/* Callers don't need to care which bitmap (local alloc or main) to
* use so we figure it out for them, but unfortunately this clutters
* things a bit. */
int ocfs2_reserve_clusters(struct ocfs2_super *osb,
u32 bits_wanted,
struct ocfs2_alloc_context **ac)
{
int status;
mlog_entry_void();
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = bits_wanted;
status = -ENOSPC;
if (ocfs2_alloc_should_use_local(osb, bits_wanted)) {
status = ocfs2_reserve_local_alloc_bits(osb,
bits_wanted,
*ac);
if ((status < 0) && (status != -ENOSPC)) {
mlog_errno(status);
goto bail;
} else if (status == -ENOSPC) {
/* reserve_local_bits will return enospc with
* the local alloc inode still locked, so we
* can change this safely here. */
mlog(0, "Disabling local alloc\n");
/* We set to OCFS2_LA_DISABLED so that umount
* can clean up what's left of the local
* allocation */
osb->local_alloc_state = OCFS2_LA_DISABLED;
}
}
if (status == -ENOSPC) {
status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
/* Callers don't need to care which bitmap (local alloc or main) to
* use so we figure it out for them, but unfortunately this clutters
* things a bit. */
static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb,
u32 bits_wanted, u64 max_block,
int flags,
struct ocfs2_alloc_context **ac)
{
int status;
mlog_entry_void();
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = bits_wanted;
(*ac)->ac_max_block = max_block;
status = -ENOSPC;
if (!(flags & ALLOC_GROUPS_FROM_GLOBAL) &&
ocfs2_alloc_should_use_local(osb, bits_wanted)) {
status = ocfs2_reserve_local_alloc_bits(osb,
bits_wanted,
*ac);
if (status == -EFBIG) {
/* The local alloc window is outside ac_max_block.
* use the main bitmap. */
status = -ENOSPC;
} else if ((status < 0) && (status != -ENOSPC)) {
mlog_errno(status);
goto bail;
}
}
if (status == -ENOSPC) {
status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
/*
* lock allocators, and reserving appropriate number of bits for
* meta blocks and data clusters.
*
* in some cases, we don't need to reserve clusters, just let data_ac
* be NULL.
*/
static int ocfs2_lock_allocators_move_extents(struct inode *inode,
struct ocfs2_extent_tree *et,
u32 clusters_to_move,
u32 extents_to_split,
struct ocfs2_alloc_context **meta_ac,
struct ocfs2_alloc_context **data_ac,
int extra_blocks,
int *credits)
{
int ret, num_free_extents;
unsigned int max_recs_needed = 2 * extents_to_split + clusters_to_move;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
num_free_extents = ocfs2_num_free_extents(osb, et);
if (num_free_extents < 0) {
ret = num_free_extents;
mlog_errno(ret);
goto out;
}
if (!num_free_extents ||
(ocfs2_sparse_alloc(osb) && num_free_extents < max_recs_needed))
extra_blocks += ocfs2_extend_meta_needed(et->et_root_el);
ret = ocfs2_reserve_new_metadata_blocks(osb, extra_blocks, meta_ac);
if (ret) {
mlog_errno(ret);
goto out;
}
if (data_ac) {
ret = ocfs2_reserve_clusters(osb, clusters_to_move, data_ac);
if (ret) {
mlog_errno(ret);
goto out;
}
}
*credits += ocfs2_calc_extend_credits(osb->sb, et->et_root_el,
clusters_to_move + 2);
mlog(0, "reserve metadata_blocks: %d, data_clusters: %u, credits: %d\n",
extra_blocks, clusters_to_move, *credits);
out:
if (ret) {
if (*meta_ac) {
ocfs2_free_alloc_context(*meta_ac);
*meta_ac = NULL;
}
}
return ret;
}
int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
struct ocfs2_dinode *fe,
struct ocfs2_alloc_context **ac)
{
int status;
u32 slot;
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = ocfs2_extend_meta_needed(fe);
(*ac)->ac_which = OCFS2_AC_USE_META;
#ifndef OCFS2_USE_ALL_METADATA_SUBALLOCATORS
slot = 0;
#else
slot = osb->slot_num;
#endif
(*ac)->ac_group_search = ocfs2_block_group_search;
status = ocfs2_reserve_suballoc_bits(osb, (*ac),
EXTENT_ALLOC_SYSTEM_INODE, slot);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
static int ocfs2_local_alloc_reserve_for_window(struct ocfs2_super *osb,
struct ocfs2_journal_handle *handle,
struct ocfs2_alloc_context **ac,
struct inode **bitmap_inode,
struct buffer_head **bitmap_bh)
{
int status;
*ac = kcalloc(1, sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_handle = handle;
(*ac)->ac_bits_wanted = ocfs2_local_alloc_window_bits(osb);
status = ocfs2_reserve_cluster_bitmap_bits(osb, *ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
*bitmap_inode = (*ac)->ac_inode;
igrab(*bitmap_inode);
*bitmap_bh = (*ac)->ac_bh;
get_bh(*bitmap_bh);
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
int ocfs2_reserve_new_metadata_blocks(struct ocfs2_super *osb,
