b2 = (unsigned char *) bitmap;
for (i = 0; i < e4b->bd_sb->s_blocksize; i++) {
if (b1[i] != b2[i]) {
- printk("corruption in group %lu at byte %u(%u):"
- " %x in copy != %x on disk/prealloc\n",
- e4b->bd_group, i, i * 8, b1[i], b2[i]);
+ printk(KERN_ERR "corruption in group %lu "
+ "at byte %u(%u): %x in copy != %x "
+ "on disk/prealloc\n",
+ e4b->bd_group, i, i * 8, b1[i], b2[i]);
BUG();
}
}
if (bh_uptodate_or_lock(bh[i]))
continue;
+ spin_lock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
ext4_init_block_bitmap(sb, bh[i],
first_group + i, desc);
set_buffer_uptodate(bh[i]);
unlock_buffer(bh[i]);
+ spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
continue;
}
+ spin_unlock(sb_bgl_lock(EXT4_SB(sb), first_group + i));
get_bh(bh[i]);
bh[i]->b_end_io = end_buffer_read_sync;
submit_bh(READ, bh[i]);
int ext4_mb_init(struct super_block *sb, int needs_recovery)
{
struct ext4_sb_info *sbi = EXT4_SB(sb);
- unsigned i;
+ unsigned i, j;
unsigned offset;
unsigned max;
int ret;
sbi->s_mb_history_filter = EXT4_MB_HISTORY_DEFAULT;
sbi->s_mb_group_prealloc = MB_DEFAULT_GROUP_PREALLOC;
- i = sizeof(struct ext4_locality_group) * NR_CPUS;
+ i = sizeof(struct ext4_locality_group) * nr_cpu_ids;
sbi->s_locality_groups = kmalloc(i, GFP_KERNEL);
if (sbi->s_locality_groups == NULL) {
clear_opt(sbi->s_mount_opt, MBALLOC);
kfree(sbi->s_mb_maxs);
return -ENOMEM;
}
- for (i = 0; i < NR_CPUS; i++) {
+ for (i = 0; i < nr_cpu_ids; i++) {
struct ext4_locality_group *lg;
lg = &sbi->s_locality_groups[i];
mutex_init(&lg->lg_mutex);
- INIT_LIST_HEAD(&lg->lg_prealloc_list);
+ for (j = 0; j < PREALLOC_TB_SIZE; j++)
+ INIT_LIST_HEAD(&lg->lg_prealloc_list[j]);
spin_lock_init(&lg->lg_prealloc_lock);
}
ext4_mb_init_per_dev_proc(sb);
ext4_mb_history_init(sb);
- printk("EXT4-fs: mballoc enabled\n");
+ printk(KERN_INFO "EXT4-fs: mballoc enabled\n");
return 0;
}
*/
static noinline_for_stack int
ext4_mb_mark_diskspace_used(struct ext4_allocation_context *ac,
- handle_t *handle)
+ handle_t *handle, unsigned long reserv_blks)
{
struct buffer_head *bitmap_bh = NULL;
struct ext4_super_block *es;
le16_add_cpu(&gdp->bg_free_blocks_count, -ac->ac_b_ex.fe_len);
gdp->bg_checksum = ext4_group_desc_csum(sbi, ac->ac_b_ex.fe_group, gdp);
spin_unlock(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group));
-
+ percpu_counter_sub(&sbi->s_freeblocks_counter, ac->ac_b_ex.fe_len);
/*
- * free blocks account has already be reduced/reserved
- * at write_begin() time for delayed allocation
- * do not double accounting
+ * Now reduce the dirty block count also. Should not go negative
*/
if (!(ac->ac_flags & EXT4_MB_DELALLOC_RESERVED))
- percpu_counter_sub(&sbi->s_freeblocks_counter,
- ac->ac_b_ex.fe_len);
+ /* release all the reserved blocks if non delalloc */
+ percpu_counter_sub(&sbi->s_dirtyblocks_counter, reserv_blks);
+ else
+ percpu_counter_sub(&sbi->s_dirtyblocks_counter,
+ ac->ac_b_ex.fe_len);
if (sbi->s_log_groups_per_flex) {
ext4_group_t flex_group = ext4_flex_group(sbi,
struct ext4_prealloc_space *pa)
{
unsigned int len = ac->ac_o_ex.fe_len;
+
ext4_get_group_no_and_offset(ac->ac_sb, pa->pa_pstart,
&ac->ac_b_ex.fe_group,
&ac->ac_b_ex.fe_start);
mb_debug("use %u/%u from group pa %p\n", pa->pa_lstart-len, len, pa);
}
+/*
+ * Return the prealloc space that have minimal distance
+ * from the goal block. @cpa is the prealloc
+ * space that is having currently known minimal distance
+ * from the goal block.
