Commit fdec29c5 authored by Alex Elder's avatar Alex Elder

Merge branch 'master' of git://oss.sgi.com/xfs/xfs into for-linus

Conflicts:
	fs/xfs/linux-2.6/xfs_lrw.c
parents 0cb583fd 9ef96da6
......@@ -5694,7 +5694,7 @@ F: include/xen/
XFS FILESYSTEM
P: Silicon Graphics Inc
M: Felix Blyakher <felixb@sgi.com>
M: Alex Elder <aelder@sgi.com>
M: xfs-masters@oss.sgi.com
L: xfs@oss.sgi.com
W: http://oss.sgi.com/projects/xfs
......
......@@ -216,7 +216,6 @@ xfs_setfilesize(
if (ip->i_d.di_size < isize) {
ip->i_d.di_size = isize;
ip->i_update_core = 1;
ip->i_update_size = 1;
xfs_mark_inode_dirty_sync(ip);
}
......
......@@ -172,12 +172,21 @@ xfs_file_release(
*/
STATIC int
xfs_file_fsync(
struct file *filp,
struct dentry *dentry,
int datasync)
struct file *file,
struct dentry *dentry,
int datasync)
{
xfs_iflags_clear(XFS_I(dentry->d_inode), XFS_ITRUNCATED);
return -xfs_fsync(XFS_I(dentry->d_inode));
struct inode *inode = dentry->d_inode;
struct xfs_inode *ip = XFS_I(inode);
int error;
/* capture size updates in I/O completion before writing the inode. */
error = filemap_fdatawait(inode->i_mapping);
if (error)
return error;
xfs_iflags_clear(ip, XFS_ITRUNCATED);
return -xfs_fsync(ip);
}
STATIC int
......
......@@ -43,7 +43,6 @@
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
......
......@@ -812,19 +812,21 @@ xfs_write(
/* Handle various SYNC-type writes */
if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) {
loff_t end = pos + ret - 1;
int error2;
xfs_iunlock(xip, iolock);
if (need_i_mutex)
mutex_unlock(&inode->i_mutex);
error2 = filemap_write_and_wait_range(mapping, pos,
pos + ret - 1);
error2 = filemap_write_and_wait_range(mapping, pos, end);
if (!error)
error = error2;
if (need_i_mutex)
mutex_lock(&inode->i_mutex);
xfs_ilock(xip, iolock);
error2 = xfs_write_sync_logforce(mp, xip);
error2 = xfs_fsync(xip);
if (!error)
error = error2;
}
......
......@@ -20,16 +20,9 @@
DEFINE_PER_CPU(struct xfsstats, xfsstats);
STATIC int
xfs_read_xfsstats(
char *buffer,
char **start,
off_t offset,
int count,
int *eof,
void *data)
static int xfs_stat_proc_show(struct seq_file *m, void *v)
{
int c, i, j, len, val;
int c, i, j, val;
__uint64_t xs_xstrat_bytes = 0;
__uint64_t xs_write_bytes = 0;
__uint64_t xs_read_bytes = 0;
......@@ -60,18 +53,18 @@ xfs_read_xfsstats(
};
/* Loop over all stats groups */
for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
len += sprintf(buffer + len, "%s", xstats[i].desc);
for (i=j = 0; i < ARRAY_SIZE(xstats); i++) {
seq_printf(m, "%s", xstats[i].desc);
/* inner loop does each group */
while (j < xstats[i].endpoint) {
val = 0;
/* sum over all cpus */
for_each_possible_cpu(c)
val += *(((__u32*)&per_cpu(xfsstats, c) + j));
len += sprintf(buffer + len, " %u", val);
seq_printf(m, " %u", val);
j++;
}
buffer[len++] = '\n';
seq_putc(m, '\n');
}
/* extra precision counters */
for_each_possible_cpu(i) {
......@@ -80,36 +73,38 @@ xfs_read_xfsstats(
xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
}
len += sprintf(buffer + len, "xpc %Lu %Lu %Lu\n",
seq_printf(m, "xpc %Lu %Lu %Lu\n",
xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
len += sprintf(buffer + len, "debug %u\n",
seq_printf(m, "debug %u\n",
#if defined(DEBUG)
1);
#else
0);
#endif
return 0;
}
if (offset >= len) {
*start = buffer;
*eof = 1;
return 0;
}
*start = buffer + offset;
if ((len -= offset) > count)
return count;
*eof = 1;
return len;
static int xfs_stat_proc_open(struct inode *inode, struct file *file)
{
return single_open(file, xfs_stat_proc_show, NULL);
}
static const struct file_operations xfs_stat_proc_fops = {
.owner = THIS_MODULE,
.open = xfs_stat_proc_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
int
xfs_init_procfs(void)
{
if (!proc_mkdir("fs/xfs", NULL))
goto out;
if (!create_proc_read_entry("fs/xfs/stat", 0, NULL,
xfs_read_xfsstats, NULL))
if (!proc_create("fs/xfs/stat", 0, NULL,
&xfs_stat_proc_fops))
goto out_remove_entry;
return 0;
......
......@@ -579,15 +579,19 @@ xfs_showargs(
else if (mp->m_qflags & XFS_UQUOTA_ACCT)
seq_puts(m, "," MNTOPT_UQUOTANOENF);
if (mp->m_qflags & (XFS_PQUOTA_ACCT|XFS_OQUOTA_ENFD))
seq_puts(m, "," MNTOPT_PRJQUOTA);
else if (mp->m_qflags & XFS_PQUOTA_ACCT)
seq_puts(m, "," MNTOPT_PQUOTANOENF);
if (mp->m_qflags & (XFS_GQUOTA_ACCT|XFS_OQUOTA_ENFD))
seq_puts(m, "," MNTOPT_GRPQUOTA);
else if (mp->m_qflags & XFS_GQUOTA_ACCT)
seq_puts(m, "," MNTOPT_GQUOTANOENF);
/* Either project or group quotas can be active, not both */
if (mp->m_qflags & XFS_PQUOTA_ACCT) {
if (mp->m_qflags & XFS_OQUOTA_ENFD)
seq_puts(m, "," MNTOPT_PRJQUOTA);
else
seq_puts(m, "," MNTOPT_PQUOTANOENF);
} else if (mp->m_qflags & XFS_GQUOTA_ACCT) {
if (mp->m_qflags & XFS_OQUOTA_ENFD)
seq_puts(m, "," MNTOPT_GRPQUOTA);
else
seq_puts(m, "," MNTOPT_GQUOTANOENF);
}
if (!(mp->m_qflags & XFS_ALL_QUOTA_ACCT))
seq_puts(m, "," MNTOPT_NOQUOTA);
......@@ -687,7 +691,7 @@ xfs_barrier_test(
return error;
}
void
STATIC void
xfs_mountfs_check_barriers(xfs_mount_t *mp)
{
int error;
......
......@@ -749,21 +749,6 @@ __xfs_inode_clear_reclaim_tag(
XFS_INO_TO_AGINO(mp, ip->i_ino), XFS_ICI_RECLAIM_TAG);
}
void
xfs_inode_clear_reclaim_tag(
xfs_inode_t *ip)
{
xfs_mount_t *mp = ip->i_mount;
xfs_perag_t *pag = xfs_get_perag(mp, ip->i_ino);
read_lock(&pag->pag_ici_lock);
spin_lock(&ip->i_flags_lock);
__xfs_inode_clear_reclaim_tag(mp, pag, ip);
spin_unlock(&ip->i_flags_lock);
read_unlock(&pag->pag_ici_lock);
xfs_put_perag(mp, pag);
}
STATIC int
xfs_reclaim_inode_now(
struct xfs_inode *ip,
......
......@@ -49,7 +49,6 @@ int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
void xfs_inode_set_reclaim_tag(struct xfs_inode *ip);
void __xfs_inode_set_reclaim_tag(struct xfs_perag *pag, struct xfs_inode *ip);
void xfs_inode_clear_reclaim_tag(struct xfs_inode *ip);
void __xfs_inode_clear_reclaim_tag(struct xfs_mount *mp, struct xfs_perag *pag,
struct xfs_inode *ip);
......
