linux/fs/xfs/libxfs/xfs_rmap_btree.c

729 lines
19 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (c) 2014 Red Hat, Inc.
* All Rights Reserved.
*/
#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_trans.h"
#include "xfs_alloc.h"
#include "xfs_btree.h"
#include "xfs_btree_staging.h"
#include "xfs_rmap.h"
#include "xfs_rmap_btree.h"
#include "xfs_health.h"
#include "xfs_trace.h"
#include "xfs_error.h"
#include "xfs_extent_busy.h"
#include "xfs_ag.h"
#include "xfs_ag_resv.h"
static struct kmem_cache *xfs_rmapbt_cur_cache;
/*
* Reverse map btree.
*
* This is a per-ag tree used to track the owner(s) of a given extent. With
* reflink it is possible for there to be multiple owners, which is a departure
* from classic XFS. Owner records for data extents are inserted when the
* extent is mapped and removed when an extent is unmapped. Owner records for
* all other block types (i.e. metadata) are inserted when an extent is
* allocated and removed when an extent is freed. There can only be one owner
* of a metadata extent, usually an inode or some other metadata structure like
* an AG btree.
*
* The rmap btree is part of the free space management, so blocks for the tree
* are sourced from the agfl. Hence we need transaction reservation support for
* this tree so that the freelist is always large enough. This also impacts on
* the minimum space we need to leave free in the AG.
*
* The tree is ordered by [ag block, owner, offset]. This is a large key size,
* but it is the only way to enforce unique keys when a block can be owned by
* multiple files at any offset. There's no need to order/search by extent
* size for online updating/management of the tree. It is intended that most
* reverse lookups will be to find the owner(s) of a particular block, or to
* try to recover tree and file data from corrupt primary metadata.
*/
static struct xfs_btree_cur *
xfs_rmapbt_dup_cursor(
struct xfs_btree_cur *cur)
{
return xfs_rmapbt_init_cursor(cur->bc_mp, cur->bc_tp,
cur->bc_ag.agbp, cur->bc_ag.pag);
}
STATIC void
xfs_rmapbt_set_root(
struct xfs_btree_cur *cur,
const union xfs_btree_ptr *ptr,
int inc)
{
struct xfs_buf *agbp = cur->bc_ag.agbp;
struct xfs_agf *agf = agbp->b_addr;
ASSERT(ptr->s != 0);
agf->agf_rmap_root = ptr->s;
be32_add_cpu(&agf->agf_rmap_level, inc);
cur->bc_ag.pag->pagf_rmap_level += inc;
xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
}
STATIC int
xfs_rmapbt_alloc_block(
struct xfs_btree_cur *cur,
const union xfs_btree_ptr *start,
union xfs_btree_ptr *new,
int *stat)
{
struct xfs_buf *agbp = cur->bc_ag.agbp;
struct xfs_agf *agf = agbp->b_addr;
struct xfs_perag *pag = cur->bc_ag.pag;
int error;
xfs_agblock_t bno;
/* Allocate the new block from the freelist. If we can't, give up. */
error = xfs_alloc_get_freelist(pag, cur->bc_tp, cur->bc_ag.agbp,
&bno, 1);
if (error)
return error;
if (bno == NULLAGBLOCK) {
*stat = 0;
return 0;
}
xfs_extent_busy_reuse(cur->bc_mp, pag, bno, 1, false);
new->s = cpu_to_be32(bno);
be32_add_cpu(&agf->agf_rmap_blocks, 1);
xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
xfs_ag_resv_rmapbt_alloc(cur->bc_mp, pag->pag_agno);
