]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blame_incremental - fs/ext4/extents.c
ext4: cleanup ext4_ext_grow_indepth code
[mirror_ubuntu-artful-kernel.git] / fs / ext4 / extents.c
... / ...
CommitLineData
1/*
2 * Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com
3 * Written by Alex Tomas <alex@clusterfs.com>
4 *
5 * Architecture independence:
6 * Copyright (c) 2005, Bull S.A.
7 * Written by Pierre Peiffer <pierre.peiffer@bull.net>
8 *
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.
12 *
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.
17 *
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-
21 */
22
23/*
24 * Extents support for EXT4
25 *
26 * TODO:
27 * - ext4*_error() should be used in some situations
28 * - analyze all BUG()/BUG_ON(), use -EIO where appropriate
29 * - smart tree reduction
30 */
31
32#include <linux/module.h>
33#include <linux/fs.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
46#include <trace/events/ext4.h>
47
48static int ext4_split_extent(handle_t *handle,
49 struct inode *inode,
50 struct ext4_ext_path *path,
51 struct ext4_map_blocks *map,
52 int split_flag,
53 int flags);
54
55static int ext4_ext_truncate_extend_restart(handle_t *handle,
56 struct inode *inode,
57 int needed)
58{
59 int err;
60
61 if (!ext4_handle_valid(handle))
62 return 0;
63 if (handle->h_buffer_credits > needed)
64 return 0;
65 err = ext4_journal_extend(handle, needed);
66 if (err <= 0)
67 return err;
68 err = ext4_truncate_restart_trans(handle, inode, needed);
69 if (err == 0)
70 err = -EAGAIN;
71
72 return err;
73}
74
75/*
76 * could return:
77 * - EROFS
78 * - ENOMEM
79 */
80static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
81 struct ext4_ext_path *path)
82{
83 if (path->p_bh) {
84 /* path points to block */
85 return ext4_journal_get_write_access(handle, path->p_bh);
86 }
87 /* path points to leaf/index in inode body */
88 /* we use in-core data, no need to protect them */
89 return 0;
90}
91
92/*
93 * could return:
94 * - EROFS
95 * - ENOMEM
96 * - EIO
97 */
98#define ext4_ext_dirty(handle, inode, path) \
99 __ext4_ext_dirty(__func__, __LINE__, (handle), (inode), (path))
100static int __ext4_ext_dirty(const char *where, unsigned int line,
101 handle_t *handle, struct inode *inode,
102 struct ext4_ext_path *path)
103{
104 int err;
105 if (path->p_bh) {
106 /* path points to block */
107 err = __ext4_handle_dirty_metadata(where, line, handle,
108 inode, path->p_bh);
109 } else {
110 /* path points to leaf/index in inode body */
111 err = ext4_mark_inode_dirty(handle, inode);
112 }
113 return err;
114}
115
116static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
117 struct ext4_ext_path *path,
118 ext4_lblk_t block)
119{
120 int depth;
121
122 if (path) {
123 struct ext4_extent *ex;
124 depth = path->p_depth;
125
126 /*
127 * Try to predict block placement assuming that we are
128 * filling in a file which will eventually be
129 * non-sparse --- i.e., in the case of libbfd writing
130 * an ELF object sections out-of-order but in a way
131 * the eventually results in a contiguous object or
132 * executable file, or some database extending a table
133 * space file. However, this is actually somewhat
134 * non-ideal if we are writing a sparse file such as
135 * qemu or KVM writing a raw image file that is going
136 * to stay fairly sparse, since it will end up
137 * fragmenting the file system's free space. Maybe we
138 * should have some hueristics or some way to allow
139 * userspace to pass a hint to file system,
140 * especially if the latter case turns out to be
141 * common.
142 */
143 ex = path[depth].p_ext;
144 if (ex) {
145 ext4_fsblk_t ext_pblk = ext4_ext_pblock(ex);
146 ext4_lblk_t ext_block = le32_to_cpu(ex->ee_block);
147
148 if (block > ext_block)
149 return ext_pblk + (block - ext_block);
150 else
151 return ext_pblk - (ext_block - block);
152 }
153
154 /* it looks like index is empty;
155 * try to find starting block from index itself */
156 if (path[depth].p_bh)
157 return path[depth].p_bh->b_blocknr;
158 }
159
160 /* OK. use inode's group */
161 return ext4_inode_to_goal_block(inode);
162}
163
164/*
165 * Allocation for a meta data block
166 */
167static ext4_fsblk_t
168ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
169 struct ext4_ext_path *path,
170 struct ext4_extent *ex, int *err, unsigned int flags)
171{
172 ext4_fsblk_t goal, newblock;
173
174 goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
175 newblock = ext4_new_meta_blocks(handle, inode, goal, flags,
176 NULL, err);
177 return newblock;
178}
179
180static inline int ext4_ext_space_block(struct inode *inode, int check)
181{
182 int size;
183
184 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
185 / sizeof(struct ext4_extent);
186 if (!check) {
187#ifdef AGGRESSIVE_TEST
188 if (size > 6)
189 size = 6;
190#endif
191 }
192 return size;
193}
194
195static inline int ext4_ext_space_block_idx(struct inode *inode, int check)
196{
197 int size;
198
199 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
200 / sizeof(struct ext4_extent_idx);
201 if (!check) {
202#ifdef AGGRESSIVE_TEST
203 if (size > 5)
204 size = 5;
205#endif
206 }
207 return size;
208}
209
210static inline int ext4_ext_space_root(struct inode *inode, int check)
211{
212 int size;
213
214 size = sizeof(EXT4_I(inode)->i_data);
215 size -= sizeof(struct ext4_extent_header);
216 size /= sizeof(struct ext4_extent);
217 if (!check) {
218#ifdef AGGRESSIVE_TEST
219 if (size > 3)
220 size = 3;
221#endif
222 }
223 return size;
224}
225
226static inline int ext4_ext_space_root_idx(struct inode *inode, int check)
227{
228 int size;
229
230 size = sizeof(EXT4_I(inode)->i_data);
231 size -= sizeof(struct ext4_extent_header);
232 size /= sizeof(struct ext4_extent_idx);
233 if (!check) {
234#ifdef AGGRESSIVE_TEST
235 if (size > 4)
236 size = 4;
237#endif
238 }
239 return size;
240}
241
242/*
243 * Calculate the number of metadata blocks needed
244 * to allocate @blocks
245 * Worse case is one block per extent
246 */
247int ext4_ext_calc_metadata_amount(struct inode *inode, ext4_lblk_t lblock)
248{
249 struct ext4_inode_info *ei = EXT4_I(inode);
250 int idxs, num = 0;
251
252 idxs = ((inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
253 / sizeof(struct ext4_extent_idx));
254
255 /*
256 * If the new delayed allocation block is contiguous with the
257 * previous da block, it can share index blocks with the
258 * previous block, so we only need to allocate a new index
259 * block every idxs leaf blocks. At ldxs**2 blocks, we need
260 * an additional index block, and at ldxs**3 blocks, yet
261 * another index blocks.
262 */
263 if (ei->i_da_metadata_calc_len &&
264 ei->i_da_metadata_calc_last_lblock+1 == lblock) {
265 if ((ei->i_da_metadata_calc_len % idxs) == 0)
266 num++;
267 if ((ei->i_da_metadata_calc_len % (idxs*idxs)) == 0)
268 num++;
269 if ((ei->i_da_metadata_calc_len % (idxs*idxs*idxs)) == 0) {
270 num++;
271 ei->i_da_metadata_calc_len = 0;
272 } else
273 ei->i_da_metadata_calc_len++;
274 ei->i_da_metadata_calc_last_lblock++;
275 return num;
276 }
277
278 /*
279 * In the worst case we need a new set of index blocks at
280 * every level of the inode's extent tree.