......@@ -48,50 +48,34 @@
struct xqmstats xqmstats;
STATIC int
xfs_qm_read_xfsquota(
char *buffer,
char **start,
off_t offset,
int count,
int *eof,
void *data)
static int xqm_proc_show(struct seq_file *m, void *v)
{
int len;
/* maximum; incore; ratio free to inuse; freelist */
len = sprintf(buffer, "%d\t%d\t%d\t%u\n",
seq_printf(m, "%d\t%d\t%d\t%u\n",
ndquot,
xfs_Gqm? atomic_read(&xfs_Gqm->qm_totaldquots) : 0,
xfs_Gqm? xfs_Gqm->qm_dqfree_ratio : 0,
xfs_Gqm? xfs_Gqm->qm_dqfreelist.qh_nelems : 0);
if (offset >= len) {
*start = buffer;
*eof = 1;
return 0;
}
*start = buffer + offset;
if ((len -= offset) > count)
return count;
*eof = 1;
return len;
return 0;
}
STATIC int
xfs_qm_read_stats(
char *buffer,
char **start,
off_t offset,
int count,
int *eof,
void *data)
static int xqm_proc_open(struct inode *inode, struct file *file)
{
int len;
return single_open(file, xqm_proc_show, NULL);
}
static const struct file_operations xqm_proc_fops = {
.owner = THIS_MODULE,
.open = xqm_proc_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
static int xqmstat_proc_show(struct seq_file *m, void *v)
{
/* quota performance statistics */
len = sprintf(buffer, "qm %u %u %u %u %u %u %u %u\n",
seq_printf(m, "qm %u %u %u %u %u %u %u %u\n",
xqmstats.xs_qm_dqreclaims,
xqmstats.xs_qm_dqreclaim_misses,
xqmstats.xs_qm_dquot_dups,
......@@ -100,25 +84,27 @@ xfs_qm_read_stats(
xqmstats.xs_qm_dqwants,
xqmstats.xs_qm_dqshake_reclaims,
xqmstats.xs_qm_dqinact_reclaims);
return 0;
}
if (offset >= len) {
*start = buffer;
*eof = 1;
return 0;
}
*start = buffer + offset;
if ((len -= offset) > count)
return count;
*eof = 1;
return len;
static int xqmstat_proc_open(struct inode *inode, struct file *file)
{
return single_open(file, xqmstat_proc_show, NULL);
}
static const struct file_operations xqmstat_proc_fops = {
.owner = THIS_MODULE,
.open = xqmstat_proc_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
void
xfs_qm_init_procfs(void)
{
create_proc_read_entry("fs/xfs/xqmstat", 0, NULL, xfs_qm_read_stats, NULL);
create_proc_read_entry("fs/xfs/xqm", 0, NULL, xfs_qm_read_xfsquota, NULL);
proc_create("fs/xfs/xqmstat", 0, NULL, &xqmstat_proc_fops);
proc_create("fs/xfs/xqm", 0, NULL, &xqm_proc_fops);
}
void
......
......@@ -198,6 +198,15 @@ typedef struct xfs_perag
xfs_agino_t pagi_count; /* number of allocated inodes */
int pagb_count; /* pagb slots in use */
xfs_perag_busy_t *pagb_list; /* unstable blocks */
/*
* Inode allocation search lookup optimisation.
* If the pagino matches, the search for new inodes
* doesn't need to search the near ones again straight away
*/
xfs_agino_t pagl_pagino;
xfs_agino_t pagl_leftrec;
xfs_agino_t pagl_rightrec;
#ifdef __KERNEL__
spinlock_t pagb_lock; /* lock for pagb_list */
......
......@@ -3713,7 +3713,7 @@ xfs_bmap_local_to_extents(
* entry (null if none). Else, *lastxp will be set to the index
* of the found entry; *gotp will contain the entry.
*/
xfs_bmbt_rec_host_t * /* pointer to found extent entry */
STATIC xfs_bmbt_rec_host_t * /* pointer to found extent entry */
xfs_bmap_search_multi_extents(
xfs_ifork_t *ifp, /* inode fork pointer */
xfs_fileoff_t bno, /* block number searched for */
......
......@@ -392,17 +392,6 @@ xfs_bmap_count_blocks(
int whichfork,
int *count);
/*
* Search the extent records for the entry containing block bno.
* If bno lies in a hole, point to the next entry. If bno lies
* past eof, *eofp will be set, and *prevp will contain the last
* entry (null if none). Else, *lastxp will be set to the index
* of the found entry; *gotp will contain the entry.
*/
xfs_bmbt_rec_host_t *
xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
#endif /* __KERNEL__ */
#endif /* __XFS_BMAP_H__ */
......@@ -202,16 +202,6 @@ xfs_bmbt_get_state(
ext_flag);
}
/* Endian flipping versions of the bmbt extraction functions */
void
xfs_bmbt_disk_get_all(
xfs_bmbt_rec_t *r,
xfs_bmbt_irec_t *s)
{
__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
get_unaligned_be64(&r->l1), s);
}
/*
* Extract the blockcount field from an on disk bmap extent record.
*/
......@@ -816,6 +806,16 @@ xfs_bmbt_trace_key(
*l1 = 0;
}
/* Endian flipping versions of the bmbt extraction functions */
STATIC void
xfs_bmbt_disk_get_all(
xfs_bmbt_rec_t *r,
xfs_bmbt_irec_t *s)
{
__xfs_bmbt_get_all(get_unaligned_be64(&r->l0),
get_unaligned_be64(&r->l1), s);
}
STATIC void
xfs_bmbt_trace_record(
struct xfs_btree_cur *cur,
......
......@@ -220,7 +220,6 @@ extern xfs_fsblock_t xfs_bmbt_get_startblock(xfs_bmbt_rec_host_t *r);
extern xfs_fileoff_t xfs_bmbt_get_startoff(xfs_bmbt_rec_host_t *r);
extern xfs_exntst_t xfs_bmbt_get_state(xfs_bmbt_rec_host_t *r);
extern void xfs_bmbt_disk_get_all(xfs_bmbt_rec_t *r, xfs_bmbt_irec_t *s);
extern xfs_filblks_t xfs_bmbt_disk_get_blockcount(xfs_bmbt_rec_t *r);
extern xfs_fileoff_t xfs_bmbt_disk_get_startoff(xfs_bmbt_rec_t *r);
......
......@@ -645,46 +645,6 @@ xfs_btree_read_bufl(
return 0;
}
/*
* Get a buffer for the block, return it read in.
* Short-form addressing.
*/
int /* error */
xfs_btree_read_bufs(
xfs_mount_t *mp, /* file system mount point */
xfs_trans_t *tp, /* transaction pointer */
xfs_agnumber_t agno, /* allocation group number */
xfs_agblock_t agbno, /* allocation group block number */
uint lock, /* lock flags for read_buf */
xfs_buf_t **bpp, /* buffer for agno/agbno */
int refval) /* ref count value for buffer */
{
xfs_buf_t *bp; /* return value */
xfs_daddr_t d; /* real disk block address */
int error;
ASSERT(agno != NULLAGNUMBER);
ASSERT(agbno != NULLAGBLOCK);
d = XFS_AGB_TO_DADDR(mp, agno, agbno);
if ((error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d,
mp->m_bsize, lock, &bp))) {
return error;
}
ASSERT(!bp || !XFS_BUF_GETERROR(bp));
if (bp != NULL) {
switch (refval) {
case XFS_ALLOC_BTREE_REF:
XFS_BUF_SET_VTYPE_REF(bp, B_FS_MAP, refval);
break;
case XFS_INO_BTREE_REF:
XFS_BUF_SET_VTYPE_REF(bp, B_FS_INOMAP, refval);
break;
}
}
*bpp = bp;
return 0;
}
/*
* Read-ahead the block, don't wait for it, don't return a buffer.
* Long-form addressing.
......@@ -2951,7 +2911,7 @@ xfs_btree_insert(
* inode we have to copy the single block it was pointing to into the
* inode.
*/
int
STATIC int
xfs_btree_kill_iroot(
struct xfs_btree_cur *cur)
{
......
......@@ -378,20 +378,6 @@ xfs_btree_read_bufl(
struct xfs_buf **bpp, /* buffer for fsbno */
int refval);/* ref count value for buffer */
/*
* Get a buffer for the block, return it read in.
* Short-form addressing.
*/
int /* error */
xfs_btree_read_bufs(
struct xfs_mount *mp, /* file system mount point */
struct xfs_trans *tp, /* transaction pointer */
xfs_agnumber_t agno, /* allocation group number */
xfs_agblock_t agbno, /* allocation group block number */
uint lock, /* lock flags for read_buf */
struct xfs_buf **bpp, /* buffer for agno/agbno */
int refval);/* ref count value for buffer */
/*
* Read-ahead the block, don't wait for it, don't return a buffer.
* Long-form addressing.
......@@ -432,7 +418,6 @@ int xfs_btree_decrement(struct xfs_btree_cur *, int, int *);
int xfs_btree_lookup(struct xfs_btree_cur *, xfs_lookup_t, int *);
int xfs_btree_update(struct xfs_btree_cur *, union xfs_btree_rec *);
int xfs_btree_new_iroot(struct xfs_btree_cur *, int *, int *);
int xfs_btree_kill_iroot(struct xfs_btree_cur *);
int xfs_btree_insert(struct xfs_btree_cur *, int *);
int xfs_btree_delete(struct xfs_btree_cur *, int *);
int xfs_btree_get_rec(struct xfs_btree_cur *, union xfs_btree_rec **, int *);
......
......@@ -57,75 +57,35 @@ xfs_ialloc_cluster_alignment(
}
/*
* Lookup the record equal to ino in the btree given by cur.
*/
STATIC int /* error */
xfs_inobt_lookup_eq(
struct xfs_btree_cur *cur, /* btree cursor */
xfs_agino_t ino, /* starting inode of chunk */
__int32_t fcnt, /* free inode count */
xfs_inofree_t free, /* free inode mask */
int *stat) /* success/failure */
{
cur->bc_rec.i.ir_startino = ino;
cur->bc_rec.i.ir_freecount = fcnt;
cur->bc_rec.i.ir_free = free;
return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
}
/*
* Lookup the first record greater than or equal to ino
* in the btree given by cur.