xfs: account only rmapbt-used blocks against rmapbt perag res The rmapbt perag metadata reservation reserves blocks for the reverse mapping btree (rmapbt). Since the rmapbt uses blocks from the agfl and perag accounting is updated as blocks are allocated from the allocation btrees, the reservation actually accounts blocks as they are allocated to (or freed from) the agfl rather than the rmapbt itself. While this works for blocks that are eventually used for the rmapbt, not all agfl blocks are destined for the rmapbt. Blocks that are allocated to the agfl (and thus "reserved" for the rmapbt) but then used by another structure leads to a growing inconsistency over time between the runtime tracking of rmapbt usage vs. actual rmapbt usage. Since the runtime tracking thinks all agfl blocks are rmapbt blocks, it essentially believes that less future reservation is required to satisfy the rmapbt than what is actually necessary. The inconsistency is rectified across mount cycles because the perag reservation is initialized based on the actual rmapbt usage at mount time. The problem, however, is that the excessive drain of the reservation at runtime opens a window to allocate blocks for other purposes that might be required for the rmapbt on a subsequent mount. This problem can be demonstrated by a simple test that runs an allocation workload to consume agfl blocks over time and then observe the difference in the agfl reservation requirement across an unmount/mount cycle: mount ...: xfs_ag_resv_init: ... resv 3193 ask 3194 len 3194 ... ... : xfs_ag_resv_alloc_extent: ... resv 2957 ask 3194 len 1 umount...: xfs_ag_resv_free: ... resv 2956 ask 3194 len 0 mount ...: xfs_ag_resv_init: ... resv 3052 ask 3194 len 3194 As the above tracepoints show, the reservation requirement reduces from 3194 blocks to 2956 blocks as the workload runs. Without any other changes in the filesystem, the same reservation requirement jumps from 2956 to 3052 blocks over a umount/mount cycle. To address this divergence, update the RMAPBT reservation to account blocks used for the rmapbt only rather than all blocks filled into the agfl. This patch makes several high-level changes toward that end: 1.) Reintroduce an AGFL reservation type to serve as an accounting no-op for blocks allocated to (or freed from) the AGFL. 2.) Invoke RMAPBT usage accounting from the actual rmapbt block allocation path rather than the AGFL allocation path. The first change is required because agfl blocks are considered free blocks throughout their lifetime. The perag reservation subsystem is invoked unconditionally by the allocation subsystem, so we need a way to tell the perag subsystem (via the allocation subsystem) to not make any accounting changes for blocks filled into the AGFL. The second change causes the in-core RMAPBT reservation usage accounting to remain consistent with the on-disk state at all times and eliminates the risk of leaving the rmapbt reservation underfilled. Signed-off-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-03-09 22:02:32 +00:00
*stat = 1;
return 0;
}
STATIC int
xfs_rmapbt_free_block(
struct xfs_btree_cur *cur,
struct xfs_buf *bp)
{
struct xfs_buf *agbp = cur->bc_ag.agbp;
struct xfs_agf *agf = agbp->b_addr;
struct xfs_perag *pag = cur->bc_ag.pag;
xfs_agblock_t bno;
int error;
bno = xfs_daddr_to_agbno(cur->bc_mp, xfs_buf_daddr(bp));
be32_add_cpu(&agf->agf_rmap_blocks, -1);
xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
error = xfs_alloc_put_freelist(pag, cur->bc_tp, agbp, NULL, bno, 1);
if (error)
return error;
xfs_extent_busy_insert(cur->bc_tp, pag, bno, 1,
XFS_EXTENT_BUSY_SKIP_DISCARD);
xfs_ag_resv_free_extent(pag, XFS_AG_RESV_RMAPBT, NULL, 1);
return 0;
}
STATIC int
xfs_rmapbt_get_minrecs(
struct xfs_btree_cur *cur,
int level)
{
return cur->bc_mp->m_rmap_mnr[level != 0];
}
STATIC int
xfs_rmapbt_get_maxrecs(
struct xfs_btree_cur *cur,
int level)
{
return cur->bc_mp->m_rmap_mxr[level != 0];
}
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
/*
* Convert the ondisk record's offset field into the ondisk key's offset field.
* Fork and bmbt are significant parts of the rmap record key, but written
* status is merely a record attribute.
