2018-06-06 09:42:14 +07:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2005-11-02 10:58:39 +07:00
|
|
|
* Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
|
|
|
|
* All Rights Reserved.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
#include "xfs.h"
|
2005-11-02 10:38:42 +07:00
|
|
|
#include "xfs_fs.h"
|
2013-10-23 06:36:05 +07:00
|
|
|
#include "xfs_shared.h"
|
2013-10-23 06:50:10 +07:00
|
|
|
#include "xfs_format.h"
|
|
|
|
#include "xfs_log_format.h"
|
|
|
|
#include "xfs_trans_resv.h"
|
2005-11-02 10:38:42 +07:00
|
|
|
#include "xfs_bit.h"
|
2005-04-17 05:20:36 +07:00
|
|
|
#include "xfs_sb.h"
|
|
|
|
#include "xfs_mount.h"
|
2016-08-03 08:15:38 +07:00
|
|
|
#include "xfs_defer.h"
|
2005-04-17 05:20:36 +07:00
|
|
|
#include "xfs_inode.h"
|
2005-11-02 10:38:42 +07:00
|
|
|
#include "xfs_btree.h"
|
|
|
|
#include "xfs_ialloc.h"
|
2013-10-23 06:51:50 +07:00
|
|
|
#include "xfs_ialloc_btree.h"
|
2005-04-17 05:20:36 +07:00
|
|
|
#include "xfs_alloc.h"
|
|
|
|
#include "xfs_rtalloc.h"
|
2017-11-01 02:04:49 +07:00
|
|
|
#include "xfs_errortag.h"
|
2005-04-17 05:20:36 +07:00
|
|
|
#include "xfs_error.h"
|
|
|
|
#include "xfs_bmap.h"
|
2013-04-03 12:11:15 +07:00
|
|
|
#include "xfs_cksum.h"
|
2013-10-23 06:50:10 +07:00
|
|
|
#include "xfs_trans.h"
|
2013-04-03 12:11:15 +07:00
|
|
|
#include "xfs_buf_item.h"
|
2013-06-27 13:04:56 +07:00
|
|
|
#include "xfs_icreate_item.h"
|
2013-08-12 17:49:34 +07:00
|
|
|
#include "xfs_icache.h"
|
2013-11-01 11:27:19 +07:00
|
|
|
#include "xfs_trace.h"
|
2015-10-12 11:59:25 +07:00
|
|
|
#include "xfs_log.h"
|
2016-08-03 08:33:42 +07:00
|
|
|
#include "xfs_rmap.h"
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocation group level functions.
|
|
|
|
*/
|
2017-06-17 01:00:08 +07:00
|
|
|
int
|
2008-03-27 14:00:38 +07:00
|
|
|
xfs_ialloc_cluster_alignment(
|
xfs: fix premature enospc on inode allocation
After growing a filesystem, XFS can fail to allocate inodes even
though there is a large amount of space available in the filesystem
for inodes. The issue is caused by a nearly full allocation group
having enough free space in it to be considered for inode
allocation, but not enough contiguous free space to actually
allocation inodes. This situation results in successful selection
of the AG for allocation, then failure of the allocation resulting
in ENOSPC being reported to the caller.
It is caused by two possible issues. Firstly, we only consider the
lognest free extent and whether it would fit an inode chunk. If the
extent is not correctly aligned, then we can't allocate an inode
chunk in it regardless of the fact that it is large enough. This
tends to be a permanent error until space in the AG is freed.
The second issue is that we don't actually lock the AGI or AGF when
we are doing these checks, and so by the time we get to actually
allocating the inode chunk the space we thought we had in the AG may
have been allocated. This tends to be a spurious error as it
requires a race to trigger. Hence this case is ignored in this patch
as the reported problem is for permanent errors.
The first issue could be addressed by simply taking into account the
alignment when checking the longest extent. This, however, would
prevent allocation in AGs that have aligned, exact sized extents
free. However, this case should be fairly rare compared to the
number of allocations that occur near ENOSPC that would trigger this
condition.
Hence, when selecting the inode AG, take into account the inode
cluster alignment when checking the lognest free extent in the AG.
If we can't find any AGs with a contiguous free space large
enough to be aligned, drop the alignment addition and just try for
an AG that has enough contiguous free space available for an inode
chunk. This won't prevent issues from occurring, but should avoid
situations where other AGs have lots of free space but the selected
AG can't allocate due to alignment constraints.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-12-04 05:42:21 +07:00
|
|
|
struct xfs_mount *mp)
|
2008-03-27 14:00:38 +07:00
|
|
|
{
|
xfs: fix premature enospc on inode allocation
After growing a filesystem, XFS can fail to allocate inodes even
though there is a large amount of space available in the filesystem
for inodes. The issue is caused by a nearly full allocation group
having enough free space in it to be considered for inode
allocation, but not enough contiguous free space to actually
allocation inodes. This situation results in successful selection
of the AG for allocation, then failure of the allocation resulting
in ENOSPC being reported to the caller.
It is caused by two possible issues. Firstly, we only consider the
lognest free extent and whether it would fit an inode chunk. If the
extent is not correctly aligned, then we can't allocate an inode
chunk in it regardless of the fact that it is large enough. This
tends to be a permanent error until space in the AG is freed.
The second issue is that we don't actually lock the AGI or AGF when
we are doing these checks, and so by the time we get to actually
allocating the inode chunk the space we thought we had in the AG may
have been allocated. This tends to be a spurious error as it
requires a race to trigger. Hence this case is ignored in this patch
as the reported problem is for permanent errors.
The first issue could be addressed by simply taking into account the
alignment when checking the longest extent. This, however, would
prevent allocation in AGs that have aligned, exact sized extents
free. However, this case should be fairly rare compared to the
number of allocations that occur near ENOSPC that would trigger this
condition.
Hence, when selecting the inode AG, take into account the inode
cluster alignment when checking the lognest free extent in the AG.
If we can't find any AGs with a contiguous free space large
enough to be aligned, drop the alignment addition and just try for
an AG that has enough contiguous free space available for an inode
chunk. This won't prevent issues from occurring, but should avoid
situations where other AGs have lots of free space but the selected
AG can't allocate due to alignment constraints.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-12-04 05:42:21 +07:00
|
|
|
if (xfs_sb_version_hasalign(&mp->m_sb) &&
|
xfs: Use xfs_icluster_size_fsb() to calculate inode chunk alignment
On a ppc64 system, executing generic/256 test with 32k block size gives the following call trace,
XFS: Assertion failed: args->maxlen > 0, file: /root/repos/linux/fs/xfs/libxfs/xfs_alloc.c, line: 2026
kernel BUG at /root/repos/linux/fs/xfs/xfs_message.c:113!
Oops: Exception in kernel mode, sig: 5 [#1]
SMP NR_CPUS=2048
DEBUG_PAGEALLOC
NUMA
pSeries
Modules linked in:
CPU: 2 PID: 19361 Comm: mkdir Not tainted 4.10.0-rc5 #58
task: c000000102606d80 task.stack: c0000001026b8000
NIP: c0000000004ef798 LR: c0000000004ef798 CTR: c00000000082b290
REGS: c0000001026bb090 TRAP: 0700 Not tainted (4.10.0-rc5)
MSR: 8000000000029032 <SF,EE,ME,IR,DR,RI>
CR: 28004428 XER: 00000000
CFAR: c0000000004ef180 SOFTE: 1
GPR00: c0000000004ef798 c0000001026bb310 c000000001157300 ffffffffffffffea
GPR04: 000000000000000a c0000001026bb130 0000000000000000 ffffffffffffffc0
GPR08: 00000000000000d1 0000000000000021 00000000ffffffd1 c000000000dd4990
GPR12: 0000000022004444 c00000000fe00800 0000000020000000 0000000000000000
GPR16: 0000000000000000 0000000043a606fc 0000000043a76c08 0000000043a1b3d0
GPR20: 000001002a35cd60 c0000001026bbb80 0000000000000000 0000000000000001
GPR24: 0000000000000240 0000000000000004 c00000062dc55000 0000000000000000
GPR28: 0000000000000004 c00000062ecd9200 0000000000000000 c0000001026bb6c0
NIP [c0000000004ef798] .assfail+0x28/0x30
LR [c0000000004ef798] .assfail+0x28/0x30
Call Trace:
[c0000001026bb310] [c0000000004ef798] .assfail+0x28/0x30 (unreliable)
[c0000001026bb380] [c000000000455d74] .xfs_alloc_space_available+0x194/0x1b0
[c0000001026bb410] [c00000000045b914] .xfs_alloc_fix_freelist+0x144/0x480
[c0000001026bb580] [c00000000045c368] .xfs_alloc_vextent+0x698/0xa90
[c0000001026bb650] [c0000000004a6200] .xfs_ialloc_ag_alloc+0x170/0x820
[c0000001026bb7c0] [c0000000004a9098] .xfs_dialloc+0x158/0x320
[c0000001026bb8a0] [c0000000004e628c] .xfs_ialloc+0x7c/0x610
[c0000001026bb990] [c0000000004e8138] .xfs_dir_ialloc+0xa8/0x2f0
[c0000001026bbaa0] [c0000000004e8814] .xfs_create+0x494/0x790
[c0000001026bbbf0] [c0000000004e5ebc] .xfs_generic_create+0x2bc/0x410
[c0000001026bbce0] [c0000000002b4a34] .vfs_mkdir+0x154/0x230
[c0000001026bbd70] [c0000000002bc444] .SyS_mkdirat+0x94/0x120
[c0000001026bbe30] [c00000000000b760] system_call+0x38/0xfc
Instruction dump:
4e800020 60000000 7c0802a6 7c862378 3c82ffca 7ca72b78 38841c18 7c651b78
38600000 f8010010 f821ff91 4bfff94d <0fe00000> 60000000 7c0802a6 7c892378
When block size is larger than inode cluster size, the call to
XFS_B_TO_FSBT(mp, mp->m_inode_cluster_size) returns 0. Also, mkfs.xfs
would have set xfs_sb->sb_inoalignmt to 0. This causes
xfs_ialloc_cluster_alignment() to return 0. Due to this
args.minalignslop (in xfs_ialloc_ag_alloc()) gets the unsigned
equivalent of -1 assigned to it. This later causes alloc_len in
xfs_alloc_space_available() to have a value of 0. In such a scenario
when args.total is also 0, the assert statement "ASSERT(args->maxlen >
0);" fails.
This commit fixes the bug by replacing the call to XFS_B_TO_FSBT() in
xfs_ialloc_cluster_alignment() with a call to xfs_icluster_size_fsb().
Suggested-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-02-17 08:12:16 +07:00
|
|
|
mp->m_sb.sb_inoalignmt >= xfs_icluster_size_fsb(mp))
|
xfs: fix premature enospc on inode allocation
After growing a filesystem, XFS can fail to allocate inodes even
though there is a large amount of space available in the filesystem
for inodes. The issue is caused by a nearly full allocation group
having enough free space in it to be considered for inode
allocation, but not enough contiguous free space to actually
allocation inodes. This situation results in successful selection
of the AG for allocation, then failure of the allocation resulting
in ENOSPC being reported to the caller.
It is caused by two possible issues. Firstly, we only consider the
lognest free extent and whether it would fit an inode chunk. If the
extent is not correctly aligned, then we can't allocate an inode
chunk in it regardless of the fact that it is large enough. This
tends to be a permanent error until space in the AG is freed.
The second issue is that we don't actually lock the AGI or AGF when
we are doing these checks, and so by the time we get to actually
allocating the inode chunk the space we thought we had in the AG may
have been allocated. This tends to be a spurious error as it
requires a race to trigger. Hence this case is ignored in this patch
as the reported problem is for permanent errors.
The first issue could be addressed by simply taking into account the
alignment when checking the longest extent. This, however, would
prevent allocation in AGs that have aligned, exact sized extents
free. However, this case should be fairly rare compared to the
number of allocations that occur near ENOSPC that would trigger this
condition.
Hence, when selecting the inode AG, take into account the inode
cluster alignment when checking the lognest free extent in the AG.
If we can't find any AGs with a contiguous free space large
enough to be aligned, drop the alignment addition and just try for
an AG that has enough contiguous free space available for an inode
chunk. This won't prevent issues from occurring, but should avoid
situations where other AGs have lots of free space but the selected
AG can't allocate due to alignment constraints.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-12-04 05:42:21 +07:00
|
|
|
return mp->m_sb.sb_inoalignmt;
|
2008-03-27 14:00:38 +07:00
|
|
|
return 1;
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2008-10-30 12:56:09 +07:00
|
|
|
/*
|
2009-09-01 06:58:21 +07:00
|
|
|
* Lookup a record by ino in the btree given by cur.
|
2008-10-30 12:56:09 +07:00
|
|
|
*/
|
2009-09-02 06:56:55 +07:00
|
|
|
int /* error */
|
2009-09-01 06:58:21 +07:00
|
|
|
xfs_inobt_lookup(
|
2008-10-30 12:56:09 +07:00
|
|
|
struct xfs_btree_cur *cur, /* btree cursor */
|
|
|
|
xfs_agino_t ino, /* starting inode of chunk */
|
2009-09-01 06:58:21 +07:00
|
|
|
xfs_lookup_t dir, /* <=, >=, == */
|
2008-10-30 12:56:09 +07:00
|
|
|
int *stat) /* success/failure */
|
|
|
|
{
|
|
|
|
cur->bc_rec.i.ir_startino = ino;
|
2015-05-29 06:03:04 +07:00
|
|
|
cur->bc_rec.i.ir_holemask = 0;
|
|
|
|
cur->bc_rec.i.ir_count = 0;
|
2009-09-01 06:58:21 +07:00
|
|
|
cur->bc_rec.i.ir_freecount = 0;
|
|
|
|
cur->bc_rec.i.ir_free = 0;
|
|
|
|
return xfs_btree_lookup(cur, dir, stat);
|
2008-10-30 12:56:09 +07:00
|
|
|
}
|
|
|
|
|
2008-10-30 12:56:32 +07:00
|
|
|
/*
|
2009-09-01 06:57:03 +07:00
|
|
|
* Update the record referred to by cur to the value given.
|
2008-10-30 12:56:32 +07:00
|
|
|
* This either works (return 0) or gets an EFSCORRUPTED error.
|
|
|
|
*/
|
|
|
|
STATIC int /* error */
|
|
|
|
xfs_inobt_update(
|
|
|
|
struct xfs_btree_cur *cur, /* btree cursor */
|
2009-09-01 06:57:03 +07:00
|
|
|
xfs_inobt_rec_incore_t *irec) /* btree record */
|
2008-10-30 12:56:32 +07:00
|
|
|
{
|
|
|
|
union xfs_btree_rec rec;
|
|
|
|
|
2009-09-01 06:57:03 +07:00
|
|
|
rec.inobt.ir_startino = cpu_to_be32(irec->ir_startino);
|
2015-05-29 06:03:04 +07:00
|
|
|
if (xfs_sb_version_hassparseinodes(&cur->bc_mp->m_sb)) {
|
|
|
|
rec.inobt.ir_u.sp.ir_holemask = cpu_to_be16(irec->ir_holemask);
|
|
|
|
rec.inobt.ir_u.sp.ir_count = irec->ir_count;
|
|
|
|
rec.inobt.ir_u.sp.ir_freecount = irec->ir_freecount;
|
|
|
|
} else {
|
|
|
|
/* ir_holemask/ir_count not supported on-disk */
|
|
|
|
rec.inobt.ir_u.f.ir_freecount = cpu_to_be32(irec->ir_freecount);
|
|
|
|
}
|
2009-09-01 06:57:03 +07:00
|
|
|
rec.inobt.ir_free = cpu_to_be64(irec->ir_free);
|
2008-10-30 12:56:32 +07:00
|
|
|
return xfs_btree_update(cur, &rec);
|
|
|
|
}
|
|
|
|
|
2017-06-17 01:00:08 +07:00
|
|
|
/* Convert on-disk btree record to incore inobt record. */
|
|
|
|
void
|
|
|
|
xfs_inobt_btrec_to_irec(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
union xfs_btree_rec *rec,
|
|
|
|
struct xfs_inobt_rec_incore *irec)
|
2008-10-30 12:58:11 +07:00
|
|
|
{
|
2015-05-29 06:03:04 +07:00
|
|
|
irec->ir_startino = be32_to_cpu(rec->inobt.ir_startino);
|
2017-06-17 01:00:08 +07:00
|
|
|
if (xfs_sb_version_hassparseinodes(&mp->m_sb)) {
|
2015-05-29 06:03:04 +07:00
|
|
|
irec->ir_holemask = be16_to_cpu(rec->inobt.ir_u.sp.ir_holemask);
|
|
|
|
irec->ir_count = rec->inobt.ir_u.sp.ir_count;
|
|
|
|
irec->ir_freecount = rec->inobt.ir_u.sp.ir_freecount;
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* ir_holemask/ir_count not supported on-disk. Fill in hardcoded
|
|
|
|
* values for full inode chunks.
|
|
|
|
*/
|
|
|
|
irec->ir_holemask = XFS_INOBT_HOLEMASK_FULL;
|
|
|
|
irec->ir_count = XFS_INODES_PER_CHUNK;
|
|
|
|
irec->ir_freecount =
|
|
|
|
be32_to_cpu(rec->inobt.ir_u.f.ir_freecount);
|
2008-10-30 12:58:11 +07:00
|
|
|
}
|
2015-05-29 06:03:04 +07:00
|
|
|
irec->ir_free = be64_to_cpu(rec->inobt.ir_free);
|
2017-06-17 01:00:08 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get the data from the pointed-to record.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_inobt_get_rec(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
struct xfs_inobt_rec_incore *irec,
|
|
|
|
int *stat)
|
|
|
|
{
|
2018-06-06 09:42:13 +07:00
|
|
|
struct xfs_mount *mp = cur->bc_mp;
|
|
|
|
xfs_agnumber_t agno = cur->bc_private.a.agno;
|
2017-06-17 01:00:08 +07:00
|
|
|
union xfs_btree_rec *rec;
|
|
|
|
int error;
|
2018-06-06 09:42:13 +07:00
|
|
|
uint64_t realfree;
|
2017-06-17 01:00:08 +07:00
|
|
|
|
|
|
|
error = xfs_btree_get_rec(cur, &rec, stat);
|
|
|
|
if (error || *stat == 0)
|
|
|
|
return error;
|
|
|
|
|
2018-06-06 09:42:13 +07:00
|
|
|
xfs_inobt_btrec_to_irec(mp, rec, irec);
|
|
|
|
|
|
|
|
if (!xfs_verify_agino(mp, agno, irec->ir_startino))
|
|
|
|
goto out_bad_rec;
|
|
|
|
if (irec->ir_count < XFS_INODES_PER_HOLEMASK_BIT ||
|
|
|
|
irec->ir_count > XFS_INODES_PER_CHUNK)
|
|
|
|
goto out_bad_rec;
|
|
|
|
if (irec->ir_freecount > XFS_INODES_PER_CHUNK)
|
|
|
|
goto out_bad_rec;
|
|
|
|
|
|
|
|
/* if there are no holes, return the first available offset */
|
|
|
|
if (!xfs_inobt_issparse(irec->ir_holemask))
|
|
|
|
realfree = irec->ir_free;
|
|
|
|
else
|
|
|
|
realfree = irec->ir_free & xfs_inobt_irec_to_allocmask(irec);
|
|
|
|
if (hweight64(realfree) != irec->ir_freecount)
|
|
|
|
goto out_bad_rec;
|
2015-05-29 06:03:04 +07:00
|
|
|
|
|
|
|
return 0;
|
2018-06-06 09:42:13 +07:00
|
|
|
|
|
|
|
out_bad_rec:
|
|
|
|
xfs_warn(mp,
|
|
|
|
"%s Inode BTree record corruption in AG %d detected!",
|
|
|
|
cur->bc_btnum == XFS_BTNUM_INO ? "Used" : "Free", agno);
|
|
|
|
xfs_warn(mp,
|
|
|
|
"start inode 0x%x, count 0x%x, free 0x%x freemask 0x%llx, holemask 0x%x",
|
|
|
|
irec->ir_startino, irec->ir_count, irec->ir_freecount,
|
|
|
|
irec->ir_free, irec->ir_holemask);
|
|
|
|
return -EFSCORRUPTED;
|
2008-10-30 12:58:11 +07:00
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
/*
|
|
|
|
* Insert a single inobt record. Cursor must already point to desired location.