* Lookup a record by ino in the btree given by cur.
*/
int /* error */
xfs_inobt_lookup_ge(
xfs_inobt_lookup(
struct xfs_btree_cur *cur, /* btree cursor */
xfs_agino_t ino, /* starting inode of chunk */
__int32_t fcnt, /* free inode count */
xfs_inofree_t free, /* free inode mask */
xfs_lookup_t dir, /* <=, >=, == */
int *stat) /* success/failure */
{
cur->bc_rec.i.ir_startino = ino;
cur->bc_rec.i.ir_freecount = fcnt;
cur->bc_rec.i.ir_free = free;
return xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
cur->bc_rec.i.ir_freecount = 0;
cur->bc_rec.i.ir_free = 0;
return xfs_btree_lookup(cur, dir, stat);
}
/*
* Lookup the first record less than or equal to ino
* in the btree given by cur.
*/
int /* error */
xfs_inobt_lookup_le(
struct xfs_btree_cur *cur, /* btree cursor */
xfs_agino_t ino, /* starting inode of chunk */
__int32_t fcnt, /* free inode count */
xfs_inofree_t free, /* free inode mask */
int *stat) /* success/failure */
{
cur->bc_rec.i.ir_startino = ino;
cur->bc_rec.i.ir_freecount = fcnt;
cur->bc_rec.i.ir_free = free;
return xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
}
/*
* Update the record referred to by cur to the value given
* by [ino, fcnt, free].
* Update the record referred to by cur to the value given.
* This either works (return 0) or gets an EFSCORRUPTED error.
*/
STATIC int /* error */
xfs_inobt_update(
struct xfs_btree_cur *cur, /* btree cursor */
xfs_agino_t ino, /* starting inode of chunk */
__int32_t fcnt, /* free inode count */
xfs_inofree_t free) /* free inode mask */
xfs_inobt_rec_incore_t *irec) /* btree record */
{
union xfs_btree_rec rec;
rec.inobt.ir_startino = cpu_to_be32(ino);
rec.inobt.ir_freecount = cpu_to_be32(fcnt);
rec.inobt.ir_free = cpu_to_be64(free);
rec.inobt.ir_startino = cpu_to_be32(irec->ir_startino);
rec.inobt.ir_freecount = cpu_to_be32(irec->ir_freecount);
rec.inobt.ir_free = cpu_to_be64(irec->ir_free);
return xfs_btree_update(cur, &rec);
}
......@@ -135,9 +95,7 @@ xfs_inobt_update(
int /* error */
xfs_inobt_get_rec(
struct xfs_btree_cur *cur, /* btree cursor */
xfs_agino_t *ino, /* output: starting inode of chunk */
__int32_t *fcnt, /* output: number of free inodes */
xfs_inofree_t *free, /* output: free inode mask */
xfs_inobt_rec_incore_t *irec, /* btree record */
int *stat) /* output: success/failure */
{
union xfs_btree_rec *rec;
......@@ -145,13 +103,135 @@ xfs_inobt_get_rec(
error = xfs_btree_get_rec(cur, &rec, stat);
if (!error && *stat == 1) {
*ino = be32_to_cpu(rec->inobt.ir_startino);
*fcnt = be32_to_cpu(rec->inobt.ir_freecount);
*free = be64_to_cpu(rec->inobt.ir_free);
irec->ir_startino = be32_to_cpu(rec->inobt.ir_startino);
irec->ir_freecount = be32_to_cpu(rec->inobt.ir_freecount);
irec->ir_free = be64_to_cpu(rec->inobt.ir_free);
}
return error;
}
/*
* 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;
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
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
/*
* Initialise a new set of inodes.
*/
STATIC void
xfs_ialloc_inode_init(
struct xfs_mount *mp,
struct xfs_trans *tp,
xfs_agnumber_t agno,
xfs_agblock_t agbno,
xfs_agblock_t length,
unsigned int gen)
{
struct xfs_buf *fbuf;
struct xfs_dinode *free;
int blks_per_cluster, nbufs, ninodes;
int version;
int i, j;
xfs_daddr_t d;
/*
* 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.
*/
if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
blks_per_cluster = 1;
nbufs = length;
ninodes = mp->m_sb.sb_inopblock;
} else {
blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
mp->m_sb.sb_blocksize;
nbufs = length / blks_per_cluster;
ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
}
/*
* 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.
*/
if (xfs_sb_version_hasnlink(&mp->m_sb))
version = 2;
else
version = 1;
for (j = 0; j < nbufs; j++) {
/*
* Get the block.
*/
d = XFS_AGB_TO_DADDR(mp, agno, agbno + (j * blks_per_cluster));
fbuf = xfs_trans_get_buf(tp, mp->m_ddev_targp, d,
mp->m_bsize * blks_per_cluster,
XFS_BUF_LOCK);
ASSERT(fbuf);
ASSERT(!XFS_BUF_GETERROR(fbuf));
/*
* Initialize all inodes in this buffer and then log them.
*
* XXX: It would be much better if we had just one transaction
* to log a whole cluster of inodes instead of all the
* individual transactions causing a lot of log traffic.
*/
xfs_biozero(fbuf, 0, ninodes << mp->m_sb.sb_inodelog);
for (i = 0; i < ninodes; i++) {
int ioffset = i << mp->m_sb.sb_inodelog;
uint isize = sizeof(struct xfs_dinode);
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);
xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
}
xfs_trans_inode_alloc_buf(tp, fbuf);
}
}
/*
* Allocate new inodes in the allocation group specified by agbp.
* Return 0 for success, else error code.
......@@ -164,24 +244,15 @@ xfs_ialloc_ag_alloc(
{
xfs_agi_t *agi; /* allocation group header */
xfs_alloc_arg_t args; /* allocation argument structure */
int blks_per_cluster; /* fs blocks per inode cluster */
xfs_btree_cur_t *cur; /* inode btree cursor */
xfs_daddr_t d; /* disk addr of buffer */
xfs_agnumber_t agno;
int error;
xfs_buf_t *fbuf; /* new free inodes' buffer */
xfs_dinode_t *free; /* new free inode structure */
int i; /* inode counter */
int j; /* block counter */
int nbufs; /* num bufs of new inodes */
int i;
xfs_agino_t newino; /* new first inode's number */
xfs_agino_t newlen; /* new number of inodes */
int ninodes; /* num inodes per buf */
xfs_agino_t thisino; /* current inode number, for loop */
int version; /* inode version number to use */
int isaligned = 0; /* inode allocation at stripe unit */
/* boundary */
unsigned int gen;
args.tp = tp;
args.mp = tp->t_mountp;
......@@ -202,12 +273,12 @@ xfs_ialloc_ag_alloc(
*/
agi = XFS_BUF_TO_AGI(agbp);
newino = be32_to_cpu(agi->agi_newino);
agno = be32_to_cpu(agi->agi_seqno);
args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
XFS_IALLOC_BLOCKS(args.mp);
if (likely(newino != NULLAGINO &&
(args.agbno < be32_to_cpu(agi->agi_length)))) {
args.fsbno = XFS_AGB_TO_FSB(args.mp,
be32_to_cpu(agi->agi_seqno), args.agbno);
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
args.type = XFS_ALLOCTYPE_THIS_BNO;
args.mod = args.total = args.wasdel = args.isfl =
args.userdata = args.minalignslop = 0;
......@@ -258,8 +329,7 @@ xfs_ialloc_ag_alloc(
* For now, just allocate blocks up front.
*/
args.agbno = be32_to_cpu(agi->agi_root);
args.fsbno = XFS_AGB_TO_FSB(args.mp,
be32_to_cpu(agi->agi_seqno), args.agbno);
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
/*
* Allocate a fixed-size extent of inodes.
*/
......@@ -282,8 +352,7 @@ xfs_ialloc_ag_alloc(
if (isaligned && args.fsbno == NULLFSBLOCK) {
args.type = XFS_ALLOCTYPE_NEAR_BNO;
args.agbno = be32_to_cpu(agi->agi_root);
args.fsbno = XFS_AGB_TO_FSB(args.mp,
be32_to_cpu(agi->agi_seqno), args.agbno);
args.fsbno = XFS_AGB_TO_FSB(args.mp, agno, args.agbno);
args.alignment = xfs_ialloc_cluster_alignment(&args);
if ((error = xfs_alloc_vextent(&args)))
return error;
......@@ -294,85 +363,30 @@ xfs_ialloc_ag_alloc(
return 0;
}
ASSERT(args.len == args.minlen);
/*
* Convert the results.
*/
newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
/*
* 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.
*/
if (args.mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(args.mp)) {
blks_per_cluster = 1;
nbufs = (int)args.len;
ninodes = args.mp->m_sb.sb_inopblock;
} else {
blks_per_cluster = XFS_INODE_CLUSTER_SIZE(args.mp) /
args.mp->m_sb.sb_blocksize;
nbufs = (int)args.len / blks_per_cluster;
ninodes = blks_per_cluster * args.mp->m_sb.sb_inopblock;
}
/*
* 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.