int blocks,
struct ocfs2_alloc_context **ac)
{
int status;
u32 slot;
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = blocks;
(*ac)->ac_which = OCFS2_AC_USE_META;
slot = osb->slot_num;
(*ac)->ac_group_search = ocfs2_block_group_search;
status = ocfs2_reserve_suballoc_bits(osb, (*ac),
EXTENT_ALLOC_SYSTEM_INODE,
slot, NULL, ALLOC_NEW_GROUP);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
struct ocfs2_alloc_context **ac)
{
int status;
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = 1;
(*ac)->ac_which = OCFS2_AC_USE_INODE;
(*ac)->ac_group_search = ocfs2_block_group_search;
status = ocfs2_reserve_suballoc_bits(osb, *ac,
INODE_ALLOC_SYSTEM_INODE,
osb->slot_num);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
mlog_exit(status);
return status;
}
static int ocfs2_extend_allocation(struct inode *inode,
u32 clusters_to_add)
{
int status = 0;
int restart_func = 0;
int drop_alloc_sem = 0;
int credits;
u32 prev_clusters, logical_start;
struct buffer_head *bh = NULL;
struct ocfs2_dinode *fe = NULL;
handle_t *handle = NULL;
struct ocfs2_alloc_context *data_ac = NULL;
struct ocfs2_alloc_context *meta_ac = NULL;
enum ocfs2_alloc_restarted why;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
mlog_entry("(clusters_to_add = %u)\n", clusters_to_add);
/*
* This function only exists for file systems which don't
* support holes.
*/
BUG_ON(ocfs2_sparse_alloc(osb));
status = ocfs2_read_block(osb, OCFS2_I(inode)->ip_blkno, &bh,
OCFS2_BH_CACHED, inode);
if (status < 0) {
mlog_errno(status);
goto leave;
}
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
OCFS2_RO_ON_INVALID_DINODE(inode->i_sb, fe);
status = -EIO;
goto leave;
}
logical_start = OCFS2_I(inode)->ip_clusters;
restart_all:
BUG_ON(le32_to_cpu(fe->i_clusters) != OCFS2_I(inode)->ip_clusters);
/* blocks peope in read/write from reading our allocation
* until we're done changing it. We depend on i_mutex to block
* other extend/truncate calls while we're here. Ordering wrt
* start_trans is important here -- always do it before! */
down_write(&OCFS2_I(inode)->ip_alloc_sem);
drop_alloc_sem = 1;
status = ocfs2_lock_allocators(inode, fe, clusters_to_add, &data_ac,
&meta_ac);
if (status) {
mlog_errno(status);
goto leave;
}
credits = ocfs2_calc_extend_credits(osb->sb, fe, clusters_to_add);
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
handle = NULL;
mlog_errno(status);
goto leave;
}
restarted_transaction:
/* reserve a write to the file entry early on - that we if we
* run out of credits in the allocation path, we can still
* update i_size. */
status = ocfs2_journal_access(handle, inode, bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto leave;
}
prev_clusters = OCFS2_I(inode)->ip_clusters;
status = ocfs2_do_extend_allocation(osb,
inode,
&logical_start,
clusters_to_add,
bh,
handle,
data_ac,
meta_ac,
&why);
if ((status < 0) && (status != -EAGAIN)) {
if (status != -ENOSPC)
mlog_errno(status);
goto leave;
}
status = ocfs2_journal_dirty(handle, bh);
if (status < 0) {
mlog_errno(status);
goto leave;
}
spin_lock(&OCFS2_I(inode)->ip_lock);
clusters_to_add -= (OCFS2_I(inode)->ip_clusters - prev_clusters);
spin_unlock(&OCFS2_I(inode)->ip_lock);
if (why != RESTART_NONE && clusters_to_add) {
if (why == RESTART_META) {
mlog(0, "restarting function.\n");
restart_func = 1;
} else {
BUG_ON(why != RESTART_TRANS);
mlog(0, "restarting transaction.\n");
/* TODO: This can be more intelligent. */
credits = ocfs2_calc_extend_credits(osb->sb,
fe,
clusters_to_add);
status = ocfs2_extend_trans(handle, credits);
if (status < 0) {
/* handle still has to be committed at
* this point. */
status = -ENOMEM;
mlog_errno(status);
goto leave;
}
goto restarted_transaction;
}
}
mlog(0, "fe: i_clusters = %u, i_size=%llu\n",
le32_to_cpu(fe->i_clusters),
(unsigned long long)le64_to_cpu(fe->i_size));
mlog(0, "inode: ip_clusters=%u, i_size=%lld\n",
OCFS2_I(inode)->ip_clusters, i_size_read(inode));
leave:
if (drop_alloc_sem) {
up_write(&OCFS2_I(inode)->ip_alloc_sem);
drop_alloc_sem = 0;
}
if (handle) {
ocfs2_commit_trans(osb, handle);
handle = NULL;
}
if (data_ac) {
ocfs2_free_alloc_context(data_ac);
data_ac = NULL;
}
if (meta_ac) {
ocfs2_free_alloc_context(meta_ac);
meta_ac = NULL;
}
if ((!status) && restart_func) {
restart_func = 0;
goto restart_all;
}
if (bh) {
brelse(bh);
bh = NULL;
}
mlog_exit(status);
return status;
}
/*
* For a given allocation, determine which allocators will need to be
* accessed, and lock them, reserving the appropriate number of bits.
