blocknr = ext4_group_first_block_no(sb, e4b->bd_group);
blocknr += first + i;
ext4_grp_locked_error(sb, e4b->bd_group,
- __func__, "double-free of inode"
- " %lu's block %llu(bit %u in group %u)",
- inode ? inode->i_ino : 0, blocknr,
- first + i, e4b->bd_group);
+ inode ? inode->i_ino : 0,
+ blocknr,
+ "freeing block already freed "
+ "(bit %u)",
+ first + i);
}
mb_clear_bit(first + i, e4b->bd_info->bb_bitmap);
}
}
}
+/*
+ * Cache the order of the largest free extent we have available in this block
+ * group.
+ */
+static void
+mb_set_largest_free_order(struct super_block *sb, struct ext4_group_info *grp)
+{
+ int i;
+ int bits;
+
+ grp->bb_largest_free_order = -1; /* uninit */
+
+ bits = sb->s_blocksize_bits + 1;
+ for (i = bits; i >= 0; i--) {
+ if (grp->bb_counters[i] > 0) {
+ grp->bb_largest_free_order = i;
+ break;
+ }
+ }
+}
+
static noinline_for_stack
void ext4_mb_generate_buddy(struct super_block *sb,
void *buddy, void *bitmap, ext4_group_t group)
grp->bb_fragments = fragments;
if (free != grp->bb_free) {
- ext4_grp_locked_error(sb, group, __func__,
- "EXT4-fs: group %u: %u blocks in bitmap, %u in gd",
- group, free, grp->bb_free);
+ ext4_grp_locked_error(sb, group, 0, 0,
+ "%u blocks in bitmap, %u in gd",
+ free, grp->bb_free);
/*
* If we intent to continue, we consider group descritor
* corrupt and update bb_free using bitmap value
*/
grp->bb_free = free;
}
+ mb_set_largest_free_order(sb, grp);
clear_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
* contain blocks_per_page (PAGE_CACHE_SIZE / blocksize) blocks.
* So it can have information regarding groups_per_page which
* is blocks_per_page/2
+ *
+ * Locking note: This routine takes the block group lock of all groups
+ * for this page; do not hold this lock when calling this routine!
*/
static int ext4_mb_init_cache(struct page *page, char *incore)
BUG_ON(incore == NULL);
mb_debug(1, "put buddy for group %u in page %lu/%x\n",
group, page->index, i * blocksize);
+ trace_ext4_mb_buddy_bitmap_load(sb, group);
grinfo = ext4_get_group_info(sb, group);
grinfo->bb_fragments = 0;
memset(grinfo->bb_counters, 0,
BUG_ON(incore != NULL);
mb_debug(1, "put bitmap for group %u in page %lu/%x\n",
group, page->index, i * blocksize);
+ trace_ext4_mb_bitmap_load(sb, group);
/* see comments in ext4_mb_put_pa() */
ext4_lock_group(sb, group);
return err;
}
+/*
+ * Locking note: This routine calls ext4_mb_init_cache(), which takes the
+ * block group lock of all groups for this page; do not hold the BG lock when
+ * calling this routine!
+ */
static noinline_for_stack
int ext4_mb_init_group(struct super_block *sb, ext4_group_t group)
{
return ret;
}
+/*
+ * Locking note: This routine calls ext4_mb_init_cache(), which takes the
+ * block group lock of all groups for this page; do not hold the BG lock when
+ * calling this routine!