|
|
|
|
*/
|
2018-05-10 00:02:02 +07:00
|
|
|
int
|
2014-04-24 13:00:53 +07:00
|
|
|
xfs_inobt_insert_rec(
|
|
|
|
struct xfs_btree_cur *cur,
|
2017-06-17 01:00:05 +07:00
|
|
|
uint16_t holemask,
|
|
|
|
uint8_t count,
|
|
|
|
int32_t freecount,
|
2014-04-24 13:00:53 +07:00
|
|
|
xfs_inofree_t free,
|
|
|
|
int *stat)
|
|
|
|
{
|
2015-05-29 06:03:04 +07:00
|
|
|
cur->bc_rec.i.ir_holemask = holemask;
|
|
|
|
cur->bc_rec.i.ir_count = count;
|
2014-04-24 13:00:53 +07:00
|
|
|
cur->bc_rec.i.ir_freecount = freecount;
|
|
|
|
cur->bc_rec.i.ir_free = free;
|
|
|
|
return xfs_btree_insert(cur, stat);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Insert records describing a newly allocated inode chunk into the inobt.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_inobt_insert(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
xfs_agino_t newino,
|
|
|
|
xfs_agino_t newlen,
|
|
|
|
xfs_btnum_t btnum)
|
|
|
|
{
|
|
|
|
struct xfs_btree_cur *cur;
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
xfs_agino_t thisino;
|
|
|
|
int i;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, btnum);
|
|
|
|
|
|
|
|
for (thisino = newino;
|
|
|
|
thisino < newino + newlen;
|
|
|
|
thisino += XFS_INODES_PER_CHUNK) {
|
|
|
|
error = xfs_inobt_lookup(cur, thisino, XFS_LOOKUP_EQ, &i);
|
|
|
|
if (error) {
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
ASSERT(i == 0);
|
|
|
|
|
2015-05-29 06:03:04 +07:00
|
|
|
error = xfs_inobt_insert_rec(cur, XFS_INOBT_HOLEMASK_FULL,
|
|
|
|
XFS_INODES_PER_CHUNK,
|
|
|
|
XFS_INODES_PER_CHUNK,
|
2014-04-24 13:00:53 +07:00
|
|
|
XFS_INOBT_ALL_FREE, &i);
|
|
|
|
if (error) {
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
ASSERT(i == 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-09-01 06:57:09 +07:00
|
|
|
/*
|
|
|
|
* Verify that the number of free inodes in the AGI is correct.
|
|
|
|
*/
|
|
|
|
#ifdef DEBUG
|
|
|
|
STATIC int
|
|
|
|
xfs_check_agi_freecount(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
struct xfs_agi *agi)
|
|
|
|
{
|
|
|
|
if (cur->bc_nlevels == 1) {
|
|
|
|
xfs_inobt_rec_incore_t rec;
|
|
|
|
int freecount = 0;
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
2009-09-01 06:58:21 +07:00
|
|
|
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
|
2009-09-01 06:57:09 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
do {
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
if (i) {
|
|
|
|
freecount += rec.ir_freecount;
|
|
|
|
error = xfs_btree_increment(cur, 0, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
} while (i == 1);
|
|
|
|
|
|
|
|
if (!XFS_FORCED_SHUTDOWN(cur->bc_mp))
|
|
|
|
ASSERT(freecount == be32_to_cpu(agi->agi_freecount));
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
#define xfs_check_agi_freecount(cur, agi) 0
|
|
|
|
#endif
|
|
|
|
|
2009-09-01 06:56:51 +07:00
|
|
|
/*
|
2013-06-27 13:04:55 +07:00
|
|
|
* Initialise a new set of inodes. When called without a transaction context
|
|
|
|
* (e.g. from recovery) we initiate a delayed write of the inode buffers rather
|
|
|
|
* than logging them (which in a transaction context puts them into the AIL
|
|
|
|
* for writeback rather than the xfsbufd queue).
|
2009-09-01 06:56:51 +07:00
|
|
|
*/
|
2013-06-27 13:04:56 +07:00
|
|
|
int
|
2009-09-01 06:56:51 +07:00
|
|
|
xfs_ialloc_inode_init(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
2013-06-27 13:04:55 +07:00
|
|
|
struct list_head *buffer_list,
|
2015-05-29 06:05:49 +07:00
|
|
|
int icount,
|
2009-09-01 06:56:51 +07:00
|
|
|
xfs_agnumber_t agno,
|
|
|
|
xfs_agblock_t agbno,
|
|
|
|
xfs_agblock_t length,
|
|
|
|
unsigned int gen)
|
|
|
|
{
|
|
|
|
struct xfs_buf *fbuf;
|
|
|
|
struct xfs_dinode *free;
|
2018-12-12 23:46:25 +07:00
|
|
|
int nbufs;
|
2009-09-01 06:56:51 +07:00
|
|
|
int version;
|
|
|
|
int i, j;
|
|
|
|
xfs_daddr_t d;
|
2013-04-03 12:11:17 +07:00
|
|
|
xfs_ino_t ino = 0;
|
2009-09-01 06:56:51 +07:00
|
|
|
|
|
|
|
/*
|
2013-12-13 11:51:49 +07:00
|
|
|
* Loop over the new block(s), filling in the inodes. For small block
|
|
|
|
* sizes, manipulate the inodes in buffers which are multiples of the
|
|
|
|
* blocks size.
|
2009-09-01 06:56:51 +07:00
|
|
|
*/
|
2018-12-12 23:46:25 +07:00
|
|
|
nbufs = length / mp->m_blocks_per_cluster;
|
2009-09-01 06:56:51 +07:00
|
|
|
|
|
|
|
/*
|
2013-04-03 12:11:17 +07:00
|
|
|
* Figure out what version number to use in the inodes we create. If
|
|
|
|
* the superblock version has caught up to the one that supports the new
|
|
|
|
* inode format, then use the new inode version. Otherwise use the old
|
|
|
|
* version so that old kernels will continue to be able to use the file
|
|
|
|
* system.
|
|
|
|
*
|
|
|
|
* For v3 inodes, we also need to write the inode number into the inode,
|
|
|
|
* so calculate the first inode number of the chunk here as
|
2018-12-12 23:46:24 +07:00
|
|
|
* XFS_AGB_TO_AGINO() only works within a filesystem block, not
|
2013-04-03 12:11:17 +07:00
|
|
|
* across multiple filesystem blocks (such as a cluster) and so cannot
|
|
|
|
* be used in the cluster buffer loop below.
|
|
|
|
*
|
|
|
|
* Further, because we are writing the inode directly into the buffer
|
|
|
|
* and calculating a CRC on the entire inode, we have ot log the entire
|
|
|
|
* inode so that the entire range the CRC covers is present in the log.
|
|
|
|
* That means for v3 inode we log the entire buffer rather than just the
|
|
|
|
* inode cores.
|
2009-09-01 06:56:51 +07:00
|
|
|
*/
|
2013-04-03 12:11:17 +07:00
|
|
|
if (xfs_sb_version_hascrc(&mp->m_sb)) {
|
|
|
|
version = 3;
|
2018-12-12 23:46:24 +07:00
|
|
|
ino = XFS_AGINO_TO_INO(mp, agno, XFS_AGB_TO_AGINO(mp, agbno));
|
2013-06-27 13:04:56 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* log the initialisation that is about to take place as an
|
|
|
|
* logical operation. This means the transaction does not
|
|
|
|
* need to log the physical changes to the inode buffers as log
|
|
|
|
* recovery will know what initialisation is actually needed.
|
|
|
|
* Hence we only need to log the buffers as "ordered" buffers so
|
|
|
|
* they track in the AIL as if they were physically logged.
|
|
|
|
*/
|
|
|
|
if (tp)
|
2015-05-29 06:05:49 +07:00
|
|
|
xfs_icreate_log(tp, agno, agbno, icount,
|
2013-06-27 13:04:56 +07:00
|
|
|
mp->m_sb.sb_inodesize, length, gen);
|
2014-05-20 04:46:40 +07:00
|
|
|
} else
|
2009-09-01 06:56:51 +07:00
|
|
|
version = 2;
|
|
|
|
|
|
|
|
for (j = 0; j < nbufs; j++) {
|
|
|
|
/*
|
|
|
|
* Get the block.
|
|
|
|
*/
|
2018-12-12 23:46:25 +07:00
|
|
|
d = XFS_AGB_TO_DADDR(mp, agno, agbno +
|
|
|
|
(j * mp->m_blocks_per_cluster));
|
2009-09-01 06:56:51 +07:00
|
|
|
fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
|
2018-12-12 23:46:25 +07:00
|
|
|
mp->m_bsize * mp->m_blocks_per_cluster,
|
2012-11-23 10:24:23 +07:00
|
|
|
XBF_UNMAPPED);
|
2011-09-20 20:56:55 +07:00
|
|
|
if (!fbuf)
|
2014-06-25 11:58:08 +07:00
|
|
|
return -ENOMEM;
|
2013-06-27 13:04:56 +07:00
|
|
|
|
|
|
|
/* Initialize the inode buffers and log them appropriately. */
|
2012-11-14 13:54:40 +07:00
|
|
|
fbuf->b_ops = &xfs_inode_buf_ops;
|
2013-04-03 12:11:17 +07:00
|
|
|
xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
|
2018-12-12 23:46:25 +07:00
|
|
|
for (i = 0; i < mp->m_inodes_per_cluster; i++) {
|
2009-09-01 06:56:51 +07:00
|
|
|
int ioffset = i << mp->m_sb.sb_inodelog;
|
2013-04-03 12:11:17 +07:00
|
|
|
uint isize = xfs_dinode_size(version);
|
2009-09-01 06:56:51 +07:00
|
|
|
|
|
|
|
free = xfs_make_iptr(mp, fbuf, i);
|
|
|
|
free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
|
|
|
|
free->di_version = version;
|
|
|
|
free->di_gen = cpu_to_be32(gen);
|
|
|
|
free->di_next_unlinked = cpu_to_be32(NULLAGINO);
|
2013-04-03 12:11:17 +07:00
|
|
|
|
|
|
|
if (version == 3) {
|
|
|
|
free->di_ino = cpu_to_be64(ino);
|
|
|
|
ino++;
|
2015-07-29 08:53:31 +07:00
|
|
|
uuid_copy(&free->di_uuid,
|
|
|
|
&mp->m_sb.sb_meta_uuid);
|
2013-04-03 12:11:17 +07:00
|
|
|
xfs_dinode_calc_crc(mp, free);
|
2013-06-27 13:04:55 +07:00
|
|
|
} else if (tp) {
|
2013-04-03 12:11:17 +07:00
|
|
|
/* just log the inode core */
|
|
|
|
xfs_trans_log_buf(tp, fbuf, ioffset,
|
|
|
|
ioffset + isize - 1);
|
|
|
|
}
|
|
|
|
}
|
2013-06-27 13:04:55 +07:00
|
|
|
|
|
|
|
if (tp) {
|
|
|
|
/*
|
|
|
|
* Mark the buffer as an inode allocation buffer so it
|
|
|
|
* sticks in AIL at the point of this allocation
|
|
|
|
* transaction. This ensures the they are on disk before
|
|
|
|
* the tail of the log can be moved past this
|
|
|
|
* transaction (i.e. by preventing relogging from moving
|
|
|
|
* it forward in the log).
|
|
|
|
*/
|
|
|
|
xfs_trans_inode_alloc_buf(tp, fbuf);
|
|
|
|
if (version == 3) {
|
2013-06-27 13:04:56 +07:00
|
|
|
/*
|
|
|
|
* Mark the buffer as ordered so that they are
|
|
|
|
* not physically logged in the transaction but
|
|
|
|
* still tracked in the AIL as part of the
|
|
|
|
* transaction and pin the log appropriately.
|
|
|
|
*/
|
|
|
|
xfs_trans_ordered_buf(tp, fbuf);
|
2013-06-27 13:04:55 +07:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
fbuf->b_flags |= XBF_DONE;
|
|
|
|
xfs_buf_delwri_queue(fbuf, buffer_list);
|
|
|
|
xfs_buf_relse(fbuf);
|
2009-09-01 06:56:51 +07:00
|
|
|
}
|
|
|
|
}
|
2011-09-20 20:56:55 +07:00
|
|
|
return 0;
|
2009-09-01 06:56:51 +07:00
|
|
|
}
|
|
|
|
|
2015-05-29 06:18:32 +07:00
|
|
|
/*
|
|
|
|
* Align startino and allocmask for a recently allocated sparse chunk such that
|
|
|
|
* they are fit for insertion (or merge) into the on-disk inode btrees.
|
|
|
|
*
|
|
|
|
* Background:
|
|
|
|
*
|
|
|
|
* When enabled, sparse inode support increases the inode alignment from cluster
|
|
|
|
* size to inode chunk size. This means that the minimum range between two
|
|
|
|
* non-adjacent inode records in the inobt is large enough for a full inode
|
|
|
|
* record. This allows for cluster sized, cluster aligned block allocation
|
|
|
|
* without need to worry about whether the resulting inode record overlaps with
|
|
|
|
* another record in the tree. Without this basic rule, we would have to deal
|
|
|
|
* with the consequences of overlap by potentially undoing recent allocations in
|
|
|
|
* the inode allocation codepath.
|
|
|
|
*
|
|
|
|
* Because of this alignment rule (which is enforced on mount), there are two
|
|
|
|
* inobt possibilities for newly allocated sparse chunks. One is that the
|
|
|
|
* aligned inode record for the chunk covers a range of inodes not already
|
|
|
|
* covered in the inobt (i.e., it is safe to insert a new sparse record). The
|
|
|
|
* other is that a record already exists at the aligned startino that considers
|
|
|
|
* the newly allocated range as sparse. In the latter case, record content is
|
|
|
|
* merged in hope that sparse inode chunks fill to full chunks over time.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_align_sparse_ino(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
xfs_agino_t *startino,
|
|
|
|
uint16_t *allocmask)
|
|
|
|
{
|
|
|
|
xfs_agblock_t agbno;
|
|
|
|
xfs_agblock_t mod;
|
|
|
|
int offset;
|
|
|
|
|
|
|
|
agbno = XFS_AGINO_TO_AGBNO(mp, *startino);
|
|
|
|
mod = agbno % mp->m_sb.sb_inoalignmt;
|
|
|
|
if (!mod)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/* calculate the inode offset and align startino */
|
2018-12-12 23:46:24 +07:00
|
|
|
offset = XFS_AGB_TO_AGINO(mp, mod);
|
2015-05-29 06:18:32 +07:00
|
|
|
*startino -= offset;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Since startino has been aligned down, left shift allocmask such that
|
|
|
|
* it continues to represent the same physical inodes relative to the
|
|
|
|
* new startino.
|
|
|
|
*/
|
|
|
|
*allocmask <<= offset / XFS_INODES_PER_HOLEMASK_BIT;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Determine whether the source inode record can merge into the target. Both
|
|
|
|
* records must be sparse, the inode ranges must match and there must be no
|
|
|
|
* allocation overlap between the records.
|
|
|
|
*/
|
|
|
|
STATIC bool
|
|
|
|
__xfs_inobt_can_merge(
|
|
|
|
struct xfs_inobt_rec_incore *trec, /* tgt record */
|
|
|
|
struct xfs_inobt_rec_incore *srec) /* src record */
|
|
|
|
{
|
|
|
|
uint64_t talloc;
|
|
|
|
uint64_t salloc;
|
|
|
|
|
|
|
|
/* records must cover the same inode range */
|
|
|
|
if (trec->ir_startino != srec->ir_startino)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* both records must be sparse */
|
|
|
|
if (!xfs_inobt_issparse(trec->ir_holemask) ||
|
|
|
|
!xfs_inobt_issparse(srec->ir_holemask))
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* both records must track some inodes */
|
|
|
|
if (!trec->ir_count || !srec->ir_count)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* can't exceed capacity of a full record */
|
|
|
|
if (trec->ir_count + srec->ir_count > XFS_INODES_PER_CHUNK)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
/* verify there is no allocation overlap */
|
|
|
|
talloc = xfs_inobt_irec_to_allocmask(trec);
|
|
|
|
salloc = xfs_inobt_irec_to_allocmask(srec);
|
|
|
|
if (talloc & salloc)
|
|
|
|
return false;
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Merge the source inode record into the target. The caller must call
|
|
|
|
* __xfs_inobt_can_merge() to ensure the merge is valid.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
__xfs_inobt_rec_merge(
|
|
|
|
struct xfs_inobt_rec_incore *trec, /* target */
|
|
|
|
struct xfs_inobt_rec_incore *srec) /* src */
|
|
|
|
{
|
|
|
|
ASSERT(trec->ir_startino == srec->ir_startino);
|
|
|
|
|
|
|
|
/* combine the counts */
|
|
|
|
trec->ir_count += srec->ir_count;
|
|
|
|
trec->ir_freecount += srec->ir_freecount;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Merge the holemask and free mask. For both fields, 0 bits refer to
|
|
|
|
* allocated inodes. We combine the allocated ranges with bitwise AND.
|
|
|
|
*/
|
|
|
|
trec->ir_holemask &= srec->ir_holemask;
|
|
|
|
trec->ir_free &= srec->ir_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Insert a new sparse inode chunk into the associated inode btree. The inode
|
|
|
|
* record for the sparse chunk is pre-aligned to a startino that should match
|
|
|
|
* any pre-existing sparse inode record in the tree. This allows sparse chunks
|
|
|
|
* to fill over time.
|
|
|
|
*
|
|
|
|
* This function supports two modes of handling preexisting records depending on
|
|
|
|
* the merge flag. If merge is true, the provided record is merged with the
|
|
|
|
* existing record and updated in place. The merged record is returned in nrec.
|
|
|
|
* If merge is false, an existing record is replaced with the provided record.
|
|
|
|
* If no preexisting record exists, the provided record is always inserted.
|
|
|
|
*
|
|
|
|
* It is considered corruption if a merge is requested and not possible. Given
|
|
|
|
* the sparse inode alignment constraints, this should never happen.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_inobt_insert_sprec(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
int btnum,
|
|
|
|
struct xfs_inobt_rec_incore *nrec, /* in/out: new/merged rec. */
|
|
|
|
bool merge) /* merge or replace */
|
|
|
|
{
|
|
|
|
struct xfs_btree_cur *cur;
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
|
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, btnum);
|
|
|
|
|
|
|
|
/* the new record is pre-aligned so we know where to look */
|
|
|
|
error = xfs_inobt_lookup(cur, nrec->ir_startino, XFS_LOOKUP_EQ, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
/* if nothing there, insert a new record and return */
|
|
|
|
if (i == 0) {
|
|
|
|
error = xfs_inobt_insert_rec(cur, nrec->ir_holemask,
|
|
|
|
nrec->ir_count, nrec->ir_freecount,
|
|
|
|
nrec->ir_free, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error);
|
|
|
|
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A record exists at this startino. Merge or replace the record
|
|
|
|
* depending on what we've been asked to do.
|
|
|
|
*/
|
|
|
|
if (merge) {
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error);
|
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp,
|
|
|
|
rec.ir_startino == nrec->ir_startino,
|
|
|
|
error);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This should never fail. If we have coexisting records that
|
|
|
|
* cannot merge, something is seriously wrong.
|
|
|
|
*/
|
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, __xfs_inobt_can_merge(nrec, &rec),
|
|
|
|
error);
|
|
|
|
|
|
|
|
trace_xfs_irec_merge_pre(mp, agno, rec.ir_startino,
|
|
|
|
rec.ir_holemask, nrec->ir_startino,
|
|
|
|
nrec->ir_holemask);
|
|
|
|
|
|
|
|
/* merge to nrec to output the updated record */
|
|
|
|
__xfs_inobt_rec_merge(nrec, &rec);
|
|
|
|
|
|
|
|
trace_xfs_irec_merge_post(mp, agno, nrec->ir_startino,
|
|
|
|
nrec->ir_holemask);
|
|
|
|
|
|
|
|
error = xfs_inobt_rec_check_count(mp, nrec);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
error = xfs_inobt_update(cur, nrec);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
out:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
return 0;
|
|
|
|
error:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Allocate new inodes in the allocation group specified by agbp.
|
|
|
|
* Return 0 for success, else error code.
|
|
|
|
*/
|
|
|
|
STATIC int /* error code or 0 */
|
|
|
|
xfs_ialloc_ag_alloc(
|
|
|
|
xfs_trans_t *tp, /* transaction pointer */
|
|
|
|
xfs_buf_t *agbp, /* alloc group buffer */
|
|
|
|
int *alloc)
|
|
|
|
{
|
|
|
|
xfs_agi_t *agi; /* allocation group header */
|
|
|
|
xfs_alloc_arg_t args; /* allocation argument structure */
|
[XFS] Lazy Superblock Counters
When we have a couple of hundred transactions on the fly at once, they all
typically modify the on disk superblock in some way.
create/unclink/mkdir/rmdir modify inode counts, allocation/freeing modify
free block counts.