*
* Called from ocfs2_extend_allocation() for file systems which don't
* support holes, and from ocfs2_write() for file systems which
* understand sparse inodes.
*/
int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_dinode *di,
u32 clusters_to_add,
struct ocfs2_alloc_context **data_ac,
struct ocfs2_alloc_context **meta_ac)
{
int ret, num_free_extents;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
*meta_ac = NULL;
*data_ac = NULL;
mlog(0, "extend inode %llu, i_size = %lld, di->i_clusters = %u, "
"clusters_to_add = %u\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno, i_size_read(inode),
le32_to_cpu(di->i_clusters), clusters_to_add);
num_free_extents = ocfs2_num_free_extents(osb, inode, di);
if (num_free_extents < 0) {
ret = num_free_extents;
mlog_errno(ret);
goto out;
}
/*
* Sparse allocation file systems need to be more conservative
* with reserving room for expansion - the actual allocation
* happens while we've got a journal handle open so re-taking
* a cluster lock (because we ran out of room for another
* extent) will violate ordering rules.
*
* Most of the time we'll only be seeing this 1 cluster at a time
* anyway.
*/
if (!num_free_extents ||
(ocfs2_sparse_alloc(osb) && num_free_extents < clusters_to_add)) {
ret = ocfs2_reserve_new_metadata(osb, di, meta_ac);
if (ret < 0) {
if (ret != -ENOSPC)
mlog_errno(ret);
goto out;
}
}
ret = ocfs2_reserve_clusters(osb, clusters_to_add, data_ac);
if (ret < 0) {
if (ret != -ENOSPC)
mlog_errno(ret);
goto out;
}
out:
if (ret) {
if (*meta_ac) {
ocfs2_free_alloc_context(*meta_ac);
*meta_ac = NULL;
}
/*
* We cannot have an error and a non null *data_ac.
*/
}
return ret;
}
int ocfs2_reserve_new_metadata_blocks(struct ocfs2_super *osb,
int blocks,
struct ocfs2_alloc_context **ac)
{
int status;
int slot = ocfs2_get_meta_steal_slot(osb);
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = blocks;
(*ac)->ac_which = OCFS2_AC_USE_META;
(*ac)->ac_group_search = ocfs2_block_group_search;
if (slot != OCFS2_INVALID_SLOT &&
atomic_read(&osb->s_num_meta_stolen) < OCFS2_MAX_TO_STEAL)
goto extent_steal;
atomic_set(&osb->s_num_meta_stolen, 0);
status = ocfs2_reserve_suballoc_bits(osb, (*ac),
EXTENT_ALLOC_SYSTEM_INODE,
(u32)osb->slot_num, NULL,
ALLOC_GROUPS_FROM_GLOBAL|ALLOC_NEW_GROUP);
if (status >= 0) {
status = 0;
if (slot != OCFS2_INVALID_SLOT)
ocfs2_init_meta_steal_slot(osb);
goto bail;
} else if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
}
ocfs2_free_ac_resource(*ac);
extent_steal:
status = ocfs2_steal_meta(osb, *ac);
atomic_inc(&osb->s_num_meta_stolen);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
if (status)
mlog_errno(status);
return status;
}
/*
* We expect the block group allocator to already be locked.