*/
if (xfs_sb_version_hasnlink(&args.mp->m_sb))
version = 2;
else
version = 1;
/*
* Stamp and write the inode buffers.
*
* 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.
*/
gen = random32();
for (j = 0; j < nbufs; j++) {
/*
* Get the block.
*/
d = XFS_AGB_TO_DADDR(args.mp, be32_to_cpu(agi->agi_seqno),
args.agbno + (j * blks_per_cluster));
fbuf = xfs_trans_get_buf(tp, args.mp->m_ddev_targp, d,
args.mp->m_bsize * blks_per_cluster,
XFS_BUF_LOCK);
ASSERT(fbuf);
ASSERT(!XFS_BUF_GETERROR(fbuf));
xfs_ialloc_inode_init(args.mp, tp, agno, args.agbno, args.len,
random32());
/*
* Initialize all inodes in this buffer and then log them.
*
* XXX: It would be much better if we had just one transaction to
* log a whole cluster of inodes instead of all the individual
* transactions causing a lot of log traffic.
*/
xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
for (i = 0; i < ninodes; i++) {
int ioffset = i << args.mp->m_sb.sb_inodelog;
uint isize = sizeof(struct xfs_dinode);
free = xfs_make_iptr(args.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);
xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
}
xfs_trans_inode_alloc_buf(tp, fbuf);
}
/*
* Convert the results.
*/
newino = XFS_OFFBNO_TO_AGINO(args.mp, args.agbno, 0);
be32_add_cpu(&agi->agi_count, newlen);
be32_add_cpu(&agi->agi_freecount, newlen);
agno = be32_to_cpu(agi->agi_seqno);
down_read(&args.mp->m_peraglock);
args.mp->m_perag[agno].pagi_freecount += newlen;
up_read(&args.mp->m_peraglock);
agi->agi_newino = cpu_to_be32(newino);
/*
* Insert records describing the new inode chunk into the btree.
*/
......@@ -380,13 +394,17 @@ xfs_ialloc_ag_alloc(
for (thisino = newino;
thisino < newino + newlen;
thisino += XFS_INODES_PER_CHUNK) {
if ((error = xfs_inobt_lookup_eq(cur, thisino,
XFS_INODES_PER_CHUNK, XFS_INOBT_ALL_FREE, &i))) {
cur->bc_rec.i.ir_startino = thisino;
cur->bc_rec.i.ir_freecount = XFS_INODES_PER_CHUNK;
cur->bc_rec.i.ir_free = XFS_INOBT_ALL_FREE;
error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, &i);
if (error) {
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
return error;
}
ASSERT(i == 0);
if ((error = xfs_btree_insert(cur, &i))) {
error = xfs_btree_insert(cur, &i);
if (error) {
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
return error;
}
......@@ -538,6 +556,62 @@ xfs_ialloc_ag_select(
}
}
/*
* 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;
XFS_WANT_CORRUPTED_RETURN(i == 1);
}
return 0;
}
STATIC int
xfs_ialloc_get_rec(
struct xfs_btree_cur *cur,
xfs_agino_t agino,
xfs_inobt_rec_incore_t *rec,
int *done,
int left)
{
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;
XFS_WANT_CORRUPTED_RETURN(i == 1);
}
return 0;
}
/*
* Visible inode allocation functions.
*/
......@@ -592,8 +666,8 @@ xfs_dialloc(
int j; /* result code */
xfs_mount_t *mp; /* file system mount structure */
int offset; /* index of inode in chunk */
xfs_agino_t pagino; /* parent's a.g. relative inode # */
xfs_agnumber_t pagno; /* parent's allocation group number */
xfs_agino_t pagino; /* parent's AG relative inode # */
xfs_agnumber_t pagno; /* parent's AG number */
xfs_inobt_rec_incore_t rec; /* inode allocation record */
xfs_agnumber_t tagno; /* testing allocation group number */
xfs_btree_cur_t *tcur; /* temp cursor */
......@@ -716,6 +790,8 @@ xfs_dialloc(
*/
agno = tagno;
*IO_agbp = NULL;
restart_pagno:
cur = xfs_inobt_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno));
/*
* If pagino is 0 (this is the root inode allocation) use newino.
......@@ -723,220 +799,199 @@ xfs_dialloc(
*/
if (!pagino)
pagino = be32_to_cpu(agi->agi_newino);
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
do {
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
} while (i == 1);
error = xfs_check_agi_freecount(cur, agi);
if (error)
goto error0;
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
/*
* If in the same a.g. as the parent, try to get near the parent.
* If in the same AG as the parent, try to get near the parent.
*/
if (pagno == agno) {
if ((error = xfs_inobt_lookup_le(cur, pagino, 0, 0, &i)))
xfs_perag_t *pag = &mp->m_perag[agno];
int doneleft; /* done, to the left */
int doneright; /* done, to the right */
int searchdistance = 10;
error = xfs_inobt_lookup(cur, pagino, XFS_LOOKUP_LE, &i);
if (error)
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
error = xfs_inobt_get_rec(cur, &rec, &j);
if (error)
goto error0;
if (i != 0 &&
(error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
j == 1 &&
rec.ir_freecount > 0) {
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if (rec.ir_freecount > 0) {
/*
* Found a free inode in the same chunk
* as parent, done.
* as the parent, done.
*/
goto alloc_inode;
}
/*
* In the same AG as parent, but parent's chunk is full.
*/
/* duplicate the cursor, search left & right simultaneously */
error = xfs_btree_dup_cursor(cur, &tcur);
if (error)
goto error0;
/*
* In the same a.g. as parent, but parent's chunk is full.
* Skip to last blocks looked up if same parent inode.
*/
else {
int doneleft; /* done, to the left */
int doneright; /* done, to the right */
if (pagino != NULLAGINO &&
pag->pagl_pagino == pagino &&
pag->pagl_leftrec != NULLAGINO &&
pag->pagl_rightrec != NULLAGINO) {
error = xfs_ialloc_get_rec(tcur, pag->pagl_leftrec,
&trec, &doneleft, 1);
if (error)
goto error1;
error = xfs_ialloc_get_rec(cur, pag->pagl_rightrec,
&rec, &doneright, 0);
if (error)
goto error0;
ASSERT(i == 1);
ASSERT(j == 1);
/*
* Duplicate the cursor, search left & right
* simultaneously.
*/
if ((error = xfs_btree_dup_cursor(cur, &tcur)))
goto error0;
/*
* Search left with tcur, back up 1 record.
*/
if ((error = xfs_btree_decrement(tcur, 0, &i)))
goto error1;
doneleft = !i;
if (!doneleft) {
if ((error = xfs_inobt_get_rec(tcur,
&trec.ir_startino,
&trec.ir_freecount,
&trec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
}
/*
* Search right with cur, go forward 1 record.
*/
if ((error = xfs_btree_increment(cur, 0, &i)))
} else {
/* search left with tcur, back up 1 record */
error = xfs_ialloc_next_rec(tcur, &trec, &doneleft, 1);
if (error)
goto error1;
doneright = !i;
if (!doneright) {
if ((error = xfs_inobt_get_rec(cur,
&rec.ir_startino,
&rec.ir_freecount,
&rec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
}
/*
* Loop until we find the closest inode chunk
* with a free one.
*/
while (!doneleft || !doneright) {
int useleft; /* using left inode
chunk this time */
/* search right with cur, go forward 1 record. */
error = xfs_ialloc_next_rec(cur, &rec, &doneright, 0);
if (error)
goto error1;
}
/*
* Loop until we find an inode chunk with a free inode.
*/
while (!doneleft || !doneright) {
int useleft; /* using left inode chunk this time */
if (!--searchdistance) {
/*
* Figure out which block is closer,
* if both are valid.
*/
if (!doneleft && !doneright)
useleft =
pagino -
(trec.ir_startino +
XFS_INODES_PER_CHUNK - 1) <
rec.ir_startino - pagino;
else
useleft = !doneleft;
/*
* If checking the left, does it have
* free inodes?
*/
if (useleft && trec.ir_freecount) {
/*
* Yes, set it up as the chunk to use.
*/
rec = trec;
xfs_btree_del_cursor(cur,
XFS_BTREE_NOERROR);
cur = tcur;
break;
}
/*
* If checking the right, does it have
* free inodes?
*/
if (!useleft && rec.ir_freecount) {
/*
* Yes, it's already set up.
*/
xfs_btree_del_cursor(tcur,
XFS_BTREE_NOERROR);
break;
}
/*
* If used the left, get another one
* further left.
*/
if (useleft) {
if ((error = xfs_btree_decrement(tcur, 0,
&i)))
goto error1;
doneleft = !i;
if (!doneleft) {
if ((error = xfs_inobt_get_rec(
tcur,
&trec.ir_startino,
&trec.ir_freecount,
&trec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1,
error1);
}
}
/*
* If used the right, get another one
* further right.