+ */
+static struct ext4_prealloc_space *
+ext4_mb_check_group_pa(ext4_fsblk_t goal_block,
+ struct ext4_prealloc_space *pa,
+ struct ext4_prealloc_space *cpa)
+{
+ ext4_fsblk_t cur_distance, new_distance;
+
+ if (cpa == NULL) {
+ atomic_inc(&pa->pa_count);
+ return pa;
+ }
+ cur_distance = abs(goal_block - cpa->pa_pstart);
+ new_distance = abs(goal_block - pa->pa_pstart);
+
+ if (cur_distance < new_distance)
+ return cpa;
+
+ /* drop the previous reference */
+ atomic_dec(&cpa->pa_count);
+ atomic_inc(&pa->pa_count);
+ return pa;
+}
+
/*
* search goal blocks in preallocated space
*/
static noinline_for_stack int
ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
{
+ int order, i;
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
struct ext4_locality_group *lg;
- struct ext4_prealloc_space *pa;
+ struct ext4_prealloc_space *pa, *cpa = NULL;
+ ext4_fsblk_t goal_block;
/* only data can be preallocated */
if (!(ac->ac_flags & EXT4_MB_HINT_DATA))
lg = ac->ac_lg;
if (lg == NULL)
return 0;
+ order = fls(ac->ac_o_ex.fe_len) - 1;
+ if (order > PREALLOC_TB_SIZE - 1)
+ /* The max size of hash table is PREALLOC_TB_SIZE */
+ order = PREALLOC_TB_SIZE - 1;
+
+ goal_block = ac->ac_g_ex.fe_group * EXT4_BLOCKS_PER_GROUP(ac->ac_sb) +
+ ac->ac_g_ex.fe_start +
+ le32_to_cpu(EXT4_SB(ac->ac_sb)->s_es->s_first_data_block);
+ /*
+ * search for the prealloc space that is having
+ * minimal distance from the goal block.
+ */
+ for (i = order; i < PREALLOC_TB_SIZE; i++) {
+ rcu_read_lock();
+ list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[i],
+ pa_inode_list) {
+ spin_lock(&pa->pa_lock);
+ if (pa->pa_deleted == 0 &&
+ pa->pa_free >= ac->ac_o_ex.fe_len) {
- rcu_read_lock();
- list_for_each_entry_rcu(pa, &lg->lg_prealloc_list, pa_inode_list) {
- spin_lock(&pa->pa_lock);
- if (pa->pa_deleted == 0 && pa->pa_free >= ac->ac_o_ex.fe_len) {
- atomic_inc(&pa->pa_count);
- ext4_mb_use_group_pa(ac, pa);
+ cpa = ext4_mb_check_group_pa(goal_block,
+ pa, cpa);
+ }
spin_unlock(&pa->pa_lock);
- ac->ac_criteria = 20;
- rcu_read_unlock();
- return 1;
}
- spin_unlock(&pa->pa_lock);
+ rcu_read_unlock();
+ }
+ if (cpa) {
+ ext4_mb_use_group_pa(ac, cpa);
+ ac->ac_criteria = 20;
+ return 1;
}
- rcu_read_unlock();
-
return 0;
}
pa->pa_free = pa->pa_len;
atomic_set(&pa->pa_count, 1);
spin_lock_init(&pa->pa_lock);
+ INIT_LIST_HEAD(&pa->pa_inode_list);
pa->pa_deleted = 0;
pa->pa_linear = 1;
list_add(&pa->pa_group_list, &grp->bb_prealloc_list);
ext4_unlock_group(sb, ac->ac_b_ex.fe_group);
- spin_lock(pa->pa_obj_lock);
- list_add_tail_rcu(&pa->pa_inode_list, &lg->lg_prealloc_list);
- spin_unlock(pa->pa_obj_lock);
-
+ /*
+ * We will later add the new pa to the right bucket
+ * after updating the pa_free in ext4_mb_release_context
+ */
return 0;
}
bitmap_bh = ext4_read_block_bitmap(sb, group);
if (bitmap_bh == NULL) {
- /* error handling here */
- ext4_mb_release_desc(&e4b);
- BUG_ON(bitmap_bh == NULL);
+ ext4_error(sb, __func__, "Error in reading block "
+ "bitmap for %lu\n", group);
+ return 0;
}
err = ext4_mb_load_buddy(sb, group, &e4b);
- BUG_ON(err != 0); /* error handling here */
+ if (err) {