*/
static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
struct inode *alloc_inode,
struct buffer_head *bh,
u64 max_block,
u64 *last_alloc_group,
int flags)
{
int status, credits;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
struct ocfs2_chain_list *cl;
struct ocfs2_alloc_context *ac = NULL;
handle_t *handle = NULL;
u16 alloc_rec;
struct buffer_head *bg_bh = NULL;
struct ocfs2_group_desc *bg;
BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode));
cl = &fe->id2.i_chain;
status = ocfs2_reserve_clusters_with_limit(osb,
le16_to_cpu(cl->cl_cpg),
max_block, flags, &ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
credits = ocfs2_calc_group_alloc_credits(osb->sb,
le16_to_cpu(cl->cl_cpg));
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
handle = NULL;
mlog_errno(status);
goto bail;
}
if (last_alloc_group && *last_alloc_group != 0) {
trace_ocfs2_block_group_alloc(
(unsigned long long)*last_alloc_group);
ac->ac_last_group = *last_alloc_group;
}
bg_bh = ocfs2_block_group_alloc_contig(osb, handle, alloc_inode,
ac, cl);
if (IS_ERR(bg_bh) && (PTR_ERR(bg_bh) == -ENOSPC))
bg_bh = ocfs2_block_group_alloc_discontig(handle,
alloc_inode,
ac, cl);
if (IS_ERR(bg_bh)) {
status = PTR_ERR(bg_bh);
bg_bh = NULL;
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
bg = (struct ocfs2_group_desc *) bg_bh->b_data;
status = ocfs2_journal_access_di(handle, INODE_CACHE(alloc_inode),
bh, OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto bail;
}
alloc_rec = le16_to_cpu(bg->bg_chain);
le32_add_cpu(&cl->cl_recs[alloc_rec].c_free,
le16_to_cpu(bg->bg_free_bits_count));
le32_add_cpu(&cl->cl_recs[alloc_rec].c_total,
le16_to_cpu(bg->bg_bits));
cl->cl_recs[alloc_rec].c_blkno = bg->bg_blkno;
if (le16_to_cpu(cl->cl_next_free_rec) < le16_to_cpu(cl->cl_count))
le16_add_cpu(&cl->cl_next_free_rec, 1);
le32_add_cpu(&fe->id1.bitmap1.i_used, le16_to_cpu(bg->bg_bits) -
le16_to_cpu(bg->bg_free_bits_count));
le32_add_cpu(&fe->id1.bitmap1.i_total, le16_to_cpu(bg->bg_bits));
le32_add_cpu(&fe->i_clusters, le16_to_cpu(cl->cl_cpg));
ocfs2_journal_dirty(handle, bh);
spin_lock(&OCFS2_I(alloc_inode)->ip_lock);
OCFS2_I(alloc_inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
fe->i_size = cpu_to_le64(ocfs2_clusters_to_bytes(alloc_inode->i_sb,
le32_to_cpu(fe->i_clusters)));
spin_unlock(&OCFS2_I(alloc_inode)->ip_lock);
i_size_write(alloc_inode, le64_to_cpu(fe->i_size));
alloc_inode->i_blocks = ocfs2_inode_sector_count(alloc_inode);
ocfs2_update_inode_fsync_trans(handle, alloc_inode, 0);
status = 0;
/* save the new last alloc group so that the caller can cache it. */
if (last_alloc_group)
*last_alloc_group = ac->ac_last_group;
bail:
if (handle)
ocfs2_commit_trans(osb, handle);
if (ac)
ocfs2_free_alloc_context(ac);
brelse(bg_bh);
if (status)
mlog_errno(status);
return status;
}
int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
struct ocfs2_alloc_context **ac)
{
int status;
int slot = ocfs2_get_inode_steal_slot(osb);
u64 alloc_group;
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
(*ac)->ac_bits_wanted = 1;
(*ac)->ac_which = OCFS2_AC_USE_INODE;
(*ac)->ac_group_search = ocfs2_block_group_search;
/*
* stat(2) can't handle i_ino > 32bits, so we tell the
* lower levels not to allocate us a block group past that
* limit. The 'inode64' mount option avoids this behavior.
*/
if (!(osb->s_mount_opt & OCFS2_MOUNT_INODE64))
(*ac)->ac_max_block = (u32)~0U;
/*
* slot is set when we successfully steal inode from other nodes.
* It is reset in 3 places:
* 1. when we flush the truncate log
* 2. when we complete local alloc recovery.
* 3. when we successfully allocate from our own slot.
* After it is set, we will go on stealing inodes until we find the
* need to check our slots to see whether there is some space for us.