+ */
static noinline_for_stack int
ext4_mb_load_buddy(struct super_block *sb, ext4_group_t group,
struct ext4_buddy *e4b)
return ret;
}
-static void ext4_mb_release_desc(struct ext4_buddy *e4b)
+static void ext4_mb_unload_buddy(struct ext4_buddy *e4b)
{
if (e4b->bd_bitmap_page)
page_cache_release(e4b->bd_bitmap_page);
blocknr = ext4_group_first_block_no(sb, e4b->bd_group);
blocknr += block;
ext4_grp_locked_error(sb, e4b->bd_group,
- __func__, "double-free of inode"
- " %lu's block %llu(bit %u in group %u)",
- inode ? inode->i_ino : 0, blocknr, block,
- e4b->bd_group);
+ inode ? inode->i_ino : 0,
+ blocknr,
+ "freeing already freed block "
+ "(bit %u)", block);
}
mb_clear_bit(block, EXT4_MB_BITMAP(e4b));
e4b->bd_info->bb_counters[order]++;
buddy = buddy2;
} while (1);
}
+ mb_set_largest_free_order(sb, e4b->bd_info);
mb_check_buddy(e4b);
}
e4b->bd_info->bb_counters[ord]++;
e4b->bd_info->bb_counters[ord]++;
}
+ mb_set_largest_free_order(e4b->bd_sb, e4b->bd_info);
mb_set_bits(EXT4_MB_BITMAP(e4b), ex->fe_start, len0);
mb_check_buddy(e4b);
}
ext4_unlock_group(ac->ac_sb, group);
- ext4_mb_release_desc(e4b);
+ ext4_mb_unload_buddy(e4b);
return 0;
}
ext4_mb_use_best_found(ac, e4b);
}
ext4_unlock_group(ac->ac_sb, group);
- ext4_mb_release_desc(e4b);
+ ext4_mb_unload_buddy(e4b);
return 0;
}
* free blocks even though group info says we
* we have free blocks
*/
- ext4_grp_locked_error(sb, e4b->bd_group,
- __func__, "%d free blocks as per "
+ ext4_grp_locked_error(sb, e4b->bd_group, 0, 0,
+ "%d free blocks as per "
"group info. But bitmap says 0",
free);
break;
mb_find_extent(e4b, 0, i, ac->ac_g_ex.fe_len, &ex);
BUG_ON(ex.fe_len <= 0);
if (free < ex.fe_len) {
- ext4_grp_locked_error(sb, e4b->bd_group,
- __func__, "%d free blocks as per "
+ ext4_grp_locked_error(sb, e4b->bd_group, 0, 0,
+ "%d free blocks as per "
"group info. But got %d blocks",
free, ex.fe_len);
/*
/*
* This is a special case for storages like raid5
- * we try to find stripe-aligned chunks for stripe-size requests
- * XXX should do so at least for multiples of stripe size as well
+ * we try to find stripe-aligned chunks for stripe-size-multiple requests
*/
static noinline_for_stack
void ext4_mb_scan_aligned(struct ext4_allocation_context *ac,
}
}
+/* This is now called BEFORE we load the buddy bitmap. */
static int ext4_mb_good_group(struct ext4_allocation_context *ac,
ext4_group_t group, int cr)
{
unsigned free, fragments;
- unsigned i, bits;
int flex_size = ext4_flex_bg_size(EXT4_SB(ac->ac_sb));
struct ext4_group_info *grp = ext4_get_group_info(ac->ac_sb, group);
BUG_ON(cr < 0 || cr >= 4);
- BUG_ON(EXT4_MB_GRP_NEED_INIT(grp));
+
+ /* We only do this if the grp has never been initialized */
+ if (unlikely(EXT4_MB_GRP_NEED_INIT(grp))) {
+ int ret = ext4_mb_init_group(ac->ac_sb, group);
+ if (ret)
+ return 0;
+ }
free = grp->bb_free;
fragments = grp->bb_fragments;
case 0:
BUG_ON(ac->ac_2order == 0);
+ if (grp->bb_largest_free_order < ac->ac_2order)
+ return 0;
+
/* Avoid using the first bg of a flexgroup for data files */
if ((ac->ac_flags & EXT4_MB_HINT_DATA) &&
(flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) &&
((group % flex_size) == 0))
return 0;
- bits = ac->ac_sb->s_blocksize_bits + 1;
- for (i = ac->ac_2order; i <= bits; i++)
- if (grp->bb_counters[i] > 0)
- return 1;
- break;
+ return 1;
case 1:
if ((free / fragments) >= ac->ac_g_ex.fe_len)
return 1;
ext4_group_t ngroups, group, i;
int cr;
int err = 0;
- int bsbits;
struct ext4_sb_info *sbi;
struct super_block *sb;
struct ext4_buddy e4b;
sbi = EXT4_SB(sb);
ngroups = ext4_get_groups_count(sb);
/* non-extent files are limited to low blocks/groups */
- if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL))