When these counts are modified in a transaction, they must eventually lock
the superblock buffer and apply the mods. The buffer then remains locked
until the transaction is committed into the incore log buffer. The result
of this is that with enough transactions on the fly the incore superblock
buffer becomes a bottleneck.
The result of contention on the incore superblock buffer is that
transaction rates fall - the more pressure that is put on the superblock
buffer, the slower things go.
The key to removing the contention is to not require the superblock fields
in question to be locked. We do that by not marking the superblock dirty
in the transaction. IOWs, we modify the incore superblock but do not
modify the cached superblock buffer. In short, we do not log superblock
modifications to critical fields in the superblock on every transaction.
In fact we only do it just before we write the superblock to disk every
sync period or just before unmount.
This creates an interesting problem - if we don't log or write out the
fields in every transaction, then how do the values get recovered after a
crash? the answer is simple - we keep enough duplicate, logged information
in other structures that we can reconstruct the correct count after log
recovery has been performed.
It is the AGF and AGI structures that contain the duplicate information;
after recovery, we walk every AGI and AGF and sum their individual
counters to get the correct value, and we do a transaction into the log to
correct them. An optimisation of this is that if we have a clean unmount
record, we know the value in the superblock is correct, so we can avoid
the summation walk under normal conditions and so mount/recovery times do
not change under normal operation.
One wrinkle that was discovered during development was that the blocks
used in the freespace btrees are never accounted for in the AGF counters.
This was once a valid optimisation to make; when the filesystem is full,
the free space btrees are empty and consume no space. Hence when it
matters, the "accounting" is correct. But that means the when we do the
AGF summations, we would not have a correct count and xfs_check would
complain. Hence a new counter was added to track the number of blocks used
by the free space btrees. This is an *on-disk format change*.
As a result of this, lazy superblock counters are a mkfs option and at the
moment on linux there is no way to convert an old filesystem. This is
possible - xfs_db can be used to twiddle the right bits and then
xfs_repair will do the format conversion for you. Similarly, you can
convert backwards as well. At some point we'll add functionality to
xfs_admin to do the bit twiddling easily....
SGI-PV: 964999
SGI-Modid: xfs-linux-melb:xfs-kern:28652a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
2007-05-24 12:26:31 +07:00
|
|
|
xfs_agnumber_t agno;
|
2005-04-17 05:20:36 +07:00
|
|
|
int error;
|
|
|
|
xfs_agino_t newino; /* new first inode's number */
|
|
|
|
xfs_agino_t newlen; /* new number of inodes */
|
2006-03-29 06:52:28 +07:00
|
|
|
int isaligned = 0; /* inode allocation at stripe unit */
|
2005-04-17 05:20:36 +07:00
|
|
|
/* boundary */
|
2015-05-29 06:18:32 +07:00
|
|
|
uint16_t allocmask = (uint16_t) -1; /* init. to full chunk */
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
2010-01-11 18:47:43 +07:00
|
|
|
struct xfs_perag *pag;
|
2015-05-29 06:19:29 +07:00
|
|
|
int do_sparse = 0;
|
|
|
|
|
2012-09-21 01:16:45 +07:00
|
|
|
memset(&args, 0, sizeof(args));
|
2005-04-17 05:20:36 +07:00
|
|
|
args.tp = tp;
|
|
|
|
args.mp = tp->t_mountp;
|
2015-05-29 06:19:29 +07:00
|
|
|
args.fsbno = NULLFSBLOCK;
|
2018-12-12 23:46:23 +07:00
|
|
|
args.oinfo = XFS_RMAP_OINFO_INODES;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2015-06-04 10:03:34 +07:00
|
|
|
#ifdef DEBUG
|
|
|
|
/* randomly do sparse inode allocations */
|
|
|
|
if (xfs_sb_version_hassparseinodes(&tp->t_mountp->m_sb) &&
|
|
|
|
args.mp->m_ialloc_min_blks < args.mp->m_ialloc_blks)
|
|
|
|
do_sparse = prandom_u32() & 1;
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Locking will ensure that we don't have two callers in here
|
|
|
|
* at one time.
|
|
|
|
*/
|
2013-12-13 11:51:46 +07:00
|
|
|
newlen = args.mp->m_ialloc_inos;
|
2005-04-17 05:20:36 +07:00
|
|
|
if (args.mp->m_maxicount &&
|
2015-05-29 04:39:34 +07:00
|
|
|
percpu_counter_read_positive(&args.mp->m_icount) + newlen >
|
2015-02-23 17:19:28 +07:00
|
|
|
args.mp->m_maxicount)
|
2014-06-25 11:58:08 +07:00
|
|
|
return -ENOSPC;
|
2013-12-13 11:51:48 +07:00
|
|
|
args.minlen = args.maxlen = args.mp->m_ialloc_blks;
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2006-03-29 06:52:28 +07:00
|
|
|
* First try to allocate inodes contiguous with the last-allocated
|
|
|
|
* chunk of inodes. If the filesystem is striped, this will fill
|
|
|
|
* an entire stripe unit with inodes.
|
2013-06-27 13:04:55 +07:00
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
agi = XFS_BUF_TO_AGI(agbp);
|
2006-03-29 06:52:28 +07:00
|
|
|
newino = be32_to_cpu(agi->agi_newino);
|
2009-09-01 06:56:51 +07:00
|
|
|
agno = be32_to_cpu(agi->agi_seqno);
|
2006-04-11 12:45:05 +07:00
|
|
|
args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
|
2013-12-13 11:51:48 +07:00
|
|
|
args.mp->m_ialloc_blks;
|
2015-05-29 06:19:29 +07:00
|
|
|
if (do_sparse)
|
|
|
|
goto sparse_alloc;
|
2006-04-11 12:45:05 +07:00
|
|
|
if (likely(newino != NULLAGINO &&
|
|
|
|
(args.agbno < be32_to_cpu(agi->agi_length)))) {
|
2009-09-01 06:56:51 +07:00
|
|
|
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
|
2006-03-29 06:52:28 +07:00
|
|
|
args.type = XFS_ALLOCTYPE_THIS_BNO;
|
|
|
|
args.prod = 1;
|
2008-03-27 14:00:38 +07:00
|
|
|
|
2006-03-29 06:52:28 +07:00
|
|
|
/*
|
2008-03-27 14:00:38 +07:00
|
|
|
* We need to take into account alignment here to ensure that
|
|
|
|
* we don't modify the free list if we fail to have an exact
|
|
|
|
* block. If we don't have an exact match, and every oher
|
|
|
|
* attempt allocation attempt fails, we'll end up cancelling
|
|
|
|
* a dirty transaction and shutting down.
|
|
|
|
*
|
|
|
|
* For an exact allocation, alignment must be 1,
|
|
|
|
* however we need to take cluster alignment into account when
|
|
|
|
* fixing up the freelist. Use the minalignslop field to
|
|
|
|
* indicate that extra blocks might be required for alignment,
|
|
|
|
* but not to use them in the actual exact allocation.
|
2006-03-29 06:52:28 +07:00
|
|
|
*/
|
2008-03-27 14:00:38 +07:00
|
|
|
args.alignment = 1;
|
2018-12-12 23:46:25 +07:00
|
|
|
args.minalignslop = args.mp->m_cluster_align - 1;
|
2008-03-27 14:00:38 +07:00
|
|
|
|
|
|
|
/* Allow space for the inode btree to split. */
|
2009-02-09 14:37:14 +07:00
|
|
|
args.minleft = args.mp->m_in_maxlevels - 1;
|
2006-03-29 06:52:28 +07:00
|
|
|
if ((error = xfs_alloc_vextent(&args)))
|
|
|
|
return error;
|
xfs: avoid AGI/AGF deadlock scenario for inode chunk allocation
The inode chunk allocation path can lead to deadlock conditions if
a transaction is dirtied with an AGF (to fix up the freelist) for
an AG that cannot satisfy the actual allocation request. This code
path is written to try and avoid this scenario, but it can be
reproduced by running xfstests generic/270 in a loop on a 512b fs.
An example situation is:
- process A attempts an inode allocation on AG 3, modifies
the freelist, fails the allocation and ultimately moves on to
AG 0 with the AG 3 AGF held
- process B is doing a free space operation (i.e., truncate) and
acquires the AG 0 AGF, waits on the AG 3 AGF
- process A acquires the AG 0 AGI, waits on the AG 0 AGF (deadlock)
The problem here is that process A acquired the AG 3 AGF while
moving on to AG 0 (and releasing the AG 3 AGI with the AG 3 AGF
held). xfs_dialloc() makes one pass through each of the AGs when
attempting to allocate an inode chunk. The expectation is a clean
transaction if a particular AG cannot satisfy the allocation
request. xfs_ialloc_ag_alloc() is written to support this through
use of the minalignslop allocation args field.
When using the agi->agi_newino optimization, we attempt an exact
bno allocation request based on the location of the previously
allocated chunk. minalignslop is set to inform the allocator that
we will require alignment on this chunk, and thus to not allow the
request for this AG if the extra space is not available. Suppose
that the AG in question has just enough space for this request, but
not at the requested bno. xfs_alloc_fix_freelist() will proceed as
normal as it determines the request should succeed, and thus it is
allowed to modify the agf. xfs_alloc_ag_vextent() ultimately fails
because the requested bno is not available. In response, the caller
moves on to a NEAR_BNO allocation request for the same AG. The
alignment is set, but the minalignslop field is never reset. This
increases the overall requirement of the request from the first
attempt. If this delta is the difference between allocation success
and failure for the AG, xfs_alloc_fix_freelist() rejects this
request outright the second time around and causes the allocation
request to unnecessarily fail for this AG.
To address this situation, reset the minalignslop field immediately
after use and prevent it from leaking into subsequent requests.
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-03-07 12:19:14 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* This request might have dirtied the transaction if the AG can
|
|
|
|
* satisfy the request, but the exact block was not available.
|
|
|
|
* If the allocation did fail, subsequent requests will relax
|
|
|
|
* the exact agbno requirement and increase the alignment
|
|
|
|
* instead. It is critical that the total size of the request
|
|
|
|
* (len + alignment + slop) does not increase from this point
|
|
|
|
* on, so reset minalignslop to ensure it is not included in
|
|
|
|
* subsequent requests.
|
|
|
|
*/
|
|
|
|
args.minalignslop = 0;
|
2015-05-29 06:19:29 +07:00
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2006-03-29 06:52:28 +07:00
|
|
|
if (unlikely(args.fsbno == NULLFSBLOCK)) {
|
|
|
|
/*
|
|
|
|
* Set the alignment for the allocation.
|
|
|
|
* If stripe alignment is turned on then align at stripe unit
|
|
|
|
* boundary.
|
2006-04-11 12:45:05 +07:00
|
|
|
* If the cluster size is smaller than a filesystem block
|
|
|
|
* then we're doing I/O for inodes in filesystem block size
|
2006-03-29 06:52:28 +07:00
|
|
|
* pieces, so don't need alignment anyway.
|
|
|
|
*/
|
|
|
|
isaligned = 0;
|
|
|
|
if (args.mp->m_sinoalign) {
|
|
|
|
ASSERT(!(args.mp->m_flags & XFS_MOUNT_NOALIGN));
|
|
|
|
args.alignment = args.mp->m_dalign;
|
|
|
|
isaligned = 1;
|
2008-03-27 14:00:38 +07:00
|
|
|
} else
|
2018-12-12 23:46:25 +07:00
|
|
|
args.alignment = args.mp->m_cluster_align;
|
2006-03-29 06:52:28 +07:00
|
|
|
/*
|
|
|
|
* Need to figure out where to allocate the inode blocks.
|
|
|
|
* Ideally they should be spaced out through the a.g.
|
|
|
|
* For now, just allocate blocks up front.
|
|
|
|
*/
|
|
|
|
args.agbno = be32_to_cpu(agi->agi_root);
|
2009-09-01 06:56:51 +07:00
|
|
|
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
|
2006-03-29 06:52:28 +07:00
|
|
|
/*
|
|
|
|
* Allocate a fixed-size extent of inodes.
|
|
|
|
*/
|
|
|
|
args.type = XFS_ALLOCTYPE_NEAR_BNO;
|
|
|
|
args.prod = 1;
|
|
|
|
/*
|
|
|
|
* Allow space for the inode btree to split.
|
|
|
|
*/
|
2009-02-09 14:37:14 +07:00
|
|
|
args.minleft = args.mp->m_in_maxlevels - 1;
|
2006-03-29 06:52:28 +07:00
|
|
|
if ((error = xfs_alloc_vextent(&args)))
|
|
|
|
return error;
|
|
|
|
}
|
2006-04-11 12:45:05 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* If stripe alignment is turned on, then try again with cluster
|
|
|
|
* alignment.
|
|
|
|
*/
|
|
|
|
if (isaligned && args.fsbno == NULLFSBLOCK) {
|
|
|
|
args.type = XFS_ALLOCTYPE_NEAR_BNO;
|
2005-11-02 11:11:25 +07:00
|
|
|
args.agbno = be32_to_cpu(agi->agi_root);
|
2009-09-01 06:56:51 +07:00
|
|
|
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
|
2018-12-12 23:46:25 +07:00
|
|
|
args.alignment = args.mp->m_cluster_align;
|
2005-04-17 05:20:36 +07:00
|
|
|
if ((error = xfs_alloc_vextent(&args)))
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2015-05-29 06:18:32 +07:00
|
|
|
/*
|
|
|
|
* Finally, try a sparse allocation if the filesystem supports it and
|
|
|
|
* the sparse allocation length is smaller than a full chunk.
|
|
|
|
*/
|
|
|
|
if (xfs_sb_version_hassparseinodes(&args.mp->m_sb) &&
|
|
|
|
args.mp->m_ialloc_min_blks < args.mp->m_ialloc_blks &&
|
|
|
|
args.fsbno == NULLFSBLOCK) {
|
2015-05-29 06:19:29 +07:00
|
|
|
sparse_alloc:
|
2015-05-29 06:18:32 +07:00
|
|
|
args.type = XFS_ALLOCTYPE_NEAR_BNO;
|
|
|
|
args.agbno = be32_to_cpu(agi->agi_root);
|
|
|
|
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
|
|
|
|
args.alignment = args.mp->m_sb.sb_spino_align;
|
|
|
|
args.prod = 1;
|
|
|
|
|
|
|
|
args.minlen = args.mp->m_ialloc_min_blks;
|
|
|
|
args.maxlen = args.minlen;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The inode record will be aligned to full chunk size. We must
|
|
|
|
* prevent sparse allocation from AG boundaries that result in
|
|
|
|
* invalid inode records, such as records that start at agbno 0
|
|
|
|
* or extend beyond the AG.
|
|
|
|
*
|
|
|
|
* Set min agbno to the first aligned, non-zero agbno and max to
|
|
|
|
* the last aligned agbno that is at least one full chunk from
|
|
|
|
* the end of the AG.
|
|
|
|
*/
|
|
|
|
args.min_agbno = args.mp->m_sb.sb_inoalignmt;
|
|
|
|
args.max_agbno = round_down(args.mp->m_sb.sb_agblocks,
|
|
|
|
args.mp->m_sb.sb_inoalignmt) -
|
|
|
|
args.mp->m_ialloc_blks;
|
|
|
|
|
|
|
|
error = xfs_alloc_vextent(&args);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
2018-12-12 23:46:24 +07:00
|
|
|
newlen = XFS_AGB_TO_AGINO(args.mp, args.len);
|
2015-06-04 10:03:34 +07:00
|
|
|
ASSERT(newlen <= XFS_INODES_PER_CHUNK);
|
2015-05-29 06:18:32 +07:00
|
|
|
allocmask = (1 << (newlen / XFS_INODES_PER_HOLEMASK_BIT)) - 1;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
if (args.fsbno == NULLFSBLOCK) {
|
|
|
|
*alloc = 0;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
ASSERT(args.len == args.minlen);
|
|
|
|
|
2008-04-29 09:53:32 +07:00
|
|
|
/*
|
2009-09-01 06:56:51 +07:00
|
|
|
* Stamp and write the inode buffers.
|
|
|
|
*
|
2008-04-29 09:53:32 +07:00
|
|
|
* Seed the new inode cluster with a random generation number. This
|
|
|
|
* prevents short-term reuse of generation numbers if a chunk is
|
|
|
|
* freed and then immediately reallocated. We use random numbers
|
|
|
|
* rather than a linear progression to prevent the next generation
|
|
|
|
* number from being easily guessable.
|
|
|
|
*/
|
2015-05-29 06:05:49 +07:00
|
|
|
error = xfs_ialloc_inode_init(args.mp, tp, NULL, newlen, agno,
|
|
|
|
args.agbno, args.len, prandom_u32());
|
2008-11-28 10:23:38 +07:00
|
|
|
|
2011-09-20 20:56:55 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
2009-09-01 06:56:51 +07:00
|
|
|
/*
|
|
|
|
* Convert the results.
|
|
|
|
*/
|
2018-12-12 23:46:24 +07:00
|
|
|
newino = XFS_AGB_TO_AGINO(args.mp, args.agbno);
|
2015-05-29 06:18:32 +07:00
|
|
|
|
|
|
|
if (xfs_inobt_issparse(~allocmask)) {
|
|
|
|
/*
|
|
|
|
* We've allocated a sparse chunk. Align the startino and mask.
|
|
|
|
*/
|
|
|
|
xfs_align_sparse_ino(args.mp, &newino, &allocmask);
|
|
|
|
|
|
|
|
rec.ir_startino = newino;
|
|
|
|
rec.ir_holemask = ~allocmask;
|
|
|
|
rec.ir_count = newlen;
|
|
|
|
rec.ir_freecount = newlen;
|
|
|
|
rec.ir_free = XFS_INOBT_ALL_FREE;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Insert the sparse record into the inobt and allow for a merge
|
|
|
|
* if necessary. If a merge does occur, rec is updated to the
|
|
|
|
* merged record.
|
|
|
|
*/
|
|
|
|
error = xfs_inobt_insert_sprec(args.mp, tp, agbp, XFS_BTNUM_INO,
|
|
|
|
&rec, true);
|
|
|
|
if (error == -EFSCORRUPTED) {
|
|
|
|
xfs_alert(args.mp,
|
|
|
|
"invalid sparse inode record: ino 0x%llx holemask 0x%x count %u",
|
|
|
|
XFS_AGINO_TO_INO(args.mp, agno,
|
|
|
|
rec.ir_startino),
|
|
|
|
rec.ir_holemask, rec.ir_count);
|
|
|
|
xfs_force_shutdown(args.mp, SHUTDOWN_CORRUPT_INCORE);
|
|
|
|
}
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We can't merge the part we've just allocated as for the inobt
|
|
|
|
* due to finobt semantics. The original record may or may not
|
|
|
|
* exist independent of whether physical inodes exist in this
|
|
|
|
* sparse chunk.
|
|
|
|
*
|
|
|
|
* We must update the finobt record based on the inobt record.
|
|
|
|
* rec contains the fully merged and up to date inobt record
|
|
|
|
* from the previous call. Set merge false to replace any
|
|
|
|
* existing record with this one.
