREASE: + error = xfs_refcount_increase_extent( + tp->t_mountp, &dfops, &irec); + break; + case XFS_REFCOUNT_DECREASE: + error = xfs_refcount_decrease_extent( + tp->t_mountp, &dfops, &irec); + break; + case XFS_REFCOUNT_ALLOC_COW: + error = xfs_refcount_alloc_cow_extent( + tp->t_mountp, &dfops, + irec.br_startblock, + irec.br_blockcount); + break; + case XFS_REFCOUNT_FREE_COW: + error = xfs_refcount_free_cow_extent( + tp->t_mountp, &dfops, + irec.br_startblock, + irec.br_blockcount); + break; + default: + ASSERT(0); + } + if (error) + goto abort_error; + requeue_only = true; + } + } + + xfs_refcount_finish_one_cleanup(tp, rcur, error); + error = xfs_defer_finish(&tp, &dfops, NULL); + if (error) + goto abort_defer; + set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags); + error = xfs_trans_commit(tp); + return error; + +abort_error: + xfs_refcount_finish_one_cleanup(tp, rcur, error); +abort_defer: + xfs_defer_cancel(&dfops); + xfs_trans_cancel(tp); + return error; +} diff --git a/fs/xfs/xfs_refcount_item.h b/fs/xfs/xfs_refcount_item.h new file mode 100644 index 000000000000..5b74dddfa64b --- /dev/null +++ b/fs/xfs/xfs_refcount_item.h @@ -0,0 +1,101 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_REFCOUNT_ITEM_H__ +#define __XFS_REFCOUNT_ITEM_H__ + +/* + * There are (currently) two pairs of refcount btree redo item types: + * increase and decrease. The log items for these are CUI (refcount + * update intent) and CUD (refcount update done). The redo item type + * is encoded in the flags field of each xfs_map_extent. + * + * *I items should be recorded in the *first* of a series of rolled + * transactions, and the *D items should be recorded in the same + * transaction that records the associated refcountbt updates. + * + * Should the system crash after the commit of the first transaction + * but before the commit of the final transaction in a series, log + * recovery will use the redo information recorded by the intent items + * to replay the refcountbt metadata updates. + */ + +/* kernel only CUI/CUD definitions */ + +struct xfs_mount; +struct kmem_zone; + +/* + * Max number of extents in fast allocation path. + */ +#define XFS_CUI_MAX_FAST_EXTENTS 16 + +/* + * Define CUI flag bits. Manipulated by set/clear/test_bit operators. + */ +#define XFS_CUI_RECOVERED 1 + +/* + * This is the "refcount update intent" log item. It is used to log + * the fact that some reverse mappings need to change. It is used in + * conjunction with the "refcount update done" log item described + * below. + * + * These log items follow the same rules as struct xfs_efi_log_item; + * see the comments about that structure (in xfs_extfree_item.h) for + * more details. + */ +struct xfs_cui_log_item { + struct xfs_log_item cui_item; + atomic_t cui_refcount; + atomic_t cui_next_extent; + unsigned long cui_flags; /* misc flags */ + struct xfs_cui_log_format cui_format; +}; + +static inline size_t +xfs_cui_log_item_sizeof( + unsigned int nr) +{ + return offsetof(struct xfs_cui_log_item, cui_format) + + xfs_cui_log_format_sizeof(nr); +} + +/* + * This is the "refcount update done" log item. It is used to log the + * fact that some refcountbt updates mentioned in an earlier cui item + * have been performed. + */ +struct xfs_cud_log_item { + struct xfs_log_item cud_item; + struct xfs_cui_log_item *cud_cuip; + struct xfs_cud_log_format cud_format; +}; + +extern struct kmem_zone *xfs_cui_zone; +extern struct kmem_zone *xfs_cud_zone; + +struct xfs_cui_log_item *xfs_cui_init(struct xfs_mount *, uint); +struct xfs_cud_log_item *xfs_cud_init(struct xfs_mount *, + struct xfs_cui_log_item *); +void xfs_cui_item_free(struct xfs_cui_log_item *); +void xfs_cui_release(struct xfs_cui_log_item *); +int xfs_cui_recover(struct xfs_mount *mp, struct xfs_cui_log_item *cuip); + +#endif /* __XFS_REFCOUNT_ITEM_H__ */ diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c new file mode 100644 index 000000000000..07593a362cd0 --- /dev/null +++ b/fs/xfs/xfs_reflink.c @@ -0,0 +1,1449 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_da_format.h" +#include "xfs_da_btree.h" +#include "xfs_inode.h" +#include "xfs_trans.h" +#include "xfs_inode_item.h" +#include "xfs_bmap.h" +#include "xfs_bmap_util.h" +#include "xfs_error.h" +#include "xfs_dir2.h" +#include "xfs_dir2_priv.h" +#include "xfs_ioctl.h" +#include "xfs_trace.h" +#include "xfs_log.h" +#include "xfs_icache.h" +#include "xfs_pnfs.h" +#include "xfs_btree.h" +#include "xfs_refcount_btree.h" +#include "xfs_refcount.h" +#include "xfs_bmap_btree.h" +#include "xfs_trans_space.h" +#include "xfs_bit.h" +#include "xfs_alloc.h" +#include "xfs_quota_defs.h" +#include "xfs_quota.h" +#include "xfs_btree.h" +#include "xfs_bmap_btree.h" +#include "xfs_reflink.h" +#include "xfs_iomap.h" +#include "xfs_rmap_btree.h" +#include "xfs_sb.h" +#include "xfs_ag_resv.h" + +/* + * Copy on Write of Shared Blocks + * + * XFS must preserve "the usual" file semantics even when two files share + * the same physical blocks. This means that a write to one file must not + * alter the blocks in a different file; the way that we'll do that is + * through the use of a copy-on-write mechanism. At a high level, that + * means that when we want to write to a shared block, we allocate a new + * block, write the data to the new block, and if that succeeds we map the + * new block into the file. + * + * XFS provides a "delayed allocation" mechanism that defers the allocation + * of disk blocks to dirty-but-not-yet-mapped file blocks as long as + * possible. This reduces fragmentation by enabling the filesystem to ask + * for bigger chunks less often, which is exactly what we want for CoW. + * + * The delalloc mechanism begins when the kernel wants to make a block + * writable (write_begin or page_mkwrite). If the offset is not mapped, we + * create a delalloc mapping, which is a regular in-core extent, but without + * a real startblock. (For delalloc mappings, the startblock encodes both + * a flag that this is a delalloc mapping, and a worst-case estimate of how + * many blocks might be required to put the mapping into the BMBT.) delalloc + * mappings are a reservation against the free space in the filesystem; + * adjacent mappings can also be combined into fewer larger mappings. + * + * When dirty pages are being written out (typically in writepage), the + * delalloc reservations are converted into real mappings by allocating + * blocks and replacing the delalloc mapping with real ones. A delalloc + * mapping can be replaced by several real ones if the free space is + * fragmented. + * + * We want to adapt the delalloc mechanism for copy-on-write, since the + * write paths are similar. The first two steps (creating the reservation + * and allocating the blocks) are exactly the same as delalloc except that + * the mappings must be stored in a separate CoW fork because we do not want + * to disturb the mapping in the data fork until we're sure that the write + * succeeded. IO completion in this case is the process of removing the old + * mapping from the data fork and moving the new mapping from the CoW fork to + * the data fork. This will be discussed shortly. + * + * For now, unaligned directio writes will be bounced back to the page cache. + * Block-aligned directio writes will use the same mechanism as buffered + * writes. + * + * CoW remapping must be done after the data block write completes, + * because we don't want to destroy the old data fork map until we're sure + * the new block has been written. Since the new mappings are kept in a + * separate fork, we can simply iterate these mappings to find the ones + * that cover the file blocks that we just CoW'd. For each extent, simply + * unmap the corresponding range in the data fork, map the new range into + * the data fork, and remove the extent from the CoW fork. + * + * Since the remapping operation can be applied to an arbitrary file + * range, we record the need for the remap step as a flag in the ioend + * instead of declaring a new IO type. This is required for direct io + * because we only have ioend for the whole dio, and we have to be able to + * remember the presence of unwritten blocks and CoW blocks with a single + * ioend structure. Better yet, the more ground we can cover with one + * ioend, the better. + */ + +/* + * Given an AG extent, find the lowest-numbered run of shared blocks + * within that range and return the range in fbno/flen. If + * find_end_of_shared is true, return the longest contiguous extent of + * shared blocks. If there are no shared extents, fbno and flen will + * be set to NULLAGBLOCK and 0, respectively. + */ +int +xfs_reflink_find_shared( + struct xfs_mount *mp, + xfs_agnumber_t agno, + xfs_agblock_t agbno, + xfs_extlen_t aglen, + xfs_agblock_t *fbno, + xfs_extlen_t *flen, + bool find_end_of_shared) +{ + struct xfs_buf *agbp; + struct xfs_btree_cur *cur; + int error; + + error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp); + if (error) + return error; + + cur = xfs_refcountbt_init_cursor(mp, NULL, agbp, agno, NULL); + + error = xfs_refcount_find_shared(cur, agbno, aglen, fbno, flen, + find_end_of_shared); + + xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR); + + xfs_buf_relse(agbp); + return error; +} + +/* + * Trim the mapping to the next block where there's a change in the + * shared/unshared status. More specifically, this means that we + * find the lowest-numbered extent of shared blocks that coincides with + * the given block mapping. If the shared extent overlaps the start of + * the mapping, trim the mapping to the end of the shared extent. If + * the shared region intersects the mapping, trim the mapping to the + * start of the shared extent. If there are no shared regions that + * overlap, just return the original extent. + */ +int +xfs_reflink_trim_around_shared( + struct xfs_inode *ip, + struct xfs_bmbt_irec *irec, + bool *shared, + bool *trimmed) +{ + xfs_agnumber_t agno; + xfs_agblock_t agbno; + xfs_extlen_t aglen; + xfs_agblock_t fbno; + xfs_extlen_t flen; + int error = 0; + + /* Holes, unwritten, and delalloc extents cannot be shared */ + if (!xfs_is_reflink_inode(ip) || + ISUNWRITTEN(irec) || + irec->br_startblock == HOLESTARTBLOCK || + irec->br_startblock == DELAYSTARTBLOCK || + isnullstartblock(irec->br_startblock)) { + *shared = false; + return 0; + } + + trace_xfs_reflink_trim_around_shared(ip, irec); + + agno = XFS_FSB_TO_AGNO(ip->i_mount, irec->br_startblock); + agbno = XFS_FSB_TO_AGBNO(ip->i_mount, irec->br_startblock); + aglen = irec->br_blockcount; + + error = xfs_reflink_find_shared(ip->i_mount, agno, agbno, + aglen, &fbno, &flen, true); + if (error) + return error; + + *shared = *trimmed = false; + if (fbno == NULLAGBLOCK) { + /* No shared blocks at all. */ + return 0; + } else if (fbno == agbno) { + /* + * The start of this extent is shared. Truncate the + * mapping at the end of the shared region so that a + * subsequent iteration starts at the start of the + * unshared region. + */ + irec->br_blockcount = flen; + *shared = true; + if (flen != aglen) + *trimmed = true; + return 0; + } else { + /* + * There's a shared extent midway through this extent. + * Truncate the mapping at the start of the shared + * extent so that a subsequent iteration starts at the + * start of the shared region. + */ + irec->br_blockcount = fbno - agbno; + *trimmed = true; + return 0; + } +} + +/* + * Trim the passed in imap to the next shared/unshared extent boundary, and + * if imap->br_startoff points to a shared extent reserve space for it in the + * COW fork. In this case *shared is set to true, else to false. + * + * Note that imap will always contain the block numbers for the existing blocks + * in the data fork, as the upper layers need them for read-modify-write + * operations. + */ +int +xfs_reflink_reserve_cow( + struct xfs_inode *ip, + struct xfs_bmbt_irec *imap, + bool *shared) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + struct xfs_bmbt_irec got; + int error = 0; + bool eof = false, trimmed; + xfs_extnum_t idx; + + /* + * Search the COW fork extent list first. This serves two purposes: + * first this implement the speculative preallocation using cowextisze, + * so that we also unshared block adjacent to shared blocks instead + * of just the shared blocks themselves. Second the lookup in the + * extent list is generally faster than going out to the shared extent + * tree. + */ + + if (!xfs_iext_lookup_extent(ip, ifp, imap->br_startoff, &idx, &got)) + eof = true; + if (!eof && got.br_startoff <= imap->br_startoff) { + trace_xfs_reflink_cow_found(ip, imap); + xfs_trim_extent(imap, got.br_startoff, got.br_blockcount); + + *shared = true; + return 0; + } + + /* Trim the mapping to the nearest shared extent boundary. */ + error = xfs_reflink_trim_around_shared(ip, imap, shared, &trimmed); + if (error) + return error; + + /* Not shared? Just report the (potentially capped) extent. */ + if (!