+ ext4_error(sb, __func__, "Error in loading buddy "
+ "information for %lu\n", group);
+ put_bh(bitmap_bh);
+ return 0;
+ }
if (needed == 0)
needed = EXT4_BLOCKS_PER_GROUP(sb) + 1;
- grp = ext4_get_group_info(sb, group);
INIT_LIST_HEAD(&list);
-
ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS);
repeat:
ext4_lock_group(sb, group);
ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, NULL);
err = ext4_mb_load_buddy(sb, group, &e4b);
- BUG_ON(err != 0); /* error handling here */
+ if (err) {
+ ext4_error(sb, __func__, "Error in loading buddy "
+ "information for %lu\n", group);
+ continue;
+ }
bitmap_bh = ext4_read_block_bitmap(sb, group);
if (bitmap_bh == NULL) {
- /* error handling here */
+ ext4_error(sb, __func__, "Error in reading block "
+ "bitmap for %lu\n", group);
ext4_mb_release_desc(&e4b);
- BUG_ON(bitmap_bh == NULL);
+ continue;
}
ext4_lock_group(sb, group);
}
+static noinline_for_stack void
+ext4_mb_discard_lg_preallocations(struct super_block *sb,
+ struct ext4_locality_group *lg,
+ int order, int total_entries)
+{
+ ext4_group_t group = 0;
+ struct ext4_buddy e4b;
+ struct list_head discard_list;
+ struct ext4_prealloc_space *pa, *tmp;
+ struct ext4_allocation_context *ac;
+
+ mb_debug("discard locality group preallocation\n");
+
+ INIT_LIST_HEAD(&discard_list);
+ ac = kmem_cache_alloc(ext4_ac_cachep, GFP_NOFS);
+
+ spin_lock(&lg->lg_prealloc_lock);
+ list_for_each_entry_rcu(pa, &lg->lg_prealloc_list[order],
+ pa_inode_list) {
+ spin_lock(&pa->pa_lock);
+ if (atomic_read(&pa->pa_count)) {
+ /*
+ * This is the pa that we just used
+ * for block allocation. So don't
+ * free that
+ */
+ spin_unlock(&pa->pa_lock);
+ continue;
+ }
+ if (pa->pa_deleted) {
+ spin_unlock(&pa->pa_lock);
+ continue;
+ }
+ /* only lg prealloc space */
+ BUG_ON(!pa->pa_linear);
+
+ /* seems this one can be freed ... */
+ pa->pa_deleted = 1;
+ spin_unlock(&pa->pa_lock);
+
+ list_del_rcu(&pa->pa_inode_list);
+ list_add(&pa->u.pa_tmp_list, &discard_list);
+
+ total_entries--;
+ if (total_entries <= 5) {
+ /*
+ * we want to keep only 5 entries
+ * allowing it to grow to 8. This
+ * mak sure we don't call discard
+ * soon for this list.
+ */
+ break;
+ }
+ }
+ spin_unlock(&lg->lg_prealloc_lock);
+
+ list_for_each_entry_safe(pa, tmp, &discard_list, u.pa_tmp_list) {
+
+ ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, NULL);
+ if (ext4_mb_load_buddy(sb, group, &e4b)) {
+ ext4_error(sb, __func__, "Error in loading buddy "
+ "information for %lu\n", group);
+ continue;
+ }
+ ext4_lock_group(sb, group);
+ list_del(&pa->pa_group_list);
+ ext4_mb_release_group_pa(&e4b, pa, ac);
+ ext4_unlock_group(sb, group);
+
+ ext4_mb_release_desc(&e4b);
+ list_del(&pa->u.pa_tmp_list);
+ call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback);
+ }
+ if (ac)
+ kmem_cache_free(ext4_ac_cachep, ac);
+}
+
+/*
+ * We have incremented pa_count. So it cannot be freed at this
+ * point. Also we hold lg_mutex. So no parallel allocation is
+ * possible from this lg. That means pa_free cannot be updated.
+ *
+ * A parallel ext4_mb_discard_group_preallocations is possible.
+ * which can cause the lg_prealloc_list to be updated.
