2 * Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com
3 * Written by Alex Tomas <alex@clusterfs.com>
5 * Architecture independence:
6 * Copyright (c) 2005, Bull S.A.
7 * Written by Pierre Peiffer <pierre.peiffer@bull.net>
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License version 2 as
11 * published by the Free Software Foundation.
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public Licens
19 * along with this program; if not, write to the Free Software
20 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-
24 * Extents support for EXT4
27 * - ext4*_error() should be used in some situations
28 * - analyze all BUG()/BUG_ON(), use -EIO where appropriate
29 * - smart tree reduction
33 #include <linux/time.h>
34 #include <linux/jbd2.h>
35 #include <linux/highuid.h>
36 #include <linux/pagemap.h>
37 #include <linux/quotaops.h>
38 #include <linux/string.h>
39 #include <linux/slab.h>
40 #include <asm/uaccess.h>
41 #include <linux/fiemap.h>
42 #include "ext4_jbd2.h"
43 #include "ext4_extents.h"
46 #include <trace/events/ext4.h>
49 * used by extent splitting.
51 #define EXT4_EXT_MAY_ZEROOUT 0x1 /* safe to zeroout if split fails \
53 #define EXT4_EXT_MARK_UNWRIT1 0x2 /* mark first half unwritten */
54 #define EXT4_EXT_MARK_UNWRIT2 0x4 /* mark second half unwritten */
56 #define EXT4_EXT_DATA_VALID1 0x8 /* first half contains valid data */
57 #define EXT4_EXT_DATA_VALID2 0x10 /* second half contains valid data */
59 static __le32 ext4_extent_block_csum(struct inode *inode,
60 struct ext4_extent_header *eh)
62 struct ext4_inode_info *ei = EXT4_I(inode);
63 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
66 csum = ext4_chksum(sbi, ei->i_csum_seed, (__u8 *)eh,
67 EXT4_EXTENT_TAIL_OFFSET(eh));
68 return cpu_to_le32(csum);
71 static int ext4_extent_block_csum_verify(struct inode *inode,
72 struct ext4_extent_header *eh)
74 struct ext4_extent_tail *et;
76 if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
77 EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
80 et = find_ext4_extent_tail(eh);
81 if (et->et_checksum != ext4_extent_block_csum(inode, eh))
86 static void ext4_extent_block_csum_set(struct inode *inode,
87 struct ext4_extent_header *eh)
89 struct ext4_extent_tail *et;
91 if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
92 EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
95 et = find_ext4_extent_tail(eh);
96 et->et_checksum = ext4_extent_block_csum(inode, eh);
99 static int ext4_split_extent(handle_t *handle,
101 struct ext4_ext_path **ppath,
102 struct ext4_map_blocks *map,
106 static int ext4_split_extent_at(handle_t *handle,
108 struct ext4_ext_path **ppath,
113 static int ext4_find_delayed_extent(struct inode *inode,
114 struct extent_status *newes);
116 static int ext4_ext_truncate_extend_restart(handle_t *handle,
122 if (!ext4_handle_valid(handle))
124 if (handle->h_buffer_credits > needed)
126 err = ext4_journal_extend(handle, needed);
129 err = ext4_truncate_restart_trans(handle, inode, needed);
141 static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
142 struct ext4_ext_path *path)
145 /* path points to block */
146 BUFFER_TRACE(path->p_bh, "get_write_access");
147 return ext4_journal_get_write_access(handle, path->p_bh);
149 /* path points to leaf/index in inode body */
150 /* we use in-core data, no need to protect them */
160 int __ext4_ext_dirty(const char *where, unsigned int line, handle_t *handle,
161 struct inode *inode, struct ext4_ext_path *path)
165 WARN_ON(!rwsem_is_locked(&EXT4_I(inode)->i_data_sem));
167 ext4_extent_block_csum_set(inode, ext_block_hdr(path->p_bh));
168 /* path points to block */
169 err = __ext4_handle_dirty_metadata(where, line, handle,
172 /* path points to leaf/index in inode body */
173 err = ext4_mark_inode_dirty(handle, inode);
178 static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
179 struct ext4_ext_path *path,
183 int depth = path->p_depth;
184 struct ext4_extent *ex;
187 * Try to predict block placement assuming that we are
188 * filling in a file which will eventually be
189 * non-sparse --- i.e., in the case of libbfd writing
190 * an ELF object sections out-of-order but in a way
191 * the eventually results in a contiguous object or
192 * executable file, or some database extending a table
193 * space file. However, this is actually somewhat
194 * non-ideal if we are writing a sparse file such as
195 * qemu or KVM writing a raw image file that is going
196 * to stay fairly sparse, since it will end up
197 * fragmenting the file system's free space. Maybe we
198 * should have some hueristics or some way to allow
199 * userspace to pass a hint to file system,
200 * especially if the latter case turns out to be
203 ex = path[depth].p_ext;
205 ext4_fsblk_t ext_pblk = ext4_ext_pblock(ex);
206 ext4_lblk_t ext_block = le32_to_cpu(ex->ee_block);
208 if (block > ext_block)
209 return ext_pblk + (block - ext_block);
211 return ext_pblk - (ext_block - block);
214 /* it looks like index is empty;
215 * try to find starting block from index itself */
216 if (path[depth].p_bh)
217 return path[depth].p_bh->b_blocknr;
220 /* OK. use inode's group */
221 return ext4_inode_to_goal_block(inode);
225 * Allocation for a meta data block
228 ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
229 struct ext4_ext_path *path,
230 struct ext4_extent *ex, int *err, unsigned int flags)
232 ext4_fsblk_t goal, newblock;
234 goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
235 newblock = ext4_new_meta_blocks(handle, inode, goal, flags,
240 static inline int ext4_ext_space_block(struct inode *inode, int check)
244 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
245 / sizeof(struct ext4_extent);
246 #ifdef AGGRESSIVE_TEST
247 if (!check && size > 6)
253 static inline int ext4_ext_space_block_idx(struct inode *inode, int check)
257 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
258 / sizeof(struct ext4_extent_idx);
259 #ifdef AGGRESSIVE_TEST
260 if (!check && size > 5)
266 static inline int ext4_ext_space_root(struct inode *inode, int check)
270 size = sizeof(EXT4_I(inode)->i_data);
271 size -= sizeof(struct ext4_extent_header);
272 size /= sizeof(struct ext4_extent);
273 #ifdef AGGRESSIVE_TEST
274 if (!check && size > 3)
280 static inline int ext4_ext_space_root_idx(struct inode *inode, int check)
284 size = sizeof(EXT4_I(inode)->i_data);
285 size -= sizeof(struct ext4_extent_header);
286 size /= sizeof(struct ext4_extent_idx);
287 #ifdef AGGRESSIVE_TEST
288 if (!check && size > 4)
295 ext4_force_split_extent_at(handle_t *handle, struct inode *inode,
296 struct ext4_ext_path **ppath, ext4_lblk_t lblk,
299 struct ext4_ext_path *path = *ppath;
300 int unwritten = ext4_ext_is_unwritten(path[path->p_depth].p_ext);
302 return ext4_split_extent_at(handle, inode, ppath, lblk, unwritten ?
303 EXT4_EXT_MARK_UNWRIT1|EXT4_EXT_MARK_UNWRIT2 : 0,
304 EXT4_EX_NOCACHE | EXT4_GET_BLOCKS_PRE_IO |
305 (nofail ? EXT4_GET_BLOCKS_METADATA_NOFAIL:0));
309 * Calculate the number of metadata blocks needed
310 * to allocate @blocks
311 * Worse case is one block per extent
313 int ext4_ext_calc_metadata_amount(struct inode *inode, ext4_lblk_t lblock)
315 struct ext4_inode_info *ei = EXT4_I(inode);
318 idxs = ((inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
319 / sizeof(struct ext4_extent_idx));
322 * If the new delayed allocation block is contiguous with the
323 * previous da block, it can share index blocks with the
324 * previous block, so we only need to allocate a new index
325 * block every idxs leaf blocks. At ldxs**2 blocks, we need
326 * an additional index block, and at ldxs**3 blocks, yet
327 * another index blocks.
329 if (ei->i_da_metadata_calc_len &&
330 ei->i_da_metadata_calc_last_lblock+1 == lblock) {
333 if ((ei->i_da_metadata_calc_len % idxs) == 0)
335 if ((ei->i_da_metadata_calc_len % (idxs*idxs)) == 0)
337 if ((ei->i_da_metadata_calc_len % (idxs*idxs*idxs)) == 0) {
339 ei->i_da_metadata_calc_len = 0;
341 ei->i_da_metadata_calc_len++;
342 ei->i_da_metadata_calc_last_lblock++;
347 * In the worst case we need a new set of index blocks at
348 * every level of the inode's extent tree.
350 ei->i_da_metadata_calc_len = 1;
351 ei->i_da_metadata_calc_last_lblock = lblock;
352 return ext_depth(inode) + 1;
356 ext4_ext_max_entries(struct inode *inode, int depth)
360 if (depth == ext_depth(inode)) {
362 max = ext4_ext_space_root(inode, 1);
364 max = ext4_ext_space_root_idx(inode, 1);
367 max = ext4_ext_space_block(inode, 1);
369 max = ext4_ext_space_block_idx(inode, 1);
375 static int ext4_valid_extent(struct inode *inode, struct ext4_extent *ext)
377 ext4_fsblk_t block = ext4_ext_pblock(ext);
378 int len = ext4_ext_get_actual_len(ext);
379 ext4_lblk_t lblock = le32_to_cpu(ext->ee_block);
380 ext4_lblk_t last = lblock + len - 1;
384 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, len);
387 static int ext4_valid_extent_idx(struct inode *inode,
388 struct ext4_extent_idx *ext_idx)
390 ext4_fsblk_t block = ext4_idx_pblock(ext_idx);
392 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, 1);
395 static int ext4_valid_extent_entries(struct inode *inode,
396 struct ext4_extent_header *eh,
399 unsigned short entries;
400 if (eh->eh_entries == 0)
403 entries = le16_to_cpu(eh->eh_entries);
407 struct ext4_extent *ext = EXT_FIRST_EXTENT(eh);
408 struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es;
409 ext4_fsblk_t pblock = 0;
410 ext4_lblk_t lblock = 0;
411 ext4_lblk_t prev = 0;
414 if (!ext4_valid_extent(inode, ext))
417 /* Check for overlapping extents */
418 lblock = le32_to_cpu(ext->ee_block);
419 len = ext4_ext_get_actual_len(ext);
420 if ((lblock <= prev) && prev) {
421 pblock = ext4_ext_pblock(ext);
422 es->s_last_error_block = cpu_to_le64(pblock);
427 prev = lblock + len - 1;
430 struct ext4_extent_idx *ext_idx = EXT_FIRST_INDEX(eh);
432 if (!ext4_valid_extent_idx(inode, ext_idx))
441 static int __ext4_ext_check(const char *function, unsigned int line,
442 struct inode *inode, struct ext4_extent_header *eh,
443 int depth, ext4_fsblk_t pblk)
445 const char *error_msg;
448 if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
449 error_msg = "invalid magic";
452 if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
453 error_msg = "unexpected eh_depth";
456 if (unlikely(eh->eh_max == 0)) {
457 error_msg = "invalid eh_max";
460 max = ext4_ext_max_entries(inode, depth);
461 if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
462 error_msg = "too large eh_max";
465 if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
466 error_msg = "invalid eh_entries";
469 if (!ext4_valid_extent_entries(inode, eh, depth)) {
470 error_msg = "invalid extent entries";
473 /* Verify checksum on non-root extent tree nodes */
474 if (ext_depth(inode) != depth &&
475 !ext4_extent_block_csum_verify(inode, eh)) {
476 error_msg = "extent tree corrupted";
482 ext4_error_inode(inode, function, line, 0,
483 "pblk %llu bad header/extent: %s - magic %x, "
484 "entries %u, max %u(%u), depth %u(%u)",
485 (unsigned long long) pblk, error_msg,
486 le16_to_cpu(eh->eh_magic),
487 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
488 max, le16_to_cpu(eh->eh_depth), depth);
492 #define ext4_ext_check(inode, eh, depth, pblk) \
493 __ext4_ext_check(__func__, __LINE__, (inode), (eh), (depth), (pblk))
495 int ext4_ext_check_inode(struct inode *inode)
497 return ext4_ext_check(inode, ext_inode_hdr(inode), ext_depth(inode), 0);
500 static struct buffer_head *
501 __read_extent_tree_block(const char *function, unsigned int line,
502 struct inode *inode, ext4_fsblk_t pblk, int depth,
505 struct buffer_head *bh;
508 bh = sb_getblk(inode->i_sb, pblk);
510 return ERR_PTR(-ENOMEM);
512 if (!bh_uptodate_or_lock(bh)) {
513 trace_ext4_ext_load_extent(inode, pblk, _RET_IP_);
514 err = bh_submit_read(bh);
518 if (buffer_verified(bh) && !(flags & EXT4_EX_FORCE_CACHE))
520 err = __ext4_ext_check(function, line, inode,
521 ext_block_hdr(bh), depth, pblk);
524 set_buffer_verified(bh);
526 * If this is a leaf block, cache all of its entries
528 if (!(flags & EXT4_EX_NOCACHE) && depth == 0) {
529 struct ext4_extent_header *eh = ext_block_hdr(bh);
530 struct ext4_extent *ex = EXT_FIRST_EXTENT(eh);
531 ext4_lblk_t prev = 0;
534 for (i = le16_to_cpu(eh->eh_entries); i > 0; i--, ex++) {
535 unsigned int status = EXTENT_STATUS_WRITTEN;
536 ext4_lblk_t lblk = le32_to_cpu(ex->ee_block);
537 int len = ext4_ext_get_actual_len(ex);
539 if (prev && (prev != lblk))
540 ext4_es_cache_extent(inode, prev,
544 if (ext4_ext_is_unwritten(ex))
545 status = EXTENT_STATUS_UNWRITTEN;
546 ext4_es_cache_extent(inode, lblk, len,
547 ext4_ext_pblock(ex), status);
558 #define read_extent_tree_block(inode, pblk, depth, flags) \
559 __read_extent_tree_block(__func__, __LINE__, (inode), (pblk), \
563 * This function is called to cache a file's extent information in the
566 int ext4_ext_precache(struct inode *inode)
568 struct ext4_inode_info *ei = EXT4_I(inode);
569 struct ext4_ext_path *path = NULL;
570 struct buffer_head *bh;
571 int i = 0, depth, ret = 0;
573 if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
574 return 0; /* not an extent-mapped inode */
576 down_read(&ei->i_data_sem);
577 depth = ext_depth(inode);
579 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1),
582 up_read(&ei->i_data_sem);
586 /* Don't cache anything if there are no external extent blocks */
589 path[0].p_hdr = ext_inode_hdr(inode);
590 ret = ext4_ext_check(inode, path[0].p_hdr, depth, 0);
593 path[0].p_idx = EXT_FIRST_INDEX(path[0].p_hdr);
596 * If this is a leaf block or we've reached the end of
597 * the index block, go up
600 path[i].p_idx > EXT_LAST_INDEX(path[i].p_hdr)) {
601 brelse(path[i].p_bh);
606 bh = read_extent_tree_block(inode,
607 ext4_idx_pblock(path[i].p_idx++),
609 EXT4_EX_FORCE_CACHE);
616 path[i].p_hdr = ext_block_hdr(bh);
617 path[i].p_idx = EXT_FIRST_INDEX(path[i].p_hdr);
619 ext4_set_inode_state(inode, EXT4_STATE_EXT_PRECACHED);
621 up_read(&ei->i_data_sem);
622 ext4_ext_drop_refs(path);
628 static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
630 int k, l = path->p_depth;
633 for (k = 0; k <= l; k++, path++) {
635 ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block),
636 ext4_idx_pblock(path->p_idx));
637 } else if (path->p_ext) {
638 ext_debug(" %d:[%d]%d:%llu ",
639 le32_to_cpu(path->p_ext->ee_block),
640 ext4_ext_is_unwritten(path->p_ext),
641 ext4_ext_get_actual_len(path->p_ext),
642 ext4_ext_pblock(path->p_ext));
649 static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
651 int depth = ext_depth(inode);
652 struct ext4_extent_header *eh;
653 struct ext4_extent *ex;
659 eh = path[depth].p_hdr;
660 ex = EXT_FIRST_EXTENT(eh);
662 ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
664 for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
665 ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
666 ext4_ext_is_unwritten(ex),
667 ext4_ext_get_actual_len(ex), ext4_ext_pblock(ex));
672 static void ext4_ext_show_move(struct inode *inode, struct ext4_ext_path *path,
673 ext4_fsblk_t newblock, int level)
675 int depth = ext_depth(inode);
676 struct ext4_extent *ex;
678 if (depth != level) {
679 struct ext4_extent_idx *idx;
680 idx = path[level].p_idx;
681 while (idx <= EXT_MAX_INDEX(path[level].p_hdr)) {
682 ext_debug("%d: move %d:%llu in new index %llu\n", level,
683 le32_to_cpu(idx->ei_block),
684 ext4_idx_pblock(idx),
692 ex = path[depth].p_ext;
693 while (ex <= EXT_MAX_EXTENT(path[depth].p_hdr)) {
694 ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
695 le32_to_cpu(ex->ee_block),
697 ext4_ext_is_unwritten(ex),
698 ext4_ext_get_actual_len(ex),
705 #define ext4_ext_show_path(inode, path)
706 #define ext4_ext_show_leaf(inode, path)
707 #define ext4_ext_show_move(inode, path, newblock, level)
710 void ext4_ext_drop_refs(struct ext4_ext_path *path)
716 depth = path->p_depth;
717 for (i = 0; i <= depth; i++, path++)
725 * ext4_ext_binsearch_idx:
726 * binary search for the closest index of the given block
727 * the header must be checked before calling this
730 ext4_ext_binsearch_idx(struct inode *inode,
731 struct ext4_ext_path *path, ext4_lblk_t block)
733 struct ext4_extent_header *eh = path->p_hdr;
734 struct ext4_extent_idx *r, *l, *m;
737 ext_debug("binsearch for %u(idx): ", block);
739 l = EXT_FIRST_INDEX(eh) + 1;
740 r = EXT_LAST_INDEX(eh);
743 if (block < le32_to_cpu(m->ei_block))
747 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
748 m, le32_to_cpu(m->ei_block),
749 r, le32_to_cpu(r->ei_block));
753 ext_debug(" -> %u->%lld ", le32_to_cpu(path->p_idx->ei_block),
754 ext4_idx_pblock(path->p_idx));
756 #ifdef CHECK_BINSEARCH
758 struct ext4_extent_idx *chix, *ix;
761 chix = ix = EXT_FIRST_INDEX(eh);
762 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
764 le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
765 printk(KERN_DEBUG "k=%d, ix=0x%p, "
767 ix, EXT_FIRST_INDEX(eh));
768 printk(KERN_DEBUG "%u <= %u\n",
769 le32_to_cpu(ix->ei_block),
770 le32_to_cpu(ix[-1].ei_block));
772 BUG_ON(k && le32_to_cpu(ix->ei_block)
773 <= le32_to_cpu(ix[-1].ei_block));
774 if (block < le32_to_cpu(ix->ei_block))
778 BUG_ON(chix != path->p_idx);
785 * ext4_ext_binsearch:
786 * binary search for closest extent of the given block
787 * the header must be checked before calling this
790 ext4_ext_binsearch(struct inode *inode,
791 struct ext4_ext_path *path, ext4_lblk_t block)
793 struct ext4_extent_header *eh = path->p_hdr;
794 struct ext4_extent *r, *l, *m;
796 if (eh->eh_entries == 0) {
798 * this leaf is empty:
799 * we get such a leaf in split/add case
804 ext_debug("binsearch for %u: ", block);
806 l = EXT_FIRST_EXTENT(eh) + 1;
807 r = EXT_LAST_EXTENT(eh);
811 if (block < le32_to_cpu(m->ee_block))
815 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
816 m, le32_to_cpu(m->ee_block),
817 r, le32_to_cpu(r->ee_block));
821 ext_debug(" -> %d:%llu:[%d]%d ",
822 le32_to_cpu(path->p_ext->ee_block),
823 ext4_ext_pblock(path->p_ext),
824 ext4_ext_is_unwritten(path->p_ext),
825 ext4_ext_get_actual_len(path->p_ext));
827 #ifdef CHECK_BINSEARCH
829 struct ext4_extent *chex, *ex;
832 chex = ex = EXT_FIRST_EXTENT(eh);
833 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
834 BUG_ON(k && le32_to_cpu(ex->ee_block)
835 <= le32_to_cpu(ex[-1].ee_block));
836 if (block < le32_to_cpu(ex->ee_block))
840 BUG_ON(chex != path->p_ext);
846 int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
848 struct ext4_extent_header *eh;
850 eh = ext_inode_hdr(inode);
853 eh->eh_magic = EXT4_EXT_MAGIC;
854 eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode, 0));
855 ext4_mark_inode_dirty(handle, inode);
859 struct ext4_ext_path *
860 ext4_find_extent(struct inode *inode, ext4_lblk_t block,
861 struct ext4_ext_path **orig_path, int flags)
863 struct ext4_extent_header *eh;
864 struct buffer_head *bh;
865 struct ext4_ext_path *path = orig_path ? *orig_path : NULL;
866 short int depth, i, ppos = 0;
869 eh = ext_inode_hdr(inode);
870 depth = ext_depth(inode);
873 ext4_ext_drop_refs(path);
874 if (depth > path[0].p_maxdepth) {
876 *orig_path = path = NULL;
880 /* account possible depth increase */
881 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
884 return ERR_PTR(-ENOMEM);
885 path[0].p_maxdepth = depth + 1;
891 /* walk through the tree */
893 ext_debug("depth %d: num %d, max %d\n",
894 ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
896 ext4_ext_binsearch_idx(inode, path + ppos, block);
897 path[ppos].p_block = ext4_idx_pblock(path[ppos].p_idx);
898 path[ppos].p_depth = i;
899 path[ppos].p_ext = NULL;
901 bh = read_extent_tree_block(inode, path[ppos].p_block, --i,
903 if (unlikely(IS_ERR(bh))) {
908 eh = ext_block_hdr(bh);
910 if (unlikely(ppos > depth)) {
912 EXT4_ERROR_INODE(inode,
913 "ppos %d > depth %d", ppos, depth);
917 path[ppos].p_bh = bh;
918 path[ppos].p_hdr = eh;
921 path[ppos].p_depth = i;
922 path[ppos].p_ext = NULL;
923 path[ppos].p_idx = NULL;
926 ext4_ext_binsearch(inode, path + ppos, block);
927 /* if not an empty leaf */
928 if (path[ppos].p_ext)
929 path[ppos].p_block = ext4_ext_pblock(path[ppos].p_ext);
931 ext4_ext_show_path(inode, path);
936 ext4_ext_drop_refs(path);
944 * ext4_ext_insert_index:
945 * insert new index [@logical;@ptr] into the block at @curp;
946 * check where to insert: before @curp or after @curp
948 static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
949 struct ext4_ext_path *curp,
950 int logical, ext4_fsblk_t ptr)
952 struct ext4_extent_idx *ix;
955 err = ext4_ext_get_access(handle, inode, curp);
959 if (unlikely(logical == le32_to_cpu(curp->p_idx->ei_block))) {
960 EXT4_ERROR_INODE(inode,
961 "logical %d == ei_block %d!",
962 logical, le32_to_cpu(curp->p_idx->ei_block));
966 if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries)
967 >= le16_to_cpu(curp->p_hdr->eh_max))) {
968 EXT4_ERROR_INODE(inode,
969 "eh_entries %d >= eh_max %d!",
970 le16_to_cpu(curp->p_hdr->eh_entries),
971 le16_to_cpu(curp->p_hdr->eh_max));
975 if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
977 ext_debug("insert new index %d after: %llu\n", logical, ptr);
978 ix = curp->p_idx + 1;
981 ext_debug("insert new index %d before: %llu\n", logical, ptr);
985 len = EXT_LAST_INDEX(curp->p_hdr) - ix + 1;
988 ext_debug("insert new index %d: "
989 "move %d indices from 0x%p to 0x%p\n",
990 logical, len, ix, ix + 1);
991 memmove(ix + 1, ix, len * sizeof(struct ext4_extent_idx));
994 if (unlikely(ix > EXT_MAX_INDEX(curp->p_hdr))) {
995 EXT4_ERROR_INODE(inode, "ix > EXT_MAX_INDEX!");
999 ix->ei_block = cpu_to_le32(logical);
1000 ext4_idx_store_pblock(ix, ptr);
1001 le16_add_cpu(&curp->p_hdr->eh_entries, 1);
1003 if (unlikely(ix > EXT_LAST_INDEX(curp->p_hdr))) {
1004 EXT4_ERROR_INODE(inode, "ix > EXT_LAST_INDEX!");
1008 err = ext4_ext_dirty(handle, inode, curp);
1009 ext4_std_error(inode->i_sb, err);
1016 * inserts new subtree into the path, using free index entry
1018 * - allocates all needed blocks (new leaf and all intermediate index blocks)
1019 * - makes decision where to split
1020 * - moves remaining extents and index entries (right to the split point)
1021 * into the newly allocated blocks
1022 * - initializes subtree
1024 static int ext4_ext_split(handle_t *handle, struct inode *inode,
1026 struct ext4_ext_path *path,
1027 struct ext4_extent *newext, int at)
1029 struct buffer_head *bh = NULL;
1030 int depth = ext_depth(inode);
1031 struct ext4_extent_header *neh;
1032 struct ext4_extent_idx *fidx;
1033 int i = at, k, m, a;
1034 ext4_fsblk_t newblock, oldblock;
1036 ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
1039 /* make decision: where to split? */
1040 /* FIXME: now decision is simplest: at current extent */
1042 /* if current leaf will be split, then we should use
1043 * border from split point */
1044 if (unlikely(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr))) {
1045 EXT4_ERROR_INODE(inode, "p_ext > EXT_MAX_EXTENT!");
1048 if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
1049 border = path[depth].p_ext[1].ee_block;
1050 ext_debug("leaf will be split."