*shared) + return 0; + + /* + * Fork all the shared blocks from our write offset until the end of + * the extent. + */ + error = xfs_qm_dqattach_locked(ip, 0); + if (error) + return error; + + error = xfs_bmapi_reserve_delalloc(ip, XFS_COW_FORK, imap->br_startoff, + imap->br_blockcount, 0, &got, &idx, eof); + if (error == -ENOSPC || error == -EDQUOT) + trace_xfs_reflink_cow_enospc(ip, imap); + if (error) + return error; + + trace_xfs_reflink_cow_alloc(ip, &got); + return 0; +} + +/* Allocate all CoW reservations covering a range of blocks in a file. */ +static int +__xfs_reflink_allocate_cow( + struct xfs_inode *ip, + xfs_fileoff_t *offset_fsb, + xfs_fileoff_t end_fsb) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_bmbt_irec imap; + struct xfs_defer_ops dfops; + struct xfs_trans *tp; + xfs_fsblock_t first_block; + int nimaps = 1, error; + bool shared; + + xfs_defer_init(&dfops, &first_block); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, + XFS_TRANS_RESERVE, &tp); + if (error) + return error; + + xfs_ilock(ip, XFS_ILOCK_EXCL); + + /* Read extent from the source file. */ + nimaps = 1; + error = xfs_bmapi_read(ip, *offset_fsb, end_fsb - *offset_fsb, + &imap, &nimaps, 0); + if (error) + goto out_unlock; + ASSERT(nimaps == 1); + + error = xfs_reflink_reserve_cow(ip, &imap, &shared); + if (error) + goto out_trans_cancel; + + if (!shared) { + *offset_fsb = imap.br_startoff + imap.br_blockcount; + goto out_trans_cancel; + } + + xfs_trans_ijoin(tp, ip, 0); + error = xfs_bmapi_write(tp, ip, imap.br_startoff, imap.br_blockcount, + XFS_BMAPI_COWFORK, &first_block, + XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK), + &imap, &nimaps, &dfops); + if (error) + goto out_trans_cancel; + + error = xfs_defer_finish(&tp, &dfops, NULL); + if (error) + goto out_trans_cancel; + + error = xfs_trans_commit(tp); + + *offset_fsb = imap.br_startoff + imap.br_blockcount; +out_unlock: + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return error; +out_trans_cancel: + xfs_defer_cancel(&dfops); + xfs_trans_cancel(tp); + goto out_unlock; +} + +/* Allocate all CoW reservations covering a part of a file. */ +int +xfs_reflink_allocate_cow_range( + struct xfs_inode *ip, + xfs_off_t offset, + xfs_off_t count) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_fileoff_t offset_fsb = XFS_B_TO_FSBT(mp, offset); + xfs_fileoff_t end_fsb = XFS_B_TO_FSB(mp, offset + count); + int error; + + ASSERT(xfs_is_reflink_inode(ip)); + + trace_xfs_reflink_allocate_cow_range(ip, offset, count); + + /* + * Make sure that the dquots are there. + */ + error = xfs_qm_dqattach(ip, 0); + if (error) + return error; + + while (offset_fsb < end_fsb) { + error = __xfs_reflink_allocate_cow(ip, &offset_fsb, end_fsb); + if (error) { + trace_xfs_reflink_allocate_cow_range_error(ip, error, + _RET_IP_); + break; + } + } + + return error; +} + +/* + * Find the CoW reservation for a given byte offset of a file. + */ +bool +xfs_reflink_find_cow_mapping( + struct xfs_inode *ip, + xfs_off_t offset, + struct xfs_bmbt_irec *imap) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + xfs_fileoff_t offset_fsb; + struct xfs_bmbt_irec got; + xfs_extnum_t idx; + + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED)); + ASSERT(xfs_is_reflink_inode(ip)); + + offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset); + if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got)) + return false; + if (got.br_startoff > offset_fsb) + return false; + + trace_xfs_reflink_find_cow_mapping(ip, offset, 1, XFS_IO_OVERWRITE, + &got); + *imap = got; + return true; +} + +/* + * Trim an extent to end at the next CoW reservation past offset_fsb. + */ +void +xfs_reflink_trim_irec_to_next_cow( + struct xfs_inode *ip, + xfs_fileoff_t offset_fsb, + struct xfs_bmbt_irec *imap) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + struct xfs_bmbt_irec got; + xfs_extnum_t idx; + + if (!xfs_is_reflink_inode(ip)) + return; + + /* Find the extent in the CoW fork. */ + if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got)) + return; + + /* This is the extent before; try sliding up one. */ + if (got.br_startoff < offset_fsb) { + if (!xfs_iext_get_extent(ifp, idx + 1, &got)) + return; + } + + if (got.br_startoff >= imap->br_startoff + imap->br_blockcount) + return; + + imap->br_blockcount = got.br_startoff - imap->br_startoff; + trace_xfs_reflink_trim_irec(ip, imap); +} + +/* + * Cancel all pending CoW reservations for some block range of an inode. + */ +int +xfs_reflink_cancel_cow_blocks( + struct xfs_inode *ip, + struct xfs_trans **tpp, + xfs_fileoff_t offset_fsb, + xfs_fileoff_t end_fsb) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + struct xfs_bmbt_irec got, del; + xfs_extnum_t idx; + xfs_fsblock_t firstfsb; + struct xfs_defer_ops dfops; + int error = 0; + + if (!xfs_is_reflink_inode(ip)) + return 0; + if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &idx, &got)) + return 0; + + while (got.br_startoff < end_fsb) { + del = got; + xfs_trim_extent(&del, offset_fsb, end_fsb - offset_fsb); + trace_xfs_reflink_cancel_cow(ip, &del); + + if (isnullstartblock(del.br_startblock)) { + error = xfs_bmap_del_extent_delay(ip, XFS_COW_FORK, + &idx, &got, &del); + if (error) + break; + } else { + xfs_trans_ijoin(*tpp, ip, 0); + xfs_defer_init(&dfops, &firstfsb); + + /* Free the CoW orphan record. */ + error = xfs_refcount_free_cow_extent(ip->i_mount, + &dfops, del.br_startblock, + del.br_blockcount); + if (error) + break; + + xfs_bmap_add_free(ip->i_mount, &dfops, + del.br_startblock, del.br_blockcount, + NULL); + + /* Update quota accounting */ + xfs_trans_mod_dquot_byino(*tpp, ip, XFS_TRANS_DQ_BCOUNT, + -(long)del.br_blockcount); + + /* Roll the transaction */ + error = xfs_defer_finish(tpp, &dfops, ip); + if (error) { + xfs_defer_cancel(&dfops); + break; + } + + /* Remove the mapping from the CoW fork. */ + xfs_bmap_del_extent_cow(ip, &idx, &got, &del); + } + + if (!xfs_iext_get_extent(ifp, ++idx, &got)) + break; + } + + /* clear tag if cow fork is emptied */ + if (!ifp->if_bytes) + xfs_inode_clear_cowblocks_tag(ip); + + return error; +} + +/* + * Cancel all pending CoW reservations for some byte range of an inode. + */ +int +xfs_reflink_cancel_cow_range( + struct xfs_inode *ip, + xfs_off_t offset, + xfs_off_t count) +{ + struct xfs_trans *tp; + xfs_fileoff_t offset_fsb; + xfs_fileoff_t end_fsb; + int error; + + trace_xfs_reflink_cancel_cow_range(ip, offset, count); + ASSERT(xfs_is_reflink_inode(ip)); + + offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset); + if (count == NULLFILEOFF) + end_fsb = NULLFILEOFF; + else + end_fsb = XFS_B_TO_FSB(ip->i_mount, offset + count); + + /* Start a rolling transaction to remove the mappings */ + error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_write, + 0, 0, 0, &tp); + if (error) + goto out; + + xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, 0); + + /* Scrape out the old CoW reservations */ + error = xfs_reflink_cancel_cow_blocks(ip, &tp, offset_fsb, end_fsb); + if (error) + goto out_cancel; + + error = xfs_trans_commit(tp); + + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return error; + +out_cancel: + xfs_trans_cancel(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); +out: + trace_xfs_reflink_cancel_cow_range_error(ip, error, _RET_IP_); + return error; +} + +/* + * Remap parts of a file's data fork after a successful CoW. + */ +int +xfs_reflink_end_cow( + struct xfs_inode *ip, + xfs_off_t offset, + xfs_off_t count) +{ + struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK); + struct xfs_bmbt_irec got, del; + struct xfs_trans *tp; + xfs_fileoff_t offset_fsb; + xfs_fileoff_t end_fsb; + xfs_fsblock_t firstfsb; + struct xfs_defer_ops dfops; + int error; + unsigned int resblks; + xfs_filblks_t rlen; + xfs_extnum_t idx; + + trace_xfs_reflink_end_cow(ip, offset, count); + + /* No COW extents? That's easy! */ + if (ifp->if_bytes == 0) + return 0; + + offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset); + end_fsb = XFS_B_TO_FSB(ip->i_mount, offset + count); + + /* Start a rolling transaction to switch the mappings */ + resblks = XFS_EXTENTADD_SPACE_RES(ip->i_mount, XFS_DATA_FORK); + error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_write, + resblks, 0, 0, &tp); + if (error) + goto out; + + xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, 0); + + /* If there is a hole at end_fsb - 1 go to the previous extent */ + if (!xfs_iext_lookup_extent(ip, ifp, end_fsb - 1, &idx, &got) || + got.br_startoff > end_fsb) { + ASSERT(idx > 0); + xfs_iext_get_extent(ifp, --idx, &got); + } + + /* Walk backwards until we're out of the I/O range... */ + while (got.br_startoff + got.br_blockcount > offset_fsb) { + del = got; + xfs_trim_extent(&del, offset_fsb, end_fsb - offset_fsb); + + /* Extent delete may have bumped idx forward */ + if (!del.br_blockcount) { + idx--; + goto next_extent; + } + + ASSERT(!isnullstartblock(got.br_startblock)); + + /* Unmap the old blocks in the data fork. */ + xfs_defer_init(&dfops, &firstfsb); + rlen = del.br_blockcount; + error = __xfs_bunmapi(tp, ip, del.br_startoff, &rlen, 0, 1, + &firstfsb, &dfops); + if (error) + goto out_defer; + + /* Trim the extent to whatever got unmapped. */ + if (rlen) { + xfs_trim_extent(&del, del.br_startoff + rlen, + del.br_blockcount - rlen); + } + trace_xfs_reflink_cow_remap(ip, &del); + + /* Free the CoW orphan record. */ + error = xfs_refcount_free_cow_extent(tp->t_mountp, &dfops, + del.br_startblock, del.br_blockcount); + if (error) + goto out_defer; + + /* Map the new blocks into the data fork. */ + error = xfs_bmap_map_extent(tp->t_mountp, &dfops, ip, &del); + if (error) + goto out_defer; + + /* Remove the mapping from the CoW fork. */ + xfs_bmap_del_extent_cow(ip, &idx, &got, &del); + + error = xfs_defer_finish(&tp, &dfops, ip); + if (error) + goto out_defer; +next_extent: + if (!xfs_iext_get_extent(ifp, idx, &got)) + break; + } + + error = xfs_trans_commit(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + if (error) + goto out; + return 0; + +out_defer: + xfs_defer_cancel(&dfops); + xfs_trans_cancel(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); +out: + trace_xfs_reflink_end_cow_error(ip, error, _RET_IP_); + return error; +} + +/* + * Free leftover CoW reservations that didn't get cleaned out. + */ +int +xfs_reflink_recover_cow( + struct xfs_mount *mp) +{ + xfs_agnumber_t agno; + int error = 0; + + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return 0; + + for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) { + error = xfs_refcount_recover_cow_leftovers(mp, agno); + if (error) + break; + } + + return error; +} + +/* + * Reflinking (Block) Ranges of Two Files Together + * + * First, ensure that the reflink flag is set on both inodes. The flag is an + * optimization to avoid unnecessary refcount btree lookups in the write path. + * + * Now we can iteratively remap the range of extents (and holes) in src to the + * corresponding ranges in dest. Let drange and srange denote the ranges of + * logical blocks in dest and src touched by the reflink operation. + * + * While the length of drange is greater than zero, + * - Read src's bmbt at the start of srange ("imap") + * - If imap doesn't exist, make imap appear to start at the end of srange + * with zero length. + * - If imap starts before srange, advance imap to start at srange. + * - If imap goes beyond srange, truncate imap to end at the end of srange. + * - Punch (imap start - srange start + imap len) blocks from dest at + * offset (drange start). + * - If imap points to a real range of pblks, + * > Increase the refcount of the imap's pblks + * > Map imap's pblks into dest at the offset + * (drange start + imap start - srange start) + * - Advance drange and srange by (imap start - srange start + imap len) + * + * Finally, if the reflink made dest longer, update both the in-core and + * on-disk file sizes. + * + * ASCII Art Demonstration: + * + * Let's say we want to reflink this source file: + * + * ----SSSSSSS-SSSSS----SSSSSS (src file) + * <--------------------> + * + * into this destination file: + * + * --DDDDDDDDDDDDDDDDDDD--DDD (dest file) + * <--------------------> + * '-' means a hole, and 'S' and 'D' are written blocks in the src and dest. + * Observe that the range has different logical offsets in either file. + * + * Consider that the first extent in the source file doesn't line up with our + * reflink range. Unmapping and remapping are separate operations, so we can + * unmap more blocks from the destination file than we remap. + * + * ----SSSSSSS-SSSSS----SSSSSS + * <-------> + * --DDDDD---------DDDDD--DDD + * <-------> + * + * Now remap the source extent into the destination file: + * + * ----SSSSSSS-SSSSS----SSSSSS + * <-------> + * --DDDDD--SSSSSSSDDDDD--DDD + * <-------> + * + * Do likewise with the second hole and extent in our range. Holes in the + * unmap range don't affect our operation. + * + * ----SSSSSSS-SSSSS----SSSSSS + * <----> + * --DDDDD--SSSSSSS-SSSSS-DDD + * <----> + * + * Finally, unmap and remap part of the third extent. This will increase the + * size of the destination file. + * + * ----SSSSSSS-SSSSS----SSSSSS + * <-----> + * --DDDDD--SSSSSSS-SSSSS----SSS + * <-----> + * + * Once we update the destination file's i_size, we're done. + */ + +/* + * Ensure the reflink bit is set in both inodes. + */ +STATIC int +xfs_reflink_set_inode_flag( + struct xfs_inode *src, + struct xfs_inode *dest) +{ + struct xfs_mount *mp = src->i_mount; + int error; + struct xfs_trans *tp; + + if (xfs_is_reflink_inode(src) && xfs_is_reflink_inode(dest)) + return 0; + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); + if (error) + goto out_error; + + /* Lock both files against IO */ + if (src->i_ino == dest->i_ino) + xfs_ilock(src, XFS_ILOCK_EXCL); + else + xfs_lock_two_inodes(src, dest, XFS_ILOCK_EXCL); + + if (!xfs_is_reflink_inode(src)) { + trace_xfs_reflink_set_inode_flag(src); + xfs_trans_ijoin(tp, src, XFS_ILOCK_EXCL); + src->i_d.di_flags2 |= XFS_DIFLAG2_REFLINK; + xfs_trans_log_inode(tp, src, XFS_ILOG_CORE); + xfs_ifork_init_cow(src); + } else + xfs_iunlock(src, XFS_ILOCK_EXCL); + + if (src->i_ino == dest->i_ino) + goto commit_flags; + + if (!xfs_is_reflink_inode(dest)) { + trace_xfs_reflink_set_inode_flag(dest); + xfs_trans_ijoin(tp, dest, XFS_ILOCK_EXCL); + dest->i_d.di_flags2 |= XFS_DIFLAG2_REFLINK; + xfs_trans_log_inode(tp, dest, XFS_ILOG_CORE); + xfs_ifork_init_cow(dest); + } else + xfs_iunlock(dest, XFS_ILOCK_EXCL); + +commit_flags: + error = xfs_trans_commit(tp); + if (error) + goto out_error; + return error; + +out_error: + trace_xfs_reflink_set_inode_flag_error(dest, error, _RET_IP_); + return error; +} + +/* + * Update destination inode size & cowextsize hint, if necessary. + */ +STATIC int +xfs_reflink_update_dest( + struct xfs_inode *dest, + xfs_off_t newlen, + xfs_extlen_t cowextsize) +{ + struct xfs_mount *mp = dest->i_mount; + struct xfs_trans *tp; + int error; + + if (newlen <= i_size_read(VFS_I(dest)) && cowextsize == 0) + return 0; + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp); + if (error) + goto out_error; + + xfs_ilock(dest, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dest, XFS_ILOCK_EXCL); + + if (newlen > i_size_read(VFS_I(dest))) { + trace_xfs_reflink_update_inode_size(dest, newlen); + i_size_write(VFS_I(dest), newlen); + dest->i_d.di_size = newlen; + } + + if (cowextsize) { + dest->i_d.di_cowextsize = cowextsize; + dest->i_d.di_flags2 |= XFS_DIFLAG2_COWEXTSIZE; + } + + xfs_trans_log_inode(tp, dest, XFS_ILOG_CORE); + + error = xfs_trans_commit(tp); + if (error) + goto out_error; + return error; + +out_error: + trace_xfs_reflink_update_inode_size_error(dest, error, _RET_IP_); + return error; +} + +/* + * Do we have enough reserve in this AG to handle a reflink? The refcount + * btree already reserved all the space it needs, but the rmap btree can grow + * infinitely, so we won't allow more reflinks when the AG is down to the + * btree reserves. + */ +static int +xfs_reflink_ag_has_free_space( + struct xfs_mount *mp, + xfs_agnumber_t agno) +{ + struct xfs_perag *pag; + int error = 0; + + if (!xfs_sb_version_hasrmapbt(&mp->m_sb)) + return 0; + + pag = xfs_perag_get(mp, agno); + if (xfs_ag_resv_critical(pag, XFS_AG_RESV_AGFL) || + xfs_ag_resv_critical(pag, XFS_AG_RESV_METADATA)) + error = -ENOSPC; + xfs_perag_put(pag); + return error; +} + +/* + * Unmap a range of blocks from a file, then map other blocks into the hole. + * The range to unmap is (destoff : destoff + srcioff + irec->br_blockcount). + * The extent irec is mapped into dest at irec->br_startoff. + */ +STATIC int +xfs_reflink_remap_extent( + struct xfs_inode *ip, + struct xfs_bmbt_irec *irec, + xfs_fileoff_t destoff, + xfs_off_t new_isize) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + xfs_fsblock_t firstfsb; + unsigned int resblks; + struct xfs_defer_ops dfops; + struct xfs_bmbt_irec uirec; + bool real_extent; + xfs_filblks_t rlen; + xfs_filblks_t unmap_len; + xfs_off_t newlen; + int