#include "buffer_head_io.h"
+#define NOT_ALLOC_NEW_GROUP 0
+#define ALLOC_NEW_GROUP 1
+
+#define OCFS2_MAX_INODES_TO_STEAL 1024
+
static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg);
static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe);
static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl);
static inline u32 ocfs2_desc_bitmap_to_cluster_off(struct inode *inode,
u64 bg_blkno,
u16 bg_bit_off);
-static inline u64 ocfs2_which_cluster_group(struct inode *inode,
- u32 cluster);
static inline void ocfs2_block_to_cluster_group(struct inode *inode,
u64 data_blkno,
u64 *bg_blkno,
u16 *bg_bit_off);
-void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac)
+void ocfs2_free_ac_resource(struct ocfs2_alloc_context *ac)
{
struct inode *inode = ac->ac_inode;
if (inode) {
if (ac->ac_which != OCFS2_AC_USE_LOCAL)
- ocfs2_meta_unlock(inode, 1);
+ ocfs2_inode_unlock(inode, 1);
mutex_unlock(&inode->i_mutex);
iput(inode);
+ ac->ac_inode = NULL;
}
- if (ac->ac_bh)
+ if (ac->ac_bh) {
brelse(ac->ac_bh);
+ ac->ac_bh = NULL;
+ }
+}
+
+void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac)
+{
+ ocfs2_free_ac_resource(ac);
kfree(ac);
}
}
/* somewhat more expensive than our other checks, so use sparingly. */
-static int ocfs2_check_group_descriptor(struct super_block *sb,
- struct ocfs2_dinode *di,
- struct ocfs2_group_desc *gd)
+int ocfs2_check_group_descriptor(struct super_block *sb,
+ struct ocfs2_dinode *di,
+ struct ocfs2_group_desc *gd)
{
unsigned int max_bits;
static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_alloc_context *ac,
int type,
- u32 slot)
+ u32 slot,
+ int alloc_new_group)
{
int status;
u32 bits_wanted = ac->ac_bits_wanted;
mutex_lock(&alloc_inode->i_mutex);
- status = ocfs2_meta_lock(alloc_inode, &bh, 1);
+ status = ocfs2_inode_lock(alloc_inode, &bh, 1);
if (status < 0) {
mutex_unlock(&alloc_inode->i_mutex);
iput(alloc_inode);
}
ac->ac_inode = alloc_inode;
+ ac->ac_alloc_slot = slot;
fe = (struct ocfs2_dinode *) bh->b_data;
if (!OCFS2_IS_VALID_DINODE(fe)) {
goto bail;
}
+ if (alloc_new_group != ALLOC_NEW_GROUP) {
+ mlog(0, "Alloc File %u Full: wanted=%u, free_bits=%u, "
+ "and we don't alloc a new group for it.\n",
+ slot, bits_wanted, free_bits);
+ status = -ENOSPC;
+ goto bail;
+ }
+
status = ocfs2_block_group_alloc(osb, alloc_inode, bh);
if (status < 0) {
if (status != -ENOSPC)
return status;
}
-int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
- struct ocfs2_dinode *fe,
- struct ocfs2_alloc_context **ac)
+int ocfs2_reserve_new_metadata_blocks(struct ocfs2_super *osb,
+ int blocks,
+ struct ocfs2_alloc_context **ac)
{
int status;
u32 slot;
goto bail;
}
- (*ac)->ac_bits_wanted = ocfs2_extend_meta_needed(fe);
+ (*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);
+ EXTENT_ALLOC_SYSTEM_INODE,
+ slot, ALLOC_NEW_GROUP);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
return status;
}
+int ocfs2_reserve_new_metadata(struct ocfs2_super *osb,
+ struct ocfs2_extent_list *root_el,
+ struct ocfs2_alloc_context **ac)
+{
+ return ocfs2_reserve_new_metadata_blocks(osb,
+ ocfs2_extend_meta_needed(root_el),
+ ac);
+}
+
+static int ocfs2_steal_inode_from_other_nodes(struct ocfs2_super *osb,
+ struct ocfs2_alloc_context *ac)
+{
+ int i, status = -ENOSPC;
+ s16 slot = ocfs2_get_inode_steal_slot(osb);
+
+ /* Start to steal inodes from the first slot after ours. */
+ if (slot == OCFS2_INVALID_SLOT)
+ slot = osb->slot_num + 1;
+
+ for (i = 0; i < osb->max_slots; i++, slot++) {
+ if (slot == osb->max_slots)
+ slot = 0;
+
+ if (slot == osb->slot_num)
+ continue;
+
+ status = ocfs2_reserve_suballoc_bits(osb, ac,
+ INODE_ALLOC_SYSTEM_INODE,
+ slot, NOT_ALLOC_NEW_GROUP);
+ if (status >= 0) {
+ ocfs2_set_inode_steal_slot(osb, slot);
+ break;
+ }
+
+ ocfs2_free_ac_resource(ac);
+ }
+
+ return status;
+}
+
int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
struct ocfs2_alloc_context **ac)
{
int status;
+ s16 slot = ocfs2_get_inode_steal_slot(osb);
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
(*ac)->ac_group_search = ocfs2_block_group_search;
+ /*
+ * 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_INODES_TO_STEAL)
+ goto inode_steal;
+
+ atomic_set(&osb->s_num_inodes_stolen, 0);
status = ocfs2_reserve_suballoc_bits(osb, *ac,
INODE_ALLOC_SYSTEM_INODE,
- osb->slot_num);
+ osb->slot_num, ALLOC_NEW_GROUP);
+ if (status >= 0) {
+ status = 0;
+
+ /*
+ * 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_from_other_nodes(osb, *ac);
+ atomic_inc(&osb->s_num_inodes_stolen);
if (status < 0) {
if (status != -ENOSPC)
mlog_errno(status);
status = ocfs2_reserve_suballoc_bits(osb, ac,
GLOBAL_BITMAP_SYSTEM_INODE,
- OCFS2_INVALID_SLOT);
+ OCFS2_INVALID_SLOT,
+ ALLOC_NEW_GROUP);
if (status < 0 && status != -ENOSPC) {
mlog_errno(status);
goto bail;
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;
}
}
* sync-data inodes."