281 */
282 ei->i_da_metadata_calc_len = 1;
283 ei->i_da_metadata_calc_last_lblock = lblock;
284 return ext_depth(inode) + 1;
285}
286
287static int
288ext4_ext_max_entries(struct inode *inode, int depth)
289{
290 int max;
291
292 if (depth == ext_depth(inode)) {
293 if (depth == 0)
294 max = ext4_ext_space_root(inode, 1);
295 else
296 max = ext4_ext_space_root_idx(inode, 1);
297 } else {
298 if (depth == 0)
299 max = ext4_ext_space_block(inode, 1);
300 else
301 max = ext4_ext_space_block_idx(inode, 1);
302 }
303
304 return max;
305}
306
307static int ext4_valid_extent(struct inode *inode, struct ext4_extent *ext)
308{
309 ext4_fsblk_t block = ext4_ext_pblock(ext);
310 int len = ext4_ext_get_actual_len(ext);
311
312 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, len);
313}
314
315static int ext4_valid_extent_idx(struct inode *inode,
316 struct ext4_extent_idx *ext_idx)
317{
318 ext4_fsblk_t block = ext4_idx_pblock(ext_idx);
319
320 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, 1);
321}
322
323static int ext4_valid_extent_entries(struct inode *inode,
324 struct ext4_extent_header *eh,
325 int depth)
326{
327 struct ext4_extent *ext;
328 struct ext4_extent_idx *ext_idx;
329 unsigned short entries;
330 if (eh->eh_entries == 0)
331 return 1;
332
333 entries = le16_to_cpu(eh->eh_entries);
334
335 if (depth == 0) {
336 /* leaf entries */
337 ext = EXT_FIRST_EXTENT(eh);
338 while (entries) {
339 if (!ext4_valid_extent(inode, ext))
340 return 0;
341 ext++;
342 entries--;
343 }
344 } else {
345 ext_idx = EXT_FIRST_INDEX(eh);
346 while (entries) {
347 if (!ext4_valid_extent_idx(inode, ext_idx))
348 return 0;
349 ext_idx++;
350 entries--;
351 }
352 }
353 return 1;
354}
355
356static int __ext4_ext_check(const char *function, unsigned int line,
357 struct inode *inode, struct ext4_extent_header *eh,
358 int depth)
359{
360 const char *error_msg;
361 int max = 0;
362
363 if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
364 error_msg = "invalid magic";
365 goto corrupted;
366 }
367 if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
368 error_msg = "unexpected eh_depth";
369 goto corrupted;
370 }
371 if (unlikely(eh->eh_max == 0)) {
372 error_msg = "invalid eh_max";
373 goto corrupted;
374 }
375 max = ext4_ext_max_entries(inode, depth);
376 if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
377 error_msg = "too large eh_max";
378 goto corrupted;
379 }
380 if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
381 error_msg = "invalid eh_entries";
382 goto corrupted;
383 }
384 if (!ext4_valid_extent_entries(inode, eh, depth)) {
385 error_msg = "invalid extent entries";
386 goto corrupted;
387 }
388 return 0;
389
390corrupted:
391 ext4_error_inode(inode, function, line, 0,
392 "bad header/extent: %s - magic %x, "
393 "entries %u, max %u(%u), depth %u(%u)",
394 error_msg, le16_to_cpu(eh->eh_magic),
395 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
396 max, le16_to_cpu(eh->eh_depth), depth);
397
398 return -EIO;
399}
400
401#define ext4_ext_check(inode, eh, depth) \
402 __ext4_ext_check(__func__, __LINE__, inode, eh, depth)
403
404int ext4_ext_check_inode(struct inode *inode)
405{
406 return ext4_ext_check(inode, ext_inode_hdr(inode), ext_depth(inode));
407}
408
409#ifdef EXT_DEBUG
410static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
411{
412 int k, l = path->p_depth;
413
414 ext_debug("path:");
415 for (k = 0; k <= l; k++, path++) {
416 if (path->p_idx) {
417 ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block),
418 ext4_idx_pblock(path->p_idx));
419 } else if (path->p_ext) {
420 ext_debug(" %d:[%d]%d:%llu ",
421 le32_to_cpu(path->p_ext->ee_block),
422 ext4_ext_is_uninitialized(path->p_ext),
423 ext4_ext_get_actual_len(path->p_ext),
424 ext4_ext_pblock(path->p_ext));
425 } else
426 ext_debug(" []");
427 }
428 ext_debug("\n");
429}
430
431static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
432{
433 int depth = ext_depth(inode);
434 struct ext4_extent_header *eh;
435 struct ext4_extent *ex;
436 int i;
437
438 if (!path)
439 return;
440
441 eh = path[depth].p_hdr;
442 ex = EXT_FIRST_EXTENT(eh);
443
444 ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
445
446 for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
447 ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
448 ext4_ext_is_uninitialized(ex),
449 ext4_ext_get_actual_len(ex), ext4_ext_pblock(ex));
450 }
451 ext_debug("\n");
452}
453
454static void ext4_ext_show_move(struct inode *inode, struct ext4_ext_path *path,
455 ext4_fsblk_t newblock, int level)
456{
457 int depth = ext_depth(inode);
458 struct ext4_extent *ex;
459
460 if (depth != level) {
461 struct ext4_extent_idx *idx;
462 idx = path[level].p_idx;
463 while (idx <= EXT_MAX_INDEX(path[level].p_hdr)) {
464 ext_debug("%d: move %d:%llu in new index %llu\n", level,
465 le32_to_cpu(idx->ei_block),
466 ext4_idx_pblock(idx),
467 newblock);
468 idx++;
469 }
470
471 return;
472 }
473
474 ex = path[depth].p_ext;
475 while (ex <= EXT_MAX_EXTENT(path[depth].p_hdr)) {
476 ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
477 le32_to_cpu(ex->ee_block),
478 ext4_ext_pblock(ex),
479 ext4_ext_is_uninitialized(ex),
480 ext4_ext_get_actual_len(ex),
481 newblock);
482 ex++;
483 }
484}
485
486#else
487#define ext4_ext_show_path(inode, path)
488#define ext4_ext_show_leaf(inode, path)
489#define ext4_ext_show_move(inode, path, newblock, level)
490#endif
491
492void ext4_ext_drop_refs(struct ext4_ext_path *path)
493{
494 int depth = path->p_depth;
495 int i;
496
497 for (i = 0; i <= depth; i++, path++)
498 if (path->p_bh) {
499 brelse(path->p_bh);
500 path->p_bh = NULL;
501 }
502}
503
504/*
505 * ext4_ext_binsearch_idx:
506 * binary search for the closest index of the given block
507 * the header must be checked before calling this
508 */
509static void
510ext4_ext_binsearch_idx(struct inode *inode,
511 struct ext4_ext_path *path, ext4_lblk_t block)
512{
513 struct ext4_extent_header *eh = path->p_hdr;
514 struct ext4_extent_idx *r, *l, *m;
515
516
517 ext_debug("binsearch for %u(idx): ", block);
518
519 l = EXT_FIRST_INDEX(eh) + 1;
520 r = EXT_LAST_INDEX(eh);
521 while (l <= r) {
522 m = l + (r - l) / 2;
523 if (block < le32_to_cpu(m->ei_block))
524 r = m - 1;
525 else
526 l = m + 1;
527 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
528 m, le32_to_cpu(m->ei_block),
529 r, le32_to_cpu(r->ei_block));
530 }
531
532 path->p_idx = l - 1;
533 ext_debug(" -> %d->%lld ", le32_to_cpu(path->p_idx->ei_block),
534 ext4_idx_pblock(path->p_idx));
535
536#ifdef CHECK_BINSEARCH
537 {
538 struct ext4_extent_idx *chix, *ix;
539 int k;
540
541 chix = ix = EXT_FIRST_INDEX(eh);
542 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
543 if (k != 0 &&
544 le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
545 printk(KERN_DEBUG "k=%d, ix=0x%p, "
546 "first=0x%p\n", k,
547 ix, EXT_FIRST_INDEX(eh));
548 printk(KERN_DEBUG "%u <= %u\n",
549 le32_to_cpu(ix->ei_block),
550 le32_to_cpu(ix[-1].ei_block));
551 }
552 BUG_ON(k && le32_to_cpu(ix->ei_block)
553 <= le32_to_cpu(ix[-1].ei_block));
554 if (block < le32_to_cpu(ix->ei_block))
555 break;
556 chix = ix;
557 }
558 BUG_ON(chix != path->p_idx);
559 }
560#endif
561
562}
563
564/*
565 * ext4_ext_binsearch:
566 * binary search for closest extent of the given block
567 * the header must be checked before calling this
568 */
569static void
570ext4_ext_binsearch(struct inode *inode,
571 struct ext4_ext_path *path, ext4_lblk_t block)
572{
573 struct ext4_extent_header *eh = path->p_hdr;
574 struct ext4_extent *r, *l, *m;
575
576 if (eh->eh_entries == 0) {
577 /*
578 * this leaf is empty:
579 * we get such a leaf in split/add case
580 */
581 return;
582 }
583
584 ext_debug("binsearch for %u: ", block);
585
586 l = EXT_FIRST_EXTENT(eh) + 1;
587 r = EXT_LAST_EXTENT(eh);
588
589 while (l <= r) {
590 m = l + (r - l) / 2;
591 if (block < le32_to_cpu(m->ee_block))
592 r = m - 1;
593 else
594 l = m + 1;
595 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
596 m, le32_to_cpu(m->ee_block),
597 r, le32_to_cpu(r->ee_block));
598 }
599
600 path->p_ext = l - 1;
601 ext_debug(" -> %d:%llu:[%d]%d ",
602 le32_to_cpu(path->p_ext->ee_block),
603 ext4_ext_pblock(path->p_ext),
604 ext4_ext_is_uninitialized(path->p_ext),
605 ext4_ext_get_actual_len(path->p_ext));
606
607#ifdef CHECK_BINSEARCH
608 {
609 struct ext4_extent *chex, *ex;
610 int k;
611
612 chex = ex = EXT_FIRST_EXTENT(eh);
613 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
614 BUG_ON(k && le32_to_cpu(ex->ee_block)
615 <= le32_to_cpu(ex[-1].ee_block));
616 if (block < le32_to_cpu(ex->ee_block))
617 break;
618 chex = ex;
619 }
620 BUG_ON(chex != path->p_ext);
621 }
622#endif
623
624}
625
626int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
627{
628 struct ext4_extent_header *eh;
629
630 eh = ext_inode_hdr(inode);
631 eh->eh_depth = 0;
632 eh->eh_entries = 0;
633 eh->eh_magic = EXT4_EXT_MAGIC;
634 eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode, 0));
635 ext4_mark_inode_dirty(handle, inode);
636 ext4_ext_invalidate_cache(inode);
637 return 0;
638}
639
640struct ext4_ext_path *
641ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block,
642 struct ext4_ext_path *path)
643{
644 struct ext4_extent_header *eh;
645 struct buffer_head *bh;
646 short int depth, i, ppos = 0, alloc = 0;
647
648 eh = ext_inode_hdr(inode);
649 depth = ext_depth(inode);
650
651 /* account possible depth increase */
652 if (!path) {
653 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
654 GFP_NOFS);
655 if (!path)
656 return ERR_PTR(-ENOMEM);
657 alloc = 1;
658 }
659 path[0].p_hdr = eh;
660 path[0].p_bh = NULL;
661
662 i = depth;
663 /* walk through the tree */
664 while (i) {
665 int need_to_validate = 0;
666
667 ext_debug("depth %d: num %d, max %d\n",
668 ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
669
670 ext4_ext_binsearch_idx(inode, path + ppos, block);
671 path[ppos].p_block = ext4_idx_pblock(path[ppos].p_idx);
672 path[ppos].p_depth = i;
673 path[ppos].p_ext = NULL;
674
675 bh = sb_getblk(inode->i_sb, path[ppos].p_block);
676 if (unlikely(!bh))
677 goto err;
678 if (!bh_uptodate_or_lock(bh)) {
679 trace_ext4_ext_load_extent(inode, block,
680 path[ppos].p_block);
681 if (bh_submit_read(bh) < 0) {
682 put_bh(bh);
683 goto err;
684 }
685 /* validate the extent entries */
686 need_to_validate = 1;
687 }
688 eh = ext_block_hdr(bh);
689 ppos++;
690 if (unlikely(ppos > depth)) {
691 put_bh(bh);
692 EXT4_ERROR_INODE(inode,
693 "ppos %d > depth %d", ppos, depth);
694 goto err;
695 }
696 path[ppos].p_bh = bh;
697 path[ppos].p_hdr = eh;
698 i--;
699
700 if (need_to_validate && ext4_ext_check(inode, eh, i))
701 goto err;
702 }
703
704 path[ppos].p_depth = i;
705 path[ppos].p_ext = NULL;
706 path[ppos].p_idx = NULL;
707
708 /* find extent */
709 ext4_ext_binsearch(inode, path + ppos, block);
710 /* if not an empty leaf */
711 if (path[ppos].p_ext)
712 path[ppos].p_block = ext4_ext_pblock(path[ppos].p_ext);
713
714 ext4_ext_show_path(inode, path);
715
716 return path;
717
718err:
719 ext4_ext_drop_refs(path);
720 if (alloc)
721 kfree(path);
722 return ERR_PTR(-EIO);
723}
724
725/*
726 * ext4_ext_insert_index:
727 * insert new index [@logical;@ptr] into the block at @curp;
728 * check where to insert: before @curp or after @curp
729 */
730static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
731 struct ext4_ext_path *curp,
732 int logical, ext4_fsblk_t ptr)
733{
734 struct ext4_extent_idx *ix;
735 int len, err;
736
737 err = ext4_ext_get_access(handle, inode, curp);
738 if (err)
739 return err;
740
741 if (unlikely(logical == le32_to_cpu(curp->p_idx->ei_block))) {
742 EXT4_ERROR_INODE(inode,
743 "logical %d == ei_block %d!",
744 logical, le32_to_cpu(curp->p_idx->ei_block));
745 return -EIO;
746 }
747
748 if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries)
749 >= le16_to_cpu(curp->p_hdr->eh_max))) {
750 EXT4_ERROR_INODE(inode,
751 "eh_entries %d >= eh_max %d!",
752 le16_to_cpu(curp->p_hdr->eh_entries),
753 le16_to_cpu(curp->p_hdr->eh_max));
754 return -EIO;
755 }
756
757 len = EXT_MAX_INDEX(curp->p_hdr) - curp->p_idx;
758 if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
759 /* insert after */
760 if (curp->p_idx != EXT_LAST_INDEX(curp->p_hdr)) {
761 len = (len - 1) * sizeof(struct ext4_extent_idx);
762 len = len < 0 ? 0 : len;
763 ext_debug("insert new index %d after: %llu. "
764 "move %d from 0x%p to 0x%p\n",
765 logical, ptr, len,
766 (curp->p_idx + 1), (curp->p_idx + 2));
767 memmove(curp->p_idx + 2, curp->p_idx + 1, len);
768 }
769 ix = curp->p_idx + 1;
770 } else {
771 /* insert before */
772 len = len * sizeof(struct ext4_extent_idx);
773 len = len < 0 ? 0 : len;
774 ext_debug("insert new index %d before: %llu. "
775 "move %d from 0x%p to 0x%p\n",
776 logical, ptr, len,
777 curp->p_idx, (curp->p_idx + 1));
778 memmove(curp->p_idx + 1, curp->p_idx, len);
779 ix = curp->p_idx;
780 }
781
782 if (unlikely(ix > EXT_MAX_INDEX(curp->p_hdr))) {
783 EXT4_ERROR_INODE(inode, "ix > EXT_MAX_INDEX!");
784 return -EIO;
785 }
786
787 ix->ei_block = cpu_to_le32(logical);
788 ext4_idx_store_pblock(ix, ptr);
789 le16_add_cpu(&curp->p_hdr->eh_entries, 1);
790
791 if (unlikely(ix > EXT_LAST_INDEX(curp->p_hdr))) {
792 EXT4_ERROR_INODE(inode, "ix > EXT_LAST_INDEX!");
793 return -EIO;
794 }
795
796 err = ext4_ext_dirty(handle, inode, curp);
797 ext4_std_error(inode->i_sb, err);
798
799 return err;
800}
801
802/*
803 * ext4_ext_split:
804 * inserts new subtree into the path, using free index entry
805 * at depth @at:
806 * - allocates all needed blocks (new leaf and all intermediate index blocks)
807 * - makes decision where to split
808 * - moves remaining extents and index entries (right to the split point)
809 * into the newly allocated blocks
810 * - initializes subtree
811 */
812static int ext4_ext_split(handle_t *handle, struct inode *inode,
813 unsigned int flags,
814 struct ext4_ext_path *path,
815 struct ext4_extent *newext, int at)
816{
817 struct buffer_head *bh = NULL;
818 int depth = ext_depth(inode);
819 struct ext4_extent_header *neh;
820 struct ext4_extent_idx *fidx;
821 int i = at, k, m, a;
822 ext4_fsblk_t newblock, oldblock;
823 __le32 border;
824 ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
825 int err = 0;
826
827 /* make decision: where to split? */
828 /* FIXME: now decision is simplest: at current extent */
829
830 /* if current leaf will be split, then we should use
831 * border from split point */
832 if (unlikely(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr))) {
833 EXT4_ERROR_INODE(inode, "p_ext > EXT_MAX_EXTENT!");
834 return -EIO;
835 }
836 if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
837 border = path[depth].p_ext[1].ee_block;
838 ext_debug("leaf will be split."
839 " next leaf starts at %d\n",
840 le32_to_cpu(border));
841 } else {
842 border = newext->ee_block;
843 ext_debug("leaf will be added."
844 " next leaf starts at %d\n",
845 le32_to_cpu(border));
846 }
847
848 /*
849 * If error occurs, then we break processing
850 * and mark filesystem read-only. index won't
851 * be inserted and tree will be in consistent
852 * state. Next mount will repair buffers too.
853 */
854
855 /*
856 * Get array to track all allocated blocks.
857 * We need this to handle errors and free blocks
858 * upon them.
859 */
860 ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
861 if (!ablocks)
862 return -ENOMEM;
863
864 /* allocate all needed blocks */
865 ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
866 for (a = 0; a < depth - at; a++) {
867 newblock = ext4_ext_new_meta_block(handle, inode, path,
868 newext, &err, flags);
869 if (newblock == 0)
870 goto cleanup;
871 ablocks[a] = newblock;
872 }
873
874 /* initialize new leaf */
875 newblock = ablocks[--a];
876 if (unlikely(newblock == 0)) {
877 EXT4_ERROR_INODE(inode, "newblock == 0!");
878 err = -EIO;
879 goto cleanup;
880 }
881 bh = sb_getblk(inode->i_sb, newblock);
882 if (!bh) {
883 err = -EIO;
884 goto cleanup;
885 }
886 lock_buffer(bh);
887
888 err = ext4_journal_get_create_access(handle, bh);
889 if (err)
890 goto cleanup;
891
892 neh = ext_block_hdr(bh);
893 neh->eh_entries = 0;
894 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
895 neh->eh_magic = EXT4_EXT_MAGIC;
896 neh->eh_depth = 0;
897
898 /* move remainder of path[depth] to the new leaf */
899 if (unlikely(path[depth].p_hdr->eh_entries !=
900 path[depth].p_hdr->eh_max)) {
901 EXT4_ERROR_INODE(inode, "eh_entries %d != eh_max %d!",
902 path[depth].p_hdr->eh_entries,
903 path[depth].p_hdr->eh_max);
904 err = -EIO;
905 goto cleanup;
906 }
907 /* start copy from next extent */
908 m = EXT_MAX_EXTENT(path[depth].p_hdr) - path[depth].p_ext++;
909 ext4_ext_show_move(inode, path, newblock, depth);
910 if (m) {
911 struct ext4_extent *ex;
912 ex = EXT_FIRST_EXTENT(neh);
913 memmove(ex, path[depth].p_ext, sizeof(struct ext4_extent) * m);
914 le16_add_cpu(&neh->eh_entries, m);
915 }
916
917 set_buffer_uptodate(bh);
918 unlock_buffer(bh);
919
920 err = ext4_handle_dirty_metadata(handle, inode, bh);
921 if (err)
922 goto cleanup;
923 brelse(bh);
924 bh = NULL;
925
926 /* correct old leaf */
927 if (m) {
928 err = ext4_ext_get_access(handle, inode, path + depth);
929 if (err)
930 goto cleanup;