error; + + unmap_len = irec->br_startoff + irec->br_blockcount - destoff; + trace_xfs_reflink_punch_range(ip, destoff, unmap_len); + + /* Only remap normal extents. */ + real_extent = (irec->br_startblock != HOLESTARTBLOCK && + irec->br_startblock != DELAYSTARTBLOCK && + !ISUNWRITTEN(irec)); + + /* No reflinking if we're low on space */ + if (real_extent) { + error = xfs_reflink_ag_has_free_space(mp, + XFS_FSB_TO_AGNO(mp, irec->br_startblock)); + if (error) + goto out; + } + + /* Start a rolling transaction to switch the mappings */ + resblks = XFS_EXTENTADD_SPACE_RES(ip->i_mount, XFS_DATA_FORK); + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, &tp); + if (error) + goto out; + + xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, 0); + + /* If we're not just clearing space, then do we have enough quota? */ + if (real_extent) { + error = xfs_trans_reserve_quota_nblks(tp, ip, + irec->br_blockcount, 0, XFS_QMOPT_RES_REGBLKS); + if (error) + goto out_cancel; + } + + trace_xfs_reflink_remap(ip, irec->br_startoff, + irec->br_blockcount, irec->br_startblock); + + /* Unmap the old blocks in the data fork. */ + rlen = unmap_len; + while (rlen) { + xfs_defer_init(&dfops, &firstfsb); + error = __xfs_bunmapi(tp, ip, destoff, &rlen, 0, 1, + &firstfsb, &dfops); + if (error) + goto out_defer; + + /* + * Trim the extent to whatever got unmapped. + * Remember, bunmapi works backwards. + */ + uirec.br_startblock = irec->br_startblock + rlen; + uirec.br_startoff = irec->br_startoff + rlen; + uirec.br_blockcount = unmap_len - rlen; + unmap_len = rlen; + + /* If this isn't a real mapping, we're done. */ + if (!real_extent || uirec.br_blockcount == 0) + goto next_extent; + + trace_xfs_reflink_remap(ip, uirec.br_startoff, + uirec.br_blockcount, uirec.br_startblock); + + /* Update the refcount tree */ + error = xfs_refcount_increase_extent(mp, &dfops, &uirec); + if (error) + goto out_defer; + + /* Map the new blocks into the data fork. */ + error = xfs_bmap_map_extent(mp, &dfops, ip, &uirec); + if (error) + goto out_defer; + + /* Update quota accounting. */ + xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, + uirec.br_blockcount); + + /* Update dest isize if needed. */ + newlen = XFS_FSB_TO_B(mp, + uirec.br_startoff + uirec.br_blockcount); + newlen = min_t(xfs_off_t, newlen, new_isize); + if (newlen > i_size_read(VFS_I(ip))) { + trace_xfs_reflink_update_inode_size(ip, newlen); + i_size_write(VFS_I(ip), newlen); + ip->i_d.di_size = newlen; + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + } + +next_extent: + /* Process all the deferred stuff. */ + error = xfs_defer_finish(&tp, &dfops, ip); + if (error) + goto out_defer; + } + + error = xfs_trans_commit(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + if (error) + goto out; + return 0; + +out_defer: + xfs_defer_cancel(&dfops); +out_cancel: + xfs_trans_cancel(tp); + xfs_iunlock(ip, XFS_ILOCK_EXCL); +out: + trace_xfs_reflink_remap_extent_error(ip, error, _RET_IP_); + return error; +} + +/* + * Iteratively remap one file's extents (and holes) to another's. + */ +STATIC int +xfs_reflink_remap_blocks( + struct xfs_inode *src, + xfs_fileoff_t srcoff, + struct xfs_inode *dest, + xfs_fileoff_t destoff, + xfs_filblks_t len, + xfs_off_t new_isize) +{ + struct xfs_bmbt_irec imap; + int nimaps; + int error = 0; + xfs_filblks_t range_len; + + /* drange = (destoff, destoff + len); srange = (srcoff, srcoff + len) */ + while (len) { + trace_xfs_reflink_remap_blocks_loop(src, srcoff, len, + dest, destoff); + /* Read extent from the source file */ + nimaps = 1; + xfs_ilock(src, XFS_ILOCK_EXCL); + error = xfs_bmapi_read(src, srcoff, len, &imap, &nimaps, 0); + xfs_iunlock(src, XFS_ILOCK_EXCL); + if (error) + goto err; + ASSERT(nimaps == 1); + + trace_xfs_reflink_remap_imap(src, srcoff, len, XFS_IO_OVERWRITE, + &imap); + + /* Translate imap into the destination file. */ + range_len = imap.br_startoff + imap.br_blockcount - srcoff; + imap.br_startoff += destoff - srcoff; + + /* Clear dest from destoff to the end of imap and map it in. */ + error = xfs_reflink_remap_extent(dest, &imap, destoff, + new_isize); + if (error) + goto err; + + if (fatal_signal_pending(current)) { + error = -EINTR; + goto err; + } + + /* Advance drange/srange */ + srcoff += range_len; + destoff += range_len; + len -= range_len; + } + + return 0; + +err: + trace_xfs_reflink_remap_blocks_error(dest, error, _RET_IP_); + return error; +} + +/* + * Link a range of blocks from one file to another. + */ +int +xfs_reflink_remap_range( + struct file *file_in, + loff_t pos_in, + struct file *file_out, + loff_t pos_out, + u64 len, + bool is_dedupe) +{ + struct inode *inode_in = file_inode(file_in); + struct xfs_inode *src = XFS_I(inode_in); + struct inode *inode_out = file_inode(file_out); + struct xfs_inode *dest = XFS_I(inode_out); + struct xfs_mount *mp = src->i_mount; + bool same_inode = (inode_in == inode_out); + xfs_fileoff_t sfsbno, dfsbno; + xfs_filblks_t fsblen; + xfs_extlen_t cowextsize; + ssize_t ret; + + if (!xfs_sb_version_hasreflink(&mp->m_sb)) + return -EOPNOTSUPP; + + if (XFS_FORCED_SHUTDOWN(mp)) + return -EIO; + + /* Lock both files against IO */ + lock_two_nondirectories(inode_in, inode_out); + if (same_inode) + xfs_ilock(src, XFS_MMAPLOCK_EXCL); + else + xfs_lock_two_inodes(src, dest, XFS_MMAPLOCK_EXCL); + + /* Check file eligibility and prepare for block sharing. */ + ret = -EINVAL; + /* Don't reflink realtime inodes */ + if (XFS_IS_REALTIME_INODE(src) || XFS_IS_REALTIME_INODE(dest)) + goto out_unlock; + + /* Don't share DAX file data for now. */ + if (IS_DAX(inode_in) || IS_DAX(inode_out)) + goto out_unlock; + + ret = vfs_clone_file_prep_inodes(inode_in, pos_in, inode_out, pos_out, + &len, is_dedupe); + if (ret <= 0) + goto out_unlock; + + trace_xfs_reflink_remap_range(src, pos_in, len, dest, pos_out); + + /* Set flags and remap blocks. */ + ret = xfs_reflink_set_inode_flag(src, dest); + if (ret) + goto out_unlock; + + dfsbno = XFS_B_TO_FSBT(mp, pos_out); + sfsbno = XFS_B_TO_FSBT(mp, pos_in); + fsblen = XFS_B_TO_FSB(mp, len); + ret = xfs_reflink_remap_blocks(src, sfsbno, dest, dfsbno, fsblen, + pos_out + len); + if (ret) + goto out_unlock; + + /* Zap any page cache for the destination file's range. */ + truncate_inode_pages_range(&inode_out->i_data, pos_out, + PAGE_ALIGN(pos_out + len) - 1); + + /* + * Carry the cowextsize hint from src to dest if we're sharing the + * entire source file to the entire destination file, the source file + * has a cowextsize hint, and the destination file does not. + */ + cowextsize = 0; + if (pos_in == 0 && len == i_size_read(inode_in) && + (src->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) && + pos_out == 0 && len >= i_size_read(inode_out) && + !(dest->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE)) + cowextsize = src->i_d.di_cowextsize; + + ret = xfs_reflink_update_dest(dest, pos_out + len, cowextsize); + +out_unlock: + xfs_iunlock(src, XFS_MMAPLOCK_EXCL); + if (!same_inode) + xfs_iunlock(dest, XFS_MMAPLOCK_EXCL); + unlock_two_nondirectories(inode_in, inode_out); + if (ret) + trace_xfs_reflink_remap_range_error(dest, ret, _RET_IP_); + return ret; +} + +/* + * The user wants to preemptively CoW all shared blocks in this file, + * which enables us to turn off the reflink flag. Iterate all + * extents which are not prealloc/delalloc to see which ranges are + * mentioned in the refcount tree, then read those blocks into the + * pagecache, dirty them, fsync them back out, and then we can update + * the inode flag. What happens if we run out of memory? :) + */ +STATIC int +xfs_reflink_dirty_extents( + struct xfs_inode *ip, + xfs_fileoff_t fbno, + xfs_filblks_t end, + xfs_off_t isize) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_agnumber_t agno; + xfs_agblock_t agbno; + xfs_extlen_t aglen; + xfs_agblock_t rbno; + xfs_extlen_t rlen; + xfs_off_t fpos; + xfs_off_t flen; + struct xfs_bmbt_irec map[2]; + int nmaps; + int error = 0; + + while (end - fbno > 0) { + nmaps = 1; + /* + * Look for extents in the file. Skip holes, delalloc, or + * unwritten extents; they can't be reflinked. + */ + error = xfs_bmapi_read(ip, fbno, end - fbno, map, &nmaps, 0); + if (error) + goto out; + if (nmaps == 0) + break; + if (map[0].br_startblock == HOLESTARTBLOCK || + map[0].br_startblock == DELAYSTARTBLOCK || + ISUNWRITTEN(&map[0])) + goto next; + + map[1] = map[0]; + while (map[1].br_blockcount) { + agno = XFS_FSB_TO_AGNO(mp, map[1].br_startblock); + agbno = XFS_FSB_TO_AGBNO(mp, map[1].br_startblock); + aglen = map[1].br_blockcount; + + error = xfs_reflink_find_shared(mp, agno, agbno, aglen, + &rbno, &rlen, true); + if (error) + goto out; + if (rbno == NULLAGBLOCK) + break; + + /* Dirty the pages */ + xfs_iunlock(ip, XFS_ILOCK_EXCL); + fpos = XFS_FSB_TO_B(mp, map[1].br_startoff + + (rbno - agbno)); + flen = XFS_FSB_TO_B(mp, rlen); + if (fpos + flen > isize) + flen = isize - fpos; + error = iomap_file_dirty(VFS_I(ip), fpos, flen, + &xfs_iomap_ops); + xfs_ilock(ip, XFS_ILOCK_EXCL); + if (error) + goto out; + + map[1].br_blockcount -= (rbno - agbno + rlen); + map[1].br_startoff += (rbno - agbno + rlen); + map[1].br_startblock += (rbno - agbno + rlen); + } + +next: + fbno = map[0].br_startoff + map[0].br_blockcount; + } +out: + return error; +} + +/* Clear the inode reflink flag if there are no shared extents. */ +int +xfs_reflink_clear_inode_flag( + struct xfs_inode *ip, + struct xfs_trans **tpp) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_fileoff_t fbno; + xfs_filblks_t end; + xfs_agnumber_t agno; + xfs_agblock_t agbno; + xfs_extlen_t aglen; + xfs_agblock_t rbno; + xfs_extlen_t rlen; + struct xfs_bmbt_irec map; + int nmaps; + int error = 0; + + ASSERT(xfs_is_reflink_inode(ip)); + + fbno = 0; + end = XFS_B_TO_FSB(mp, i_size_read(VFS_I(ip))); + while (end - fbno > 0) { + nmaps = 1; + /* + * Look for extents in the file. Skip holes, delalloc, or + * unwritten extents; they can't be reflinked. + */ + error = xfs_bmapi_read(ip, fbno, end - fbno, &map, &nmaps, 0); + if (error) + return error; + if (nmaps == 0) + break; + if (map.br_startblock == HOLESTARTBLOCK || + map.br_startblock == DELAYSTARTBLOCK || + ISUNWRITTEN(&map)) + goto next; + + agno = XFS_FSB_TO_AGNO(mp, map.br_startblock); + agbno = XFS_FSB_TO_AGBNO(mp, map.br_startblock); + aglen = map.br_blockcount; + + error = xfs_reflink_find_shared(mp, agno, agbno, aglen, + &rbno, &rlen, false); + if (error) + return error; + /* Is there still a shared block here? */ + if (rbno != NULLAGBLOCK) + return 0; +next: + fbno = map.br_startoff + map.br_blockcount; + } + + /* + * We didn't find any shared blocks so turn off the reflink flag. + * First, get rid of any leftover CoW mappings. + */ + error = xfs_reflink_cancel_cow_blocks(ip, tpp, 0, NULLFILEOFF); + if (error) + return error; + + /* Clear the inode flag. */ + trace_xfs_reflink_unset_inode_flag(ip); + ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK; + xfs_inode_clear_cowblocks_tag(ip); + xfs_trans_ijoin(*tpp, ip, 0); + xfs_trans_log_inode(*tpp, ip, XFS_ILOG_CORE); + + return error; +} + +/* + * Clear the inode reflink flag if there are no shared extents and the size + * hasn't changed. + */ +STATIC int +xfs_reflink_try_clear_inode_flag( + struct xfs_inode *ip) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp; + int error = 0; + + /* Start a rolling transaction to remove the mappings */ + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0, &tp); + if (error) + return error; + + xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, 0); + + error = xfs_reflink_clear_inode_flag(ip, &tp); + if (error) + goto cancel; + + error = xfs_trans_commit(tp); + if (error) + goto out; + + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return 0; +cancel: + xfs_trans_cancel(tp); +out: + xfs_iunlock(ip, XFS_ILOCK_EXCL); + return error; +} + +/* + * Pre-COW all shared blocks within a given byte range of a file and turn off + * the reflink flag if we unshare all of the file's blocks. + */ +int +xfs_reflink_unshare( + struct xfs_inode *ip, + xfs_off_t offset, + xfs_off_t len) +{ + struct xfs_mount *mp = ip->i_mount; + xfs_fileoff_t fbno; + xfs_filblks_t end; + xfs_off_t isize; + int error; + + if (!xfs_is_reflink_inode(ip)) + return 0; + + trace_xfs_reflink_unshare(ip, offset, len); + + inode_dio_wait(VFS_I(ip)); + + /* Try to CoW the selected ranges */ + xfs_ilock(ip, XFS_ILOCK_EXCL); + fbno = XFS_B_TO_FSBT(mp, offset); + isize = i_size_read(VFS_I(ip)); + end = XFS_B_TO_FSB(mp, offset + len); + error = xfs_reflink_dirty_extents(ip, fbno, end, isize); + if (error) + goto out_unlock; + xfs_iunlock(ip, XFS_ILOCK_EXCL); + + /* Wait for the IO to finish */ + error = filemap_write_and_wait(VFS_I(ip)->i_mapping); + if (error) + goto out; + + /* Turn off the reflink flag if possible. */ + error = xfs_reflink_try_clear_inode_flag(ip); + if (error) + goto out; + + return 0; + +out_unlock: + xfs_iunlock(ip, XFS_ILOCK_EXCL); +out: + trace_xfs_reflink_unshare_error(ip, error, _RET_IP_); + return error; +} diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h new file mode 100644 index 000000000000..aa6a4d64bd35 --- /dev/null +++ b/fs/xfs/xfs_reflink.h @@ -0,0 +1,53 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#ifndef __XFS_REFLINK_H +#define __XFS_REFLINK_H 1 + +extern int xfs_reflink_find_shared(struct xfs_mount *mp, xfs_agnumber_t agno, + xfs_agblock_t agbno, xfs_extlen_t aglen, xfs_agblock_t *fbno, + xfs_extlen_t *flen, bool find_maximal); +extern int xfs_reflink_trim_around_shared(struct xfs_inode *ip, + struct xfs_bmbt_irec *irec, bool *shared, bool *trimmed); + +extern int xfs_reflink_reserve_cow(struct xfs_inode *ip, + struct xfs_bmbt_irec *imap, bool *shared); +extern int xfs_reflink_allocate_cow_range(struct xfs_inode *ip, + xfs_off_t offset, xfs_off_t count); +extern bool xfs_reflink_find_cow_mapping(struct xfs_inode *ip, xfs_off_t offset, + struct xfs_bmbt_irec *imap); +extern void xfs_reflink_trim_irec_to_next_cow(struct xfs_inode *ip, + xfs_fileoff_t offset_fsb, struct xfs_bmbt_irec *imap); + +extern int xfs_reflink_cancel_cow_blocks(struct xfs_inode *ip, + struct xfs_trans **tpp, xfs_fileoff_t offset_fsb, + xfs_fileoff_t end_fsb); +extern int xfs_reflink_cancel_cow_range(struct xfs_inode *ip, xfs_off_t offset, + xfs_off_t count); +extern int xfs_reflink_end_cow(struct xfs_inode *ip, xfs_off_t offset, + xfs_off_t count); +extern int xfs_reflink_recover_cow(struct xfs_mount *mp); +extern int xfs_reflink_remap_range(struct file *file_in, loff_t pos_in, + struct file *file_out, loff_t pos_out, u64 len, bool is_dedupe); +extern int xfs_reflink_clear_inode_flag(struct xfs_inode *ip, + struct xfs_trans **tpp); +extern int xfs_reflink_unshare(struct xfs_inode *ip, xfs_off_t offset, + xfs_off_t len); + +#endif /* __XFS_REFLINK_H */ diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 2500f28689d5..73c827831551 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -51,28 +51,16 @@ xfs_rui_item_free( kmem_zone_free(xfs_rui_zone, ruip); } -/* - * This