*/
if (slot != OCFS2_INVALID_SLOT &&
atomic_read(&osb->s_num_inodes_stolen) < OCFS2_MAX_TO_STEAL)
goto inode_steal;
atomic_set(&osb->s_num_inodes_stolen, 0);
alloc_group = osb->osb_inode_alloc_group;
status = ocfs2_reserve_suballoc_bits(osb, *ac,
INODE_ALLOC_SYSTEM_INODE,
(u32)osb->slot_num,
&alloc_group,
ALLOC_NEW_GROUP |
ALLOC_GROUPS_FROM_GLOBAL);
if (status >= 0) {
status = 0;
spin_lock(&osb->osb_lock);
osb->osb_inode_alloc_group = alloc_group;
spin_unlock(&osb->osb_lock);
trace_ocfs2_reserve_new_inode_new_group(
(unsigned long long)alloc_group);
/*
* Some inodes must be freed by us, so try to allocate
* from our own next time.
*/
if (slot != OCFS2_INVALID_SLOT)
ocfs2_init_inode_steal_slot(osb);
goto bail;
} else if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
}
ocfs2_free_ac_resource(*ac);
inode_steal:
status = ocfs2_steal_inode(osb, *ac);
atomic_inc(&osb->s_num_inodes_stolen);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
status = 0;
bail:
if ((status < 0) && *ac) {
ocfs2_free_alloc_context(*ac);
*ac = NULL;
}
if (status)
mlog_errno(status);
return status;
}
/* Note that we do *NOT* lock the local alloc inode here as
* it's been locked already for us. */
static int ocfs2_local_alloc_slide_window(struct ocfs2_super *osb,
struct inode *local_alloc_inode)
{
int status = 0;
struct buffer_head *main_bm_bh = NULL;
struct inode *main_bm_inode = NULL;
struct ocfs2_journal_handle *handle = NULL;
struct ocfs2_dinode *alloc;
struct ocfs2_dinode *alloc_copy = NULL;
struct ocfs2_alloc_context *ac = NULL;
mlog_entry_void();
handle = ocfs2_alloc_handle(osb);
if (!handle) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
/* This will lock the main bitmap for us. */
status = ocfs2_local_alloc_reserve_for_window(osb,
handle,
&ac,
&main_bm_inode,
&main_bm_bh);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
handle = ocfs2_start_trans(osb, handle, OCFS2_WINDOW_MOVE_CREDITS);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
handle = NULL;
mlog_errno(status);
goto bail;
}
alloc = (struct ocfs2_dinode *) osb->local_alloc_bh->b_data;
/* We want to clear the local alloc before doing anything
* else, so that if we error later during this operation,
* local alloc shutdown won't try to double free main bitmap
* bits. Make a copy so the sync function knows which bits to
* free. */
alloc_copy = kmalloc(osb->local_alloc_bh->b_size, GFP_KERNEL);
if (!alloc_copy) {
status = -ENOMEM;
mlog_errno(status);
goto bail;
}
memcpy(alloc_copy, alloc, osb->local_alloc_bh->b_size);
status = ocfs2_journal_access(handle, local_alloc_inode,
osb->local_alloc_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto bail;
}
ocfs2_clear_local_alloc(alloc);
status = ocfs2_journal_dirty(handle, osb->local_alloc_bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = ocfs2_sync_local_to_main(osb, handle, alloc_copy,
main_bm_inode, main_bm_bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = ocfs2_local_alloc_new_window(osb, handle, ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
atomic_inc(&osb->alloc_stats.moves);
status = 0;
bail:
if (handle)
ocfs2_commit_trans(handle);
if (main_bm_bh)
brelse(main_bm_bh);
if (main_bm_inode)
iput(main_bm_inode);
if (alloc_copy)
kfree(alloc_copy);
if (ac)
ocfs2_free_alloc_context(ac);
mlog_exit(status);
return status;
}
static int ocfs2_move_extent(struct ocfs2_move_extents_context *context,
u32 cpos, u32 phys_cpos, u32 *new_phys_cpos,
u32 len, int ext_flags)
{
int ret, credits = 0, extra_blocks = 0, goal_bit = 0;
handle_t *handle;
struct inode *inode = context->inode;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct inode *tl_inode = osb->osb_tl_inode;
struct inode *gb_inode = NULL;
struct buffer_head *gb_bh = NULL;
struct buffer_head *gd_bh = NULL;
struct ocfs2_group_desc *gd;
struct ocfs2_refcount_tree *ref_tree = NULL;
u32 move_max_hop = ocfs2_blocks_to_clusters(inode->i_sb,
context->range->me_threshold);
u64 phys_blkno, new_phys_blkno;
phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
if ((ext_flags & OCFS2_EXT_REFCOUNTED) && len) {
BUG_ON(!(OCFS2_I(inode)->ip_dyn_features &
OCFS2_HAS_REFCOUNT_FL));
BUG_ON(!context->refcount_loc);
ret = ocfs2_lock_refcount_tree(osb, context->refcount_loc, 1,
&ref_tree, NULL);
if (ret) {
mlog_errno(ret);
return ret;
}
ret = ocfs2_prepare_refcount_change_for_del(inode,
context->refcount_loc,
phys_blkno,
len,
&credits,
&extra_blocks);
if (ret) {
mlog_errno(ret);
goto out;
}
}
ret = ocfs2_lock_allocators_move_extents(inode, &context->et, len, 1,
&context->meta_ac,
NULL, extra_blocks, &credits);
if (ret) {
mlog_errno(ret);
goto out;
}
/*
* need to count 2 extra credits for global_bitmap inode and
* group descriptor.