* Not in range - save last search
* location and allocate a new inode
*/
else {
if ((error = xfs_btree_increment(cur, 0,
&i)))
goto error1;
doneright = !i;
if (!doneright) {
if ((error = xfs_inobt_get_rec(
cur,
&rec.ir_startino,
&rec.ir_freecount,
&rec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1,
error1);
}
}
pag->pagl_leftrec = trec.ir_startino;
pag->pagl_rightrec = rec.ir_startino;
pag->pagl_pagino = pagino;
goto newino;
}
/* 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;
}
ASSERT(!doneleft || !doneright);
/* free inodes to the left? */
if (useleft && trec.ir_freecount) {
rec = trec;
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
cur = tcur;
pag->pagl_leftrec = trec.ir_startino;
pag->pagl_rightrec = rec.ir_startino;
pag->pagl_pagino = pagino;
goto alloc_inode;
}
/* free inodes to the right? */
if (!useleft && rec.ir_freecount) {
xfs_btree_del_cursor(tcur, XFS_BTREE_NOERROR);
pag->pagl_leftrec = trec.ir_startino;
pag->pagl_rightrec = rec.ir_startino;
pag->pagl_pagino = pagino;
goto alloc_inode;
}
/* 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;
}
/*
* 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;
}
/*
* In a different a.g. from the parent.
* In a different AG from the parent.
* See if the most recently allocated block has any free.
*/
else if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
if ((error = xfs_inobt_lookup_eq(cur,
be32_to_cpu(agi->agi_newino), 0, 0, &i)))
newino:
if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
error = xfs_inobt_lookup(cur, be32_to_cpu(agi->agi_newino),
XFS_LOOKUP_EQ, &i);
if (error)
goto error0;
if (i == 1 &&
(error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
j == 1 &&
rec.ir_freecount > 0) {
/*
* The last chunk allocated in the group still has
* a free inode.
*/
}
/*
* None left in the last group, search the whole a.g.
*/
else {
if (i == 1) {
error = xfs_inobt_get_rec(cur, &rec, &j);
if (error)
goto error0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
ASSERT(i == 1);
for (;;) {
if ((error = xfs_inobt_get_rec(cur,
&rec.ir_startino,
&rec.ir_freecount, &rec.ir_free,
&i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if (rec.ir_freecount > 0)
break;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if (j == 1 && rec.ir_freecount > 0) {
/*
* The last chunk allocated in the group
* still has a free inode.
*/
goto alloc_inode;
}
}
}
/*
* None left in the last group, search the whole AG
*/
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
if (error)
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
for (;;) {
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error)
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if (rec.ir_freecount > 0)
break;
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
}
alloc_inode:
offset = xfs_ialloc_find_free(&rec.ir_free);
ASSERT(offset >= 0);
ASSERT(offset < XFS_INODES_PER_CHUNK);
......@@ -945,33 +1000,19 @@ xfs_dialloc(
ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
rec.ir_free &= ~XFS_INOBT_MASK(offset);
rec.ir_freecount--;
if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount,
rec.ir_free)))
error = xfs_inobt_update(cur, &rec);
if (error)
goto error0;
be32_add_cpu(&agi->agi_freecount, -1);
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
down_read(&mp->m_peraglock);
mp->m_perag[tagno].pagi_freecount--;
up_read(&mp->m_peraglock);
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
do {
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
} while (i == 1);
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
error = xfs_check_agi_freecount(cur, agi);
if (error)
goto error0;
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
*inop = ino;
......@@ -1062,38 +1103,23 @@ xfs_difree(
* Initialize the cursor.
*/
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
do {
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &i)))
goto error0;
if (i) {
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
}
} while (i == 1);
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
error = xfs_check_agi_freecount(cur, agi);
if (error)
goto error0;
/*
* Look for the entry describing this inode.
*/
if ((error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i))) {
if ((error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i))) {
cmn_err(CE_WARN,
"xfs_difree: xfs_inobt_lookup_le returned() an error %d on %s. Returning error.",
"xfs_difree: xfs_inobt_lookup returned() an error %d on %s. Returning error.",
error, mp->m_fsname);
goto error0;
}
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino, &rec.ir_freecount,
&rec.ir_free, &i))) {
error = xfs_inobt_get_rec(cur, &rec, &i);
if (error) {
cmn_err(CE_WARN,
"xfs_difree: xfs_inobt_get_rec() returned an error %d on %s. Returning error.",
error, mp->m_fsname);
......@@ -1148,12 +1174,14 @@ xfs_difree(
} else {
*delete = 0;
if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount, rec.ir_free))) {
error = xfs_inobt_update(cur, &rec);
if (error) {
cmn_err(CE_WARN,
"xfs_difree: xfs_inobt_update() returned an error %d on %s. Returning error.",
"xfs_difree: xfs_inobt_update returned an error %d on %s.",
error, mp->m_fsname);
goto error0;
}
/*
* Change the inode free counts and log the ag/sb changes.
*/
......@@ -1165,28 +1193,10 @@ xfs_difree(
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
}
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
error = xfs_check_agi_freecount(cur, agi);
if (error)
goto error0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
do {
if ((error = xfs_inobt_get_rec(cur,
&rec.ir_startino,
&rec.ir_freecount,
&rec.ir_free, &i)))
goto error0;
if (i) {
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
}
} while (i == 1);
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
return 0;
......@@ -1297,9 +1307,7 @@ xfs_imap(
chunk_agbno = agbno - offset_agbno;
} else {
xfs_btree_cur_t *cur; /* inode btree cursor */
xfs_agino_t chunk_agino; /* first agino in inode chunk */
__int32_t chunk_cnt; /* count of free inodes in chunk */
xfs_inofree_t chunk_free; /* mask of free inodes in chunk */
xfs_inobt_rec_incore_t chunk_rec;
xfs_buf_t *agbp; /* agi buffer */
int i; /* temp state */
......@@ -1315,15 +1323,14 @@ xfs_imap(
}
cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
error = xfs_inobt_lookup_le(cur, agino, 0, 0, &i);
error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
if (error) {
xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
"xfs_inobt_lookup_le() failed");
"xfs_inobt_lookup() failed");
goto error0;
}
error = xfs_inobt_get_rec(cur, &chunk_agino, &chunk_cnt,
&chunk_free, &i);
error = xfs_inobt_get_rec(cur, &chunk_rec, &i);
if (error) {
xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: "
"xfs_inobt_get_rec() failed");
......@@ -1341,7 +1348,7 @@ xfs_imap(
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
if (error)
return error;
chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_agino);
chunk_agbno = XFS_AGINO_TO_AGBNO(mp, chunk_rec.ir_startino);
offset_agbno = agbno - chunk_agbno;
}
......
......@@ -150,23 +150,15 @@ xfs_ialloc_pagi_init(
xfs_agnumber_t agno); /* allocation group number */
/*
* Lookup the first record greater than or equal to ino
* in the btree given by cur.
* Lookup a record by ino in the btree given by cur.
*/
int xfs_inobt_lookup_ge(struct xfs_btree_cur *cur, xfs_agino_t ino,
__int32_t fcnt, xfs_inofree_t free, int *stat);
/*
* Lookup the first record less than or equal to ino
* in the btree given by cur.
*/
int xfs_inobt_lookup_le(struct xfs_btree_cur *cur, xfs_agino_t ino,
__int32_t fcnt, xfs_inofree_t free, int *stat);
int xfs_inobt_lookup(struct xfs_btree_cur *cur, xfs_agino_t ino,
xfs_lookup_t dir, int *stat);
/*
* Get the data from the pointed-to record.
*/
extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur, xfs_agino_t *ino,
__int32_t *fcnt, xfs_inofree_t *free, int *stat);
extern int xfs_inobt_get_rec(struct xfs_btree_cur *cur,
xfs_inobt_rec_incore_t *rec, int *stat);
#endif /* __XFS_IALLOC_H__ */
......@@ -82,7 +82,6 @@ xfs_inode_alloc(
memset(&ip->i_df, 0, sizeof(xfs_ifork_t));
ip->i_flags = 0;
ip->i_update_core = 0;
ip->i_update_size = 0;
ip->i_delayed_blks = 0;
memset(&ip->i_d, 0, sizeof(xfs_icdinode_t));
ip->i_size = 0;
......@@ -456,32 +455,6 @@ xfs_iget(
return error;
}
/*
* Look for the inode corresponding to the given ino in the hash table.
* If it is there and its i_transp pointer matches tp, return it.
* Otherwise, return NULL.
*/
xfs_inode_t *
xfs_inode_incore(xfs_mount_t *mp,
xfs_ino_t ino,
xfs_trans_t *tp)
{
xfs_inode_t *ip;
xfs_perag_t *pag;
pag = xfs_get_perag(mp, ino);
read_lock(&pag->pag_ici_lock);
ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, ino));
read_unlock(&pag->pag_ici_lock);
xfs_put_perag(mp, pag);
/* the returned inode must match the transaction */
if (ip && (ip->i_transp != tp))
return NULL;
return ip;
}
/*
* Decrement reference count of an inode structure and unlock it.
*
......