+ */
+
+static void ext4_mb_add_n_trim(struct ext4_allocation_context *ac)
+{
+ int order, added = 0, lg_prealloc_count = 1;
+ struct super_block *sb = ac->ac_sb;
+ struct ext4_locality_group *lg = ac->ac_lg;
+ struct ext4_prealloc_space *tmp_pa, *pa = ac->ac_pa;
+
+ order = fls(pa->pa_free) - 1;
+ if (order > PREALLOC_TB_SIZE - 1)
+ /* The max size of hash table is PREALLOC_TB_SIZE */
+ order = PREALLOC_TB_SIZE - 1;
+ /* Add the prealloc space to lg */
+ rcu_read_lock();
+ list_for_each_entry_rcu(tmp_pa, &lg->lg_prealloc_list[order],
+ pa_inode_list) {
+ spin_lock(&tmp_pa->pa_lock);
+ if (tmp_pa->pa_deleted) {
+ spin_unlock(&pa->pa_lock);
+ continue;
+ }
+ if (!added && pa->pa_free < tmp_pa->pa_free) {
+ /* Add to the tail of the previous entry */
+ list_add_tail_rcu(&pa->pa_inode_list,
+ &tmp_pa->pa_inode_list);
+ added = 1;
+ /*
+ * we want to count the total
+ * number of entries in the list
+ */
+ }
+ spin_unlock(&tmp_pa->pa_lock);
+ lg_prealloc_count++;
+ }
+ if (!added)
+ list_add_tail_rcu(&pa->pa_inode_list,
+ &lg->lg_prealloc_list[order]);
+ rcu_read_unlock();
+
+ /* Now trim the list to be not more than 8 elements */
+ if (lg_prealloc_count > 8) {
+ ext4_mb_discard_lg_preallocations(sb, lg,
+ order, lg_prealloc_count);
+ return;
+ }
+ return ;
+}
+
/*
* release all resource we used in allocation
*/
static int ext4_mb_release_context(struct ext4_allocation_context *ac)
{
- if (ac->ac_pa) {
- if (ac->ac_pa->pa_linear) {
+ struct ext4_prealloc_space *pa = ac->ac_pa;
+ if (pa) {
+ if (pa->pa_linear) {
/* see comment in ext4_mb_use_group_pa() */
- spin_lock(&ac->ac_pa->pa_lock);
- ac->ac_pa->pa_pstart += ac->ac_b_ex.fe_len;
- ac->ac_pa->pa_lstart += ac->ac_b_ex.fe_len;
- ac->ac_pa->pa_free -= ac->ac_b_ex.fe_len;
- ac->ac_pa->pa_len -= ac->ac_b_ex.fe_len;
- spin_unlock(&ac->ac_pa->pa_lock);
+ spin_lock(&pa->pa_lock);
+ pa->pa_pstart += ac->ac_b_ex.fe_len;
+ pa->pa_lstart += ac->ac_b_ex.fe_len;
+ pa->pa_free -= ac->ac_b_ex.fe_len;
+ pa->pa_len -= ac->ac_b_ex.fe_len;
+ spin_unlock(&pa->pa_lock);
+ /*
+ * We want to add the pa to the right bucket.
+ * Remove it from the list and while adding
+ * make sure the list to which we are adding
+ * doesn't grow big.
+ */
+ if (likely(pa->pa_free)) {
+ spin_lock(pa->pa_obj_lock);
+ list_del_rcu(&pa->pa_inode_list);
+ spin_unlock(pa->pa_obj_lock);
+ ext4_mb_add_n_trim(ac);
+ }
}
- ext4_mb_put_pa(ac, ac->ac_sb, ac->ac_pa);
+ ext4_mb_put_pa(ac, ac->ac_sb, pa);
}
if (ac->ac_bitmap_page)
page_cache_release(ac->ac_bitmap_page);
ext4_fsblk_t ext4_mb_new_blocks(handle_t *handle,
struct ext4_allocation_request *ar, int *errp)
{
+ int freed;
struct ext4_allocation_context *ac = NULL;
struct ext4_sb_info *sbi;
struct super_block *sb;
ext4_fsblk_t block = 0;
- int freed;
- int inquota;
+ unsigned long inquota;
+ unsigned long reserv_blks = 0;
sb = ar->inode->i_sb;
sbi = EXT4_SB(sb);
/*
* With delalloc we already reserved the blocks
*/
- ar->len = ext4_has_free_blocks(sbi, ar->len);
- }
-
- if (ar->len == 0) {
- *errp = -ENOSPC;
- return 0;
+ while (ar->len && ext4_claim_free_blocks(sbi, ar->len)) {
+ /* let others to free the space */
+ yield();
+ ar->len = ar->len >> 1;
+ }
+ if (!ar->len) {
+ *errp = -ENOSPC;
+ return 0;
+ }
+ reserv_blks = ar->len;
}
-
while (ar->len && DQUOT_ALLOC_BLOCK(ar->inode, ar->len)) {
ar->flags |= EXT4_MB_HINT_NOPREALLOC;
ar->len--;
}
if (likely(ac->ac_status == AC_STATUS_FOUND)) {
- *errp = ext4_mb_mark_diskspace_used(ac, handle);
+ *errp = ext4_mb_mark_diskspace_used(ac, handle, reserv_blks);
if (*errp == -EAGAIN) {
ac->ac_b_ex.fe_group = 0;
ac->ac_b_ex.fe_start = 0;
count -= overflow;
}
bitmap_bh = ext4_read_block_bitmap(sb, block_group);
- if (!bitmap_bh)
+ if (!bitmap_bh) {
+ err = -EIO;
goto error_return;
+ }
gdp = ext4_get_group_desc(sb, block_group, &gd_bh);
- if (!gdp)
+ if (!gdp) {
+ err = -EIO;
goto error_return;
+ }
if (in_range(ext4_block_bitmap(sb, gdp), block, count) ||
in_range(ext4_inode_bitmap(sb, gdp), block, count) ||