1051 " next leaf starts at %d\n",
1052 le32_to_cpu(border));
1054 border = newext->ee_block;
1055 ext_debug("leaf will be added."
1056 " next leaf starts at %d\n",
1057 le32_to_cpu(border));
1061 * If error occurs, then we break processing
1062 * and mark filesystem read-only. index won't
1063 * be inserted and tree will be in consistent
1064 * state. Next mount will repair buffers too.
1068 * Get array to track all allocated blocks.
1069 * We need this to handle errors and free blocks
1072 ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
1076 /* allocate all needed blocks */
1077 ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
1078 for (a = 0; a < depth - at; a++) {
1079 newblock = ext4_ext_new_meta_block(handle, inode, path,
1080 newext, &err, flags);
1083 ablocks[a] = newblock;
1086 /* initialize new leaf */
1087 newblock = ablocks[--a];
1088 if (unlikely(newblock == 0)) {
1089 EXT4_ERROR_INODE(inode, "newblock == 0!");
1093 bh = sb_getblk(inode->i_sb, newblock);
1094 if (unlikely(!bh)) {
1100 err = ext4_journal_get_create_access(handle, bh);
1104 neh = ext_block_hdr(bh);
1105 neh->eh_entries = 0;
1106 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
1107 neh->eh_magic = EXT4_EXT_MAGIC;
1110 /* move remainder of path[depth] to the new leaf */
1111 if (unlikely(path[depth].p_hdr->eh_entries !=
1112 path[depth].p_hdr->eh_max)) {
1113 EXT4_ERROR_INODE(inode, "eh_entries %d != eh_max %d!",
1114 path[depth].p_hdr->eh_entries,
1115 path[depth].p_hdr->eh_max);
1119 /* start copy from next extent */
1120 m = EXT_MAX_EXTENT(path[depth].p_hdr) - path[depth].p_ext++;
1121 ext4_ext_show_move(inode, path, newblock, depth);
1123 struct ext4_extent *ex;
1124 ex = EXT_FIRST_EXTENT(neh);
1125 memmove(ex, path[depth].p_ext, sizeof(struct ext4_extent) * m);
1126 le16_add_cpu(&neh->eh_entries, m);
1129 ext4_extent_block_csum_set(inode, neh);
1130 set_buffer_uptodate(bh);
1133 err = ext4_handle_dirty_metadata(handle, inode, bh);
1139 /* correct old leaf */
1141 err = ext4_ext_get_access(handle, inode, path + depth);
1144 le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
1145 err = ext4_ext_dirty(handle, inode, path + depth);
1151 /* create intermediate indexes */
1153 if (unlikely(k < 0)) {
1154 EXT4_ERROR_INODE(inode, "k %d < 0!", k);
1159 ext_debug("create %d intermediate indices\n", k);
1160 /* insert new index into current index block */
1161 /* current depth stored in i var */
1164 oldblock = newblock;
1165 newblock = ablocks[--a];
1166 bh = sb_getblk(inode->i_sb, newblock);
1167 if (unlikely(!bh)) {
1173 err = ext4_journal_get_create_access(handle, bh);
1177 neh = ext_block_hdr(bh);
1178 neh->eh_entries = cpu_to_le16(1);
1179 neh->eh_magic = EXT4_EXT_MAGIC;
1180 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
1181 neh->eh_depth = cpu_to_le16(depth - i);
1182 fidx = EXT_FIRST_INDEX(neh);
1183 fidx->ei_block = border;
1184 ext4_idx_store_pblock(fidx, oldblock);
1186 ext_debug("int.index at %d (block %llu): %u -> %llu\n",
1187 i, newblock, le32_to_cpu(border), oldblock);
1189 /* move remainder of path[i] to the new index block */
1190 if (unlikely(EXT_MAX_INDEX(path[i].p_hdr) !=
1191 EXT_LAST_INDEX(path[i].p_hdr))) {
1192 EXT4_ERROR_INODE(inode,
1193 "EXT_MAX_INDEX != EXT_LAST_INDEX ee_block %d!",
1194 le32_to_cpu(path[i].p_ext->ee_block));
1198 /* start copy indexes */
1199 m = EXT_MAX_INDEX(path[i].p_hdr) - path[i].p_idx++;
1200 ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
1201 EXT_MAX_INDEX(path[i].p_hdr));
1202 ext4_ext_show_move(inode, path, newblock, i);
1204 memmove(++fidx, path[i].p_idx,
1205 sizeof(struct ext4_extent_idx) * m);
1206 le16_add_cpu(&neh->eh_entries, m);
1208 ext4_extent_block_csum_set(inode, neh);
1209 set_buffer_uptodate(bh);
1212 err = ext4_handle_dirty_metadata(handle, inode, bh);
1218 /* correct old index */
1220 err = ext4_ext_get_access(handle, inode, path + i);
1223 le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
1224 err = ext4_ext_dirty(handle, inode, path + i);
1232 /* insert new index */
1233 err = ext4_ext_insert_index(handle, inode, path + at,
1234 le32_to_cpu(border), newblock);
1238 if (buffer_locked(bh))
1244 /* free all allocated blocks in error case */
1245 for (i = 0; i < depth; i++) {
1248 ext4_free_blocks(handle, inode, NULL, ablocks[i], 1,
1249 EXT4_FREE_BLOCKS_METADATA);
1258 * ext4_ext_grow_indepth:
1259 * implements tree growing procedure:
1260 * - allocates new block
1261 * - moves top-level data (index block or leaf) into the new block
1262 * - initializes new top-level, creating index that points to the
1263 * just created block
1265 static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
1267 struct ext4_extent *newext)
1269 struct ext4_extent_header *neh;
1270 struct buffer_head *bh;
1271 ext4_fsblk_t newblock;
1274 newblock = ext4_ext_new_meta_block(handle, inode, NULL,
1275 newext, &err, flags);
1279 bh = sb_getblk(inode->i_sb, newblock);
1284 err = ext4_journal_get_create_access(handle, bh);
1290 /* move top-level index/leaf into new block */
1291 memmove(bh->b_data, EXT4_I(inode)->i_data,
1292 sizeof(EXT4_I(inode)->i_data));
1294 /* set size of new block */
1295 neh = ext_block_hdr(bh);
1296 /* old root could have indexes or leaves
1297 * so calculate e_max right way */
1298 if (ext_depth(inode))
1299 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
1301 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
1302 neh->eh_magic = EXT4_EXT_MAGIC;
1303 ext4_extent_block_csum_set(inode, neh);
1304 set_buffer_uptodate(bh);
1307 err = ext4_handle_dirty_metadata(handle, inode, bh);
1311 /* Update top-level index: num,max,pointer */
1312 neh = ext_inode_hdr(inode);
1313 neh->eh_entries = cpu_to_le16(1);
1314 ext4_idx_store_pblock(EXT_FIRST_INDEX(neh), newblock);
1315 if (neh->eh_depth == 0) {
1316 /* Root extent block becomes index block */
1317 neh->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode, 0));
1318 EXT_FIRST_INDEX(neh)->ei_block =
1319 EXT_FIRST_EXTENT(neh)->ee_block;
1321 ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
1322 le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
1323 le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block),
1324 ext4_idx_pblock(EXT_FIRST_INDEX(neh)));
1326 le16_add_cpu(&neh->eh_depth, 1);
1327 ext4_mark_inode_dirty(handle, inode);
1335 * ext4_ext_create_new_leaf:
1336 * finds empty index and adds new leaf.
1337 * if no free index is found, then it requests in-depth growing.
1339 static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
1340 unsigned int mb_flags,
1341 unsigned int gb_flags,
1342 struct ext4_ext_path **ppath,
1343 struct ext4_extent *newext)
1345 struct ext4_ext_path *path = *ppath;
1346 struct ext4_ext_path *curp;
1347 int depth, i, err = 0;
1350 i = depth = ext_depth(inode);
1352 /* walk up to the tree and look for free index entry */
1353 curp = path + depth;
1354 while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
1359 /* we use already allocated block for index block,
1360 * so subsequent data blocks should be contiguous */
1361 if (EXT_HAS_FREE_INDEX(curp)) {
1362 /* if we found index with free entry, then use that
1363 * entry: create all needed subtree and add new leaf */
1364 err = ext4_ext_split(handle, inode, mb_flags, path, newext, i);
1369 path = ext4_find_extent(inode,
1370 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1373 err = PTR_ERR(path);
1375 /* tree is full, time to grow in depth */
1376 err = ext4_ext_grow_indepth(handle, inode, mb_flags, newext);
1381 path = ext4_find_extent(inode,
1382 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1385 err = PTR_ERR(path);
1390 * only first (depth 0 -> 1) produces free space;
1391 * in all other cases we have to split the grown tree
1393 depth = ext_depth(inode);
1394 if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
1395 /* now we need to split */
1405 * search the closest allocated block to the left for *logical
1406 * and returns it at @logical + it's physical address at @phys
1407 * if *logical is the smallest allocated block, the function
1408 * returns 0 at @phys
1409 * return value contains 0 (success) or error code
1411 static int ext4_ext_search_left(struct inode *inode,
1412 struct ext4_ext_path *path,
1413 ext4_lblk_t *logical, ext4_fsblk_t *phys)
1415 struct ext4_extent_idx *ix;
1416 struct ext4_extent *ex;
1419 if (unlikely(path == NULL)) {
1420 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1423 depth = path->p_depth;
1426 if (depth == 0 && path->p_ext == NULL)
1429 /* usually extent in the path covers blocks smaller
1430 * then *logical, but it can be that extent is the
1431 * first one in the file */
1433 ex = path[depth].p_ext;
1434 ee_len = ext4_ext_get_actual_len(ex);
1435 if (*logical < le32_to_cpu(ex->ee_block)) {
1436 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1437 EXT4_ERROR_INODE(inode,
1438 "EXT_FIRST_EXTENT != ex *logical %d ee_block %d!",
1439 *logical, le32_to_cpu(ex->ee_block));
1442 while (--depth >= 0) {
1443 ix = path[depth].p_idx;
1444 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1445 EXT4_ERROR_INODE(inode,
1446 "ix (%d) != EXT_FIRST_INDEX (%d) (depth %d)!",
1447 ix != NULL ? le32_to_cpu(ix->ei_block) : 0,
1448 EXT_FIRST_INDEX(path[depth].p_hdr) != NULL ?
1449 le32_to_cpu(EXT_FIRST_INDEX(path[depth].p_hdr)->ei_block) : 0,
1457 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1458 EXT4_ERROR_INODE(inode,
1459 "logical %d < ee_block %d + ee_len %d!",
1460 *logical, le32_to_cpu(ex->ee_block), ee_len);
1464 *logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
1465 *phys = ext4_ext_pblock(ex) + ee_len - 1;
1470 * search the closest allocated block to the right for *logical
1471 * and returns it at @logical + it's physical address at @phys
1472 * if *logical is the largest allocated block, the function
1473 * returns 0 at @phys
1474 * return value contains 0 (success) or error code
1476 static int ext4_ext_search_right(struct inode *inode,
1477 struct ext4_ext_path *path,
1478 ext4_lblk_t *logical, ext4_fsblk_t *phys,
1479 struct ext4_extent **ret_ex)
1481 struct buffer_head *bh = NULL;
1482 struct ext4_extent_header *eh;
1483 struct ext4_extent_idx *ix;
1484 struct ext4_extent *ex;
1486 int depth; /* Note, NOT eh_depth; depth from top of tree */
1489 if (unlikely(path == NULL)) {
1490 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1493 depth = path->p_depth;
1496 if (depth == 0 && path->p_ext == NULL)
1499 /* usually extent in the path covers blocks smaller
1500 * then *logical, but it can be that extent is the
1501 * first one in the file */
1503 ex = path[depth].p_ext;
1504 ee_len = ext4_ext_get_actual_len(ex);
1505 if (*logical < le32_to_cpu(ex->ee_block)) {
1506 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1507 EXT4_ERROR_INODE(inode,
1508 "first_extent(path[%d].p_hdr) != ex",
1512 while (--depth >= 0) {
1513 ix = path[depth].p_idx;
1514 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1515 EXT4_ERROR_INODE(inode,
1516 "ix != EXT_FIRST_INDEX *logical %d!",
1524 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1525 EXT4_ERROR_INODE(inode,
1526 "logical %d < ee_block %d + ee_len %d!",
1527 *logical, le32_to_cpu(ex->ee_block), ee_len);
1531 if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
1532 /* next allocated block in this leaf */
1537 /* go up and search for index to the right */
1538 while (--depth >= 0) {
1539 ix = path[depth].p_idx;
1540 if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
1544 /* we've gone up to the root and found no index to the right */
1548 /* we've found index to the right, let's
1549 * follow it and find the closest allocated
1550 * block to the right */
1552 block = ext4_idx_pblock(ix);
1553 while (++depth < path->p_depth) {
1554 /* subtract from p_depth to get proper eh_depth */
1555 bh = read_extent_tree_block(inode, block,
1556 path->p_depth - depth, 0);
1559 eh = ext_block_hdr(bh);
1560 ix = EXT_FIRST_INDEX(eh);
1561 block = ext4_idx_pblock(ix);
1565 bh = read_extent_tree_block(inode, block, path->p_depth - depth, 0);
1568 eh = ext_block_hdr(bh);
1569 ex = EXT_FIRST_EXTENT(eh);
1571 *logical = le32_to_cpu(ex->ee_block);
1572 *phys = ext4_ext_pblock(ex);
1580 * ext4_ext_next_allocated_block:
1581 * returns allocated block in subsequent extent or EXT_MAX_BLOCKS.
1582 * NOTE: it considers block number from index entry as
1583 * allocated block. Thus, index entries have to be consistent
1587 ext4_ext_next_allocated_block(struct ext4_ext_path *path)
1591 BUG_ON(path == NULL);
1592 depth = path->p_depth;
1594 if (depth == 0 && path->p_ext == NULL)
1595 return EXT_MAX_BLOCKS;
1597 while (depth >= 0) {
1598 if (depth == path->p_depth) {
1600 if (path[depth].p_ext &&
1601 path[depth].p_ext !=
1602 EXT_LAST_EXTENT(path[depth].p_hdr))
1603 return le32_to_cpu(path[depth].p_ext[1].ee_block);
1606 if (path[depth].p_idx !=
1607 EXT_LAST_INDEX(path[depth].p_hdr))
1608 return le32_to_cpu(path[depth].p_idx[1].ei_block);
1613 return EXT_MAX_BLOCKS;
1617 * ext4_ext_next_leaf_block:
1618 * returns first allocated block from next leaf or EXT_MAX_BLOCKS
1620 static ext4_lblk_t ext4_ext_next_leaf_block(struct ext4_ext_path *path)
1624 BUG_ON(path == NULL);
1625 depth = path->p_depth;
1627 /* zero-tree has no leaf blocks at all */
1629 return EXT_MAX_BLOCKS;
1631 /* go to index block */
1634 while (depth >= 0) {
1635 if (path[depth].p_idx !=
1636 EXT_LAST_INDEX(path[depth].p_hdr))
1637 return (ext4_lblk_t)
1638 le32_to_cpu(path[depth].p_idx[1].ei_block);
1642 return EXT_MAX_BLOCKS;
1646 * ext4_ext_correct_indexes:
1647 * if leaf gets modified and modified extent is first in the leaf,
1648 * then we have to correct all indexes above.
1649 * TODO: do we need to correct tree in all cases?