......@@ -651,7 +651,7 @@ xfs_iformat_btree(
return 0;
}
void
STATIC void
xfs_dinode_from_disk(
xfs_icdinode_t *to,
xfs_dinode_t *from)
......@@ -1247,7 +1247,7 @@ xfs_isize_check(
* In that case the pages will still be in memory, but the inode size
* will never have been updated.
*/
xfs_fsize_t
STATIC xfs_fsize_t
xfs_file_last_byte(
xfs_inode_t *ip)
{
......@@ -3837,7 +3837,7 @@ xfs_iext_inline_to_direct(
/*
* Resize an extent indirection array to new_size bytes.
*/
void
STATIC void
xfs_iext_realloc_indirect(
xfs_ifork_t *ifp, /* inode fork pointer */
int new_size) /* new indirection array size */
......@@ -3862,7 +3862,7 @@ xfs_iext_realloc_indirect(
/*
* Switch from indirection array to linear (direct) extent allocations.
*/
void
STATIC void
xfs_iext_indirect_to_direct(
xfs_ifork_t *ifp) /* inode fork pointer */
{
......
......@@ -261,7 +261,6 @@ typedef struct xfs_inode {
/* Miscellaneous state. */
unsigned short i_flags; /* see defined flags below */
unsigned char i_update_core; /* timestamps/size is dirty */
unsigned char i_update_size; /* di_size field is dirty */
unsigned int i_delayed_blks; /* count of delay alloc blks */
xfs_icdinode_t i_d; /* most of ondisk inode */
......@@ -468,8 +467,6 @@ static inline void xfs_ifunlock(xfs_inode_t *ip)
/*
* xfs_iget.c prototypes.
*/
xfs_inode_t *xfs_inode_incore(struct xfs_mount *, xfs_ino_t,
struct xfs_trans *);
int xfs_iget(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
uint, uint, xfs_inode_t **, xfs_daddr_t);
void xfs_iput(xfs_inode_t *, uint);
......@@ -504,7 +501,6 @@ void xfs_ipin(xfs_inode_t *);
void xfs_iunpin(xfs_inode_t *);
int xfs_iflush(xfs_inode_t *, uint);
void xfs_ichgtime(xfs_inode_t *, int);
xfs_fsize_t xfs_file_last_byte(xfs_inode_t *);
void xfs_lock_inodes(xfs_inode_t **, int, uint);
void xfs_lock_two_inodes(xfs_inode_t *, xfs_inode_t *, uint);
......@@ -572,8 +568,6 @@ int xfs_itobp(struct xfs_mount *, struct xfs_trans *,
struct xfs_buf **, uint);
int xfs_iread(struct xfs_mount *, struct xfs_trans *,
struct xfs_inode *, xfs_daddr_t, uint);
void xfs_dinode_from_disk(struct xfs_icdinode *,
struct xfs_dinode *);
void xfs_dinode_to_disk(struct xfs_dinode *,
struct xfs_icdinode *);
void xfs_idestroy_fork(struct xfs_inode *, int);
......@@ -592,8 +586,6 @@ void xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
void xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
void xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
void xfs_iext_realloc_direct(xfs_ifork_t *, int);
void xfs_iext_realloc_indirect(xfs_ifork_t *, int);
void xfs_iext_indirect_to_direct(xfs_ifork_t *);
void xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
void xfs_iext_inline_to_direct(xfs_ifork_t *, int);
void xfs_iext_destroy(xfs_ifork_t *);
......
......@@ -262,14 +262,6 @@ xfs_inode_item_format(
SYNCHRONIZE();
}
/*
* We don't have to worry about re-ordering here because
* the update_size field is protected by the inode lock
* and we have that held in exclusive mode.
*/
if (ip->i_update_size)
ip->i_update_size = 0;
/*
* Make sure to get the latest atime from the Linux inode.
*/
......@@ -712,8 +704,6 @@ xfs_inode_item_unlock(
* Clear out the fields of the inode log item particular
* to the current transaction.
*/
iip->ili_ilock_recur = 0;
iip->ili_iolock_recur = 0;
iip->ili_flags = 0;
/*
......
......@@ -137,8 +137,6 @@ typedef struct xfs_inode_log_item {
struct xfs_inode *ili_inode; /* inode ptr */
xfs_lsn_t ili_flush_lsn; /* lsn at last flush */
xfs_lsn_t ili_last_lsn; /* lsn at last transaction */
unsigned short ili_ilock_recur; /* lock recursion count */
unsigned short ili_iolock_recur; /* lock recursion count */
unsigned short ili_flags; /* misc flags */
unsigned short ili_logged; /* flushed logged data */
unsigned int ili_last_fields; /* fields when flushed */
......
......@@ -72,7 +72,6 @@ struct xfs_mount;
#if XFS_BIG_INUMS
#define XFS_MAXINUMBER ((xfs_ino_t)((1ULL << 56) - 1ULL))
#define XFS_INO64_OFFSET ((xfs_ino_t)(1ULL << 32))
#else
#define XFS_MAXINUMBER ((xfs_ino_t)((1ULL << 32) - 1ULL))
#endif
......
......@@ -39,7 +39,7 @@
#include "xfs_error.h"
#include "xfs_btree.h"
int
STATIC int
xfs_internal_inum(
xfs_mount_t *mp,
xfs_ino_t ino)
......@@ -353,9 +353,6 @@ xfs_bulkstat(
int end_of_ag; /* set if we've seen the ag end */
int error; /* error code */
int fmterror;/* bulkstat formatter result */
__int32_t gcnt; /* current btree rec's count */
xfs_inofree_t gfree; /* current btree rec's free mask */
xfs_agino_t gino; /* current btree rec's start inode */
int i; /* loop index */
int icount; /* count of inodes good in irbuf */
size_t irbsize; /* size of irec buffer in bytes */
......@@ -442,40 +439,43 @@ xfs_bulkstat(
* we need to get the remainder of the chunk we're in.
*/
if (agino > 0) {
xfs_inobt_rec_incore_t r;
/*
* Lookup the inode chunk that this inode lives in.
*/
error = xfs_inobt_lookup_le(cur, agino, 0, 0, &tmp);
error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE,
&tmp);
if (!error && /* no I/O error */
tmp && /* lookup succeeded */
/* got the record, should always work */
!(error = xfs_inobt_get_rec(cur, &gino, &gcnt,
&gfree, &i)) &&
!(error = xfs_inobt_get_rec(cur, &r, &i)) &&
i == 1 &&
/* this is the right chunk */
agino < gino + XFS_INODES_PER_CHUNK &&
agino < r.ir_startino + XFS_INODES_PER_CHUNK &&
/* lastino was not last in chunk */
(chunkidx = agino - gino + 1) <
(chunkidx = agino - r.ir_startino + 1) <
XFS_INODES_PER_CHUNK &&
/* there are some left allocated */
xfs_inobt_maskn(chunkidx,
XFS_INODES_PER_CHUNK - chunkidx) & ~gfree) {
XFS_INODES_PER_CHUNK - chunkidx) &
~r.ir_free) {
/*
* Grab the chunk record. Mark all the
* uninteresting inodes (because they're
* before our start point) free.
*/
for (i = 0; i < chunkidx; i++) {
if (XFS_INOBT_MASK(i) & ~gfree)
gcnt++;
if (XFS_INOBT_MASK(i) & ~r.ir_free)
r.ir_freecount++;
}
gfree |= xfs_inobt_maskn(0, chunkidx);
irbp->ir_startino = gino;
irbp->ir_freecount = gcnt;
irbp->ir_free = gfree;
r.ir_free |= xfs_inobt_maskn(0, chunkidx);
irbp->ir_startino = r.ir_startino;
irbp->ir_freecount = r.ir_freecount;
irbp->ir_free = r.ir_free;
irbp++;
agino = gino + XFS_INODES_PER_CHUNK;
icount = XFS_INODES_PER_CHUNK - gcnt;
agino = r.ir_startino + XFS_INODES_PER_CHUNK;
icount = XFS_INODES_PER_CHUNK - r.ir_freecount;
} else {
/*
* If any of those tests failed, bump the
......@@ -493,7 +493,7 @@ xfs_bulkstat(
/*
* Start of ag. Lookup the first inode chunk.
*/
error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &tmp);
error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &tmp);
icount = 0;
}
/*
......@@ -501,6 +501,8 @@ xfs_bulkstat(
* until we run out of inodes or space in the buffer.
*/
while (irbp < irbufend && icount < ubcount) {
xfs_inobt_rec_incore_t r;
/*
* Loop as long as we're unable to read the
* inode btree.
......@@ -510,51 +512,55 @@ xfs_bulkstat(
if (XFS_AGINO_TO_AGBNO(mp, agino) >=
be32_to_cpu(agi->agi_length))
break;
error = xfs_inobt_lookup_ge(cur, agino, 0, 0,
&tmp);
error = xfs_inobt_lookup(cur, agino,
XFS_LOOKUP_GE, &tmp);
cond_resched();
}
/*
* If ran off the end of the ag either with an error,
* or the normal way, set end and stop collecting.