*
* Note: OCFS2 already does this differently for metadata vs data
- * allocations, as those bitmaps are seperate and undo access is never
+ * allocations, as those bitmaps are separate and undo access is never
* called on a metadata group descriptor.
*/
static int ocfs2_test_bg_bit_allocatable(struct buffer_head *bg_bh,
int search = -ENOSPC;
int ret;
struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *) group_bh->b_data;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
u16 tmp_off, tmp_found;
unsigned int max_bits, gd_cluster_off;
*bit_off = tmp_off;
*bits_found = tmp_found;
search = 0; /* success */
+ } else if (tmp_found) {
+ /*
+ * Don't show bits which we'll be returning
+ * for allocation to the local alloc bitmap.
+ */
+ ocfs2_local_alloc_seen_free_bits(osb, tmp_found);
}
}
status = -ENOSPC;
/* for now, the chain search is a bit simplistic. We just use
* the 1st group with any empty bits. */
- while ((status = ac->ac_group_search(alloc_inode, group_bh,
- bits_wanted, min_bits, bit_off,
- &tmp_bits)) == -ENOSPC) {
+ while ((status = ac->ac_group_search(alloc_inode, group_bh, bits_wanted,
+ min_bits, bit_off, &tmp_bits)) == -ENOSPC) {
if (!bg->bg_next_group)
break;
/* given a cluster offset, calculate which block group it belongs to
* and return that block offset. */
-static inline u64 ocfs2_which_cluster_group(struct inode *inode,
- u32 cluster)
+u64 ocfs2_which_cluster_group(struct inode *inode, u32 cluster)
{
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
u32 group_no;
if (min_clusters > (osb->bitmap_cpg - 1)) {
/* The only paths asking for contiguousness
* should know about this already. */
- mlog(ML_ERROR, "minimum allocation requested exceeds "
- "group bitmap size!");
+ mlog(ML_ERROR, "minimum allocation requested %u exceeds "
+ "group bitmap size %u!\n", min_clusters,
+ osb->bitmap_cpg);
status = -ENOSPC;
goto bail;
}
status = ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
bg_start_bit, bg_blkno,
num_clusters);
- if (status < 0)
+ if (status < 0) {
mlog_errno(status);
+ goto out;
+ }
+
+ ocfs2_local_alloc_seen_free_bits(OCFS2_SB(bitmap_inode->i_sb),
+ num_clusters);
+out:
mlog_exit(status);
return status;
}
(unsigned long long)fe->id2.i_chain.cl_recs[i].c_blkno);
}
}
+
+/*
+ * For a given allocation, determine which allocators will need to be
+ * accessed, and lock them, reserving the appropriate number of bits.
+ *
+ * Sparse file systems call this from ocfs2_write_begin_nolock()
+ * and ocfs2_allocate_unwritten_extents().
+ *
+ * File systems which don't support holes call this from
+ * ocfs2_extend_allocation().
+ */
+int ocfs2_lock_allocators(struct inode *inode, struct buffer_head *root_bh,
+ struct ocfs2_extent_list *root_el,
+ u32 clusters_to_add, u32 extents_to_split,
+ struct ocfs2_alloc_context **data_ac,
+ struct ocfs2_alloc_context **meta_ac,
+ enum ocfs2_extent_tree_type type, void *private)
+{
+ int ret = 0, num_free_extents;
+ unsigned int max_recs_needed = clusters_to_add + 2 * extents_to_split;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+ *meta_ac = NULL;
+ if (data_ac)
+ *data_ac = NULL;
+
+ BUG_ON(clusters_to_add != 0 && data_ac == NULL);
+
+ num_free_extents = ocfs2_num_free_extents(osb, inode, root_bh,
+ type, private);
+ 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.
+ *
+ * Always lock for any unwritten extents - we might want to
+ * add blocks during a split.
+ */
+ if (!num_free_extents ||
+ (ocfs2_sparse_alloc(osb) && num_free_extents < max_recs_needed)) {
+ ret = ocfs2_reserve_new_metadata(osb, root_el, meta_ac);
+ if (ret < 0) {
+ if (ret != -ENOSPC)
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ if (clusters_to_add == 0)
+ 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;
+}