*/
credits += OCFS2_INODE_UPDATE_CREDITS + 1;
/*
* ocfs2_move_extent() didn't reserve any clusters in lock_allocators()
* logic, while we still need to lock the global_bitmap.
*/
gb_inode = ocfs2_get_system_file_inode(osb, GLOBAL_BITMAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT);
if (!gb_inode) {
mlog(ML_ERROR, "unable to get global_bitmap inode\n");
ret = -EIO;
goto out;
}
mutex_lock(&gb_inode->i_mutex);
ret = ocfs2_inode_lock(gb_inode, &gb_bh, 1);
if (ret) {
mlog_errno(ret);
goto out_unlock_gb_mutex;
}
mutex_lock(&tl_inode->i_mutex);
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
mlog_errno(ret);
goto out_unlock_tl_inode;
}
new_phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, *new_phys_cpos);
ret = ocfs2_find_victim_alloc_group(inode, new_phys_blkno,
GLOBAL_BITMAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT,
&goal_bit, &gd_bh);
if (ret) {
mlog_errno(ret);
goto out_commit;
}
/*
* probe the victim cluster group to find a proper
* region to fit wanted movement, it even will perfrom
* a best-effort attempt by compromising to a threshold
* around the goal.
*/
ocfs2_probe_alloc_group(inode, gd_bh, &goal_bit, len, move_max_hop,
new_phys_cpos);
if (!*new_phys_cpos) {
ret = -ENOSPC;
goto out_commit;
}
ret = __ocfs2_move_extent(handle, context, cpos, len, phys_cpos,
*new_phys_cpos, ext_flags);
if (ret) {
mlog_errno(ret);
goto out_commit;
}
gd = (struct ocfs2_group_desc *)gd_bh->b_data;
ret = ocfs2_alloc_dinode_update_counts(gb_inode, handle, gb_bh, len,
le16_to_cpu(gd->bg_chain));
if (ret) {
mlog_errno(ret);
goto out_commit;
}
ret = ocfs2_block_group_set_bits(handle, gb_inode, gd, gd_bh,
goal_bit, len);
if (ret)
mlog_errno(ret);
/*
* Here we should write the new page out first if we are
* in write-back mode.
*/
ret = ocfs2_cow_sync_writeback(inode->i_sb, context->inode, cpos, len);
if (ret)
mlog_errno(ret);
out_commit:
ocfs2_commit_trans(osb, handle);
brelse(gd_bh);
out_unlock_tl_inode:
mutex_unlock(&tl_inode->i_mutex);
ocfs2_inode_unlock(gb_inode, 1);
out_unlock_gb_mutex:
mutex_unlock(&gb_inode->i_mutex);
brelse(gb_bh);
iput(gb_inode);
out:
if (context->meta_ac) {
ocfs2_free_alloc_context(context->meta_ac);
context->meta_ac = NULL;
}
if (ref_tree)
ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
return ret;
}
/*
* We expect the block group allocator to already be locked.