+ if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)))
ngroups = sbi->s_blockfile_groups;
BUG_ON(ac->ac_status == AC_STATUS_FOUND);
ac->ac_2order = i - 1;
}
- bsbits = ac->ac_sb->s_blocksize_bits;
-
/* if stream allocation is enabled, use global goal */
if (ac->ac_flags & EXT4_MB_STREAM_ALLOC) {
/* TBD: may be hot point */
group = ac->ac_g_ex.fe_group;
for (i = 0; i < ngroups; group++, i++) {
- struct ext4_group_info *grp;
- struct ext4_group_desc *desc;
-
if (group == ngroups)
group = 0;
- /* quick check to skip empty groups */
- grp = ext4_get_group_info(sb, group);
- if (grp->bb_free == 0)
+ /* This now checks without needing the buddy page */
+ if (!ext4_mb_good_group(ac, group, cr))
continue;
err = ext4_mb_load_buddy(sb, group, &e4b);
goto out;
ext4_lock_group(sb, group);
+
+ /*
+ * We need to check again after locking the
+ * block group
+ */
if (!ext4_mb_good_group(ac, group, cr)) {
- /* someone did allocation from this group */
ext4_unlock_group(sb, group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
continue;
}
ac->ac_groups_scanned++;
- desc = ext4_get_group_desc(sb, group, NULL);
if (cr == 0)
ext4_mb_simple_scan_group(ac, &e4b);
- else if (cr == 1 &&
- ac->ac_g_ex.fe_len == sbi->s_stripe)
+ else if (cr == 1 && sbi->s_stripe &&
+ !(ac->ac_g_ex.fe_len % sbi->s_stripe))
ext4_mb_scan_aligned(ac, &e4b);
else
ext4_mb_complex_scan_group(ac, &e4b);
ext4_unlock_group(sb, group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
if (ac->ac_status != AC_STATUS_CONTINUE)
break;
ext4_lock_group(sb, group);
memcpy(&sg, ext4_get_group_info(sb, group), i);
ext4_unlock_group(sb, group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free,
sg.info.bb_fragments, sg.info.bb_first_free);
rc = seq_open(file, &ext4_mb_seq_groups_ops);
if (rc == 0) {
- struct seq_file *m = (struct seq_file *)file->private_data;
+ struct seq_file *m = file->private_data;
m->private = sb;
}
return rc;
INIT_LIST_HEAD(&meta_group_info[i]->bb_prealloc_list);
init_rwsem(&meta_group_info[i]->alloc_sem);
meta_group_info[i]->bb_free_root = RB_ROOT;
+ meta_group_info[i]->bb_largest_free_order = -1; /* uninit */
#ifdef DOUBLE_CHECK
{
return 0;
}
+static inline void ext4_issue_discard(struct super_block *sb,
+ ext4_group_t block_group, ext4_grpblk_t block, int count)
+{
+ int ret;
+ ext4_fsblk_t discard_block;
+
+ discard_block = block + ext4_group_first_block_no(sb, block_group);
+ trace_ext4_discard_blocks(sb,
+ (unsigned long long) discard_block, count);
+ ret = sb_issue_discard(sb, discard_block, count);
+ if (ret == EOPNOTSUPP) {
+ ext4_warning(sb, "discard not supported, disabling");
+ clear_opt(EXT4_SB(sb)->s_mount_opt, DISCARD);
+ }
+}
+
/*
* This function is called by the jbd2 layer once the commit has finished,
* so we know we can free the blocks that were released with that commit.
mb_debug(1, "gonna free %u blocks in group %u (0x%p):",
entry->count, entry->group, entry);
- if (test_opt(sb, DISCARD)) {
- ext4_fsblk_t discard_block;
-
- discard_block = entry->start_blk +
- ext4_group_first_block_no(sb, entry->group);
- trace_ext4_discard_blocks(sb,
- (unsigned long long)discard_block,
- entry->count);
- sb_issue_discard(sb, discard_block, entry->count);
- }
+ if (test_opt(sb, DISCARD))
+ ext4_issue_discard(sb, entry->group,
+ entry->start_blk, entry->count);
err = ext4_mb_load_buddy(sb, entry->group, &e4b);
/* we expect to find existing buddy because it's pinned */
}
ext4_unlock_group(sb, entry->group);
kmem_cache_free(ext4_free_ext_cachep, entry);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
}
mb_debug(1, "freed %u blocks in %u structures\n", count, count2);
void exit_ext4_mballoc(void)
{
- /*
+ /*
* Wait for completion of call_rcu()'s on ext4_pspace_cachep
* before destroying the slab cache.