returns the number of iovecs needed to log the given rui item. - * We only need 1 iovec for an rui item. It just logs the rui_log_format - * structure. - */ -static inline int -xfs_rui_item_sizeof( - struct xfs_rui_log_item *ruip) -{ - return sizeof(struct xfs_rui_log_format) + - (ruip->rui_format.rui_nextents - 1) * - sizeof(struct xfs_map_extent); -} - STATIC void xfs_rui_item_size( struct xfs_log_item *lip, int *nvecs, int *nbytes) { + struct xfs_rui_log_item *ruip = RUI_ITEM(lip); + *nvecs += 1; - *nbytes += xfs_rui_item_sizeof(RUI_ITEM(lip)); + *nbytes += xfs_rui_log_format_sizeof(ruip->rui_format.rui_nextents); } /* @@ -97,7 +85,7 @@ xfs_rui_item_format( ruip->rui_format.rui_size = 1; xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_RUI_FORMAT, &ruip->rui_format, - xfs_rui_item_sizeof(ruip)); + xfs_rui_log_format_sizeof(ruip->rui_format.rui_nextents)); } /* @@ -205,16 +193,12 @@ xfs_rui_init( { struct xfs_rui_log_item *ruip; - uint size; ASSERT(nextents > 0); - if (nextents > XFS_RUI_MAX_FAST_EXTENTS) { - size = (uint)(sizeof(struct xfs_rui_log_item) + - ((nextents - 1) * sizeof(struct xfs_map_extent))); - ruip = kmem_zalloc(size, KM_SLEEP); - } else { + if (nextents > XFS_RUI_MAX_FAST_EXTENTS) + ruip = kmem_zalloc(xfs_rui_log_item_sizeof(nextents), KM_SLEEP); + else ruip = kmem_zone_zalloc(xfs_rui_zone, KM_SLEEP); - } xfs_log_item_init(mp, &ruip->rui_item, XFS_LI_RUI, &xfs_rui_item_ops); ruip->rui_format.rui_nextents = nextents; @@ -239,14 +223,12 @@ xfs_rui_copy_format( uint len; src_rui_fmt = buf->i_addr; - len = sizeof(struct xfs_rui_log_format) + - (src_rui_fmt->rui_nextents - 1) * - sizeof(struct xfs_map_extent); + len = xfs_rui_log_format_sizeof(src_rui_fmt->rui_nextents); if (buf->i_len != len) return -EFSCORRUPTED; - memcpy((char *)dst_rui_fmt, (char *)src_rui_fmt, len); + memcpy(dst_rui_fmt, src_rui_fmt, len); return 0; } @@ -459,8 +441,11 @@ xfs_rui_recover( XFS_FSB_TO_DADDR(mp, rmap->me_startblock)); switch (rmap->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) { case XFS_RMAP_EXTENT_MAP: + case XFS_RMAP_EXTENT_MAP_SHARED: case XFS_RMAP_EXTENT_UNMAP: + case XFS_RMAP_EXTENT_UNMAP_SHARED: case XFS_RMAP_EXTENT_CONVERT: + case XFS_RMAP_EXTENT_CONVERT_SHARED: case XFS_RMAP_EXTENT_ALLOC: case XFS_RMAP_EXTENT_FREE: op_ok = true; @@ -499,12 +484,21 @@ xfs_rui_recover( case XFS_RMAP_EXTENT_MAP: type = XFS_RMAP_MAP; break; + case XFS_RMAP_EXTENT_MAP_SHARED: + type = XFS_RMAP_MAP_SHARED; + break; case XFS_RMAP_EXTENT_UNMAP: type = XFS_RMAP_UNMAP; break; + case XFS_RMAP_EXTENT_UNMAP_SHARED: + type = XFS_RMAP_UNMAP_SHARED; + break; case XFS_RMAP_EXTENT_CONVERT: type = XFS_RMAP_CONVERT; break; + case XFS_RMAP_EXTENT_CONVERT_SHARED: + type = XFS_RMAP_CONVERT_SHARED; + break; case XFS_RMAP_EXTENT_ALLOC: type = XFS_RMAP_ALLOC; break; diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h index aefcc3a318a5..340c968e1f9c 100644 --- a/fs/xfs/xfs_rmap_item.h +++ b/fs/xfs/xfs_rmap_item.h @@ -70,6 +70,14 @@ struct xfs_rui_log_item { struct xfs_rui_log_format rui_format; }; +static inline size_t +xfs_rui_log_item_sizeof( + unsigned int nr) +{ + return offsetof(struct xfs_rui_log_item, rui_format) + + xfs_rui_log_format_sizeof(nr); +} + /* * This is the "rmap update done" log item. It is used to log the fact that * some rmapbt updates mentioned in an earlier rui item have been performed. diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c index 6e812fe0fd43..f11282c96887 100644 --- a/fs/xfs/xfs_stats.c +++ b/fs/xfs/xfs_stats.c @@ -62,6 +62,7 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf) { "ibt2", XFSSTAT_END_IBT_V2 }, { "fibt2", XFSSTAT_END_FIBT_V2 }, { "rmapbt", XFSSTAT_END_RMAP_V2 }, + { "refcntbt", XFSSTAT_END_REFCOUNT }, /* we print both series of quota information together */ { "qm", XFSSTAT_END_QM }, }; @@ -79,9 +80,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf) } /* extra precision counters */ for_each_possible_cpu(i) { - xs_xstrat_bytes += per_cpu_ptr(stats, i)->xs_xstrat_bytes; - xs_write_bytes += per_cpu_ptr(stats, i)->xs_write_bytes; - xs_read_bytes += per_cpu_ptr(stats, i)->xs_read_bytes; + xs_xstrat_bytes += per_cpu_ptr(stats, i)->s.xs_xstrat_bytes; + xs_write_bytes += per_cpu_ptr(stats, i)->s.xs_write_bytes; + xs_read_bytes += per_cpu_ptr(stats, i)->s.xs_read_bytes; } len += snprintf(buf + len, PATH_MAX-len, "xpc %Lu %Lu %Lu\n", @@ -105,9 +106,9 @@ void xfs_stats_clearall(struct xfsstats __percpu *stats) for_each_possible_cpu(c) { preempt_disable(); /* save vn_active, it's a universal truth! */ - vn_active = per_cpu_ptr(stats, c)->vn_active; + vn_active = per_cpu_ptr(stats, c)->s.vn_active; memset(per_cpu_ptr(stats, c), 0, sizeof(*stats)); - per_cpu_ptr(stats, c)->vn_active = vn_active; + per_cpu_ptr(stats, c)->s.vn_active = vn_active; preempt_enable(); } } diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h index 657865f51e78..375840f5a99a 100644 --- a/fs/xfs/xfs_stats.h +++ b/fs/xfs/xfs_stats.h @@ -22,9 +22,37 @@ #include <linux/percpu.h> /* + * The btree stats arrays have fixed offsets for the different stats. We + * store the base index in the btree cursor via XFS_STATS_CALC_INDEX() and + * that allows us to use fixed offsets into the stats array for each btree + * stat. These index offsets are defined in the order they will be emitted + * in the stats files, so it is possible to add new btree stat types by + * appending to the enum list below. + */ +enum { + __XBTS_lookup = 0, + __XBTS_compare = 1, + __XBTS_insrec = 2, + __XBTS_delrec = 3, + __XBTS_newroot = 4, + __XBTS_killroot = 5, + __XBTS_increment = 6, + __XBTS_decrement = 7, + __XBTS_lshift = 8, + __XBTS_rshift = 9, + __XBTS_split = 10, + __XBTS_join = 11, + __XBTS_alloc = 12, + __XBTS_free = 13, + __XBTS_moves = 14, + + __XBTS_MAX = 15, +}; + +/* * XFS global statistics */ -struct xfsstats { +struct __xfsstats { # define XFSSTAT_END_EXTENT_ALLOC 4 __uint32_t xs_allocx; __uint32_t xs_allocb; @@ -117,103 +145,21 @@ struct xfsstats { __uint32_t xb_page_found; __uint32_t xb_get_read; /* Version 2 btree counters */ -#define XFSSTAT_END_ABTB_V2 (XFSSTAT_END_BUF+15) - __uint32_t xs_abtb_2_lookup; - __uint32_t xs_abtb_2_compare; - __uint32_t xs_abtb_2_insrec; - __uint32_t xs_abtb_2_delrec; - __uint32_t xs_abtb_2_newroot; - __uint32_t xs_abtb_2_killroot; - __uint32_t xs_abtb_2_increment; - __uint32_t xs_abtb_2_decrement; - __uint32_t xs_abtb_2_lshift; - __uint32_t xs_abtb_2_rshift; - __uint32_t xs_abtb_2_split; - __uint32_t xs_abtb_2_join; - __uint32_t xs_abtb_2_alloc; - __uint32_t xs_abtb_2_free; - __uint32_t xs_abtb_2_moves; -#define XFSSTAT_END_ABTC_V2 (XFSSTAT_END_ABTB_V2+15) - __uint32_t xs_abtc_2_lookup; - __uint32_t xs_abtc_2_compare; - __uint32_t xs_abtc_2_insrec; - __uint32_t xs_abtc_2_delrec; - __uint32_t xs_abtc_2_newroot; - __uint32_t xs_abtc_2_killroot; - __uint32_t xs_abtc_2_increment; - __uint32_t xs_abtc_2_decrement; - __uint32_t xs_abtc_2_lshift; - __uint32_t xs_abtc_2_rshift; - __uint32_t xs_abtc_2_split; - __uint32_t xs_abtc_2_join; - __uint32_t xs_abtc_2_alloc; - __uint32_t xs_abtc_2_free; - __uint32_t xs_abtc_2_moves; -#define XFSSTAT_END_BMBT_V2 (XFSSTAT_END_ABTC_V2+15) - __uint32_t xs_bmbt_2_lookup; - __uint32_t xs_bmbt_2_compare; - __uint32_t xs_bmbt_2_insrec; - __uint32_t xs_bmbt_2_delrec; - __uint32_t xs_bmbt_2_newroot; - __uint32_t xs_bmbt_2_killroot; - __uint32_t xs_bmbt_2_increment; - __uint32_t xs_bmbt_2_decrement; - __uint32_t xs_bmbt_2_lshift; - __uint32_t xs_bmbt_2_rshift; - __uint32_t xs_bmbt_2_split; - __uint32_t xs_bmbt_2_join; - __uint32_t xs_bmbt_2_alloc; - __uint32_t xs_bmbt_2_free; - __uint32_t xs_bmbt_2_moves; -#define XFSSTAT_END_IBT_V2 (XFSSTAT_END_BMBT_V2+15) - __uint32_t xs_ibt_2_lookup; - __uint32_t xs_ibt_2_compare; - __uint32_t xs_ibt_2_insrec; - __uint32_t xs_ibt_2_delrec; - __uint32_t xs_ibt_2_newroot; - __uint32_t xs_ibt_2_killroot; - __uint32_t xs_ibt_2_increment; - __uint32_t xs_ibt_2_decrement; - __uint32_t xs_ibt_2_lshift; - __uint32_t xs_ibt_2_rshift; - __uint32_t xs_ibt_2_split; - __uint32_t xs_ibt_2_join; - __uint32_t xs_ibt_2_alloc; - __uint32_t xs_ibt_2_free; - __uint32_t xs_ibt_2_moves; -#define XFSSTAT_END_FIBT_V2 (XFSSTAT_END_IBT_V2+15) - __uint32_t xs_fibt_2_lookup; - __uint32_t xs_fibt_2_compare; - __uint32_t xs_fibt_2_insrec; - __uint32_t xs_fibt_2_delrec; - __uint32_t xs_fibt_2_newroot; - __uint32_t xs_fibt_2_killroot; - __uint32_t xs_fibt_2_increment; - __uint32_t xs_fibt_2_decrement; - __uint32_t xs_fibt_2_lshift; - __uint32_t xs_fibt_2_rshift; - __uint32_t xs_fibt_2_split; - __uint32_t xs_fibt_2_join; - __uint32_t xs_fibt_2_alloc; - __uint32_t xs_fibt_2_free; - __uint32_t xs_fibt_2_moves; -#define XFSSTAT_END_RMAP_V2 (XFSSTAT_END_FIBT_V2+15) - __uint32_t xs_rmap_2_lookup; - __uint32_t xs_rmap_2_compare; - __uint32_t xs_rmap_2_insrec; - __uint32_t xs_rmap_2_delrec; - __uint32_t xs_rmap_2_newroot; - __uint32_t xs_rmap_2_killroot; - __uint32_t xs_rmap_2_increment; - __uint32_t xs_rmap_2_decrement; - __uint32_t xs_rmap_2_lshift; - __uint32_t xs_rmap_2_rshift; - __uint32_t xs_rmap_2_split; - __uint32_t xs_rmap_2_join; - __uint32_t xs_rmap_2_alloc; - __uint32_t xs_rmap_2_free; - __uint32_t xs_rmap_2_moves; -#define XFSSTAT_END_XQMSTAT (XFSSTAT_END_RMAP_V2+6) +#define XFSSTAT_END_ABTB_V2 (XFSSTAT_END_BUF + __XBTS_MAX) + __uint32_t xs_abtb_2[__XBTS_MAX]; +#define XFSSTAT_END_ABTC_V2 (XFSSTAT_END_ABTB_V2 + __XBTS_MAX) + __uint32_t xs_abtc_2[__XBTS_MAX]; +#define XFSSTAT_END_BMBT_V2 (XFSSTAT_END_ABTC_V2 + __XBTS_MAX) + __uint32_t xs_bmbt_2[__XBTS_MAX]; +#define XFSSTAT_END_IBT_V2 (XFSSTAT_END_BMBT_V2 + __XBTS_MAX) + __uint32_t xs_ibt_2[__XBTS_MAX]; +#define XFSSTAT_END_FIBT_V2 (XFSSTAT_END_IBT_V2 + __XBTS_MAX) + __uint32_t xs_fibt_2[__XBTS_MAX]; +#define XFSSTAT_END_RMAP_V2 (XFSSTAT_END_FIBT_V2 + __XBTS_MAX) + __uint32_t xs_rmap_2[__XBTS_MAX]; +#define XFSSTAT_END_REFCOUNT (XFSSTAT_END_RMAP_V2 + __XBTS_MAX) + __uint32_t xs_refcbt_2[__XBTS_MAX]; +#define XFSSTAT_END_XQMSTAT (XFSSTAT_END_REFCOUNT + 6) __uint32_t xs_qm_dqreclaims; __uint32_t xs_qm_dqreclaim_misses; __uint32_t xs_qm_dquot_dups; @@ -229,26 +175,58 @@ struct xfsstats { __uint64_t xs_read_bytes; }; +struct xfsstats { + union { + struct __xfsstats s; + uint32_t a[XFSSTAT_END_XQMSTAT]; + }; +}; + +/* + * simple wrapper for getting the array index of s struct member offset + */ +#define XFS_STATS_CALC_INDEX(member) \ + (offsetof(struct __xfsstats, member) / (int)sizeof(__uint32_t)) + + int xfs_stats_format(struct xfsstats __percpu *stats, char *buf); void xfs_stats_clearall(struct xfsstats __percpu *stats); extern struct xstats xfsstats; #define XFS_STATS_INC(mp, v) \ do { \ - per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v++; \ - per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v++; \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v++; \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v++; \ } while (0) #define XFS_STATS_DEC(mp, v) \ do { \ - per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v--; \ - per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v--; \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v--; \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v--; \ } while (0) #define XFS_STATS_ADD(mp, v, inc) \ do { \ - per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v += (inc); \ - per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v += (inc); \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v += (inc); \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v += (inc); \ +} while (0) + +#define XFS_STATS_INC_OFF(mp, off) \ +do { \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off]++; \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off]++; \ +} while (0) + +#define XFS_STATS_DEC_OFF(mp, off) \ +do { \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off]; \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off]; \ +} while (0) + +#define XFS_STATS_ADD_OFF(mp, off, inc) \ +do { \ + per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off] += (inc); \ + per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off] += (inc); \ } while (0) #if defined(CONFIG_PROC_FS) diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index fd6be45b3a1e..eecbaac08eba 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -47,6 +47,9 @@ #include "xfs_sysfs.h" #include "xfs_ondisk.h" #include "xfs_rmap_item.h" +#include "xfs_refcount_item.h" +#include "xfs_bmap_item.h" +#include "xfs_reflink.h" #include <linux/namei.h> #include <linux/init.h> @@ -101,9 +104,6 @@ static const match_table_t tokens = { {Opt_sysvgroups,"sysvgroups"}, /* group-ID from current process */ {Opt_allocsize, "allocsize=%s"},/* preferred allocation size */ {Opt_norecovery,"norecovery"}, /* don't run XFS recovery */ - {Opt_barrier, "barrier"}, /* use writer barriers for log write and - * unwritten extent conversion */ - {Opt_nobarrier, "nobarrier"}, /* .. disable */ {Opt_inode64, "inode64"}, /* inodes can be allocated anywhere */ {Opt_inode32, "inode32"}, /* inode allocation limited to * XFS_MAXINUMBER_32 */ @@ -131,6 +131,12 @@ static const match_table_t tokens = { {Opt_nodiscard, "nodiscard"}, /* Do not discard unused blocks */ {Opt_dax, "dax"}, /* Enable direct access to bdev pages */ + + /* Deprecated mount options scheduled for removal */ + {Opt_barrier, "barrier"}, /* use writer barriers for log write and + * unwritten extent conversion */ + {Opt_nobarrier, "nobarrier"}, /* .. disable */ + {Opt_err, NULL}, }; @@ -298,12 +304,6 @@ xfs_parseargs( case Opt_nouuid: mp->m_flags |= XFS_MOUNT_NOUUID; break; - case Opt_barrier: - mp->m_flags |= XFS_MOUNT_BARRIER; - break; - case Opt_nobarrier: - mp->m_flags &= ~XFS_MOUNT_BARRIER; - break; case Opt_ikeep: mp->m_flags |= XFS_MOUNT_IKEEP; break; @@ -371,6 +371,14 @@ xfs_parseargs( mp->m_flags |= XFS_MOUNT_DAX; break; #endif + case Opt_barrier: + xfs_warn(mp, "%s option is deprecated, ignoring.", p); + mp->m_flags |= XFS_MOUNT_BARRIER; + break; + case Opt_nobarrier: + xfs_warn(mp, "%s option is deprecated, ignoring.", p); + mp->m_flags &= ~XFS_MOUNT_BARRIER; + break; default: xfs_warn(mp, "unknown mount option [%s].", p); return -EINVAL; @@ -936,13 +944,22 @@ xfs_fs_destroy_inode( struct inode *inode) { struct xfs_inode *ip = XFS_I(inode); + int error; trace_xfs_destroy_inode(ip); - ASSERT(!rwsem_is_locked(&ip->i_iolock.mr_lock)); + ASSERT(!rwsem_is_locked(&inode->i_rwsem)); XFS_STATS_INC(ip->i_mount, vn_rele); XFS_STATS_INC(ip->i_mount, vn_remove); + if (xfs_is_reflink_inode(ip)) { + error = xfs_reflink_cancel_cow_range(ip, 0, NULLFILEOFF); + if (error && !XFS_FORCED_SHUTDOWN(ip->i_mount)) + xfs_warn(ip->i_mount, +"Error %d while evicting CoW blocks for inode %llu.", + error, ip->i_ino); + } + xfs_inactive(ip); ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) || ip->i_delayed_blks == 0); @@ -1006,6 +1023,16 @@ xfs_fs_drop_inode( { struct xfs_inode *ip = XFS_I(inode); + /* + * If this unlinked inode is in the middle of recovery, don't + * drop the inode just yet; log recovery will take care of + * that. See the comment for this inode flag. + */ + if (ip->i_flags & XFS_IRECOVERY) { + ASSERT(ip->i_mount->m_log->l_flags & XLOG_RECOVERY_NEEDED); + return 0; + } + return generic_drop_inode(inode) || (ip->i_flags & XFS_IDONTCACHE); } @@ -1137,7 +1164,7 @@ xfs_restore_resvblks(struct xfs_mount *mp) * Note: xfs_log_quiesce() stops background log work - the callers must ensure * it is started again when appropriate. */ -static void +void xfs_quiesce_attr( struct xfs_mount *mp) { @@ -1216,9 +1243,11 @@ xfs_fs_remount( token = match_token(p, tokens, args); switch (token) { case Opt_barrier: + xfs_warn(mp, "%s option is deprecated, ignoring.", p); mp->m_flags |= XFS_MOUNT_BARRIER; break; case Opt_nobarrier: + xfs_warn(mp, "%s option is deprecated, ignoring.", p); mp->m_flags &= ~XFS_MOUNT_BARRIER; break; case Opt_inode64: @@ -1296,10 +1325,31 @@ xfs_fs_remount( xfs_restore_resvblks(mp); xfs_log_work_queue(mp); xfs_queue_eofblocks(mp); + + /* Recover any CoW blocks that never got remapped. */ + error = xfs_reflink_recover_cow(mp); + if (error) { + xfs_err(mp, + "Error %d recovering leftover CoW allocations.", error); + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + return error; + } + + /* Create the per-AG metadata reservation pool .