*/
static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
struct inode *alloc_inode,
struct buffer_head *bh)
{
int status, credits;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
struct ocfs2_chain_list *cl;
struct ocfs2_alloc_context *ac = NULL;
handle_t *handle = NULL;
u32 bit_off, num_bits;
u16 alloc_rec;
u64 bg_blkno;
struct buffer_head *bg_bh = NULL;
struct ocfs2_group_desc *bg;
BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode));
mlog_entry_void();
cl = &fe->id2.i_chain;
status = ocfs2_reserve_clusters(osb,
le16_to_cpu(cl->cl_cpg),
&ac);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
credits = ocfs2_calc_group_alloc_credits(osb->sb,
le16_to_cpu(cl->cl_cpg));
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
status = PTR_ERR(handle);
handle = NULL;
mlog_errno(status);
goto bail;
}
status = ocfs2_claim_clusters(osb,
handle,
ac,
le16_to_cpu(cl->cl_cpg),
&bit_off,
&num_bits);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
goto bail;
}
alloc_rec = ocfs2_find_smallest_chain(cl);
/* setup the group */
bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
mlog(0, "new descriptor, record %u, at block %llu\n",
alloc_rec, (unsigned long long)bg_blkno);
bg_bh = sb_getblk(osb->sb, bg_blkno);
if (!bg_bh) {
status = -EIO;
mlog_errno(status);
goto bail;
}
ocfs2_set_new_buffer_uptodate(alloc_inode, bg_bh);
status = ocfs2_block_group_fill(handle,
alloc_inode,
bg_bh,
bg_blkno,
alloc_rec,
cl);
if (status < 0) {
mlog_errno(status);
goto bail;
}
bg = (struct ocfs2_group_desc *) bg_bh->b_data;
status = ocfs2_journal_access(handle, alloc_inode,
bh, OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
mlog_errno(status);
goto bail;
}
le32_add_cpu(&cl->cl_recs[alloc_rec].c_free,
le16_to_cpu(bg->bg_free_bits_count));
le32_add_cpu(&cl->cl_recs[alloc_rec].c_total, le16_to_cpu(bg->bg_bits));
cl->cl_recs[alloc_rec].c_blkno = cpu_to_le64(bg_blkno);
if (le16_to_cpu(cl->cl_next_free_rec) < le16_to_cpu(cl->cl_count))
le16_add_cpu(&cl->cl_next_free_rec, 1);
le32_add_cpu(&fe->id1.bitmap1.i_used, le16_to_cpu(bg->bg_bits) -
le16_to_cpu(bg->bg_free_bits_count));
le32_add_cpu(&fe->id1.bitmap1.i_total, le16_to_cpu(bg->bg_bits));
le32_add_cpu(&fe->i_clusters, le16_to_cpu(cl->cl_cpg));
status = ocfs2_journal_dirty(handle, bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
spin_lock(&OCFS2_I(alloc_inode)->ip_lock);
OCFS2_I(alloc_inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
fe->i_size = cpu_to_le64(ocfs2_clusters_to_bytes(alloc_inode->i_sb,
le32_to_cpu(fe->i_clusters)));
spin_unlock(&OCFS2_I(alloc_inode)->ip_lock);
i_size_write(alloc_inode, le64_to_cpu(fe->i_size));
alloc_inode->i_blocks = ocfs2_inode_sector_count(alloc_inode);
status = 0;
bail:
if (handle)
ocfs2_commit_trans(osb, handle);
if (ac)
ocfs2_free_alloc_context(ac);
if (bg_bh)
brelse(bg_bh);
mlog_exit(status);
return status;
}
/*
* Using one journal handle to guarantee the data consistency in case
* crash happens anywhere.
*
* XXX: defrag can end up with finishing partial extent as requested,
* due to not enough contiguous clusters can be found in allocator.
*/
static int ocfs2_defrag_extent(struct ocfs2_move_extents_context *context,
u32 cpos, u32 phys_cpos, u32 *len, int ext_flags)
{
int ret, credits = 0, extra_blocks = 0, partial = context->partial;
handle_t *handle;
struct inode *inode = context->inode;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct inode *tl_inode = osb->osb_tl_inode;
struct ocfs2_refcount_tree *ref_tree = NULL;
u32 new_phys_cpos, new_len;
u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
if ((ext_flags & OCFS2_EXT_REFCOUNTED) && *len) {
BUG_ON(!(OCFS2_I(inode)->ip_dyn_features &
OCFS2_HAS_REFCOUNT_FL));
BUG_ON(!context->refcount_loc);
ret = ocfs2_lock_refcount_tree(osb, context->refcount_loc, 1,
&ref_tree, NULL);
if (ret) {
mlog_errno(ret);
return ret;
}
ret = ocfs2_prepare_refcount_change_for_del(inode,
context->refcount_loc,
phys_blkno,
*len,
&credits,
&extra_blocks);
if (ret) {
mlog_errno(ret);
goto out;
}
}
ret = ocfs2_lock_allocators_move_extents(inode, &context->et, *len, 1,
&context->meta_ac,
&context->data_ac,
extra_blocks, &credits);
if (ret) {
mlog_errno(ret);
goto out;
}
/*
* should be using allocation reservation strategy there?
*
* if (context->data_ac)
* context->data_ac->ac_resv = &OCFS2_I(inode)->ip_la_data_resv;
*/
inode_lock(tl_inode);
if (ocfs2_truncate_log_needs_flush(osb)) {
ret = __ocfs2_flush_truncate_log(osb);
if (ret < 0) {
mlog_errno(ret);
goto out_unlock_mutex;
}
}
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
mlog_errno(ret);
goto out_unlock_mutex;
}
ret = __ocfs2_claim_clusters(handle, context->data_ac, 1, *len,
&new_phys_cpos, &new_len);
if (ret) {
mlog_errno(ret);
goto out_commit;
}
/*
* allowing partial extent moving is kind of 'pros and cons', it makes
* whole defragmentation less likely to fail, on the contrary, the bad
* thing is it may make the fs even more fragmented after moving, let
* userspace make a good decision here.