*/
if (error ||
(error = xfs_inobt_get_rec(cur, &gino, &gcnt,
&gfree, &i)) ||
i == 0) {
if (error) {
end_of_ag = 1;
break;
}
error = xfs_inobt_get_rec(cur, &r, &i);
if (error || i == 0) {
end_of_ag = 1;
break;
}
/*
* If this chunk has any allocated inodes, save it.
* Also start read-ahead now for this chunk.
*/
if (gcnt < XFS_INODES_PER_CHUNK) {
if (r.ir_freecount < XFS_INODES_PER_CHUNK) {
/*
* Loop over all clusters in the next chunk.
* Do a readahead if there are any allocated
* inodes in that cluster.
*/
for (agbno = XFS_AGINO_TO_AGBNO(mp, gino),
chunkidx = 0;
agbno = XFS_AGINO_TO_AGBNO(mp, r.ir_startino);
for (chunkidx = 0;
chunkidx < XFS_INODES_PER_CHUNK;
chunkidx += nicluster,
agbno += nbcluster) {
if (xfs_inobt_maskn(chunkidx,
nicluster) & ~gfree)
if (xfs_inobt_maskn(chunkidx, nicluster)
& ~r.ir_free)
xfs_btree_reada_bufs(mp, agno,
agbno, nbcluster);
}
irbp->ir_startino = gino;
irbp->ir_freecount = gcnt;
irbp->ir_free = gfree;
irbp->ir_startino = r.ir_startino;
irbp->ir_freecount = r.ir_freecount;
irbp->ir_free = r.ir_free;
irbp++;
icount += XFS_INODES_PER_CHUNK - gcnt;
icount += XFS_INODES_PER_CHUNK - r.ir_freecount;
}
/*
* Set agino to after this chunk and bump the cursor.
*/
agino = gino + XFS_INODES_PER_CHUNK;
agino = r.ir_startino + XFS_INODES_PER_CHUNK;
error = xfs_btree_increment(cur, 0, &tmp);
cond_resched();
}
......@@ -820,9 +826,7 @@ xfs_inumbers(
int bufidx;
xfs_btree_cur_t *cur;
int error;
__int32_t gcnt;
xfs_inofree_t gfree;
xfs_agino_t gino;
xfs_inobt_rec_incore_t r;
int i;
xfs_ino_t ino;
int left;
......@@ -855,7 +859,8 @@ xfs_inumbers(
continue;
}
cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
error = xfs_inobt_lookup_ge(cur, agino, 0, 0, &tmp);
error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_GE,
&tmp);
if (error) {
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
cur = NULL;
......@@ -870,9 +875,8 @@ xfs_inumbers(
continue;
}
}
if ((error = xfs_inobt_get_rec(cur, &gino, &gcnt, &gfree,
&i)) ||
i == 0) {
error = xfs_inobt_get_rec(cur, &r, &i);
if (error || i == 0) {
xfs_buf_relse(agbp);
agbp = NULL;
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
......@@ -881,10 +885,12 @@ xfs_inumbers(
agino = 0;
continue;
}
agino = gino + XFS_INODES_PER_CHUNK - 1;
buffer[bufidx].xi_startino = XFS_AGINO_TO_INO(mp, agno, gino);
buffer[bufidx].xi_alloccount = XFS_INODES_PER_CHUNK - gcnt;
buffer[bufidx].xi_allocmask = ~gfree;
agino = r.ir_startino + XFS_INODES_PER_CHUNK - 1;
buffer[bufidx].xi_startino =
XFS_AGINO_TO_INO(mp, agno, r.ir_startino);
buffer[bufidx].xi_alloccount =
XFS_INODES_PER_CHUNK - r.ir_freecount;
buffer[bufidx].xi_allocmask = ~r.ir_free;
bufidx++;
left--;
if (bufidx == bcount) {
......
......@@ -99,11 +99,6 @@ xfs_bulkstat_one(
void *dibuff,
int *stat);
int
xfs_internal_inum(
xfs_mount_t *mp,
xfs_ino_t ino);
typedef int (*inumbers_fmt_pf)(
void __user *ubuffer, /* buffer to write to */
const xfs_inogrp_t *buffer, /* buffer to read from */
......
......@@ -451,8 +451,6 @@ extern int xlog_find_tail(xlog_t *log,
extern int xlog_recover(xlog_t *log);
extern int xlog_recover_finish(xlog_t *log);
extern void xlog_pack_data(xlog_t *log, xlog_in_core_t *iclog, int);
extern void xlog_recover_process_iunlinks(xlog_t *log);
extern struct xfs_buf *xlog_get_bp(xlog_t *, int);
extern void xlog_put_bp(struct xfs_buf *);
......
......@@ -3263,7 +3263,7 @@ xlog_recover_process_one_iunlink(
* freeing of the inode and its removal from the list must be
* atomic.
*/
void
STATIC void
xlog_recover_process_iunlinks(
xlog_t *log)
{
......
......@@ -1568,7 +1568,7 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
*
* The m_sb_lock must be held when this routine is called.
*/
int
STATIC int
xfs_mod_incore_sb_unlocked(
xfs_mount_t *mp,
xfs_sb_field_t field,
......
......@@ -414,13 +414,10 @@ typedef struct xfs_mod_sb {
extern int xfs_log_sbcount(xfs_mount_t *, uint);
extern int xfs_mountfs(xfs_mount_t *mp);
extern void xfs_mountfs_check_barriers(xfs_mount_t *mp);
extern void xfs_unmountfs(xfs_mount_t *);
extern int xfs_unmountfs_writesb(xfs_mount_t *);
extern int xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int64_t, int);
extern int xfs_mod_incore_sb_unlocked(xfs_mount_t *, xfs_sb_field_t,
int64_t, int);
extern int xfs_mod_incore_sb_batch(xfs_mount_t *, xfs_mod_sb_t *,
uint, int);
extern int xfs_mount_log_sb(xfs_mount_t *, __int64_t);
......
......@@ -563,35 +563,6 @@ xfs_mru_cache_lookup(
return elem ? elem->value : NULL;
}
/*
* To look up an element using its key, but leave its location in the internal
* lists alone, call xfs_mru_cache_peek(). If the element isn't found, this
* function returns NULL.
*
* See the comments above the declaration of the xfs_mru_cache_lookup() function
* for important locking information pertaining to this call.
*/
void *
xfs_mru_cache_peek(
xfs_mru_cache_t *mru,
unsigned long key)
{
xfs_mru_cache_elem_t *elem;
ASSERT(mru && mru->lists);
if (!mru || !mru->lists)
return NULL;
spin_lock(&mru->lock);
elem = radix_tree_lookup(&mru->store, key);
if (!elem)
spin_unlock(&mru->lock);
else
__release(mru_lock); /* help sparse not be stupid */
return elem ? elem->value : NULL;
}
/*
* To release the internal data structure spinlock after having performed an
* xfs_mru_cache_lookup() or an xfs_mru_cache_peek(), call xfs_mru_cache_done()
......
......@@ -49,7 +49,6 @@ int xfs_mru_cache_insert(struct xfs_mru_cache *mru, unsigned long key,
void * xfs_mru_cache_remove(struct xfs_mru_cache *mru, unsigned long key);
void xfs_mru_cache_delete(struct xfs_mru_cache *mru, unsigned long key);
void *xfs_mru_cache_lookup(struct xfs_mru_cache *mru, unsigned long key);
void *xfs_mru_cache_peek(struct xfs_mru_cache *mru, unsigned long key);
void xfs_mru_cache_done(struct xfs_mru_cache *mru);
#endif /* __XFS_MRU_CACHE_H__ */
......@@ -87,90 +87,6 @@ xfs_write_clear_setuid(
return 0;
}
/*
* Handle logging requirements of various synchronous types of write.
*/
int
xfs_write_sync_logforce(
xfs_mount_t *mp,
xfs_inode_t *ip)
{
int error = 0;
/*
* If we're treating this as O_DSYNC and we have not updated the
* size, force the log.
*/
if (!(mp->m_flags & XFS_MOUNT_OSYNCISOSYNC) &&
!(ip->i_update_size)) {
xfs_inode_log_item_t *iip = ip->i_itemp;
/*
* If an allocation transaction occurred
* without extending the size, then we have to force
* the log up the proper point to ensure that the
* allocation is permanent. We can't count on
* the fact that buffered writes lock out direct I/O
* writes - the direct I/O write could have extended
* the size nontransactionally, then finished before
* we started. xfs_write_file will think that the file
* didn't grow but the update isn't safe unless the
* size change is logged.
*
* Force the log if we've committed a transaction
* against the inode or if someone else has and
* the commit record hasn't gone to disk (e.g.
* the inode is pinned). This guarantees that
* all changes affecting the inode are permanent
* when we return.
*/
if (iip && iip->ili_last_lsn) {
error = _xfs_log_force(mp, iip->ili_last_lsn,
XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
} else if (xfs_ipincount(ip) > 0) {
error = _xfs_log_force(mp, (xfs_lsn_t)0,
XFS_LOG_FORCE | XFS_LOG_SYNC, NULL);
}
} else {
xfs_trans_t *tp;
/*
* O_SYNC or O_DSYNC _with_ a size update are handled
* the same way.