*/ + error = xfs_fs_reserve_ag_blocks(mp); + if (error && error != -ENOSPC) + return error; } /* rw -> ro */ if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & MS_RDONLY)) { + /* Free the per-AG metadata reservation pool. */ + error = xfs_fs_unreserve_ag_blocks(mp); + if (error) { + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + return error; + } + /* * Before we sync the metadata, we need to free up the reserve * block pool so that the used block count in the superblock on @@ -1490,6 +1540,7 @@ xfs_fs_fill_super( atomic_set(&mp->m_active_trans, 0); INIT_DELAYED_WORK(&mp->m_reclaim_work, xfs_reclaim_worker); INIT_DELAYED_WORK(&mp->m_eofblocks_work, xfs_eofblocks_worker); + INIT_DELAYED_WORK(&mp->m_cowblocks_work, xfs_cowblocks_worker); mp->m_kobj.kobject.kset = xfs_kset; mp->m_super = sb; @@ -1572,6 +1623,9 @@ xfs_fs_fill_super( "DAX unsupported by block device. Turning off DAX."); mp->m_flags &= ~XFS_MOUNT_DAX; } + if (xfs_sb_version_hasreflink(&mp->m_sb)) + xfs_alert(mp, + "DAX and reflink have not been tested together!"); } if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { @@ -1585,6 +1639,10 @@ xfs_fs_fill_super( "EXPERIMENTAL reverse mapping btree feature enabled. Use at your own risk!"); } + if (xfs_sb_version_hasreflink(&mp->m_sb)) + xfs_alert(mp, + "EXPERIMENTAL reflink feature enabled. Use at your own risk!"); + error = xfs_mountfs(mp); if (error) goto out_filestream_unmount; @@ -1782,15 +1840,44 @@ xfs_init_zones(void) if (!xfs_rud_zone) goto out_destroy_icreate_zone; - xfs_rui_zone = kmem_zone_init((sizeof(struct xfs_rui_log_item) + - ((XFS_RUI_MAX_FAST_EXTENTS - 1) * - sizeof(struct xfs_map_extent))), + xfs_rui_zone = kmem_zone_init( + xfs_rui_log_item_sizeof(XFS_RUI_MAX_FAST_EXTENTS), "xfs_rui_item"); if (!xfs_rui_zone) goto out_destroy_rud_zone; + xfs_cud_zone = kmem_zone_init(sizeof(struct xfs_cud_log_item), + "xfs_cud_item"); + if (!xfs_cud_zone) + goto out_destroy_rui_zone; + + xfs_cui_zone = kmem_zone_init( + xfs_cui_log_item_sizeof(XFS_CUI_MAX_FAST_EXTENTS), + "xfs_cui_item"); + if (!xfs_cui_zone) + goto out_destroy_cud_zone; + + xfs_bud_zone = kmem_zone_init(sizeof(struct xfs_bud_log_item), + "xfs_bud_item"); + if (!xfs_bud_zone) + goto out_destroy_cui_zone; + + xfs_bui_zone = kmem_zone_init( + xfs_bui_log_item_sizeof(XFS_BUI_MAX_FAST_EXTENTS), + "xfs_bui_item"); + if (!xfs_bui_zone) + goto out_destroy_bud_zone; + return 0; + out_destroy_bud_zone: + kmem_zone_destroy(xfs_bud_zone); + out_destroy_cui_zone: + kmem_zone_destroy(xfs_cui_zone); + out_destroy_cud_zone: + kmem_zone_destroy(xfs_cud_zone); + out_destroy_rui_zone: + kmem_zone_destroy(xfs_rui_zone); out_destroy_rud_zone: kmem_zone_destroy(xfs_rud_zone); out_destroy_icreate_zone: @@ -1833,6 +1920,10 @@ xfs_destroy_zones(void) * destroy caches. */ rcu_barrier(); + kmem_zone_destroy(xfs_bui_zone); + kmem_zone_destroy(xfs_bud_zone); + kmem_zone_destroy(xfs_cui_zone); + kmem_zone_destroy(xfs_cud_zone); kmem_zone_destroy(xfs_rui_zone); kmem_zone_destroy(xfs_rud_zone); kmem_zone_destroy(xfs_icreate_zone); @@ -1886,6 +1977,8 @@ init_xfs_fs(void) xfs_extent_free_init_defer_op(); xfs_rmap_update_init_defer_op(); + xfs_refcount_update_init_defer_op(); + xfs_bmap_update_init_defer_op(); xfs_dir_startup(); diff --git a/fs/xfs/xfs_super.h b/fs/xfs/xfs_super.h index 529bce9fc37e..b6418abd85ad 100644 --- a/fs/xfs/xfs_super.h +++ b/fs/xfs/xfs_super.h @@ -61,6 +61,7 @@ struct xfs_mount; struct xfs_buftarg; struct block_device; +extern void xfs_quiesce_attr(struct xfs_mount *mp); extern void xfs_flush_inodes(struct xfs_mount *mp); extern void xfs_blkdev_issue_flush(struct xfs_buftarg *); extern xfs_agnumber_t xfs_set_inode_alloc(struct xfs_mount *, diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 58142aeeeea6..f2cb45ed1d54 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -238,8 +238,7 @@ xfs_symlink( if (error) goto out_release_inode; - xfs_ilock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL | - XFS_IOLOCK_PARENT | XFS_ILOCK_PARENT); + xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT); unlock_dp_on_error = true; /* @@ -287,7 +286,7 @@ xfs_symlink( * the transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - xfs_trans_ijoin(tp, dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = false; /* @@ -412,7 +411,7 @@ out_release_inode: xfs_qm_dqrele(pdqp); if (unlock_dp_on_error) - xfs_iunlock(dp, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); + xfs_iunlock(dp, XFS_ILOCK_EXCL); return error; } diff --git a/fs/xfs/xfs_sysctl.c b/fs/xfs/xfs_sysctl.c index aed74d3f8da9..afe1f66aaa69 100644 --- a/fs/xfs/xfs_sysctl.c +++ b/fs/xfs/xfs_sysctl.c @@ -184,6 +184,15 @@ static struct ctl_table xfs_table[] = { .extra1 = &xfs_params.eofb_timer.min, .extra2 = &xfs_params.eofb_timer.max, }, + { + .procname = "speculative_cow_prealloc_lifetime", + .data = &xfs_params.cowb_timer.val, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec_minmax, + .extra1 = &xfs_params.cowb_timer.min, + .extra2 = &xfs_params.cowb_timer.max, + }, /* please keep this the last entry */ #ifdef CONFIG_PROC_FS { diff --git a/fs/xfs/xfs_sysctl.h b/fs/xfs/xfs_sysctl.h index ffef45375754..984a3499cfe3 100644 --- a/fs/xfs/xfs_sysctl.h +++ b/fs/xfs/xfs_sysctl.h @@ -48,6 +48,7 @@ typedef struct xfs_param { xfs_sysctl_val_t inherit_nodfrg;/* Inherit the "nodefrag" inode flag. */ xfs_sysctl_val_t fstrm_timer; /* Filestream dir-AG assoc'n timeout. */ xfs_sysctl_val_t eofb_timer; /* Interval between eofb scan wakeups */ + xfs_sysctl_val_t cowb_timer; /* Interval between cowb scan wakeups */ } xfs_param_t; /* diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 79cfd3fc5324..de6195e38910 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -393,9 +393,15 @@ max_retries_show( struct kobject *kobject, char *buf) { + int retries; struct xfs_error_cfg *cfg = to_error_cfg(kobject); - return snprintf(buf, PAGE_SIZE, "%d\n", cfg->max_retries); + if (cfg->max_retries == XFS_ERR_RETRY_FOREVER) + retries = -1; + else + retries = cfg->max_retries; + + return snprintf(buf, PAGE_SIZE, "%d\n", retries); } static ssize_t @@ -415,7 +421,10 @@ max_retries_store( if (val < -1) return -EINVAL; - cfg->max_retries = val; + if (val == -1) + cfg->max_retries = XFS_ERR_RETRY_FOREVER; + else + cfg->max_retries = val; return count; } XFS_SYSFS_ATTR_RW(max_retries); @@ -425,10 +434,15 @@ retry_timeout_seconds_show( struct kobject *kobject, char *buf) { + int timeout; struct xfs_error_cfg *cfg = to_error_cfg(kobject); - return snprintf(buf, PAGE_SIZE, "%ld\n", - jiffies_to_msecs(cfg->retry_timeout) / MSEC_PER_SEC); + if (cfg->retry_timeout == XFS_ERR_RETRY_FOREVER) + timeout = -1; + else + timeout = jiffies_to_msecs(cfg->retry_timeout) / MSEC_PER_SEC; + + return snprintf(buf, PAGE_SIZE, "%d\n", timeout); } static ssize_t @@ -445,11 +459,16 @@ retry_timeout_seconds_store( if (ret) return ret; - /* 1 day timeout maximum */ - if (val < 0 || val > 86400) + /* 1 day timeout maximum, -1 means infinite */ + if (val < -1 || val > 86400) return -EINVAL; - cfg->retry_timeout = msecs_to_jiffies(val * MSEC_PER_SEC); + if (val == -1) + cfg->retry_timeout = XFS_ERR_RETRY_FOREVER; + else { + cfg->retry_timeout = msecs_to_jiffies(val * MSEC_PER_SEC); + ASSERT(msecs_to_jiffies(val * MSEC_PER_SEC) < LONG_MAX); + } return count; } XFS_SYSFS_ATTR_RW(retry_timeout_seconds); @@ -493,13 +512,13 @@ static struct attribute *xfs_error_attrs[] = { }; -struct kobj_type xfs_error_cfg_ktype = { +static struct kobj_type xfs_error_cfg_ktype = { .release = xfs_sysfs_release, .sysfs_ops = &xfs_sysfs_ops, .default_attrs = xfs_error_attrs, }; -struct kobj_type xfs_error_ktype = { +static struct kobj_type xfs_error_ktype = { .release = xfs_sysfs_release, .sysfs_ops = &xfs_sysfs_ops, }; @@ -519,18 +538,19 @@ struct xfs_error_init { static const struct xfs_error_init xfs_error_meta_init[XFS_ERR_ERRNO_MAX] = { { .name = "default", .max_retries = XFS_ERR_RETRY_FOREVER, - .retry_timeout = 0, + .retry_timeout = XFS_ERR_RETRY_FOREVER, }, { .name = "EIO", .max_retries = XFS_ERR_RETRY_FOREVER, - .retry_timeout = 0, + .retry_timeout = XFS_ERR_RETRY_FOREVER, }, { .name = "ENOSPC", .max_retries = XFS_ERR_RETRY_FOREVER, - .retry_timeout = 0, + .retry_timeout = XFS_ERR_RETRY_FOREVER, }, { .name = "ENODEV", - .max_retries = 0, + .max_retries = 0, /* We can't recover from devices disappearing */ + .retry_timeout = 0, }, }; @@ -561,7 +581,10 @@ xfs_error_sysfs_init_class( goto out_error; cfg->max_retries = init[i].max_retries; - cfg->retry_timeout = msecs_to_jiffies( + if (init[i].retry_timeout == XFS_ERR_RETRY_FOREVER) + cfg->retry_timeout = XFS_ERR_RETRY_FOREVER; + else + cfg->retry_timeout = msecs_to_jiffies( init[i].retry_timeout * MSEC_PER_SEC); } return 0; diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index d303a665dba9..69c5bcd9a51b 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -39,6 +39,7 @@ struct xfs_buf_log_format; struct xfs_inode_log_format; struct xfs_bmbt_irec; struct xfs_btree_cur; +struct xfs_refcount_irec; DECLARE_EVENT_CLASS(xfs_attr_list_class, TP_PROTO(struct xfs_attr_list_context *ctx), @@ -135,6 +136,8 @@ DEFINE_PERAG_REF_EVENT(xfs_perag_set_reclaim); DEFINE_PERAG_REF_EVENT(xfs_perag_clear_reclaim); DEFINE_PERAG_REF_EVENT(xfs_perag_set_eofblocks); DEFINE_PERAG_REF_EVENT(xfs_perag_clear_eofblocks); +DEFINE_PERAG_REF_EVENT(xfs_perag_set_cowblocks); +DEFINE_PERAG_REF_EVENT(xfs_perag_clear_cowblocks); DECLARE_EVENT_CLASS(xfs_ag_class, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno), @@ -268,10 +271,10 @@ DECLARE_EVENT_CLASS(xfs_bmap_class, __field(unsigned long, caller_ip) ), TP_fast_assign( - struct xfs_ifork *ifp = (state & BMAP_ATTRFORK) ? - ip->i_afp : &ip->i_df; + struct xfs_ifork *ifp; struct xfs_bmbt_irec r; + ifp = xfs_iext_state_to_fork(ip, state); xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &r); __entry->dev = VFS_I(ip)->i_sb->s_dev; __entry->ino = ip->i_ino; @@ -352,7 +355,6 @@ DEFINE_BUF_EVENT(xfs_buf_rele); DEFINE_BUF_EVENT(xfs_buf_iodone); DEFINE_BUF_EVENT(xfs_buf_submit); DEFINE_BUF_EVENT(xfs_buf_submit_wait); -DEFINE_BUF_EVENT(xfs_buf_bawrite); DEFINE_BUF_EVENT(xfs_buf_lock); DEFINE_BUF_EVENT(xfs_buf_lock_done); DEFINE_BUF_EVENT(xfs_buf_trylock_fail); @@ -364,19 +366,15 @@ DEFINE_BUF_EVENT(xfs_buf_delwri_queue); DEFINE_BUF_EVENT(xfs_buf_delwri_queued); DEFINE_BUF_EVENT(xfs_buf_delwri_split); DEFINE_BUF_EVENT(xfs_buf_get_uncached); -DEFINE_BUF_EVENT(xfs_bdstrat_shut); DEFINE_BUF_EVENT(xfs_buf_item_relse); DEFINE_BUF_EVENT(xfs_buf_item_iodone_async); DEFINE_BUF_EVENT(xfs_buf_error_relse); DEFINE_BUF_EVENT(xfs_buf_wait_buftarg); -DEFINE_BUF_EVENT(xfs_trans_read_buf_io); DEFINE_BUF_EVENT(xfs_trans_read_buf_shut); /* not really buffer traces, but the buf provides useful information */ DEFINE_BUF_EVENT(xfs_btree_corrupt); -DEFINE_BUF_EVENT(xfs_da_btree_corrupt); DEFINE_BUF_EVENT(xfs_reset_dqcounts); -DEFINE_BUF_EVENT(xfs_inode_item_push); /* pass flags explicitly */ DECLARE_EVENT_CLASS(xfs_buf_flags_class, @@ -538,7 +536,6 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release); DEFINE_BUF_ITEM_EVENT(xfs_trans_binval); -DEFINE_BUF_ITEM_EVENT(xfs_trans_buf_ordered); DECLARE_EVENT_CLASS(xfs_filestream_class, TP_PROTO(struct xfs_inode *ip, xfs_agnumber_t agno), @@ -677,7 +674,6 @@ DEFINE_INODE_EVENT(xfs_ioctl_setattr); DEFINE_INODE_EVENT(xfs_dir_fsync); DEFINE_INODE_EVENT(xfs_file_fsync); DEFINE_INODE_EVENT(xfs_destroy_inode); -DEFINE_INODE_EVENT(xfs_evict_inode); DEFINE_INODE_EVENT(xfs_update_time); DEFINE_INODE_EVENT(xfs_dquot_dqalloc); @@ -686,6 +682,9 @@ DEFINE_INODE_EVENT(xfs_dquot_dqdetach); DEFINE_INODE_EVENT(xfs_inode_set_eofblocks_tag); DEFINE_INODE_EVENT(xfs_inode_clear_eofblocks_tag); DEFINE_INODE_EVENT(xfs_inode_free_eofblocks_invalid); +DEFINE_INODE_EVENT(xfs_inode_set_cowblocks_tag); +DEFINE_INODE_EVENT(xfs_inode_clear_cowblocks_tag); +DEFINE_INODE_EVENT(xfs_inode_free_cowblocks_invalid); DEFINE_INODE_EVENT(xfs_filemap_fault); DEFINE_INODE_EVENT(xfs_filemap_pmd_fault); @@ -792,7 +791,6 @@ TRACE_EVENT(xfs_irec_merge_post, DEFINE_EVENT(xfs_iref_class, name, \ TP_PROTO(struct xfs_inode *ip, unsigned long caller_ip), \ TP_ARGS(ip, caller_ip)) -DEFINE_IREF_EVENT(xfs_ihold); DEFINE_IREF_EVENT(xfs_irele); DEFINE_IREF_EVENT(xfs_inode_pin); DEFINE_IREF_EVENT(xfs_inode_unpin); @@ -933,7 +931,6 @@ DEFINE_DQUOT_EVENT(xfs_dqget_miss); DEFINE_DQUOT_EVENT(xfs_dqget_freeing); DEFINE_DQUOT_EVENT(xfs_dqget_dup); DEFINE_DQUOT_EVENT(xfs_dqput); -DEFINE_DQUOT_EVENT(xfs_dqput_wait); DEFINE_DQUOT_EVENT(xfs_dqput_free); DEFINE_DQUOT_EVENT(xfs_dqrele); DEFINE_DQUOT_EVENT(xfs_dqflush); @@ -1170,7 +1167,6 @@ DEFINE_RW_EVENT(xfs_file_dax_read); DEFINE_RW_EVENT(xfs_file_buffered_write); DEFINE_RW_EVENT(xfs_file_direct_write); DEFINE_RW_EVENT(xfs_file_dax_write); -DEFINE_RW_EVENT(xfs_file_splice_read); DECLARE_EVENT_CLASS(xfs_page_class, TP_PROTO(struct inode *inode, struct page *page, unsigned long off, @@ -1570,14 +1566,15 @@ TRACE_EVENT(xfs_agf, TRACE_EVENT(xfs_free_extent, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno, - xfs_extlen_t len, bool isfl, int haveleft, int haveright), - TP_ARGS(mp, agno, agbno, len, isfl, haveleft, haveright), + xfs_extlen_t len, enum xfs_ag_resv_type resv, int haveleft, + int haveright), + TP_ARGS(mp, agno, agbno, len, resv, haveleft, haveright), TP_STRUCT__entry( __field(dev_t, dev) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, agbno) __field(xfs_extlen_t, len) - __field(int, isfl) + __field(int, resv) __field(int, haveleft) __field(int, haveright) ), @@ -1586,16 +1583,16 @@ TRACE_EVENT(xfs_free_extent, __entry->agno = agno; __entry->agbno = agbno; __entry->len = len; - __entry->isfl = isfl; + __entry->resv = resv; __entry->haveleft = haveleft; __entry->haveright = haveright; ), - TP_printk("dev %d:%d agno %u agbno %u len %u isfl %d %s", + TP_printk("dev %d:%d agno %u agbno %u len %u resv %d %s", MAJOR(__entry->dev), MINOR(__entry->dev), __entry->agno, __entry->agbno, __entry->len, - __entry->isfl, + __entry->resv, __entry->haveleft ? (__entry->haveright ? "both" : "left") : (__entry->haveright ? "right" : "none")) @@ -1622,8 +1619,8 @@ DECLARE_EVENT_CLASS(xfs_alloc_class, __field(short, otype) __field(char, wasdel) __field(char, wasfromfl) - __field(char, isfl) - __field(char, userdata) + __field(int, resv) + __field(int, datatype) __field(xfs_fsblock_t, firstblock) ), TP_fast_assign( @@ -1643,14 +1640,14 @@ DECLARE_EVENT_CLASS(xfs_alloc_class, __entry->otype = args->otype; __entry->wasdel = args->wasdel; __entry->wasfromfl = args->wasfromfl; - __entry->isfl = args->isfl; - __entry->userdata = args->userdata; + __entry->resv = args->resv; + __entry->datatype = args->datatype; __entry->firstblock = args->firstblock; ), TP_printk("dev %d:%d agno %u agbno %u minlen %u maxlen %u mod %u " "prod %u minleft %u total %u alignment %u minalignslop %u " - "len %u type %s otype %s wasdel %d wasfromfl %d isfl %d " - "userdata %d firstblock 0x%llx", + "len %u type %s otype %s wasdel %d wasfromfl %d resv %d " + "datatype 0x%x firstblock 0x%llx", MAJOR(__entry->dev), MINOR(__entry->dev), __entry->agno, __entry->agbno, @@ -1667,8 +1664,8 @@ DECLARE_EVENT_CLASS(xfs_alloc_class, __print_symbolic(__entry->otype, XFS_ALLOC_TYPES), __entry->wasdel, __entry->wasfromfl, - __entry->isfl, - __entry->userdata, + __entry->resv, + __entry->datatype, (unsigned long long)__entry->firstblock) ) @@ -1809,7 +1806,6 @@ DEFINE_ATTR_EVENT(xfs_attr_sf_addname); DEFINE_ATTR_EVENT(xfs_attr_sf_create); DEFINE_ATTR_EVENT(xfs_attr_sf_lookup); DEFINE_ATTR_EVENT(xfs_attr_sf_remove); -DEFINE_ATTR_EVENT(xfs_attr_sf_removename); DEFINE_ATTR_EVENT(xfs_attr_sf_to_leaf); DEFINE_ATTR_EVENT(xfs_attr_leaf_add); @@ -1838,7 +1834,6 @@ DEFINE_ATTR_EVENT(xfs_attr_leaf_toosmall); DEFINE_ATTR_EVENT(xfs_attr_node_addname); DEFINE_ATTR_EVENT(xfs_attr_node_get); -DEFINE_ATTR_EVENT(xfs_attr_node_lookup); DEFINE_ATTR_EVENT(xfs_attr_node_replace); DEFINE_ATTR_EVENT(xfs_attr_node_removename); @@ -1984,6 +1979,29 @@ DEFINE_EVENT(xfs_swap_extent_class, name, \ DEFINE_SWAPEXT_EVENT(xfs_swap_extent_before); DEFINE_SWAPEXT_EVENT(xfs_swap_extent_after); +TRACE_EVENT(xfs_log_recover_record, + TP_PROTO(struct xlog *log, struct xlog_rec_header *rhead, int pass), + TP_ARGS(log, rhead, pass), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_lsn_t, lsn) + __field(int, len) + __field(int, num_logops) + __field(int, pass) + ), + TP_fast_assign( + __entry->dev = log->l_mp->m_super->s_dev; + __entry->lsn = be64_to_cpu(rhead->h_lsn); + __entry->len = be32_to_cpu(rhead->h_len); + __entry->num_logops = be32_to_cpu(rhead->h_num_logops); + __entry->pass = pass; + ), + TP_printk("dev %d:%d lsn 0x%llx len 0x%x num_logops 0x%x pass %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->lsn, __entry->len, __entry->num_logops, + __entry->pass) +) + DECLARE_EVENT_CLASS(xfs_log_recover_item_class, TP_PROTO(struct xlog *log, struct xlog_recover *trans, struct xlog_recover_item *item, int pass), @@ -1992,6 +2010,7 @@ DECLARE_EVENT_CLASS(xfs_log_recover_item_class, __field(dev_t, dev) __field(unsigned long, item) __field(xlog_tid_t, tid) + __field(xfs_lsn_t, lsn) __field(int, type) __field(int, pass) __field(int, count) @@ -2001,15 +2020,17 @@ DECLARE_EVENT_CLASS(xfs_log_recover_item_class, __entry->dev = log->l_mp->m_super->s_dev; __entry->item = (unsigned long)item; __entry->tid = trans->r_log_tid; + __entry->lsn = trans->r_lsn; __entry->type = ITEM_TYPE(item); __entry->pass = pass; __entry->count = item->ri_cnt; __entry->total = item->ri_total; ), - TP_printk("dev %d:%d trans 0x%x, pass %d, item 0x%p, item type %s " - "item region count/total %d/%d", + TP_printk("dev %d:%d tid 0x%x lsn 0x%llx, pass %d, item 0x%p, " + "item type %s item region count/total %d/%d", MAJOR(__entry->dev), MINOR(__entry->dev), __entry->tid, + __entry->lsn, __entry->pass, (void *)__entry->item, __print_symbolic(__entry->type, XFS_LI_TYPE_DESC), @@ -2068,6 +2089,7 @@ DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel_add); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_cancel_ref_inc); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_recover); +DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_skip); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_inode_buf); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_reg_buf); DEFINE_LOG_RECOVER_BUF_ITEM(xfs_log_recover_buf_dquot_buf); @@ -2407,11 +2429,9 @@ DEFINE_DEFER_EVENT(xfs_defer_finish_done); DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error); DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error); -DEFINE_DEFER_ERROR_EVENT(xfs_defer_op_finish_error); DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_work); DEFINE_DEFER_PENDING_EVENT(xfs_defer_intake_cancel); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_commit); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_cancel); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); @@ -2554,10 +2574,698 @@ DEFINE_RMAPBT_EVENT(xfs_rmap_delete); DEFINE_AG_ERROR_EVENT(xfs_rmap_insert_error); DEFINE_AG_ERROR_EVENT(xfs_rmap_delete_error); DEFINE_AG_ERROR_EVENT(xfs_rmap_update_error); + +DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_candidate); +DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_query); +DEFINE_RMAPBT_EVENT(xfs_rmap_lookup_le_range_candidate); +DEFINE_RMAPBT_EVENT(xfs_rmap_lookup_le_range); DEFINE_RMAPBT_EVENT(xfs_rmap_lookup_le_range_result); DEFINE_RMAPBT_EVENT(xfs_rmap_find_right_neighbor_result); DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_result); +/* deferred bmbt updates */ +#define DEFINE_BMAP_DEFERRED_EVENT DEFINE_RMAP_DEFERRED_EVENT +DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_defer); +DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_deferred); + +/* per-AG reservation */ +DECLARE_EVENT_CLASS(xfs_ag_resv_class, + TP_PROTO(struct xfs_perag *pag, enum xfs_ag_resv_type resv, + xfs_extlen_t len), + TP_ARGS(pag, resv, len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(int, resv) + __field(xfs_extlen_t, freeblks) + __field(xfs_extlen_t, flcount) + __field(xfs_extlen_t, reserved) + __field(xfs_extlen_t, asked) + __field(xfs_extlen_t, len) + ), + TP_fast_assign( + struct xfs_ag_resv *r = xfs_perag_resv(pag, resv); + + __entry->dev = pag->pag_mount->m_super->s_dev; + __entry->agno = pag->pag_agno; + __entry->resv = resv; + __entry->freeblks = pag->pagf_freeblks; + __entry->flcount = pag->pagf_flcount; + __entry->reserved = r ? r->ar_reserved : 0; + __entry->asked = r ? r->ar_asked : 0; + __entry->len = len; + ), + TP_printk("dev %d:%d agno %u resv %d freeblks %u flcount %u resv %u ask %u len %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->resv, + __entry->freeblks, + __entry->flcount, + __entry->reserved, + __entry->asked, + __entry->len) +) +#define DEFINE_AG_RESV_EVENT(name) \ +DEFINE_EVENT(xfs_ag_resv_class, name, \ + TP_PROTO(struct xfs_perag *pag, enum xfs_ag_resv_type type, \ + xfs_extlen_t len), \ + TP_ARGS(pag, type, len)) + +/* per-AG reservation tracepoints */ +DEFINE_AG_RESV_EVENT(xfs_ag_resv_init); +DEFINE_AG_RESV_EVENT(xfs_ag_resv_free); +DEFINE_AG_RESV_EVENT(xfs_ag_resv_alloc_extent); +DEFINE_AG_RESV_EVENT(xfs_ag_resv_free_extent); +DEFINE_AG_RESV_EVENT(xfs_ag_resv_critical); +DEFINE_AG_RESV_EVENT(xfs_ag_resv_needed); + +DEFINE_AG_ERROR_EVENT(xfs_ag_resv_free_error); +DEFINE_AG_ERROR_EVENT(xfs_ag_resv_init_error); + +/* refcount tracepoint classes */ + +/* reuse the discard trace class for agbno/aglen-based traces */ +#define DEFINE_AG_EXTENT_EVENT(name) DEFINE_DISCARD_EVENT(name) + +/* ag btree lookup tracepoint class */ +#define XFS_AG_BTREE_CMP_FORMAT_STR \ + { XFS_LOOKUP_EQ, "eq" }, \ + { XFS_LOOKUP_LE, "le" }, \ + { XFS_LOOKUP_GE, "ge" } +DECLARE_EVENT_CLASS(xfs_ag_btree_lookup_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + xfs_agblock_t agbno, xfs_lookup_t dir), + TP_ARGS(mp, agno, agbno, dir), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, agbno) + __field(xfs_lookup_t, dir) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->agbno = agbno; + __entry->dir = dir; + ), + TP_printk("dev %d:%d agno %u agbno %u cmp %s(%d)\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->agbno, + __print_symbolic(__entry->dir, XFS_AG_BTREE_CMP_FORMAT_STR), + __entry->dir) +) + +#define DEFINE_AG_BTREE_LOOKUP_EVENT(name) \ +DEFINE_EVENT(xfs_ag_btree_lookup_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + xfs_agblock_t agbno, xfs_lookup_t dir), \ + TP_ARGS(mp, agno, agbno, dir)) + +/* single-rcext tracepoint class */ +DECLARE_EVENT_CLASS(xfs_refcount_extent_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + struct xfs_refcount_irec *irec), + TP_ARGS(mp, agno, irec), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, startblock) + __field(xfs_extlen_t, blockcount) + __field(xfs_nlink_t, refcount) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->startblock = irec->rc_startblock; + __entry->blockcount = irec->rc_blockcount; + __entry->refcount = irec->rc_refcount; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->startblock, + __entry->blockcount, + __entry->refcount) +) + +#define DEFINE_REFCOUNT_EXTENT_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_extent_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + struct xfs_refcount_irec *irec), \ + TP_ARGS(mp, agno, irec)) + +/* single-rcext and an agbno tracepoint class */ +DECLARE_EVENT_CLASS(xfs_refcount_extent_at_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + struct xfs_refcount_irec *irec, xfs_agblock_t agbno), + TP_ARGS(mp, agno, irec, agbno), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, startblock) + __field(xfs_extlen_t, blockcount) + __field(xfs_nlink_t, refcount) + __field(xfs_agblock_t, agbno) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->startblock = irec->rc_startblock; + __entry->blockcount = irec->rc_blockcount; + __entry->refcount = irec->rc_refcount; + __entry->agbno = agbno; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u @ agbno %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->startblock, + __entry->blockcount, + __entry->refcount, + __entry->agbno) +) + +#define DEFINE_REFCOUNT_EXTENT_AT_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_extent_at_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + struct xfs_refcount_irec *irec, xfs_agblock_t agbno), \ + TP_ARGS(mp, agno, irec, agbno)) + +/* double-rcext tracepoint class */ +DECLARE_EVENT_CLASS(xfs_refcount_double_extent_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2), + TP_ARGS(mp, agno, i1, i2), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, i1_startblock) + __field(xfs_extlen_t, i1_blockcount) + __field(xfs_nlink_t, i1_refcount) + __field(xfs_agblock_t, i2_startblock) + __field(xfs_extlen_t, i2_blockcount) + __field(xfs_nlink_t, i2_refcount) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->i1_startblock = i1->rc_startblock; + __entry->i1_blockcount = i1->rc_blockcount; + __entry->i1_refcount = i1->rc_refcount; + __entry->i2_startblock = i2->rc_startblock; + __entry->i2_blockcount = i2->rc_blockcount; + __entry->i2_refcount = i2->rc_refcount; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- " + "agbno %u len %u refcount %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->i1_startblock, + __entry->i1_blockcount, + __entry->i1_refcount, + __entry->i2_startblock, + __entry->i2_blockcount, + __entry->i2_refcount) +) + +#define DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_double_extent_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2), \ + TP_ARGS(mp, agno, i1, i2)) + +/* double-rcext and an agbno tracepoint class */ +DECLARE_EVENT_CLASS(xfs_refcount_double_extent_at_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, + xfs_agblock_t agbno), + TP_ARGS(mp, agno, i1, i2, agbno), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, i1_startblock) + __field(xfs_extlen_t, i1_blockcount) + __field(xfs_nlink_t, i1_refcount) + __field(xfs_agblock_t, i2_startblock) + __field(xfs_extlen_t, i2_blockcount) + __field(xfs_nlink_t, i2_refcount) + __field(xfs_agblock_t, agbno) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->i1_startblock = i1->rc_startblock; + __entry->i1_blockcount = i1->rc_blockcount; + __entry->i1_refcount = i1->rc_refcount; + __entry->i2_startblock = i2->rc_startblock; + __entry->i2_blockcount = i2->rc_blockcount; + __entry->i2_refcount = i2->rc_refcount; + __entry->agbno = agbno; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- " + "agbno %u len %u refcount %u @ agbno %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->i1_startblock, + __entry->i1_blockcount, + __entry->i1_refcount, + __entry->i2_startblock, + __entry->i2_blockcount, + __entry->i2_refcount, + __entry->agbno) +) + +#define DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_double_extent_at_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, \ + xfs_agblock_t agbno), \ + TP_ARGS(mp, agno, i1, i2, agbno)) + +/* triple-rcext tracepoint class */ +DECLARE_EVENT_CLASS(xfs_refcount_triple_extent_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, + struct xfs_refcount_irec *i3), + TP_ARGS(mp, agno, i1, i2, i3), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, i1_startblock) + __field(xfs_extlen_t, i1_blockcount) + __field(xfs_nlink_t, i1_refcount) + __field(xfs_agblock_t, i2_startblock) + __field(xfs_extlen_t, i2_blockcount) + __field(xfs_nlink_t, i2_refcount) + __field(xfs_agblock_t, i3_startblock) + __field(xfs_extlen_t, i3_blockcount) + __field(xfs_nlink_t, i3_refcount) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->i1_startblock = i1->rc_startblock; + __entry->i1_blockcount = i1->rc_blockcount; + __entry->i1_refcount = i1->rc_refcount; + __entry->i2_startblock = i2->rc_startblock; + __entry->i2_blockcount = i2->rc_blockcount; + __entry->i2_refcount = i2->rc_refcount; + __entry->i3_startblock = i3->rc_startblock; + __entry->i3_blockcount = i3->rc_blockcount; + __entry->i3_refcount = i3->rc_refcount; + ), + TP_printk("dev %d:%d agno %u agbno %u len %u refcount %u -- " + "agbno %u len %u refcount %u -- " + "agbno %u len %u refcount %u\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->i1_startblock, + __entry->i1_blockcount, + __entry->i1_refcount, + __entry->i2_startblock, + __entry->i2_blockcount, + __entry->i2_refcount, + __entry->i3_startblock, + __entry->i3_blockcount, + __entry->i3_refcount) +); + +#define DEFINE_REFCOUNT_TRIPLE_EXTENT_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_triple_extent_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + struct xfs_refcount_irec *i1, struct xfs_refcount_irec *i2, \ + struct xfs_refcount_irec *i3), \ + TP_ARGS(mp, agno, i1, i2, i3)) + +/* refcount btree tracepoints */ +DEFINE_BUSY_EVENT(xfs_refcountbt_alloc_block); +DEFINE_BUSY_EVENT(xfs_refcountbt_free_block); +DEFINE_AG_BTREE_LOOKUP_EVENT(xfs_refcount_lookup); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_get); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_update); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_insert); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_delete); +DEFINE_AG_ERROR_EVENT(xfs_refcount_insert_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_delete_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_update_error); + +/* refcount adjustment tracepoints */ +DEFINE_AG_EXTENT_EVENT(xfs_refcount_increase); +DEFINE_AG_EXTENT_EVENT(xfs_refcount_decrease); +DEFINE_AG_EXTENT_EVENT(xfs_refcount_cow_increase); +DEFINE_AG_EXTENT_EVENT(xfs_refcount_cow_decrease); +DEFINE_REFCOUNT_TRIPLE_EXTENT_EVENT(xfs_refcount_merge_center_extents); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_modify_extent); +DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_recover_extent); +DEFINE_REFCOUNT_EXTENT_AT_EVENT(xfs_refcount_split_extent); +DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(xfs_refcount_merge_left_extent); +DEFINE_REFCOUNT_DOUBLE_EXTENT_EVENT(xfs_refcount_merge_right_extent); +DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(xfs_refcount_find_left_extent); +DEFINE_REFCOUNT_DOUBLE_EXTENT_AT_EVENT(xfs_refcount_find_right_extent); +DEFINE_AG_ERROR_EVENT(xfs_refcount_adjust_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_adjust_cow_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_center_extents_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_modify_extent_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_split_extent_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_left_extent_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_merge_right_extent_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_find_left_extent_error); +DEFINE_AG_ERROR_EVENT(xfs_refcount_find_right_extent_error); + +/* reflink helpers */ +DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared); +DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared_result); +DEFINE_AG_ERROR_EVENT(xfs_refcount_find_shared_error); +#define DEFINE_REFCOUNT_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT +DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_defer); +DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_deferred); + +TRACE_EVENT(xfs_refcount_finish_one_leftover, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + int type, xfs_agblock_t agbno, xfs_extlen_t len, + xfs_agblock_t new_agbno, xfs_extlen_t new_len), + TP_ARGS(mp, agno, type, agbno, len, new_agbno, new_len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(int, type) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + __field(xfs_agblock_t, new_agbno) + __field(xfs_extlen_t, new_len) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->type = type; + __entry->agbno = agbno; + __entry->len = len; + __entry->new_agbno = new_agbno; + __entry->new_len = new_len; + ), + TP_printk("dev %d:%d type %d agno %u agbno %u len %u new_agbno %u new_len %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->type, + __entry->agno, + __entry->agbno, + __entry->len, + __entry->new_agbno, + __entry->new_len) +); + +/* simple inode-based error/%ip tracepoint class */ +DECLARE_EVENT_CLASS(xfs_inode_error_class, + TP_PROTO(struct xfs_inode *ip, int error, unsigned long caller_ip), + TP_ARGS(ip, error, caller_ip), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(int, error) + __field(unsigned long, caller_ip) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->ino = ip->i_ino; + __entry->error = error; + __entry->caller_ip = caller_ip; + ), + TP_printk("dev %d:%d ino %llx error %d caller %ps", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->error, + (char *)__entry->caller_ip) +); + +#define DEFINE_INODE_ERROR_EVENT(name) \ +DEFINE_EVENT(xfs_inode_error_class, name, \ + TP_PROTO(struct xfs_inode *ip, int error, \ + unsigned long caller_ip), \ + TP_ARGS(ip, error, caller_ip)) + +/* reflink allocator */ +TRACE_EVENT(xfs_bmap_remap_alloc, + TP_PROTO(struct xfs_inode *ip, xfs_fsblock_t fsbno, + xfs_extlen_t len), + TP_ARGS(ip, fsbno, len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_fsblock_t, fsbno) + __field(xfs_extlen_t, len) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->ino = ip->i_ino; + __entry->fsbno = fsbno; + __entry->len = len; + ), + TP_printk("dev %d:%d ino 0x%llx fsbno 0x%llx len %x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->fsbno, + __entry->len) +); +DEFINE_INODE_ERROR_EVENT(xfs_bmap_remap_alloc_error); + +/* reflink tracepoint classes */ + +/* two-file io tracepoint class */ +DECLARE_EVENT_CLASS(xfs_double_io_class, + TP_PROTO(struct xfs_inode *src, xfs_off_t soffset, xfs_off_t len, + struct xfs_inode *dest, xfs_off_t doffset), + TP_ARGS(src, soffset, len, dest, doffset), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, src_ino) + __field(loff_t, src_isize) + __field(loff_t, src_disize) + __field(loff_t, src_offset) + __field(size_t, len) + __field(xfs_ino_t, dest_ino) + __field(loff_t, dest_isize) + __field(loff_t, dest_disize) + __field(loff_t, dest_offset) + ), + TP_fast_assign( + __entry->dev = VFS_I(src)->i_sb->s_dev; + __entry->src_ino = src->i_ino; + __entry->src_isize = VFS_I(src)->i_size; + __entry->src_disize = src->i_d.di_size; + __entry->src_offset = soffset; + __entry->len = len; + __entry->dest_ino = dest->i_ino; + __entry->dest_isize = VFS_I(dest)->i_size; + __entry->dest_disize = dest->i_d.di_size; + __entry->dest_offset = doffset; + ), + TP_printk("dev %d:%d count %zd " + "ino 0x%llx isize 0x%llx disize 0x%llx offset 0x%llx -> " + "ino 0x%llx isize 0x%llx disize 0x%llx offset 0x%llx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->len, + __entry->src_ino, + __entry->src_isize, + __entry->src_disize, + __entry->src_offset, + __entry->dest_ino, + __entry->dest_isize, + __entry->dest_disize, + __entry->dest_offset) +) + +#define DEFINE_DOUBLE_IO_EVENT(name) \ +DEFINE_EVENT(xfs_double_io_class, name, \ + TP_PROTO(struct xfs_inode *src, xfs_off_t soffset, xfs_off_t len, \ + struct xfs_inode *dest, xfs_off_t doffset), \ + TP_ARGS(src, soffset, len, dest, doffset)) + +/* inode/irec events */ +DECLARE_EVENT_CLASS(xfs_inode_irec_class, + TP_PROTO(struct xfs_inode *ip, struct xfs_bmbt_irec *irec), + TP_ARGS(ip, irec), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_fileoff_t, lblk) + __field(xfs_extlen_t, len) + __field(xfs_fsblock_t, pblk) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->ino = ip->i_ino; + __entry->lblk = irec->br_startoff; + __entry->len = irec->br_blockcount; + __entry->pblk = irec->br_startblock; + ), + TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx len 0x%x pblk %llu", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->lblk, + __entry->len, + __entry->pblk) +); +#define DEFINE_INODE_IREC_EVENT(name) \ +DEFINE_EVENT(xfs_inode_irec_class, name, \ + TP_PROTO(struct xfs_inode *ip, struct xfs_bmbt_irec *irec), \ + TP_ARGS(ip, irec)) + +/* refcount/reflink tracepoint definitions */ + +/* reflink tracepoints */ +DEFINE_INODE_EVENT(xfs_reflink_set_inode_flag); +DEFINE_INODE_EVENT(xfs_reflink_unset_inode_flag); +DEFINE_ITRUNC_EVENT(xfs_reflink_update_inode_size); +DEFINE_IOMAP_EVENT(xfs_reflink_remap_imap); +TRACE_EVENT(xfs_reflink_remap_blocks_loop, + TP_PROTO(struct xfs_inode *src, xfs_fileoff_t soffset, + xfs_filblks_t len, struct xfs_inode *dest, + xfs_fileoff_t doffset), + TP_ARGS(src, soffset, len, dest, doffset), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, src_ino) + __field(xfs_fileoff_t, src_lblk) + __field(xfs_filblks_t, len) + __field(xfs_ino_t, dest_ino) + __field(xfs_fileoff_t, dest_lblk) + ), + TP_fast_assign( + __entry->dev = VFS_I(src)->i_sb->s_dev; + __entry->src_ino = src->i_ino; + __entry->src_lblk = soffset; + __entry->len = len; + __entry->dest_ino = dest->i_ino; + __entry->dest_lblk = doffset; + ), + TP_printk("dev %d:%d len 0x%llx " + "ino 0x%llx offset 0x%llx blocks -> " + "ino 0x%llx offset 0x%llx blocks", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->len, + __entry->src_ino, + __entry->src_lblk, + __entry->dest_ino, + __entry->dest_lblk) +); +TRACE_EVENT(xfs_reflink_punch_range, + TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk, + xfs_extlen_t len), + TP_ARGS(ip, lblk, len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_fileoff_t, lblk) + __field(xfs_extlen_t, len) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->ino = ip->i_ino; + __entry->lblk = lblk; + __entry->len = len; + ), + TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx len 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->lblk, + __entry->len) +); +TRACE_EVENT(xfs_reflink_remap, + TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk, + xfs_extlen_t len, xfs_fsblock_t new_pblk), + TP_ARGS(ip, lblk, len, new_pblk), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_fileoff_t, lblk) + __field(xfs_extlen_t, len) + __field(xfs_fsblock_t, new_pblk) + ), + TP_fast_assign( + __entry->dev = VFS_I(ip)->i_sb->s_dev; + __entry->ino = ip->i_ino; + __entry->lblk = lblk; + __entry->len = len; + __entry->new_pblk = new_pblk; + ), + TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx len 0x%x new_pblk %llu", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->lblk, + __entry->len, + __entry->new_pblk) +); +DEFINE_DOUBLE_IO_EVENT(xfs_reflink_remap_range); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_range_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_set_inode_flag_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_update_inode_size_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_blocks_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_extent_error); + +/* dedupe tracepoints */ +DEFINE_DOUBLE_IO_EVENT(xfs_reflink_compare_extents); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_compare_extents_error); + +/* ioctl tracepoints */ +TRACE_EVENT(xfs_ioctl_clone, + TP_PROTO(struct inode *src, struct inode *dest), + TP_ARGS(src, dest), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(unsigned long, src_ino) + __field(loff_t, src_isize) + __field(unsigned long, dest_ino) + __field(loff_t, dest_isize) + ), + TP_fast_assign( + __entry->dev = src->i_sb->s_dev; + __entry->src_ino = src->i_ino; + __entry->src_isize = i_size_read(src); + __entry->dest_ino = dest->i_ino; + __entry->dest_isize = i_size_read(dest); + ), + TP_printk("dev %d:%d " + "ino 0x%lx isize 0x%llx -> " + "ino 0x%lx isize 0x%llx\n", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->src_ino, + __entry->src_isize, + __entry->dest_ino, + __entry->dest_isize) +); + +/* unshare tracepoints */ +DEFINE_SIMPLE_IO_EVENT(xfs_reflink_unshare); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_unshare_error); + +/* copy on write */ +DEFINE_INODE_IREC_EVENT(xfs_reflink_trim_around_shared); +DEFINE_INODE_IREC_EVENT(xfs_reflink_cow_alloc); +DEFINE_INODE_IREC_EVENT(xfs_reflink_cow_found); +DEFINE_INODE_IREC_EVENT(xfs_reflink_cow_enospc); + +DEFINE_RW_EVENT(xfs_reflink_reserve_cow); +DEFINE_RW_EVENT(xfs_reflink_allocate_cow_range); + +DEFINE_INODE_IREC_EVENT(xfs_reflink_bounce_dio_write); +DEFINE_IOMAP_EVENT(xfs_reflink_find_cow_mapping); +DEFINE_INODE_IREC_EVENT(xfs_reflink_trim_irec); + +DEFINE_SIMPLE_IO_EVENT(xfs_reflink_cancel_cow_range); +DEFINE_SIMPLE_IO_EVENT(xfs_reflink_end_cow); +DEFINE_INODE_IREC_EVENT(xfs_reflink_cow_remap); + +DEFINE_INODE_ERROR_EVENT(xfs_reflink_allocate_cow_range_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_cancel_cow_range_error); +DEFINE_INODE_ERROR_EVENT(xfs_reflink_end_cow_error); + + +DEFINE_INODE_IREC_EVENT(xfs_reflink_cancel_cow); + +/* rmap swapext tracepoints */ +DEFINE_INODE_IREC_EVENT(xfs_swap_extent_rmap_remap); +DEFINE_INODE_IREC_EVENT(xfs_swap_extent_rmap_remap_piece); +DEFINE_INODE_ERROR_EVENT(xfs_swap_extent_rmap_error); + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 5f3d33d16e67..70f42ea86dfb 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -217,7 +217,7 @@ undo_log: undo_blocks: if (blocks > 0) { - xfs_mod_fdblocks(tp->t_mountp, -((int64_t)blocks), rsvd); + xfs_mod_fdblocks(tp->t_mountp, (int64_t)blocks, rsvd); tp->t_blk_res = 0; } @@ -318,7 +318,6 @@ xfs_trans_mod_sb( * in-core superblock's counter. This should only * be applied to the on-disk superblock. */ - ASSERT(delta < 0); tp->t_res_fdblocks_delta += delta; if (xfs_sb_version_haslazysbcount(&mp->m_sb)) flags &= ~XFS_TRANS_SB_DIRTY; diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index e2bf86aad33d..61b7fbdd3ebd 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -36,6 +36,11 @@ struct xfs_busy_extent; struct xfs_rud_log_item; struct xfs_rui_log_item; struct xfs_btree_cur; +struct xfs_cui_log_item; +struct xfs_cud_log_item; +struct xfs_defer_ops; +struct xfs_bui_log_item; +struct xfs_bud_log_item; typedef struct xfs_log_item { struct list_head li_ail; /* AIL pointers */ @@ -248,4 +253,28 @@ int xfs_trans_log_finish_rmap_update(struct xfs_trans *tp, xfs_fsblock_t startblock, xfs_filblks_t blockcount, xfs_exntst_t state, struct xfs_btree_cur **pcur); +/* refcount updates */ +enum xfs_refcount_intent_type; + +void xfs_refcount_update_init_defer_op(void); +struct xfs_cud_log_item *xfs_trans_get_cud(struct xfs_trans *tp, + struct xfs_cui_log_item *cuip); +int xfs_trans_log_finish_refcount_update(struct xfs_trans *tp, + struct xfs_cud_log_item *cudp, struct xfs_defer_ops *dfops, + enum xfs_refcount_intent_type type, xfs_fsblock_t startblock, + xfs_extlen_t blockcount, xfs_fsblock_t *new_fsb, + xfs_extlen_t *new_len, struct xfs_btree_cur **pcur); + +/* mapping updates */ +enum xfs_bmap_intent_type; + +void xfs_bmap_update_init_defer_op(void); +struct xfs_bud_log_item *xfs_trans_get_bud(struct xfs_trans *tp, + struct xfs_bui_log_item *buip); +int xfs_trans_log_finish_bmap_update(struct xfs_trans *tp, + struct xfs_bud_log_item *rudp, struct xfs_defer_ops *dfops, + enum xfs_bmap_intent_type type, struct xfs_inode *ip, + int whichfork, xfs_fileoff_t startoff, xfs_fsblock_t startblock, + xfs_filblks_t blockcount, xfs_exntst_t state); + #endif /* __XFS_TRANS_H__ */ diff --git a/fs/xfs/xfs_trans_bmap.c b/fs/xfs/xfs_trans_bmap.c new file mode 100644 index 000000000000..6408e7d7c08c --- /dev/null +++ b/fs/xfs/xfs_trans_bmap.c @@ -0,0 +1,249 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_bmap_item.h" +#include "xfs_alloc.h" +#include "xfs_bmap.h" +#include "xfs_inode.h" + +/* + * This routine is called to allocate a "bmap update done" + * log item. + */ +struct xfs_bud_log_item * +xfs_trans_get_bud( + struct xfs_trans *tp, + struct xfs_bui_log_item *buip) +{ + struct xfs_bud_log_item *budp; + + budp = xfs_bud_init(tp->t_mountp, buip); + xfs_trans_add_item(tp, &budp->bud_item); + return budp; +} + +/* + * Finish an bmap update and log it to the BUD. Note that the + * transaction is marked dirty regardless of whether the bmap update + * succeeds or fails to support the BUI/BUD lifecycle rules. + */ +int +xfs_trans_log_finish_bmap_update( + struct xfs_trans *tp, + struct xfs_bud_log_item *budp, + struct xfs_defer_ops *dop, + enum xfs_bmap_intent_type type, + struct xfs_inode *ip, + int whichfork, + xfs_fileoff_t startoff, + xfs_fsblock_t startblock, + xfs_filblks_t blockcount, + xfs_exntst_t state) +{ + int error; + + error = xfs_bmap_finish_one(tp, dop, ip, type, whichfork, startoff, + startblock, blockcount, state); + + /* + * Mark the transaction dirty, even on error. This ensures the + * transaction is aborted, which: + * + * 1.) releases the BUI and frees the BUD + * 2.) shuts down the filesystem + */ + tp->t_flags |= XFS_TRANS_DIRTY; + budp->bud_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + return error; +} + +/* Sort bmap intents by inode. */ +static int +xfs_bmap_update_diff_items( + void *priv, + struct list_head *a, + struct list_head *b) +{ + struct xfs_bmap_intent *ba; + struct xfs_bmap_intent *bb; + + ba = container_of(a, struct xfs_bmap_intent, bi_list); + bb = container_of(b, struct xfs_bmap_intent, bi_list); + return ba->bi_owner->i_ino - bb->bi_owner->i_ino; +} + +/* Get an BUI. */ +STATIC void * +xfs_bmap_update_create_intent( + struct xfs_trans *tp, + unsigned int count) +{ + struct xfs_bui_log_item *buip; + + ASSERT(count == XFS_BUI_MAX_FAST_EXTENTS); + ASSERT(tp != NULL); + + buip = xfs_bui_init(tp->t_mountp); + ASSERT(buip != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &buip->bui_item); + return buip; +} + +/* Set the map extent flags for this mapping. */ +static void +xfs_trans_set_bmap_flags( + struct xfs_map_extent *bmap, + enum xfs_bmap_intent_type type, + int whichfork, + xfs_exntst_t state) +{ + bmap->me_flags = 0; + switch (type) { + case XFS_BMAP_MAP: + case XFS_BMAP_UNMAP: + bmap->me_flags = type; + break; + default: + ASSERT(0); + } + if (state == XFS_EXT_UNWRITTEN) + bmap->me_flags |= XFS_BMAP_EXTENT_UNWRITTEN; + if (whichfork == XFS_ATTR_FORK) + bmap->me_flags |= XFS_BMAP_EXTENT_ATTR_FORK; +} + +/* Log bmap updates in the intent item. */ +STATIC void +xfs_bmap_update_log_item( + struct xfs_trans *tp, + void *intent, + struct list_head *item) +{ + struct xfs_bui_log_item *buip = intent; + struct xfs_bmap_intent *bmap; + uint next_extent; + struct xfs_map_extent *map; + + bmap = container_of(item, struct xfs_bmap_intent, bi_list); + + tp->t_flags |= XFS_TRANS_DIRTY; + buip->bui_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + /* + * atomic_inc_return gives us the value after the increment; + * we want to use it as an array index so we need to subtract 1 from + * it. + */ + next_extent = atomic_inc_return(&buip->bui_next_extent) - 1; + ASSERT(next_extent < buip->bui_format.bui_nextents); + map = &buip->bui_format.bui_extents[next_extent]; + map->me_owner = bmap->bi_owner->i_ino; + map->me_startblock = bmap->bi_bmap.br_startblock; + map->me_startoff = bmap->bi_bmap.br_startoff; + map->me_len = bmap->bi_bmap.br_blockcount; + xfs_trans_set_bmap_flags(map, bmap->bi_type, bmap->bi_whichfork, + bmap->bi_bmap.br_state); +} + +/* Get an BUD so we can process all the deferred rmap updates. */ +STATIC void * +xfs_bmap_update_create_done( + struct xfs_trans *tp, + void *intent, + unsigned int count) +{ + return xfs_trans_get_bud(tp, intent); +} + +/* Process a deferred rmap update. */ +STATIC int +xfs_bmap_update_finish_item( + struct xfs_trans *tp, + struct xfs_defer_ops *dop, + struct list_head *item, + void *done_item, + void **state) +{ + struct xfs_bmap_intent *bmap; + int error; + + bmap = container_of(item, struct xfs_bmap_intent, bi_list); + error = xfs_trans_log_finish_bmap_update(tp, done_item, dop, + bmap->bi_type, + bmap->bi_owner, bmap->bi_whichfork, + bmap->bi_bmap.br_startoff, + bmap->bi_bmap.br_startblock, + bmap->bi_bmap.br_blockcount, + bmap->bi_bmap.br_state); + kmem_free(bmap); + return error; +} + +/* Abort all pending BUIs. */ +STATIC void +xfs_bmap_update_abort_intent( + void *intent) +{ + xfs_bui_release(intent); +} + +/* Cancel a deferred rmap update. */ +STATIC void +xfs_bmap_update_cancel_item( + struct list_head *item) +{ + struct xfs_bmap_intent *bmap; + + bmap = container_of(item, struct xfs_bmap_intent, bi_list); + kmem_free(bmap); +} + +static const struct xfs_defer_op_type xfs_bmap_update_defer_type = { + .type = XFS_DEFER_OPS_TYPE_BMAP, + .max_items = XFS_BUI_MAX_FAST_EXTENTS, + .diff_items = xfs_bmap_update_diff_items, + .create_intent = xfs_bmap_update_create_intent, + .abort_intent = xfs_bmap_update_abort_intent, + .log_item = xfs_bmap_update_log_item, + .create_done = xfs_bmap_update_create_done, + .finish_item = xfs_bmap_update_finish_item, + .cancel_item = xfs_bmap_update_cancel_item, +}; + +/* Register the deferred op type. */ +void +xfs_bmap_update_init_defer_op(void) +{ + xfs_defer_init_op_type(&xfs_bmap_update_defer_type); +} diff --git a/fs/xfs/xfs_trans_extfree.c b/fs/xfs/xfs_trans_extfree.c index 459ddec137a4..ab438647592a 100644 --- a/fs/xfs/xfs_trans_extfree.c +++ b/fs/xfs/xfs_trans_extfree.c @@ -79,7 +79,8 @@ xfs_trans_free_extent( trace_xfs_bmap_free_deferred(tp->t_mountp, agno, 0, agbno, ext_len); - error = xfs_free_extent(tp, start_block, ext_len, oinfo); + error = xfs_free_extent(tp, start_block, ext_len, oinfo, + XFS_AG_RESV_NONE); /* * Mark the transaction dirty, even on error. This ensures the diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c index 11a3af08b5c7..dab8daa676f9 100644 --- a/fs/xfs/xfs_trans_inode.c +++ b/fs/xfs/xfs_trans_inode.c @@ -73,7 +73,7 @@ xfs_trans_ichgtime( ASSERT(tp); ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); - tv = current_fs_time(inode->i_sb); + tv = current_time(inode); if (flags & XFS_ICHGTIME_MOD) inode->i_mtime = tv; diff --git a/fs/xfs/xfs_trans_refcount.c b/fs/xfs/xfs_trans_refcount.c new file mode 100644 index 000000000000..94c1877af834 --- /dev/null +++ b/fs/xfs/xfs_trans_refcount.c @@ -0,0 +1,264 @@ +/* + * Copyright (C) 2016 Oracle. All Rights Reserved. + * + * Author: Darrick J. Wong <darrick.wong@oracle.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version 2 + * of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it would be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_trans.h" +#include "xfs_trans_priv.h" +#include "xfs_refcount_item.h" +#include "xfs_alloc.h" +#include "xfs_refcount.h" + +/* + * This routine is called to allocate a "refcount update done" + * log item. + */ +struct xfs_cud_log_item * +xfs_trans_get_cud( + struct xfs_trans *tp, + struct xfs_cui_log_item *cuip) +{ + struct xfs_cud_log_item *cudp; + + cudp = xfs_cud_init(tp->t_mountp, cuip); + xfs_trans_add_item(tp, &cudp->cud_item); + return cudp; +} + +/* + * Finish an refcount update and log it to the CUD. Note that the + * transaction is marked dirty regardless of whether the refcount + * update succeeds or fails to support the CUI/CUD lifecycle rules. + */ +int +xfs_trans_log_finish_refcount_update( + struct xfs_trans *tp, + struct xfs_cud_log_item *cudp, + struct xfs_defer_ops *dop, + enum xfs_refcount_intent_type type, + xfs_fsblock_t startblock, + xfs_extlen_t blockcount, + xfs_fsblock_t *new_fsb, + xfs_extlen_t *new_len, + struct xfs_btree_cur **pcur) +{ + int error; + + error = xfs_refcount_finish_one(tp, dop, type, startblock, + blockcount, new_fsb, new_len, pcur); + + /* + * Mark the transaction dirty, even on error. This ensures the + * transaction is aborted, which: + * + * 1.) releases the CUI and frees the CUD + * 2.) shuts down the filesystem + */ + tp->t_flags |= XFS_TRANS_DIRTY; + cudp->cud_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + return error; +} + +/* Sort refcount intents by AG. */ +static int +xfs_refcount_update_diff_items( + void *priv, + struct list_head *a, + struct list_head *b) +{ + struct xfs_mount *mp = priv; + struct xfs_refcount_intent *ra; + struct xfs_refcount_intent *rb; + + ra = container_of(a, struct xfs_refcount_intent, ri_list); + rb = container_of(b, struct xfs_refcount_intent, ri_list); + return XFS_FSB_TO_AGNO(mp, ra->ri_startblock) - + XFS_FSB_TO_AGNO(mp, rb->ri_startblock); +} + +/* Get an CUI. */ +STATIC void * +xfs_refcount_update_create_intent( + struct xfs_trans *tp, + unsigned int count) +{ + struct xfs_cui_log_item *cuip; + + ASSERT(tp != NULL); + ASSERT(count > 0); + + cuip = xfs_cui_init(tp->t_mountp, count); + ASSERT(cuip != NULL); + + /* + * Get a log_item_desc to point at the new item. + */ + xfs_trans_add_item(tp, &cuip->cui_item); + return cuip; +} + +/* Set the phys extent flags for this reverse mapping. */ +static void +xfs_trans_set_refcount_flags( + struct xfs_phys_extent *refc, + enum xfs_refcount_intent_type type) +{ + refc->pe_flags = 0; + switch (type) { + case XFS_REFCOUNT_INCREASE: + case XFS_REFCOUNT_DECREASE: + case XFS_REFCOUNT_ALLOC_COW: + case XFS_REFCOUNT_FREE_COW: + refc->pe_flags |= type; + break; + default: + ASSERT(0); + } +} + +/* Log refcount updates in the intent item. */ +STATIC void +xfs_refcount_update_log_item( + struct xfs_trans *tp, + void *intent, + struct list_head *item) +{ + struct xfs_cui_log_item *cuip = intent; + struct xfs_refcount_intent *refc; + uint next_extent; + struct xfs_phys_extent *ext; + + refc = container_of(item, struct xfs_refcount_intent, ri_list); + + tp->t_flags |= XFS_TRANS_DIRTY; + cuip->cui_item.li_desc->lid_flags |= XFS_LID_DIRTY; + + /* + * atomic_inc_return gives us the value after the increment; + * we want to use it as an array index so we need to subtract 1 from + * it. + */ + next_extent = atomic_inc_return(&cuip->cui_next_extent) - 1; + ASSERT(next_extent < cuip->cui_format.cui_nextents); + ext = &cuip->cui_format.cui_extents[next_extent]; + ext->pe_startblock = refc->ri_startblock; + ext->pe_len = refc->ri_blockcount; + xfs_trans_set_refcount_flags(ext, refc->ri_type); +} + +/* Get an CUD so we can process all the deferred refcount updates. */ +STATIC void * +xfs_refcount_update_create_done( + struct xfs_trans *tp, + void *intent, + unsigned int count) +{ + return xfs_trans_get_cud(tp, intent); +} + +/* Process a deferred refcount update. */ +STATIC int +xfs_refcount_update_finish_item( + struct xfs_trans *tp, + struct xfs_defer_ops *dop, + struct list_head *item, + void *done_item, + void **state) +{ + struct xfs_refcount_intent *refc; + xfs_fsblock_t new_fsb; + xfs_extlen_t new_aglen; + int error; + + refc = container_of(item, struct xfs_refcount_intent, ri_list); + error = xfs_trans_log_finish_refcount_update(tp, done_item, dop, + refc->ri_type, + refc->ri_startblock, + refc->ri_blockcount, + &new_fsb, &new_aglen, + (struct xfs_btree_cur **)state); + /* Did we run out of reservation? Requeue what we didn't finish. */ + if (!error && new_aglen > 0) { + ASSERT(refc->ri_type == XFS_REFCOUNT_INCREASE || + refc->ri_type == XFS_REFCOUNT_DECREASE); + refc->ri_startblock = new_fsb; + refc->ri_blockcount = new_aglen; + return -EAGAIN; + } + kmem_free(refc); + return error; +} + +/* Clean up after processing deferred refcounts. */ +STATIC void +xfs_refcount_update_finish_cleanup( + struct xfs_trans *tp, + void *state, + int error) +{ + struct xfs_btree_cur *rcur = state; + + xfs_refcount_finish_one_cleanup(tp, rcur, error); +} + +/* Abort all pending CUIs. */ +STATIC void +xfs_refcount_update_abort_intent( + void *intent) +{ + xfs_cui_release(intent); +} + +/* Cancel a deferred refcount update. */ +STATIC void +xfs_refcount_update_cancel_item( + struct list_head *item) +{ + struct xfs_refcount_intent *refc; + + refc = container_of(item, struct xfs_refcount_intent, ri_list); + kmem_free(refc); +} + +static const struct xfs_defer_op_type xfs_refcount_update_defer_type = { + .type = XFS_DEFER_OPS_TYPE_REFCOUNT, + .max_items = XFS_CUI_MAX_FAST_EXTENTS, + .diff_items = xfs_refcount_update_diff_items, + .create_intent = xfs_refcount_update_create_intent, + .abort_intent = xfs_refcount_update_abort_intent, + .log_item = xfs_refcount_update_log_item, + .create_done = xfs_refcount_update_create_done, + .finish_item = xfs_refcount_update_finish_item, + .finish_cleanup = xfs_refcount_update_finish_cleanup, + .cancel_item = xfs_refcount_update_cancel_item, +}; + +/* Register the deferred op type. */ +void +xfs_refcount_update_init_defer_op(void) +{ + xfs_defer_init_op_type(&xfs_refcount_update_defer_type); +} diff --git a/fs/xfs/xfs_trans_rmap.c b/fs/xfs/xfs_trans_rmap.c index 5a50ef881568..9ead064b5e90 100644 --- a/fs/xfs/xfs_trans_rmap.c +++ b/fs/xfs/xfs_trans_rmap.c @@ -48,12 +48,21 @@ xfs_trans_set_rmap_flags( case XFS_RMAP_MAP: rmap->me_flags |= XFS_RMAP_EXTENT_MAP; break; + case XFS_RMAP_MAP_SHARED: + rmap->me_flags |= XFS_RMAP_EXTENT_MAP_SHARED; + break; case XFS_RMAP_UNMAP: rmap->me_flags |= XFS_RMAP_EXTENT_UNMAP; break; + case XFS_RMAP_UNMAP_SHARED: + rmap->me_flags |= XFS_RMAP_EXTENT_UNMAP_SHARED; + break; case XFS_RMAP_CONVERT: rmap->me_flags |= XFS_RMAP_EXTENT_CONVERT; break; + case XFS_RMAP_CONVERT_SHARED: + rmap->me_flags |= XFS_RMAP_EXTENT_CONVERT_SHARED; + break; case XFS_RMAP_ALLOC: rmap->me_flags |= XFS_RMAP_EXTENT_ALLOC; break; diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c index ea62245fee26..0594db435972 100644 --- a/fs/xfs/xfs_xattr.c +++ b/fs/xfs/xfs_xattr.c @@ -130,7 +130,7 @@ const struct xattr_handler *xfs_xattr_handlers[] = { NULL }; -static int +static void __xfs_xattr_put_listent( struct xfs_attr_list_context *context, char *prefix, @@ -147,7 +147,8 @@ __xfs_xattr_put_listent( arraytop = context->count + prefix_len + namelen + 1; if (arraytop > context->firstu) { context->count = -1; /* insufficient space */ - return 0; + context->seen_enough = 1; + return; } offset = (char *)context->alist + context->count; strncpy(offset, prefix, prefix_len); @@ -158,10 +159,10 @@ __xfs_xattr_put_listent( compute_size: context->count += prefix_len + namelen + 1; - return 0; + return; } -static int +static void xfs_xattr_put_listent( struct xfs_attr_list_context *context, int flags, @@ -179,23 +180,19 @@ xfs_xattr_put_listent( if (namelen == SGI_ACL_FILE_SIZE && strncmp(name, SGI_ACL_FILE, SGI_ACL_FILE_SIZE) == 0) { - int ret = __xfs_xattr_put_listent( + __xfs_xattr_put_listent( context, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN, XATTR_POSIX_ACL_ACCESS, strlen(XATTR_POSIX_ACL_ACCESS)); - if (ret) - return ret; } else if (namelen == SGI_ACL_DEFAULT_SIZE && strncmp(name, SGI_ACL_DEFAULT, SGI_ACL_DEFAULT_SIZE) == 0) { - int ret = __xfs_xattr_put_listent( + __xfs_xattr_put_listent( context, XATTR_SYSTEM_PREFIX, XATTR_SYSTEM_PREFIX_LEN, XATTR_POSIX_ACL_DEFAULT, strlen(XATTR_POSIX_ACL_DEFAULT)); - if (ret) - return ret; } #endif @@ -204,7 +201,7 @@ xfs_xattr_put_listent( * see them. */ if (!capable(CAP_SYS_ADMIN)) - return 0; + return; prefix = XATTR_TRUSTED_PREFIX; prefix_len = XATTR_TRUSTED_PREFIX_LEN; @@ -216,8 +213,9 @@ xfs_xattr_put_listent( prefix_len = XATTR_USER_PREFIX_LEN; } - return __xfs_xattr_put_listent(context, prefix, prefix_len, name, - namelen); + __xfs_xattr_put_listent(context, prefix, prefix_len, name, + namelen); + return; } ssize_t |