*/
static inline __be64 ondisk_rec_offset_to_key(const union xfs_btree_rec *rec)
{
return rec->rmap.rm_offset & ~cpu_to_be64(XFS_RMAP_OFF_UNWRITTEN);
}
STATIC void
xfs_rmapbt_init_key_from_rec(
union xfs_btree_key *key,
const union xfs_btree_rec *rec)
{
key->rmap.rm_startblock = rec->rmap.rm_startblock;
key->rmap.rm_owner = rec->rmap.rm_owner;
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
key->rmap.rm_offset = ondisk_rec_offset_to_key(rec);
}
/*
* The high key for a reverse mapping record can be computed by shifting
* the startblock and offset to the highest value that would still map
* to that record. In practice this means that we add blockcount-1 to
* the startblock for all records, and if the record is for a data/attr
* fork mapping, we add blockcount-1 to the offset too.
*/
STATIC void
xfs_rmapbt_init_high_key_from_rec(
union xfs_btree_key *key,
const union xfs_btree_rec *rec)
{
uint64_t off;
int adj;
adj = be32_to_cpu(rec->rmap.rm_blockcount) - 1;
key->rmap.rm_startblock = rec->rmap.rm_startblock;
be32_add_cpu(&key->rmap.rm_startblock, adj);
key->rmap.rm_owner = rec->rmap.rm_owner;
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
key->rmap.rm_offset = ondisk_rec_offset_to_key(rec);
if (XFS_RMAP_NON_INODE_OWNER(be64_to_cpu(rec->rmap.rm_owner)) ||
XFS_RMAP_IS_BMBT_BLOCK(be64_to_cpu(rec->rmap.rm_offset)))
return;
off = be64_to_cpu(key->rmap.rm_offset);
off = (XFS_RMAP_OFF(off) + adj) | (off & ~XFS_RMAP_OFF_MASK);
key->rmap.rm_offset = cpu_to_be64(off);
}
STATIC void
xfs_rmapbt_init_rec_from_cur(
struct xfs_btree_cur *cur,
union xfs_btree_rec *rec)
{
rec->rmap.rm_startblock = cpu_to_be32(cur->bc_rec.r.rm_startblock);
rec->rmap.rm_blockcount = cpu_to_be32(cur->bc_rec.r.rm_blockcount);
rec->rmap.rm_owner = cpu_to_be64(cur->bc_rec.r.rm_owner);
rec->rmap.rm_offset = cpu_to_be64(
xfs_rmap_irec_offset_pack(&cur->bc_rec.r));
}
STATIC void
xfs_rmapbt_init_ptr_from_cur(
struct xfs_btree_cur *cur,
union xfs_btree_ptr *ptr)
{
struct xfs_agf *agf = cur->bc_ag.agbp->b_addr;
ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno));
ptr->s = agf->agf_rmap_root;
}
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
/*
* Mask the appropriate parts of the ondisk key field for a key comparison.
* Fork and bmbt are significant parts of the rmap record key, but written
* status is merely a record attribute.
*/
static inline uint64_t offset_keymask(uint64_t offset)
{
return offset & ~XFS_RMAP_OFF_UNWRITTEN;
}
STATIC int64_t
xfs_rmapbt_key_diff(
struct xfs_btree_cur *cur,
const union xfs_btree_key *key)
{
struct xfs_rmap_irec *rec = &cur->bc_rec.r;
const struct xfs_rmap_key *kp = &key->rmap;
__u64 x, y;
int64_t d;
d = (int64_t)be32_to_cpu(kp->rm_startblock) - rec->rm_startblock;
if (d)
return d;
x = be64_to_cpu(kp->rm_owner);
y = rec->rm_owner;
if (x > y)
return 1;
else if (y > x)
return -1;
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
x = offset_keymask(be64_to_cpu(kp->rm_offset));
y = offset_keymask(xfs_rmap_irec_offset_pack(rec));
if (x > y)
return 1;
else if (y > x)
return -1;
return 0;
}
STATIC int64_t
xfs_rmapbt_diff_two_keys(
struct xfs_btree_cur *cur,
const union xfs_btree_key *k1,
const union xfs_btree_key *k2,
const union xfs_btree_key *mask)
{
const struct xfs_rmap_key *kp1 = &k1->rmap;
const struct xfs_rmap_key *kp2 = &k2->rmap;
int64_t d;
__u64 x, y;
/* Doesn't make sense to mask off the physical space part */
ASSERT(!mask || mask->rmap.rm_startblock);
d = (int64_t)be32_to_cpu(kp1->rm_startblock) -
be32_to_cpu(kp2->rm_startblock);
if (d)
return d;
if (!mask || mask->rmap.rm_owner) {
x = be64_to_cpu(kp1->rm_owner);
y = be64_to_cpu(kp2->rm_owner);
if (x > y)
return 1;
else if (y > x)
return -1;
}
if (!mask || mask->rmap.rm_offset) {
/* Doesn't make sense to allow offset but not owner */
ASSERT(!mask || mask->rmap.rm_owner);
x = offset_keymask(be64_to_cpu(kp1->rm_offset));
y = offset_keymask(be64_to_cpu(kp2->rm_offset));
if (x > y)
return 1;
else if (y > x)
return -1;
}
return 0;
}
static xfs_failaddr_t
xfs_rmapbt_verify(
struct xfs_buf *bp)
{
struct xfs_mount *mp = bp->b_mount;
struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp);
struct xfs_perag *pag = bp->b_pag;
xfs_failaddr_t fa;
unsigned int level;
/*
* magic number and level verification
*
* During growfs operations, we can't verify the exact level or owner as
* the perag is not fully initialised and hence not attached to the
* buffer. In this case, check against the maximum tree depth.
*
* Similarly, during log recovery we will have a perag structure
* attached, but the agf information will not yet have been initialised
* from the on disk AGF. Again, we can only check against maximum limits
* in this case.
*/
if (!xfs_verify_magic(bp, block->bb_magic))
return __this_address;
if (!xfs_has_rmapbt(mp))
return __this_address;
fa = xfs_btree_sblock_v5hdr_verify(bp);
if (fa)
return fa;
level = be16_to_cpu(block->bb_level);
if (pag && xfs_perag_initialised_agf(pag)) {
if (level >= pag->pagf_rmap_level)
return __this_address;
} else if (level >= mp->m_rmap_maxlevels)
return __this_address;
return xfs_btree_sblock_verify(bp, mp->m_rmap_mxr[level != 0]);
}
static void
xfs_rmapbt_read_verify(
struct xfs_buf *bp)
{
xfs_failaddr_t fa;
if (!xfs_btree_sblock_verify_crc(bp))
xfs_verifier_error(bp, -EFSBADCRC, __this_address);
else {
fa = xfs_rmapbt_verify(bp);
if (fa)
xfs_verifier_error(bp, -EFSCORRUPTED, fa);
}
if (bp->b_error)
trace_xfs_btree_corrupt(bp, _RET_IP_);
}
static void
xfs_rmapbt_write_verify(
struct xfs_buf *bp)
{
xfs_failaddr_t fa;
fa = xfs_rmapbt_verify(bp);
if (fa) {
trace_xfs_btree_corrupt(bp, _RET_IP_);
xfs_verifier_error(bp, -EFSCORRUPTED, fa);
return;
}
xfs_btree_sblock_calc_crc(bp);
}
const struct xfs_buf_ops xfs_rmapbt_buf_ops = {
.name = "xfs_rmapbt",
.magic = { 0, cpu_to_be32(XFS_RMAP_CRC_MAGIC) },
.verify_read = xfs_rmapbt_read_verify,
.verify_write = xfs_rmapbt_write_verify,
.verify_struct = xfs_rmapbt_verify,
};
STATIC int
xfs_rmapbt_keys_inorder(
struct xfs_btree_cur *cur,
const union xfs_btree_key *k1,
const union xfs_btree_key *k2)
{
uint32_t x;
uint32_t y;
uint64_t a;
uint64_t b;
x = be32_to_cpu(k1->rmap.rm_startblock);
y = be32_to_cpu(k2->rmap.rm_startblock);
if (x < y)
return 1;
else if (x > y)
return 0;
a = be64_to_cpu(k1->rmap.rm_owner);
b = be64_to_cpu(k2->rmap.rm_owner);
if (a < b)
return 1;
else if (a > b)
return 0;
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
a = offset_keymask(be64_to_cpu(k1->rmap.rm_offset));
b = offset_keymask(be64_to_cpu(k2->rmap.rm_offset));
if (a <= b)
return 1;
return 0;
}
STATIC int
xfs_rmapbt_recs_inorder(
struct xfs_btree_cur *cur,
const union xfs_btree_rec *r1,
const union xfs_btree_rec *r2)
{
uint32_t x;
uint32_t y;
uint64_t a;
uint64_t b;
x = be32_to_cpu(r1->rmap.rm_startblock);
y = be32_to_cpu(r2->rmap.rm_startblock);
if (x < y)
return 1;
else if (x > y)
return 0;
a = be64_to_cpu(r1->rmap.rm_owner);
b = be64_to_cpu(r2->rmap.rm_owner);
if (a < b)
return 1;
else if (a > b)
return 0;
xfs: fix rm_offset flag handling in rmap keys Keys for extent interval records in the reverse mapping btree are supposed to be computed as follows: (physical block, owner, fork, is_btree, offset) This provides users the ability to look up a reverse mapping from a file block mapping record -- start with the physical block; then if there are multiple records for the same block, move on to the owner; then the inode fork type; and so on to the file offset. Unfortunately, the code that creates rmap lookup keys from rmap records forgot to mask off the record attribute flags, leading to ondisk keys that look like this: (physical block, owner, fork, is_btree, unwritten state, offset) Fortunately, this has all worked ok for the past six years because the key comparison functions incorrectly ignore the fork/bmbt/unwritten information that's encoded in the on-disk offset. This means that lookup comparisons are only done with: (physical block, owner, offset) Queries can (theoretically) return incorrect results because of this omission. On consistent filesystems this isn't an issue because xattr and bmbt blocks cannot be shared and hence the comparisons succeed purely on the contents of the rm_startblock field. For the one case where we support sharing (written data fork blocks) all flag bits are zero, so the omission in the comparison has no ill effects. Unfortunately, this bug prevents scrub from detecting incorrect fork and bmbt flag bits in the rmap btree, so we really do need to fix the compare code. Old filesystems with the unwritten bit erroneously set in the rmap key struct will work fine on new kernels since we still ignore the unwritten bit. New filesystems on older kernels will work fine since the old kernels never paid attention to the unwritten bit. A previous version of this patch forgot to keep the (un)written state flag masked during the comparison and caused a major regression in 5.9.x since unwritten extent conversion can update an rmap record without requiring key updates. Note that blocks cannot go directly from data fork to attr fork without being deallocated and reallocated, nor can they be added to or removed from a bmbt without a free/alloc cycle, so this should not cause any regressions. Found by fuzzing keys[1].attrfork = ones on xfs/371. Fixes: 4b8ed67794fe ("xfs: add rmap btree operations") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2023-04-12 02:00:07 +00:00
a = offset_keymask(be64_to_cpu(r1->rmap.rm_offset));
b = offset_keymask(be64_to_cpu(r2->rmap.rm_offset));
if (a <= b)
return 1;
return 0;
}
STATIC enum xbtree_key_contig
xfs_rmapbt_keys_contiguous(
struct xfs_btree_cur *cur,
const union xfs_btree_key *key1,
const union xfs_btree_key *key2,
const union xfs_btree_key *mask)
{
ASSERT(!mask || mask->rmap.rm_startblock);
/*
* We only support checking contiguity of the physical space component.
* If any callers ever need more specificity than that, they'll have to
* implement it here.
*/
ASSERT(!mask || (!mask->rmap.rm_owner && !mask->rmap.rm_offset));
return xbtree_key_contig(be32_to_cpu(key1->rmap.rm_startblock),
be32_to_cpu(key2->rmap.rm_startblock));
}
const struct xfs_btree_ops xfs_rmapbt_ops = {
.name = "rmap",
.type = XFS_BTREE_TYPE_AG,
.geom_flags = XFS_BTGEO_OVERLAPPING,
.rec_len = sizeof(struct xfs_rmap_rec),
/* Overlapping btree; 2 keys per pointer. */
.key_len = 2 * sizeof(struct xfs_rmap_key),
.ptr_len = XFS_BTREE_SHORT_PTR_LEN,
.lru_refs = XFS_RMAP_BTREE_REF,
.statoff = XFS_STATS_CALC_INDEX(xs_rmap_2),
.sick_mask = XFS_SICK_AG_RMAPBT,
.dup_cursor = xfs_rmapbt_dup_cursor,
.set_root = xfs_rmapbt_set_root,
.alloc_block = xfs_rmapbt_alloc_block,
.free_block = xfs_rmapbt_free_block,
.get_minrecs = xfs_rmapbt_get_minrecs,
.get_maxrecs = xfs_rmapbt_get_maxrecs,
.init_key_from_rec = xfs_rmapbt_init_key_from_rec,
.init_high_key_from_rec = xfs_rmapbt_init_high_key_from_rec,
.init_rec_from_cur = xfs_rmapbt_init_rec_from_cur,
.init_ptr_from_cur = xfs_rmapbt_init_ptr_from_cur,
.key_diff = xfs_rmapbt_key_diff,
.buf_ops = &xfs_rmapbt_buf_ops,
.diff_two_keys = xfs_rmapbt_diff_two_keys,
.keys_inorder = xfs_rmapbt_keys_inorder,
.recs_inorder = xfs_rmapbt_recs_inorder,
.keys_contiguous = xfs_rmapbt_keys_contiguous,
};
/*
* Create a new reverse mapping btree cursor.
*
* For staging cursors tp and agbp are NULL.
*/
struct xfs_btree_cur *
xfs_rmapbt_init_cursor(
struct xfs_mount *mp,
struct xfs_trans *tp,
struct xfs_buf *agbp,
struct xfs_perag *pag)
{
struct xfs_btree_cur *cur;
cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_RMAP, &xfs_rmapbt_ops,
mp->m_rmap_maxlevels, xfs_rmapbt_cur_cache);
cur->bc_ag.pag = xfs_perag_hold(pag);
cur->bc_ag.agbp = agbp;
if (agbp) {
struct xfs_agf *agf = agbp->b_addr;
cur->bc_nlevels = be32_to_cpu(agf->agf_rmap_level);
}
return cur;
}
/*
* Install a new reverse mapping btree root. Caller is responsible for
* invalidating and freeing the old btree blocks.
*/
void
xfs_rmapbt_commit_staged_btree(
struct xfs_btree_cur *cur,
struct xfs_trans *tp,
struct xfs_buf *agbp)
{
struct xfs_agf *agf = agbp->b_addr;
struct xbtree_afakeroot *afake = cur->bc_ag.afake;
ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
agf->agf_rmap_root = cpu_to_be32(afake->af_root);
agf->agf_rmap_level = cpu_to_be32(afake->af_levels);
agf->agf_rmap_blocks = cpu_to_be32(afake->af_blocks);
xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS |
XFS_AGF_RMAP_BLOCKS);
xfs_btree_commit_afakeroot(cur, tp, agbp);
}
/* Calculate number of records in a reverse mapping btree block. */
static inline unsigned int
xfs_rmapbt_block_maxrecs(
unsigned int blocklen,
bool leaf)
{
if (leaf)
return blocklen / sizeof(struct xfs_rmap_rec);
return blocklen /
(2 * sizeof(struct xfs_rmap_key) + sizeof(xfs_rmap_ptr_t));
}
/*
* Calculate number of records in an rmap btree block.
*/
int
xfs_rmapbt_maxrecs(
int blocklen,
int leaf)
{
blocklen -= XFS_RMAP_BLOCK_LEN;
return xfs_rmapbt_block_maxrecs(blocklen, leaf);
}
/* Compute the max possible height for reverse mapping btrees. */
unsigned int
xfs_rmapbt_maxlevels_ondisk(void)
{
unsigned int minrecs[2];
unsigned int blocklen;
blocklen = XFS_MIN_CRC_BLOCKSIZE - XFS_BTREE_SBLOCK_CRC_LEN;
minrecs[0] = xfs_rmapbt_block_maxrecs(blocklen, true) / 2;
minrecs[1] = xfs_rmapbt_block_maxrecs(blocklen, false) / 2;
/*
* Compute the asymptotic maxlevels for an rmapbt on any reflink fs.
*
* On a reflink filesystem, 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. However, we're likely to run
* out of blocks in the AG long before that happens, which means that
* we must compute the max height based on what the btree will look
* like if it consumes almost all the blocks in the AG due to maximal
* sharing factor.
*/
return xfs_btree_space_to_height(minrecs, XFS_MAX_CRC_AG_BLOCKS);
}
/* Compute the maximum height of an rmap btree. */
void
xfs_rmapbt_compute_maxlevels(
struct xfs_mount *mp)
{
if (!xfs_has_rmapbt(mp)) {
mp->m_rmap_maxlevels = 0;
return;
}
if (xfs_has_reflink(mp)) {
/*
* Compute the asymptotic maxlevels for an rmap btree on a
* filesystem that supports reflink.
*
* On a reflink filesystem, 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.
* However, we're likely to run out of blocks in the AG long
* before that happens, which means that we must compute the
* max height based on what the btree will look like if it
* consumes almost all the blocks in the AG due to maximal
* sharing factor.
*/
mp->m_rmap_maxlevels = xfs_btree_space_to_height(mp->m_rmap_mnr,
mp->m_sb.sb_agblocks);
} else {
/*
* If there's no block sharing, compute the maximum rmapbt
* height assuming one rmap record per AG block.
*/
mp->m_rmap_maxlevels = xfs_btree_compute_maxlevels(
mp->m_rmap_mnr, mp->m_sb.sb_agblocks);
}
ASSERT(mp->m_rmap_maxlevels <= xfs_rmapbt_maxlevels_ondisk());
}
/* 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->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,
struct xfs_trans *tp,
struct xfs_perag *pag,
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_has_rmapbt(mp))
return 0;
error = xfs_alloc_read_agf(pag, tp, 0, &agbp);
if (error)
return error;
agf = agbp->b_addr;
agblocks = be32_to_cpu(agf->agf_length);
tree_len = be32_to_cpu(agf->agf_rmap_blocks);
xfs_trans_brelse(tp, agbp);
/*
* The log is permanently allocated, so the space it occupies will
* never be available for the kinds of things that would require btree
* expansion. We therefore can pretend the space isn't there.
*/
if (xfs_ag_contains_log(mp, pag->pag_agno))
agblocks -= mp->m_sb.sb_logblocks;
/* 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;
}
int __init
xfs_rmapbt_init_cur_cache(void)
{
xfs_rmapbt_cur_cache = kmem_cache_create("xfs_rmapbt_cur",
xfs_btree_cur_sizeof(xfs_rmapbt_maxlevels_ondisk()),
0, 0, NULL);
if (!xfs_rmapbt_cur_cache)
return -ENOMEM;
return 0;
}
void
xfs_rmapbt_destroy_cur_cache(void)
{
kmem_cache_destroy(xfs_rmapbt_cur_cache);
xfs_rmapbt_cur_cache = NULL;
}