1651 static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
1652 struct ext4_ext_path *path)
1654 struct ext4_extent_header *eh;
1655 int depth = ext_depth(inode);
1656 struct ext4_extent *ex;
1660 eh = path[depth].p_hdr;
1661 ex = path[depth].p_ext;
1663 if (unlikely(ex == NULL || eh == NULL)) {
1664 EXT4_ERROR_INODE(inode,
1665 "ex %p == NULL or eh %p == NULL", ex, eh);
1670 /* there is no tree at all */
1674 if (ex != EXT_FIRST_EXTENT(eh)) {
1675 /* we correct tree if first leaf got modified only */
1680 * TODO: we need correction if border is smaller than current one
1683 border = path[depth].p_ext->ee_block;
1684 err = ext4_ext_get_access(handle, inode, path + k);
1687 path[k].p_idx->ei_block = border;
1688 err = ext4_ext_dirty(handle, inode, path + k);
1693 /* change all left-side indexes */
1694 if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
1696 err = ext4_ext_get_access(handle, inode, path + k);
1699 path[k].p_idx->ei_block = border;
1700 err = ext4_ext_dirty(handle, inode, path + k);
1709 ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
1710 struct ext4_extent *ex2)
1712 unsigned short ext1_ee_len, ext2_ee_len;
1715 * Make sure that both extents are initialized. We don't merge
1716 * unwritten extents so that we can be sure that end_io code has
1717 * the extent that was written properly split out and conversion to
1718 * initialized is trivial.
1720 if (ext4_ext_is_unwritten(ex1) != ext4_ext_is_unwritten(ex2))
1723 ext1_ee_len = ext4_ext_get_actual_len(ex1);
1724 ext2_ee_len = ext4_ext_get_actual_len(ex2);
1726 if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
1727 le32_to_cpu(ex2->ee_block))
1731 * To allow future support for preallocated extents to be added
1732 * as an RO_COMPAT feature, refuse to merge to extents if
1733 * this can result in the top bit of ee_len being set.
1735 if (ext1_ee_len + ext2_ee_len > EXT_INIT_MAX_LEN)
1737 if (ext4_ext_is_unwritten(ex1) &&
1738 (ext4_test_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN) ||
1739 atomic_read(&EXT4_I(inode)->i_unwritten) ||
1740 (ext1_ee_len + ext2_ee_len > EXT_UNWRITTEN_MAX_LEN)))
1742 #ifdef AGGRESSIVE_TEST
1743 if (ext1_ee_len >= 4)
1747 if (ext4_ext_pblock(ex1) + ext1_ee_len == ext4_ext_pblock(ex2))
1753 * This function tries to merge the "ex" extent to the next extent in the tree.
1754 * It always tries to merge towards right. If you want to merge towards
1755 * left, pass "ex - 1" as argument instead of "ex".
1756 * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
1757 * 1 if they got merged.
1759 static int ext4_ext_try_to_merge_right(struct inode *inode,
1760 struct ext4_ext_path *path,
1761 struct ext4_extent *ex)
1763 struct ext4_extent_header *eh;
1764 unsigned int depth, len;
1765 int merge_done = 0, unwritten;
1767 depth = ext_depth(inode);
1768 BUG_ON(path[depth].p_hdr == NULL);
1769 eh = path[depth].p_hdr;
1771 while (ex < EXT_LAST_EXTENT(eh)) {
1772 if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
1774 /* merge with next extent! */
1775 unwritten = ext4_ext_is_unwritten(ex);
1776 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1777 + ext4_ext_get_actual_len(ex + 1));
1779 ext4_ext_mark_unwritten(ex);
1781 if (ex + 1 < EXT_LAST_EXTENT(eh)) {
1782 len = (EXT_LAST_EXTENT(eh) - ex - 1)
1783 * sizeof(struct ext4_extent);
1784 memmove(ex + 1, ex + 2, len);
1786 le16_add_cpu(&eh->eh_entries, -1);
1788 WARN_ON(eh->eh_entries == 0);
1789 if (!eh->eh_entries)
1790 EXT4_ERROR_INODE(inode, "eh->eh_entries = 0!");
1797 * This function does a very simple check to see if we can collapse
1798 * an extent tree with a single extent tree leaf block into the inode.
1800 static void ext4_ext_try_to_merge_up(handle_t *handle,
1801 struct inode *inode,
1802 struct ext4_ext_path *path)
1805 unsigned max_root = ext4_ext_space_root(inode, 0);
1808 if ((path[0].p_depth != 1) ||
1809 (le16_to_cpu(path[0].p_hdr->eh_entries) != 1) ||
1810 (le16_to_cpu(path[1].p_hdr->eh_entries) > max_root))
1814 * We need to modify the block allocation bitmap and the block
1815 * group descriptor to release the extent tree block. If we
1816 * can't get the journal credits, give up.
1818 if (ext4_journal_extend(handle, 2))
1822 * Copy the extent data up to the inode
1824 blk = ext4_idx_pblock(path[0].p_idx);
1825 s = le16_to_cpu(path[1].p_hdr->eh_entries) *
1826 sizeof(struct ext4_extent_idx);
1827 s += sizeof(struct ext4_extent_header);
1829 path[1].p_maxdepth = path[0].p_maxdepth;
1830 memcpy(path[0].p_hdr, path[1].p_hdr, s);
1831 path[0].p_depth = 0;
1832 path[0].p_ext = EXT_FIRST_EXTENT(path[0].p_hdr) +
1833 (path[1].p_ext - EXT_FIRST_EXTENT(path[1].p_hdr));
1834 path[0].p_hdr->eh_max = cpu_to_le16(max_root);
1836 brelse(path[1].p_bh);
1837 ext4_free_blocks(handle, inode, NULL, blk, 1,
1838 EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
1842 * This function tries to merge the @ex extent to neighbours in the tree.
1843 * return 1 if merge left else 0.
1845 static void ext4_ext_try_to_merge(handle_t *handle,
1846 struct inode *inode,
1847 struct ext4_ext_path *path,
1848 struct ext4_extent *ex) {
1849 struct ext4_extent_header *eh;
1853 depth = ext_depth(inode);
1854 BUG_ON(path[depth].p_hdr == NULL);
1855 eh = path[depth].p_hdr;
1857 if (ex > EXT_FIRST_EXTENT(eh))
1858 merge_done = ext4_ext_try_to_merge_right(inode, path, ex - 1);
1861 (void) ext4_ext_try_to_merge_right(inode, path, ex);
1863 ext4_ext_try_to_merge_up(handle, inode, path);
1867 * check if a portion of the "newext" extent overlaps with an
1870 * If there is an overlap discovered, it updates the length of the newext
1871 * such that there will be no overlap, and then returns 1.
1872 * If there is no overlap found, it returns 0.
1874 static unsigned int ext4_ext_check_overlap(struct ext4_sb_info *sbi,
1875 struct inode *inode,
1876 struct ext4_extent *newext,
1877 struct ext4_ext_path *path)
1880 unsigned int depth, len1;
1881 unsigned int ret = 0;
1883 b1 = le32_to_cpu(newext->ee_block);
1884 len1 = ext4_ext_get_actual_len(newext);
1885 depth = ext_depth(inode);
1886 if (!path[depth].p_ext)
1888 b2 = EXT4_LBLK_CMASK(sbi, le32_to_cpu(path[depth].p_ext->ee_block));
1891 * get the next allocated block if the extent in the path
1892 * is before the requested block(s)
1895 b2 = ext4_ext_next_allocated_block(path);
1896 if (b2 == EXT_MAX_BLOCKS)
1898 b2 = EXT4_LBLK_CMASK(sbi, b2);
1901 /* check for wrap through zero on extent logical start block*/
1902 if (b1 + len1 < b1) {
1903 len1 = EXT_MAX_BLOCKS - b1;
1904 newext->ee_len = cpu_to_le16(len1);
1908 /* check for overlap */
1909 if (b1 + len1 > b2) {
1910 newext->ee_len = cpu_to_le16(b2 - b1);
1918 * ext4_ext_insert_extent:
1919 * tries to merge requsted extent into the existing extent or
1920 * inserts requested extent as new one into the tree,
1921 * creating new leaf in the no-space case.
1923 int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
1924 struct ext4_ext_path **ppath,
1925 struct ext4_extent *newext, int gb_flags)
1927 struct ext4_ext_path *path = *ppath;
1928 struct ext4_extent_header *eh;
1929 struct ext4_extent *ex, *fex;
1930 struct ext4_extent *nearex; /* nearest extent */
1931 struct ext4_ext_path *npath = NULL;
1932 int depth, len, err;
1934 int mb_flags = 0, unwritten;
1936 if (unlikely(ext4_ext_get_actual_len(newext) == 0)) {
1937 EXT4_ERROR_INODE(inode, "ext4_ext_get_actual_len(newext) == 0");
1940 depth = ext_depth(inode);
1941 ex = path[depth].p_ext;
1942 eh = path[depth].p_hdr;
1943 if (unlikely(path[depth].p_hdr == NULL)) {
1944 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
1948 /* try to insert block into found extent and return */
1949 if (ex && !(gb_flags & EXT4_GET_BLOCKS_PRE_IO)) {
1952 * Try to see whether we should rather test the extent on
1953 * right from ex, or from the left of ex. This is because
1954 * ext4_find_extent() can return either extent on the
1955 * left, or on the right from the searched position. This
1956 * will make merging more effective.
1958 if (ex < EXT_LAST_EXTENT(eh) &&
1959 (le32_to_cpu(ex->ee_block) +
1960 ext4_ext_get_actual_len(ex) <
1961 le32_to_cpu(newext->ee_block))) {
1964 } else if ((ex > EXT_FIRST_EXTENT(eh)) &&
1965 (le32_to_cpu(newext->ee_block) +
1966 ext4_ext_get_actual_len(newext) <
1967 le32_to_cpu(ex->ee_block)))
1970 /* Try to append newex to the ex */
1971 if (ext4_can_extents_be_merged(inode, ex, newext)) {
1972 ext_debug("append [%d]%d block to %u:[%d]%d"
1974 ext4_ext_is_unwritten(newext),
1975 ext4_ext_get_actual_len(newext),
1976 le32_to_cpu(ex->ee_block),
1977 ext4_ext_is_unwritten(ex),
1978 ext4_ext_get_actual_len(ex),
1979 ext4_ext_pblock(ex));
1980 err = ext4_ext_get_access(handle, inode,
1984 unwritten = ext4_ext_is_unwritten(ex);
1985 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1986 + ext4_ext_get_actual_len(newext));
1988 ext4_ext_mark_unwritten(ex);
1989 eh = path[depth].p_hdr;
1995 /* Try to prepend newex to the ex */
1996 if (ext4_can_extents_be_merged(inode, newext, ex)) {
1997 ext_debug("prepend %u[%d]%d block to %u:[%d]%d"
1999 le32_to_cpu(newext->ee_block),
2000 ext4_ext_is_unwritten(newext),
2001 ext4_ext_get_actual_len(newext),
2002 le32_to_cpu(ex->ee_block),
2003 ext4_ext_is_unwritten(ex),
2004 ext4_ext_get_actual_len(ex),
2005 ext4_ext_pblock(ex));
2006 err = ext4_ext_get_access(handle, inode,
2011 unwritten = ext4_ext_is_unwritten(ex);
2012 ex->ee_block = newext->ee_block;
2013 ext4_ext_store_pblock(ex, ext4_ext_pblock(newext));
2014 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
2015 + ext4_ext_get_actual_len(newext));
2017 ext4_ext_mark_unwritten(ex);
2018 eh = path[depth].p_hdr;
2024 depth = ext_depth(inode);
2025 eh = path[depth].p_hdr;
2026 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
2029 /* probably next leaf has space for us? */
2030 fex = EXT_LAST_EXTENT(eh);
2031 next = EXT_MAX_BLOCKS;
2032 if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
2033 next = ext4_ext_next_leaf_block(path);
2034 if (next != EXT_MAX_BLOCKS) {
2035 ext_debug("next leaf block - %u\n", next);
2036 BUG_ON(npath != NULL);
2037 npath = ext4_find_extent(inode, next, NULL, 0);
2039 return PTR_ERR(npath);
2040 BUG_ON(npath->p_depth != path->p_depth);
2041 eh = npath[depth].p_hdr;
2042 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
2043 ext_debug("next leaf isn't full(%d)\n",
2044 le16_to_cpu(eh->eh_entries));
2048 ext_debug("next leaf has no free space(%d,%d)\n",
2049 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
2053 * There is no free space in the found leaf.
2054 * We're gonna add a new leaf in the tree.
2056 if (gb_flags & EXT4_GET_BLOCKS_METADATA_NOFAIL)
2057 mb_flags = EXT4_MB_USE_RESERVED;
2058 err = ext4_ext_create_new_leaf(handle, inode, mb_flags, gb_flags,
2062 depth = ext_depth(inode);
2063 eh = path[depth].p_hdr;
2066 nearex = path[depth].p_ext;
2068 err = ext4_ext_get_access(handle, inode, path + depth);
2073 /* there is no extent in this leaf, create first one */
2074 ext_debug("first extent in the leaf: %u:%llu:[%d]%d\n",
2075 le32_to_cpu(newext->ee_block),
2076 ext4_ext_pblock(newext),
2077 ext4_ext_is_unwritten(newext),
2078 ext4_ext_get_actual_len(newext));
2079 nearex = EXT_FIRST_EXTENT(eh);
2081 if (le32_to_cpu(newext->ee_block)
2082 > le32_to_cpu(nearex->ee_block)) {
2084 ext_debug("insert %u:%llu:[%d]%d before: "
2086 le32_to_cpu(newext->ee_block),
2087 ext4_ext_pblock(newext),
2088 ext4_ext_is_unwritten(newext),
2089 ext4_ext_get_actual_len(newext),
2094 BUG_ON(newext->ee_block == nearex->ee_block);
2095 ext_debug("insert %u:%llu:[%d]%d after: "
2097 le32_to_cpu(newext->ee_block),
2098 ext4_ext_pblock(newext),
2099 ext4_ext_is_unwritten(newext),
2100 ext4_ext_get_actual_len(newext),
2103 len = EXT_LAST_EXTENT(eh) - nearex + 1;
2105 ext_debug("insert %u:%llu:[%d]%d: "
2106 "move %d extents from 0x%p to 0x%p\n",
2107 le32_to_cpu(newext->ee_block),
2108 ext4_ext_pblock(newext),
2109 ext4_ext_is_unwritten(newext),
2110 ext4_ext_get_actual_len(newext),
2111 len, nearex, nearex + 1);
2112 memmove(nearex + 1, nearex,
2113 len * sizeof(struct ext4_extent));
2117 le16_add_cpu(&eh->eh_entries, 1);
2118 path[depth].p_ext = nearex;
2119 nearex->ee_block = newext->ee_block;
2120 ext4_ext_store_pblock(nearex, ext4_ext_pblock(newext));
2121 nearex->ee_len = newext->ee_len;
2124 /* try to merge extents */
2125 if (!(gb_flags & EXT4_GET_BLOCKS_PRE_IO))
2126 ext4_ext_try_to_merge(handle, inode, path, nearex);
2129 /* time to correct all indexes above */
2130 err = ext4_ext_correct_indexes(handle, inode, path);
2134 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
2137 ext4_ext_drop_refs(npath);
2142 static int ext4_fill_fiemap_extents(struct inode *inode,
2143 ext4_lblk_t block, ext4_lblk_t num,
2144 struct fiemap_extent_info *fieinfo)
2146 struct ext4_ext_path *path = NULL;
2147 struct ext4_extent *ex;
2148 struct extent_status es;
2149 ext4_lblk_t next, next_del, start = 0, end = 0;
2150 ext4_lblk_t last = block + num;
2151 int exists, depth = 0, err = 0;
2152 unsigned int flags = 0;
2153 unsigned char blksize_bits = inode->i_sb->s_blocksize_bits;
2155 while (block < last && block != EXT_MAX_BLOCKS) {
2157 /* find extent for this block */
2158 down_read(&EXT4_I(inode)->i_data_sem);
2160 path = ext4_find_extent(inode, block, &path, 0);
2162 up_read(&EXT4_I(inode)->i_data_sem);
2163 err = PTR_ERR(path);
2168 depth = ext_depth(inode);
2169 if (unlikely(path[depth].p_hdr == NULL)) {
2170 up_read(&EXT4_I(inode)->i_data_sem);
2171 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
2175 ex = path[depth].p_ext;
2176 next = ext4_ext_next_allocated_block(path);
2181 /* there is no extent yet, so try to allocate
2182 * all requested space */
2185 } else if (le32_to_cpu(ex->ee_block) > block) {
2186 /* need to allocate space before found extent */
2188 end = le32_to_cpu(ex->ee_block);
2189 if (block + num < end)
2191 } else if (block >= le32_to_cpu(ex->ee_block)
2192 + ext4_ext_get_actual_len(ex)) {
2193 /* need to allocate space after found extent */
2198 } else if (block >= le32_to_cpu(ex->ee_block)) {
2200 * some part of requested space is covered
2204 end = le32_to_cpu(ex->ee_block)
2205 + ext4_ext_get_actual_len(ex);
2206 if (block + num < end)
2212 BUG_ON(end <= start);
2216 es.es_len = end - start;
2219 es.es_lblk = le32_to_cpu(ex->ee_block);
2220 es.es_len = ext4_ext_get_actual_len(ex);
2221 es.es_pblk = ext4_ext_pblock(ex);
2222 if (ext4_ext_is_unwritten(ex))
2223 flags |= FIEMAP_EXTENT_UNWRITTEN;
2227 * Find delayed extent and update es accordingly. We call
2228 * it even in !exists case to find out whether es is the
2229 * last existing extent or not.
2231 next_del = ext4_find_delayed_extent(inode, &es);
2232 if (!exists && next_del) {
2234 flags |= (FIEMAP_EXTENT_DELALLOC |
2235 FIEMAP_EXTENT_UNKNOWN);
2237 up_read(&EXT4_I(inode)->i_data_sem);
2239 if (unlikely(es.es_len == 0)) {
2240 EXT4_ERROR_INODE(inode, "es.es_len == 0");
2246 * This is possible iff next == next_del == EXT_MAX_BLOCKS.
2247 * we need to check next == EXT_MAX_BLOCKS because it is
2248 * possible that an extent is with unwritten and delayed
2249 * status due to when an extent is delayed allocated and
2250 * is allocated by fallocate status tree will track both of
2253 * So we could return a unwritten and delayed extent, and
2254 * its block is equal to 'next'.
2256 if (next == next_del && next == EXT_MAX_BLOCKS) {
2257 flags |= FIEMAP_EXTENT_LAST;
2258 if (unlikely(next_del != EXT_MAX_BLOCKS ||
2259 next != EXT_MAX_BLOCKS)) {
2260 EXT4_ERROR_INODE(inode,
2261 "next extent == %u, next "
2262 "delalloc extent = %u",
2270 err = fiemap_fill_next_extent(fieinfo,
2271 (__u64)es.es_lblk << blksize_bits,
2272 (__u64)es.es_pblk << blksize_bits,
2273 (__u64)es.es_len << blksize_bits,
2283 block = es.es_lblk + es.es_len;
2286 ext4_ext_drop_refs(path);
2292 * ext4_ext_put_gap_in_cache:
2293 * calculate boundaries of the gap that the requested block fits into
2294 * and cache this gap
2297 ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
2300 int depth = ext_depth(inode);
2301 unsigned long len = 0;
2302 ext4_lblk_t lblock = 0;
2303 struct ext4_extent *ex;
2305 ex = path[depth].p_ext;
2308 * there is no extent yet, so gap is [0;-] and we
2311 ext_debug("cache gap(whole file):");
2312 } else if (block < le32_to_cpu(ex->ee_block)) {
2314 len = le32_to_cpu(ex->ee_block) - block;
2315 ext_debug("cache gap(before): %u [%u:%u]",
2317 le32_to_cpu(ex->ee_block),
2318 ext4_ext_get_actual_len(ex));
2319 if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
2320 ext4_es_insert_extent(inode, lblock, len, ~0,
2321 EXTENT_STATUS_HOLE);
2322 } else if (block >= le32_to_cpu(ex->ee_block)
2323 + ext4_ext_get_actual_len(ex)) {
2325 lblock = le32_to_cpu(ex->ee_block)
2326 + ext4_ext_get_actual_len(ex);
2328 next = ext4_ext_next_allocated_block(path);
2329 ext_debug("cache gap(after): [%u:%u] %u",
2330 le32_to_cpu(ex->ee_block),
2331 ext4_ext_get_actual_len(ex),
2333 BUG_ON(next == lblock);
2334 len = next - lblock;
2335 if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
2336 ext4_es_insert_extent(inode, lblock, len, ~0,
2337 EXTENT_STATUS_HOLE);
2342 ext_debug(" -> %u:%lu\n", lblock, len);
2347 * removes index from the index block.