|
|
|
|
*/
|
|
|
|
if (xfs_sb_version_hasfinobt(&args.mp->m_sb)) {
|
|
|
|
error = xfs_inobt_insert_sprec(args.mp, tp, agbp,
|
|
|
|
XFS_BTNUM_FINO, &rec,
|
|
|
|
false);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* full chunk - insert new records to both btrees */
|
|
|
|
error = xfs_inobt_insert(args.mp, tp, agbp, newino, newlen,
|
|
|
|
XFS_BTNUM_INO);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
if (xfs_sb_version_hasfinobt(&args.mp->m_sb)) {
|
|
|
|
error = xfs_inobt_insert(args.mp, tp, agbp, newino,
|
|
|
|
newlen, XFS_BTNUM_FINO);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Update AGI counts and newino.
|
|
|
|
*/
|
2008-02-14 06:03:29 +07:00
|
|
|
be32_add_cpu(&agi->agi_count, newlen);
|
|
|
|
be32_add_cpu(&agi->agi_freecount, newlen);
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(args.mp, agno);
|
|
|
|
pag->pagi_freecount += newlen;
|
2018-06-12 11:52:01 +07:00
|
|
|
pag->pagi_count += newlen;
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-11-02 11:11:25 +07:00
|
|
|
agi->agi_newino = cpu_to_be32(newino);
|
2009-09-01 06:56:51 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Log allocation group header fields
|
|
|
|
*/
|
|
|
|
xfs_ialloc_log_agi(tp, agbp,
|
|
|
|
XFS_AGI_COUNT | XFS_AGI_FREECOUNT | XFS_AGI_NEWINO);
|
|
|
|
/*
|
|
|
|
* Modify/log superblock values for inode count and inode free count.
|
|
|
|
*/
|
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, (long)newlen);
|
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, (long)newlen);
|
|
|
|
*alloc = 1;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-11-14 23:17:22 +07:00
|
|
|
STATIC xfs_agnumber_t
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ialloc_next_ag(
|
|
|
|
xfs_mount_t *mp)
|
|
|
|
{
|
|
|
|
xfs_agnumber_t agno;
|
|
|
|
|
|
|
|
spin_lock(&mp->m_agirotor_lock);
|
|
|
|
agno = mp->m_agirotor;
|
2012-09-20 20:32:36 +07:00
|
|
|
if (++mp->m_agirotor >= mp->m_maxagi)
|
2005-04-17 05:20:36 +07:00
|
|
|
mp->m_agirotor = 0;
|
|
|
|
spin_unlock(&mp->m_agirotor_lock);
|
|
|
|
|
|
|
|
return agno;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Select an allocation group to look for a free inode in, based on the parent
|
2013-08-12 10:14:58 +07:00
|
|
|
* inode and the mode. Return the allocation group buffer.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2012-07-04 21:54:49 +07:00
|
|
|
STATIC xfs_agnumber_t
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ialloc_ag_select(
|
|
|
|
xfs_trans_t *tp, /* transaction pointer */
|
|
|
|
xfs_ino_t parent, /* parent directory inode number */
|
2017-12-05 08:32:55 +07:00
|
|
|
umode_t mode) /* bits set to indicate file type */
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
xfs_agnumber_t agcount; /* number of ag's in the filesystem */
|
|
|
|
xfs_agnumber_t agno; /* current ag number */
|
|
|
|
int flags; /* alloc buffer locking flags */
|
|
|
|
xfs_extlen_t ineed; /* blocks needed for inode allocation */
|
|
|
|
xfs_extlen_t longest = 0; /* longest extent available */
|
|
|
|
xfs_mount_t *mp; /* mount point structure */
|
|
|
|
int needspace; /* file mode implies space allocated */
|
|
|
|
xfs_perag_t *pag; /* per allocation group data */
|
|
|
|
xfs_agnumber_t pagno; /* parent (starting) ag number */
|
2012-07-04 21:54:49 +07:00
|
|
|
int error;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Files of these types need at least one block if length > 0
|
|
|
|
* (and they won't fit in the inode, but that's hard to figure out).
|
|
|
|
*/
|
|
|
|
needspace = S_ISDIR(mode) || S_ISREG(mode) || S_ISLNK(mode);
|
|
|
|
mp = tp->t_mountp;
|
|
|
|
agcount = mp->m_maxagi;
|
|
|
|
if (S_ISDIR(mode))
|
|
|
|
pagno = xfs_ialloc_next_ag(mp);
|
|
|
|
else {
|
|
|
|
pagno = XFS_INO_TO_AGNO(mp, parent);
|
|
|
|
if (pagno >= agcount)
|
|
|
|
pagno = 0;
|
|
|
|
}
|
2012-07-04 21:54:49 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
ASSERT(pagno < agcount);
|
2012-07-04 21:54:49 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Loop through allocation groups, looking for one with a little
|
|
|
|
* free space in it. Note we don't look for free inodes, exactly.
|
|
|
|
* Instead, we include whether there is a need to allocate inodes
|
|
|
|
* to mean that blocks must be allocated for them,
|
|
|
|
* if none are currently free.
|
|
|
|
*/
|
|
|
|
agno = pagno;
|
|
|
|
flags = XFS_ALLOC_FLAG_TRYLOCK;
|
|
|
|
for (;;) {
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(mp, agno);
|
2012-07-04 21:54:49 +07:00
|
|
|
if (!pag->pagi_inodeok) {
|
|
|
|
xfs_ialloc_next_ag(mp);
|
|
|
|
goto nextag;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
if (!pag->pagi_init) {
|
2012-07-04 21:54:49 +07:00
|
|
|
error = xfs_ialloc_pagi_init(mp, tp, agno);
|
|
|
|
if (error)
|
2005-04-17 05:20:36 +07:00
|
|
|
goto nextag;
|
2012-07-04 21:54:49 +07:00
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2012-07-04 21:54:49 +07:00
|
|
|
if (pag->pagi_freecount) {
|
|
|
|
xfs_perag_put(pag);
|
|
|
|
return agno;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2012-07-04 21:54:49 +07:00
|
|
|
if (!pag->pagf_init) {
|
|
|
|
error = xfs_alloc_pagf_init(mp, tp, agno, flags);
|
|
|
|
if (error)
|
2005-04-17 05:20:36 +07:00
|
|
|
goto nextag;
|
|
|
|
}
|
2012-07-04 21:54:49 +07:00
|
|
|
|
|
|
|
/*
|
xfs: fix premature enospc on inode allocation
After growing a filesystem, XFS can fail to allocate inodes even
though there is a large amount of space available in the filesystem
for inodes. The issue is caused by a nearly full allocation group
having enough free space in it to be considered for inode
allocation, but not enough contiguous free space to actually
allocation inodes. This situation results in successful selection
of the AG for allocation, then failure of the allocation resulting
in ENOSPC being reported to the caller.
It is caused by two possible issues. Firstly, we only consider the
lognest free extent and whether it would fit an inode chunk. If the
extent is not correctly aligned, then we can't allocate an inode
chunk in it regardless of the fact that it is large enough. This
tends to be a permanent error until space in the AG is freed.
The second issue is that we don't actually lock the AGI or AGF when
we are doing these checks, and so by the time we get to actually
allocating the inode chunk the space we thought we had in the AG may
have been allocated. This tends to be a spurious error as it
requires a race to trigger. Hence this case is ignored in this patch
as the reported problem is for permanent errors.
The first issue could be addressed by simply taking into account the
alignment when checking the longest extent. This, however, would
prevent allocation in AGs that have aligned, exact sized extents
free. However, this case should be fairly rare compared to the
number of allocations that occur near ENOSPC that would trigger this
condition.
Hence, when selecting the inode AG, take into account the inode
cluster alignment when checking the lognest free extent in the AG.
If we can't find any AGs with a contiguous free space large
enough to be aligned, drop the alignment addition and just try for
an AG that has enough contiguous free space available for an inode
chunk. This won't prevent issues from occurring, but should avoid
situations where other AGs have lots of free space but the selected
AG can't allocate due to alignment constraints.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-12-04 05:42:21 +07:00
|
|
|
* Check that there is enough free space for the file plus a
|
|
|
|
* chunk of inodes if we need to allocate some. If this is the
|
|
|
|
* first pass across the AGs, take into account the potential
|
|
|
|
* space needed for alignment of inode chunks when checking the
|
|
|
|
* longest contiguous free space in the AG - this prevents us
|
|
|
|
* from getting ENOSPC because we have free space larger than
|
|
|
|
* m_ialloc_blks but alignment constraints prevent us from using
|
|
|
|
* it.
|
|
|
|
*
|
|
|
|
* If we can't find an AG with space for full alignment slack to
|
|
|
|
* be taken into account, we must be near ENOSPC in all AGs.
|
|
|
|
* Hence we don't include alignment for the second pass and so
|
|
|
|
* if we fail allocation due to alignment issues then it is most
|
|
|
|
* likely a real ENOSPC condition.
|
2012-07-04 21:54:49 +07:00
|
|
|
*/
|
2015-05-29 05:55:20 +07:00
|
|
|
ineed = mp->m_ialloc_min_blks;
|
xfs: fix premature enospc on inode allocation
After growing a filesystem, XFS can fail to allocate inodes even
though there is a large amount of space available in the filesystem
for inodes. The issue is caused by a nearly full allocation group
having enough free space in it to be considered for inode
allocation, but not enough contiguous free space to actually
allocation inodes. This situation results in successful selection
of the AG for allocation, then failure of the allocation resulting
in ENOSPC being reported to the caller.
It is caused by two possible issues. Firstly, we only consider the
lognest free extent and whether it would fit an inode chunk. If the
extent is not correctly aligned, then we can't allocate an inode
chunk in it regardless of the fact that it is large enough. This
tends to be a permanent error until space in the AG is freed.
The second issue is that we don't actually lock the AGI or AGF when
we are doing these checks, and so by the time we get to actually
allocating the inode chunk the space we thought we had in the AG may
have been allocated. This tends to be a spurious error as it
requires a race to trigger. Hence this case is ignored in this patch
as the reported problem is for permanent errors.
The first issue could be addressed by simply taking into account the
alignment when checking the longest extent. This, however, would
prevent allocation in AGs that have aligned, exact sized extents
free. However, this case should be fairly rare compared to the
number of allocations that occur near ENOSPC that would trigger this
condition.
Hence, when selecting the inode AG, take into account the inode
cluster alignment when checking the lognest free extent in the AG.
If we can't find any AGs with a contiguous free space large
enough to be aligned, drop the alignment addition and just try for
an AG that has enough contiguous free space available for an inode
chunk. This won't prevent issues from occurring, but should avoid
situations where other AGs have lots of free space but the selected
AG can't allocate due to alignment constraints.
Reported-by: Arkadiusz Miskiewicz <arekm@maven.pl>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
2014-12-04 05:42:21 +07:00
|
|
|
if (flags && ineed > 1)
|
2018-12-12 23:46:25 +07:00
|
|
|
ineed += mp->m_cluster_align;
|
2012-07-04 21:54:49 +07:00
|
|
|
longest = pag->pagf_longest;
|
|
|
|
if (!longest)
|
|
|
|
longest = pag->pagf_flcount > 0;
|
|
|
|
|
|
|
|
if (pag->pagf_freeblks >= needspace + ineed &&
|
|
|
|
longest >= ineed) {
|
|
|
|
xfs_perag_put(pag);
|
|
|
|
return agno;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
nextag:
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* No point in iterating over the rest, if we're shutting
|
|
|
|
* down.
|
|
|
|
*/
|
2010-01-11 18:47:44 +07:00
|
|
|
if (XFS_FORCED_SHUTDOWN(mp))
|
2012-07-04 21:54:49 +07:00
|
|
|
return NULLAGNUMBER;
|
2005-04-17 05:20:36 +07:00
|
|
|
agno++;
|
|
|
|
if (agno >= agcount)
|
|
|
|
agno = 0;
|
|
|
|
if (agno == pagno) {
|
2010-01-11 18:47:44 +07:00
|
|
|
if (flags == 0)
|
2012-07-04 21:54:49 +07:00
|
|
|
return NULLAGNUMBER;
|
2005-04-17 05:20:36 +07:00
|
|
|
flags = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-09-01 06:57:14 +07:00
|
|
|
/*
|
|
|
|
* Try to retrieve the next record to the left/right from the current one.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_ialloc_next_rec(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
xfs_inobt_rec_incore_t *rec,
|
|
|
|
int *done,
|
|
|
|
int left)
|
|
|
|
{
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (left)
|
|
|
|
error = xfs_btree_decrement(cur, 0, &i);
|
|
|
|
else
|
|
|
|
error = xfs_btree_increment(cur, 0, &i);
|
|
|
|
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
*done = !i;
|
|
|
|
if (i) {
|
|
|
|
error = xfs_inobt_get_rec(cur, rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2009-09-01 06:57:14 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-09-01 06:58:28 +07:00
|
|
|
STATIC int
|
|
|
|
xfs_ialloc_get_rec(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
xfs_agino_t agino,
|
|
|
|
xfs_inobt_rec_incore_t *rec,
|
2013-06-24 12:37:23 +07:00
|
|
|
int *done)
|
2009-09-01 06:58:28 +07:00
|
|
|
{
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_EQ, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
*done = !i;
|
|
|
|
if (i) {
|
|
|
|
error = xfs_inobt_get_rec(cur, rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2009-09-01 06:58:28 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2009-09-01 06:57:09 +07:00
|
|
|
|
2015-05-29 05:50:21 +07:00
|
|
|
/*
|
2015-05-29 06:20:10 +07:00
|
|
|
* Return the offset of the first free inode in the record. If the inode chunk
|
|
|
|
* is sparsely allocated, we convert the record holemask to inode granularity
|
|
|
|
* and mask off the unallocated regions from the inode free mask.
|
2015-05-29 05:50:21 +07:00
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_inobt_first_free_inode(
|
|
|
|
struct xfs_inobt_rec_incore *rec)
|
|
|
|
{
|
2015-05-29 06:20:10 +07:00
|
|
|
xfs_inofree_t realfree;
|
|
|
|
|
|
|
|
/* if there are no holes, return the first available offset */
|
|
|
|
if (!xfs_inobt_issparse(rec->ir_holemask))
|
|
|
|
return xfs_lowbit64(rec->ir_free);
|
|
|
|
|
|
|
|
realfree = xfs_inobt_irec_to_allocmask(rec);
|
|
|
|
realfree &= rec->ir_free;
|
|
|
|
|
|
|
|
return xfs_lowbit64(realfree);
|
2015-05-29 05:50:21 +07:00
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2014-04-24 13:00:53 +07:00
|
|
|
* Allocate an inode using the inobt-only algorithm.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2012-07-04 21:54:46 +07:00
|
|
|
STATIC int
|
2014-04-24 13:00:53 +07:00
|
|
|
xfs_dialloc_ag_inobt(
|
2012-07-04 21:54:46 +07:00
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
xfs_ino_t parent,
|
|
|
|
xfs_ino_t *inop)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2012-07-04 21:54:46 +07:00
|
|
|
struct xfs_mount *mp = tp->t_mountp;
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
xfs_agnumber_t pagno = XFS_INO_TO_AGNO(mp, parent);
|
|
|
|
xfs_agino_t pagino = XFS_INO_TO_AGINO(mp, parent);
|
|
|
|
struct xfs_perag *pag;
|
|
|
|
struct xfs_btree_cur *cur, *tcur;
|
|
|
|
struct xfs_inobt_rec_incore rec, trec;
|
|
|
|
xfs_ino_t ino;
|
|
|
|
int error;
|
|
|
|
int offset;
|
|
|
|
int i, j;
|
xfs: stop searching for free slots in an inode chunk when there are none
In a filesystem without finobt, the Space manager selects an AG to alloc a new
inode, where xfs_dialloc_ag_inobt() will search the AG for the free slot chunk.
When the new inode is in the same AG as its parent, the btree will be searched
starting on the parent's record, and then retried from the top if no slot is
available beyond the parent's record.
To exit this loop though, xfs_dialloc_ag_inobt() relies on the fact that the
btree must have a free slot available, once its callers relied on the
agi->freecount when deciding how/where to allocate this new inode.
In the case when the agi->freecount is corrupted, showing available inodes in an
AG, when in fact there is none, this becomes an infinite loop.
Add a way to stop the loop when a free slot is not found in the btree, making
the function to fall into the whole AG scan which will then, be able to detect
the corruption and shut the filesystem down.
As pointed by Brian, this might impact performance, giving the fact we
don't reset the search distance anymore when we reach the end of the
tree, giving it fewer tries before falling back to the whole AG search, but
it will only affect searches that start within 10 records to the end of the tree.
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-08-19 08:07:04 +07:00
|
|
|
int searchdistance = 10;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(mp, agno);
|
2009-09-01 06:58:28 +07:00
|
|
|
|
2012-07-04 21:54:48 +07:00
|
|
|
ASSERT(pag->pagi_init);
|
|
|
|
ASSERT(pag->pagi_inodeok);
|
|
|
|
ASSERT(pag->pagi_freecount > 0);
|
|
|
|
|
2009-09-01 06:58:28 +07:00
|
|
|
restart_pagno:
|
2014-04-24 13:00:50 +07:00
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* If pagino is 0 (this is the root inode allocation) use newino.
|
|
|
|
* This must work because we've just allocated some.
|
|
|
|
*/
|
|
|
|
if (!pagino)
|
2005-11-02 11:11:25 +07:00
|
|
|
pagino = be32_to_cpu(agi->agi_newino);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2009-09-01 06:57:09 +07:00
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
2009-09-01 06:57:14 +07:00
|
|
|
* If in the same AG as the parent, try to get near the parent.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
if (pagno == agno) {
|
2009-09-01 06:57:14 +07:00
|
|
|
int doneleft; /* done, to the left */
|
|
|
|
int doneright; /* done, to the right */
|
|
|
|
|
2009-09-01 06:58:21 +07:00
|
|
|
error = xfs_inobt_lookup(cur, pagino, XFS_LOOKUP_LE, &i);
|
2009-09-01 06:57:14 +07:00
|
|
|
if (error)
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &j);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, j == 1, error0);
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
if (rec.ir_freecount > 0) {
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Found a free inode in the same chunk
|
2009-09-01 06:57:14 +07:00
|
|
|
* as the parent, done.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2009-09-01 06:57:14 +07:00
|
|
|
goto alloc_inode;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2009-09-01 06:57:14 +07:00
|
|
|
* In the same AG as parent, but parent's chunk is full.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
|
2009-09-01 06:57:14 +07:00
|
|
|
/* duplicate the cursor, search left & right simultaneously */
|
|
|
|
error = xfs_btree_dup_cursor(cur, &tcur);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
|
2009-09-01 06:58:28 +07:00
|
|
|
/*
|
|
|
|
* Skip to last blocks looked up if same parent inode.
|
|
|
|
*/
|
|
|
|
if (pagino != NULLAGINO &&
|
|
|
|
pag->pagl_pagino == pagino &&
|
|
|
|
pag->pagl_leftrec != NULLAGINO &&
|
|
|
|
pag->pagl_rightrec != NULLAGINO) {
|
|
|
|
error = xfs_ialloc_get_rec(tcur, pag->pagl_leftrec,
|
2013-06-24 12:37:23 +07:00
|
|
|
&trec, &doneleft);
|
2009-09-01 06:58:28 +07:00
|
|
|
if (error)
|
|
|
|
goto error1;
|
2009-09-01 06:57:14 +07:00
|
|
|
|
2009-09-01 06:58:28 +07:00
|
|
|
error = xfs_ialloc_get_rec(cur, pag->pagl_rightrec,
|
2013-06-24 12:37:23 +07:00
|
|
|
&rec, &doneright);
|
2009-09-01 06:58:28 +07:00
|
|
|
if (error)
|
|
|
|
goto error1;
|
|
|
|
} else {
|
|
|
|
/* search left with tcur, back up 1 record */
|
|
|
|
error = xfs_ialloc_next_rec(tcur, &trec, &doneleft, 1);
|
|
|
|
if (error)
|
|
|
|
goto error1;
|
|
|
|
|
|
|
|
/* search right with cur, go forward 1 record. */
|
|
|
|
error = xfs_ialloc_next_rec(cur, &rec, &doneright, 0);
|
|
|
|
if (error)
|
|
|
|
goto error1;
|
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Loop until we find an inode chunk with a free inode.
|
|
|
|
*/
|
xfs: stop searching for free slots in an inode chunk when there are none
In a filesystem without finobt, the Space manager selects an AG to alloc a new
inode, where xfs_dialloc_ag_inobt() will search the AG for the free slot chunk.