*/
handle_t *handle, unsigned int reserv_blks)
{
struct buffer_head *bitmap_bh = NULL;
- struct ext4_super_block *es;
struct ext4_group_desc *gdp;
struct buffer_head *gdp_bh;
struct ext4_sb_info *sbi;
sb = ac->ac_sb;
sbi = EXT4_SB(sb);
- es = sbi->s_es;
-
err = -EIO;
bitmap_bh = ext4_read_block_bitmap(sb, ac->ac_b_ex.fe_group);
err = ext4_handle_dirty_metadata(handle, NULL, gdp_bh);
out_err:
- sb->s_dirt = 1;
+ ext4_mark_super_dirty(sb);
brelse(bitmap_bh);
return err;
}
int bsbits, max;
ext4_lblk_t end;
loff_t size, orig_size, start_off;
- ext4_lblk_t start, orig_start;
+ ext4_lblk_t start;
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
struct ext4_prealloc_space *pa;
size = size << bsbits;
if (size < i_size_read(ac->ac_inode))
size = i_size_read(ac->ac_inode);
+ orig_size = size;
/* max size of free chunks */
max = 2 << bsbits;
start_off = (loff_t)ac->ac_o_ex.fe_logical << bsbits;
size = ac->ac_o_ex.fe_len << bsbits;
}
- orig_size = size = size >> bsbits;
- orig_start = start = start_off >> bsbits;
+ size = size >> bsbits;
+ start = start_off >> bsbits;
/* don't cover already allocated blocks in selected range */
if (ar->pleft && start <= ar->lleft) {
if (sbi->s_mb_stats && ac->ac_g_ex.fe_len > 1) {
atomic_inc(&sbi->s_bal_reqs);
atomic_add(ac->ac_b_ex.fe_len, &sbi->s_bal_allocated);
- if (ac->ac_o_ex.fe_len >= ac->ac_g_ex.fe_len)
+ if (ac->ac_b_ex.fe_len >= ac->ac_o_ex.fe_len)
atomic_inc(&sbi->s_bal_success);
atomic_add(ac->ac_found, &sbi->s_bal_ex_scanned);
if (ac->ac_g_ex.fe_start == ac->ac_b_ex.fe_start &&
continue;
/* non-extent files can't have physical blocks past 2^32 */
- if (!(EXT4_I(ac->ac_inode)->i_flags & EXT4_EXTENTS_FL) &&
+ if (!(ext4_test_inode_flag(ac->ac_inode, EXT4_INODE_EXTENTS)) &&
pa->pa_pstart + pa->pa_len > EXT4_MAX_BLOCK_FILE_PHYS)
continue;
spin_unlock(&pa->pa_lock);
grp_blk = pa->pa_pstart;
- /*
+ /*
* If doing group-based preallocation, pa_pstart may be in the
* next group when pa is used up
*/
ext4_group_t group;
ext4_grpblk_t bit;
unsigned long long grp_blk_start;
- sector_t start;
int err = 0;
int free = 0;
if (bit >= end)
break;
next = mb_find_next_bit(bitmap_bh->b_data, end, bit);
- start = ext4_group_first_block_no(sb, group) + bit;
mb_debug(1, " free preallocated %u/%u in group %u\n",
- (unsigned) start, (unsigned) next - bit,
- (unsigned) group);
+ (unsigned) ext4_group_first_block_no(sb, group) + bit,
+ (unsigned) next - bit, (unsigned) group);
free += next - bit;
if (ac) {
trace_ext4_mballoc_discard(ac);
}
- trace_ext4_mb_release_inode_pa(ac, pa, grp_blk_start + bit,
+ trace_ext4_mb_release_inode_pa(sb, ac, pa, grp_blk_start + bit,
next - bit);
mb_free_blocks(pa->pa_inode, e4b, bit, next - bit);
bit = next + 1;
pa, (unsigned long) pa->pa_lstart,
(unsigned long) pa->pa_pstart,
(unsigned long) pa->pa_len);
- ext4_grp_locked_error(sb, group,
- __func__, "free %u, pa_free %u",
+ ext4_grp_locked_error(sb, group, 0, 0, "free %u, pa_free %u",