2349 static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
2350 struct ext4_ext_path *path, int depth)
2355 /* free index block */
2357 path = path + depth;
2358 leaf = ext4_idx_pblock(path->p_idx);
2359 if (unlikely(path->p_hdr->eh_entries == 0)) {
2360 EXT4_ERROR_INODE(inode, "path->p_hdr->eh_entries == 0");
2363 err = ext4_ext_get_access(handle, inode, path);
2367 if (path->p_idx != EXT_LAST_INDEX(path->p_hdr)) {
2368 int len = EXT_LAST_INDEX(path->p_hdr) - path->p_idx;
2369 len *= sizeof(struct ext4_extent_idx);
2370 memmove(path->p_idx, path->p_idx + 1, len);
2373 le16_add_cpu(&path->p_hdr->eh_entries, -1);
2374 err = ext4_ext_dirty(handle, inode, path);
2377 ext_debug("index is empty, remove it, free block %llu\n", leaf);
2378 trace_ext4_ext_rm_idx(inode, leaf);
2380 ext4_free_blocks(handle, inode, NULL, leaf, 1,
2381 EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
2383 while (--depth >= 0) {
2384 if (path->p_idx != EXT_FIRST_INDEX(path->p_hdr))
2387 err = ext4_ext_get_access(handle, inode, path);
2390 path->p_idx->ei_block = (path+1)->p_idx->ei_block;
2391 err = ext4_ext_dirty(handle, inode, path);
2399 * ext4_ext_calc_credits_for_single_extent:
2400 * This routine returns max. credits that needed to insert an extent
2401 * to the extent tree.
2402 * When pass the actual path, the caller should calculate credits
2405 int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
2406 struct ext4_ext_path *path)
2409 int depth = ext_depth(inode);
2412 /* probably there is space in leaf? */
2413 if (le16_to_cpu(path[depth].p_hdr->eh_entries)
2414 < le16_to_cpu(path[depth].p_hdr->eh_max)) {
2417 * There are some space in the leaf tree, no
2418 * need to account for leaf block credit
2420 * bitmaps and block group descriptor blocks
2421 * and other metadata blocks still need to be
2424 /* 1 bitmap, 1 block group descriptor */
2425 ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
2430 return ext4_chunk_trans_blocks(inode, nrblocks);
2434 * How many index/leaf blocks need to change/allocate to add @extents extents?
2436 * If we add a single extent, then in the worse case, each tree level
2437 * index/leaf need to be changed in case of the tree split.
2439 * If more extents are inserted, they could cause the whole tree split more
2440 * than once, but this is really rare.
2442 int ext4_ext_index_trans_blocks(struct inode *inode, int extents)
2447 /* If we are converting the inline data, only one is needed here. */
2448 if (ext4_has_inline_data(inode))
2451 depth = ext_depth(inode);
2461 static inline int get_default_free_blocks_flags(struct inode *inode)
2463 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2464 return EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET;
2465 else if (ext4_should_journal_data(inode))
2466 return EXT4_FREE_BLOCKS_FORGET;
2470 static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
2471 struct ext4_extent *ex,
2472 long long *partial_cluster,
2473 ext4_lblk_t from, ext4_lblk_t to)
2475 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2476 unsigned short ee_len = ext4_ext_get_actual_len(ex);
2478 int flags = get_default_free_blocks_flags(inode);
2481 * For bigalloc file systems, we never free a partial cluster
2482 * at the beginning of the extent. Instead, we make a note
2483 * that we tried freeing the cluster, and check to see if we
2484 * need to free it on a subsequent call to ext4_remove_blocks,
2485 * or at the end of the ext4_truncate() operation.
2487 flags |= EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER;
2489 trace_ext4_remove_blocks(inode, ex, from, to, *partial_cluster);
2491 * If we have a partial cluster, and it's different from the
2492 * cluster of the last block, we need to explicitly free the
2493 * partial cluster here.
2495 pblk = ext4_ext_pblock(ex) + ee_len - 1;
2496 if ((*partial_cluster > 0) &&
2497 (EXT4_B2C(sbi, pblk) != *partial_cluster)) {
2498 ext4_free_blocks(handle, inode, NULL,
2499 EXT4_C2B(sbi, *partial_cluster),
2500 sbi->s_cluster_ratio, flags);
2501 *partial_cluster = 0;
2504 #ifdef EXTENTS_STATS
2506 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2507 spin_lock(&sbi->s_ext_stats_lock);
2508 sbi->s_ext_blocks += ee_len;
2509 sbi->s_ext_extents++;
2510 if (ee_len < sbi->s_ext_min)
2511 sbi->s_ext_min = ee_len;
2512 if (ee_len > sbi->s_ext_max)
2513 sbi->s_ext_max = ee_len;
2514 if (ext_depth(inode) > sbi->s_depth_max)
2515 sbi->s_depth_max = ext_depth(inode);
2516 spin_unlock(&sbi->s_ext_stats_lock);
2519 if (from >= le32_to_cpu(ex->ee_block)
2520 && to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
2523 unsigned int unaligned;
2525 num = le32_to_cpu(ex->ee_block) + ee_len - from;
2526 pblk = ext4_ext_pblock(ex) + ee_len - num;
2528 * Usually we want to free partial cluster at the end of the
2529 * extent, except for the situation when the cluster is still
2530 * used by any other extent (partial_cluster is negative).
2532 if (*partial_cluster < 0 &&
2533 -(*partial_cluster) == EXT4_B2C(sbi, pblk + num - 1))
2534 flags |= EXT4_FREE_BLOCKS_NOFREE_LAST_CLUSTER;
2536 ext_debug("free last %u blocks starting %llu partial %lld\n",
2537 num, pblk, *partial_cluster);
2538 ext4_free_blocks(handle, inode, NULL, pblk, num, flags);
2540 * If the block range to be freed didn't start at the
2541 * beginning of a cluster, and we removed the entire
2542 * extent and the cluster is not used by any other extent,
2543 * save the partial cluster here, since we might need to
2544 * delete if we determine that the truncate operation has
2545 * removed all of the blocks in the cluster.
2547 * On the other hand, if we did not manage to free the whole
2548 * extent, we have to mark the cluster as used (store negative
2549 * cluster number in partial_cluster).
2551 unaligned = EXT4_PBLK_COFF(sbi, pblk);
2552 if (unaligned && (ee_len == num) &&
2553 (*partial_cluster != -((long long)EXT4_B2C(sbi, pblk))))
2554 *partial_cluster = EXT4_B2C(sbi, pblk);
2556 *partial_cluster = -((long long)EXT4_B2C(sbi, pblk));
2557 else if (*partial_cluster > 0)
2558 *partial_cluster = 0;
2560 ext4_error(sbi->s_sb, "strange request: removal(2) "
2561 "%u-%u from %u:%u\n",
2562 from, to, le32_to_cpu(ex->ee_block), ee_len);
2568 * ext4_ext_rm_leaf() Removes the extents associated with the
2569 * blocks appearing between "start" and "end", and splits the extents
2570 * if "start" and "end" appear in the same extent
2572 * @handle: The journal handle
2573 * @inode: The files inode
2574 * @path: The path to the leaf
2575 * @partial_cluster: The cluster which we'll have to free if all extents
2576 * has been released from it. It gets negative in case
2577 * that the cluster is still used.
2578 * @start: The first block to remove
2579 * @end: The last block to remove
2582 ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
2583 struct ext4_ext_path *path,
2584 long long *partial_cluster,
2585 ext4_lblk_t start, ext4_lblk_t end)
2587 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2588 int err = 0, correct_index = 0;
2589 int depth = ext_depth(inode), credits;
2590 struct ext4_extent_header *eh;
2593 ext4_lblk_t ex_ee_block;
2594 unsigned short ex_ee_len;
2595 unsigned unwritten = 0;
2596 struct ext4_extent *ex;
2599 /* the header must be checked already in ext4_ext_remove_space() */
2600 ext_debug("truncate since %u in leaf to %u\n", start, end);
2601 if (!path[depth].p_hdr)
2602 path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
2603 eh = path[depth].p_hdr;
2604 if (unlikely(path[depth].p_hdr == NULL)) {
2605 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
2608 /* find where to start removing */
2609 ex = path[depth].p_ext;
2611 ex = EXT_LAST_EXTENT(eh);
2613 ex_ee_block = le32_to_cpu(ex->ee_block);
2614 ex_ee_len = ext4_ext_get_actual_len(ex);
2617 * If we're starting with an extent other than the last one in the
2618 * node, we need to see if it shares a cluster with the extent to
2619 * the right (towards the end of the file). If its leftmost cluster
2620 * is this extent's rightmost cluster and it is not cluster aligned,
2621 * we'll mark it as a partial that is not to be deallocated.
2624 if (ex != EXT_LAST_EXTENT(eh)) {
2625 ext4_fsblk_t current_pblk, right_pblk;
2626 long long current_cluster, right_cluster;
2628 current_pblk = ext4_ext_pblock(ex) + ex_ee_len - 1;
2629 current_cluster = (long long)EXT4_B2C(sbi, current_pblk);
2630 right_pblk = ext4_ext_pblock(ex + 1);
2631 right_cluster = (long long)EXT4_B2C(sbi, right_pblk);
2632 if (current_cluster == right_cluster &&
2633 EXT4_PBLK_COFF(sbi, right_pblk))
2634 *partial_cluster = -right_cluster;
2637 trace_ext4_ext_rm_leaf(inode, start, ex, *partial_cluster);
2639 while (ex >= EXT_FIRST_EXTENT(eh) &&
2640 ex_ee_block + ex_ee_len > start) {
2642 if (ext4_ext_is_unwritten(ex))
2647 ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
2648 unwritten, ex_ee_len);
2649 path[depth].p_ext = ex;
2651 a = ex_ee_block > start ? ex_ee_block : start;
2652 b = ex_ee_block+ex_ee_len - 1 < end ?
2653 ex_ee_block+ex_ee_len - 1 : end;
2655 ext_debug(" border %u:%u\n", a, b);
2657 /* If this extent is beyond the end of the hole, skip it */
2658 if (end < ex_ee_block) {
2660 * We're going to skip this extent and move to another,
2661 * so if this extent is not cluster aligned we have
2662 * to mark the current cluster as used to avoid
2663 * accidentally freeing it later on
2665 pblk = ext4_ext_pblock(ex);
2666 if (EXT4_PBLK_COFF(sbi, pblk))
2668 -((long long)EXT4_B2C(sbi, pblk));
2670 ex_ee_block = le32_to_cpu(ex->ee_block);
2671 ex_ee_len = ext4_ext_get_actual_len(ex);
2673 } else if (b != ex_ee_block + ex_ee_len - 1) {
2674 EXT4_ERROR_INODE(inode,
2675 "can not handle truncate %u:%u "
2677 start, end, ex_ee_block,
2678 ex_ee_block + ex_ee_len - 1);
2681 } else if (a != ex_ee_block) {
2682 /* remove tail of the extent */
2683 num = a - ex_ee_block;
2685 /* remove whole extent: excellent! */
2689 * 3 for leaf, sb, and inode plus 2 (bmap and group
2690 * descriptor) for each block group; assume two block
2691 * groups plus ex_ee_len/blocks_per_block_group for
2694 credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
2695 if (ex == EXT_FIRST_EXTENT(eh)) {
2697 credits += (ext_depth(inode)) + 1;
2699 credits += EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb);
2701 err = ext4_ext_truncate_extend_restart(handle, inode, credits);
2705 err = ext4_ext_get_access(handle, inode, path + depth);
2709 err = ext4_remove_blocks(handle, inode, ex, partial_cluster,
2715 /* this extent is removed; mark slot entirely unused */
2716 ext4_ext_store_pblock(ex, 0);
2718 ex->ee_len = cpu_to_le16(num);
2720 * Do not mark unwritten if all the blocks in the
2721 * extent have been removed.
2723 if (unwritten && num)
2724 ext4_ext_mark_unwritten(ex);
2726 * If the extent was completely released,
2727 * we need to remove it from the leaf
2730 if (end != EXT_MAX_BLOCKS - 1) {
2732 * For hole punching, we need to scoot all the
2733 * extents up when an extent is removed so that
2734 * we dont have blank extents in the middle
2736 memmove(ex, ex+1, (EXT_LAST_EXTENT(eh) - ex) *
2737 sizeof(struct ext4_extent));
2739 /* Now get rid of the one at the end */
2740 memset(EXT_LAST_EXTENT(eh), 0,
2741 sizeof(struct ext4_extent));
2743 le16_add_cpu(&eh->eh_entries, -1);
2744 } else if (*partial_cluster > 0)
2745 *partial_cluster = 0;
2747 err = ext4_ext_dirty(handle, inode, path + depth);
2751 ext_debug("new extent: %u:%u:%llu\n", ex_ee_block, num,
2752 ext4_ext_pblock(ex));
2754 ex_ee_block = le32_to_cpu(ex->ee_block);
2755 ex_ee_len = ext4_ext_get_actual_len(ex);
2758 if (correct_index && eh->eh_entries)
2759 err = ext4_ext_correct_indexes(handle, inode, path);
2762 * If there's a partial cluster and at least one extent remains in
2763 * the leaf, free the partial cluster if it isn't shared with the
2764 * current extent. If there's a partial cluster and no extents
2765 * remain in the leaf, it can't be freed here. It can only be
2766 * freed when it's possible to determine if it's not shared with
2767 * any other extent - when the next leaf is processed or when space
2768 * removal is complete.
2770 if (*partial_cluster > 0 && eh->eh_entries &&
2771 (EXT4_B2C(sbi, ext4_ext_pblock(ex) + ex_ee_len - 1) !=
2772 *partial_cluster)) {
2773 int flags = get_default_free_blocks_flags(inode);
2775 ext4_free_blocks(handle, inode, NULL,
2776 EXT4_C2B(sbi, *partial_cluster),
2777 sbi->s_cluster_ratio, flags);
2778 *partial_cluster = 0;
2781 /* if this leaf is free, then we should
2782 * remove it from index block above */
2783 if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
2784 err = ext4_ext_rm_idx(handle, inode, path, depth);
2791 * ext4_ext_more_to_rm:
2792 * returns 1 if current index has to be freed (even partial)
2795 ext4_ext_more_to_rm(struct ext4_ext_path *path)
2797 BUG_ON(path->p_idx == NULL);
2799 if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
2803 * if truncate on deeper level happened, it wasn't partial,
2804 * so we have to consider current index for truncation
2806 if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
2811 int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,
2814 struct super_block *sb = inode->i_sb;
2815 int depth = ext_depth(inode);
2816 struct ext4_ext_path *path = NULL;
2817 long long partial_cluster = 0;
2821 ext_debug("truncate since %u to %u\n", start, end);
2823 /* probably first extent we're gonna free will be last in block */
2824 handle = ext4_journal_start(inode, EXT4_HT_TRUNCATE, depth + 1);
2826 return PTR_ERR(handle);
2829 trace_ext4_ext_remove_space(inode, start, end, depth);
2832 * Check if we are removing extents inside the extent tree. If that
2833 * is the case, we are going to punch a hole inside the extent tree
2834 * so we have to check whether we need to split the extent covering
2835 * the last block to remove so we can easily remove the part of it
2836 * in ext4_ext_rm_leaf().
2838 if (end < EXT_MAX_BLOCKS - 1) {
2839 struct ext4_extent *ex;
2840 ext4_lblk_t ee_block;
2842 /* find extent for this block */
2843 path = ext4_find_extent(inode, end, NULL, EXT4_EX_NOCACHE);
2845 ext4_journal_stop(handle);
2846 return PTR_ERR(path);
2848 depth = ext_depth(inode);
2849 /* Leaf not may not exist only if inode has no blocks at all */
2850 ex = path[depth].p_ext;
2853 EXT4_ERROR_INODE(inode,
2854 "path[%d].p_hdr == NULL",
2861 ee_block = le32_to_cpu(ex->ee_block);
2864 * See if the last block is inside the extent, if so split
2865 * the extent at 'end' block so we can easily remove the
2866 * tail of the first part of the split extent in
2867 * ext4_ext_rm_leaf().
2869 if (end >= ee_block &&
2870 end < ee_block + ext4_ext_get_actual_len(ex) - 1) {
2872 * Split the extent in two so that 'end' is the last
2873 * block in the first new extent. Also we should not
2874 * fail removing space due to ENOSPC so try to use
2875 * reserved block if that happens.
2877 err = ext4_force_split_extent_at(handle, inode, &path,
2884 * We start scanning from right side, freeing all the blocks
2885 * after i_size and walking into the tree depth-wise.
2887 depth = ext_depth(inode);
2892 le16_to_cpu(path[k].p_hdr->eh_entries)+1;
2894 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1),
2897 ext4_journal_stop(handle);
2900 path[0].p_maxdepth = path[0].p_depth = depth;
2901 path[0].p_hdr = ext_inode_hdr(inode);
2904 if (ext4_ext_check(inode, path[0].p_hdr, depth, 0)) {
2911 while (i >= 0 && err == 0) {
2913 /* this is leaf block */
2914 err = ext4_ext_rm_leaf(handle, inode, path,
2915 &partial_cluster, start,
2917 /* root level has p_bh == NULL, brelse() eats this */
2918 brelse(path[i].p_bh);
2919 path[i].p_bh = NULL;
2924 /* this is index block */
2925 if (!path[i].p_hdr) {
2926 ext_debug("initialize header\n");
2927 path[i].p_hdr = ext_block_hdr(path[i].p_bh);
2930 if (!path[i].p_idx) {
2931 /* this level hasn't been touched yet */
2932 path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
2933 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
2934 ext_debug("init index ptr: hdr 0x%p, num %d\n",
2936 le16_to_cpu(path[i].p_hdr->eh_entries));
2938 /* we were already here, see at next index */
2942 ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
2943 i, EXT_FIRST_INDEX(path[i].p_hdr),
2945 if (ext4_ext_more_to_rm(path + i)) {
2946 struct buffer_head *bh;
2947 /* go to the next level */
2948 ext_debug("move to level %d (block %llu)\n",
2949 i + 1, ext4_idx_pblock(path[i].p_idx));
2950 memset(path + i + 1, 0, sizeof(*path));
2951 bh = read_extent_tree_block(inode,
2952 ext4_idx_pblock(path[i].p_idx), depth - i - 1,
2955 /* should we reset i_size? */
2959 /* Yield here to deal with large extent trees.
2960 * Should be a no-op if we did IO above. */
2962 if (WARN_ON(i + 1 > depth)) {
2966 path[i + 1].p_bh = bh;
2968 /* save actual number of indexes since this
2969 * number is changed at the next iteration */
2970 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
2973 /* we finished processing this index, go up */
2974 if (path[i].p_hdr->eh_entries == 0 && i > 0) {
2975 /* index is empty, remove it;
2976 * handle must be already prepared by the
2977 * truncatei_leaf() */
2978 err = ext4_ext_rm_idx(handle, inode, path, i);
2980 /* root level has p_bh == NULL, brelse() eats this */
2981 brelse(path[i].p_bh);
2982 path[i].p_bh = NULL;
2984 ext_debug("return to level %d\n", i);
2988 trace_ext4_ext_remove_space_done(inode, start, end, depth,
2989 partial_cluster, path->p_hdr->eh_entries);
2991 /* If we still have something in the partial cluster and we have removed
2992 * even the first extent, then we should free the blocks in the partial
2993 * cluster as well. */
2994 if (partial_cluster > 0 && path->p_hdr->eh_entries == 0) {
2995 int flags = get_default_free_blocks_flags(inode);
2997 ext4_free_blocks(handle, inode, NULL,
2998 EXT4_C2B(EXT4_SB(sb), partial_cluster),
2999 EXT4_SB(sb)->s_cluster_ratio, flags);
3000 partial_cluster = 0;
3003 /* TODO: flexible tree reduction should be here */
3004 if (path->p_hdr->eh_entries == 0) {
3006 * truncate to zero freed all the tree,
3007 * so we need to correct eh_depth
3009 err = ext4_ext_get_access(handle, inode, path);
3011 ext_inode_hdr(inode)->eh_depth = 0;
3012 ext_inode_hdr(inode)->eh_max =
3013 cpu_to_le16(ext4_ext_space_root(inode, 0));
3014 err = ext4_ext_dirty(handle, inode, path);
3018 ext4_ext_drop_refs(path);
3023 ext4_journal_stop(handle);
3029 * called at mount time
3031 void ext4_ext_init(struct super_block *sb)
3034 * possible initialization would be here
3037 if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
3038 #if defined(AGGRESSIVE_TEST) || defined(CHECK_BINSEARCH) || defined(EXTENTS_STATS)
3039 printk(KERN_INFO "EXT4-fs: file extents enabled"
3040 #ifdef AGGRESSIVE_TEST
3041 ", aggressive tests"
3043 #ifdef CHECK_BINSEARCH
3046 #ifdef EXTENTS_STATS
3051 #ifdef EXTENTS_STATS
3052 spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
3053 EXT4_SB(sb)->s_ext_min = 1 << 30;
3054 EXT4_SB(sb)->s_ext_max = 0;
3060 * called at umount time
3062 void ext4_ext_release(struct super_block *sb)
3064 if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
3067 #ifdef EXTENTS_STATS
3068 if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
3069 struct ext4_sb_info *sbi = EXT4_SB(sb);
3070 printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
3071 sbi->s_ext_blocks, sbi->s_ext_extents,
3072 sbi->s_ext_blocks / sbi->s_ext_extents);
3073 printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
3074 sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
3079 static int ext4_zeroout_es(struct inode *inode, struct ext4_extent *ex)
3081 ext4_lblk_t ee_block;
3082 ext4_fsblk_t ee_pblock;
3083 unsigned int ee_len;
3085 ee_block = le32_to_cpu(ex->ee_block);
3086 ee_len = ext4_ext_get_actual_len(ex);
3087 ee_pblock = ext4_ext_pblock(ex);
3092 return ext4_es_insert_extent(inode, ee_block, ee_len, ee_pblock,
3093 EXTENT_STATUS_WRITTEN);
3096 /* FIXME!! we need to try to merge to left or right after zero-out */
3097 static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
3099 ext4_fsblk_t ee_pblock;
3100 unsigned int ee_len;
3103 ee_len = ext4_ext_get_actual_len(ex);
3104 ee_pblock = ext4_ext_pblock(ex);
3106 ret = sb_issue_zeroout(inode->i_sb, ee_pblock, ee_len, GFP_NOFS);
3114 * ext4_split_extent_at() splits an extent at given block.
3116 * @handle: the journal handle
3117 * @inode: the file inode
3118 * @path: the path to the extent
3119 * @split: the logical block where the extent is splitted.
3120 * @split_flags: indicates if the extent could be zeroout if split fails, and
3121 * the states(init or unwritten) of new extents.
3122 * @flags: flags used to insert new extent to extent tree.
3125 * Splits extent [a, b] into two extents [a, @split) and [@split, b], states
3126 * of which are deterimined by split_flag.
3128 * There are two cases:
3129 * a> the extent are splitted into two extent.
3130 * b> split is not needed, and just mark the extent.
3132 * return 0 on success.
3134 static int ext4_split_extent_at(handle_t *handle,
3135 struct inode *inode,
3136 struct ext4_ext_path **ppath,
3141 struct ext4_ext_path *path = *ppath;
3142 ext4_fsblk_t newblock;
3143 ext4_lblk_t ee_block;
3144 struct ext4_extent *ex, newex, orig_ex, zero_ex;
3145 struct ext4_extent *ex2 = NULL;
3146 unsigned int ee_len, depth;
3149 BUG_ON((split_flag & (EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2)) ==
3150 (EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2));
3152 ext_debug("ext4_split_extents_at: inode %lu, logical"
3153 "block %llu\n", inode->i_ino, (unsigned long long)split);
3155 ext4_ext_show_leaf(inode, path);
3157 depth = ext_depth(inode);
3158 ex = path[depth].p_ext;
3159 ee_block = le32_to_cpu(ex->ee_block);
3160 ee_len = ext4_ext_get_actual_len(ex);
3161 newblock = split - ee_block + ext4_ext_pblock(ex);
3163 BUG_ON(split < ee_block || split >= (ee_block + ee_len));
3164 BUG_ON(!ext4_ext_is_unwritten(ex) &&
3165 split_flag & (EXT4_EXT_MAY_ZEROOUT |
3166 EXT4_EXT_MARK_UNWRIT1 |
3167 EXT4_EXT_MARK_UNWRIT2));
3169 err = ext4_ext_get_access(handle, inode, path + depth);
3173 if (split == ee_block) {
3175 * case b: block @split is the block that the extent begins with
3176 * then we just change the state of the extent, and splitting
3179 if (split_flag & EXT4_EXT_MARK_UNWRIT2)
3180 ext4_ext_mark_unwritten(ex);
3182 ext4_ext_mark_initialized(ex);
3184 if (!(flags & EXT4_GET_BLOCKS_PRE_IO))
3185 ext4_ext_try_to_merge(handle, inode, path, ex);
3187 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
3192 memcpy(&orig_ex, ex, sizeof(orig_ex));
3193 ex->ee_len = cpu_to_le16(split - ee_block);
3194 if (split_flag & EXT4_EXT_MARK_UNWRIT1)
3195 ext4_ext_mark_unwritten(ex);
3198 * path may lead to new leaf, not to original leaf any more
3199 * after ext4_ext_insert_extent() returns,
3201 err = ext4_ext_dirty(handle, inode, path + depth);
3203 goto fix_extent_len;
3206 ex2->ee_block = cpu_to_le32(split);
3207 ex2->ee_len = cpu_to_le16(ee_len - (split - ee_block));
3208 ext4_ext_store_pblock(ex2, newblock);
3209 if (split_flag & EXT4_EXT_MARK_UNWRIT2)
3210 ext4_ext_mark_unwritten(ex2);
3212 err = ext4_ext_insert_extent(handle, inode, ppath, &newex, flags);
3213 if (err == -ENOSPC && (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
3214 if (split_flag & (EXT4_EXT_DATA_VALID1|EXT4_EXT_DATA_VALID2)) {
3215 if (split_flag & EXT4_EXT_DATA_VALID1) {
3216 err = ext4_ext_zeroout(inode, ex2);
3217 zero_ex.ee_block = ex2->ee_block;
3218 zero_ex.ee_len = cpu_to_le16(
3219 ext4_ext_get_actual_len(ex2));
3220 ext4_ext_store_pblock(&zero_ex,
3221 ext4_ext_pblock(ex2));
3223 err = ext4_ext_zeroout(inode, ex);
3224 zero_ex.ee_block = ex->ee_block;
3225 zero_ex.ee_len = cpu_to_le16(
3226 ext4_ext_get_actual_len(ex));
3227 ext4_ext_store_pblock(&zero_ex,
3228 ext4_ext_pblock(ex));
3231 err = ext4_ext_zeroout(inode, &orig_ex);
3232 zero_ex.ee_block = orig_ex.ee_block;
3233 zero_ex.ee_len = cpu_to_le16(
3234 ext4_ext_get_actual_len(&orig_ex));
3235 ext4_ext_store_pblock(&zero_ex,
3236 ext4_ext_pblock(&orig_ex));
3240 goto fix_extent_len;
3241 /* update the extent length and mark as initialized */
3242 ex->ee_len = cpu_to_le16(ee_len);
3243 ext4_ext_try_to_merge(handle, inode, path, ex);
3244 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
3246 goto fix_extent_len;
3248 /* update extent status tree */
3249 err = ext4_zeroout_es(inode, &zero_ex);
3253 goto fix_extent_len;
3256 ext4_ext_show_leaf(inode, path);
3260 ex->ee_len = orig_ex.ee_len;
3261 ext4_ext_dirty(handle, inode, path + path->p_depth);
3266 * ext4_split_extents() splits an extent and mark extent which is covered
3267 * by @map as split_flags indicates
3269 * It may result in splitting the extent into multiple extents (up to three)
3270 * There are three possibilities:
3271 * a> There is no split required
3272 * b> Splits in two extents: Split is happening at either end of the extent
3273 * c> Splits in three extents: Somone is splitting in middle of the extent
3276 static int ext4_split_extent(handle_t *handle,
3277 struct inode *inode,
3278 struct ext4_ext_path **ppath,
3279 struct ext4_map_blocks *map,
3283 struct ext4_ext_path *path = *ppath;
3284 ext4_lblk_t ee_block;
3285 struct ext4_extent *ex;
3286 unsigned int ee_len, depth;
3289 int split_flag1, flags1;
3290 int allocated = map->m_len;
3292 depth = ext_depth(inode);
3293 ex = path[depth].p_ext;
3294 ee_block = le32_to_cpu(ex->ee_block);
3295 ee_len = ext4_ext_get_actual_len(ex);
3296 unwritten = ext4_ext_is_unwritten(ex);
3298 if (map->m_lblk + map->m_len < ee_block + ee_len) {
3299 split_flag1 = split_flag & EXT4_EXT_MAY_ZEROOUT;
3300 flags1 = flags | EXT4_GET_BLOCKS_PRE_IO;
3302 split_flag1 |= EXT4_EXT_MARK_UNWRIT1 |
3303 EXT4_EXT_MARK_UNWRIT2;
3304 if (split_flag & EXT4_EXT_DATA_VALID2)
3305 split_flag1 |= EXT4_EXT_DATA_VALID1;
3306 err = ext4_split_extent_at(handle, inode, ppath,
3307 map->m_lblk + map->m_len, split_flag1, flags1);
3311 allocated = ee_len - (map->m_lblk - ee_block);
3314 * Update path is required because previous ext4_split_extent_at() may
3315 * result in split of original leaf or extent zeroout.
3317 path = ext4_find_extent(inode, map->m_lblk, ppath, 0);
3319 return PTR_ERR(path);
3320 depth = ext_depth(inode);
3321 ex = path[depth].p_ext;
3323 EXT4_ERROR_INODE(inode, "unexpected hole at %lu",
3324 (unsigned long) map->m_lblk);
3327 unwritten = ext4_ext_is_unwritten(ex);
3330 if (map->m_lblk >= ee_block) {
3331 split_flag1 = split_flag & EXT4_EXT_DATA_VALID2;
3333 split_flag1 |= EXT4_EXT_MARK_UNWRIT1;
3334 split_flag1 |= split_flag & (EXT4_EXT_MAY_ZEROOUT |
3335 EXT4_EXT_MARK_UNWRIT2);
3337 err = ext4_split_extent_at(handle, inode, ppath,
3338 map->m_lblk, split_flag1, flags);
3343 ext4_ext_show_leaf(inode, path);
3345 return err ? err : allocated;
3349 * This function is called by ext4_ext_map_blocks() if someone tries to write
3350 * to an unwritten extent. It may result in splitting the unwritten
3351 * extent into multiple extents (up to three - one initialized and two
3353 * There are three possibilities:
3354 * a> There is no split required: Entire extent should be initialized
3355 * b> Splits in two extents: Write is happening at either end of the extent
3356 * c> Splits in three extents: Somone is writing in middle of the extent
3359 * - The extent pointed to by 'path' is unwritten.
3360 * - The extent pointed to by 'path' contains a superset
3361 * of the logical span [map->m_lblk, map->m_lblk + map->m_len).
3363 * Post-conditions on success:
3364 * - the returned value is the number of blocks beyond map->l_lblk
3365 * that are allocated and initialized.
3366 * It is guaranteed to be >= map->m_len.
3368 static int ext4_ext_convert_to_initialized(handle_t *handle,
3369 struct inode *inode,
3370 struct ext4_map_blocks *map,
3371 struct ext4_ext_path **ppath,
3374 struct ext4_ext_path *path = *ppath;
3375 struct ext4_sb_info *sbi;
3376 struct ext4_extent_header *eh;
3377 struct ext4_map_blocks split_map;
3378 struct ext4_extent zero_ex;
3379 struct ext4_extent *ex, *abut_ex;
3380 ext4_lblk_t ee_block, eof_block;
3381 unsigned int ee_len, depth, map_len = map->m_len;
3382 int allocated = 0, max_zeroout = 0;
3386 ext_debug("ext4_ext_convert_to_initialized: inode %lu, logical"
3387 "block %llu, max_blocks %u\n", inode->i_ino,
3388 (unsigned long long)map->m_lblk, map_len);
3390 sbi = EXT4_SB(inode->i_sb);
3391 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3392 inode->i_sb->s_blocksize_bits;
3393 if (eof_block < map->m_lblk + map_len)
3394 eof_block = map->m_lblk + map_len;
3396 depth = ext_depth(inode);
3397 eh = path[depth].p_hdr;
3398 ex = path[depth].p_ext;
3399 ee_block = le32_to_cpu(ex->ee_block);
3400 ee_len = ext4_ext_get_actual_len(ex);
3403 trace_ext4_ext_convert_to_initialized_enter(inode, map, ex);
3405 /* Pre-conditions */
3406 BUG_ON(!ext4_ext_is_unwritten(ex));
3407 BUG_ON(!in_range(map->m_lblk, ee_block, ee_len));
3410 * Attempt to transfer newly initialized blocks from the currently
3411 * unwritten extent to its neighbor. This is much cheaper
3412 * than an insertion followed by a merge as those involve costly
3413 * memmove() calls. Transferring to the left is the common case in
3414 * steady state for workloads doing fallocate(FALLOC_FL_KEEP_SIZE)
3415 * followed by append writes.
3417 * Limitations of the current logic:
3418 * - L1: we do not deal with writes covering the whole extent.
3419 * This would require removing the extent if the transfer
3421 * - L2: we only attempt to merge with an extent stored in the
3422 * same extent tree node.
3424 if ((map->m_lblk == ee_block) &&
3425 /* See if we can merge left */
3426 (map_len < ee_len) && /*L1*/
3427 (ex > EXT_FIRST_EXTENT(eh))) { /*L2*/
3428 ext4_lblk_t prev_lblk;
3429 ext4_fsblk_t prev_pblk, ee_pblk;
3430 unsigned int prev_len;
3433 prev_lblk = le32_to_cpu(abut_ex->ee_block);
3434 prev_len = ext4_ext_get_actual_len(abut_ex);
3435 prev_pblk = ext4_ext_pblock(abut_ex);
3436 ee_pblk = ext4_ext_pblock(ex);
3439 * A transfer of blocks from 'ex' to 'abut_ex' is allowed
3440 * upon those conditions:
3441 * - C1: abut_ex is initialized,
3442 * - C2: abut_ex is logically abutting ex,
3443 * - C3: abut_ex is physically abutting ex,
3444 * - C4: abut_ex can receive the additional blocks without
3445 * overflowing the (initialized) length limit.
3447 if ((!ext4_ext_is_unwritten(abut_ex)) && /*C1*/
3448 ((prev_lblk + prev_len) == ee_block) && /*C2*/
3449 ((prev_pblk + prev_len) == ee_pblk) && /*C3*/
3450 (prev_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
3451 err = ext4_ext_get_access(handle, inode, path + depth);
3455 trace_ext4_ext_convert_to_initialized_fastpath(inode,
3458 /* Shift the start of ex by 'map_len' blocks */
3459 ex->ee_block = cpu_to_le32(ee_block + map_len);
3460 ext4_ext_store_pblock(ex, ee_pblk + map_len);
3461 ex->ee_len = cpu_to_le16(ee_len - map_len);
3462 ext4_ext_mark_unwritten(ex); /* Restore the flag */
3464 /* Extend abut_ex by 'map_len' blocks */
3465 abut_ex->ee_len = cpu_to_le16(prev_len + map_len);
3467 /* Result: number of initialized blocks past m_lblk */
3468 allocated = map_len;
3470 } else if (((map->m_lblk + map_len) == (ee_block + ee_len)) &&
3471 (map_len < ee_len) && /*L1*/
3472 ex < EXT_LAST_EXTENT(eh)) { /*L2*/
3473 /* See if we can merge right */
3474 ext4_lblk_t next_lblk;
3475 ext4_fsblk_t next_pblk, ee_pblk;
3476 unsigned int next_len;
3479 next_lblk = le32_to_cpu(abut_ex->ee_block);
3480 next_len = ext4_ext_get_actual_len(abut_ex);
3481 next_pblk = ext4_ext_pblock(abut_ex);
3482 ee_pblk = ext4_ext_pblock(ex);
3485 * A transfer of blocks from 'ex' to 'abut_ex' is allowed
3486 * upon those conditions:
3487 * - C1: abut_ex is initialized,
3488 * - C2: abut_ex is logically abutting ex,
3489 * - C3: abut_ex is physically abutting ex,
3490 * - C4: abut_ex can receive the additional blocks without
3491 * overflowing the (initialized) length limit.
3493 if ((!ext4_ext_is_unwritten(abut_ex)) && /*C1*/
3494 ((map->m_lblk + map_len) == next_lblk) && /*C2*/
3495 ((ee_pblk + ee_len) == next_pblk) && /*C3*/
3496 (next_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
3497 err = ext4_ext_get_access(handle, inode, path + depth);
3501 trace_ext4_ext_convert_to_initialized_fastpath(inode,
3504 /* Shift the start of abut_ex by 'map_len' blocks */
3505 abut_ex->ee_block = cpu_to_le32(next_lblk - map_len);
3506 ext4_ext_store_pblock(abut_ex, next_pblk - map_len);
3507 ex->ee_len = cpu_to_le16(ee_len - map_len);
3508 ext4_ext_mark_unwritten(ex); /* Restore the flag */
3510 /* Extend abut_ex by 'map_len' blocks */
3511 abut_ex->ee_len = cpu_to_le16(next_len + map_len);
3513 /* Result: number of initialized blocks past m_lblk */
3514 allocated = map_len;
3518 /* Mark the block containing both extents as dirty */
3519 ext4_ext_dirty(handle, inode, path + depth);
3521 /* Update path to point to the right extent */
3522 path[depth].p_ext = abut_ex;
3525 allocated = ee_len - (map->m_lblk - ee_block);
3527 WARN_ON(map->m_lblk < ee_block);
3529 * It is safe to convert extent to initialized via explicit
3530 * zeroout only if extent is fully inside i_size or new_size.
3532 split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
3534 if (EXT4_EXT_MAY_ZEROOUT & split_flag)
3535 max_zeroout = sbi->s_extent_max_zeroout_kb >>
3536 (inode->i_sb->s_blocksize_bits - 10);
3538 /* If extent is less than s_max_zeroout_kb, zeroout directly */
3539 if (max_zeroout && (ee_len <= max_zeroout)) {
3540 err = ext4_ext_zeroout(inode, ex);
3543 zero_ex.ee_block = ex->ee_block;
3544 zero_ex.ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex));
3545 ext4_ext_store_pblock(&zero_ex, ext4_ext_pblock(ex));
3547 err = ext4_ext_get_access(handle, inode, path + depth);
3550 ext4_ext_mark_initialized(ex);
3551 ext4_ext_try_to_merge(handle, inode, path, ex);
3552 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
3558 * 1. split the extent into three extents.
3559 * 2. split the extent into two extents, zeroout the first half.
3560 * 3. split the extent into two extents, zeroout the second half.
3561 * 4. split the extent into two extents with out zeroout.
3563 split_map.m_lblk = map->m_lblk;
3564 split_map.m_len = map->m_len;
3566 if (max_zeroout && (allocated > map->m_len)) {
3567 if (allocated <= max_zeroout) {
3570 cpu_to_le32(map->m_lblk);
3571 zero_ex.ee_len = cpu_to_le16(allocated);
3572 ext4_ext_store_pblock(&zero_ex,
3573 ext4_ext_pblock(ex) + map->m_lblk - ee_block);
3574 err = ext4_ext_zeroout(inode, &zero_ex);
3577 split_map.m_lblk = map->m_lblk;
3578 split_map.m_len = allocated;
3579 } else if (map->m_lblk - ee_block + map->m_len < max_zeroout) {
3581 if (map->m_lblk != ee_block) {
3582 zero_ex.ee_block = ex->ee_block;
3583 zero_ex.ee_len = cpu_to_le16(map->m_lblk -
3585 ext4_ext_store_pblock(&zero_ex,
3586 ext4_ext_pblock(ex));
3587 err = ext4_ext_zeroout(inode, &zero_ex);
3592 split_map.m_lblk = ee_block;
3593 split_map.m_len = map->m_lblk - ee_block + map->m_len;
3594 allocated = map->m_len;
3598 allocated = ext4_split_extent(handle, inode, ppath,
3599 &split_map, split_flag, flags);
3604 /* If we have gotten a failure, don't zero out status tree */
3606 err = ext4_zeroout_es(inode, &zero_ex);
3607 return err ? err : allocated;
3611 * This function is called by ext4_ext_map_blocks() from
3612 * ext4_get_blocks_dio_write() when DIO to write
3613 * to an unwritten extent.
3615 * Writing to an unwritten extent may result in splitting the unwritten
3616 * extent into multiple initialized/unwritten extents (up to three)
3617 * There are three possibilities:
3618 * a> There is no split required: Entire extent should be unwritten
3619 * b> Splits in two extents: Write is happening at either end of the extent
3620 * c> Splits in three extents: Somone is writing in middle of the extent
3622 * This works the same way in the case of initialized -> unwritten conversion.
3624 * One of more index blocks maybe needed if the extent tree grow after
3625 * the unwritten extent split. To prevent ENOSPC occur at the IO
3626 * complete, we need to split the unwritten extent before DIO submit
3627 * the IO. The unwritten extent called at this time will be split
3628 * into three unwritten extent(at most). After IO complete, the part
3629 * being filled will be convert to initialized by the end_io callback function
3630 * via ext4_convert_unwritten_extents().
3632 * Returns the size of unwritten extent to be written on success.
3634 static int ext4_split_convert_extents(handle_t *handle,
3635 struct inode *inode,
3636 struct ext4_map_blocks *map,
3637 struct ext4_ext_path **ppath,
3640 struct ext4_ext_path *path = *ppath;
3641 ext4_lblk_t eof_block;
3642 ext4_lblk_t ee_block;
3643 struct ext4_extent *ex;
3644 unsigned int ee_len;
3645 int split_flag = 0, depth;
3647 ext_debug("%s: inode %lu, logical block %llu, max_blocks %u\n",
3648 __func__, inode->i_ino,
3649 (unsigned long long)map->m_lblk, map->m_len);
3651 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3652 inode->i_sb->s_blocksize_bits;
3653 if (eof_block < map->m_lblk + map->m_len)
3654 eof_block = map->m_lblk + map->m_len;
3656 * It is safe to convert extent to initialized via explicit
3657 * zeroout only if extent is fully insde i_size or new_size.
3659 depth = ext_depth(inode);
3660 ex = path[depth].p_ext;
3661 ee_block = le32_to_cpu(ex->ee_block);
3662 ee_len = ext4_ext_get_actual_len(ex);
3664 /* Convert to unwritten */
3665 if (flags & EXT4_GET_BLOCKS_CONVERT_UNWRITTEN) {
3666 split_flag |= EXT4_EXT_DATA_VALID1;
3667 /* Convert to initialized */
3668 } else if (flags & EXT4_GET_BLOCKS_CONVERT) {
3669 split_flag |= ee_block + ee_len <= eof_block ?
3670 EXT4_EXT_MAY_ZEROOUT : 0;
3671 split_flag |= (EXT4_EXT_MARK_UNWRIT2 | EXT4_EXT_DATA_VALID2);
3673 flags |= EXT4_GET_BLOCKS_PRE_IO;
3674 return ext4_split_extent(handle, inode, ppath, map, split_flag, flags);
3677 static int ext4_convert_unwritten_extents_endio(handle_t *handle,
3678 struct inode *inode,
3679 struct ext4_map_blocks *map,
3680 struct ext4_ext_path **ppath)
3682 struct ext4_ext_path *path = *ppath;
3683 struct ext4_extent *ex;
3684 ext4_lblk_t ee_block;
3685 unsigned int ee_len;
3689 depth = ext_depth(inode);
3690 ex = path[depth].p_ext;
3691 ee_block = le32_to_cpu(ex->ee_block);
3692 ee_len = ext4_ext_get_actual_len(ex);
3694 ext_debug("ext4_convert_unwritten_extents_endio: inode %lu, logical"
3695 "block %llu, max_blocks %u\n", inode->i_ino,
3696 (unsigned long long)ee_block, ee_len);
3698 /* If extent is larger than requested it is a clear sign that we still
3699 * have some extent state machine issues left. So extent_split is still
3701 * TODO: Once all related issues will be fixed this situation should be
3704 if (ee_block != map->m_lblk || ee_len > map->m_len) {
3706 ext4_warning("Inode (%ld) finished: extent logical block %llu,"
3707 " len %u; IO logical block %llu, len %u\n",
3708 inode->i_ino, (unsigned long long)ee_block, ee_len,
3709 (unsigned long long)map->m_lblk, map->m_len);
3711 err = ext4_split_convert_extents(handle, inode, map, ppath,
3712 EXT4_GET_BLOCKS_CONVERT);
3715 path = ext4_find_extent(inode, map->m_lblk, ppath, 0);
3717 return PTR_ERR(path);
3718 depth = ext_depth(inode);
3719 ex = path[depth].p_ext;
3722 err = ext4_ext_get_access(handle, inode, path + depth);
3725 /* first mark the extent as initialized */
3726 ext4_ext_mark_initialized(ex);
3728 /* note: ext4_ext_correct_indexes() isn't needed here because
3729 * borders are not changed
3731 ext4_ext_try_to_merge(handle, inode, path, ex);
3733 /* Mark modified extent as dirty */
3734 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
3736 ext4_ext_show_leaf(inode, path);
3740 static void unmap_underlying_metadata_blocks(struct block_device *bdev,
3741 sector_t block, int count)
3744 for (i = 0; i < count; i++)
3745 unmap_underlying_metadata(bdev, block + i);
3749 * Handle EOFBLOCKS_FL flag, clearing it if necessary
3751 static int check_eofblocks_fl(handle_t *handle, struct inode *inode,
3753 struct ext4_ext_path *path,
3757 struct ext4_extent_header *eh;
3758 struct ext4_extent *last_ex;
3760 if (!ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS))
3763 depth = ext_depth(inode);
3764 eh = path[depth].p_hdr;
3767 * We're going to remove EOFBLOCKS_FL entirely in future so we
3768 * do not care for this case anymore. Simply remove the flag
3769 * if there are no extents.
3771 if (unlikely(!eh->eh_entries))
3773 last_ex = EXT_LAST_EXTENT(eh);
3775 * We should clear the EOFBLOCKS_FL flag if we are writing the
3776 * last block in the last extent in the file. We test this by
3777 * first checking to see if the caller to
3778 * ext4_ext_get_blocks() was interested in the last block (or
3779 * a block beyond the last block) in the current extent. If
3780 * this turns out to be false, we can bail out from this
3781 * function immediately.
3783 if (lblk + len < le32_to_cpu(last_ex->ee_block) +
3784 ext4_ext_get_actual_len(last_ex))
3787 * If the caller does appear to be planning to write at or
3788 * beyond the end of the current extent, we then test to see
3789 * if the current extent is the last extent in the file, by
3790 * checking to make sure it was reached via the rightmost node
3791 * at each level of the tree.
3793 for (i = depth-1; i >= 0; i--)
3794 if (path[i].p_idx != EXT_LAST_INDEX(path[i].p_hdr))
3797 ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
3798 return ext4_mark_inode_dirty(handle, inode);
3802 * ext4_find_delalloc_range: find delayed allocated block in the given range.
3804 * Return 1 if there is a delalloc block in the range, otherwise 0.
3806 int ext4_find_delalloc_range(struct inode *inode,
3807 ext4_lblk_t lblk_start,
3808 ext4_lblk_t lblk_end)
3810 struct extent_status es;
3812 ext4_es_find_delayed_extent_range(inode, lblk_start, lblk_end, &es);
3814 return 0; /* there is no delay extent in this tree */
3815 else if (es.es_lblk <= lblk_start &&
3816 lblk_start < es.es_lblk + es.es_len)
3818 else if (lblk_start <= es.es_lblk && es.es_lblk <= lblk_end)
3824 int ext4_find_delalloc_cluster(struct inode *inode, ext4_lblk_t lblk)
3826 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3827 ext4_lblk_t lblk_start, lblk_end;
3828 lblk_start = EXT4_LBLK_CMASK(sbi, lblk);
3829 lblk_end = lblk_start + sbi->s_cluster_ratio - 1;
3831 return ext4_find_delalloc_range(inode, lblk_start, lblk_end);
3835 * Determines how many complete clusters (out of those specified by the 'map')
3836 * are under delalloc and were reserved quota for.
3837 * This function is called when we are writing out the blocks that were
3838 * originally written with their allocation delayed, but then the space was
3839 * allocated using fallocate() before the delayed allocation could be resolved.
3840 * The cases to look for are:
3841 * ('=' indicated delayed allocated blocks
3842 * '-' indicates non-delayed allocated blocks)
3843 * (a) partial clusters towards beginning and/or end outside of allocated range
3844 * are not delalloc'ed.
3846 * |----c---=|====c====|====c====|===-c----|
3847 * |++++++ allocated ++++++|
3848 * ==> 4 complete clusters in above example
3850 * (b) partial cluster (outside of allocated range) towards either end is
3851 * marked for delayed allocation. In this case, we will exclude that
3854 * |----====c========|========c========|
3855 * |++++++ allocated ++++++|
3856 * ==> 1 complete clusters in above example
3859 * |================c================|
3860 * |++++++ allocated ++++++|
3861 * ==> 0 complete clusters in above example
3863 * The ext4_da_update_reserve_space will be called only if we
3864 * determine here that there were some "entire" clusters that span
3865 * this 'allocated' range.
3866 * In the non-bigalloc case, this function will just end up returning num_blks
3867 * without ever calling ext4_find_delalloc_range.
3870 get_reserved_cluster_alloc(struct inode *inode, ext4_lblk_t lblk_start,
3871 unsigned int num_blks)
3873 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3874 ext4_lblk_t alloc_cluster_start, alloc_cluster_end;
3875 ext4_lblk_t lblk_from, lblk_to, c_offset;
3876 unsigned int allocated_clusters = 0;
3878 alloc_cluster_start = EXT4_B2C(sbi, lblk_start);
3879 alloc_cluster_end = EXT4_B2C(sbi, lblk_start + num_blks - 1);
3881 /* max possible clusters for this allocation */
3882 allocated_clusters = alloc_cluster_end - alloc_cluster_start + 1;
3884 trace_ext4_get_reserved_cluster_alloc(inode, lblk_start, num_blks);
3886 /* Check towards left side */
3887 c_offset = EXT4_LBLK_COFF(sbi, lblk_start);
3889 lblk_from = EXT4_LBLK_CMASK(sbi, lblk_start);
3890 lblk_to = lblk_from + c_offset - 1;
3892 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
3893 allocated_clusters--;
3896 /* Now check towards right. */
3897 c_offset = EXT4_LBLK_COFF(sbi, lblk_start + num_blks);
3898 if (allocated_clusters && c_offset) {
3899 lblk_from = lblk_start + num_blks;
3900 lblk_to = lblk_from + (sbi->s_cluster_ratio - c_offset) - 1;
3902 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
3903 allocated_clusters--;
3906 return allocated_clusters;
3910 convert_initialized_extent(handle_t *handle, struct inode *inode,
3911 struct ext4_map_blocks *map,
3912 struct ext4_ext_path **ppath, int flags,
3913 unsigned int allocated, ext4_fsblk_t newblock)
3915 struct ext4_ext_path *path = *ppath;
3916 struct ext4_extent *ex;
3917 ext4_lblk_t ee_block;
3918 unsigned int ee_len;
3923 * Make sure that the extent is no bigger than we support with
3926 if (map->m_len > EXT_UNWRITTEN_MAX_LEN)
3927 map->m_len = EXT_UNWRITTEN_MAX_LEN / 2;
3929 depth = ext_depth(inode);
3930 ex = path[depth].p_ext;
3931 ee_block = le32_to_cpu(ex->ee_block);
3932 ee_len = ext4_ext_get_actual_len(ex);
3934 ext_debug("%s: inode %lu, logical"
3935 "block %llu, max_blocks %u\n", __func__, inode->i_ino,
3936 (unsigned long long)ee_block, ee_len);
3938 if (ee_block != map->m_lblk || ee_len > map->m_len) {
3939 err = ext4_split_convert_extents(handle, inode, map, ppath,
3940 EXT4_GET_BLOCKS_CONVERT_UNWRITTEN);
3943 path = ext4_find_extent(inode, map->m_lblk, ppath, 0);
3945 return PTR_ERR(path);
3946 depth = ext_depth(inode);
3947 ex = path[depth].p_ext;
3949 EXT4_ERROR_INODE(inode, "unexpected hole at %lu",
3950 (unsigned long) map->m_lblk);
3955 err = ext4_ext_get_access(handle, inode, path + depth);
3958 /* first mark the extent as unwritten */
3959 ext4_ext_mark_unwritten(ex);
3961 /* note: ext4_ext_correct_indexes() isn't needed here because
3962 * borders are not changed
3964 ext4_ext_try_to_merge(handle, inode, path, ex);
3966 /* Mark modified extent as dirty */
3967 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
3970 ext4_ext_show_leaf(inode, path);
3972 ext4_update_inode_fsync_trans(handle, inode, 1);
3973 err = check_eofblocks_fl(handle, inode, map->m_lblk, path, map->m_len);
3976 map->m_flags |= EXT4_MAP_UNWRITTEN;
3977 if (allocated > map->m_len)
3978 allocated = map->m_len;
3979 map->m_len = allocated;
3984 ext4_ext_handle_unwritten_extents(handle_t *handle, struct inode *inode,
3985 struct ext4_map_blocks *map,
3986 struct ext4_ext_path **ppath, int flags,
3987 unsigned int allocated, ext4_fsblk_t newblock)
3989 struct ext4_ext_path *path = *ppath;
3992 ext4_io_end_t *io = ext4_inode_aio(inode);
3994 ext_debug("ext4_ext_handle_unwritten_extents: inode %lu, logical "
3995 "block %llu, max_blocks %u, flags %x, allocated %u\n",
3996 inode->i_ino, (unsigned long long)map->m_lblk, map->m_len,
3998 ext4_ext_show_leaf(inode, path);
4001 * When writing into unwritten space, we should not fail to
4002 * allocate metadata blocks for the new extent block if needed.
4004 flags |= EXT4_GET_BLOCKS_METADATA_NOFAIL;
4006 trace_ext4_ext_handle_unwritten_extents(inode, map, flags,
4007 allocated, newblock);
4009 /* get_block() before submit the IO, split the extent */
4010 if (flags & EXT4_GET_BLOCKS_PRE_IO) {
4011 ret = ext4_split_convert_extents(handle, inode, map, ppath,
4012 flags | EXT4_GET_BLOCKS_CONVERT);
4016 * Flag the inode(non aio case) or end_io struct (aio case)
4017 * that this IO needs to conversion to written when IO is
4021 ext4_set_io_unwritten_flag(inode, io);
4023 ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
4024 map->m_flags |= EXT4_MAP_UNWRITTEN;
4027 /* IO end_io complete, convert the filled extent to written */
4028 if (flags & EXT4_GET_BLOCKS_CONVERT) {
4029 ret = ext4_convert_unwritten_extents_endio(handle, inode, map,
4032 ext4_update_inode_fsync_trans(handle, inode, 1);
4033 err = check_eofblocks_fl(handle, inode, map->m_lblk,
4037 map->m_flags |= EXT4_MAP_MAPPED;
4038 map->m_pblk = newblock;
4039 if (allocated > map->m_len)
4040 allocated = map->m_len;
4041 map->m_len = allocated;
4044 /* buffered IO case */
4046 * repeat fallocate creation request
4047 * we already have an unwritten extent
4049 if (flags & EXT4_GET_BLOCKS_UNWRIT_EXT) {
4050 map->m_flags |= EXT4_MAP_UNWRITTEN;
4054 /* buffered READ or buffered write_begin() lookup */
4055 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
4057 * We have blocks reserved already. We
4058 * return allocated blocks so that delalloc
4059 * won't do block reservation for us. But
4060 * the buffer head will be unmapped so that
4061 * a read from the block returns 0s.
4063 map->m_flags |= EXT4_MAP_UNWRITTEN;
4067 /* buffered write, writepage time, convert*/
4068 ret = ext4_ext_convert_to_initialized(handle, inode, map, ppath, flags);
4070 ext4_update_inode_fsync_trans(handle, inode, 1);
4077 map->m_flags |= EXT4_MAP_NEW;
4079 * if we allocated more blocks than requested
4080 * we need to make sure we unmap the extra block
4081 * allocated. The actual needed block will get
4082 * unmapped later when we find the buffer_head marked
4085 if (allocated > map->m_len) {
4086 unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
4087 newblock + map->m_len,
4088 allocated - map->m_len);
4089 allocated = map->m_len;
4091 map->m_len = allocated;
4094 * If we have done fallocate with the offset that is already
4095 * delayed allocated, we would have block reservation
4096 * and quota reservation done in the delayed write path.
4097 * But fallocate would have already updated quota and block
4098 * count for this offset. So cancel these reservation
4100 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
4101 unsigned int reserved_clusters;
4102 reserved_clusters = get_reserved_cluster_alloc(inode,
4103 map->m_lblk, map->m_len);
4104 if (reserved_clusters)
4105 ext4_da_update_reserve_space(inode,
4111 map->m_flags |= EXT4_MAP_MAPPED;
4112 if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0) {
4113 err = check_eofblocks_fl(handle, inode, map->m_lblk, path,
4119 if (allocated > map->m_len)
4120 allocated = map->m_len;
4121 ext4_ext_show_leaf(inode, path);
4122 map->m_pblk = newblock;
4123 map->m_len = allocated;
4125 return err ? err : allocated;
4129 * get_implied_cluster_alloc - check to see if the requested
4130 * allocation (in the map structure) overlaps with a cluster already
4131 * allocated in an extent.
4132 * @sb The filesystem superblock structure
4133 * @map The requested lblk->pblk mapping
4134 * @ex The extent structure which might contain an implied
4135 * cluster allocation
4137 * This function is called by ext4_ext_map_blocks() after we failed to
4138 * find blocks that were already in the inode's extent tree. Hence,
4139 * we know that the beginning of the requested region cannot overlap
4140 * the extent from the inode's extent tree. There are three cases we
4141 * want to catch. The first is this case:
4143 * |--- cluster # N--|
4144 * |--- extent ---| |---- requested region ---|
4147 * The second case that we need to test for is this one:
4149 * |--------- cluster # N ----------------|
4150 * |--- requested region --| |------- extent ----|
4151 * |=======================|
4153 * The third case is when the requested region lies between two extents
4154 * within the same cluster:
4155 * |------------- cluster # N-------------|
4156 * |----- ex -----| |---- ex_right ----|
4157 * |------ requested region ------|
4158 * |================|
4160 * In each of the above cases, we need to set the map->m_pblk and
4161 * map->m_len so it corresponds to the return the extent labelled as
4162 * "|====|" from cluster #N, since it is already in use for data in
4163 * cluster EXT4_B2C(sbi, map->m_lblk). We will then return 1 to
4164 * signal to ext4_ext_map_blocks() that map->m_pblk should be treated
4165 * as a new "allocated" block region. Otherwise, we will return 0 and
4166 * ext4_ext_map_blocks() will then allocate one or more new clusters
4167 * by calling ext4_mb_new_blocks().
4169 static int get_implied_cluster_alloc(struct super_block *sb,
4170 struct ext4_map_blocks *map,
4171 struct ext4_extent *ex,
4172 struct ext4_ext_path *path)
4174 struct ext4_sb_info *sbi = EXT4_SB(sb);
4175 ext4_lblk_t c_offset = EXT4_LBLK_COFF(sbi, map->m_lblk);
4176 ext4_lblk_t ex_cluster_start, ex_cluster_end;
4177 ext4_lblk_t rr_cluster_start;
4178 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
4179 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
4180 unsigned short ee_len = ext4_ext_get_actual_len(ex);
4182 /* The extent passed in that we are trying to match */
4183 ex_cluster_start = EXT4_B2C(sbi, ee_block);
4184 ex_cluster_end = EXT4_B2C(sbi, ee_block + ee_len - 1);
4186 /* The requested region passed into ext4_map_blocks() */
4187 rr_cluster_start = EXT4_B2C(sbi, map->m_lblk);
4189 if ((rr_cluster_start == ex_cluster_end) ||
4190 (rr_cluster_start == ex_cluster_start)) {
4191 if (rr_cluster_start == ex_cluster_end)
4192 ee_start += ee_len - 1;
4193 map->m_pblk = EXT4_PBLK_CMASK(sbi, ee_start) + c_offset;
4194 map->m_len = min(map->m_len,
4195 (unsigned) sbi->s_cluster_ratio - c_offset);
4197 * Check for and handle this case:
4199 * |--------- cluster # N-------------|
4200 * |------- extent ----|
4201 * |--- requested region ---|
4205 if (map->m_lblk < ee_block)
4206 map->m_len = min(map->m_len, ee_block - map->m_lblk);
4209 * Check for the case where there is already another allocated
4210 * block to the right of 'ex' but before the end of the cluster.
4212 * |------------- cluster # N-------------|
4213 * |----- ex -----| |---- ex_right ----|
4214 * |------ requested region ------|
4215 * |================|
4217 if (map->m_lblk > ee_block) {
4218 ext4_lblk_t next = ext4_ext_next_allocated_block(path);
4219 map->m_len = min(map->m_len, next - map->m_lblk);
4222 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 1);
4226 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 0);
4232 * Block allocation/map/preallocation routine for extents based files
4235 * Need to be called with
4236 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
4237 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
4239 * return > 0, number of of blocks already mapped/allocated
4240 * if create == 0 and these are pre-allocated blocks
4241 * buffer head is unmapped
4242 * otherwise blocks are mapped
4244 * return = 0, if plain look up failed (blocks have not been allocated)
4245 * buffer head is unmapped
4247 * return < 0, error case.
4249 int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
4250 struct ext4_map_blocks *map, int flags)
4252 struct ext4_ext_path *path = NULL;
4253 struct ext4_extent newex, *ex, *ex2;
4254 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
4255 ext4_fsblk_t newblock = 0;
4256 int free_on_err = 0, err = 0, depth, ret;
4257 unsigned int allocated = 0, offset = 0;
4258 unsigned int allocated_clusters = 0;
4259 struct ext4_allocation_request ar;
4260 ext4_io_end_t *io = ext4_inode_aio(inode);
4261 ext4_lblk_t cluster_offset;
4262 int set_unwritten = 0;
4264 ext_debug("blocks %u/%u requested for inode %lu\n",
4265 map->m_lblk, map->m_len, inode->i_ino);
4266 trace_ext4_ext_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
4268 /* find extent for this block */
4269 path = ext4_find_extent(inode, map->m_lblk, NULL, 0);
4271 err = PTR_ERR(path);
4276 depth = ext_depth(inode);
4279 * consistent leaf must not be empty;
4280 * this situation is possible, though, _during_ tree modification;
4281 * this is why assert can't be put in ext4_find_extent()
4283 if (unlikely(path[depth].p_ext == NULL && depth != 0)) {
4284 EXT4_ERROR_INODE(inode, "bad extent address "
4285 "lblock: %lu, depth: %d pblock %lld",
4286 (unsigned long) map->m_lblk, depth,
4287 path[depth].p_block);
4292 ex = path[depth].p_ext;
4294 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
4295 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
4296 unsigned short ee_len;
4300 * unwritten extents are treated as holes, except that
4301 * we split out initialized portions during a write.
4303 ee_len = ext4_ext_get_actual_len(ex);
4305 trace_ext4_ext_show_extent(inode, ee_block, ee_start, ee_len);
4307 /* if found extent covers block, simply return it */
4308 if (in_range(map->m_lblk, ee_block, ee_len)) {
4309 newblock = map->m_lblk - ee_block + ee_start;
4310 /* number of remaining blocks in the extent */
4311 allocated = ee_len - (map->m_lblk - ee_block);
4312 ext_debug("%u fit into %u:%d -> %llu\n", map->m_lblk,
4313 ee_block, ee_len, newblock);
4316 * If the extent is initialized check whether the
4317 * caller wants to convert it to unwritten.
4319 if ((!ext4_ext_is_unwritten(ex)) &&
4320 (flags & EXT4_GET_BLOCKS_CONVERT_UNWRITTEN)) {
4321 allocated = convert_initialized_extent(
4322 handle, inode, map, &path,
4323 flags, allocated, newblock);
4325 } else if (!ext4_ext_is_unwritten(ex))
4328 ret = ext4_ext_handle_unwritten_extents(
4329 handle, inode, map, &path, flags,
4330 allocated, newblock);
4339 if ((sbi->s_cluster_ratio > 1) &&
4340 ext4_find_delalloc_cluster(inode, map->m_lblk))
4341 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
4344 * requested block isn't allocated yet;
4345 * we couldn't try to create block if create flag is zero
4347 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
4349 * put just found gap into cache to speed up
4350 * subsequent requests
4352 if ((flags & EXT4_GET_BLOCKS_NO_PUT_HOLE) == 0)
4353 ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
4358 * Okay, we need to do block allocation.
4360 map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
4361 newex.ee_block = cpu_to_le32(map->m_lblk);
4362 cluster_offset = EXT4_LBLK_COFF(sbi, map->m_lblk);
4365 * If we are doing bigalloc, check to see if the extent returned
4366 * by ext4_find_extent() implies a cluster we can use.
4368 if (cluster_offset && ex &&
4369 get_implied_cluster_alloc(inode->i_sb, map, ex, path)) {
4370 ar.len = allocated = map->m_len;
4371 newblock = map->m_pblk;
4372 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
4373 goto got_allocated_blocks;
4376 /* find neighbour allocated blocks */
4377 ar.lleft = map->m_lblk;
4378 err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
4381 ar.lright = map->m_lblk;
4383 err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright, &ex2);
4387 /* Check if the extent after searching to the right implies a
4388 * cluster we can use. */
4389 if ((sbi->s_cluster_ratio > 1) && ex2 &&
4390 get_implied_cluster_alloc(inode->i_sb, map, ex2, path)) {
4391 ar.len = allocated = map->m_len;
4392 newblock = map->m_pblk;
4393 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
4394 goto got_allocated_blocks;
4398 * See if request is beyond maximum number of blocks we can have in
4399 * a single extent. For an initialized extent this limit is
4400 * EXT_INIT_MAX_LEN and for an unwritten extent this limit is
4401 * EXT_UNWRITTEN_MAX_LEN.
4403 if (map->m_len > EXT_INIT_MAX_LEN &&
4404 !(flags & EXT4_GET_BLOCKS_UNWRIT_EXT))
4405 map->m_len = EXT_INIT_MAX_LEN;
4406 else if (map->m_len > EXT_UNWRITTEN_MAX_LEN &&
4407 (flags & EXT4_GET_BLOCKS_UNWRIT_EXT))
4408 map->m_len = EXT_UNWRITTEN_MAX_LEN;
4410 /* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
4411 newex.ee_len = cpu_to_le16(map->m_len);
4412 err = ext4_ext_check_overlap(sbi, inode, &newex, path);
4414 allocated = ext4_ext_get_actual_len(&newex);
4416 allocated = map->m_len;
4418 /* allocate new block */
4420 ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
4421 ar.logical = map->m_lblk;
4423 * We calculate the offset from the beginning of the cluster
4424 * for the logical block number, since when we allocate a
4425 * physical cluster, the physical block should start at the
4426 * same offset from the beginning of the cluster. This is
4427 * needed so that future calls to get_implied_cluster_alloc()
4430 offset = EXT4_LBLK_COFF(sbi, map->m_lblk);
4431 ar.len = EXT4_NUM_B2C(sbi, offset+allocated);
4433 ar.logical -= offset;
4434 if (S_ISREG(inode->i_mode))
4435 ar.flags = EXT4_MB_HINT_DATA;
4437 /* disable in-core preallocation for non-regular files */
4439 if (flags & EXT4_GET_BLOCKS_NO_NORMALIZE)
4440 ar.flags |= EXT4_MB_HINT_NOPREALLOC;
4441 newblock = ext4_mb_new_blocks(handle, &ar, &err);
4444 ext_debug("allocate new block: goal %llu, found %llu/%u\n",
4445 ar.goal, newblock, allocated);
4447 allocated_clusters = ar.len;
4448 ar.len = EXT4_C2B(sbi, ar.len) - offset;
4449 if (ar.len > allocated)
4452 got_allocated_blocks:
4453 /* try to insert new extent into found leaf and return */
4454 ext4_ext_store_pblock(&newex, newblock + offset);
4455 newex.ee_len = cpu_to_le16(ar.len);
4456 /* Mark unwritten */
4457 if (flags & EXT4_GET_BLOCKS_UNWRIT_EXT){
4458 ext4_ext_mark_unwritten(&newex);
4459 map->m_flags |= EXT4_MAP_UNWRITTEN;
4461 * io_end structure was created for every IO write to an
4462 * unwritten extent. To avoid unnecessary conversion,
4463 * here we flag the IO that really needs the conversion.
4464 * For non asycn direct IO case, flag the inode state
4465 * that we need to perform conversion when IO is done.
4467 if (flags & EXT4_GET_BLOCKS_PRE_IO)
4472 if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0)
4473 err = check_eofblocks_fl(handle, inode, map->m_lblk,
4476 err = ext4_ext_insert_extent(handle, inode, &path,
4479 if (!err && set_unwritten) {
4481 ext4_set_io_unwritten_flag(inode, io);
4483 ext4_set_inode_state(inode,
4484 EXT4_STATE_DIO_UNWRITTEN);
4487 if (err && free_on_err) {
4488 int fb_flags = flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE ?
4489 EXT4_FREE_BLOCKS_NO_QUOT_UPDATE : 0;
4490 /* free data blocks we just allocated */
4491 /* not a good idea to call discard here directly,
4492 * but otherwise we'd need to call it every free() */
4493 ext4_discard_preallocations(inode);
4494 ext4_free_blocks(handle, inode, NULL, newblock,
4495 EXT4_C2B(sbi, allocated_clusters), fb_flags);
4499 /* previous routine could use block we allocated */
4500 newblock = ext4_ext_pblock(&newex);
4501 allocated = ext4_ext_get_actual_len(&newex);
4502 if (allocated > map->m_len)
4503 allocated = map->m_len;
4504 map->m_flags |= EXT4_MAP_NEW;
4507 * Update reserved blocks/metadata blocks after successful
4508 * block allocation which had been deferred till now.
4510 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
4511 unsigned int reserved_clusters;
4513 * Check how many clusters we had reserved this allocated range
4515 reserved_clusters = get_reserved_cluster_alloc(inode,
4516 map->m_lblk, allocated);
4517 if (map->m_flags & EXT4_MAP_FROM_CLUSTER) {
4518 if (reserved_clusters) {
4520 * We have clusters reserved for this range.
4521 * But since we are not doing actual allocation
4522 * and are simply using blocks from previously
4523 * allocated cluster, we should release the
4524 * reservation and not claim quota.
4526 ext4_da_update_reserve_space(inode,
4527 reserved_clusters, 0);
4530 BUG_ON(allocated_clusters < reserved_clusters);
4531 if (reserved_clusters < allocated_clusters) {
4532 struct ext4_inode_info *ei = EXT4_I(inode);
4533 int reservation = allocated_clusters -
4536 * It seems we claimed few clusters outside of
4537 * the range of this allocation. We should give
4538 * it back to the reservation pool. This can
4539 * happen in the following case:
4541 * * Suppose s_cluster_ratio is 4 (i.e., each
4542 * cluster has 4 blocks. Thus, the clusters
4543 * are [0-3],[4-7],[8-11]...
4544 * * First comes delayed allocation write for
4545 * logical blocks 10 & 11. Since there were no
4546 * previous delayed allocated blocks in the
4547 * range [8-11], we would reserve 1 cluster
4549 * * Next comes write for logical blocks 3 to 8.
4550 * In this case, we will reserve 2 clusters
4551 * (for [0-3] and [4-7]; and not for [8-11] as
4552 * that range has a delayed allocated blocks.
4553 * Thus total reserved clusters now becomes 3.
4554 * * Now, during the delayed allocation writeout
4555 * time, we will first write blocks [3-8] and
4556 * allocate 3 clusters for writing these
4557 * blocks. Also, we would claim all these
4558 * three clusters above.
4559 * * Now when we come here to writeout the
4560 * blocks [10-11], we would expect to claim
4561 * the reservation of 1 cluster we had made
4562 * (and we would claim it since there are no
4563 * more delayed allocated blocks in the range
4564 * [8-11]. But our reserved cluster count had
4565 * already gone to 0.
4567 * Thus, at the step 4 above when we determine
4568 * that there are still some unwritten delayed
4569 * allocated blocks outside of our current
4570 * block range, we should increment the
4571 * reserved clusters count so that when the
4572 * remaining blocks finally gets written, we
4575 dquot_reserve_block(inode,
4576 EXT4_C2B(sbi, reservation));
4577 spin_lock(&ei->i_block_reservation_lock);
4578 ei->i_reserved_data_blocks += reservation;
4579 spin_unlock(&ei->i_block_reservation_lock);
4582 * We will claim quota for all newly allocated blocks.
4583 * We're updating the reserved space *after* the
4584 * correction above so we do not accidentally free
4585 * all the metadata reservation because we might
4586 * actually need it later on.
4588 ext4_da_update_reserve_space(inode, allocated_clusters,
4594 * Cache the extent and update transaction to commit on fdatasync only
4595 * when it is _not_ an unwritten extent.
4597 if ((flags & EXT4_GET_BLOCKS_UNWRIT_EXT) == 0)
4598 ext4_update_inode_fsync_trans(handle, inode, 1);
4600 ext4_update_inode_fsync_trans(handle, inode, 0);
4602 if (allocated > map->m_len)
4603 allocated = map->m_len;
4604 ext4_ext_show_leaf(inode, path);
4605 map->m_flags |= EXT4_MAP_MAPPED;
4606 map->m_pblk = newblock;
4607 map->m_len = allocated;
4609 ext4_ext_drop_refs(path);
4612 trace_ext4_ext_map_blocks_exit(inode, flags, map,
4613 err ? err : allocated);
4614 ext4_es_lru_add(inode);
4615 return err ? err : allocated;
4618 void ext4_ext_truncate(handle_t *handle, struct inode *inode)
4620 struct super_block *sb = inode->i_sb;
4621 ext4_lblk_t last_block;
4625 * TODO: optimization is possible here.
4626 * Probably we need not scan at all,
4627 * because page truncation is enough.
4630 /* we have to know where to truncate from in crash case */
4631 EXT4_I(inode)->i_disksize = inode->i_size;
4632 ext4_mark_inode_dirty(handle, inode);
4634 last_block = (inode->i_size + sb->s_blocksize - 1)
4635 >> EXT4_BLOCK_SIZE_BITS(sb);
4637 err = ext4_es_remove_extent(inode, last_block,
4638 EXT_MAX_BLOCKS - last_block);
4639 if (err == -ENOMEM) {
4641 congestion_wait(BLK_RW_ASYNC, HZ/50);
4645 ext4_std_error(inode->i_sb, err);
4648 err = ext4_ext_remove_space(inode, last_block, EXT_MAX_BLOCKS - 1);
4649 ext4_std_error(inode->i_sb, err);
4652 static int ext4_alloc_file_blocks(struct file *file, ext4_lblk_t offset,
4653 ext4_lblk_t len, loff_t new_size,
4654 int flags, int mode)
4656 struct inode *inode = file_inode(file);
4661 struct ext4_map_blocks map;
4662 unsigned int credits;
4665 map.m_lblk = offset;
4668 * Don't normalize the request if it can fit in one extent so
4669 * that it doesn't get unnecessarily split into multiple
4672 if (len <= EXT_UNWRITTEN_MAX_LEN)
4673 flags |= EXT4_GET_BLOCKS_NO_NORMALIZE;
4676 * credits to insert 1 extent into extent tree
4678 credits = ext4_chunk_trans_blocks(inode, len);
4681 while (ret >= 0 && len) {
4682 handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS,
4684 if (IS_ERR(handle)) {
4685 ret = PTR_ERR(handle);
4688 ret = ext4_map_blocks(handle, inode, &map, flags);
4690 ext4_debug("inode #%lu: block %u: len %u: "
4691 "ext4_ext_map_blocks returned %d",
4692 inode->i_ino, map.m_lblk,
4694 ext4_mark_inode_dirty(handle, inode);
4695 ret2 = ext4_journal_stop(handle);
4699 map.m_len = len = len - ret;
4700 epos = (loff_t)map.m_lblk << inode->i_blkbits;
4701 inode->i_ctime = ext4_current_time(inode);
4703 if (epos > new_size)
4705 if (ext4_update_inode_size(inode, epos) & 0x1)
4706 inode->i_mtime = inode->i_ctime;
4708 if (epos > inode->i_size)
4709 ext4_set_inode_flag(inode,
4710 EXT4_INODE_EOFBLOCKS);
4712 ext4_mark_inode_dirty(handle, inode);
4713 ret2 = ext4_journal_stop(handle);
4717 if (ret == -ENOSPC &&
4718 ext4_should_retry_alloc(inode->i_sb, &retries)) {
4723 return ret > 0 ? ret2 : ret;
4726 static long ext4_zero_range(struct file *file, loff_t offset,
4727 loff_t len, int mode)
4729 struct inode *inode = file_inode(file);
4730 handle_t *handle = NULL;
4731 unsigned int max_blocks;
4732 loff_t new_size = 0;
4736 int partial_begin, partial_end;
4739 struct address_space *mapping = inode->i_mapping;
4740 unsigned int blkbits = inode->i_blkbits;
4742 trace_ext4_zero_range(inode, offset, len, mode);
4744 if (!S_ISREG(inode->i_mode))
4747 /* Call ext4_force_commit to flush all data in case of data=journal. */
4748 if (ext4_should_journal_data(inode)) {
4749 ret = ext4_force_commit(inode->i_sb);
4755 * Write out all dirty pages to avoid race conditions
4756 * Then release them.
4758 if (mapping->nrpages && mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
4759 ret = filemap_write_and_wait_range(mapping, offset,
4766 * Round up offset. This is not fallocate, we neet to zero out
4767 * blocks, so convert interior block aligned part of the range to
4768 * unwritten and possibly manually zero out unaligned parts of the
4771 start = round_up(offset, 1 << blkbits);
4772 end = round_down((offset + len), 1 << blkbits);
4774 if (start < offset || end > offset + len)
4776 partial_begin = offset & ((1 << blkbits) - 1);
4777 partial_end = (offset + len) & ((1 << blkbits) - 1);
4779 lblk = start >> blkbits;
4780 max_blocks = (end >> blkbits);
4781 if (max_blocks < lblk)
4786 flags = EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT |
4787 EXT4_GET_BLOCKS_CONVERT_UNWRITTEN |
4789 if (mode & FALLOC_FL_KEEP_SIZE)
4790 flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
4792 mutex_lock(&inode->i_mutex);
4795 * Indirect files do not support unwritten extnets
4797 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
4802 if (!(mode & FALLOC_FL_KEEP_SIZE) &&
4803 offset + len > i_size_read(inode)) {
4804 new_size = offset + len;
4805 ret = inode_newsize_ok(inode, new_size);
4809 * If we have a partial block after EOF we have to allocate
4816 if (max_blocks > 0) {
4818 /* Now release the pages and zero block aligned part of pages*/
4819 truncate_pagecache_range(inode, start, end - 1);
4820 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
4822 /* Wait all existing dio workers, newcomers will block on i_mutex */
4823 ext4_inode_block_unlocked_dio(inode);
4824 inode_dio_wait(inode);
4826 ret = ext4_alloc_file_blocks(file, lblk, max_blocks, new_size,
4831 * Remove entire range from the extent status tree.
4833 * ext4_es_remove_extent(inode, lblk, max_blocks) is
4834 * NOT sufficient. I'm not sure why this is the case,
4835 * but let's be conservative and remove the extent
4836 * status tree for the entire inode. There should be
4837 * no outstanding delalloc extents thanks to the
4838 * filemap_write_and_wait_range() call above.
4840 ret = ext4_es_remove_extent(inode, 0, EXT_MAX_BLOCKS);
4844 if (!partial_begin && !partial_end)
4848 * In worst case we have to writeout two nonadjacent unwritten
4849 * blocks and update the inode
4851 credits = (2 * ext4_ext_index_trans_blocks(inode, 2)) + 1;
4852 if (ext4_should_journal_data(inode))
4854 handle = ext4_journal_start(inode, EXT4_HT_MISC, credits);
4855 if (IS_ERR(handle)) {
4856 ret = PTR_ERR(handle);
4857 ext4_std_error(inode->i_sb, ret);
4861 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
4863 ext4_update_inode_size(inode, new_size);
4866 * Mark that we allocate beyond EOF so the subsequent truncate
4867 * can proceed even if the new size is the same as i_size.
4869 if ((offset + len) > i_size_read(inode))
4870 ext4_set_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
4872 ext4_mark_inode_dirty(handle, inode);
4874 /* Zero out partial block at the edges of the range */
4875 ret = ext4_zero_partial_blocks(handle, inode, offset, len);
4877 if (file->f_flags & O_SYNC)
4878 ext4_handle_sync(handle);
4880 ext4_journal_stop(handle);
4882 ext4_inode_resume_unlocked_dio(inode);
4884 mutex_unlock(&inode->i_mutex);
4889 * preallocate space for a file. This implements ext4's fallocate file
4890 * operation, which gets called from sys_fallocate system call.
4891 * For block-mapped files, posix_fallocate should fall back to the method
4892 * of writing zeroes to the required new blocks (the same behavior which is
4893 * expected for file systems which do not support fallocate() system call).
4895 long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
4897 struct inode *inode = file_inode(file);
4898 loff_t new_size = 0;
4899 unsigned int max_blocks;
4903 unsigned int blkbits = inode->i_blkbits;
4905 /* Return error if mode is not supported */
4906 if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE |
4907 FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE))
4910 if (mode & FALLOC_FL_PUNCH_HOLE)
4911 return ext4_punch_hole(inode, offset, len);
4913 ret = ext4_convert_inline_data(inode);
4918 * currently supporting (pre)allocate mode for extent-based
4921 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
4924 if (mode & FALLOC_FL_COLLAPSE_RANGE)
4925 return ext4_collapse_range(inode, offset, len);
4927 if (mode & FALLOC_FL_ZERO_RANGE)
4928 return ext4_zero_range(file, offset, len, mode);
4930 trace_ext4_fallocate_enter(inode, offset, len, mode);
4931 lblk = offset >> blkbits;
4933 * We can't just convert len to max_blocks because
4934 * If blocksize = 4096 offset = 3072 and len = 2048
4936 max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
4939 flags = EXT4_GET_BLOCKS_CREATE_UNWRIT_EXT;
4940 if (mode & FALLOC_FL_KEEP_SIZE)
4941 flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
4943 mutex_lock(&inode->i_mutex);
4945 if (!(mode & FALLOC_FL_KEEP_SIZE) &&
4946 offset + len > i_size_read(inode)) {
4947 new_size = offset + len;
4948 ret = inode_newsize_ok(inode, new_size);
4953 ret = ext4_alloc_file_blocks(file, lblk, max_blocks, new_size,
4958 if (file->f_flags & O_SYNC && EXT4_SB(inode->i_sb)->s_journal) {
4959 ret = jbd2_complete_transaction(EXT4_SB(inode->i_sb)->s_journal,
4960 EXT4_I(inode)->i_sync_tid);
4963 mutex_unlock(&inode->i_mutex);
4964 trace_ext4_fallocate_exit(inode, offset, max_blocks, ret);
4969 * This function convert a range of blocks to written extents
4970 * The caller of this function will pass the start offset and the size.
4971 * all unwritten extents within this range will be converted to
4974 * This function is called from the direct IO end io call back
4975 * function, to convert the fallocated extents after IO is completed.
4976 * Returns 0 on success.
4978 int ext4_convert_unwritten_extents(handle_t *handle, struct inode *inode,
4979 loff_t offset, ssize_t len)
4981 unsigned int max_blocks;
4984 struct ext4_map_blocks map;
4985 unsigned int credits, blkbits = inode->i_blkbits;
4987 map.m_lblk = offset >> blkbits;
4989 * We can't just convert len to max_blocks because
4990 * If blocksize = 4096 offset = 3072 and len = 2048
4992 max_blocks = ((EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) -
4995 * This is somewhat ugly but the idea is clear: When transaction is
4996 * reserved, everything goes into it. Otherwise we rather start several
4997 * smaller transactions for conversion of each extent separately.
5000 handle = ext4_journal_start_reserved(handle,
5001 EXT4_HT_EXT_CONVERT);
5003 return PTR_ERR(handle);
5007 * credits to insert 1 extent into extent tree
5009 credits = ext4_chunk_trans_blocks(inode, max_blocks);
5011 while (ret >= 0 && ret < max_blocks) {
5013 map.m_len = (max_blocks -= ret);
5015 handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS,
5017 if (IS_ERR(handle)) {
5018 ret = PTR_ERR(handle);
5022 ret = ext4_map_blocks(handle, inode, &map,
5023 EXT4_GET_BLOCKS_IO_CONVERT_EXT);
5025 ext4_warning(inode->i_sb,
5026 "inode #%lu: block %u: len %u: "
5027 "ext4_ext_map_blocks returned %d",
5028 inode->i_ino, map.m_lblk,
5030 ext4_mark_inode_dirty(handle, inode);
5032 ret2 = ext4_journal_stop(handle);
5033 if (ret <= 0 || ret2)
5037 ret2 = ext4_journal_stop(handle);
5038 return ret > 0 ? ret2 : ret;
5042 * If newes is not existing extent (newes->ec_pblk equals zero) find
5043 * delayed extent at start of newes and update newes accordingly and
5044 * return start of the next delayed extent.
5046 * If newes is existing extent (newes->ec_pblk is not equal zero)
5047 * return start of next delayed extent or EXT_MAX_BLOCKS if no delayed
5048 * extent found. Leave newes unmodified.
5050 static int ext4_find_delayed_extent(struct inode *inode,
5051 struct extent_status *newes)
5053 struct extent_status es;
5054 ext4_lblk_t block, next_del;
5056 if (newes->es_pblk == 0) {
5057 ext4_es_find_delayed_extent_range(inode, newes->es_lblk,
5058 newes->es_lblk + newes->es_len - 1, &es);
5061 * No extent in extent-tree contains block @newes->es_pblk,
5062 * then the block may stay in 1)a hole or 2)delayed-extent.
5068 if (es.es_lblk > newes->es_lblk) {
5070 newes->es_len = min(es.es_lblk - newes->es_lblk,
5075 newes->es_len = es.es_lblk + es.es_len - newes->es_lblk;
5078 block = newes->es_lblk + newes->es_len;
5079 ext4_es_find_delayed_extent_range(inode, block, EXT_MAX_BLOCKS, &es);
5081 next_del = EXT_MAX_BLOCKS;
5083 next_del = es.es_lblk;
5087 /* fiemap flags we can handle specified here */
5088 #define EXT4_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
5090 static int ext4_xattr_fiemap(struct inode *inode,
5091 struct fiemap_extent_info *fieinfo)
5095 __u32 flags = FIEMAP_EXTENT_LAST;
5096 int blockbits = inode->i_sb->s_blocksize_bits;
5100 if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
5101 struct ext4_iloc iloc;
5102 int offset; /* offset of xattr in inode */
5104 error = ext4_get_inode_loc(inode, &iloc);
5107 physical = (__u64)iloc.bh->b_blocknr << blockbits;
5108 offset = EXT4_GOOD_OLD_INODE_SIZE +
5109 EXT4_I(inode)->i_extra_isize;
5111 length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
5112 flags |= FIEMAP_EXTENT_DATA_INLINE;
5114 } else { /* external block */
5115 physical = (__u64)EXT4_I(inode)->i_file_acl << blockbits;
5116 length = inode->i_sb->s_blocksize;
5120 error = fiemap_fill_next_extent(fieinfo, 0, physical,
5122 return (error < 0 ? error : 0);
5125 int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
5126 __u64 start, __u64 len)
5128 ext4_lblk_t start_blk;
5131 if (ext4_has_inline_data(inode)) {
5134 error = ext4_inline_data_fiemap(inode, fieinfo, &has_inline);
5140 if (fieinfo->fi_flags & FIEMAP_FLAG_CACHE) {
5141 error = ext4_ext_precache(inode);
5146 /* fallback to generic here if not in extents fmt */
5147 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
5148 return generic_block_fiemap(inode, fieinfo, start, len,
5151 if (fiemap_check_flags(fieinfo, EXT4_FIEMAP_FLAGS))
5154 if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
5155 error = ext4_xattr_fiemap(inode, fieinfo);
5157 ext4_lblk_t len_blks;
5160 start_blk = start >> inode->i_sb->s_blocksize_bits;
5161 last_blk = (start + len - 1) >> inode->i_sb->s_blocksize_bits;
5162 if (last_blk >= EXT_MAX_BLOCKS)
5163 last_blk = EXT_MAX_BLOCKS-1;
5164 len_blks = ((ext4_lblk_t) last_blk) - start_blk + 1;
5167 * Walk the extent tree gathering extent information
5168 * and pushing extents back to the user.
5170 error = ext4_fill_fiemap_extents(inode, start_blk,
5173 ext4_es_lru_add(inode);
5179 * Function to access the path buffer for marking it dirty.
5180 * It also checks if there are sufficient credits left in the journal handle
5184 ext4_access_path(handle_t *handle, struct inode *inode,
5185 struct ext4_ext_path *path)
5189 if (!ext4_handle_valid(handle))
5193 * Check if need to extend journal credits
5194 * 3 for leaf, sb, and inode plus 2 (bmap and group
5195 * descriptor) for each block group; assume two block
5198 if (handle->h_buffer_credits < 7) {
5199 credits = ext4_writepage_trans_blocks(inode);
5200 err = ext4_ext_truncate_extend_restart(handle, inode, credits);
5201 /* EAGAIN is success */
5202 if (err && err != -EAGAIN)
5206 err = ext4_ext_get_access(handle, inode, path);
5211 * ext4_ext_shift_path_extents:
5212 * Shift the extents of a path structure lying between path[depth].p_ext
5213 * and EXT_LAST_EXTENT(path[depth].p_hdr) downwards, by subtracting shift
5214 * from starting block for each extent.
5217 ext4_ext_shift_path_extents(struct ext4_ext_path *path, ext4_lblk_t shift,
5218 struct inode *inode, handle_t *handle,
5222 struct ext4_extent *ex_start, *ex_last;
5224 depth = path->p_depth;
5226 while (depth >= 0) {
5227 if (depth == path->p_depth) {
5228 ex_start = path[depth].p_ext;
5232 ex_last = EXT_LAST_EXTENT(path[depth].p_hdr);
5236 err = ext4_access_path(handle, inode, path + depth);
5240 if (ex_start == EXT_FIRST_EXTENT(path[depth].p_hdr))
5243 *start = le32_to_cpu(ex_last->ee_block) +
5244 ext4_ext_get_actual_len(ex_last);
5246 while (ex_start <= ex_last) {
5247 le32_add_cpu(&ex_start->ee_block, -shift);
5248 /* Try to merge to the left. */
5250 EXT_FIRST_EXTENT(path[depth].p_hdr)) &&
5251 ext4_ext_try_to_merge_right(inode,
5252 path, ex_start - 1))
5257 err = ext4_ext_dirty(handle, inode, path + depth);
5261 if (--depth < 0 || !update)
5265 /* Update index too */
5266 err = ext4_access_path(handle, inode, path + depth);
5270 le32_add_cpu(&path[depth].p_idx->ei_block, -shift);
5271 err = ext4_ext_dirty(handle, inode, path + depth);
5275 /* we are done if current index is not a starting index */
5276 if (path[depth].p_idx != EXT_FIRST_INDEX(path[depth].p_hdr))
5287 * ext4_ext_shift_extents:
5288 * All the extents which lies in the range from start to the last allocated
5289 * block for the file are shifted downwards by shift blocks.
5290 * On success, 0 is returned, error otherwise.
5293 ext4_ext_shift_extents(struct inode *inode, handle_t *handle,
5294 ext4_lblk_t start, ext4_lblk_t shift)
5296 struct ext4_ext_path *path;
5298 struct ext4_extent *extent;
5299 ext4_lblk_t stop_block;
5300 ext4_lblk_t ex_start, ex_end;
5302 /* Let path point to the last extent */
5303 path = ext4_find_extent(inode, EXT_MAX_BLOCKS - 1, NULL, 0);
5305 return PTR_ERR(path);
5307 depth = path->p_depth;
5308 extent = path[depth].p_ext;
5312 stop_block = le32_to_cpu(extent->ee_block) +
5313 ext4_ext_get_actual_len(extent);
5315 /* Nothing to shift, if hole is at the end of file */
5316 if (start >= stop_block)
5320 * Don't start shifting extents until we make sure the hole is big
5321 * enough to accomodate the shift.
5323 path = ext4_find_extent(inode, start - 1, &path, 0);
5325 return PTR_ERR(path);
5326 depth = path->p_depth;
5327 extent = path[depth].p_ext;
5329 ex_start = le32_to_cpu(extent->ee_block);
5330 ex_end = le32_to_cpu(extent->ee_block) +
5331 ext4_ext_get_actual_len(extent);
5337 if ((start == ex_start && shift > ex_start) ||
5338 (shift > start - ex_end))
5341 /* Its safe to start updating extents */
5342 while (start < stop_block) {
5343 path = ext4_find_extent(inode, start, &path, 0);
5345 return PTR_ERR(path);
5346 depth = path->p_depth;
5347 extent = path[depth].p_ext;
5349 EXT4_ERROR_INODE(inode, "unexpected hole at %lu",
5350 (unsigned long) start);
5353 if (start > le32_to_cpu(extent->ee_block)) {
5354 /* Hole, move to the next extent */
5355 if (extent < EXT_LAST_EXTENT(path[depth].p_hdr)) {
5356 path[depth].p_ext++;
5358 start = ext4_ext_next_allocated_block(path);
5362 ret = ext4_ext_shift_path_extents(path, shift, inode,
5368 ext4_ext_drop_refs(path);
5374 * ext4_collapse_range:
5375 * This implements the fallocate's collapse range functionality for ext4
5376 * Returns: 0 and non-zero on error.
5378 int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
5380 struct super_block *sb = inode->i_sb;
5381 ext4_lblk_t punch_start, punch_stop;
5383 unsigned int credits;
5384 loff_t new_size, ioffset;
5387 /* Collapse range works only on fs block size aligned offsets. */
5388 if (offset & (EXT4_CLUSTER_SIZE(sb) - 1) ||
5389 len & (EXT4_CLUSTER_SIZE(sb) - 1))
5392 if (!S_ISREG(inode->i_mode))
5395 trace_ext4_collapse_range(inode, offset, len);
5397 punch_start = offset >> EXT4_BLOCK_SIZE_BITS(sb);
5398 punch_stop = (offset + len) >> EXT4_BLOCK_SIZE_BITS(sb);
5400 /* Call ext4_force_commit to flush all data in case of data=journal. */
5401 if (ext4_should_journal_data(inode)) {
5402 ret = ext4_force_commit(inode->i_sb);
5408 * Need to round down offset to be aligned with page size boundary
5409 * for page size > block size.
5411 ioffset = round_down(offset, PAGE_SIZE);
5413 /* Write out all dirty pages */
5414 ret = filemap_write_and_wait_range(inode->i_mapping, ioffset,
5419 /* Take mutex lock */
5420 mutex_lock(&inode->i_mutex);
5423 * There is no need to overlap collapse range with EOF, in which case
5424 * it is effectively a truncate operation
5426 if (offset + len >= i_size_read(inode)) {
5431 /* Currently just for extent based files */
5432 if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
5437 truncate_pagecache(inode, ioffset);
5439 /* Wait for existing dio to complete */
5440 ext4_inode_block_unlocked_dio(inode);
5441 inode_dio_wait(inode);
5443 credits = ext4_writepage_trans_blocks(inode);
5444 handle = ext4_journal_start(inode, EXT4_HT_TRUNCATE, credits);
5445 if (IS_ERR(handle)) {
5446 ret = PTR_ERR(handle);
5450 down_write(&EXT4_I(inode)->i_data_sem);
5451 ext4_discard_preallocations(inode);
5453 ret = ext4_es_remove_extent(inode, punch_start,
5454 EXT_MAX_BLOCKS - punch_start);
5456 up_write(&EXT4_I(inode)->i_data_sem);
5460 ret = ext4_ext_remove_space(inode, punch_start, punch_stop - 1);
5462 up_write(&EXT4_I(inode)->i_data_sem);
5465 ext4_discard_preallocations(inode);
5467 ret = ext4_ext_shift_extents(inode, handle, punch_stop,
5468 punch_stop - punch_start);
5470 up_write(&EXT4_I(inode)->i_data_sem);
5474 new_size = i_size_read(inode) - len;
5475 i_size_write(inode, new_size);
5476 EXT4_I(inode)->i_disksize = new_size;
5478 up_write(&EXT4_I(inode)->i_data_sem);
5480 ext4_handle_sync(handle);
5481 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
5482 ext4_mark_inode_dirty(handle, inode);
5485 ext4_journal_stop(handle);
5487 ext4_inode_resume_unlocked_dio(inode);
5489 mutex_unlock(&inode->i_mutex);
5494 * ext4_swap_extents - Swap extents between two inodes
5496 * @inode1: First inode
5497 * @inode2: Second inode
5498 * @lblk1: Start block for first inode
5499 * @lblk2: Start block for second inode
5500 * @count: Number of blocks to swap
5501 * @mark_unwritten: Mark second inode's extents as unwritten after swap
5502 * @erp: Pointer to save error value
5504 * This helper routine does exactly what is promise "swap extents". All other
5505 * stuff such as page-cache locking consistency, bh mapping consistency or
5506 * extent's data copying must be performed by caller.
5508 * i_mutex is held for both inodes
5509 * i_data_sem is locked for write for both inodes
5511 * All pages from requested range are locked for both inodes
5514 ext4_swap_extents(handle_t *handle, struct inode *inode1,
5515 struct inode *inode2, ext4_lblk_t lblk1, ext4_lblk_t lblk2,
5516 ext4_lblk_t count, int unwritten, int *erp)
5518 struct ext4_ext_path *path1 = NULL;
5519 struct ext4_ext_path *path2 = NULL;
5520 int replaced_count = 0;
5522 BUG_ON(!rwsem_is_locked(&EXT4_I(inode1)->i_data_sem));
5523 BUG_ON(!rwsem_is_locked(&EXT4_I(inode2)->i_data_sem));
5524 BUG_ON(!mutex_is_locked(&inode1->i_mutex));
5525 BUG_ON(!mutex_is_locked(&inode1->i_mutex));
5527 *erp = ext4_es_remove_extent(inode1, lblk1, count);
5530 *erp = ext4_es_remove_extent(inode2, lblk2, count);
5535 struct ext4_extent *ex1, *ex2, tmp_ex;
5536 ext4_lblk_t e1_blk, e2_blk;
5537 int e1_len, e2_len, len;
5540 path1 = ext4_find_extent(inode1, lblk1, NULL, EXT4_EX_NOCACHE);
5541 if (unlikely(IS_ERR(path1))) {
5542 *erp = PTR_ERR(path1);
5548 path2 = ext4_find_extent(inode2, lblk2, NULL, EXT4_EX_NOCACHE);
5549 if (unlikely(IS_ERR(path2))) {
5550 *erp = PTR_ERR(path2);
5554 ex1 = path1[path1->p_depth].p_ext;
5555 ex2 = path2[path2->p_depth].p_ext;
5556 /* Do we have somthing to swap ? */
5557 if (unlikely(!ex2 || !ex1))
5560 e1_blk = le32_to_cpu(ex1->ee_block);
5561 e2_blk = le32_to_cpu(ex2->ee_block);
5562 e1_len = ext4_ext_get_actual_len(ex1);
5563 e2_len = ext4_ext_get_actual_len(ex2);
5566 if (!in_range(lblk1, e1_blk, e1_len) ||
5567 !in_range(lblk2, e2_blk, e2_len)) {
5568 ext4_lblk_t next1, next2;
5570 /* if hole after extent, then go to next extent */
5571 next1 = ext4_ext_next_allocated_block(path1);
5572 next2 = ext4_ext_next_allocated_block(path2);
5573 /* If hole before extent, then shift to that extent */
5578 /* Do we have something to swap */
5579 if (next1 == EXT_MAX_BLOCKS || next2 == EXT_MAX_BLOCKS)
5581 /* Move to the rightest boundary */
5582 len = next1 - lblk1;
5583 if (len < next2 - lblk2)
5584 len = next2 - lblk2;
5593 /* Prepare left boundary */
5594 if (e1_blk < lblk1) {
5596 *erp = ext4_force_split_extent_at(handle, inode1,
5601 if (e2_blk < lblk2) {
5603 *erp = ext4_force_split_extent_at(handle, inode2,
5608 /* ext4_split_extent_at() may result in leaf extent split,
5609 * path must to be revalidated. */
5613 /* Prepare right boundary */
5615 if (len > e1_blk + e1_len - lblk1)
5616 len = e1_blk + e1_len - lblk1;
5617 if (len > e2_blk + e2_len - lblk2)
5618 len = e2_blk + e2_len - lblk2;
5620 if (len != e1_len) {
5622 *erp = ext4_force_split_extent_at(handle, inode1,
5623 &path1, lblk1 + len, 0);
5627 if (len != e2_len) {
5629 *erp = ext4_force_split_extent_at(handle, inode2,
5630 &path2, lblk2 + len, 0);
5634 /* ext4_split_extent_at() may result in leaf extent split,
5635 * path must to be revalidated. */
5639 BUG_ON(e2_len != e1_len);
5640 *erp = ext4_ext_get_access(handle, inode1, path1 + path1->p_depth);
5643 *erp = ext4_ext_get_access(handle, inode2, path2 + path2->p_depth);
5647 /* Both extents are fully inside boundaries. Swap it now */
5649 ext4_ext_store_pblock(ex1, ext4_ext_pblock(ex2));
5650 ext4_ext_store_pblock(ex2, ext4_ext_pblock(&tmp_ex));
5651 ex1->ee_len = cpu_to_le16(e2_len);
5652 ex2->ee_len = cpu_to_le16(e1_len);
5654 ext4_ext_mark_unwritten(ex2);
5655 if (ext4_ext_is_unwritten(&tmp_ex))
5656 ext4_ext_mark_unwritten(ex1);
5658 ext4_ext_try_to_merge(handle, inode2, path2, ex2);
5659 ext4_ext_try_to_merge(handle, inode1, path1, ex1);
5660 *erp = ext4_ext_dirty(handle, inode2, path2 +
5664 *erp = ext4_ext_dirty(handle, inode1, path1 +
5667 * Looks scarry ah..? second inode already points to new blocks,
5668 * and it was successfully dirtied. But luckily error may happen
5669 * only due to journal error, so full transaction will be
5676 replaced_count += len;
5680 ext4_ext_drop_refs(path1);
5682 ext4_ext_drop_refs(path2);
5684 path1 = path2 = NULL;
5686 return replaced_count;