/*
* Allocation group level functions.
*/
static inline int
xfs_ialloc_cluster_alignment(
xfs_alloc_arg_t *args)
{
if (xfs_sb_version_hasalign(&args->mp->m_sb) &&
args->mp->m_sb.sb_inoalignmt >=
XFS_B_TO_FSBT(args->mp, XFS_INODE_CLUSTER_SIZE(args->mp)))
return args->mp->m_sb.sb_inoalignmt;
return 1;
}
/*
* 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 */
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 */
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;
/*
* Locking will ensure that we don't have two callers in here
* at one time.
*/
newlen = XFS_IALLOC_INODES(args.mp);
if (args.mp->m_maxicount &&
args.mp->m_sb.sb_icount + newlen > args.mp->m_maxicount)
return XFS_ERROR(ENOSPC);
args.minlen = args.maxlen = XFS_IALLOC_BLOCKS(args.mp);
/*
* 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.
*/
agi = XFS_BUF_TO_AGI(agbp);
newino = be32_to_cpu(agi->agi_newino);
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.type = XFS_ALLOCTYPE_THIS_BNO;
args.mod = args.total = args.wasdel = args.isfl =
args.userdata = args.minalignslop = 0;
args.prod = 1;
/*
* 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.
*/
args.alignment = 1;
args.minalignslop = xfs_ialloc_cluster_alignment(&args) - 1;
/* Allow space for the inode btree to split. */
args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1;
if ((error = xfs_alloc_vextent(&args)))
return error;
} else
args.fsbno = NULLFSBLOCK;
if (unlikely(args.fsbno == NULLFSBLOCK)) {
/*
* Set the alignment for the allocation.
* If stripe alignment is turned on then align at stripe unit
* boundary.
* If the cluster size is smaller than a filesystem block
* then we're doing I/O for inodes in filesystem block size
* 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;
} else
args.alignment = xfs_ialloc_cluster_alignment(&args);
/*
* 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);
args.fsbno = XFS_AGB_TO_FSB(args.mp,
be32_to_cpu(agi->agi_seqno), args.agbno);
/*
* Allocate a fixed-size extent of inodes.
*/
args.type = XFS_ALLOCTYPE_NEAR_BNO;
args.mod = args.total = args.wasdel = args.isfl =
args.userdata = args.minalignslop = 0;
args.prod = 1;
/*
* Allow space for the inode btree to split.
*/
args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1;
if ((error = xfs_alloc_vextent(&args)))
return error;
}
/*
* If stripe alignment is turned on, then try again with cluster
* alignment.
*/
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.alignment = xfs_ialloc_cluster_alignment(&args);
if ((error = xfs_alloc_vextent(&args)))
return error;
}
if (args.fsbno == NULLFSBLOCK) {
*alloc = 0;
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 = XFS_DINODE_VERSION_2;
else
version = XFS_DINODE_VERSION_1;
/*
* 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));
/*
* Set initial values for the inodes in this buffer.
*/
xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
for (i = 0; i < ninodes; i++) {
free = XFS_MAKE_IPTR(args.mp, fbuf, i);
free->di_core.di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
free->di_core.di_version = version;
free->di_core.di_gen = cpu_to_be32(gen);
free->di_next_unlinked = cpu_to_be32(NULLAGINO);
xfs_ialloc_log_di(tp, fbuf, i,
XFS_DI_CORE_BITS | XFS_DI_NEXT_UNLINKED);
}
xfs_trans_inode_alloc_buf(tp, fbuf);
}
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.
*/
cur = xfs_btree_init_cursor(args.mp, tp, agbp, agno,
XFS_BTNUM_INO, (xfs_inode_t *)0, 0);
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))) {
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
return error;
}
ASSERT(i == 0);
if ((error = xfs_inobt_insert(cur, &i))) {
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
return error;
}
ASSERT(i == 1);
}
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
/*
* 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;
}
/*
* 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).
*/
int
xfs_ialloc_inode_init(
struct xfs_mount *mp,
struct xfs_trans *tp,
struct list_head *buffer_list,
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;
xfs_ino_t ino = 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 (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.
*
* 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
* XFS_OFFBNO_TO_AGINO() only works within a filesystem block, not
* 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.
*/
if (xfs_sb_version_hascrc(&mp->m_sb)) {
version = 3;
ino = XFS_AGINO_TO_INO(mp, agno,
XFS_OFFBNO_TO_AGINO(mp, agbno, 0));
/*
* 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)
xfs_icreate_log(tp, agno, agbno, XFS_IALLOC_INODES(mp),
mp->m_sb.sb_inodesize, length, gen);
} else 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,
XBF_UNMAPPED);
if (!fbuf)
return ENOMEM;
/* Initialize the inode buffers and log them appropriately. */
fbuf->b_ops = &xfs_inode_buf_ops;
xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
for (i = 0; i < ninodes; i++) {
int ioffset = i << mp->m_sb.sb_inodelog;
uint isize = xfs_dinode_size(version);
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);
if (version == 3) {
free->di_ino = cpu_to_be64(ino);
ino++;
uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid);
xfs_dinode_calc_crc(mp, free);
} else if (tp) {
/* just log the inode core */
xfs_trans_log_buf(tp, fbuf, ioffset,
ioffset + isize - 1);
}
}
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) {
/*
* 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);
xfs_trans_log_buf(tp, fbuf, 0,
BBTOB(fbuf->b_length) - 1);
}
} else {
fbuf->b_flags |= XBF_DONE;
xfs_buf_delwri_queue(fbuf, buffer_list);
xfs_buf_relse(fbuf);
}
}
return 0;
}
/*
* Initialise a new set of inodes.
*/
STATIC int
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,
XBF_LOCK);
if (!fbuf)
return ENOMEM;
/*
* 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_buf_zero(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);
}
return 0;
}
/*
* Initialise a new set of inodes.
*/
STATIC int
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;
xfs_ino_t ino = 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 (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.
*
* 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
* XFS_OFFBNO_TO_AGINO() only works within a filesystem block, not
* 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.
*/
if (xfs_sb_version_hascrc(&mp->m_sb)) {
version = 3;
ino = XFS_AGINO_TO_INO(mp, agno,
XFS_OFFBNO_TO_AGINO(mp, agbno, 0));
} else 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,
XBF_UNMAPPED);
if (!fbuf)
return ENOMEM;
/*
* 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.
*/
fbuf->b_ops = &xfs_inode_buf_ops;
xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
for (i = 0; i < ninodes; i++) {
int ioffset = i << mp->m_sb.sb_inodelog;
uint isize = xfs_dinode_size(version);
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);
if (version == 3) {
free->di_ino = cpu_to_be64(ino);
ino++;
uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid);
xfs_dinode_calc_crc(mp, free);
} else {
/* just log the inode core */
xfs_trans_log_buf(tp, fbuf, ioffset,
ioffset + isize - 1);
}
}
if (version == 3) {
/* need to log the entire buffer */
xfs_trans_log_buf(tp, fbuf, 0,
BBTOB(fbuf->b_length) - 1);
}
xfs_trans_inode_alloc_buf(tp, fbuf);
}
return 0;
}