*
* If the write was synchronous then we need to make
* sure that the inode modification time is permanent.
* We'll have updated the timestamp above, so here
* we use a synchronous transaction to log the inode.
* It's not fast, but it's necessary.
*
* If this a dsync write and the size got changed
* non-transactionally, then we need to ensure that
* the size change gets logged in a synchronous
* transaction.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_WRITE_SYNC);
if ((error = xfs_trans_reserve(tp, 0,
XFS_SWRITE_LOG_RES(mp),
0, 0, 0))) {
/* Transaction reserve failed */
xfs_trans_cancel(tp, 0);
} else {
/* Transaction reserve successful */
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
}
return error;
}
/*
* Force a shutdown of the filesystem instantly while keeping
* the filesystem consistent. We don't do an unmount here; just shutdown
......
......@@ -68,7 +68,6 @@ xfs_get_extsz_hint(
* Prototypes for functions in xfs_rw.c.
*/
extern int xfs_write_clear_setuid(struct xfs_inode *ip);
extern int xfs_write_sync_logforce(struct xfs_mount *mp, struct xfs_inode *ip);
extern int xfs_bwrite(struct xfs_mount *mp, struct xfs_buf *bp);
extern int xfs_bioerror(struct xfs_buf *bp);
extern int xfs_bioerror_relse(struct xfs_buf *bp);
......@@ -78,10 +77,4 @@ extern int xfs_read_buf(struct xfs_mount *mp, xfs_buftarg_t *btp,
extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
xfs_buf_t *bp, xfs_daddr_t blkno);
/*
* Prototypes for functions in xfs_vnodeops.c.
*/
extern int xfs_free_eofblocks(struct xfs_mount *mp, struct xfs_inode *ip,
int flags);
#endif /* __XFS_RW_H__ */
......@@ -68,7 +68,7 @@ typedef struct xfs_trans_header {
#define XFS_TRANS_GROWFS 14
#define XFS_TRANS_STRAT_WRITE 15
#define XFS_TRANS_DIOSTRAT 16
#define XFS_TRANS_WRITE_SYNC 17
/* 17 was XFS_TRANS_WRITE_SYNC */
#define XFS_TRANS_WRITEID 18
#define XFS_TRANS_ADDAFORK 19
#define XFS_TRANS_ATTRINVAL 20
......
......@@ -307,7 +307,7 @@ xfs_trans_read_buf(
return (flags & XFS_BUF_TRYLOCK) ?
EAGAIN : XFS_ERROR(ENOMEM);
if ((bp != NULL) && (XFS_BUF_GETERROR(bp) != 0)) {
if (XFS_BUF_GETERROR(bp) != 0) {
xfs_ioerror_alert("xfs_trans_read_buf", mp,
bp, blkno);
error = XFS_BUF_GETERROR(bp);
......@@ -315,7 +315,7 @@ xfs_trans_read_buf(
return error;
}
#ifdef DEBUG
if (xfs_do_error && (bp != NULL)) {
if (xfs_do_error) {
if (xfs_error_target == target) {
if (((xfs_req_num++) % xfs_error_mod) == 0) {
xfs_buf_relse(bp);
......
......@@ -49,30 +49,7 @@ xfs_trans_inode_broot_debug(
/*
* Get and lock the inode for the caller if it is not already
* locked within the given transaction. If it is already locked
* within the transaction, just increment its lock recursion count
* and return a pointer to it.
*
* For an inode to be locked in a transaction, the inode lock, as
* opposed to the io lock, must be taken exclusively. This ensures
* that the inode can be involved in only 1 transaction at a time.
* Lock recursion is handled on the io lock, but only for lock modes
* of equal or lesser strength. That is, you can recur on the io lock
* held EXCL with a SHARED request but not vice versa. Also, if
* the inode is already a part of the transaction then you cannot
* go from not holding the io lock to having it EXCL or SHARED.
*
* Use the inode cache routine xfs_inode_incore() to find the inode
* if it is already owned by this transaction.
*
* If we don't already own the inode, use xfs_iget() to get it.
* Since the inode log item structure is embedded in the incore
* inode structure and is initialized when the inode is brought
* into memory, there is nothing to do with it here.
*
* If the given transaction pointer is NULL, just call xfs_iget().
* This simplifies code which must handle both cases.
* Get an inode and join it to the transaction.
*/
int
xfs_trans_iget(
......@@ -84,62 +61,11 @@ xfs_trans_iget(
xfs_inode_t **ipp)
{
int error;
xfs_inode_t *ip;
/*
* If the transaction pointer is NULL, just call the normal
* xfs_iget().
*/
if (tp == NULL)
return xfs_iget(mp, NULL, ino, flags, lock_flags, ipp, 0);
/*
* If we find the inode in core with this transaction
* pointer in its i_transp field, then we know we already
* have it locked. In this case we just increment the lock
* recursion count and return the inode to the caller.
* Assert that the inode is already locked in the mode requested
* by the caller. We cannot do lock promotions yet, so
* die if someone gets this wrong.
*/
if ((ip = xfs_inode_incore(tp->t_mountp, ino, tp)) != NULL) {
/*
* Make sure that the inode lock is held EXCL and
* that the io lock is never upgraded when the inode
* is already a part of the transaction.
*/
ASSERT(ip->i_itemp != NULL);
ASSERT(lock_flags & XFS_ILOCK_EXCL);
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
xfs_isilocked(ip, XFS_IOLOCK_EXCL));
ASSERT((!(lock_flags & XFS_IOLOCK_EXCL)) ||
(ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_EXCL));
ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
xfs_isilocked(ip, XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED));
ASSERT((!(lock_flags & XFS_IOLOCK_SHARED)) ||
(ip->i_itemp->ili_flags & XFS_ILI_IOLOCKED_ANY));
if (lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) {
ip->i_itemp->ili_iolock_recur++;
}
if (lock_flags & XFS_ILOCK_EXCL) {
ip->i_itemp->ili_ilock_recur++;
}
*ipp = ip;
return 0;
}
ASSERT(lock_flags & XFS_ILOCK_EXCL);
error = xfs_iget(tp->t_mountp, tp, ino, flags, lock_flags, &ip, 0);
if (error) {
return error;
}
ASSERT(ip != NULL);
xfs_trans_ijoin(tp, ip, lock_flags);
*ipp = ip;
return 0;
error = xfs_iget(mp, tp, ino, flags, lock_flags, ipp, 0);
if (!error && tp)
xfs_trans_ijoin(tp, *ipp, lock_flags);
return error;
}
/*
......@@ -163,8 +89,6 @@ xfs_trans_ijoin(
xfs_inode_item_init(ip, ip->i_mount);
iip = ip->i_itemp;
ASSERT(iip->ili_flags == 0);
ASSERT(iip->ili_ilock_recur == 0);
ASSERT(iip->ili_iolock_recur == 0);
/*
* Get a log_item_desc to point at the new item.
......
......@@ -611,7 +611,7 @@ xfs_fsync(
xfs_inode_t *ip)
{
xfs_trans_t *tp;
int error;
int error = 0;
int log_flushed = 0, changed = 1;
xfs_itrace_entry(ip);
......@@ -619,14 +619,9 @@ xfs_fsync(
if (XFS_FORCED_SHUTDOWN(ip->i_mount))
return XFS_ERROR(EIO);
/* capture size updates in I/O completion before writing the inode. */
error = xfs_wait_on_pages(ip, 0, -1);
if (error)
return XFS_ERROR(error);
/*
* We always need to make sure that the required inode state is safe on
* disk. The vnode might be clean but we still might need to force the
* disk. The inode might be clean but we still might need to force the
* log because of committed transactions that haven't hit the disk yet.
* Likewise, there could be unflushed non-transactional changes to the
* inode core that have to go to disk and this requires us to issue
......@@ -638,7 +633,7 @@ xfs_fsync(
*/
xfs_ilock(ip, XFS_ILOCK_SHARED);
if (!(ip->i_update_size || ip->i_update_core)) {
if (!ip->i_update_core) {
/*
* Timestamps/size haven't changed since last inode flush or
* inode transaction commit. That means either nothing got
......@@ -718,7 +713,7 @@ xfs_fsync(
* when the link count isn't zero and by xfs_dm_punch_hole() when
* punching a hole to EOF.
*/
int
STATIC int
xfs_free_eofblocks(
xfs_mount_t *mp,
xfs_inode_t *ip,
......@@ -1476,8 +1471,8 @@ xfs_create(
if (error == ENOSPC) {
/* flush outstanding delalloc blocks and retry */
xfs_flush_inodes(dp);
error = xfs_trans_reserve(tp, resblks, XFS_CREATE_LOG_RES(mp), 0,
XFS_TRANS_PERM_LOG_RES, XFS_CREATE_LOG_COUNT);
error = xfs_trans_reserve(tp, resblks, log_res, 0,
XFS_TRANS_PERM_LOG_RES, log_count);
}
if (error == ENOSPC) {
/* No space at all so try a "no-allocation" reservation */
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment