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
32 #include <linux/module.h>
34 #include <linux/time.h>
35 #include <linux/jbd2.h>
36 #include <linux/highuid.h>
37 #include <linux/pagemap.h>
38 #include <linux/quotaops.h>
39 #include <linux/string.h>
40 #include <linux/slab.h>
41 #include <linux/falloc.h>
42 #include <asm/uaccess.h>
43 #include <linux/fiemap.h>
44 #include "ext4_jbd2.h"
45 #include "ext4_extents.h"
50 * combine low and high parts of physical block number into ext4_fsblk_t
52 ext4_fsblk_t ext_pblock(struct ext4_extent *ex)
56 block = le32_to_cpu(ex->ee_start_lo);
57 block |= ((ext4_fsblk_t) le16_to_cpu(ex->ee_start_hi) << 31) << 1;
63 * combine low and high parts of a leaf physical block number into ext4_fsblk_t
65 ext4_fsblk_t idx_pblock(struct ext4_extent_idx *ix)
69 block = le32_to_cpu(ix->ei_leaf_lo);
70 block |= ((ext4_fsblk_t) le16_to_cpu(ix->ei_leaf_hi) << 31) << 1;
75 * ext4_ext_store_pblock:
76 * stores a large physical block number into an extent struct,
77 * breaking it into parts
79 void ext4_ext_store_pblock(struct ext4_extent *ex, ext4_fsblk_t pb)
81 ex->ee_start_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff));
82 ex->ee_start_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
86 * ext4_idx_store_pblock:
87 * stores a large physical block number into an index struct,
88 * breaking it into parts
90 static void ext4_idx_store_pblock(struct ext4_extent_idx *ix, ext4_fsblk_t pb)
92 ix->ei_leaf_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff));
93 ix->ei_leaf_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
96 static int ext4_ext_journal_restart(handle_t *handle, int needed)
100 if (!ext4_handle_valid(handle))
102 if (handle->h_buffer_credits > needed)
104 err = ext4_journal_extend(handle, needed);
107 return ext4_journal_restart(handle, needed);
115 static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
116 struct ext4_ext_path *path)
119 /* path points to block */
120 return ext4_journal_get_write_access(handle, path->p_bh);
122 /* path points to leaf/index in inode body */
123 /* we use in-core data, no need to protect them */
133 static int ext4_ext_dirty(handle_t *handle, struct inode *inode,
134 struct ext4_ext_path *path)
138 /* path points to block */
139 err = ext4_handle_dirty_metadata(handle, inode, path->p_bh);
141 /* path points to leaf/index in inode body */
142 err = ext4_mark_inode_dirty(handle, inode);
147 static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
148 struct ext4_ext_path *path,
151 struct ext4_inode_info *ei = EXT4_I(inode);
152 ext4_fsblk_t bg_start;
153 ext4_fsblk_t last_block;
154 ext4_grpblk_t colour;
155 ext4_group_t block_group;
156 int flex_size = ext4_flex_bg_size(EXT4_SB(inode->i_sb));
160 struct ext4_extent *ex;
161 depth = path->p_depth;
163 /* try to predict block placement */
164 ex = path[depth].p_ext;
166 return ext_pblock(ex)+(block-le32_to_cpu(ex->ee_block));
168 /* it looks like index is empty;
169 * try to find starting block from index itself */
170 if (path[depth].p_bh)
171 return path[depth].p_bh->b_blocknr;
174 /* OK. use inode's group */
175 block_group = ei->i_block_group;
176 if (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) {
178 * If there are at least EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME
179 * block groups per flexgroup, reserve the first block
180 * group for directories and special files. Regular
181 * files will start at the second block group. This
182 * tends to speed up directory access and improves
185 block_group &= ~(flex_size-1);
186 if (S_ISREG(inode->i_mode))
189 bg_start = (block_group * EXT4_BLOCKS_PER_GROUP(inode->i_sb)) +
190 le32_to_cpu(EXT4_SB(inode->i_sb)->s_es->s_first_data_block);
191 last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;
194 * If we are doing delayed allocation, we don't need take
195 * colour into account.
197 if (test_opt(inode->i_sb, DELALLOC))
200 if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
201 colour = (current->pid % 16) *
202 (EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
204 colour = (current->pid % 16) * ((last_block - bg_start) / 16);
205 return bg_start + colour + block;
209 * Allocation for a meta data block
212 ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
213 struct ext4_ext_path *path,
214 struct ext4_extent *ex, int *err)
216 ext4_fsblk_t goal, newblock;
218 goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
219 newblock = ext4_new_meta_blocks(handle, inode, goal, NULL, err);
223 static int ext4_ext_space_block(struct inode *inode)
227 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
228 / sizeof(struct ext4_extent);
229 #ifdef AGGRESSIVE_TEST
236 static int ext4_ext_space_block_idx(struct inode *inode)
240 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
241 / sizeof(struct ext4_extent_idx);
242 #ifdef AGGRESSIVE_TEST
249 static int ext4_ext_space_root(struct inode *inode)
253 size = sizeof(EXT4_I(inode)->i_data);
254 size -= sizeof(struct ext4_extent_header);
255 size /= sizeof(struct ext4_extent);
256 #ifdef AGGRESSIVE_TEST
263 static int ext4_ext_space_root_idx(struct inode *inode)
267 size = sizeof(EXT4_I(inode)->i_data);
268 size -= sizeof(struct ext4_extent_header);
269 size /= sizeof(struct ext4_extent_idx);
270 #ifdef AGGRESSIVE_TEST
278 * Calculate the number of metadata blocks needed
279 * to allocate @blocks
280 * Worse case is one block per extent
282 int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks)
284 int lcap, icap, rcap, leafs, idxs, num;
285 int newextents = blocks;
287 rcap = ext4_ext_space_root_idx(inode);
288 lcap = ext4_ext_space_block(inode);
289 icap = ext4_ext_space_block_idx(inode);
291 /* number of new leaf blocks needed */
292 num = leafs = (newextents + lcap - 1) / lcap;
295 * Worse case, we need separate index block(s)
296 * to link all new leaf blocks
298 idxs = (leafs + icap - 1) / icap;
301 idxs = (idxs + icap - 1) / icap;
302 } while (idxs > rcap);
308 ext4_ext_max_entries(struct inode *inode, int depth)
312 if (depth == ext_depth(inode)) {
314 max = ext4_ext_space_root(inode);
316 max = ext4_ext_space_root_idx(inode);
319 max = ext4_ext_space_block(inode);
321 max = ext4_ext_space_block_idx(inode);
327 static int ext4_valid_extent(struct inode *inode, struct ext4_extent *ext)
329 ext4_fsblk_t block = ext_pblock(ext);
330 int len = ext4_ext_get_actual_len(ext);
332 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, len);
335 static int ext4_valid_extent_idx(struct inode *inode,
336 struct ext4_extent_idx *ext_idx)
338 ext4_fsblk_t block = idx_pblock(ext_idx);
340 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, 1);
343 static int ext4_valid_extent_entries(struct inode *inode,
344 struct ext4_extent_header *eh,
347 struct ext4_extent *ext;
348 struct ext4_extent_idx *ext_idx;
349 unsigned short entries;
350 if (eh->eh_entries == 0)
353 entries = le16_to_cpu(eh->eh_entries);
357 ext = EXT_FIRST_EXTENT(eh);
359 if (!ext4_valid_extent(inode, ext))
365 ext_idx = EXT_FIRST_INDEX(eh);
367 if (!ext4_valid_extent_idx(inode, ext_idx))
376 static int __ext4_ext_check(const char *function, struct inode *inode,
377 struct ext4_extent_header *eh,
380 const char *error_msg;
383 if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
384 error_msg = "invalid magic";
387 if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
388 error_msg = "unexpected eh_depth";
391 if (unlikely(eh->eh_max == 0)) {
392 error_msg = "invalid eh_max";
395 max = ext4_ext_max_entries(inode, depth);
396 if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
397 error_msg = "too large eh_max";
400 if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
401 error_msg = "invalid eh_entries";
404 if (!ext4_valid_extent_entries(inode, eh, depth)) {
405 error_msg = "invalid extent entries";
411 ext4_error(inode->i_sb, function,
412 "bad header/extent in inode #%lu: %s - magic %x, "
413 "entries %u, max %u(%u), depth %u(%u)",
414 inode->i_ino, error_msg, le16_to_cpu(eh->eh_magic),
415 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
416 max, le16_to_cpu(eh->eh_depth), depth);
421 #define ext4_ext_check(inode, eh, depth) \
422 __ext4_ext_check(__func__, inode, eh, depth)
424 int ext4_ext_check_inode(struct inode *inode)
426 return ext4_ext_check(inode, ext_inode_hdr(inode), ext_depth(inode));
430 static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
432 int k, l = path->p_depth;
435 for (k = 0; k <= l; k++, path++) {
437 ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block),
438 idx_pblock(path->p_idx));
439 } else if (path->p_ext) {
440 ext_debug(" %d:[%d]%d:%llu ",
441 le32_to_cpu(path->p_ext->ee_block),
442 ext4_ext_is_uninitialized(path->p_ext),
443 ext4_ext_get_actual_len(path->p_ext),
444 ext_pblock(path->p_ext));
451 static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
453 int depth = ext_depth(inode);
454 struct ext4_extent_header *eh;
455 struct ext4_extent *ex;
461 eh = path[depth].p_hdr;
462 ex = EXT_FIRST_EXTENT(eh);
464 ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
466 for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
467 ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
468 ext4_ext_is_uninitialized(ex),
469 ext4_ext_get_actual_len(ex), ext_pblock(ex));
474 #define ext4_ext_show_path(inode, path)
475 #define ext4_ext_show_leaf(inode, path)
478 void ext4_ext_drop_refs(struct ext4_ext_path *path)
480 int depth = path->p_depth;
483 for (i = 0; i <= depth; i++, path++)
491 * ext4_ext_binsearch_idx:
492 * binary search for the closest index of the given block
493 * the header must be checked before calling this
496 ext4_ext_binsearch_idx(struct inode *inode,
497 struct ext4_ext_path *path, ext4_lblk_t block)
499 struct ext4_extent_header *eh = path->p_hdr;
500 struct ext4_extent_idx *r, *l, *m;
503 ext_debug("binsearch for %u(idx): ", block);
505 l = EXT_FIRST_INDEX(eh) + 1;
506 r = EXT_LAST_INDEX(eh);
509 if (block < le32_to_cpu(m->ei_block))
513 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
514 m, le32_to_cpu(m->ei_block),
515 r, le32_to_cpu(r->ei_block));
519 ext_debug(" -> %d->%lld ", le32_to_cpu(path->p_idx->ei_block),
520 idx_pblock(path->p_idx));
522 #ifdef CHECK_BINSEARCH
524 struct ext4_extent_idx *chix, *ix;
527 chix = ix = EXT_FIRST_INDEX(eh);
528 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
530 le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
531 printk(KERN_DEBUG "k=%d, ix=0x%p, "
533 ix, EXT_FIRST_INDEX(eh));
534 printk(KERN_DEBUG "%u <= %u\n",
535 le32_to_cpu(ix->ei_block),
536 le32_to_cpu(ix[-1].ei_block));
538 BUG_ON(k && le32_to_cpu(ix->ei_block)
539 <= le32_to_cpu(ix[-1].ei_block));
540 if (block < le32_to_cpu(ix->ei_block))
544 BUG_ON(chix != path->p_idx);
551 * ext4_ext_binsearch:
552 * binary search for closest extent of the given block
553 * the header must be checked before calling this
556 ext4_ext_binsearch(struct inode *inode,
557 struct ext4_ext_path *path, ext4_lblk_t block)
559 struct ext4_extent_header *eh = path->p_hdr;
560 struct ext4_extent *r, *l, *m;
562 if (eh->eh_entries == 0) {
564 * this leaf is empty:
565 * we get such a leaf in split/add case
570 ext_debug("binsearch for %u: ", block);
572 l = EXT_FIRST_EXTENT(eh) + 1;
573 r = EXT_LAST_EXTENT(eh);
577 if (block < le32_to_cpu(m->ee_block))
581 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
582 m, le32_to_cpu(m->ee_block),
583 r, le32_to_cpu(r->ee_block));
587 ext_debug(" -> %d:%llu:[%d]%d ",
588 le32_to_cpu(path->p_ext->ee_block),
589 ext_pblock(path->p_ext),
590 ext4_ext_is_uninitialized(path->p_ext),
591 ext4_ext_get_actual_len(path->p_ext));
593 #ifdef CHECK_BINSEARCH
595 struct ext4_extent *chex, *ex;
598 chex = ex = EXT_FIRST_EXTENT(eh);
599 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
600 BUG_ON(k && le32_to_cpu(ex->ee_block)
601 <= le32_to_cpu(ex[-1].ee_block));
602 if (block < le32_to_cpu(ex->ee_block))
606 BUG_ON(chex != path->p_ext);
612 int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
614 struct ext4_extent_header *eh;
616 eh = ext_inode_hdr(inode);
619 eh->eh_magic = EXT4_EXT_MAGIC;
620 eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode));
621 ext4_mark_inode_dirty(handle, inode);
622 ext4_ext_invalidate_cache(inode);
626 struct ext4_ext_path *
627 ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block,
628 struct ext4_ext_path *path)
630 struct ext4_extent_header *eh;
631 struct buffer_head *bh;
632 short int depth, i, ppos = 0, alloc = 0;
634 eh = ext_inode_hdr(inode);
635 depth = ext_depth(inode);
637 /* account possible depth increase */
639 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
642 return ERR_PTR(-ENOMEM);
649 /* walk through the tree */
651 int need_to_validate = 0;
653 ext_debug("depth %d: num %d, max %d\n",
654 ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
656 ext4_ext_binsearch_idx(inode, path + ppos, block);
657 path[ppos].p_block = idx_pblock(path[ppos].p_idx);
658 path[ppos].p_depth = i;
659 path[ppos].p_ext = NULL;
661 bh = sb_getblk(inode->i_sb, path[ppos].p_block);
664 if (!bh_uptodate_or_lock(bh)) {
665 if (bh_submit_read(bh) < 0) {
669 /* validate the extent entries */
670 need_to_validate = 1;
672 eh = ext_block_hdr(bh);
674 BUG_ON(ppos > depth);
675 path[ppos].p_bh = bh;
676 path[ppos].p_hdr = eh;
679 if (need_to_validate && ext4_ext_check(inode, eh, i))
683 path[ppos].p_depth = i;
684 path[ppos].p_ext = NULL;
685 path[ppos].p_idx = NULL;
688 ext4_ext_binsearch(inode, path + ppos, block);
689 /* if not an empty leaf */
690 if (path[ppos].p_ext)
691 path[ppos].p_block = ext_pblock(path[ppos].p_ext);
693 ext4_ext_show_path(inode, path);
698 ext4_ext_drop_refs(path);
701 return ERR_PTR(-EIO);
705 * ext4_ext_insert_index:
706 * insert new index [@logical;@ptr] into the block at @curp;
707 * check where to insert: before @curp or after @curp
709 static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
710 struct ext4_ext_path *curp,
711 int logical, ext4_fsblk_t ptr)
713 struct ext4_extent_idx *ix;
716 err = ext4_ext_get_access(handle, inode, curp);
720 BUG_ON(logical == le32_to_cpu(curp->p_idx->ei_block));
721 len = EXT_MAX_INDEX(curp->p_hdr) - curp->p_idx;
722 if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
724 if (curp->p_idx != EXT_LAST_INDEX(curp->p_hdr)) {
725 len = (len - 1) * sizeof(struct ext4_extent_idx);
726 len = len < 0 ? 0 : len;
727 ext_debug("insert new index %d after: %llu. "
728 "move %d from 0x%p to 0x%p\n",
730 (curp->p_idx + 1), (curp->p_idx + 2));
731 memmove(curp->p_idx + 2, curp->p_idx + 1, len);
733 ix = curp->p_idx + 1;
736 len = len * sizeof(struct ext4_extent_idx);
737 len = len < 0 ? 0 : len;
738 ext_debug("insert new index %d before: %llu. "
739 "move %d from 0x%p to 0x%p\n",
741 curp->p_idx, (curp->p_idx + 1));
742 memmove(curp->p_idx + 1, curp->p_idx, len);
746 ix->ei_block = cpu_to_le32(logical);
747 ext4_idx_store_pblock(ix, ptr);
748 le16_add_cpu(&curp->p_hdr->eh_entries, 1);
750 BUG_ON(le16_to_cpu(curp->p_hdr->eh_entries)
751 > le16_to_cpu(curp->p_hdr->eh_max));
752 BUG_ON(ix > EXT_LAST_INDEX(curp->p_hdr));
754 err = ext4_ext_dirty(handle, inode, curp);
755 ext4_std_error(inode->i_sb, err);
762 * inserts new subtree into the path, using free index entry
764 * - allocates all needed blocks (new leaf and all intermediate index blocks)
765 * - makes decision where to split
766 * - moves remaining extents and index entries (right to the split point)
767 * into the newly allocated blocks
768 * - initializes subtree
770 static int ext4_ext_split(handle_t *handle, struct inode *inode,
771 struct ext4_ext_path *path,
772 struct ext4_extent *newext, int at)
774 struct buffer_head *bh = NULL;
775 int depth = ext_depth(inode);
776 struct ext4_extent_header *neh;
777 struct ext4_extent_idx *fidx;
778 struct ext4_extent *ex;
780 ext4_fsblk_t newblock, oldblock;
782 ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
785 /* make decision: where to split? */
786 /* FIXME: now decision is simplest: at current extent */
788 /* if current leaf will be split, then we should use
789 * border from split point */
790 BUG_ON(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr));
791 if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
792 border = path[depth].p_ext[1].ee_block;
793 ext_debug("leaf will be split."
794 " next leaf starts at %d\n",
795 le32_to_cpu(border));
797 border = newext->ee_block;
798 ext_debug("leaf will be added."
799 " next leaf starts at %d\n",
800 le32_to_cpu(border));
804 * If error occurs, then we break processing
805 * and mark filesystem read-only. index won't
806 * be inserted and tree will be in consistent
807 * state. Next mount will repair buffers too.
811 * Get array to track all allocated blocks.
812 * We need this to handle errors and free blocks
815 ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
819 /* allocate all needed blocks */
820 ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
821 for (a = 0; a < depth - at; a++) {
822 newblock = ext4_ext_new_meta_block(handle, inode, path,
826 ablocks[a] = newblock;
829 /* initialize new leaf */
830 newblock = ablocks[--a];
831 BUG_ON(newblock == 0);
832 bh = sb_getblk(inode->i_sb, newblock);
839 err = ext4_journal_get_create_access(handle, bh);
843 neh = ext_block_hdr(bh);
845 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
846 neh->eh_magic = EXT4_EXT_MAGIC;
848 ex = EXT_FIRST_EXTENT(neh);
850 /* move remainder of path[depth] to the new leaf */
851 BUG_ON(path[depth].p_hdr->eh_entries != path[depth].p_hdr->eh_max);
852 /* start copy from next extent */
853 /* TODO: we could do it by single memmove */
856 while (path[depth].p_ext <=
857 EXT_MAX_EXTENT(path[depth].p_hdr)) {
858 ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
859 le32_to_cpu(path[depth].p_ext->ee_block),
860 ext_pblock(path[depth].p_ext),
861 ext4_ext_is_uninitialized(path[depth].p_ext),
862 ext4_ext_get_actual_len(path[depth].p_ext),
864 /*memmove(ex++, path[depth].p_ext++,
865 sizeof(struct ext4_extent));
871 memmove(ex, path[depth].p_ext-m, sizeof(struct ext4_extent)*m);
872 le16_add_cpu(&neh->eh_entries, m);
875 set_buffer_uptodate(bh);
878 err = ext4_handle_dirty_metadata(handle, inode, bh);
884 /* correct old leaf */
886 err = ext4_ext_get_access(handle, inode, path + depth);
889 le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
890 err = ext4_ext_dirty(handle, inode, path + depth);
896 /* create intermediate indexes */
900 ext_debug("create %d intermediate indices\n", k);
901 /* insert new index into current index block */
902 /* current depth stored in i var */
906 newblock = ablocks[--a];
907 bh = sb_getblk(inode->i_sb, newblock);
914 err = ext4_journal_get_create_access(handle, bh);
918 neh = ext_block_hdr(bh);
919 neh->eh_entries = cpu_to_le16(1);
920 neh->eh_magic = EXT4_EXT_MAGIC;
921 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
922 neh->eh_depth = cpu_to_le16(depth - i);
923 fidx = EXT_FIRST_INDEX(neh);
924 fidx->ei_block = border;
925 ext4_idx_store_pblock(fidx, oldblock);
927 ext_debug("int.index at %d (block %llu): %u -> %llu\n",
928 i, newblock, le32_to_cpu(border), oldblock);
933 ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
934 EXT_MAX_INDEX(path[i].p_hdr));
935 BUG_ON(EXT_MAX_INDEX(path[i].p_hdr) !=
936 EXT_LAST_INDEX(path[i].p_hdr));
937 while (path[i].p_idx <= EXT_MAX_INDEX(path[i].p_hdr)) {
938 ext_debug("%d: move %d:%llu in new index %llu\n", i,
939 le32_to_cpu(path[i].p_idx->ei_block),
940 idx_pblock(path[i].p_idx),
942 /*memmove(++fidx, path[i].p_idx++,
943 sizeof(struct ext4_extent_idx));
945 BUG_ON(neh->eh_entries > neh->eh_max);*/
950 memmove(++fidx, path[i].p_idx - m,
951 sizeof(struct ext4_extent_idx) * m);
952 le16_add_cpu(&neh->eh_entries, m);
954 set_buffer_uptodate(bh);
957 err = ext4_handle_dirty_metadata(handle, inode, bh);
963 /* correct old index */
965 err = ext4_ext_get_access(handle, inode, path + i);
968 le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
969 err = ext4_ext_dirty(handle, inode, path + i);
977 /* insert new index */
978 err = ext4_ext_insert_index(handle, inode, path + at,
979 le32_to_cpu(border), newblock);
983 if (buffer_locked(bh))
989 /* free all allocated blocks in error case */
990 for (i = 0; i < depth; i++) {
993 ext4_free_blocks(handle, inode, ablocks[i], 1, 1);
1002 * ext4_ext_grow_indepth:
1003 * implements tree growing procedure:
1004 * - allocates new block
1005 * - moves top-level data (index block or leaf) into the new block
1006 * - initializes new top-level, creating index that points to the
1007 * just created block
1009 static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
1010 struct ext4_ext_path *path,
1011 struct ext4_extent *newext)
1013 struct ext4_ext_path *curp = path;
1014 struct ext4_extent_header *neh;
1015 struct ext4_extent_idx *fidx;
1016 struct buffer_head *bh;
1017 ext4_fsblk_t newblock;
1020 newblock = ext4_ext_new_meta_block(handle, inode, path, newext, &err);
1024 bh = sb_getblk(inode->i_sb, newblock);
1027 ext4_std_error(inode->i_sb, err);
1032 err = ext4_journal_get_create_access(handle, bh);
1038 /* move top-level index/leaf into new block */
1039 memmove(bh->b_data, curp->p_hdr, sizeof(EXT4_I(inode)->i_data));
1041 /* set size of new block */
1042 neh = ext_block_hdr(bh);
1043 /* old root could have indexes or leaves
1044 * so calculate e_max right way */
1045 if (ext_depth(inode))
1046 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
1048 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
1049 neh->eh_magic = EXT4_EXT_MAGIC;
1050 set_buffer_uptodate(bh);
1053 err = ext4_handle_dirty_metadata(handle, inode, bh);
1057 /* create index in new top-level index: num,max,pointer */
1058 err = ext4_ext_get_access(handle, inode, curp);
1062 curp->p_hdr->eh_magic = EXT4_EXT_MAGIC;
1063 curp->p_hdr->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode));
1064 curp->p_hdr->eh_entries = cpu_to_le16(1);
1065 curp->p_idx = EXT_FIRST_INDEX(curp->p_hdr);
1067 if (path[0].p_hdr->eh_depth)
1068 curp->p_idx->ei_block =
1069 EXT_FIRST_INDEX(path[0].p_hdr)->ei_block;
1071 curp->p_idx->ei_block =
1072 EXT_FIRST_EXTENT(path[0].p_hdr)->ee_block;
1073 ext4_idx_store_pblock(curp->p_idx, newblock);
1075 neh = ext_inode_hdr(inode);
1076 fidx = EXT_FIRST_INDEX(neh);
1077 ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
1078 le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
1079 le32_to_cpu(fidx->ei_block), idx_pblock(fidx));
1081 neh->eh_depth = cpu_to_le16(path->p_depth + 1);
1082 err = ext4_ext_dirty(handle, inode, curp);
1090 * ext4_ext_create_new_leaf:
1091 * finds empty index and adds new leaf.
1092 * if no free index is found, then it requests in-depth growing.
1094 static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
1095 struct ext4_ext_path *path,
1096 struct ext4_extent *newext)
1098 struct ext4_ext_path *curp;
1099 int depth, i, err = 0;
1102 i = depth = ext_depth(inode);
1104 /* walk up to the tree and look for free index entry */
1105 curp = path + depth;
1106 while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
1111 /* we use already allocated block for index block,
1112 * so subsequent data blocks should be contiguous */
1113 if (EXT_HAS_FREE_INDEX(curp)) {
1114 /* if we found index with free entry, then use that
1115 * entry: create all needed subtree and add new leaf */
1116 err = ext4_ext_split(handle, inode, path, newext, i);
1121 ext4_ext_drop_refs(path);
1122 path = ext4_ext_find_extent(inode,
1123 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1126 err = PTR_ERR(path);
1128 /* tree is full, time to grow in depth */
1129 err = ext4_ext_grow_indepth(handle, inode, path, newext);
1134 ext4_ext_drop_refs(path);
1135 path = ext4_ext_find_extent(inode,
1136 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1139 err = PTR_ERR(path);
1144 * only first (depth 0 -> 1) produces free space;
1145 * in all other cases we have to split the grown tree
1147 depth = ext_depth(inode);
1148 if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
1149 /* now we need to split */
1159 * search the closest allocated block to the left for *logical
1160 * and returns it at @logical + it's physical address at @phys
1161 * if *logical is the smallest allocated block, the function
1162 * returns 0 at @phys
1163 * return value contains 0 (success) or error code
1166 ext4_ext_search_left(struct inode *inode, struct ext4_ext_path *path,
1167 ext4_lblk_t *logical, ext4_fsblk_t *phys)
1169 struct ext4_extent_idx *ix;
1170 struct ext4_extent *ex;
1173 BUG_ON(path == NULL);
1174 depth = path->p_depth;
1177 if (depth == 0 && path->p_ext == NULL)
1180 /* usually extent in the path covers blocks smaller
1181 * then *logical, but it can be that extent is the
1182 * first one in the file */
1184 ex = path[depth].p_ext;
1185 ee_len = ext4_ext_get_actual_len(ex);
1186 if (*logical < le32_to_cpu(ex->ee_block)) {
1187 BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex);
1188 while (--depth >= 0) {
1189 ix = path[depth].p_idx;
1190 BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr));
1195 BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len));
1197 *logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
1198 *phys = ext_pblock(ex) + ee_len - 1;
1203 * search the closest allocated block to the right for *logical
1204 * and returns it at @logical + it's physical address at @phys
1205 * if *logical is the smallest allocated block, the function
1206 * returns 0 at @phys
1207 * return value contains 0 (success) or error code
1210 ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path,
1211 ext4_lblk_t *logical, ext4_fsblk_t *phys)
1213 struct buffer_head *bh = NULL;
1214 struct ext4_extent_header *eh;
1215 struct ext4_extent_idx *ix;
1216 struct ext4_extent *ex;
1218 int depth; /* Note, NOT eh_depth; depth from top of tree */
1221 BUG_ON(path == NULL);
1222 depth = path->p_depth;
1225 if (depth == 0 && path->p_ext == NULL)
1228 /* usually extent in the path covers blocks smaller
1229 * then *logical, but it can be that extent is the
1230 * first one in the file */
1232 ex = path[depth].p_ext;
1233 ee_len = ext4_ext_get_actual_len(ex);
1234 if (*logical < le32_to_cpu(ex->ee_block)) {
1235 BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex);
1236 while (--depth >= 0) {
1237 ix = path[depth].p_idx;
1238 BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr));
1240 *logical = le32_to_cpu(ex->ee_block);
1241 *phys = ext_pblock(ex);
1245 BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len));
1247 if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
1248 /* next allocated block in this leaf */
1250 *logical = le32_to_cpu(ex->ee_block);
1251 *phys = ext_pblock(ex);
1255 /* go up and search for index to the right */
1256 while (--depth >= 0) {
1257 ix = path[depth].p_idx;
1258 if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
1262 /* we've gone up to the root and found no index to the right */
1266 /* we've found index to the right, let's
1267 * follow it and find the closest allocated
1268 * block to the right */
1270 block = idx_pblock(ix);
1271 while (++depth < path->p_depth) {
1272 bh = sb_bread(inode->i_sb, block);
1275 eh = ext_block_hdr(bh);
1276 /* subtract from p_depth to get proper eh_depth */
1277 if (ext4_ext_check(inode, eh, path->p_depth - depth)) {
1281 ix = EXT_FIRST_INDEX(eh);
1282 block = idx_pblock(ix);
1286 bh = sb_bread(inode->i_sb, block);
1289 eh = ext_block_hdr(bh);
1290 if (ext4_ext_check(inode, eh, path->p_depth - depth)) {
1294 ex = EXT_FIRST_EXTENT(eh);
1295 *logical = le32_to_cpu(ex->ee_block);
1296 *phys = ext_pblock(ex);
1302 * ext4_ext_next_allocated_block:
1303 * returns allocated block in subsequent extent or EXT_MAX_BLOCK.
1304 * NOTE: it considers block number from index entry as
1305 * allocated block. Thus, index entries have to be consistent
1309 ext4_ext_next_allocated_block(struct ext4_ext_path *path)
1313 BUG_ON(path == NULL);
1314 depth = path->p_depth;
1316 if (depth == 0 && path->p_ext == NULL)
1317 return EXT_MAX_BLOCK;
1319 while (depth >= 0) {
1320 if (depth == path->p_depth) {
1322 if (path[depth].p_ext !=
1323 EXT_LAST_EXTENT(path[depth].p_hdr))
1324 return le32_to_cpu(path[depth].p_ext[1].ee_block);
1327 if (path[depth].p_idx !=
1328 EXT_LAST_INDEX(path[depth].p_hdr))
1329 return le32_to_cpu(path[depth].p_idx[1].ei_block);
1334 return EXT_MAX_BLOCK;
1338 * ext4_ext_next_leaf_block:
1339 * returns first allocated block from next leaf or EXT_MAX_BLOCK
1341 static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode,
1342 struct ext4_ext_path *path)
1346 BUG_ON(path == NULL);
1347 depth = path->p_depth;
1349 /* zero-tree has no leaf blocks at all */
1351 return EXT_MAX_BLOCK;
1353 /* go to index block */
1356 while (depth >= 0) {
1357 if (path[depth].p_idx !=
1358 EXT_LAST_INDEX(path[depth].p_hdr))
1359 return (ext4_lblk_t)
1360 le32_to_cpu(path[depth].p_idx[1].ei_block);
1364 return EXT_MAX_BLOCK;
1368 * ext4_ext_correct_indexes:
1369 * if leaf gets modified and modified extent is first in the leaf,
1370 * then we have to correct all indexes above.
1371 * TODO: do we need to correct tree in all cases?
1373 static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
1374 struct ext4_ext_path *path)
1376 struct ext4_extent_header *eh;
1377 int depth = ext_depth(inode);
1378 struct ext4_extent *ex;
1382 eh = path[depth].p_hdr;
1383 ex = path[depth].p_ext;
1388 /* there is no tree at all */
1392 if (ex != EXT_FIRST_EXTENT(eh)) {
1393 /* we correct tree if first leaf got modified only */
1398 * TODO: we need correction if border is smaller than current one
1401 border = path[depth].p_ext->ee_block;
1402 err = ext4_ext_get_access(handle, inode, path + k);
1405 path[k].p_idx->ei_block = border;
1406 err = ext4_ext_dirty(handle, inode, path + k);
1411 /* change all left-side indexes */
1412 if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
1414 err = ext4_ext_get_access(handle, inode, path + k);
1417 path[k].p_idx->ei_block = border;
1418 err = ext4_ext_dirty(handle, inode, path + k);
1427 ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
1428 struct ext4_extent *ex2)
1430 unsigned short ext1_ee_len, ext2_ee_len, max_len;
1433 * Make sure that either both extents are uninitialized, or
1436 if (ext4_ext_is_uninitialized(ex1) ^ ext4_ext_is_uninitialized(ex2))
1439 if (ext4_ext_is_uninitialized(ex1))
1440 max_len = EXT_UNINIT_MAX_LEN;
1442 max_len = EXT_INIT_MAX_LEN;
1444 ext1_ee_len = ext4_ext_get_actual_len(ex1);
1445 ext2_ee_len = ext4_ext_get_actual_len(ex2);
1447 if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
1448 le32_to_cpu(ex2->ee_block))
1452 * To allow future support for preallocated extents to be added
1453 * as an RO_COMPAT feature, refuse to merge to extents if
1454 * this can result in the top bit of ee_len being set.
1456 if (ext1_ee_len + ext2_ee_len > max_len)
1458 #ifdef AGGRESSIVE_TEST
1459 if (ext1_ee_len >= 4)
1463 if (ext_pblock(ex1) + ext1_ee_len == ext_pblock(ex2))
1469 * This function tries to merge the "ex" extent to the next extent in the tree.
1470 * It always tries to merge towards right. If you want to merge towards
1471 * left, pass "ex - 1" as argument instead of "ex".
1472 * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
1473 * 1 if they got merged.
1475 int ext4_ext_try_to_merge(struct inode *inode,
1476 struct ext4_ext_path *path,
1477 struct ext4_extent *ex)
1479 struct ext4_extent_header *eh;
1480 unsigned int depth, len;
1482 int uninitialized = 0;
1484 depth = ext_depth(inode);
1485 BUG_ON(path[depth].p_hdr == NULL);
1486 eh = path[depth].p_hdr;
1488 while (ex < EXT_LAST_EXTENT(eh)) {
1489 if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
1491 /* merge with next extent! */
1492 if (ext4_ext_is_uninitialized(ex))
1494 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1495 + ext4_ext_get_actual_len(ex + 1));
1497 ext4_ext_mark_uninitialized(ex);
1499 if (ex + 1 < EXT_LAST_EXTENT(eh)) {
1500 len = (EXT_LAST_EXTENT(eh) - ex - 1)
1501 * sizeof(struct ext4_extent);
1502 memmove(ex + 1, ex + 2, len);
1504 le16_add_cpu(&eh->eh_entries, -1);
1506 WARN_ON(eh->eh_entries == 0);
1507 if (!eh->eh_entries)
1508 ext4_error(inode->i_sb, "ext4_ext_try_to_merge",
1509 "inode#%lu, eh->eh_entries = 0!", inode->i_ino);
1516 * check if a portion of the "newext" extent overlaps with an
1519 * If there is an overlap discovered, it updates the length of the newext
1520 * such that there will be no overlap, and then returns 1.
1521 * If there is no overlap found, it returns 0.
1523 unsigned int ext4_ext_check_overlap(struct inode *inode,
1524 struct ext4_extent *newext,
1525 struct ext4_ext_path *path)
1528 unsigned int depth, len1;
1529 unsigned int ret = 0;
1531 b1 = le32_to_cpu(newext->ee_block);
1532 len1 = ext4_ext_get_actual_len(newext);
1533 depth = ext_depth(inode);
1534 if (!path[depth].p_ext)
1536 b2 = le32_to_cpu(path[depth].p_ext->ee_block);
1539 * get the next allocated block if the extent in the path
1540 * is before the requested block(s)
1543 b2 = ext4_ext_next_allocated_block(path);
1544 if (b2 == EXT_MAX_BLOCK)
1548 /* check for wrap through zero on extent logical start block*/
1549 if (b1 + len1 < b1) {
1550 len1 = EXT_MAX_BLOCK - b1;
1551 newext->ee_len = cpu_to_le16(len1);
1555 /* check for overlap */
1556 if (b1 + len1 > b2) {
1557 newext->ee_len = cpu_to_le16(b2 - b1);
1565 * ext4_ext_insert_extent:
1566 * tries to merge requsted extent into the existing extent or
1567 * inserts requested extent as new one into the tree,
1568 * creating new leaf in the no-space case.
1570 int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
1571 struct ext4_ext_path *path,
1572 struct ext4_extent *newext)
1574 struct ext4_extent_header *eh;
1575 struct ext4_extent *ex, *fex;
1576 struct ext4_extent *nearex; /* nearest extent */
1577 struct ext4_ext_path *npath = NULL;
1578 int depth, len, err;
1580 unsigned uninitialized = 0;
1582 BUG_ON(ext4_ext_get_actual_len(newext) == 0);
1583 depth = ext_depth(inode);
1584 ex = path[depth].p_ext;
1585 BUG_ON(path[depth].p_hdr == NULL);
1587 /* try to insert block into found extent and return */
1588 if (ex && ext4_can_extents_be_merged(inode, ex, newext)) {
1589 ext_debug("append [%d]%d block to %d:[%d]%d (from %llu)\n",
1590 ext4_ext_is_uninitialized(newext),
1591 ext4_ext_get_actual_len(newext),
1592 le32_to_cpu(ex->ee_block),
1593 ext4_ext_is_uninitialized(ex),
1594 ext4_ext_get_actual_len(ex), ext_pblock(ex));
1595 err = ext4_ext_get_access(handle, inode, path + depth);
1600 * ext4_can_extents_be_merged should have checked that either
1601 * both extents are uninitialized, or both aren't. Thus we
1602 * need to check only one of them here.
1604 if (ext4_ext_is_uninitialized(ex))
1606 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1607 + ext4_ext_get_actual_len(newext));
1609 ext4_ext_mark_uninitialized(ex);
1610 eh = path[depth].p_hdr;
1616 depth = ext_depth(inode);
1617 eh = path[depth].p_hdr;
1618 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
1621 /* probably next leaf has space for us? */
1622 fex = EXT_LAST_EXTENT(eh);
1623 next = ext4_ext_next_leaf_block(inode, path);
1624 if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
1625 && next != EXT_MAX_BLOCK) {
1626 ext_debug("next leaf block - %d\n", next);
1627 BUG_ON(npath != NULL);
1628 npath = ext4_ext_find_extent(inode, next, NULL);
1630 return PTR_ERR(npath);
1631 BUG_ON(npath->p_depth != path->p_depth);
1632 eh = npath[depth].p_hdr;
1633 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
1634 ext_debug("next leaf isnt full(%d)\n",
1635 le16_to_cpu(eh->eh_entries));
1639 ext_debug("next leaf has no free space(%d,%d)\n",
1640 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
1644 * There is no free space in the found leaf.
1645 * We're gonna add a new leaf in the tree.
1647 err = ext4_ext_create_new_leaf(handle, inode, path, newext);
1650 depth = ext_depth(inode);
1651 eh = path[depth].p_hdr;
1654 nearex = path[depth].p_ext;
1656 err = ext4_ext_get_access(handle, inode, path + depth);
1661 /* there is no extent in this leaf, create first one */
1662 ext_debug("first extent in the leaf: %d:%llu:[%d]%d\n",
1663 le32_to_cpu(newext->ee_block),
1665 ext4_ext_is_uninitialized(newext),
1666 ext4_ext_get_actual_len(newext));
1667 path[depth].p_ext = EXT_FIRST_EXTENT(eh);
1668 } else if (le32_to_cpu(newext->ee_block)
1669 > le32_to_cpu(nearex->ee_block)) {
1670 /* BUG_ON(newext->ee_block == nearex->ee_block); */
1671 if (nearex != EXT_LAST_EXTENT(eh)) {
1672 len = EXT_MAX_EXTENT(eh) - nearex;
1673 len = (len - 1) * sizeof(struct ext4_extent);
1674 len = len < 0 ? 0 : len;
1675 ext_debug("insert %d:%llu:[%d]%d after: nearest 0x%p, "
1676 "move %d from 0x%p to 0x%p\n",
1677 le32_to_cpu(newext->ee_block),
1679 ext4_ext_is_uninitialized(newext),
1680 ext4_ext_get_actual_len(newext),
1681 nearex, len, nearex + 1, nearex + 2);
1682 memmove(nearex + 2, nearex + 1, len);
1684 path[depth].p_ext = nearex + 1;
1686 BUG_ON(newext->ee_block == nearex->ee_block);
1687 len = (EXT_MAX_EXTENT(eh) - nearex) * sizeof(struct ext4_extent);
1688 len = len < 0 ? 0 : len;
1689 ext_debug("insert %d:%llu:[%d]%d before: nearest 0x%p, "
1690 "move %d from 0x%p to 0x%p\n",
1691 le32_to_cpu(newext->ee_block),
1693 ext4_ext_is_uninitialized(newext),
1694 ext4_ext_get_actual_len(newext),
1695 nearex, len, nearex + 1, nearex + 2);
1696 memmove(nearex + 1, nearex, len);
1697 path[depth].p_ext = nearex;
1700 le16_add_cpu(&eh->eh_entries, 1);
1701 nearex = path[depth].p_ext;
1702 nearex->ee_block = newext->ee_block;
1703 ext4_ext_store_pblock(nearex, ext_pblock(newext));
1704 nearex->ee_len = newext->ee_len;
1707 /* try to merge extents to the right */
1708 ext4_ext_try_to_merge(inode, path, nearex);
1710 /* try to merge extents to the left */
1712 /* time to correct all indexes above */
1713 err = ext4_ext_correct_indexes(handle, inode, path);
1717 err = ext4_ext_dirty(handle, inode, path + depth);
1721 ext4_ext_drop_refs(npath);
1724 ext4_ext_invalidate_cache(inode);
1728 int ext4_ext_walk_space(struct inode *inode, ext4_lblk_t block,
1729 ext4_lblk_t num, ext_prepare_callback func,
1732 struct ext4_ext_path *path = NULL;
1733 struct ext4_ext_cache cbex;
1734 struct ext4_extent *ex;
1735 ext4_lblk_t next, start = 0, end = 0;
1736 ext4_lblk_t last = block + num;
1737 int depth, exists, err = 0;
1739 BUG_ON(func == NULL);
1740 BUG_ON(inode == NULL);
1742 while (block < last && block != EXT_MAX_BLOCK) {
1744 /* find extent for this block */
1745 path = ext4_ext_find_extent(inode, block, path);
1747 err = PTR_ERR(path);
1752 depth = ext_depth(inode);
1753 BUG_ON(path[depth].p_hdr == NULL);
1754 ex = path[depth].p_ext;
1755 next = ext4_ext_next_allocated_block(path);
1759 /* there is no extent yet, so try to allocate
1760 * all requested space */
1763 } else if (le32_to_cpu(ex->ee_block) > block) {
1764 /* need to allocate space before found extent */
1766 end = le32_to_cpu(ex->ee_block);
1767 if (block + num < end)
1769 } else if (block >= le32_to_cpu(ex->ee_block)
1770 + ext4_ext_get_actual_len(ex)) {
1771 /* need to allocate space after found extent */
1776 } else if (block >= le32_to_cpu(ex->ee_block)) {
1778 * some part of requested space is covered
1782 end = le32_to_cpu(ex->ee_block)
1783 + ext4_ext_get_actual_len(ex);
1784 if (block + num < end)
1790 BUG_ON(end <= start);
1793 cbex.ec_block = start;
1794 cbex.ec_len = end - start;
1796 cbex.ec_type = EXT4_EXT_CACHE_GAP;
1798 cbex.ec_block = le32_to_cpu(ex->ee_block);
1799 cbex.ec_len = ext4_ext_get_actual_len(ex);
1800 cbex.ec_start = ext_pblock(ex);
1801 cbex.ec_type = EXT4_EXT_CACHE_EXTENT;
1804 BUG_ON(cbex.ec_len == 0);
1805 err = func(inode, path, &cbex, ex, cbdata);
1806 ext4_ext_drop_refs(path);
1811 if (err == EXT_REPEAT)
1813 else if (err == EXT_BREAK) {
1818 if (ext_depth(inode) != depth) {
1819 /* depth was changed. we have to realloc path */
1824 block = cbex.ec_block + cbex.ec_len;
1828 ext4_ext_drop_refs(path);
1836 ext4_ext_put_in_cache(struct inode *inode, ext4_lblk_t block,
1837 __u32 len, ext4_fsblk_t start, int type)
1839 struct ext4_ext_cache *cex;
1841 spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
1842 cex = &EXT4_I(inode)->i_cached_extent;
1843 cex->ec_type = type;
1844 cex->ec_block = block;
1846 cex->ec_start = start;
1847 spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
1851 * ext4_ext_put_gap_in_cache:
1852 * calculate boundaries of the gap that the requested block fits into
1853 * and cache this gap
1856 ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
1859 int depth = ext_depth(inode);
1862 struct ext4_extent *ex;
1864 ex = path[depth].p_ext;
1866 /* there is no extent yet, so gap is [0;-] */
1868 len = EXT_MAX_BLOCK;
1869 ext_debug("cache gap(whole file):");
1870 } else if (block < le32_to_cpu(ex->ee_block)) {
1872 len = le32_to_cpu(ex->ee_block) - block;
1873 ext_debug("cache gap(before): %u [%u:%u]",
1875 le32_to_cpu(ex->ee_block),
1876 ext4_ext_get_actual_len(ex));
1877 } else if (block >= le32_to_cpu(ex->ee_block)
1878 + ext4_ext_get_actual_len(ex)) {
1880 lblock = le32_to_cpu(ex->ee_block)
1881 + ext4_ext_get_actual_len(ex);
1883 next = ext4_ext_next_allocated_block(path);
1884 ext_debug("cache gap(after): [%u:%u] %u",
1885 le32_to_cpu(ex->ee_block),
1886 ext4_ext_get_actual_len(ex),
1888 BUG_ON(next == lblock);
1889 len = next - lblock;
1895 ext_debug(" -> %u:%lu\n", lblock, len);
1896 ext4_ext_put_in_cache(inode, lblock, len, 0, EXT4_EXT_CACHE_GAP);
1900 ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block,
1901 struct ext4_extent *ex)
1903 struct ext4_ext_cache *cex;
1904 int ret = EXT4_EXT_CACHE_NO;
1907 * We borrow i_block_reservation_lock to protect i_cached_extent
1909 spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
1910 cex = &EXT4_I(inode)->i_cached_extent;
1912 /* has cache valid data? */
1913 if (cex->ec_type == EXT4_EXT_CACHE_NO)
1916 BUG_ON(cex->ec_type != EXT4_EXT_CACHE_GAP &&
1917 cex->ec_type != EXT4_EXT_CACHE_EXTENT);
1918 if (block >= cex->ec_block && block < cex->ec_block + cex->ec_len) {
1919 ex->ee_block = cpu_to_le32(cex->ec_block);
1920 ext4_ext_store_pblock(ex, cex->ec_start);
1921 ex->ee_len = cpu_to_le16(cex->ec_len);
1922 ext_debug("%u cached by %u:%u:%llu\n",
1924 cex->ec_block, cex->ec_len, cex->ec_start);
1928 spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
1934 * removes index from the index block.
1935 * It's used in truncate case only, thus all requests are for
1936 * last index in the block only.
1938 static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
1939 struct ext4_ext_path *path)
1941 struct buffer_head *bh;
1945 /* free index block */
1947 leaf = idx_pblock(path->p_idx);
1948 BUG_ON(path->p_hdr->eh_entries == 0);
1949 err = ext4_ext_get_access(handle, inode, path);
1952 le16_add_cpu(&path->p_hdr->eh_entries, -1);
1953 err = ext4_ext_dirty(handle, inode, path);
1956 ext_debug("index is empty, remove it, free block %llu\n", leaf);
1957 bh = sb_find_get_block(inode->i_sb, leaf);
1958 ext4_forget(handle, 1, inode, bh, leaf);
1959 ext4_free_blocks(handle, inode, leaf, 1, 1);
1964 * ext4_ext_calc_credits_for_single_extent:
1965 * This routine returns max. credits that needed to insert an extent
1966 * to the extent tree.
1967 * When pass the actual path, the caller should calculate credits
1970 int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
1971 struct ext4_ext_path *path)
1974 int depth = ext_depth(inode);
1977 /* probably there is space in leaf? */
1978 if (le16_to_cpu(path[depth].p_hdr->eh_entries)
1979 < le16_to_cpu(path[depth].p_hdr->eh_max)) {
1982 * There are some space in the leaf tree, no
1983 * need to account for leaf block credit
1985 * bitmaps and block group descriptor blocks
1986 * and other metadat blocks still need to be
1989 /* 1 bitmap, 1 block group descriptor */
1990 ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
1995 return ext4_chunk_trans_blocks(inode, nrblocks);
1999 * How many index/leaf blocks need to change/allocate to modify nrblocks?
2001 * if nrblocks are fit in a single extent (chunk flag is 1), then
2002 * in the worse case, each tree level index/leaf need to be changed
2003 * if the tree split due to insert a new extent, then the old tree
2004 * index/leaf need to be updated too
2006 * If the nrblocks are discontiguous, they could cause
2007 * the whole tree split more than once, but this is really rare.
2009 int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
2012 int depth = ext_depth(inode);
2022 static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
2023 struct ext4_extent *ex,
2024 ext4_lblk_t from, ext4_lblk_t to)
2026 struct buffer_head *bh;
2027 unsigned short ee_len = ext4_ext_get_actual_len(ex);
2028 int i, metadata = 0;
2030 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2032 #ifdef EXTENTS_STATS
2034 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2035 spin_lock(&sbi->s_ext_stats_lock);
2036 sbi->s_ext_blocks += ee_len;
2037 sbi->s_ext_extents++;
2038 if (ee_len < sbi->s_ext_min)
2039 sbi->s_ext_min = ee_len;
2040 if (ee_len > sbi->s_ext_max)
2041 sbi->s_ext_max = ee_len;
2042 if (ext_depth(inode) > sbi->s_depth_max)
2043 sbi->s_depth_max = ext_depth(inode);
2044 spin_unlock(&sbi->s_ext_stats_lock);
2047 if (from >= le32_to_cpu(ex->ee_block)
2048 && to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
2053 num = le32_to_cpu(ex->ee_block) + ee_len - from;
2054 start = ext_pblock(ex) + ee_len - num;
2055 ext_debug("free last %u blocks starting %llu\n", num, start);
2056 for (i = 0; i < num; i++) {
2057 bh = sb_find_get_block(inode->i_sb, start + i);
2058 ext4_forget(handle, 0, inode, bh, start + i);
2060 ext4_free_blocks(handle, inode, start, num, metadata);
2061 } else if (from == le32_to_cpu(ex->ee_block)
2062 && to <= le32_to_cpu(ex->ee_block) + ee_len - 1) {
2063 printk(KERN_INFO "strange request: removal %u-%u from %u:%u\n",
2064 from, to, le32_to_cpu(ex->ee_block), ee_len);
2066 printk(KERN_INFO "strange request: removal(2) "
2067 "%u-%u from %u:%u\n",
2068 from, to, le32_to_cpu(ex->ee_block), ee_len);
2074 ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
2075 struct ext4_ext_path *path, ext4_lblk_t start)
2077 int err = 0, correct_index = 0;
2078 int depth = ext_depth(inode), credits;
2079 struct ext4_extent_header *eh;
2080 ext4_lblk_t a, b, block;
2082 ext4_lblk_t ex_ee_block;
2083 unsigned short ex_ee_len;
2084 unsigned uninitialized = 0;
2085 struct ext4_extent *ex;
2087 /* the header must be checked already in ext4_ext_remove_space() */
2088 ext_debug("truncate since %u in leaf\n", start);
2089 if (!path[depth].p_hdr)
2090 path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
2091 eh = path[depth].p_hdr;
2094 /* find where to start removing */
2095 ex = EXT_LAST_EXTENT(eh);
2097 ex_ee_block = le32_to_cpu(ex->ee_block);
2098 ex_ee_len = ext4_ext_get_actual_len(ex);
2100 while (ex >= EXT_FIRST_EXTENT(eh) &&
2101 ex_ee_block + ex_ee_len > start) {
2103 if (ext4_ext_is_uninitialized(ex))
2108 ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
2109 uninitialized, ex_ee_len);
2110 path[depth].p_ext = ex;
2112 a = ex_ee_block > start ? ex_ee_block : start;
2113 b = ex_ee_block + ex_ee_len - 1 < EXT_MAX_BLOCK ?
2114 ex_ee_block + ex_ee_len - 1 : EXT_MAX_BLOCK;
2116 ext_debug(" border %u:%u\n", a, b);
2118 if (a != ex_ee_block && b != ex_ee_block + ex_ee_len - 1) {
2122 } else if (a != ex_ee_block) {
2123 /* remove tail of the extent */
2124 block = ex_ee_block;
2126 } else if (b != ex_ee_block + ex_ee_len - 1) {
2127 /* remove head of the extent */
2130 /* there is no "make a hole" API yet */
2133 /* remove whole extent: excellent! */
2134 block = ex_ee_block;
2136 BUG_ON(a != ex_ee_block);
2137 BUG_ON(b != ex_ee_block + ex_ee_len - 1);
2141 * 3 for leaf, sb, and inode plus 2 (bmap and group
2142 * descriptor) for each block group; assume two block
2143 * groups plus ex_ee_len/blocks_per_block_group for
2146 credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
2147 if (ex == EXT_FIRST_EXTENT(eh)) {
2149 credits += (ext_depth(inode)) + 1;
2151 credits += 2 * EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb);
2153 err = ext4_ext_journal_restart(handle, credits);
2157 err = ext4_ext_get_access(handle, inode, path + depth);
2161 err = ext4_remove_blocks(handle, inode, ex, a, b);
2166 /* this extent is removed; mark slot entirely unused */
2167 ext4_ext_store_pblock(ex, 0);
2168 le16_add_cpu(&eh->eh_entries, -1);
2171 ex->ee_block = cpu_to_le32(block);
2172 ex->ee_len = cpu_to_le16(num);
2174 * Do not mark uninitialized if all the blocks in the
2175 * extent have been removed.
2177 if (uninitialized && num)
2178 ext4_ext_mark_uninitialized(ex);
2180 err = ext4_ext_dirty(handle, inode, path + depth);
2184 ext_debug("new extent: %u:%u:%llu\n", block, num,
2187 ex_ee_block = le32_to_cpu(ex->ee_block);
2188 ex_ee_len = ext4_ext_get_actual_len(ex);
2191 if (correct_index && eh->eh_entries)
2192 err = ext4_ext_correct_indexes(handle, inode, path);
2194 /* if this leaf is free, then we should
2195 * remove it from index block above */
2196 if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
2197 err = ext4_ext_rm_idx(handle, inode, path + depth);
2204 * ext4_ext_more_to_rm:
2205 * returns 1 if current index has to be freed (even partial)
2208 ext4_ext_more_to_rm(struct ext4_ext_path *path)
2210 BUG_ON(path->p_idx == NULL);
2212 if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
2216 * if truncate on deeper level happened, it wasn't partial,
2217 * so we have to consider current index for truncation
2219 if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
2224 static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
2226 struct super_block *sb = inode->i_sb;
2227 int depth = ext_depth(inode);
2228 struct ext4_ext_path *path;
2232 ext_debug("truncate since %u\n", start);
2234 /* probably first extent we're gonna free will be last in block */
2235 handle = ext4_journal_start(inode, depth + 1);
2237 return PTR_ERR(handle);
2239 ext4_ext_invalidate_cache(inode);
2242 * We start scanning from right side, freeing all the blocks
2243 * after i_size and walking into the tree depth-wise.
2245 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1), GFP_NOFS);
2247 ext4_journal_stop(handle);
2250 path[0].p_hdr = ext_inode_hdr(inode);
2251 if (ext4_ext_check(inode, path[0].p_hdr, depth)) {
2255 path[0].p_depth = depth;
2257 while (i >= 0 && err == 0) {
2259 /* this is leaf block */
2260 err = ext4_ext_rm_leaf(handle, inode, path, start);
2261 /* root level has p_bh == NULL, brelse() eats this */
2262 brelse(path[i].p_bh);
2263 path[i].p_bh = NULL;
2268 /* this is index block */
2269 if (!path[i].p_hdr) {
2270 ext_debug("initialize header\n");
2271 path[i].p_hdr = ext_block_hdr(path[i].p_bh);
2274 if (!path[i].p_idx) {
2275 /* this level hasn't been touched yet */
2276 path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
2277 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
2278 ext_debug("init index ptr: hdr 0x%p, num %d\n",
2280 le16_to_cpu(path[i].p_hdr->eh_entries));
2282 /* we were already here, see at next index */
2286 ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
2287 i, EXT_FIRST_INDEX(path[i].p_hdr),
2289 if (ext4_ext_more_to_rm(path + i)) {
2290 struct buffer_head *bh;
2291 /* go to the next level */
2292 ext_debug("move to level %d (block %llu)\n",
2293 i + 1, idx_pblock(path[i].p_idx));
2294 memset(path + i + 1, 0, sizeof(*path));
2295 bh = sb_bread(sb, idx_pblock(path[i].p_idx));
2297 /* should we reset i_size? */
2301 if (WARN_ON(i + 1 > depth)) {
2305 if (ext4_ext_check(inode, ext_block_hdr(bh),
2310 path[i + 1].p_bh = bh;
2312 /* save actual number of indexes since this
2313 * number is changed at the next iteration */
2314 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
2317 /* we finished processing this index, go up */
2318 if (path[i].p_hdr->eh_entries == 0 && i > 0) {
2319 /* index is empty, remove it;
2320 * handle must be already prepared by the
2321 * truncatei_leaf() */
2322 err = ext4_ext_rm_idx(handle, inode, path + i);
2324 /* root level has p_bh == NULL, brelse() eats this */
2325 brelse(path[i].p_bh);
2326 path[i].p_bh = NULL;
2328 ext_debug("return to level %d\n", i);
2332 /* TODO: flexible tree reduction should be here */
2333 if (path->p_hdr->eh_entries == 0) {
2335 * truncate to zero freed all the tree,
2336 * so we need to correct eh_depth
2338 err = ext4_ext_get_access(handle, inode, path);
2340 ext_inode_hdr(inode)->eh_depth = 0;
2341 ext_inode_hdr(inode)->eh_max =
2342 cpu_to_le16(ext4_ext_space_root(inode));
2343 err = ext4_ext_dirty(handle, inode, path);
2347 ext4_ext_drop_refs(path);
2349 ext4_journal_stop(handle);
2355 * called at mount time
2357 void ext4_ext_init(struct super_block *sb)
2360 * possible initialization would be here
2363 if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
2364 printk(KERN_INFO "EXT4-fs: file extents enabled");
2365 #ifdef AGGRESSIVE_TEST
2366 printk(", aggressive tests");
2368 #ifdef CHECK_BINSEARCH
2369 printk(", check binsearch");
2371 #ifdef EXTENTS_STATS
2375 #ifdef EXTENTS_STATS
2376 spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
2377 EXT4_SB(sb)->s_ext_min = 1 << 30;
2378 EXT4_SB(sb)->s_ext_max = 0;
2384 * called at umount time
2386 void ext4_ext_release(struct super_block *sb)
2388 if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
2391 #ifdef EXTENTS_STATS
2392 if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
2393 struct ext4_sb_info *sbi = EXT4_SB(sb);
2394 printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
2395 sbi->s_ext_blocks, sbi->s_ext_extents,
2396 sbi->s_ext_blocks / sbi->s_ext_extents);
2397 printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
2398 sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
2403 static void bi_complete(struct bio *bio, int error)
2405 complete((struct completion *)bio->bi_private);
2408 /* FIXME!! we need to try to merge to left or right after zero-out */
2409 static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
2413 int blkbits, blocksize;
2415 struct completion event;
2416 unsigned int ee_len, len, done, offset;
2419 blkbits = inode->i_blkbits;
2420 blocksize = inode->i_sb->s_blocksize;
2421 ee_len = ext4_ext_get_actual_len(ex);
2422 ee_pblock = ext_pblock(ex);
2424 /* convert ee_pblock to 512 byte sectors */
2425 ee_pblock = ee_pblock << (blkbits - 9);
2427 while (ee_len > 0) {
2429 if (ee_len > BIO_MAX_PAGES)
2430 len = BIO_MAX_PAGES;
2434 bio = bio_alloc(GFP_NOIO, len);
2435 bio->bi_sector = ee_pblock;
2436 bio->bi_bdev = inode->i_sb->s_bdev;
2440 while (done < len) {
2441 ret = bio_add_page(bio, ZERO_PAGE(0),
2443 if (ret != blocksize) {
2445 * We can't add any more pages because of
2446 * hardware limitations. Start a new bio.
2451 offset += blocksize;
2452 if (offset >= PAGE_CACHE_SIZE)
2456 init_completion(&event);
2457 bio->bi_private = &event;
2458 bio->bi_end_io = bi_complete;
2459 submit_bio(WRITE, bio);
2460 wait_for_completion(&event);
2462 if (test_bit(BIO_UPTODATE, &bio->bi_flags))
2470 ee_pblock += done << (blkbits - 9);
2475 #define EXT4_EXT_ZERO_LEN 7
2478 * This function is called by ext4_ext_get_blocks() if someone tries to write
2479 * to an uninitialized extent. It may result in splitting the uninitialized
2480 * extent into multiple extents (upto three - one initialized and two
2482 * There are three possibilities:
2483 * a> There is no split required: Entire extent should be initialized
2484 * b> Splits in two extents: Write is happening at either end of the extent
2485 * c> Splits in three extents: Somone is writing in middle of the extent
2487 static int ext4_ext_convert_to_initialized(handle_t *handle,
2488 struct inode *inode,
2489 struct ext4_ext_path *path,
2491 unsigned int max_blocks)
2493 struct ext4_extent *ex, newex, orig_ex;
2494 struct ext4_extent *ex1 = NULL;
2495 struct ext4_extent *ex2 = NULL;
2496 struct ext4_extent *ex3 = NULL;
2497 struct ext4_extent_header *eh;
2498 ext4_lblk_t ee_block;
2499 unsigned int allocated, ee_len, depth;
2500 ext4_fsblk_t newblock;
2504 depth = ext_depth(inode);
2505 eh = path[depth].p_hdr;
2506 ex = path[depth].p_ext;
2507 ee_block = le32_to_cpu(ex->ee_block);
2508 ee_len = ext4_ext_get_actual_len(ex);
2509 allocated = ee_len - (iblock - ee_block);
2510 newblock = iblock - ee_block + ext_pblock(ex);
2512 orig_ex.ee_block = ex->ee_block;
2513 orig_ex.ee_len = cpu_to_le16(ee_len);
2514 ext4_ext_store_pblock(&orig_ex, ext_pblock(ex));
2516 err = ext4_ext_get_access(handle, inode, path + depth);
2519 /* If extent has less than 2*EXT4_EXT_ZERO_LEN zerout directly */
2520 if (ee_len <= 2*EXT4_EXT_ZERO_LEN) {
2521 err = ext4_ext_zeroout(inode, &orig_ex);
2523 goto fix_extent_len;
2524 /* update the extent length and mark as initialized */
2525 ex->ee_block = orig_ex.ee_block;
2526 ex->ee_len = orig_ex.ee_len;
2527 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2528 ext4_ext_dirty(handle, inode, path + depth);
2529 /* zeroed the full extent */
2533 /* ex1: ee_block to iblock - 1 : uninitialized */
2534 if (iblock > ee_block) {
2536 ex1->ee_len = cpu_to_le16(iblock - ee_block);
2537 ext4_ext_mark_uninitialized(ex1);
2541 * for sanity, update the length of the ex2 extent before
2542 * we insert ex3, if ex1 is NULL. This is to avoid temporary
2543 * overlap of blocks.
2545 if (!ex1 && allocated > max_blocks)
2546 ex2->ee_len = cpu_to_le16(max_blocks);
2547 /* ex3: to ee_block + ee_len : uninitialised */
2548 if (allocated > max_blocks) {
2549 unsigned int newdepth;
2550 /* If extent has less than EXT4_EXT_ZERO_LEN zerout directly */
2551 if (allocated <= EXT4_EXT_ZERO_LEN) {
2553 * iblock == ee_block is handled by the zerouout
2555 * Mark first half uninitialized.
2556 * Mark second half initialized and zero out the
2557 * initialized extent
2559 ex->ee_block = orig_ex.ee_block;
2560 ex->ee_len = cpu_to_le16(ee_len - allocated);
2561 ext4_ext_mark_uninitialized(ex);
2562 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2563 ext4_ext_dirty(handle, inode, path + depth);
2566 ex3->ee_block = cpu_to_le32(iblock);
2567 ext4_ext_store_pblock(ex3, newblock);
2568 ex3->ee_len = cpu_to_le16(allocated);
2569 err = ext4_ext_insert_extent(handle, inode, path, ex3);
2570 if (err == -ENOSPC) {
2571 err = ext4_ext_zeroout(inode, &orig_ex);
2573 goto fix_extent_len;
2574 ex->ee_block = orig_ex.ee_block;
2575 ex->ee_len = orig_ex.ee_len;
2576 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2577 ext4_ext_dirty(handle, inode, path + depth);
2578 /* blocks available from iblock */
2582 goto fix_extent_len;
2585 * We need to zero out the second half because
2586 * an fallocate request can update file size and
2587 * converting the second half to initialized extent
2588 * implies that we can leak some junk data to user
2591 err = ext4_ext_zeroout(inode, ex3);
2594 * We should actually mark the
2595 * second half as uninit and return error
2596 * Insert would have changed the extent
2598 depth = ext_depth(inode);
2599 ext4_ext_drop_refs(path);
2600 path = ext4_ext_find_extent(inode,
2603 err = PTR_ERR(path);
2606 /* get the second half extent details */
2607 ex = path[depth].p_ext;
2608 err = ext4_ext_get_access(handle, inode,
2612 ext4_ext_mark_uninitialized(ex);
2613 ext4_ext_dirty(handle, inode, path + depth);
2617 /* zeroed the second half */
2621 ex3->ee_block = cpu_to_le32(iblock + max_blocks);
2622 ext4_ext_store_pblock(ex3, newblock + max_blocks);
2623 ex3->ee_len = cpu_to_le16(allocated - max_blocks);
2624 ext4_ext_mark_uninitialized(ex3);
2625 err = ext4_ext_insert_extent(handle, inode, path, ex3);
2626 if (err == -ENOSPC) {
2627 err = ext4_ext_zeroout(inode, &orig_ex);
2629 goto fix_extent_len;
2630 /* update the extent length and mark as initialized */
2631 ex->ee_block = orig_ex.ee_block;
2632 ex->ee_len = orig_ex.ee_len;
2633 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2634 ext4_ext_dirty(handle, inode, path + depth);
2635 /* zeroed the full extent */
2636 /* blocks available from iblock */
2640 goto fix_extent_len;
2642 * The depth, and hence eh & ex might change
2643 * as part of the insert above.
2645 newdepth = ext_depth(inode);
2647 * update the extent length after successful insert of the
2650 orig_ex.ee_len = cpu_to_le16(ee_len -
2651 ext4_ext_get_actual_len(ex3));
2653 ext4_ext_drop_refs(path);
2654 path = ext4_ext_find_extent(inode, iblock, path);
2656 err = PTR_ERR(path);
2659 eh = path[depth].p_hdr;
2660 ex = path[depth].p_ext;
2664 err = ext4_ext_get_access(handle, inode, path + depth);
2668 allocated = max_blocks;
2670 /* If extent has less than EXT4_EXT_ZERO_LEN and we are trying
2671 * to insert a extent in the middle zerout directly
2672 * otherwise give the extent a chance to merge to left
2674 if (le16_to_cpu(orig_ex.ee_len) <= EXT4_EXT_ZERO_LEN &&
2675 iblock != ee_block) {
2676 err = ext4_ext_zeroout(inode, &orig_ex);
2678 goto fix_extent_len;
2679 /* update the extent length and mark as initialized */
2680 ex->ee_block = orig_ex.ee_block;
2681 ex->ee_len = orig_ex.ee_len;
2682 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2683 ext4_ext_dirty(handle, inode, path + depth);
2684 /* zero out the first half */
2685 /* blocks available from iblock */
2690 * If there was a change of depth as part of the
2691 * insertion of ex3 above, we need to update the length
2692 * of the ex1 extent again here
2694 if (ex1 && ex1 != ex) {
2696 ex1->ee_len = cpu_to_le16(iblock - ee_block);
2697 ext4_ext_mark_uninitialized(ex1);
2700 /* ex2: iblock to iblock + maxblocks-1 : initialised */
2701 ex2->ee_block = cpu_to_le32(iblock);
2702 ext4_ext_store_pblock(ex2, newblock);
2703 ex2->ee_len = cpu_to_le16(allocated);
2707 * New (initialized) extent starts from the first block
2708 * in the current extent. i.e., ex2 == ex
2709 * We have to see if it can be merged with the extent
2712 if (ex2 > EXT_FIRST_EXTENT(eh)) {
2714 * To merge left, pass "ex2 - 1" to try_to_merge(),
2715 * since it merges towards right _only_.
2717 ret = ext4_ext_try_to_merge(inode, path, ex2 - 1);
2719 err = ext4_ext_correct_indexes(handle, inode, path);
2722 depth = ext_depth(inode);
2727 * Try to Merge towards right. This might be required
2728 * only when the whole extent is being written to.
2729 * i.e. ex2 == ex and ex3 == NULL.
2732 ret = ext4_ext_try_to_merge(inode, path, ex2);
2734 err = ext4_ext_correct_indexes(handle, inode, path);
2739 /* Mark modified extent as dirty */
2740 err = ext4_ext_dirty(handle, inode, path + depth);
2743 err = ext4_ext_insert_extent(handle, inode, path, &newex);
2744 if (err == -ENOSPC) {
2745 err = ext4_ext_zeroout(inode, &orig_ex);
2747 goto fix_extent_len;
2748 /* update the extent length and mark as initialized */
2749 ex->ee_block = orig_ex.ee_block;
2750 ex->ee_len = orig_ex.ee_len;
2751 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2752 ext4_ext_dirty(handle, inode, path + depth);
2753 /* zero out the first half */
2756 goto fix_extent_len;
2758 ext4_ext_show_leaf(inode, path);
2759 return err ? err : allocated;
2762 ex->ee_block = orig_ex.ee_block;
2763 ex->ee_len = orig_ex.ee_len;
2764 ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
2765 ext4_ext_mark_uninitialized(ex);
2766 ext4_ext_dirty(handle, inode, path + depth);
2771 * Block allocation/map/preallocation routine for extents based files
2774 * Need to be called with
2775 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
2776 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
2778 * return > 0, number of of blocks already mapped/allocated
2779 * if create == 0 and these are pre-allocated blocks
2780 * buffer head is unmapped
2781 * otherwise blocks are mapped
2783 * return = 0, if plain look up failed (blocks have not been allocated)
2784 * buffer head is unmapped
2786 * return < 0, error case.
2788 int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
2790 unsigned int max_blocks, struct buffer_head *bh_result,
2793 struct ext4_ext_path *path = NULL;
2794 struct ext4_extent_header *eh;
2795 struct ext4_extent newex, *ex;
2796 ext4_fsblk_t newblock;
2797 int err = 0, depth, ret, cache_type;
2798 unsigned int allocated = 0;
2799 struct ext4_allocation_request ar;
2801 __clear_bit(BH_New, &bh_result->b_state);
2802 ext_debug("blocks %u/%u requested for inode %lu\n",
2803 iblock, max_blocks, inode->i_ino);
2805 /* check in cache */
2806 cache_type = ext4_ext_in_cache(inode, iblock, &newex);
2808 if (cache_type == EXT4_EXT_CACHE_GAP) {
2809 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
2811 * block isn't allocated yet and
2812 * user doesn't want to allocate it
2816 /* we should allocate requested block */
2817 } else if (cache_type == EXT4_EXT_CACHE_EXTENT) {
2818 /* block is already allocated */
2820 - le32_to_cpu(newex.ee_block)
2821 + ext_pblock(&newex);
2822 /* number of remaining blocks in the extent */
2823 allocated = ext4_ext_get_actual_len(&newex) -
2824 (iblock - le32_to_cpu(newex.ee_block));
2831 /* find extent for this block */
2832 path = ext4_ext_find_extent(inode, iblock, NULL);
2834 err = PTR_ERR(path);
2839 depth = ext_depth(inode);
2842 * consistent leaf must not be empty;
2843 * this situation is possible, though, _during_ tree modification;
2844 * this is why assert can't be put in ext4_ext_find_extent()
2846 BUG_ON(path[depth].p_ext == NULL && depth != 0);
2847 eh = path[depth].p_hdr;
2849 ex = path[depth].p_ext;
2851 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
2852 ext4_fsblk_t ee_start = ext_pblock(ex);
2853 unsigned short ee_len;
2856 * Uninitialized extents are treated as holes, except that
2857 * we split out initialized portions during a write.
2859 ee_len = ext4_ext_get_actual_len(ex);
2860 /* if found extent covers block, simply return it */
2861 if (iblock >= ee_block && iblock < ee_block + ee_len) {
2862 newblock = iblock - ee_block + ee_start;
2863 /* number of remaining blocks in the extent */
2864 allocated = ee_len - (iblock - ee_block);
2865 ext_debug("%u fit into %u:%d -> %llu\n", iblock,
2866 ee_block, ee_len, newblock);
2868 /* Do not put uninitialized extent in the cache */
2869 if (!ext4_ext_is_uninitialized(ex)) {
2870 ext4_ext_put_in_cache(inode, ee_block,
2872 EXT4_EXT_CACHE_EXTENT);
2875 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT)
2877 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
2878 if (allocated > max_blocks)
2879 allocated = max_blocks;
2881 * We have blocks reserved already. We
2882 * return allocated blocks so that delalloc
2883 * won't do block reservation for us. But
2884 * the buffer head will be unmapped so that
2885 * a read from the block returns 0s.
2887 set_buffer_unwritten(bh_result);
2888 bh_result->b_bdev = inode->i_sb->s_bdev;
2889 bh_result->b_blocknr = newblock;
2893 ret = ext4_ext_convert_to_initialized(handle, inode,
2906 * requested block isn't allocated yet;
2907 * we couldn't try to create block if create flag is zero
2909 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
2911 * put just found gap into cache to speed up
2912 * subsequent requests
2914 ext4_ext_put_gap_in_cache(inode, path, iblock);
2918 * Okay, we need to do block allocation.
2921 /* find neighbour allocated blocks */
2923 err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
2927 err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright);
2932 * See if request is beyond maximum number of blocks we can have in
2933 * a single extent. For an initialized extent this limit is
2934 * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
2935 * EXT_UNINIT_MAX_LEN.
2937 if (max_blocks > EXT_INIT_MAX_LEN &&
2938 !(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
2939 max_blocks = EXT_INIT_MAX_LEN;
2940 else if (max_blocks > EXT_UNINIT_MAX_LEN &&
2941 (flags & EXT4_GET_BLOCKS_UNINIT_EXT))
2942 max_blocks = EXT_UNINIT_MAX_LEN;
2944 /* Check if we can really insert (iblock)::(iblock+max_blocks) extent */
2945 newex.ee_block = cpu_to_le32(iblock);
2946 newex.ee_len = cpu_to_le16(max_blocks);
2947 err = ext4_ext_check_overlap(inode, &newex, path);
2949 allocated = ext4_ext_get_actual_len(&newex);
2951 allocated = max_blocks;
2953 /* allocate new block */
2955 ar.goal = ext4_ext_find_goal(inode, path, iblock);
2956 ar.logical = iblock;
2958 if (S_ISREG(inode->i_mode))
2959 ar.flags = EXT4_MB_HINT_DATA;
2961 /* disable in-core preallocation for non-regular files */
2963 newblock = ext4_mb_new_blocks(handle, &ar, &err);
2966 ext_debug("allocate new block: goal %llu, found %llu/%u\n",
2967 ar.goal, newblock, allocated);
2969 /* try to insert new extent into found leaf and return */
2970 ext4_ext_store_pblock(&newex, newblock);
2971 newex.ee_len = cpu_to_le16(ar.len);
2972 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT) /* Mark uninitialized */
2973 ext4_ext_mark_uninitialized(&newex);
2974 err = ext4_ext_insert_extent(handle, inode, path, &newex);
2976 /* free data blocks we just allocated */
2977 /* not a good idea to call discard here directly,
2978 * but otherwise we'd need to call it every free() */
2979 ext4_discard_preallocations(inode);
2980 ext4_free_blocks(handle, inode, ext_pblock(&newex),
2981 ext4_ext_get_actual_len(&newex), 0);
2985 /* previous routine could use block we allocated */
2986 newblock = ext_pblock(&newex);
2987 allocated = ext4_ext_get_actual_len(&newex);
2989 set_buffer_new(bh_result);
2991 /* Cache only when it is _not_ an uninitialized extent */
2992 if ((flags & EXT4_GET_BLOCKS_UNINIT_EXT) == 0)
2993 ext4_ext_put_in_cache(inode, iblock, allocated, newblock,
2994 EXT4_EXT_CACHE_EXTENT);
2996 if (allocated > max_blocks)
2997 allocated = max_blocks;
2998 ext4_ext_show_leaf(inode, path);
2999 set_buffer_mapped(bh_result);
3000 bh_result->b_bdev = inode->i_sb->s_bdev;
3001 bh_result->b_blocknr = newblock;
3004 ext4_ext_drop_refs(path);
3007 return err ? err : allocated;
3010 void ext4_ext_truncate(struct inode *inode)
3012 struct address_space *mapping = inode->i_mapping;
3013 struct super_block *sb = inode->i_sb;
3014 ext4_lblk_t last_block;
3019 * probably first extent we're gonna free will be last in block
3021 err = ext4_writepage_trans_blocks(inode);
3022 handle = ext4_journal_start(inode, err);
3026 if (inode->i_size & (sb->s_blocksize - 1))
3027 ext4_block_truncate_page(handle, mapping, inode->i_size);
3029 if (ext4_orphan_add(handle, inode))
3032 down_write(&EXT4_I(inode)->i_data_sem);
3033 ext4_ext_invalidate_cache(inode);
3035 ext4_discard_preallocations(inode);
3038 * TODO: optimization is possible here.
3039 * Probably we need not scan at all,
3040 * because page truncation is enough.
3043 /* we have to know where to truncate from in crash case */
3044 EXT4_I(inode)->i_disksize = inode->i_size;
3045 ext4_mark_inode_dirty(handle, inode);
3047 last_block = (inode->i_size + sb->s_blocksize - 1)
3048 >> EXT4_BLOCK_SIZE_BITS(sb);
3049 err = ext4_ext_remove_space(inode, last_block);
3051 /* In a multi-transaction truncate, we only make the final
3052 * transaction synchronous.
3055 ext4_handle_sync(handle);
3058 up_write(&EXT4_I(inode)->i_data_sem);
3060 * If this was a simple ftruncate() and the file will remain alive,
3061 * then we need to clear up the orphan record which we created above.
3062 * However, if this was a real unlink then we were called by
3063 * ext4_delete_inode(), and we allow that function to clean up the
3064 * orphan info for us.
3067 ext4_orphan_del(handle, inode);
3069 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
3070 ext4_mark_inode_dirty(handle, inode);
3071 ext4_journal_stop(handle);
3074 static void ext4_falloc_update_inode(struct inode *inode,
3075 int mode, loff_t new_size, int update_ctime)
3077 struct timespec now;
3080 now = current_fs_time(inode->i_sb);
3081 if (!timespec_equal(&inode->i_ctime, &now))
3082 inode->i_ctime = now;
3085 * Update only when preallocation was requested beyond
3088 if (!(mode & FALLOC_FL_KEEP_SIZE)) {
3089 if (new_size > i_size_read(inode))
3090 i_size_write(inode, new_size);
3091 if (new_size > EXT4_I(inode)->i_disksize)
3092 ext4_update_i_disksize(inode, new_size);
3098 * preallocate space for a file. This implements ext4's fallocate inode
3099 * operation, which gets called from sys_fallocate system call.
3100 * For block-mapped files, posix_fallocate should fall back to the method
3101 * of writing zeroes to the required new blocks (the same behavior which is
3102 * expected for file systems which do not support fallocate() system call).
3104 long ext4_fallocate(struct inode *inode, int mode, loff_t offset, loff_t len)
3109 unsigned int max_blocks;
3113 struct buffer_head map_bh;
3114 unsigned int credits, blkbits = inode->i_blkbits;
3117 * currently supporting (pre)allocate mode for extent-based
3120 if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
3123 /* preallocation to directories is currently not supported */
3124 if (S_ISDIR(inode->i_mode))
3127 block = offset >> blkbits;
3129 * We can't just convert len to max_blocks because
3130 * If blocksize = 4096 offset = 3072 and len = 2048
3132 max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
3135 * credits to insert 1 extent into extent tree
3137 credits = ext4_chunk_trans_blocks(inode, max_blocks);
3138 mutex_lock(&inode->i_mutex);
3140 while (ret >= 0 && ret < max_blocks) {
3141 block = block + ret;
3142 max_blocks = max_blocks - ret;
3143 handle = ext4_journal_start(inode, credits);
3144 if (IS_ERR(handle)) {
3145 ret = PTR_ERR(handle);
3149 ret = ext4_get_blocks(handle, inode, block,
3150 max_blocks, &map_bh,
3151 EXT4_GET_BLOCKS_CREATE_UNINIT_EXT);
3155 printk(KERN_ERR "%s: ext4_ext_get_blocks "
3156 "returned error inode#%lu, block=%u, "
3157 "max_blocks=%u", __func__,
3158 inode->i_ino, block, max_blocks);
3160 ext4_mark_inode_dirty(handle, inode);
3161 ret2 = ext4_journal_stop(handle);
3164 if ((block + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
3165 blkbits) >> blkbits))
3166 new_size = offset + len;
3168 new_size = (block + ret) << blkbits;
3170 ext4_falloc_update_inode(inode, mode, new_size,
3171 buffer_new(&map_bh));
3172 ext4_mark_inode_dirty(handle, inode);
3173 ret2 = ext4_journal_stop(handle);
3177 if (ret == -ENOSPC &&
3178 ext4_should_retry_alloc(inode->i_sb, &retries)) {
3182 mutex_unlock(&inode->i_mutex);
3183 return ret > 0 ? ret2 : ret;
3187 * Callback function called for each extent to gather FIEMAP information.
3189 static int ext4_ext_fiemap_cb(struct inode *inode, struct ext4_ext_path *path,
3190 struct ext4_ext_cache *newex, struct ext4_extent *ex,
3193 struct fiemap_extent_info *fieinfo = data;
3194 unsigned char blksize_bits = inode->i_sb->s_blocksize_bits;
3201 logical = (__u64)newex->ec_block << blksize_bits;
3203 if (newex->ec_type == EXT4_EXT_CACHE_GAP) {
3206 struct buffer_head *bh = NULL;
3208 offset = logical >> PAGE_SHIFT;
3209 page = find_get_page(inode->i_mapping, offset);
3210 if (!page || !page_has_buffers(page))
3211 return EXT_CONTINUE;
3213 bh = page_buffers(page);
3216 return EXT_CONTINUE;
3218 if (buffer_delay(bh)) {
3219 flags |= FIEMAP_EXTENT_DELALLOC;
3220 page_cache_release(page);
3222 page_cache_release(page);
3223 return EXT_CONTINUE;
3227 physical = (__u64)newex->ec_start << blksize_bits;
3228 length = (__u64)newex->ec_len << blksize_bits;
3230 if (ex && ext4_ext_is_uninitialized(ex))
3231 flags |= FIEMAP_EXTENT_UNWRITTEN;
3234 * If this extent reaches EXT_MAX_BLOCK, it must be last.
3236 * Or if ext4_ext_next_allocated_block is EXT_MAX_BLOCK,
3237 * this also indicates no more allocated blocks.
3239 * XXX this might miss a single-block extent at EXT_MAX_BLOCK
3241 if (ext4_ext_next_allocated_block(path) == EXT_MAX_BLOCK ||
3242 newex->ec_block + newex->ec_len - 1 == EXT_MAX_BLOCK) {
3243 loff_t size = i_size_read(inode);
3244 loff_t bs = EXT4_BLOCK_SIZE(inode->i_sb);
3246 flags |= FIEMAP_EXTENT_LAST;
3247 if ((flags & FIEMAP_EXTENT_DELALLOC) &&
3248 logical+length > size)
3249 length = (size - logical + bs - 1) & ~(bs-1);
3252 error = fiemap_fill_next_extent(fieinfo, logical, physical,
3259 return EXT_CONTINUE;
3262 /* fiemap flags we can handle specified here */
3263 #define EXT4_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
3265 static int ext4_xattr_fiemap(struct inode *inode,
3266 struct fiemap_extent_info *fieinfo)
3270 __u32 flags = FIEMAP_EXTENT_LAST;
3271 int blockbits = inode->i_sb->s_blocksize_bits;
3275 if (EXT4_I(inode)->i_state & EXT4_STATE_XATTR) {
3276 struct ext4_iloc iloc;
3277 int offset; /* offset of xattr in inode */
3279 error = ext4_get_inode_loc(inode, &iloc);
3282 physical = iloc.bh->b_blocknr << blockbits;
3283 offset = EXT4_GOOD_OLD_INODE_SIZE +
3284 EXT4_I(inode)->i_extra_isize;
3286 length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
3287 flags |= FIEMAP_EXTENT_DATA_INLINE;
3288 } else { /* external block */
3289 physical = EXT4_I(inode)->i_file_acl << blockbits;
3290 length = inode->i_sb->s_blocksize;
3294 error = fiemap_fill_next_extent(fieinfo, 0, physical,
3296 return (error < 0 ? error : 0);
3299 int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
3300 __u64 start, __u64 len)
3302 ext4_lblk_t start_blk;
3303 ext4_lblk_t len_blks;
3306 /* fallback to generic here if not in extents fmt */
3307 if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
3308 return generic_block_fiemap(inode, fieinfo, start, len,
3311 if (fiemap_check_flags(fieinfo, EXT4_FIEMAP_FLAGS))
3314 if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
3315 error = ext4_xattr_fiemap(inode, fieinfo);
3317 start_blk = start >> inode->i_sb->s_blocksize_bits;
3318 len_blks = len >> inode->i_sb->s_blocksize_bits;
3321 * Walk the extent tree gathering extent information.
3322 * ext4_ext_fiemap_cb will push extents back to user.
3324 down_read(&EXT4_I(inode)->i_data_sem);
3325 error = ext4_ext_walk_space(inode, start_blk, len_blks,
3326 ext4_ext_fiemap_cb, fieinfo);
3327 up_read(&EXT4_I(inode)->i_data_sem);