When the new inode is in the same AG as its parent, the btree will be searched
starting on the parent's record, and then retried from the top if no slot is
available beyond the parent's record.
To exit this loop though, xfs_dialloc_ag_inobt() relies on the fact that the
btree must have a free slot available, once its callers relied on the
agi->freecount when deciding how/where to allocate this new inode.
In the case when the agi->freecount is corrupted, showing available inodes in an
AG, when in fact there is none, this becomes an infinite loop.
Add a way to stop the loop when a free slot is not found in the btree, making
the function to fall into the whole AG scan which will then, be able to detect
the corruption and shut the filesystem down.
As pointed by Brian, this might impact performance, giving the fact we
don't reset the search distance anymore when we reach the end of the
tree, giving it fewer tries before falling back to the whole AG search, but
it will only affect searches that start within 10 records to the end of the tree.
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-08-19 08:07:04 +07:00
|
|
|
while (--searchdistance > 0 && (!doneleft || !doneright)) {
|
2009-09-01 06:57:14 +07:00
|
|
|
int useleft; /* using left inode chunk this time */
|
|
|
|
|
|
|
|
/* figure out the closer block if both are valid. */
|
|
|
|
if (!doneleft && !doneright) {
|
|
|
|
useleft = pagino -
|
|
|
|
(trec.ir_startino + XFS_INODES_PER_CHUNK - 1) <
|
|
|
|
rec.ir_startino - pagino;
|
|
|
|
} else {
|
|
|
|
useleft = !doneleft;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
/* free inodes to the left? */
|
|
|
|
if (useleft && trec.ir_freecount) {
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
cur = tcur;
|
2009-09-01 06:58:28 +07:00
|
|
|
|
|
|
|
pag->pagl_leftrec = trec.ir_startino;
|
|
|
|
pag->pagl_rightrec = rec.ir_startino;
|
|
|
|
pag->pagl_pagino = pagino;
|
2017-08-11 23:00:06 +07:00
|
|
|
rec = trec;
|
2009-09-01 06:57:14 +07:00
|
|
|
goto alloc_inode;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2009-09-01 06:57:14 +07:00
|
|
|
/* free inodes to the right? */
|
|
|
|
if (!useleft && rec.ir_freecount) {
|
|
|
|
xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
|
2009-09-01 06:58:28 +07:00
|
|
|
|
|
|
|
pag->pagl_leftrec = trec.ir_startino;
|
|
|
|
pag->pagl_rightrec = rec.ir_startino;
|
|
|
|
pag->pagl_pagino = pagino;
|
2009-09-01 06:57:14 +07:00
|
|
|
goto alloc_inode;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
/* get next record to check */
|
|
|
|
if (useleft) {
|
|
|
|
error = xfs_ialloc_next_rec(tcur, &trec,
|
|
|
|
&doneleft, 1);
|
|
|
|
} else {
|
|
|
|
error = xfs_ialloc_next_rec(cur, &rec,
|
|
|
|
&doneright, 0);
|
|
|
|
}
|
|
|
|
if (error)
|
|
|
|
goto error1;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:58:28 +07:00
|
|
|
|
xfs: stop searching for free slots in an inode chunk when there are none
In a filesystem without finobt, the Space manager selects an AG to alloc a new
inode, where xfs_dialloc_ag_inobt() will search the AG for the free slot chunk.
When the new inode is in the same AG as its parent, the btree will be searched
starting on the parent's record, and then retried from the top if no slot is
available beyond the parent's record.
To exit this loop though, xfs_dialloc_ag_inobt() relies on the fact that the
btree must have a free slot available, once its callers relied on the
agi->freecount when deciding how/where to allocate this new inode.
In the case when the agi->freecount is corrupted, showing available inodes in an
AG, when in fact there is none, this becomes an infinite loop.
Add a way to stop the loop when a free slot is not found in the btree, making
the function to fall into the whole AG scan which will then, be able to detect
the corruption and shut the filesystem down.
As pointed by Brian, this might impact performance, giving the fact we
don't reset the search distance anymore when we reach the end of the
tree, giving it fewer tries before falling back to the whole AG search, but
it will only affect searches that start within 10 records to the end of the tree.
Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-08-19 08:07:04 +07:00
|
|
|
if (searchdistance <= 0) {
|
|
|
|
/*
|
|
|
|
* Not in range - save last search
|
|
|
|
* location and allocate a new inode
|
|
|
|
*/
|
|
|
|
xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
|
|
|
|
pag->pagl_leftrec = trec.ir_startino;
|
|
|
|
pag->pagl_rightrec = rec.ir_startino;
|
|
|
|
pag->pagl_pagino = pagino;
|
|
|
|
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* We've reached the end of the btree. because
|
|
|
|
* we are only searching a small chunk of the
|
|
|
|
* btree each search, there is obviously free
|
|
|
|
* inodes closer to the parent inode than we
|
|
|
|
* are now. restart the search again.
|
|
|
|
*/
|
|
|
|
pag->pagl_pagino = NULLAGINO;
|
|
|
|
pag->pagl_leftrec = NULLAGINO;
|
|
|
|
pag->pagl_rightrec = NULLAGINO;
|
|
|
|
xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
goto restart_pagno;
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2009-09-01 06:57:14 +07:00
|
|
|
* In a different AG from the parent.
|
2005-04-17 05:20:36 +07:00
|
|
|
* See if the most recently allocated block has any free.
|
|
|
|
*/
|
2011-07-08 19:36:05 +07:00
|
|
|
if (agi->agi_newino != cpu_to_be32(NULLAGINO)) {
|
2009-09-01 06:58:21 +07:00
|
|
|
error = xfs_inobt_lookup(cur, be32_to_cpu(agi->agi_newino),
|
|
|
|
XFS_LOOKUP_EQ, &i);
|
2009-09-01 06:57:14 +07:00
|
|
|
if (error)
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
if (i == 1) {
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &j);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
|
|
|
|
if (j == 1 && rec.ir_freecount > 0) {
|
|
|
|
/*
|
|
|
|
* The last chunk allocated in the group
|
|
|
|
* still has a free inode.
|
|
|
|
*/
|
|
|
|
goto alloc_inode;
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:58:28 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
2009-09-01 06:58:28 +07:00
|
|
|
/*
|
|
|
|
* None left in the last group, search the whole AG
|
|
|
|
*/
|
|
|
|
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2009-09-01 06:58:28 +07:00
|
|
|
|
|
|
|
for (;;) {
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2009-09-01 06:58:28 +07:00
|
|
|
if (rec.ir_freecount > 0)
|
|
|
|
break;
|
|
|
|
error = xfs_btree_increment(cur, 0, &i);
|
2009-09-01 06:57:14 +07:00
|
|
|
if (error)
|
|
|
|
goto error0;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2009-09-01 06:57:14 +07:00
|
|
|
|
|
|
|
alloc_inode:
|
2015-05-29 05:50:21 +07:00
|
|
|
offset = xfs_inobt_first_free_inode(&rec);
|
2005-04-17 05:20:36 +07:00
|
|
|
ASSERT(offset >= 0);
|
|
|
|
ASSERT(offset < XFS_INODES_PER_CHUNK);
|
|
|
|
ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
|
|
|
|
XFS_INODES_PER_CHUNK) == 0);
|
|
|
|
ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
|
2009-02-09 14:37:14 +07:00
|
|
|
rec.ir_free &= ~XFS_INOBT_MASK(offset);
|
2005-04-17 05:20:36 +07:00
|
|
|
rec.ir_freecount--;
|
2009-09-01 06:57:03 +07:00
|
|
|
error = xfs_inobt_update(cur, &rec);
|
|
|
|
if (error)
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
2008-02-14 06:03:29 +07:00
|
|
|
be32_add_cpu(&agi->agi_freecount, -1);
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
|
2010-01-11 18:47:43 +07:00
|
|
|
pag->pagi_freecount--;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2009-09-01 06:57:09 +07:00
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
*inop = ino;
|
|
|
|
return 0;
|
|
|
|
error1:
|
|
|
|
xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
|
|
|
|
error0:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
/*
|
|
|
|
* Use the free inode btree to allocate an inode based on distance from the
|
|
|
|
* parent. Note that the provided cursor may be deleted and replaced.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_dialloc_ag_finobt_near(
|
|
|
|
xfs_agino_t pagino,
|
|
|
|
struct xfs_btree_cur **ocur,
|
|
|
|
struct xfs_inobt_rec_incore *rec)
|
|
|
|
{
|
|
|
|
struct xfs_btree_cur *lcur = *ocur; /* left search cursor */
|
|
|
|
struct xfs_btree_cur *rcur; /* right search cursor */
|
|
|
|
struct xfs_inobt_rec_incore rrec;
|
|
|
|
int error;
|
|
|
|
int i, j;
|
|
|
|
|
|
|
|
error = xfs_inobt_lookup(lcur, pagino, XFS_LOOKUP_LE, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
if (i == 1) {
|
|
|
|
error = xfs_inobt_get_rec(lcur, rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(lcur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* See if we've landed in the parent inode record. The finobt
|
|
|
|
* only tracks chunks with at least one free inode, so record
|
|
|
|
* existence is enough.
|
|
|
|
*/
|
|
|
|
if (pagino >= rec->ir_startino &&
|
|
|
|
pagino < (rec->ir_startino + XFS_INODES_PER_CHUNK))
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
error = xfs_btree_dup_cursor(lcur, &rcur);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
error = xfs_inobt_lookup(rcur, pagino, XFS_LOOKUP_GE, &j);
|
|
|
|
if (error)
|
|
|
|
goto error_rcur;
|
|
|
|
if (j == 1) {
|
|
|
|
error = xfs_inobt_get_rec(rcur, &rrec, &j);
|
|
|
|
if (error)
|
|
|
|
goto error_rcur;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(lcur->bc_mp, j == 1, error_rcur);
|
2014-04-24 13:00:53 +07:00
|
|
|
}
|
|
|
|
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(lcur->bc_mp, i == 1 || j == 1, error_rcur);
|
2014-04-24 13:00:53 +07:00
|
|
|
if (i == 1 && j == 1) {
|
|
|
|
/*
|
|
|
|
* Both the left and right records are valid. Choose the closer
|
|
|
|
* inode chunk to the target.
|
|
|
|
*/
|
|
|
|
if ((pagino - rec->ir_startino + XFS_INODES_PER_CHUNK - 1) >
|
|
|
|
(rrec.ir_startino - pagino)) {
|
|
|
|
*rec = rrec;
|
|
|
|
xfs_btree_del_cursor(lcur, XFS_BTREE_NOERROR);
|
|
|
|
*ocur = rcur;
|
|
|
|
} else {
|
|
|
|
xfs_btree_del_cursor(rcur, XFS_BTREE_NOERROR);
|
|
|
|
}
|
|
|
|
} else if (j == 1) {
|
|
|
|
/* only the right record is valid */
|
|
|
|
*rec = rrec;
|
|
|
|
xfs_btree_del_cursor(lcur, XFS_BTREE_NOERROR);
|
|
|
|
*ocur = rcur;
|
|
|
|
} else if (i == 1) {
|
|
|
|
/* only the left record is valid */
|
|
|
|
xfs_btree_del_cursor(rcur, XFS_BTREE_NOERROR);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
error_rcur:
|
|
|
|
xfs_btree_del_cursor(rcur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Use the free inode btree to find a free inode based on a newino hint. If
|
|
|
|
* the hint is NULL, find the first free inode in the AG.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_dialloc_ag_finobt_newino(
|
|
|
|
struct xfs_agi *agi,
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
struct xfs_inobt_rec_incore *rec)
|
|
|
|
{
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (agi->agi_newino != cpu_to_be32(NULLAGINO)) {
|
2014-09-29 07:43:15 +07:00
|
|
|
error = xfs_inobt_lookup(cur, be32_to_cpu(agi->agi_newino),
|
|
|
|
XFS_LOOKUP_EQ, &i);
|
2014-04-24 13:00:53 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
if (i == 1) {
|
|
|
|
error = xfs_inobt_get_rec(cur, rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Find the first inode available in the AG.
|
|
|
|
*/
|
|
|
|
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
|
|
|
error = xfs_inobt_get_rec(cur, rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Update the inobt based on a modification made to the finobt. Also ensure that
|
|
|
|
* the records from both trees are equivalent post-modification.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_dialloc_ag_update_inobt(
|
|
|
|
struct xfs_btree_cur *cur, /* inobt cursor */
|
|
|
|
struct xfs_inobt_rec_incore *frec, /* finobt record */
|
|
|
|
int offset) /* inode offset */
|
|
|
|
{
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
error = xfs_inobt_lookup(cur, frec->ir_startino, XFS_LOOKUP_EQ, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error)
|
|
|
|
return error;
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, i == 1);
|
2014-04-24 13:00:53 +07:00
|
|
|
ASSERT((XFS_AGINO_TO_OFFSET(cur->bc_mp, rec.ir_startino) %
|
|
|
|
XFS_INODES_PER_CHUNK) == 0);
|
|
|
|
|
|
|
|
rec.ir_free &= ~XFS_INOBT_MASK(offset);
|
|
|
|
rec.ir_freecount--;
|
|
|
|
|
2015-02-23 18:39:13 +07:00
|
|
|
XFS_WANT_CORRUPTED_RETURN(cur->bc_mp, (rec.ir_free == frec->ir_free) &&
|
2014-04-24 13:00:53 +07:00
|
|
|
(rec.ir_freecount == frec->ir_freecount));
|
|
|
|
|
2014-12-01 04:24:58 +07:00
|
|
|
return xfs_inobt_update(cur, &rec);
|
2014-04-24 13:00:53 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate an inode using the free inode btree, if available. Otherwise, fall
|
|
|
|
* back to the inobt search algorithm.
|
|
|
|
*
|
|
|
|
* The caller selected an AG for us, and made sure that free inodes are
|
|
|
|
* available.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_dialloc_ag(
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
xfs_ino_t parent,
|
|
|
|
xfs_ino_t *inop)
|
|
|
|
{
|
|
|
|
struct xfs_mount *mp = tp->t_mountp;
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
xfs_agnumber_t pagno = XFS_INO_TO_AGNO(mp, parent);
|
|
|
|
xfs_agino_t pagino = XFS_INO_TO_AGINO(mp, parent);
|
|
|
|
struct xfs_perag *pag;
|
|
|
|
struct xfs_btree_cur *cur; /* finobt cursor */
|
|
|
|
struct xfs_btree_cur *icur; /* inobt cursor */
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
xfs_ino_t ino;
|
|
|
|
int error;
|
|
|
|
int offset;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (!xfs_sb_version_hasfinobt(&mp->m_sb))
|
|
|
|
return xfs_dialloc_ag_inobt(tp, agbp, parent, inop);
|
|
|
|
|
|
|
|
pag = xfs_perag_get(mp, agno);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If pagino is 0 (this is the root inode allocation) use newino.
|
|
|
|
* This must work because we've just allocated some.
|
|
|
|
*/
|
|
|
|
if (!pagino)
|
|
|
|
pagino = be32_to_cpu(agi->agi_newino);
|
|
|
|
|
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_FINO);
|
|
|
|
|
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error_cur;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The search algorithm depends on whether we're in the same AG as the
|
|
|
|
* parent. If so, find the closest available inode to the parent. If
|
|
|
|
* not, consider the agi hint or find the first free inode in the AG.
|
|
|
|
*/
|
|
|
|
if (agno == pagno)
|
|
|
|
error = xfs_dialloc_ag_finobt_near(pagino, &cur, &rec);
|
|
|
|
else
|
|
|
|
error = xfs_dialloc_ag_finobt_newino(agi, cur, &rec);
|
|
|
|
if (error)
|
|
|
|
goto error_cur;
|
|
|
|
|
2015-05-29 05:50:21 +07:00
|
|
|
offset = xfs_inobt_first_free_inode(&rec);
|
2014-04-24 13:00:53 +07:00
|
|
|
ASSERT(offset >= 0);
|
|
|
|
ASSERT(offset < XFS_INODES_PER_CHUNK);
|
|
|
|
ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
|
|
|
|
XFS_INODES_PER_CHUNK) == 0);
|
|
|
|
ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Modify or remove the finobt record.
|
|
|
|
*/
|
|
|
|
rec.ir_free &= ~XFS_INOBT_MASK(offset);
|
|
|
|
rec.ir_freecount--;
|
|
|
|
if (rec.ir_freecount)
|
|
|
|
error = xfs_inobt_update(cur, &rec);
|
|
|
|
else
|
|
|
|
error = xfs_btree_delete(cur, &i);
|
|
|
|
if (error)
|
|
|
|
goto error_cur;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The finobt has now been updated appropriately. We haven't updated the
|
|
|
|
* agi and superblock yet, so we can create an inobt cursor and validate
|
|
|
|
* the original freecount. If all is well, make the equivalent update to
|
|
|
|
* the inobt using the finobt record and offset information.
|
|
|
|
*/
|
|
|
|
icur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
|
|
|
|
|
|
|
|
error = xfs_check_agi_freecount(icur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error_icur;
|
|
|
|
|
|
|
|
error = xfs_dialloc_ag_update_inobt(icur, &rec, offset);
|
|
|
|
if (error)
|
|
|
|
goto error_icur;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Both trees have now been updated. We must update the perag and
|
|
|
|
* superblock before we can check the freecount for each btree.
|
|
|
|
*/
|
|
|
|
be32_add_cpu(&agi->agi_freecount, -1);
|
|
|
|
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
|
|
|
|
pag->pagi_freecount--;
|
|
|
|
|
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
|
|
|
|
|
|
|
|
error = xfs_check_agi_freecount(icur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error_icur;
|
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error_icur;
|
|
|
|
|
|
|
|
xfs_btree_del_cursor(icur, XFS_BTREE_NOERROR);
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
xfs_perag_put(pag);
|
|
|
|
*inop = ino;
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
error_icur:
|
|
|
|
xfs_btree_del_cursor(icur, XFS_BTREE_ERROR);
|
|
|
|
error_cur:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
xfs_perag_put(pag);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2012-07-04 21:54:46 +07:00
|
|
|
/*
|
|
|
|
* Allocate an inode on disk.
|
|
|
|
*
|
|
|
|
* Mode is used to tell whether the new inode will need space, and whether it
|
|
|
|
* is a directory.
|
|
|
|
*
|
|
|
|
* This function is designed to be called twice if it has to do an allocation
|
|
|
|
* to make more free inodes. On the first call, *IO_agbp should be set to NULL.
|
|
|
|
* If an inode is available without having to performn an allocation, an inode
|
2012-10-20 21:08:19 +07:00
|
|
|
* number is returned. In this case, *IO_agbp is set to NULL. If an allocation
|
|
|
|
* needs to be done, xfs_dialloc returns the current AGI buffer in *IO_agbp.
|
|
|
|
* The caller should then commit the current transaction, allocate a
|
2012-07-04 21:54:46 +07:00
|
|
|
* new transaction, and call xfs_dialloc() again, passing in the previous value
|
|
|
|
* of *IO_agbp. IO_agbp should be held across the transactions. Since the AGI
|
|
|
|
* buffer is locked across the two calls, the second call is guaranteed to have
|
|
|
|
* a free inode available.