*/
if (new_len != *len) {
mlog(0, "len_claimed: %u, len: %u\n", new_len, *len);
if (!partial) {
context->range->me_flags &= ~OCFS2_MOVE_EXT_FL_COMPLETE;
ret = -ENOSPC;
goto out_commit;
}
}
mlog(0, "cpos: %u, phys_cpos: %u, new_phys_cpos: %u\n", cpos,
phys_cpos, new_phys_cpos);
ret = __ocfs2_move_extent(handle, context, cpos, new_len, phys_cpos,
new_phys_cpos, ext_flags);
if (ret)
mlog_errno(ret);
if (partial && (new_len != *len))
*len = new_len;
/*
* Here we should write the new page out first if we are
* in write-back mode.
*/
ret = ocfs2_cow_sync_writeback(inode->i_sb, context->inode, cpos, *len);
if (ret)
mlog_errno(ret);
out_commit:
ocfs2_commit_trans(osb, handle);
out_unlock_mutex:
inode_unlock(tl_inode);
if (context->data_ac) {
ocfs2_free_alloc_context(context->data_ac);
context->data_ac = NULL;
}
if (context->meta_ac) {
ocfs2_free_alloc_context(context->meta_ac);
context->meta_ac = NULL;
}
out:
if (ref_tree)
ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
return ret;
}
static int ocfs2_defrag_extent(struct ocfs2_move_extents_context *context,
u32 cpos, u32 phys_cpos, u32 *len, int ext_flags)
{
int ret, credits = 0, extra_blocks = 0, partial = context->partial;
handle_t *handle;
struct inode *inode = context->inode;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct inode *tl_inode = osb->osb_tl_inode;
struct ocfs2_refcount_tree *ref_tree = NULL;
u32 new_phys_cpos, new_len;
u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
if ((ext_flags & OCFS2_EXT_REFCOUNTED) && *len) {
BUG_ON(!(OCFS2_I(inode)->ip_dyn_features &
OCFS2_HAS_REFCOUNT_FL));
BUG_ON(!context->refcount_loc);
ret = ocfs2_lock_refcount_tree(osb, context->refcount_loc, 1,
&ref_tree, NULL);
if (ret) {
mlog_errno(ret);
return ret;
}
ret = ocfs2_prepare_refcount_change_for_del(inode,
context->refcount_loc,
phys_blkno,
*len,
&credits,
&extra_blocks);
if (ret) {
mlog_errno(ret);
goto out;
}
}
ret = ocfs2_lock_allocators_move_extents(inode, &context->et, *len, 1,
&context->meta_ac,
&context->data_ac,
extra_blocks, &credits);
if (ret) {
mlog_errno(ret);
goto out;
}
mutex_lock(&tl_inode->i_mutex);
if (ocfs2_truncate_log_needs_flush(osb)) {
ret = __ocfs2_flush_truncate_log(osb);
if (ret < 0) {
mlog_errno(ret);
goto out_unlock_mutex;
}
}
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
mlog_errno(ret);
goto out_unlock_mutex;
}
ret = __ocfs2_claim_clusters(handle, context->data_ac, 1, *len,
&new_phys_cpos, &new_len);
if (ret) {
mlog_errno(ret);
goto out_commit;
}
if (new_len != *len) {
mlog(0, "len_claimed: %u, len: %u\n", new_len, *len);
if (!partial) {
context->range->me_flags &= ~OCFS2_MOVE_EXT_FL_COMPLETE;
ret = -ENOSPC;
goto out_commit;
}
}
mlog(0, "cpos: %u, phys_cpos: %u, new_phys_cpos: %u\n", cpos,
phys_cpos, new_phys_cpos);
ret = __ocfs2_move_extent(handle, context, cpos, new_len, phys_cpos,
new_phys_cpos, ext_flags);
if (ret)
mlog_errno(ret);
if (partial && (new_len != *len))
*len = new_len;
ret = ocfs2_cow_sync_writeback(inode->i_sb, context->inode, cpos, *len);
if (ret)
mlog_errno(ret);
out_commit:
ocfs2_commit_trans(osb, handle);
out_unlock_mutex:
mutex_unlock(&tl_inode->i_mutex);
if (context->data_ac) {
ocfs2_free_alloc_context(context->data_ac);
context->data_ac = NULL;
}
if (context->meta_ac) {
ocfs2_free_alloc_context(context->meta_ac);
context->meta_ac = NULL;
}
out:
if (ref_tree)
ocfs2_unlock_refcount_tree(osb, ref_tree, 1);
return ret;
}