free, pa->pa_free);
/*
* pa is already deleted so we use the value obtained
ext4_group_t group;
ext4_grpblk_t bit;
- trace_ext4_mb_release_group_pa(ac, pa);
+ trace_ext4_mb_release_group_pa(sb, ac, pa);
BUG_ON(pa->pa_deleted == 0);
ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit);
BUG_ON(group != e4b->bd_group && pa->pa_len != 0);
ext4_unlock_group(sb, group);
if (ac)
kmem_cache_free(ext4_ac_cachep, ac);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
put_bh(bitmap_bh);
return free;
}
if (bitmap_bh == NULL) {
ext4_error(sb, "Error reading block bitmap for %u",
group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
continue;
}
ext4_mb_release_inode_pa(&e4b, bitmap_bh, pa, ac);
ext4_unlock_group(sb, group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
put_bh(bitmap_bh);
list_del(&pa->u.pa_tmp_list);
ext4_mb_release_group_pa(&e4b, pa, ac);
ext4_unlock_group(sb, group);
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
list_del(&pa->u.pa_tmp_list);
call_rcu(&(pa)->u.pa_rcu, ext4_mb_pa_callback);
}
ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b,
struct ext4_free_data *new_entry)
{
+ ext4_group_t group = e4b->bd_group;
ext4_grpblk_t block;
struct ext4_free_data *entry;
struct ext4_group_info *db = e4b->bd_info;
else if (block >= (entry->start_blk + entry->count))
n = &(*n)->rb_right;
else {
- ext4_grp_locked_error(sb, e4b->bd_group, __func__,
- "Double free of blocks %d (%d %d)",
- block, entry->start_blk, entry->count);
+ ext4_grp_locked_error(sb, group, 0,
+ ext4_group_first_block_no(sb, group) + block,
+ "Block already on to-be-freed list");
return 0;
}
}
struct super_block *sb = inode->i_sb;
struct ext4_allocation_context *ac = NULL;
struct ext4_group_desc *gdp;
- struct ext4_super_block *es;
unsigned long freed = 0;
unsigned int overflow;
ext4_grpblk_t bit;
}
sbi = EXT4_SB(sb);
- es = EXT4_SB(sb)->s_es;
if (!(flags & EXT4_FREE_BLOCKS_VALIDATED) &&
!ext4_data_block_valid(sbi, block, count)) {
ext4_error(sb, "Freeing blocks not in datazone - "
if (!bh)
tbh = sb_find_get_block(inode->i_sb,
block + i);
- ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA,
+ ext4_forget(handle, flags & EXT4_FREE_BLOCKS_METADATA,
inode, tbh, block + i);
}
}
- /*
+ /*
* We need to make sure we don't reuse the freed block until
* after the transaction is committed, which we can do by
* treating the block as metadata, below. We make an
mb_clear_bits(bitmap_bh->b_data, bit, count);
mb_free_blocks(inode, &e4b, bit, count);
ext4_mb_return_to_preallocation(inode, &e4b, block, count);
+ if (test_opt(sb, DISCARD))
+ ext4_issue_discard(sb, block_group, bit, count);
}
ret = ext4_free_blks_count(sb, gdp) + count;
atomic_add(count, &sbi->s_flex_groups[flex_group].free_blocks);
}
- ext4_mb_release_desc(&e4b);
+ ext4_mb_unload_buddy(&e4b);
freed += count;
put_bh(bitmap_bh);
goto do_more;
}
- sb->s_dirt = 1;
+ ext4_mark_super_dirty(sb);
error_return:
if (freed)
dquot_free_block(inode, freed);