|
|
|
|
*
|
|
|
|
* Once we successfully pick an inode its number is returned and the on-disk
|
|
|
|
* data structures are updated. The inode itself is not read in, since doing so
|
|
|
|
* would break ordering constraints with xfs_reclaim.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_dialloc(
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
xfs_ino_t parent,
|
|
|
|
umode_t mode,
|
|
|
|
struct xfs_buf **IO_agbp,
|
|
|
|
xfs_ino_t *inop)
|
|
|
|
{
|
2012-07-04 21:54:49 +07:00
|
|
|
struct xfs_mount *mp = tp->t_mountp;
|
2012-07-04 21:54:46 +07:00
|
|
|
struct xfs_buf *agbp;
|
|
|
|
xfs_agnumber_t agno;
|
|
|
|
int error;
|
|
|
|
int ialloced;
|
|
|
|
int noroom = 0;
|
2012-07-04 21:54:50 +07:00
|
|
|
xfs_agnumber_t start_agno;
|
2012-07-04 21:54:46 +07:00
|
|
|
struct xfs_perag *pag;
|
2017-12-05 08:32:55 +07:00
|
|
|
int okalloc = 1;
|
2012-07-04 21:54:46 +07:00
|
|
|
|
2012-07-04 21:54:48 +07:00
|
|
|
if (*IO_agbp) {
|
2012-07-04 21:54:46 +07:00
|
|
|
/*
|
2012-07-04 21:54:48 +07:00
|
|
|
* If the caller passes in a pointer to the AGI buffer,
|
|
|
|
* continue where we left off before. In this case, we
|
2012-07-04 21:54:46 +07:00
|
|
|
* know that the allocation group has free inodes.
|
|
|
|
*/
|
|
|
|
agbp = *IO_agbp;
|
2012-07-04 21:54:48 +07:00
|
|
|
goto out_alloc;
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
2012-07-04 21:54:48 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* We do not have an agbp, so select an initial allocation
|
|
|
|
* group for inode allocation.
|
|
|
|
*/
|
2017-12-05 08:32:55 +07:00
|
|
|
start_agno = xfs_ialloc_ag_select(tp, parent, mode);
|
2012-07-04 21:54:50 +07:00
|
|
|
if (start_agno == NULLAGNUMBER) {
|
2012-07-04 21:54:48 +07:00
|
|
|
*inop = NULLFSINO;
|
|
|
|
return 0;
|
|
|
|
}
|
2012-07-04 21:54:49 +07:00
|
|
|
|
2012-07-04 21:54:46 +07:00
|
|
|
/*
|
|
|
|
* If we have already hit the ceiling of inode blocks then clear
|
|
|
|
* okalloc so we scan all available agi structures for a free
|
|
|
|
* inode.
|
2015-05-29 04:39:34 +07:00
|
|
|
*
|
|
|
|
* Read rough value of mp->m_icount by percpu_counter_read_positive,
|
|
|
|
* which will sacrifice the preciseness but improve the performance.
|
2012-07-04 21:54:46 +07:00
|
|
|
*/
|
|
|
|
if (mp->m_maxicount &&
|
2015-05-29 04:39:34 +07:00
|
|
|
percpu_counter_read_positive(&mp->m_icount) + mp->m_ialloc_inos
|
|
|
|
> mp->m_maxicount) {
|
2012-07-04 21:54:46 +07:00
|
|
|
noroom = 1;
|
|
|
|
okalloc = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Loop until we find an allocation group that either has free inodes
|
|
|
|
* or in which we can allocate some inodes. Iterate through the
|
|
|
|
* allocation groups upward, wrapping at the end.
|
|
|
|
*/
|
2012-07-04 21:54:50 +07:00
|
|
|
agno = start_agno;
|
|
|
|
for (;;) {
|
|
|
|
pag = xfs_perag_get(mp, agno);
|
|
|
|
if (!pag->pagi_inodeok) {
|
|
|
|
xfs_ialloc_next_ag(mp);
|
|
|
|
goto nextag;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!pag->pagi_init) {
|
|
|
|
error = xfs_ialloc_pagi_init(mp, tp, agno);
|
|
|
|
if (error)
|
|
|
|
goto out_error;
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
2012-07-04 21:54:50 +07:00
|
|
|
|
2012-07-04 21:54:46 +07:00
|
|
|
/*
|
2012-07-04 21:54:50 +07:00
|
|
|
* Do a first racy fast path check if this AG is usable.
|
2012-07-04 21:54:46 +07:00
|
|
|
*/
|
2012-07-04 21:54:50 +07:00
|
|
|
if (!pag->pagi_freecount && !okalloc)
|
|
|
|
goto nextag;
|
|
|
|
|
2012-08-07 13:02:02 +07:00
|
|
|
/*
|
|
|
|
* Then read in the AGI buffer and recheck with the AGI buffer
|
|
|
|
* lock held.
|
|
|
|
*/
|
2012-07-04 21:54:50 +07:00
|
|
|
error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
|
|
|
|
if (error)
|
|
|
|
goto out_error;
|
|
|
|
|
|
|
|
if (pag->pagi_freecount) {
|
|
|
|
xfs_perag_put(pag);
|
|
|
|
goto out_alloc;
|
|
|
|
}
|
|
|
|
|
2012-08-07 13:02:02 +07:00
|
|
|
if (!okalloc)
|
|
|
|
goto nextag_relse_buffer;
|
|
|
|
|
2012-07-04 21:54:50 +07:00
|
|
|
|
|
|
|
error = xfs_ialloc_ag_alloc(tp, agbp, &ialloced);
|
|
|
|
if (error) {
|
|
|
|
xfs_trans_brelse(tp, agbp);
|
|
|
|
|
2014-06-25 11:58:08 +07:00
|
|
|
if (error != -ENOSPC)
|
2012-07-04 21:54:50 +07:00
|
|
|
goto out_error;
|
|
|
|
|
|
|
|
xfs_perag_put(pag);
|
2012-07-04 21:54:46 +07:00
|
|
|
*inop = NULLFSINO;
|
2012-07-04 21:54:50 +07:00
|
|
|
return 0;
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
2012-07-04 21:54:50 +07:00
|
|
|
|
|
|
|
if (ialloced) {
|
|
|
|
/*
|
|
|
|
* We successfully allocated some inodes, return
|
|
|
|
* the current context to the caller so that it
|
|
|
|
* can commit the current transaction and call
|
|
|
|
* us again where we left off.
|
|
|
|
*/
|
|
|
|
ASSERT(pag->pagi_freecount > 0);
|
2012-07-04 21:54:46 +07:00
|
|
|
xfs_perag_put(pag);
|
2012-07-04 21:54:50 +07:00
|
|
|
|
|
|
|
*IO_agbp = agbp;
|
|
|
|
*inop = NULLFSINO;
|
|
|
|
return 0;
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
2012-07-04 21:54:50 +07:00
|
|
|
|
2012-08-07 13:02:02 +07:00
|
|
|
nextag_relse_buffer:
|
|
|
|
xfs_trans_brelse(tp, agbp);
|
2012-07-04 21:54:50 +07:00
|
|
|
nextag:
|
2012-07-04 21:54:46 +07:00
|
|
|
xfs_perag_put(pag);
|
2012-07-04 21:54:50 +07:00
|
|
|
if (++agno == mp->m_sb.sb_agcount)
|
|
|
|
agno = 0;
|
|
|
|
if (agno == start_agno) {
|
|
|
|
*inop = NULLFSINO;
|
2014-06-25 11:58:08 +07:00
|
|
|
return noroom ? -ENOSPC : 0;
|
2012-07-04 21:54:50 +07:00
|
|
|
}
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
|
|
|
|
2012-07-04 21:54:48 +07:00
|
|
|
out_alloc:
|
2012-07-04 21:54:46 +07:00
|
|
|
*IO_agbp = NULL;
|
|
|
|
return xfs_dialloc_ag(tp, agbp, parent, inop);
|
2012-07-04 21:54:50 +07:00
|
|
|
out_error:
|
|
|
|
xfs_perag_put(pag);
|
2014-06-22 12:04:54 +07:00
|
|
|
return error;
|
2012-07-04 21:54:46 +07:00
|
|
|
}
|
|
|
|
|
2015-05-29 06:22:52 +07:00
|
|
|
/*
|
|
|
|
* Free the blocks of an inode chunk. We must consider that the inode chunk
|
|
|
|
* might be sparse and only free the regions that are allocated as part of the
|
|
|
|
* chunk.
|
|
|
|
*/
|
|
|
|
STATIC void
|
|
|
|
xfs_difree_inode_chunk(
|
2018-08-01 21:20:34 +07:00
|
|
|
struct xfs_trans *tp,
|
2015-05-29 06:22:52 +07:00
|
|
|
xfs_agnumber_t agno,
|
2018-08-01 21:20:34 +07:00
|
|
|
struct xfs_inobt_rec_incore *rec)
|
2015-05-29 06:22:52 +07:00
|
|
|
{
|
2018-08-01 21:20:34 +07:00
|
|
|
struct xfs_mount *mp = tp->t_mountp;
|
|
|
|
xfs_agblock_t sagbno = XFS_AGINO_TO_AGBNO(mp,
|
|
|
|
rec->ir_startino);
|
|
|
|
int startidx, endidx;
|
|
|
|
int nextbit;
|
|
|
|
xfs_agblock_t agbno;
|
|
|
|
int contigblk;
|
2015-05-29 06:22:52 +07:00
|
|
|
DECLARE_BITMAP(holemask, XFS_INOBT_HOLEMASK_BITS);
|
|
|
|
|
|
|
|
if (!xfs_inobt_issparse(rec->ir_holemask)) {
|
|
|
|
/* not sparse, calculate extent info directly */
|
2018-08-01 21:20:34 +07:00
|
|
|
xfs_bmap_add_free(tp, XFS_AGB_TO_FSB(mp, agno, sagbno),
|
2018-12-12 23:46:23 +07:00
|
|
|
mp->m_ialloc_blks, &XFS_RMAP_OINFO_INODES);
|
2015-05-29 06:22:52 +07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* holemask is only 16-bits (fits in an unsigned long) */
|
|
|
|
ASSERT(sizeof(rec->ir_holemask) <= sizeof(holemask[0]));
|
|
|
|
holemask[0] = rec->ir_holemask;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Find contiguous ranges of zeroes (i.e., allocated regions) in the
|
|
|
|
* holemask and convert the start/end index of each range to an extent.
|
|
|
|
* We start with the start and end index both pointing at the first 0 in
|
|
|
|
* the mask.
|
|
|
|
*/
|
|
|
|
startidx = endidx = find_first_zero_bit(holemask,
|
|
|
|
XFS_INOBT_HOLEMASK_BITS);
|
|
|
|
nextbit = startidx + 1;
|
|
|
|
while (startidx < XFS_INOBT_HOLEMASK_BITS) {
|
|
|
|
nextbit = find_next_zero_bit(holemask, XFS_INOBT_HOLEMASK_BITS,
|
|
|
|
nextbit);
|
|
|
|
/*
|
|
|
|
* If the next zero bit is contiguous, update the end index of
|
|
|
|
* the current range and continue.
|
|
|
|
*/
|
|
|
|
if (nextbit != XFS_INOBT_HOLEMASK_BITS &&
|
|
|
|
nextbit == endidx + 1) {
|
|
|
|
endidx = nextbit;
|
|
|
|
goto next;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* nextbit is not contiguous with the current end index. Convert
|
|
|
|
* the current start/end to an extent and add it to the free
|
|
|
|
* list.
|
|
|
|
*/
|
|
|
|
agbno = sagbno + (startidx * XFS_INODES_PER_HOLEMASK_BIT) /
|
|
|
|
mp->m_sb.sb_inopblock;
|
|
|
|
contigblk = ((endidx - startidx + 1) *
|
|
|
|
XFS_INODES_PER_HOLEMASK_BIT) /
|
|
|
|
mp->m_sb.sb_inopblock;
|
|
|
|
|
|
|
|
ASSERT(agbno % mp->m_sb.sb_spino_align == 0);
|
|
|
|
ASSERT(contigblk % mp->m_sb.sb_spino_align == 0);
|
2018-08-01 21:20:34 +07:00
|
|
|
xfs_bmap_add_free(tp, XFS_AGB_TO_FSB(mp, agno, agbno),
|
2018-12-12 23:46:23 +07:00
|
|
|
contigblk, &XFS_RMAP_OINFO_INODES);
|
2015-05-29 06:22:52 +07:00
|
|
|
|
|
|
|
/* reset range to current bit and carry on... */
|
|
|
|
startidx = endidx = nextbit;
|
|
|
|
|
|
|
|
next:
|
|
|
|
nextbit++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
STATIC int
|
|
|
|
xfs_difree_inobt(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
xfs_agino_t agino,
|
2015-05-29 06:26:03 +07:00
|
|
|
struct xfs_icluster *xic,
|
2014-04-24 13:00:53 +07:00
|
|
|
struct xfs_inobt_rec_incore *orec)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2014-04-24 13:00:53 +07:00
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
struct xfs_perag *pag;
|
|
|
|
struct xfs_btree_cur *cur;
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
int ilen;
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
int off;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2011-07-08 19:36:05 +07:00
|
|
|
ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
|
2014-04-24 13:00:53 +07:00
|
|
|
ASSERT(XFS_AGINO_TO_AGBNO(mp, agino) < be32_to_cpu(agi->agi_length));
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Initialize the cursor.
|
|
|
|
*/
|
2014-04-24 13:00:50 +07:00
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2009-09-01 06:57:09 +07:00
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Look for the entry describing this inode.
|
|
|
|
*/
|
2009-09-01 06:58:21 +07:00
|
|
|
if ((error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i))) {
|
2011-03-07 06:08:35 +07:00
|
|
|
xfs_warn(mp, "%s: xfs_inobt_lookup() returned error %d.",
|
|
|
|
__func__, error);
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
|
|
|
}
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2009-09-01 06:56:58 +07:00
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error) {
|
2011-03-07 06:08:35 +07:00
|
|
|
xfs_warn(mp, "%s: xfs_inobt_get_rec() returned error %d.",
|
|
|
|
__func__, error);
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
|
|
|
}
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error0);
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Get the offset in the inode chunk.
|
|
|
|
*/
|
|
|
|
off = agino - rec.ir_startino;
|
|
|
|
ASSERT(off >= 0 && off < XFS_INODES_PER_CHUNK);
|
2009-02-09 14:37:14 +07:00
|
|
|
ASSERT(!(rec.ir_free & XFS_INOBT_MASK(off)));
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Mark the inode free & increment the count.
|
|
|
|
*/
|
2009-02-09 14:37:14 +07:00
|
|
|
rec.ir_free |= XFS_INOBT_MASK(off);
|
2005-04-17 05:20:36 +07:00
|
|
|
rec.ir_freecount++;
|
|
|
|
|
|
|
|
/*
|
2015-05-29 05:51:37 +07:00
|
|
|
* When an inode chunk is free, it becomes eligible for removal. Don't
|
|
|
|
* remove the chunk if the block size is large enough for multiple inode
|
|
|
|
* chunks (that might not be free).
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2008-02-29 09:58:40 +07:00
|
|
|
if (!(mp->m_flags & XFS_MOUNT_IKEEP) &&
|
2015-05-29 05:51:37 +07:00
|
|
|
rec.ir_free == XFS_INOBT_ALL_FREE &&
|
|
|
|
mp->m_sb.sb_inopblock <= XFS_INODES_PER_CHUNK) {
|
2017-10-10 01:38:54 +07:00
|
|
|
xic->deleted = true;
|
2015-05-29 06:26:03 +07:00
|
|
|
xic->first_ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino);
|
|
|
|
xic->alloc = xfs_inobt_irec_to_allocmask(&rec);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Remove the inode cluster from the AGI B+Tree, adjust the
|
|
|
|
* AGI and Superblock inode counts, and mark the disk space
|
|
|
|
* to be freed when the transaction is committed.
|
|
|
|
*/
|
2015-05-29 05:51:37 +07:00
|
|
|
ilen = rec.ir_freecount;
|
2008-02-14 06:03:29 +07:00
|
|
|
be32_add_cpu(&agi->agi_count, -ilen);
|
|
|
|
be32_add_cpu(&agi->agi_freecount, -(ilen - 1));
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT);
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(mp, agno);
|
|
|
|
pag->pagi_freecount -= ilen - 1;
|
2018-06-12 11:52:01 +07:00
|
|
|
pag->pagi_count -= ilen;
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen);
|
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1));
|
|
|
|
|
2008-10-30 12:58:01 +07:00
|
|
|
if ((error = xfs_btree_delete(cur, &i))) {
|
2011-03-07 06:08:35 +07:00
|
|
|
xfs_warn(mp, "%s: xfs_btree_delete returned error %d.",
|
|
|
|
__func__, error);
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
|
|
|
}
|
|
|
|
|
2018-08-01 21:20:34 +07:00
|
|
|
xfs_difree_inode_chunk(tp, agno, &rec);
|
2005-04-17 05:20:36 +07:00
|
|
|
} else {
|
2017-10-10 01:38:54 +07:00
|
|
|
xic->deleted = false;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2009-09-01 06:57:03 +07:00
|
|
|
error = xfs_inobt_update(cur, &rec);
|
|
|
|
if (error) {
|
2011-03-07 06:08:35 +07:00
|
|
|
xfs_warn(mp, "%s: xfs_inobt_update returned error %d.",
|
|
|
|
__func__, error);
|
2005-04-17 05:20:36 +07:00
|
|
|
goto error0;
|
|
|
|
}
|
2009-09-01 06:57:03 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Change the inode free counts and log the ag/sb changes.
|
|
|
|
*/
|
2008-02-14 06:03:29 +07:00
|
|
|
be32_add_cpu(&agi->agi_freecount, 1);
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(mp, agno);
|
|
|
|
pag->pagi_freecount++;
|
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
|
|
|
|
}
|
|
|
|
|
2009-09-01 06:57:09 +07:00
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
*orec = rec;
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
error0:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
/*
|
|
|
|
* Free an inode in the free inode btree.
|
|
|
|
*/
|
|
|
|
STATIC int
|
|
|
|
xfs_difree_finobt(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
struct xfs_buf *agbp,
|
|
|
|
xfs_agino_t agino,
|
|
|
|
struct xfs_inobt_rec_incore *ibtrec) /* inobt record */
|
|
|
|
{
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp);
|
|
|
|
xfs_agnumber_t agno = be32_to_cpu(agi->agi_seqno);
|
|
|
|
struct xfs_btree_cur *cur;
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
int offset = agino - ibtrec->ir_startino;
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_FINO);
|
|
|
|
|
|
|
|
error = xfs_inobt_lookup(cur, ibtrec->ir_startino, XFS_LOOKUP_EQ, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
if (i == 0) {
|
|
|
|
/*
|
|
|
|
* If the record does not exist in the finobt, we must have just
|
|
|
|
* freed an inode in a previously fully allocated chunk. If not,
|
|
|
|
* something is out of sync.
|
|
|
|
*/
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, ibtrec->ir_freecount == 1, error);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
2015-05-29 06:03:04 +07:00
|
|
|
error = xfs_inobt_insert_rec(cur, ibtrec->ir_holemask,
|
|
|
|
ibtrec->ir_count,
|
|
|
|
ibtrec->ir_freecount,
|
2014-04-24 13:00:53 +07:00
|
|
|
ibtrec->ir_free, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
ASSERT(i == 1);
|
|
|
|
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Read and update the existing record. We could just copy the ibtrec
|
|
|
|
* across here, but that would defeat the purpose of having redundant
|
|
|
|
* metadata. By making the modifications independently, we can catch
|
|
|
|
* corruptions that we wouldn't see if we just copied from one record
|
|
|
|
* to another.
|
|
|
|
*/
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, error);
|
2014-04-24 13:00:53 +07:00
|
|
|
|
|
|
|
rec.ir_free |= XFS_INOBT_MASK(offset);
|
|
|
|
rec.ir_freecount++;
|
|
|
|
|
2015-02-23 18:39:08 +07:00
|
|
|
XFS_WANT_CORRUPTED_GOTO(mp, (rec.ir_free == ibtrec->ir_free) &&
|
2014-04-24 13:00:53 +07:00
|
|
|
(rec.ir_freecount == ibtrec->ir_freecount),
|
|
|
|
error);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The content of inobt records should always match between the inobt
|
|
|
|
* and finobt. The lifecycle of records in the finobt is different from
|
|
|
|
* the inobt in that the finobt only tracks records with at least one
|
|
|
|
* free inode. Hence, if all of the inodes are free and we aren't
|
|
|
|
* keeping inode chunks permanently on disk, remove the record.
|
|
|
|
* Otherwise, update the record with the new information.
|
2015-05-29 05:51:37 +07:00
|
|
|
*
|
|
|
|
* Note that we currently can't free chunks when the block size is large
|
|
|
|
* enough for multiple chunks. Leave the finobt record to remain in sync
|
|
|
|
* with the inobt.
|
2014-04-24 13:00:53 +07:00
|
|
|
*/
|
2015-05-29 05:51:37 +07:00
|
|
|
if (rec.ir_free == XFS_INOBT_ALL_FREE &&
|
|
|
|
mp->m_sb.sb_inopblock <= XFS_INODES_PER_CHUNK &&
|
2014-04-24 13:00:53 +07:00
|
|
|
!(mp->m_flags & XFS_MOUNT_IKEEP)) {
|
|
|
|
error = xfs_btree_delete(cur, &i);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
ASSERT(i == 1);
|
|
|
|
} else {
|
|
|
|
error = xfs_inobt_update(cur, &rec);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
error = xfs_check_agi_freecount(cur, agi);
|
|
|
|
if (error)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
error:
|
|
|
|
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
/*
|
|
|
|
* Free disk inode. Carefully avoids touching the incore inode, all
|
|
|
|
* manipulations incore are the caller's responsibility.
|
|
|
|
* The on-disk inode is not changed by this operation, only the
|
|
|
|
* btree (free inode mask) is changed.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_difree(
|
|
|
|
struct xfs_trans *tp, /* transaction pointer */
|
|
|
|
xfs_ino_t inode, /* inode to be freed */
|
2015-05-29 06:26:03 +07:00
|
|
|
struct xfs_icluster *xic) /* cluster info if deleted */
|
2014-04-24 13:00:53 +07:00
|
|
|
{
|
|
|
|
/* REFERENCED */
|
|
|
|
xfs_agblock_t agbno; /* block number containing inode */
|
|
|
|
struct xfs_buf *agbp; /* buffer for allocation group header */
|
|
|
|
xfs_agino_t agino; /* allocation group inode number */
|
|
|
|
xfs_agnumber_t agno; /* allocation group number */
|
|
|
|
int error; /* error return value */
|
|
|
|
struct xfs_mount *mp; /* mount structure for filesystem */
|
|
|
|
struct xfs_inobt_rec_incore rec;/* btree record */
|
|
|
|
|
|
|
|
mp = tp->t_mountp;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Break up inode number into its components.
|
|
|
|
*/
|
|
|
|
agno = XFS_INO_TO_AGNO(mp, inode);
|
|
|
|
if (agno >= mp->m_sb.sb_agcount) {
|
|
|
|
xfs_warn(mp, "%s: agno >= mp->m_sb.sb_agcount (%d >= %d).",
|
|
|
|
__func__, agno, mp->m_sb.sb_agcount);
|
|
|
|
ASSERT(0);
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2014-04-24 13:00:53 +07:00
|
|
|
}
|
|
|
|
agino = XFS_INO_TO_AGINO(mp, inode);
|
|
|
|
if (inode != XFS_AGINO_TO_INO(mp, agno, agino)) {
|
|
|
|
xfs_warn(mp, "%s: inode != XFS_AGINO_TO_INO() (%llu != %llu).",
|
|
|
|
__func__, (unsigned long long)inode,
|
|
|
|
(unsigned long long)XFS_AGINO_TO_INO(mp, agno, agino));
|
|
|
|
ASSERT(0);
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2014-04-24 13:00:53 +07:00
|
|
|
}
|
|
|
|
agbno = XFS_AGINO_TO_AGBNO(mp, agino);
|
|
|
|
if (agbno >= mp->m_sb.sb_agblocks) {
|
|
|
|
xfs_warn(mp, "%s: agbno >= mp->m_sb.sb_agblocks (%d >= %d).",
|
|
|
|
__func__, agbno, mp->m_sb.sb_agblocks);
|
|
|
|
ASSERT(0);
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2014-04-24 13:00:53 +07:00
|
|
|
}
|
|
|
|
/*
|
|
|
|
* Get the allocation group header.
|
|
|
|
*/
|
|
|
|
error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
|
|
|
|
if (error) {
|
|
|
|
xfs_warn(mp, "%s: xfs_ialloc_read_agi() returned error %d.",
|
|
|
|
__func__, error);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Fix up the inode allocation btree.
|
|
|
|
*/
|
2018-07-12 12:26:07 +07:00
|
|
|
error = xfs_difree_inobt(mp, tp, agbp, agino, xic, &rec);
|
2014-04-24 13:00:53 +07:00
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
/*
|
|
|
|
* Fix up the free inode btree.
|
|
|
|
*/
|
|
|
|
if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
|
|
|
|
error = xfs_difree_finobt(mp, tp, agbp, agino, &rec);
|
|
|
|
if (error)
|
|
|
|
goto error0;
|
|
|
|
}
|
|
|
|
|
2014-04-24 13:00:53 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
error0:
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2010-06-24 08:15:33 +07:00
|
|
|
STATIC int
|
|
|
|
xfs_imap_lookup(
|
|
|
|
struct xfs_mount *mp,
|
|
|
|
struct xfs_trans *tp,
|
|
|
|
xfs_agnumber_t agno,
|
|
|
|
xfs_agino_t agino,
|
|
|
|
xfs_agblock_t agbno,
|
|
|
|
xfs_agblock_t *chunk_agbno,
|
|
|
|
xfs_agblock_t *offset_agbno,
|
|
|
|
int flags)
|
|
|
|
{
|
|
|
|
struct xfs_inobt_rec_incore rec;
|
|
|
|
struct xfs_btree_cur *cur;
|
|
|
|
struct xfs_buf *agbp;
|
|
|
|
int error;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
|
|
|
|
if (error) {
|
2011-03-07 06:05:35 +07:00
|
|
|
xfs_alert(mp,
|
|
|
|
"%s: xfs_ialloc_read_agi() returned error %d, agno %d",
|
|
|
|
__func__, error, agno);
|
2010-06-24 08:15:33 +07:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2010-08-24 08:42:30 +07:00
|
|
|
* Lookup the inode record for the given agino. If the record cannot be
|
|
|
|
* found, then it's an invalid inode number and we should abort. Once
|
|
|
|
* we have a record, we need to ensure it contains the inode number
|
|
|
|
* we are looking up.
|
2010-06-24 08:15:33 +07:00
|
|
|
*/
|
2014-04-24 13:00:50 +07:00
|
|
|
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
|
2010-08-24 08:42:30 +07:00
|
|
|
error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
|
2010-06-24 08:15:33 +07:00
|
|
|
if (!error) {
|
|
|
|
if (i)
|
|
|
|
error = xfs_inobt_get_rec(cur, &rec, &i);
|
|
|
|
if (!error && i == 0)
|
2014-06-25 11:58:08 +07:00
|
|
|
error = -EINVAL;
|
2010-06-24 08:15:33 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
xfs_trans_brelse(tp, agbp);
|
2018-07-20 02:26:31 +07:00
|
|
|
xfs_btree_del_cursor(cur, error);
|
2010-06-24 08:15:33 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
2010-08-24 08:42:30 +07:00
|
|
|
/* check that the returned record contains the required inode */
|
|
|
|
if (rec.ir_startino > agino ||
|
2013-12-13 11:51:46 +07:00
|
|
|
rec.ir_startino + mp->m_ialloc_inos <= agino)
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2010-08-24 08:42:30 +07:00
|
|
|
|
2010-06-24 08:15:33 +07:00
|
|
|
/* for untrusted inodes check it is allocated first */
|
2010-06-24 08:15:47 +07:00
|
|
|
if ((flags & XFS_IGET_UNTRUSTED) &&
|
2010-06-24 08:15:33 +07:00
|
|
|
(rec.ir_free & XFS_INOBT_MASK(agino - rec.ir_startino)))
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2010-06-24 08:15:33 +07:00
|
|
|
|
|
|
|
*chunk_agbno = XFS_AGINO_TO_AGBNO(mp, rec.ir_startino);
|
|
|
|
*offset_agbno = agbno - *chunk_agbno;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2008-11-28 10:23:41 +07:00
|
|
|
* Return the location of the inode in imap, for mapping it into a buffer.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
int
|
2008-11-28 10:23:41 +07:00
|
|
|
xfs_imap(
|
|
|
|
xfs_mount_t *mp, /* file system mount structure */
|
|
|
|
xfs_trans_t *tp, /* transaction pointer */
|
2005-04-17 05:20:36 +07:00
|
|
|
xfs_ino_t ino, /* inode to locate */
|
2008-11-28 10:23:41 +07:00
|
|
|
struct xfs_imap *imap, /* location map structure */
|
|
|
|
uint flags) /* flags for inode btree lookup */
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
xfs_agblock_t agbno; /* block number of inode in the alloc group */
|
|
|
|
xfs_agino_t agino; /* inode number within alloc group */
|
|
|
|
xfs_agnumber_t agno; /* allocation group number */
|
|
|
|
xfs_agblock_t chunk_agbno; /* first block in inode chunk */
|
|
|
|
xfs_agblock_t cluster_agbno; /* first block in inode cluster */
|
|
|
|
int error; /* error code */
|
|
|
|
int offset; /* index of inode in its buffer */
|
2013-08-12 17:49:44 +07:00
|
|
|
xfs_agblock_t offset_agbno; /* blks from chunk start to inode */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
ASSERT(ino != NULLFSINO);
|
2008-11-28 10:23:41 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Split up the inode number into its parts.
|
|
|
|
*/
|
|
|
|
agno = XFS_INO_TO_AGNO(mp, ino);
|
|
|
|
agino = XFS_INO_TO_AGINO(mp, ino);
|
|
|
|
agbno = XFS_AGINO_TO_AGBNO(mp, agino);
|
|
|
|
if (agno >= mp->m_sb.sb_agcount || agbno >= mp->m_sb.sb_agblocks ||
|
|
|
|
ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
|
|
|
|
#ifdef DEBUG
|
2010-06-24 08:15:47 +07:00
|
|
|
/*
|
|
|
|
* Don't output diagnostic information for untrusted inodes
|
|
|
|
* as they can be invalid without implying corruption.
|
|
|
|
*/
|
|
|
|
if (flags & XFS_IGET_UNTRUSTED)
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2005-04-17 05:20:36 +07:00
|
|
|
if (agno >= mp->m_sb.sb_agcount) {
|
2011-03-07 06:05:35 +07:00
|
|
|
xfs_alert(mp,
|
|
|
|
"%s: agno (%d) >= mp->m_sb.sb_agcount (%d)",
|
|
|
|
__func__, agno, mp->m_sb.sb_agcount);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
if (agbno >= mp->m_sb.sb_agblocks) {
|
2011-03-07 06:05:35 +07:00
|
|
|
xfs_alert(mp,
|
|
|
|
"%s: agbno (0x%llx) >= mp->m_sb.sb_agblocks (0x%lx)",
|
|
|
|
__func__, (unsigned long long)agbno,
|
|
|
|
(unsigned long)mp->m_sb.sb_agblocks);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
if (ino != XFS_AGINO_TO_INO(mp, agno, agino)) {
|
2011-03-07 06:05:35 +07:00
|
|
|
xfs_alert(mp,
|
|
|
|
"%s: ino (0x%llx) != XFS_AGINO_TO_INO() (0x%llx)",
|
|
|
|
__func__, ino,
|
|
|
|
XFS_AGINO_TO_INO(mp, agno, agino));
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2006-09-28 08:02:23 +07:00
|
|
|
xfs_stack_trace();
|
2005-04-17 05:20:36 +07:00
|
|
|
#endif /* DEBUG */
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2008-11-28 10:23:41 +07:00
|
|
|
|
2010-06-24 08:15:33 +07:00
|
|
|
/*
|
|
|
|
* For bulkstat and handle lookups, we have an untrusted inode number
|
|
|
|
* that we have to verify is valid. We cannot do this just by reading
|
|
|
|
* the inode buffer as it may have been unlinked and removed leaving
|
|
|
|
* inodes in stale state on disk. Hence we have to do a btree lookup
|
|
|
|
* in all cases where an untrusted inode number is passed.
|
|
|
|
*/
|
2010-06-24 08:15:47 +07:00
|
|
|
if (flags & XFS_IGET_UNTRUSTED) {
|
2010-06-24 08:15:33 +07:00
|
|
|
error = xfs_imap_lookup(mp, tp, agno, agino, agbno,
|
|
|
|
&chunk_agbno, &offset_agbno, flags);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
goto out_map;
|
|
|
|
}
|
|
|
|
|
2008-11-28 10:23:41 +07:00
|
|
|
/*
|
|
|
|
* If the inode cluster size is the same as the blocksize or
|
|
|
|
* smaller we get to the buffer by simple arithmetics.
|
|
|
|
*/
|
2018-12-12 23:46:25 +07:00
|
|
|
if (mp->m_blocks_per_cluster == 1) {
|
2005-04-17 05:20:36 +07:00
|
|
|
offset = XFS_INO_TO_OFFSET(mp, ino);
|
|
|
|
ASSERT(offset < mp->m_sb.sb_inopblock);
|
2008-11-28 10:23:41 +07:00
|
|
|
|
|
|
|
imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, agbno);
|
|
|
|
imap->im_len = XFS_FSB_TO_BB(mp, 1);
|
2016-11-08 07:55:48 +07:00
|
|
|
imap->im_boffset = (unsigned short)(offset <<
|
|
|
|
mp->m_sb.sb_inodelog);
|
2005-04-17 05:20:36 +07:00
|
|
|
return 0;
|
|
|
|
}
|
2008-11-28 10:23:41 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the inode chunks are aligned then use simple maths to
|
|
|
|
* find the location. Otherwise we have to do a btree
|
|
|
|
* lookup to find the location.
|
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
if (mp->m_inoalign_mask) {
|
|
|
|
offset_agbno = agbno & mp->m_inoalign_mask;
|
|
|
|
chunk_agbno = agbno - offset_agbno;
|
|
|
|
} else {
|
2010-06-24 08:15:33 +07:00
|
|
|
error = xfs_imap_lookup(mp, tp, agno, agino, agbno,
|
|
|
|
&chunk_agbno, &offset_agbno, flags);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
}
|
2008-11-28 10:23:41 +07:00
|
|
|
|
2010-06-24 08:15:33 +07:00
|
|
|
out_map:
|
2005-04-17 05:20:36 +07:00
|
|
|
ASSERT(agbno >= chunk_agbno);
|
|
|
|
cluster_agbno = chunk_agbno +
|
2018-12-12 23:46:25 +07:00
|
|
|
((offset_agbno / mp->m_blocks_per_cluster) *
|
|
|
|
mp->m_blocks_per_cluster);
|
2005-04-17 05:20:36 +07:00
|
|
|
offset = ((agbno - cluster_agbno) * mp->m_sb.sb_inopblock) +
|
|
|
|
XFS_INO_TO_OFFSET(mp, ino);
|
2008-11-28 10:23:41 +07:00
|
|
|
|
|
|
|
imap->im_blkno = XFS_AGB_TO_DADDR(mp, agno, cluster_agbno);
|
2018-12-12 23:46:25 +07:00
|
|
|
imap->im_len = XFS_FSB_TO_BB(mp, mp->m_blocks_per_cluster);
|
2016-11-08 07:55:48 +07:00
|
|
|
imap->im_boffset = (unsigned short)(offset << mp->m_sb.sb_inodelog);
|
2008-11-28 10:23:41 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If the inode number maps to a block outside the bounds
|
|
|
|
* of the file system then return NULL rather than calling
|
|
|
|
* read_buf and panicing when we get an error from the
|
|
|
|
* driver.
|
|
|
|
*/
|
|
|
|
if ((imap->im_blkno + imap->im_len) >
|
|
|
|
XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) {
|
2011-03-07 06:05:35 +07:00
|
|
|
xfs_alert(mp,
|
|
|
|
"%s: (im_blkno (0x%llx) + im_len (0x%llx)) > sb_dblocks (0x%llx)",
|
|
|
|
__func__, (unsigned long long) imap->im_blkno,
|
2008-11-28 10:23:41 +07:00
|
|
|
(unsigned long long) imap->im_len,
|
|
|
|
XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks));
|
2014-06-25 11:58:08 +07:00
|
|
|
return -EINVAL;
|
2008-11-28 10:23:41 +07:00
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Compute and fill in value of m_in_maxlevels.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_ialloc_compute_maxlevels(
|
|
|
|
xfs_mount_t *mp) /* file system mount structure */
|
|
|
|
{
|
2016-06-21 08:53:28 +07:00
|
|
|
uint inodes;
|
|
|
|
|
|
|
|
inodes = (1LL << XFS_INO_AGINO_BITS(mp)) >> XFS_INODES_PER_CHUNK_LOG;
|
2018-04-07 00:09:42 +07:00
|
|
|
mp->m_in_maxlevels = xfs_btree_compute_maxlevels(mp->m_inobt_mnr,
|
2016-06-21 08:53:28 +07:00
|
|
|
inodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2014-04-24 13:00:52 +07:00
|
|
|
* Log specified fields for the ag hdr (inode section). The growth of the agi
|
|
|
|
* structure over time requires that we interpret the buffer as two logical
|
|
|
|
* regions delineated by the end of the unlinked list. This is due to the size
|
|
|
|
* of the hash table and its location in the middle of the agi.
|
|
|
|
*
|
|
|
|
* For example, a request to log a field before agi_unlinked and a field after
|
|
|
|
* agi_unlinked could cause us to log the entire hash table and use an excessive
|
|
|
|
* amount of log space. To avoid this behavior, log the region up through
|
|
|
|
* agi_unlinked in one call and the region after agi_unlinked through the end of
|
|
|
|
* the structure in another.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
void
|
|
|
|
xfs_ialloc_log_agi(
|
|
|
|
xfs_trans_t *tp, /* transaction pointer */
|
|
|
|
xfs_buf_t *bp, /* allocation group header buffer */
|
|
|
|
int fields) /* bitmask of fields to log */
|
|
|
|
{
|
|
|
|
int first; /* first byte number */
|
|
|
|
int last; /* last byte number */
|
|
|
|
static const short offsets[] = { /* field starting offsets */
|
|
|
|
/* keep in sync with bit definitions */
|
|
|
|
offsetof(xfs_agi_t, agi_magicnum),
|
|
|
|
offsetof(xfs_agi_t, agi_versionnum),
|
|
|
|
offsetof(xfs_agi_t, agi_seqno),
|
|
|
|
offsetof(xfs_agi_t, agi_length),
|
|
|
|
offsetof(xfs_agi_t, agi_count),
|
|
|
|
offsetof(xfs_agi_t, agi_root),
|
|
|
|
offsetof(xfs_agi_t, agi_level),
|
|
|
|
offsetof(xfs_agi_t, agi_freecount),
|
|
|
|
offsetof(xfs_agi_t, agi_newino),
|
|
|
|
offsetof(xfs_agi_t, agi_dirino),
|
|
|
|
offsetof(xfs_agi_t, agi_unlinked),
|
2014-04-24 13:00:52 +07:00
|
|
|
offsetof(xfs_agi_t, agi_free_root),
|
|
|
|
offsetof(xfs_agi_t, agi_free_level),
|
2005-04-17 05:20:36 +07:00
|
|
|
sizeof(xfs_agi_t)
|
|
|
|
};
|
|
|
|
#ifdef DEBUG
|
|
|
|
xfs_agi_t *agi; /* allocation group header */
|
|
|
|
|
|
|
|
agi = XFS_BUF_TO_AGI(bp);
|
2011-07-08 19:36:05 +07:00
|
|
|
ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
|
2005-04-17 05:20:36 +07:00
|
|
|
#endif
|
2014-04-24 13:00:52 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2014-04-24 13:00:52 +07:00
|
|
|
* Compute byte offsets for the first and last fields in the first
|
|
|
|
* region and log the agi buffer. This only logs up through
|
|
|
|
* agi_unlinked.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2014-04-24 13:00:52 +07:00
|
|
|
if (fields & XFS_AGI_ALL_BITS_R1) {
|
|
|
|
xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS_R1,
|
|
|
|
&first, &last);
|
|
|
|
xfs_trans_log_buf(tp, bp, first, last);
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
2014-04-24 13:00:52 +07:00
|
|
|
* Mask off the bits in the first region and calculate the first and
|
|
|
|
* last field offsets for any bits in the second region.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2014-04-24 13:00:52 +07:00
|
|
|
fields &= ~XFS_AGI_ALL_BITS_R1;
|
|
|
|
if (fields) {
|
|
|
|
xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS_R2,
|
|
|
|
&first, &last);
|
|
|
|
xfs_trans_log_buf(tp, bp, first, last);
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2018-01-09 01:51:03 +07:00
|
|
|
static xfs_failaddr_t
|
2012-11-14 13:52:32 +07:00
|
|
|
xfs_agi_verify(
|
2012-11-12 18:54:05 +07:00
|
|
|
struct xfs_buf *bp)
|
|
|
|
{
|
|
|
|
struct xfs_mount *mp = bp->b_target->bt_mount;
|
|
|
|
struct xfs_agi *agi = XFS_BUF_TO_AGI(bp);
|
2018-06-04 06:12:16 +07:00
|
|
|
int i;
|
2012-11-12 18:54:05 +07:00
|
|
|
|
2015-10-12 11:59:25 +07:00
|
|
|
if (xfs_sb_version_hascrc(&mp->m_sb)) {
|
|
|
|
if (!uuid_equal(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid))
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2015-10-12 11:59:25 +07:00
|
|
|
if (!xfs_log_check_lsn(mp,
|
|
|
|
be64_to_cpu(XFS_BUF_TO_AGI(bp)->agi_lsn)))
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2015-10-12 11:59:25 +07:00
|
|
|
}
|
|
|
|
|
2012-11-12 18:54:05 +07:00
|
|
|
/*
|
|
|
|
* Validate the magic number of the agi block.
|
|
|
|
*/
|
2013-04-03 12:11:15 +07:00
|
|
|
if (agi->agi_magicnum != cpu_to_be32(XFS_AGI_MAGIC))
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2013-04-03 12:11:15 +07:00
|
|
|
if (!XFS_AGI_GOOD_VERSION(be32_to_cpu(agi->agi_versionnum)))
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2012-11-12 18:54:05 +07:00
|
|
|
|
2016-12-05 08:32:50 +07:00
|
|
|
if (be32_to_cpu(agi->agi_level) < 1 ||
|
|
|
|
be32_to_cpu(agi->agi_level) > XFS_BTREE_MAXLEVELS)
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2016-12-05 08:32:50 +07:00
|
|
|
|
|
|
|
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))
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2016-12-05 08:32:50 +07:00
|
|
|
|
2012-11-12 18:54:05 +07:00
|
|
|
/*
|
|
|
|
* during growfs operations, the perag is not fully initialised,
|
|
|
|
* so we can't use it for any useful checking. growfs ensures we can't
|
|
|
|
* use it by using uncached buffers that don't have the perag attached
|
|
|
|
* so we can detect and avoid this problem.
|
|
|
|
*/
|
2013-04-03 12:11:15 +07:00
|
|
|
if (bp->b_pag && be32_to_cpu(agi->agi_seqno) != bp->b_pag->pag_agno)
|
2018-01-09 01:51:03 +07:00
|
|
|
return __this_address;
|
2012-11-12 18:54:05 +07:00
|
|
|
|
2018-06-04 06:12:16 +07:00
|
|
|
for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
|
2018-07-18 04:25:20 +07:00
|
|
|
if (agi->agi_unlinked[i] == cpu_to_be32(NULLAGINO))
|
2018-06-04 06:12:16 +07:00
|
|
|
continue;
|
|
|
|
if (!xfs_verify_ino(mp, be32_to_cpu(agi->agi_unlinked[i])))
|
|
|
|
return __this_address;
|
|
|
|
}
|
|
|
|
|
2018-01-09 01:51:03 +07:00
|
|
|
return NULL;
|
2012-11-14 13:52:32 +07:00
|
|
|
}
|
|
|
|
|
2012-11-14 13:54:40 +07:00
|
|
|
static void
|
|
|
|
xfs_agi_read_verify(
|
2012-11-14 13:52:32 +07:00
|
|
|
struct xfs_buf *bp)
|
|
|
|
{
|
2013-04-03 12:11:15 +07:00
|
|
|
struct xfs_mount *mp = bp->b_target->bt_mount;
|
2018-01-09 01:51:03 +07:00
|
|
|
xfs_failaddr_t fa;
|
2013-04-03 12:11:15 +07:00
|
|
|
|
2014-02-27 11:23:10 +07:00
|
|
|
if (xfs_sb_version_hascrc(&mp->m_sb) &&
|
|
|
|
!xfs_buf_verify_cksum(bp, XFS_AGI_CRC_OFF))
|
2018-01-09 01:51:03 +07:00
|
|
|
xfs_verifier_error(bp, -EFSBADCRC, __this_address);
|
|
|
|
else {
|
|
|
|
fa = xfs_agi_verify(bp);
|
|
|
|
if (XFS_TEST_ERROR(fa, mp, XFS_ERRTAG_IALLOC_READ_AGI))
|
|
|
|
xfs_verifier_error(bp, -EFSCORRUPTED, fa);
|
|
|
|
}
|
2012-11-14 13:52:32 +07:00
|
|
|
}
|
|
|
|
|
2012-11-14 13:53:49 +07:00
|
|
|
static void
|
2012-11-14 13:54:40 +07:00
|
|
|
xfs_agi_write_verify(
|
2012-11-14 13:52:32 +07:00
|
|
|
struct xfs_buf *bp)
|
|
|
|
{
|
2018-01-25 04:38:48 +07:00
|
|
|
struct xfs_mount *mp = bp->b_target->bt_mount;
|
|
|
|
struct xfs_buf_log_item *bip = bp->b_log_item;
|
2018-01-09 01:51:03 +07:00
|
|
|
xfs_failaddr_t fa;
|
2013-04-03 12:11:15 +07:00
|
|
|
|
2018-01-09 01:51:03 +07:00
|
|
|
fa = xfs_agi_verify(bp);
|
|
|
|
if (fa) {
|
|
|
|
xfs_verifier_error(bp, -EFSCORRUPTED, fa);
|
2013-04-03 12:11:15 +07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!xfs_sb_version_hascrc(&mp->m_sb))
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (bip)
|
|
|
|
XFS_BUF_TO_AGI(bp)->agi_lsn = cpu_to_be64(bip->bli_item.li_lsn);
|
2014-02-27 11:18:23 +07:00
|
|
|
xfs_buf_update_cksum(bp, XFS_AGI_CRC_OFF);
|
2012-11-12 18:54:05 +07:00
|
|
|
}
|
|
|
|
|
2012-11-14 13:54:40 +07:00
|
|
|
const struct xfs_buf_ops xfs_agi_buf_ops = {
|
2016-01-04 12:10:19 +07:00
|
|
|
.name = "xfs_agi",
|
2012-11-14 13:54:40 +07:00
|
|
|
.verify_read = xfs_agi_read_verify,
|
|
|
|
.verify_write = xfs_agi_write_verify,
|
2018-01-09 01:51:08 +07:00
|
|
|
.verify_struct = xfs_agi_verify,
|
2012-11-14 13:54:40 +07:00
|
|
|
};
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Read in the allocation group header (inode allocation section)
|
|
|
|
*/
|
|
|
|
int
|
2008-11-28 10:23:37 +07:00
|
|
|
xfs_read_agi(
|
|
|
|
struct xfs_mount *mp, /* file system mount structure */
|
|
|
|
struct xfs_trans *tp, /* transaction pointer */
|
|
|
|
xfs_agnumber_t agno, /* allocation group number */
|
|
|
|
struct xfs_buf **bpp) /* allocation group hdr buf */
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2008-11-28 10:23:37 +07:00
|
|
|
int error;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2013-11-01 11:27:19 +07:00
|
|
|
trace_xfs_read_agi(mp, agno);
|
2008-11-28 10:23:37 +07:00
|
|
|
|
2013-11-01 11:27:19 +07:00
|
|
|
ASSERT(agno != NULLAGNUMBER);
|
2008-11-28 10:23:37 +07:00
|
|
|
error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp,
|
2005-04-17 05:20:36 +07:00
|
|
|
XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)),
|
2012-11-14 13:54:40 +07:00
|
|
|
XFS_FSS_TO_BB(mp, 1), 0, bpp, &xfs_agi_buf_ops);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (error)
|
|
|
|
return error;
|
2016-12-05 08:31:31 +07:00
|
|
|
if (tp)
|
|
|
|
xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_AGI_BUF);
|
2008-11-28 10:23:37 +07:00
|
|
|
|
2011-10-10 23:52:45 +07:00
|
|
|
xfs_buf_set_ref(*bpp, XFS_AGI_REF);
|
2008-11-28 10:23:37 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
xfs_ialloc_read_agi(
|
|
|
|
struct xfs_mount *mp, /* file system mount structure */
|
|
|
|
struct xfs_trans *tp, /* transaction pointer */
|
|
|
|
xfs_agnumber_t agno, /* allocation group number */
|
|
|
|
struct xfs_buf **bpp) /* allocation group hdr buf */
|
|
|
|
{
|
|
|
|
struct xfs_agi *agi; /* allocation group header */
|
|
|
|
struct xfs_perag *pag; /* per allocation group data */
|
|
|
|
int error;
|
|
|
|
|
2013-11-01 11:27:19 +07:00
|
|
|
trace_xfs_ialloc_read_agi(mp, agno);
|
|
|
|
|
2008-11-28 10:23:37 +07:00
|
|
|
error = xfs_read_agi(mp, tp, agno, bpp);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
agi = XFS_BUF_TO_AGI(*bpp);
|
2010-01-11 18:47:43 +07:00
|
|
|
pag = xfs_perag_get(mp, agno);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (!pag->pagi_init) {
|
2005-11-02 11:11:25 +07:00
|
|
|
pag->pagi_freecount = be32_to_cpu(agi->agi_freecount);
|
[XFS] Lazy Superblock Counters
When we have a couple of hundred transactions on the fly at once, they all
typically modify the on disk superblock in some way.
create/unclink/mkdir/rmdir modify inode counts, allocation/freeing modify
free block counts.
When these counts are modified in a transaction, they must eventually lock
the superblock buffer and apply the mods. The buffer then remains locked
until the transaction is committed into the incore log buffer. The result
of this is that with enough transactions on the fly the incore superblock
buffer becomes a bottleneck.
The result of contention on the incore superblock buffer is that
transaction rates fall - the more pressure that is put on the superblock
buffer, the slower things go.
The key to removing the contention is to not require the superblock fields
in question to be locked. We do that by not marking the superblock dirty
in the transaction. IOWs, we modify the incore superblock but do not
modify the cached superblock buffer. In short, we do not log superblock
modifications to critical fields in the superblock on every transaction.
In fact we only do it just before we write the superblock to disk every
sync period or just before unmount.
This creates an interesting problem - if we don't log or write out the
fields in every transaction, then how do the values get recovered after a
crash? the answer is simple - we keep enough duplicate, logged information
in other structures that we can reconstruct the correct count after log
recovery has been performed.
It is the AGF and AGI structures that contain the duplicate information;
after recovery, we walk every AGI and AGF and sum their individual
counters to get the correct value, and we do a transaction into the log to
correct them. An optimisation of this is that if we have a clean unmount
record, we know the value in the superblock is correct, so we can avoid
the summation walk under normal conditions and so mount/recovery times do
not change under normal operation.
One wrinkle that was discovered during development was that the blocks
used in the freespace btrees are never accounted for in the AGF counters.
This was once a valid optimisation to make; when the filesystem is full,
the free space btrees are empty and consume no space. Hence when it
matters, the "accounting" is correct. But that means the when we do the
AGF summations, we would not have a correct count and xfs_check would
complain. Hence a new counter was added to track the number of blocks used
by the free space btrees. This is an *on-disk format change*.
As a result of this, lazy superblock counters are a mkfs option and at the
moment on linux there is no way to convert an old filesystem. This is
possible - xfs_db can be used to twiddle the right bits and then
xfs_repair will do the format conversion for you. Similarly, you can
convert backwards as well. At some point we'll add functionality to
xfs_admin to do the bit twiddling easily....
SGI-PV: 964999
SGI-Modid: xfs-linux-melb:xfs-kern:28652a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
2007-05-24 12:26:31 +07:00
|
|
|
pag->pagi_count = be32_to_cpu(agi->agi_count);
|
2005-04-17 05:20:36 +07:00
|
|
|
pag->pagi_init = 1;
|
|
|
|
}
|
|
|
|
|
2008-11-28 10:23:37 +07:00
|
|
|
/*
|
|
|
|
* It's possible for these to be out of sync if
|
|
|
|
* we are in the middle of a forced shutdown.
|
|
|
|
*/
|
|
|
|
ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) ||
|
|
|
|
XFS_FORCED_SHUTDOWN(mp));
|
2010-01-11 18:47:43 +07:00
|
|
|
xfs_perag_put(pag);
|
2005-04-17 05:20:36 +07:00
|
|
|
return 0;
|
|
|
|
}
|
[XFS] Lazy Superblock Counters
When we have a couple of hundred transactions on the fly at once, they all
typically modify the on disk superblock in some way.
create/unclink/mkdir/rmdir modify inode counts, allocation/freeing modify
free block counts.
When these counts are modified in a transaction, they must eventually lock
the superblock buffer and apply the mods. The buffer then remains locked
until the transaction is committed into the incore log buffer. The result
of this is that with enough transactions on the fly the incore superblock
buffer becomes a bottleneck.
The result of contention on the incore superblock buffer is that
transaction rates fall - the more pressure that is put on the superblock
buffer, the slower things go.
The key to removing the contention is to not require the superblock fields
in question to be locked. We do that by not marking the superblock dirty
in the transaction. IOWs, we modify the incore superblock but do not
modify the cached superblock buffer. In short, we do not log superblock
modifications to critical fields in the superblock on every transaction.
In fact we only do it just before we write the superblock to disk every
sync period or just before unmount.
This creates an interesting problem - if we don't log or write out the
fields in every transaction, then how do the values get recovered after a
crash? the answer is simple - we keep enough duplicate, logged information
in other structures that we can reconstruct the correct count after log
recovery has been performed.
It is the AGF and AGI structures that contain the duplicate information;
after recovery, we walk every AGI and AGF and sum their individual
counters to get the correct value, and we do a transaction into the log to
correct them. An optimisation of this is that if we have a clean unmount
record, we know the value in the superblock is correct, so we can avoid
the summation walk under normal conditions and so mount/recovery times do
not change under normal operation.
One wrinkle that was discovered during development was that the blocks
used in the freespace btrees are never accounted for in the AGF counters.
This was once a valid optimisation to make; when the filesystem is full,
the free space btrees are empty and consume no space. Hence when it
matters, the "accounting" is correct. But that means the when we do the
AGF summations, we would not have a correct count and xfs_check would
complain. Hence a new counter was added to track the number of blocks used
by the free space btrees. This is an *on-disk format change*.
As a result of this, lazy superblock counters are a mkfs option and at the
moment on linux there is no way to convert an old filesystem. This is
possible - xfs_db can be used to twiddle the right bits and then
xfs_repair will do the format conversion for you. Similarly, you can
convert backwards as well. At some point we'll add functionality to
xfs_admin to do the bit twiddling easily....
SGI-PV: 964999
SGI-Modid: xfs-linux-melb:xfs-kern:28652a
Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Tim Shimmin <tes@sgi.com>
2007-05-24 12:26:31 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Read in the agi to initialise the per-ag data in the mount structure
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
xfs_ialloc_pagi_init(
|
|
|
|
xfs_mount_t *mp, /* file system mount structure */
|
|
|
|
xfs_trans_t *tp, /* transaction pointer */
|
|
|
|
xfs_agnumber_t agno) /* allocation group number */
|
|
|
|
{
|
|
|
|
xfs_buf_t *bp = NULL;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
error = xfs_ialloc_read_agi(mp, tp, agno, &bp);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
if (bp)
|
|
|
|
xfs_trans_brelse(tp, bp);
|
|
|
|
return 0;
|
|
|
|
}
|
2017-10-18 11:37:34 +07:00
|
|
|
|
2018-01-17 09:52:12 +07:00
|
|
|
/* Is there an inode record covering a given range of inode numbers? */
|
|
|
|
int
|
|
|
|
xfs_ialloc_has_inode_record(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
xfs_agino_t low,
|
|
|
|
xfs_agino_t high,
|
|
|
|
bool *exists)
|
|
|
|
{
|
|
|
|
struct xfs_inobt_rec_incore irec;
|
|
|
|
xfs_agino_t agino;
|
|
|
|
uint16_t holemask;
|
|
|
|
int has_record;
|
|
|
|
int i;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
*exists = false;
|
|
|
|
error = xfs_inobt_lookup(cur, low, XFS_LOOKUP_LE, &has_record);
|
|
|
|
while (error == 0 && has_record) {
|
|
|
|
error = xfs_inobt_get_rec(cur, &irec, &has_record);
|
|
|
|
if (error || irec.ir_startino > high)
|
|
|
|
break;
|
|
|
|
|
|
|
|
agino = irec.ir_startino;
|
|
|
|
holemask = irec.ir_holemask;
|
|
|
|
for (i = 0; i < XFS_INOBT_HOLEMASK_BITS; holemask >>= 1,
|
|
|
|
i++, agino += XFS_INODES_PER_HOLEMASK_BIT) {
|
|
|
|
if (holemask & 1)
|
|
|
|
continue;
|
|
|
|
if (agino + XFS_INODES_PER_HOLEMASK_BIT > low &&
|
|
|
|
agino <= high) {
|
|
|
|
*exists = true;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
error = xfs_btree_increment(cur, 0, &has_record);
|
|
|
|
}
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is there an inode record covering a given extent? */
|
|
|
|
int
|
|
|
|
xfs_ialloc_has_inodes_at_extent(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
xfs_agblock_t bno,
|
|
|
|
xfs_extlen_t len,
|
|
|
|
bool *exists)
|
|
|
|
{
|
|
|
|
xfs_agino_t low;
|
|
|
|
xfs_agino_t high;
|
|
|
|
|
2018-12-12 23:46:24 +07:00
|
|
|
low = XFS_AGB_TO_AGINO(cur->bc_mp, bno);
|
|
|
|
high = XFS_AGB_TO_AGINO(cur->bc_mp, bno + len) - 1;
|
2018-01-17 09:52:12 +07:00
|
|
|
|
|
|
|
return xfs_ialloc_has_inode_record(cur, low, high, exists);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct xfs_ialloc_count_inodes {
|
|
|
|
xfs_agino_t count;
|
|
|
|
xfs_agino_t freecount;
|
|
|
|
};
|
|
|
|
|
|
|
|
/* Record inode counts across all inobt records. */
|
|
|
|
STATIC int
|
|
|
|
xfs_ialloc_count_inodes_rec(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
union xfs_btree_rec *rec,
|
|
|
|
void *priv)
|
|
|
|
{
|
|
|
|
struct xfs_inobt_rec_incore irec;
|
|
|
|
struct xfs_ialloc_count_inodes *ci = priv;
|
|
|
|
|
|
|
|
xfs_inobt_btrec_to_irec(cur->bc_mp, rec, &irec);
|
|
|
|
ci->count += irec.ir_count;
|
|
|
|
ci->freecount += irec.ir_freecount;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Count allocated and free inodes under an inobt. */
|
|
|
|
int
|
|
|
|
xfs_ialloc_count_inodes(
|
|
|
|
struct xfs_btree_cur *cur,
|
|
|
|
xfs_agino_t *count,
|
|
|
|
xfs_agino_t *freecount)
|
|
|
|
{
|
|
|
|
struct xfs_ialloc_count_inodes ci = {0};
|
|
|
|
int error;
|
|
|
|
|
|
|
|
ASSERT(cur->bc_btnum == XFS_BTNUM_INO);
|
|
|
|
error = xfs_btree_query_all(cur, xfs_ialloc_count_inodes_rec, &ci);
|
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
*count = ci.count;
|
|
|
|
*freecount = ci.freecount;
|
|
|
|
return 0;
|
|
|
|
}
|