931 le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
932 err = ext4_ext_dirty(handle, inode, path + depth);
933 if (err)
934 goto cleanup;
935
936 }
937
938 /* create intermediate indexes */
939 k = depth - at - 1;
940 if (unlikely(k < 0)) {
941 EXT4_ERROR_INODE(inode, "k %d < 0!", k);
942 err = -EIO;
943 goto cleanup;
944 }
945 if (k)
946 ext_debug("create %d intermediate indices\n", k);
947 /* insert new index into current index block */
948 /* current depth stored in i var */
949 i = depth - 1;
950 while (k--) {
951 oldblock = newblock;
952 newblock = ablocks[--a];
953 bh = sb_getblk(inode->i_sb, newblock);
954 if (!bh) {
955 err = -EIO;
956 goto cleanup;
957 }
958 lock_buffer(bh);
959
960 err = ext4_journal_get_create_access(handle, bh);
961 if (err)
962 goto cleanup;
963
964 neh = ext_block_hdr(bh);
965 neh->eh_entries = cpu_to_le16(1);
966 neh->eh_magic = EXT4_EXT_MAGIC;
967 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
968 neh->eh_depth = cpu_to_le16(depth - i);
969 fidx = EXT_FIRST_INDEX(neh);
970 fidx->ei_block = border;
971 ext4_idx_store_pblock(fidx, oldblock);
972
973 ext_debug("int.index at %d (block %llu): %u -> %llu\n",
974 i, newblock, le32_to_cpu(border), oldblock);
975
976 /* move remainder of path[i] to the new index block */
977 if (unlikely(EXT_MAX_INDEX(path[i].p_hdr) !=
978 EXT_LAST_INDEX(path[i].p_hdr))) {
979 EXT4_ERROR_INODE(inode,
980 "EXT_MAX_INDEX != EXT_LAST_INDEX ee_block %d!",
981 le32_to_cpu(path[i].p_ext->ee_block));
982 err = -EIO;
983 goto cleanup;
984 }
985 /* start copy indexes */
986 m = EXT_MAX_INDEX(path[i].p_hdr) - path[i].p_idx++;
987 ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
988 EXT_MAX_INDEX(path[i].p_hdr));
989 ext4_ext_show_move(inode, path, newblock, i);
990 if (m) {
991 memmove(++fidx, path[i].p_idx,
992 sizeof(struct ext4_extent_idx) * m);
993 le16_add_cpu(&neh->eh_entries, m);
994 }
995 set_buffer_uptodate(bh);
996 unlock_buffer(bh);
997
998 err = ext4_handle_dirty_metadata(handle, inode, bh);
999 if (err)
1000 goto cleanup;
1001 brelse(bh);
1002 bh = NULL;
1003
1004 /* correct old index */
1005 if (m) {
1006 err = ext4_ext_get_access(handle, inode, path + i);
1007 if (err)
1008 goto cleanup;
1009 le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
1010 err = ext4_ext_dirty(handle, inode, path + i);
1011 if (err)
1012 goto cleanup;
1013 }
1014
1015 i--;
1016 }
1017
1018 /* insert new index */
1019 err = ext4_ext_insert_index(handle, inode, path + at,
1020 le32_to_cpu(border), newblock);
1021
1022cleanup:
1023 if (bh) {
1024 if (buffer_locked(bh))
1025 unlock_buffer(bh);
1026 brelse(bh);
1027 }
1028
1029 if (err) {
1030 /* free all allocated blocks in error case */
1031 for (i = 0; i < depth; i++) {
1032 if (!ablocks[i])
1033 continue;
1034 ext4_free_blocks(handle, inode, NULL, ablocks[i], 1,
1035 EXT4_FREE_BLOCKS_METADATA);
1036 }
1037 }
1038 kfree(ablocks);
1039
1040 return err;
1041}
1042
1043/*
1044 * ext4_ext_grow_indepth:
1045 * implements tree growing procedure:
1046 * - allocates new block
1047 * - moves top-level data (index block or leaf) into the new block
1048 * - initializes new top-level, creating index that points to the
1049 * just created block
1050 */
1051static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
1052 unsigned int flags,
1053 struct ext4_extent *newext)
1054{
1055 struct ext4_extent_header *neh;
1056 struct buffer_head *bh;
1057 ext4_fsblk_t newblock;
1058 int err = 0;
1059
1060 newblock = ext4_ext_new_meta_block(handle, inode, NULL,
1061 newext, &err, flags);
1062 if (newblock == 0)
1063 return err;
1064
1065 bh = sb_getblk(inode->i_sb, newblock);
1066 if (!bh) {
1067 err = -EIO;
1068 ext4_std_error(inode->i_sb, err);
1069 return err;
1070 }
1071 lock_buffer(bh);
1072
1073 err = ext4_journal_get_create_access(handle, bh);
1074 if (err) {
1075 unlock_buffer(bh);
1076 goto out;
1077 }
1078
1079 /* move top-level index/leaf into new block */
1080 memmove(bh->b_data, EXT4_I(inode)->i_data,
1081 sizeof(EXT4_I(inode)->i_data));
1082
1083 /* set size of new block */
1084 neh = ext_block_hdr(bh);
1085 /* old root could have indexes or leaves
1086 * so calculate e_max right way */
1087 if (ext_depth(inode))
1088 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
1089 else
1090 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
1091 neh->eh_magic = EXT4_EXT_MAGIC;
1092 set_buffer_uptodate(bh);
1093 unlock_buffer(bh);
1094
1095 err = ext4_handle_dirty_metadata(handle, inode, bh);
1096 if (err)
1097 goto out;
1098
1099 /* Update top-level index: num,max,pointer */
1100 neh = ext_inode_hdr(inode);
1101 neh->eh_entries = cpu_to_le16(1);
1102 ext4_idx_store_pblock(EXT_FIRST_INDEX(neh), newblock);
1103 if (neh->eh_depth == 0) {
1104 /* Root extent block becomes index block */
1105 neh->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode, 0));
1106 EXT_FIRST_INDEX(neh)->ei_block =
1107 EXT_FIRST_EXTENT(neh)->ee_block;
1108 }
1109 ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
1110 le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
1111 le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block),
1112 ext4_idx_pblock(EXT_FIRST_INDEX(neh)));
1113
1114 neh->eh_depth = cpu_to_le16(neh->eh_depth + 1);
1115 ext4_mark_inode_dirty(handle, inode);
1116out:
1117 brelse(bh);
1118
1119 return err;
1120}
1121
1122/*
1123 * ext4_ext_create_new_leaf:
1124 * finds empty index and adds new leaf.
1125 * if no free index is found, then it requests in-depth growing.
1126 */
1127static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
1128 unsigned int flags,
1129 struct ext4_ext_path *path,
1130 struct ext4_extent *newext)
1131{
1132 struct ext4_ext_path *curp;
1133 int depth, i, err = 0;
1134
1135repeat:
1136 i = depth = ext_depth(inode);
1137
1138 /* walk up to the tree and look for free index entry */
1139 curp = path + depth;
1140 while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
1141 i--;
1142 curp--;
1143 }
1144
1145 /* we use already allocated block for index block,
1146 * so subsequent data blocks should be contiguous */
1147 if (EXT_HAS_FREE_INDEX(curp)) {
1148 /* if we found index with free entry, then use that
1149 * entry: create all needed subtree and add new leaf */
1150 err = ext4_ext_split(handle, inode, flags, path, newext, i);
1151 if (err)
1152 goto out;
1153
1154 /* refill path */
1155 ext4_ext_drop_refs(path);
1156 path = ext4_ext_find_extent(inode,
1157 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1158 path);
1159 if (IS_ERR(path))
1160 err = PTR_ERR(path);
1161 } else {
1162 /* tree is full, time to grow in depth */
1163 err = ext4_ext_grow_indepth(handle, inode, flags, newext);
1164 if (err)
1165 goto out;
1166
1167 /* refill path */
1168 ext4_ext_drop_refs(path);
1169 path = ext4_ext_find_extent(inode,
1170 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
1171 path);
1172 if (IS_ERR(path)) {
1173 err = PTR_ERR(path);
1174 goto out;
1175 }
1176
1177 /*
1178 * only first (depth 0 -> 1) produces free space;
1179 * in all other cases we have to split the grown tree
1180 */
1181 depth = ext_depth(inode);
1182 if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
1183 /* now we need to split */
1184 goto repeat;
1185 }
1186 }
1187
1188out:
1189 return err;
1190}
1191
1192/*
1193 * search the closest allocated block to the left for *logical
1194 * and returns it at @logical + it's physical address at @phys
1195 * if *logical is the smallest allocated block, the function
1196 * returns 0 at @phys
1197 * return value contains 0 (success) or error code
1198 */
1199static int ext4_ext_search_left(struct inode *inode,
1200 struct ext4_ext_path *path,
1201 ext4_lblk_t *logical, ext4_fsblk_t *phys)
1202{
1203 struct ext4_extent_idx *ix;
1204 struct ext4_extent *ex;
1205 int depth, ee_len;
1206
1207 if (unlikely(path == NULL)) {
1208 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1209 return -EIO;
1210 }
1211 depth = path->p_depth;
1212 *phys = 0;
1213
1214 if (depth == 0 && path->p_ext == NULL)
1215 return 0;
1216
1217 /* usually extent in the path covers blocks smaller
1218 * then *logical, but it can be that extent is the
1219 * first one in the file */
1220
1221 ex = path[depth].p_ext;
1222 ee_len = ext4_ext_get_actual_len(ex);
1223 if (*logical < le32_to_cpu(ex->ee_block)) {
1224 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1225 EXT4_ERROR_INODE(inode,
1226 "EXT_FIRST_EXTENT != ex *logical %d ee_block %d!",
1227 *logical, le32_to_cpu(ex->ee_block));
1228 return -EIO;
1229 }
1230 while (--depth >= 0) {
1231 ix = path[depth].p_idx;
1232 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1233 EXT4_ERROR_INODE(inode,
1234 "ix (%d) != EXT_FIRST_INDEX (%d) (depth %d)!",
1235 ix != NULL ? le32_to_cpu(ix->ei_block) : 0,
1236 EXT_FIRST_INDEX(path[depth].p_hdr) != NULL ?
1237 le32_to_cpu(EXT_FIRST_INDEX(path[depth].p_hdr)->ei_block) : 0,
1238 depth);
1239 return -EIO;
1240 }
1241 }
1242 return 0;
1243 }
1244
1245 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1246 EXT4_ERROR_INODE(inode,
1247 "logical %d < ee_block %d + ee_len %d!",
1248 *logical, le32_to_cpu(ex->ee_block), ee_len);
1249 return -EIO;
1250 }
1251
1252 *logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
1253 *phys = ext4_ext_pblock(ex) + ee_len - 1;
1254 return 0;
1255}
1256
1257/*
1258 * search the closest allocated block to the right for *logical
1259 * and returns it at @logical + it's physical address at @phys
1260 * if *logical is the largest allocated block, the function
1261 * returns 0 at @phys
1262 * return value contains 0 (success) or error code
1263 */
1264static int ext4_ext_search_right(struct inode *inode,
1265 struct ext4_ext_path *path,
1266 ext4_lblk_t *logical, ext4_fsblk_t *phys,
1267 struct ext4_extent **ret_ex)
1268{
1269 struct buffer_head *bh = NULL;
1270 struct ext4_extent_header *eh;
1271 struct ext4_extent_idx *ix;
1272 struct ext4_extent *ex;
1273 ext4_fsblk_t block;
1274 int depth; /* Note, NOT eh_depth; depth from top of tree */
1275 int ee_len;
1276
1277 if (unlikely(path == NULL)) {
1278 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1279 return -EIO;
1280 }
1281 depth = path->p_depth;
1282 *phys = 0;
1283
1284 if (depth == 0 && path->p_ext == NULL)
1285 return 0;
1286
1287 /* usually extent in the path covers blocks smaller
1288 * then *logical, but it can be that extent is the
1289 * first one in the file */
1290
1291 ex = path[depth].p_ext;
1292 ee_len = ext4_ext_get_actual_len(ex);
1293 if (*logical < le32_to_cpu(ex->ee_block)) {
1294 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1295 EXT4_ERROR_INODE(inode,
1296 "first_extent(path[%d].p_hdr) != ex",
1297 depth);
1298 return -EIO;
1299 }
1300 while (--depth >= 0) {
1301 ix = path[depth].p_idx;
1302 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1303 EXT4_ERROR_INODE(inode,
1304 "ix != EXT_FIRST_INDEX *logical %d!",
1305 *logical);
1306 return -EIO;
1307 }
1308 }
1309 goto found_extent;
1310 }
1311
1312 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1313 EXT4_ERROR_INODE(inode,
1314 "logical %d < ee_block %d + ee_len %d!",
1315 *logical, le32_to_cpu(ex->ee_block), ee_len);
1316 return -EIO;
1317 }
1318
1319 if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
1320 /* next allocated block in this leaf */
1321 ex++;
1322 goto found_extent;
1323 }
1324
1325 /* go up and search for index to the right */
1326 while (--depth >= 0) {
1327 ix = path[depth].p_idx;
1328 if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
1329 goto got_index;
1330 }
1331
1332 /* we've gone up to the root and found no index to the right */
1333 return 0;
1334
1335got_index:
1336 /* we've found index to the right, let's
1337 * follow it and find the closest allocated
1338 * block to the right */
1339 ix++;
1340 block = ext4_idx_pblock(ix);
1341 while (++depth < path->p_depth) {
1342 bh = sb_bread(inode->i_sb, block);
1343 if (bh == NULL)
1344 return -EIO;
1345 eh = ext_block_hdr(bh);
1346 /* subtract from p_depth to get proper eh_depth */
1347 if (ext4_ext_check(inode, eh, path->p_depth - depth)) {
1348 put_bh(bh);
1349 return -EIO;
1350 }
1351 ix = EXT_FIRST_INDEX(eh);
1352 block = ext4_idx_pblock(ix);
1353 put_bh(bh);
1354 }
1355
1356 bh = sb_bread(inode->i_sb, block);
1357 if (bh == NULL)
1358 return -EIO;
1359 eh = ext_block_hdr(bh);
1360 if (ext4_ext_check(inode, eh, path->p_depth - depth)) {
1361 put_bh(bh);
1362 return -EIO;
1363 }
1364 ex = EXT_FIRST_EXTENT(eh);
1365found_extent:
1366 *logical = le32_to_cpu(ex->ee_block);
1367 *phys = ext4_ext_pblock(ex);
1368 *ret_ex = ex;
1369 if (bh)
1370 put_bh(bh);
1371 return 0;
1372}
1373
1374/*
1375 * ext4_ext_next_allocated_block:
1376 * returns allocated block in subsequent extent or EXT_MAX_BLOCKS.
1377 * NOTE: it considers block number from index entry as
1378 * allocated block. Thus, index entries have to be consistent
1379 * with leaves.
1380 */
1381static ext4_lblk_t
1382ext4_ext_next_allocated_block(struct ext4_ext_path *path)
1383{
1384 int depth;
1385
1386 BUG_ON(path == NULL);
1387 depth = path->p_depth;
1388
1389 if (depth == 0 && path->p_ext == NULL)
1390 return EXT_MAX_BLOCKS;
1391
1392 while (depth >= 0) {
1393 if (depth == path->p_depth) {
1394 /* leaf */
1395 if (path[depth].p_ext !=
1396 EXT_LAST_EXTENT(path[depth].p_hdr))
1397 return le32_to_cpu(path[depth].p_ext[1].ee_block);
1398 } else {
1399 /* index */
1400 if (path[depth].p_idx !=
1401 EXT_LAST_INDEX(path[depth].p_hdr))
1402 return le32_to_cpu(path[depth].p_idx[1].ei_block);
1403 }
1404 depth--;
1405 }
1406
1407 return EXT_MAX_BLOCKS;
1408}
1409
1410/*
1411 * ext4_ext_next_leaf_block:
1412 * returns first allocated block from next leaf or EXT_MAX_BLOCKS
1413 */
1414static ext4_lblk_t ext4_ext_next_leaf_block(struct ext4_ext_path *path)
1415{
1416 int depth;
1417
1418 BUG_ON(path == NULL);
1419 depth = path->p_depth;
1420
1421 /* zero-tree has no leaf blocks at all */
1422 if (depth == 0)
1423 return EXT_MAX_BLOCKS;
1424
1425 /* go to index block */
1426 depth--;
1427
1428 while (depth >= 0) {
1429 if (path[depth].p_idx !=
1430 EXT_LAST_INDEX(path[depth].p_hdr))
1431 return (ext4_lblk_t)
1432 le32_to_cpu(path[depth].p_idx[1].ei_block);
1433 depth--;
1434 }
1435
1436 return EXT_MAX_BLOCKS;
1437}
1438
1439/*
1440 * ext4_ext_correct_indexes:
1441 * if leaf gets modified and modified extent is first in the leaf,
1442 * then we have to correct all indexes above.
1443 * TODO: do we need to correct tree in all cases?
1444 */
1445static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
1446 struct ext4_ext_path *path)
1447{
1448 struct ext4_extent_header *eh;
1449 int depth = ext_depth(inode);
1450 struct ext4_extent *ex;
1451 __le32 border;
1452 int k, err = 0;
1453
1454 eh = path[depth].p_hdr;
1455 ex = path[depth].p_ext;
1456
1457 if (unlikely(ex == NULL || eh == NULL)) {
1458 EXT4_ERROR_INODE(inode,
1459 "ex %p == NULL or eh %p == NULL", ex, eh);
1460 return -EIO;
1461 }
1462
1463 if (depth == 0) {
1464 /* there is no tree at all */
1465 return 0;
1466 }
1467
1468 if (ex != EXT_FIRST_EXTENT(eh)) {
1469 /* we correct tree if first leaf got modified only */
1470 return 0;
1471 }
1472
1473 /*
1474 * TODO: we need correction if border is smaller than current one
1475 */
1476 k = depth - 1;
1477 border = path[depth].p_ext->ee_block;
1478 err = ext4_ext_get_access(handle, inode, path + k);
1479 if (err)
1480 return err;
1481 path[k].p_idx->ei_block = border;
1482 err = ext4_ext_dirty(handle, inode, path + k);
1483 if (err)
1484 return err;
1485
1486 while (k--) {
1487 /* change all left-side indexes */
1488 if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
1489 break;
1490 err = ext4_ext_get_access(handle, inode, path + k);
1491 if (err)
1492 break;
1493 path[k].p_idx->ei_block = border;
1494 err = ext4_ext_dirty(handle, inode, path + k);
1495 if (err)
1496 break;
1497 }
1498
1499 return err;
1500}
1501
1502int
1503ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
1504 struct ext4_extent *ex2)
1505{
1506 unsigned short ext1_ee_len, ext2_ee_len, max_len;
1507
1508 /*
1509 * Make sure that either both extents are uninitialized, or
1510 * both are _not_.
1511 */
1512 if (ext4_ext_is_uninitialized(ex1) ^ ext4_ext_is_uninitialized(ex2))
1513 return 0;
1514
1515 if (ext4_ext_is_uninitialized(ex1))
1516 max_len = EXT_UNINIT_MAX_LEN;
1517 else
1518 max_len = EXT_INIT_MAX_LEN;
1519
1520 ext1_ee_len = ext4_ext_get_actual_len(ex1);
1521 ext2_ee_len = ext4_ext_get_actual_len(ex2);
1522
1523 if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
1524 le32_to_cpu(ex2->ee_block))
1525 return 0;
1526
1527 /*
1528 * To allow future support for preallocated extents to be added
1529 * as an RO_COMPAT feature, refuse to merge to extents if
1530 * this can result in the top bit of ee_len being set.
1531 */
1532 if (ext1_ee_len + ext2_ee_len > max_len)
1533 return 0;
1534#ifdef AGGRESSIVE_TEST
1535 if (ext1_ee_len >= 4)
1536 return 0;
1537#endif
1538
1539 if (ext4_ext_pblock(ex1) + ext1_ee_len == ext4_ext_pblock(ex2))
1540 return 1;
1541 return 0;
1542}
1543
1544/*
1545 * This function tries to merge the "ex" extent to the next extent in the tree.
1546 * It always tries to merge towards right. If you want to merge towards
1547 * left, pass "ex - 1" as argument instead of "ex".
1548 * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
1549 * 1 if they got merged.
1550 */
1551static int ext4_ext_try_to_merge_right(struct inode *inode,
1552 struct ext4_ext_path *path,
1553 struct ext4_extent *ex)
1554{
1555 struct ext4_extent_header *eh;
1556 unsigned int depth, len;
1557 int merge_done = 0;
1558 int uninitialized = 0;
1559
1560 depth = ext_depth(inode);
1561 BUG_ON(path[depth].p_hdr == NULL);
1562 eh = path[depth].p_hdr;
1563
1564 while (ex < EXT_LAST_EXTENT(eh)) {
1565 if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
1566 break;
1567 /* merge with next extent! */
1568 if (ext4_ext_is_uninitialized(ex))
1569 uninitialized = 1;
1570 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1571 + ext4_ext_get_actual_len(ex + 1));
1572 if (uninitialized)
1573 ext4_ext_mark_uninitialized(ex);
1574
1575 if (ex + 1 < EXT_LAST_EXTENT(eh)) {
1576 len = (EXT_LAST_EXTENT(eh) - ex - 1)
1577 * sizeof(struct ext4_extent);
1578 memmove(ex + 1, ex + 2, len);
1579 }
1580 le16_add_cpu(&eh->eh_entries, -1);
1581 merge_done = 1;
1582 WARN_ON(eh->eh_entries == 0);
1583 if (!eh->eh_entries)
1584 EXT4_ERROR_INODE(inode, "eh->eh_entries = 0!");
1585 }
1586
1587 return merge_done;
1588}
1589
1590/*
1591 * This function tries to merge the @ex extent to neighbours in the tree.
1592 * return 1 if merge left else 0.
1593 */
1594static int ext4_ext_try_to_merge(struct inode *inode,
1595 struct ext4_ext_path *path,
1596 struct ext4_extent *ex) {
1597 struct ext4_extent_header *eh;
1598 unsigned int depth;
1599 int merge_done = 0;
1600 int ret = 0;
1601
1602 depth = ext_depth(inode);
1603 BUG_ON(path[depth].p_hdr == NULL);
1604 eh = path[depth].p_hdr;
1605
1606 if (ex > EXT_FIRST_EXTENT(eh))
1607 merge_done = ext4_ext_try_to_merge_right(inode, path, ex - 1);
1608
1609 if (!merge_done)
1610 ret = ext4_ext_try_to_merge_right(inode, path, ex);
1611
1612 return ret;
1613}
1614
1615/*
1616 * check if a portion of the "newext" extent overlaps with an
1617 * existing extent.
1618 *
1619 * If there is an overlap discovered, it updates the length of the newext
1620 * such that there will be no overlap, and then returns 1.
1621 * If there is no overlap found, it returns 0.
1622 */
1623static unsigned int ext4_ext_check_overlap(struct ext4_sb_info *sbi,
1624 struct inode *inode,
1625 struct ext4_extent *newext,
1626 struct ext4_ext_path *path)
1627{
1628 ext4_lblk_t b1, b2;
1629 unsigned int depth, len1;
1630 unsigned int ret = 0;
1631
1632 b1 = le32_to_cpu(newext->ee_block);
1633 len1 = ext4_ext_get_actual_len(newext);
1634 depth = ext_depth(inode);
1635 if (!path[depth].p_ext)
1636 goto out;
1637 b2 = le32_to_cpu(path[depth].p_ext->ee_block);
1638 b2 &= ~(sbi->s_cluster_ratio - 1);
1639
1640 /*
1641 * get the next allocated block if the extent in the path
1642 * is before the requested block(s)
1643 */
1644 if (b2 < b1) {
1645 b2 = ext4_ext_next_allocated_block(path);
1646 if (b2 == EXT_MAX_BLOCKS)
1647 goto out;
1648 b2 &= ~(sbi->s_cluster_ratio - 1);
1649 }
1650
1651 /* check for wrap through zero on extent logical start block*/
1652 if (b1 + len1 < b1) {
1653 len1 = EXT_MAX_BLOCKS - b1;
1654 newext->ee_len = cpu_to_le16(len1);
1655 ret = 1;
1656 }
1657
1658 /* check for overlap */
1659 if (b1 + len1 > b2) {
1660 newext->ee_len = cpu_to_le16(b2 - b1);
1661 ret = 1;
1662 }
1663out:
1664 return ret;
1665}
1666
1667/*
1668 * ext4_ext_insert_extent:
1669 * tries to merge requsted extent into the existing extent or
1670 * inserts requested extent as new one into the tree,
1671 * creating new leaf in the no-space case.
1672 */
1673int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
1674 struct ext4_ext_path *path,
1675 struct ext4_extent *newext, int flag)
1676{
1677 struct ext4_extent_header *eh;
1678 struct ext4_extent *ex, *fex;
1679 struct ext4_extent *nearex; /* nearest extent */
1680 struct ext4_ext_path *npath = NULL;
1681 int depth, len, err;
1682 ext4_lblk_t next;
1683 unsigned uninitialized = 0;
1684 int flags = 0;
1685
1686 if (unlikely(ext4_ext_get_actual_len(newext) == 0)) {
1687 EXT4_ERROR_INODE(inode, "ext4_ext_get_actual_len(newext) == 0");
1688 return -EIO;
1689 }
1690 depth = ext_depth(inode);
1691 ex = path[depth].p_ext;
1692 if (unlikely(path[depth].p_hdr == NULL)) {
1693 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
1694 return -EIO;
1695 }
1696
1697 /* try to insert block into found extent and return */
1698 if (ex && !(flag & EXT4_GET_BLOCKS_PRE_IO)
1699 && ext4_can_extents_be_merged(inode, ex, newext)) {
1700 ext_debug("append [%d]%d block to %d:[%d]%d (from %llu)\n",
1701 ext4_ext_is_uninitialized(newext),
1702 ext4_ext_get_actual_len(newext),
1703 le32_to_cpu(ex->ee_block),
1704 ext4_ext_is_uninitialized(ex),
1705 ext4_ext_get_actual_len(ex),
1706 ext4_ext_pblock(ex));
1707 err = ext4_ext_get_access(handle, inode, path + depth);
1708 if (err)
1709 return err;
1710
1711 /*
1712 * ext4_can_extents_be_merged should have checked that either
1713 * both extents are uninitialized, or both aren't. Thus we
1714 * need to check only one of them here.
1715 */
1716 if (ext4_ext_is_uninitialized(ex))
1717 uninitialized = 1;
1718 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1719 + ext4_ext_get_actual_len(newext));
1720 if (uninitialized)
1721 ext4_ext_mark_uninitialized(ex);
1722 eh = path[depth].p_hdr;
1723 nearex = ex;
1724 goto merge;
1725 }
1726
1727 depth = ext_depth(inode);
1728 eh = path[depth].p_hdr;
1729 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
1730 goto has_space;
1731
1732 /* probably next leaf has space for us? */
1733 fex = EXT_LAST_EXTENT(eh);
1734 next = EXT_MAX_BLOCKS;
1735 if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
1736 next = ext4_ext_next_leaf_block(path);
1737 if (next != EXT_MAX_BLOCKS) {
1738 ext_debug("next leaf block - %d\n", next);
1739 BUG_ON(npath != NULL);
1740 npath = ext4_ext_find_extent(inode, next, NULL);
1741 if (IS_ERR(npath))
1742 return PTR_ERR(npath);
1743 BUG_ON(npath->p_depth != path->p_depth);
1744 eh = npath[depth].p_hdr;
1745 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
1746 ext_debug("next leaf isn't full(%d)\n",
1747 le16_to_cpu(eh->eh_entries));
1748 path = npath;
1749 goto has_space;
1750 }
1751 ext_debug("next leaf has no free space(%d,%d)\n",
1752 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
1753 }
1754
1755 /*
1756 * There is no free space in the found leaf.
1757 * We're gonna add a new leaf in the tree.
1758 */
1759 if (flag & EXT4_GET_BLOCKS_PUNCH_OUT_EXT)
1760 flags = EXT4_MB_USE_ROOT_BLOCKS;
1761 err = ext4_ext_create_new_leaf(handle, inode, flags, path, newext);
1762 if (err)
1763 goto cleanup;
1764 depth = ext_depth(inode);
1765 eh = path[depth].p_hdr;
1766
1767has_space:
1768 nearex = path[depth].p_ext;
1769
1770 err = ext4_ext_get_access(handle, inode, path + depth);
1771 if (err)
1772 goto cleanup;
1773
1774 if (!nearex) {
1775 /* there is no extent in this leaf, create first one */
1776 ext_debug("first extent in the leaf: %d:%llu:[%d]%d\n",
1777 le32_to_cpu(newext->ee_block),
1778 ext4_ext_pblock(newext),
1779 ext4_ext_is_uninitialized(newext),
1780 ext4_ext_get_actual_len(newext));
1781 path[depth].p_ext = EXT_FIRST_EXTENT(eh);
1782 } else if (le32_to_cpu(newext->ee_block)
1783 > le32_to_cpu(nearex->ee_block)) {
1784/* BUG_ON(newext->ee_block == nearex->ee_block); */
1785 if (nearex != EXT_LAST_EXTENT(eh)) {
1786 len = EXT_MAX_EXTENT(eh) - nearex;
1787 len = (len - 1) * sizeof(struct ext4_extent);
1788 len = len < 0 ? 0 : len;
1789 ext_debug("insert %d:%llu:[%d]%d after: nearest 0x%p, "
1790 "move %d from 0x%p to 0x%p\n",
1791 le32_to_cpu(newext->ee_block),
1792 ext4_ext_pblock(newext),
1793 ext4_ext_is_uninitialized(newext),
1794 ext4_ext_get_actual_len(newext),
1795 nearex, len, nearex + 1, nearex + 2);
1796 memmove(nearex + 2, nearex + 1, len);
1797 }
1798 path[depth].p_ext = nearex + 1;
1799 } else {
1800 BUG_ON(newext->ee_block == nearex->ee_block);
1801 len = (EXT_MAX_EXTENT(eh) - nearex) * sizeof(struct ext4_extent);
1802 len = len < 0 ? 0 : len;
1803 ext_debug("insert %d:%llu:[%d]%d before: nearest 0x%p, "
1804 "move %d from 0x%p to 0x%p\n",
1805 le32_to_cpu(newext->ee_block),
1806 ext4_ext_pblock(newext),
1807 ext4_ext_is_uninitialized(newext),
1808 ext4_ext_get_actual_len(newext),
1809 nearex, len, nearex, nearex + 1);
1810 memmove(nearex + 1, nearex, len);
1811 path[depth].p_ext = nearex;
1812 }
1813
1814 le16_add_cpu(&eh->eh_entries, 1);
1815 nearex = path[depth].p_ext;
1816 nearex->ee_block = newext->ee_block;
1817 ext4_ext_store_pblock(nearex, ext4_ext_pblock(newext));
1818 nearex->ee_len = newext->ee_len;
1819
1820merge:
1821 /* try to merge extents to the right */
1822 if (!(flag & EXT4_GET_BLOCKS_PRE_IO))
1823 ext4_ext_try_to_merge(inode, path, nearex);
1824
1825 /* try to merge extents to the left */
1826
1827 /* time to correct all indexes above */
1828 err = ext4_ext_correct_indexes(handle, inode, path);
1829 if (err)
1830 goto cleanup;
1831
1832 err = ext4_ext_dirty(handle, inode, path + depth);
1833
1834cleanup:
1835 if (npath) {
1836 ext4_ext_drop_refs(npath);
1837 kfree(npath);
1838 }
1839 ext4_ext_invalidate_cache(inode);
1840 return err;
1841}
1842
1843static int ext4_ext_walk_space(struct inode *inode, ext4_lblk_t block,
1844 ext4_lblk_t num, ext_prepare_callback func,
1845 void *cbdata)
1846{
1847 struct ext4_ext_path *path = NULL;
1848 struct ext4_ext_cache cbex;
1849 struct ext4_extent *ex;
1850 ext4_lblk_t next, start = 0, end = 0;
1851 ext4_lblk_t last = block + num;
1852 int depth, exists, err = 0;
1853
1854 BUG_ON(func == NULL);
1855 BUG_ON(inode == NULL);
1856
1857 while (block < last && block != EXT_MAX_BLOCKS) {
1858 num = last - block;
1859 /* find extent for this block */
1860 down_read(&EXT4_I(inode)->i_data_sem);
1861 path = ext4_ext_find_extent(inode, block, path);
1862 up_read(&EXT4_I(inode)->i_data_sem);
1863 if (IS_ERR(path)) {
1864 err = PTR_ERR(path);
1865 path = NULL;
1866 break;
1867 }
1868
1869 depth = ext_depth(inode);
1870 if (unlikely(path[depth].p_hdr == NULL)) {
1871 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
1872 err = -EIO;
1873 break;
1874 }
1875 ex = path[depth].p_ext;
1876 next = ext4_ext_next_allocated_block(path);
1877
1878 exists = 0;
1879 if (!ex) {
1880 /* there is no extent yet, so try to allocate
1881 * all requested space */
1882 start = block;
1883 end = block + num;
1884 } else if (le32_to_cpu(ex->ee_block) > block) {
1885 /* need to allocate space before found extent */
1886 start = block;
1887 end = le32_to_cpu(ex->ee_block);
1888 if (block + num < end)
1889 end = block + num;
1890 } else if (block >= le32_to_cpu(ex->ee_block)
1891 + ext4_ext_get_actual_len(ex)) {
1892 /* need to allocate space after found extent */
1893 start = block;
1894 end = block + num;
1895 if (end >= next)
1896 end = next;
1897 } else if (block >= le32_to_cpu(ex->ee_block)) {
1898 /*
1899 * some part of requested space is covered
1900 * by found extent
1901 */
1902 start = block;
1903 end = le32_to_cpu(ex->ee_block)
1904 + ext4_ext_get_actual_len(ex);
1905 if (block + num < end)
1906 end = block + num;
1907 exists = 1;
1908 } else {
1909 BUG();
1910 }
1911 BUG_ON(end <= start);
1912
1913 if (!exists) {
1914 cbex.ec_block = start;
1915 cbex.ec_len = end - start;
1916 cbex.ec_start = 0;
1917 } else {
1918 cbex.ec_block = le32_to_cpu(ex->ee_block);
1919 cbex.ec_len = ext4_ext_get_actual_len(ex);
1920 cbex.ec_start = ext4_ext_pblock(ex);
1921 }
1922
1923 if (unlikely(cbex.ec_len == 0)) {
1924 EXT4_ERROR_INODE(inode, "cbex.ec_len == 0");
1925 err = -EIO;
1926 break;
1927 }
1928 err = func(inode, next, &cbex, ex, cbdata);
1929 ext4_ext_drop_refs(path);
1930
1931 if (err < 0)
1932 break;
1933
1934 if (err == EXT_REPEAT)
1935 continue;
1936 else if (err == EXT_BREAK) {
1937 err = 0;
1938 break;
1939 }
1940
1941 if (ext_depth(inode) != depth) {
1942 /* depth was changed. we have to realloc path */
1943 kfree(path);
1944 path = NULL;
1945 }
1946
1947 block = cbex.ec_block + cbex.ec_len;
1948 }
1949
1950 if (path) {
1951 ext4_ext_drop_refs(path);
1952 kfree(path);
1953 }
1954
1955 return err;
1956}
1957
1958static void
1959ext4_ext_put_in_cache(struct inode *inode, ext4_lblk_t block,
1960 __u32 len, ext4_fsblk_t start)
1961{
1962 struct ext4_ext_cache *cex;
1963 BUG_ON(len == 0);
1964 spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
1965 trace_ext4_ext_put_in_cache(inode, block, len, start);
1966 cex = &EXT4_I(inode)->i_cached_extent;
1967 cex->ec_block = block;
1968 cex->ec_len = len;
1969 cex->ec_start = start;
1970 spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
1971}
1972
1973/*
1974 * ext4_ext_put_gap_in_cache:
1975 * calculate boundaries of the gap that the requested block fits into
1976 * and cache this gap
1977 */
1978static void
1979ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
1980 ext4_lblk_t block)
1981{
1982 int depth = ext_depth(inode);
1983 unsigned long len;
1984 ext4_lblk_t lblock;
1985 struct ext4_extent *ex;
1986
1987 ex = path[depth].p_ext;
1988 if (ex == NULL) {
1989 /* there is no extent yet, so gap is [0;-] */
1990 lblock = 0;
1991 len = EXT_MAX_BLOCKS;
1992 ext_debug("cache gap(whole file):");
1993 } else if (block < le32_to_cpu(ex->ee_block)) {
1994 lblock = block;
1995 len = le32_to_cpu(ex->ee_block) - block;
1996 ext_debug("cache gap(before): %u [%u:%u]",
1997 block,
1998 le32_to_cpu(ex->ee_block),
1999 ext4_ext_get_actual_len(ex));
2000 } else if (block >= le32_to_cpu(ex->ee_block)
2001 + ext4_ext_get_actual_len(ex)) {
2002 ext4_lblk_t next;
2003 lblock = le32_to_cpu(ex->ee_block)
2004 + ext4_ext_get_actual_len(ex);
2005
2006 next = ext4_ext_next_allocated_block(path);
2007 ext_debug("cache gap(after): [%u:%u] %u",
2008 le32_to_cpu(ex->ee_block),
2009 ext4_ext_get_actual_len(ex),
2010 block);
2011 BUG_ON(next == lblock);
2012 len = next - lblock;
2013 } else {
2014 lblock = len = 0;
2015 BUG();
2016 }
2017
2018 ext_debug(" -> %u:%lu\n", lblock, len);
2019 ext4_ext_put_in_cache(inode, lblock, len, 0);
2020}
2021
2022/*
2023 * ext4_ext_check_cache()
2024 * Checks to see if the given block is in the cache.
2025 * If it is, the cached extent is stored in the given
2026 * cache extent pointer. If the cached extent is a hole,
2027 * this routine should be used instead of
2028 * ext4_ext_in_cache if the calling function needs to
2029 * know the size of the hole.
2030 *
2031 * @inode: The files inode
2032 * @block: The block to look for in the cache
2033 * @ex: Pointer where the cached extent will be stored
2034 * if it contains block
2035 *
2036 * Return 0 if cache is invalid; 1 if the cache is valid
2037 */
2038static int ext4_ext_check_cache(struct inode *inode, ext4_lblk_t block,
2039 struct ext4_ext_cache *ex){
2040 struct ext4_ext_cache *cex;
2041 struct ext4_sb_info *sbi;
2042 int ret = 0;
2043
2044 /*
2045 * We borrow i_block_reservation_lock to protect i_cached_extent
2046 */
2047 spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
2048 cex = &EXT4_I(inode)->i_cached_extent;
2049 sbi = EXT4_SB(inode->i_sb);
2050
2051 /* has cache valid data? */
2052 if (cex->ec_len == 0)
2053 goto errout;
2054
2055 if (in_range(block, cex->ec_block, cex->ec_len)) {
2056 memcpy(ex, cex, sizeof(struct ext4_ext_cache));
2057 ext_debug("%u cached by %u:%u:%llu\n",
2058 block,
2059 cex->ec_block, cex->ec_len, cex->ec_start);
2060 ret = 1;
2061 }
2062errout:
2063 if (!ret)
2064 sbi->extent_cache_misses++;
2065 else
2066 sbi->extent_cache_hits++;
2067 trace_ext4_ext_in_cache(inode, block, ret);
2068 spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
2069 return ret;
2070}
2071
2072/*
2073 * ext4_ext_in_cache()
2074 * Checks to see if the given block is in the cache.
2075 * If it is, the cached extent is stored in the given
2076 * extent pointer.
2077 *
2078 * @inode: The files inode
2079 * @block: The block to look for in the cache
2080 * @ex: Pointer where the cached extent will be stored
2081 * if it contains block
2082 *
2083 * Return 0 if cache is invalid; 1 if the cache is valid
2084 */
2085static int
2086ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block,
2087 struct ext4_extent *ex)
2088{
2089 struct ext4_ext_cache cex;
2090 int ret = 0;
2091
2092 if (ext4_ext_check_cache(inode, block, &cex)) {
2093 ex->ee_block = cpu_to_le32(cex.ec_block);
2094 ext4_ext_store_pblock(ex, cex.ec_start);
2095 ex->ee_len = cpu_to_le16(cex.ec_len);
2096 ret = 1;
2097 }
2098
2099 return ret;
2100}
2101
2102
2103/*
2104 * ext4_ext_rm_idx:
2105 * removes index from the index block.
2106 */
2107static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
2108 struct ext4_ext_path *path)
2109{
2110 int err;
2111 ext4_fsblk_t leaf;
2112
2113 /* free index block */
2114 path--;
2115 leaf = ext4_idx_pblock(path->p_idx);
2116 if (unlikely(path->p_hdr->eh_entries == 0)) {
2117 EXT4_ERROR_INODE(inode, "path->p_hdr->eh_entries == 0");
2118 return -EIO;
2119 }
2120 err = ext4_ext_get_access(handle, inode, path);
2121 if (err)
2122 return err;
2123
2124 if (path->p_idx != EXT_LAST_INDEX(path->p_hdr)) {
2125 int len = EXT_LAST_INDEX(path->p_hdr) - path->p_idx;
2126 len *= sizeof(struct ext4_extent_idx);
2127 memmove(path->p_idx, path->p_idx + 1, len);
2128 }
2129
2130 le16_add_cpu(&path->p_hdr->eh_entries, -1);
2131 err = ext4_ext_dirty(handle, inode, path);
2132 if (err)
2133 return err;
2134 ext_debug("index is empty, remove it, free block %llu\n", leaf);
2135 trace_ext4_ext_rm_idx(inode, leaf);
2136
2137 ext4_free_blocks(handle, inode, NULL, leaf, 1,
2138 EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
2139 return err;
2140}
2141
2142/*
2143 * ext4_ext_calc_credits_for_single_extent:
2144 * This routine returns max. credits that needed to insert an extent
2145 * to the extent tree.
2146 * When pass the actual path, the caller should calculate credits
2147 * under i_data_sem.
2148 */
2149int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
2150 struct ext4_ext_path *path)
2151{
2152 if (path) {
2153 int depth = ext_depth(inode);
2154 int ret = 0;
2155
2156 /* probably there is space in leaf? */
2157 if (le16_to_cpu(path[depth].p_hdr->eh_entries)
2158 < le16_to_cpu(path[depth].p_hdr->eh_max)) {
2159
2160 /*
2161 * There are some space in the leaf tree, no
2162 * need to account for leaf block credit
2163 *
2164 * bitmaps and block group descriptor blocks
2165 * and other metadata blocks still need to be
2166 * accounted.
2167 */
2168 /* 1 bitmap, 1 block group descriptor */
2169 ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
2170 return ret;
2171 }
2172 }
2173
2174 return ext4_chunk_trans_blocks(inode, nrblocks);
2175}
2176
2177/*
2178 * How many index/leaf blocks need to change/allocate to modify nrblocks?
2179 *
2180 * if nrblocks are fit in a single extent (chunk flag is 1), then
2181 * in the worse case, each tree level index/leaf need to be changed
2182 * if the tree split due to insert a new extent, then the old tree
2183 * index/leaf need to be updated too
2184 *
2185 * If the nrblocks are discontiguous, they could cause
2186 * the whole tree split more than once, but this is really rare.
2187 */
2188int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
2189{
2190 int index;
2191 int depth = ext_depth(inode);
2192
2193 if (chunk)
2194 index = depth * 2;
2195 else
2196 index = depth * 3;
2197
2198 return index;
2199}
2200
2201static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
2202 struct ext4_extent *ex,
2203 ext4_fsblk_t *partial_cluster,
2204 ext4_lblk_t from, ext4_lblk_t to)
2205{
2206 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2207 unsigned short ee_len = ext4_ext_get_actual_len(ex);
2208 ext4_fsblk_t pblk;
2209 int flags = EXT4_FREE_BLOCKS_FORGET;
2210
2211 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2212 flags |= EXT4_FREE_BLOCKS_METADATA;
2213 /*
2214 * For bigalloc file systems, we never free a partial cluster
2215 * at the beginning of the extent. Instead, we make a note
2216 * that we tried freeing the cluster, and check to see if we
2217 * need to free it on a subsequent call to ext4_remove_blocks,
2218 * or at the end of the ext4_truncate() operation.
2219 */
2220 flags |= EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER;
2221
2222 trace_ext4_remove_blocks(inode, ex, from, to, *partial_cluster);
2223 /*
2224 * If we have a partial cluster, and it's different from the
2225 * cluster of the last block, we need to explicitly free the
2226 * partial cluster here.
2227 */
2228 pblk = ext4_ext_pblock(ex) + ee_len - 1;
2229 if (*partial_cluster && (EXT4_B2C(sbi, pblk) != *partial_cluster)) {
2230 ext4_free_blocks(handle, inode, NULL,
2231 EXT4_C2B(sbi, *partial_cluster),
2232 sbi->s_cluster_ratio, flags);
2233 *partial_cluster = 0;
2234 }
2235
2236#ifdef EXTENTS_STATS
2237 {
2238 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2239 spin_lock(&sbi->s_ext_stats_lock);
2240 sbi->s_ext_blocks += ee_len;
2241 sbi->s_ext_extents++;
2242 if (ee_len < sbi->s_ext_min)
2243 sbi->s_ext_min = ee_len;
2244 if (ee_len > sbi->s_ext_max)
2245 sbi->s_ext_max = ee_len;
2246 if (ext_depth(inode) > sbi->s_depth_max)
2247 sbi->s_depth_max = ext_depth(inode);
2248 spin_unlock(&sbi->s_ext_stats_lock);
2249 }
2250#endif
2251 if (from >= le32_to_cpu(ex->ee_block)
2252 && to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
2253 /* tail removal */
2254 ext4_lblk_t num;
2255
2256 num = le32_to_cpu(ex->ee_block) + ee_len - from;
2257 pblk = ext4_ext_pblock(ex) + ee_len - num;
2258 ext_debug("free last %u blocks starting %llu\n", num, pblk);
2259 ext4_free_blocks(handle, inode, NULL, pblk, num, flags);
2260 /*
2261 * If the block range to be freed didn't start at the
2262 * beginning of a cluster, and we removed the entire
2263 * extent, save the partial cluster here, since we
2264 * might need to delete if we determine that the
2265 * truncate operation has removed all of the blocks in
2266 * the cluster.
2267 */
2268 if (pblk & (sbi->s_cluster_ratio - 1) &&
2269 (ee_len == num))
2270 *partial_cluster = EXT4_B2C(sbi, pblk);
2271 else
2272 *partial_cluster = 0;
2273 } else if (from == le32_to_cpu(ex->ee_block)
2274 && to <= le32_to_cpu(ex->ee_block) + ee_len - 1) {
2275 /* head removal */
2276 ext4_lblk_t num;
2277 ext4_fsblk_t start;
2278
2279 num = to - from;
2280 start = ext4_ext_pblock(ex);
2281
2282 ext_debug("free first %u blocks starting %llu\n", num, start);
2283 ext4_free_blocks(handle, inode, NULL, start, num, flags);
2284
2285 } else {
2286 printk(KERN_INFO "strange request: removal(2) "
2287 "%u-%u from %u:%u\n",
2288 from, to, le32_to_cpu(ex->ee_block), ee_len);
2289 }
2290 return 0;
2291}
2292
2293
2294/*
2295 * ext4_ext_rm_leaf() Removes the extents associated with the
2296 * blocks appearing between "start" and "end", and splits the extents
2297 * if "start" and "end" appear in the same extent
2298 *
2299 * @handle: The journal handle
2300 * @inode: The files inode
2301 * @path: The path to the leaf
2302 * @start: The first block to remove
2303 * @end: The last block to remove
2304 */
2305static int
2306ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
2307 struct ext4_ext_path *path, ext4_fsblk_t *partial_cluster,
2308 ext4_lblk_t start, ext4_lblk_t end)
2309{
2310 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
2311 int err = 0, correct_index = 0;
2312 int depth = ext_depth(inode), credits;
2313 struct ext4_extent_header *eh;
2314 ext4_lblk_t a, b, block;
2315 unsigned num;
2316 ext4_lblk_t ex_ee_block;
2317 unsigned short ex_ee_len;
2318 unsigned uninitialized = 0;
2319 struct ext4_extent *ex;
2320 struct ext4_map_blocks map;
2321
2322 /* the header must be checked already in ext4_ext_remove_space() */
2323 ext_debug("truncate since %u in leaf\n", start);
2324 if (!path[depth].p_hdr)
2325 path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
2326 eh = path[depth].p_hdr;
2327 if (unlikely(path[depth].p_hdr == NULL)) {
2328 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
2329 return -EIO;
2330 }
2331 /* find where to start removing */
2332 ex = EXT_LAST_EXTENT(eh);
2333
2334 ex_ee_block = le32_to_cpu(ex->ee_block);
2335 ex_ee_len = ext4_ext_get_actual_len(ex);
2336
2337 trace_ext4_ext_rm_leaf(inode, start, ex, *partial_cluster);
2338
2339 while (ex >= EXT_FIRST_EXTENT(eh) &&
2340 ex_ee_block + ex_ee_len > start) {
2341
2342 if (ext4_ext_is_uninitialized(ex))
2343 uninitialized = 1;
2344 else
2345 uninitialized = 0;
2346
2347 ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
2348 uninitialized, ex_ee_len);
2349 path[depth].p_ext = ex;
2350
2351 a = ex_ee_block > start ? ex_ee_block : start;
2352 b = ex_ee_block+ex_ee_len - 1 < end ?
2353 ex_ee_block+ex_ee_len - 1 : end;
2354
2355 ext_debug(" border %u:%u\n", a, b);
2356
2357 /* If this extent is beyond the end of the hole, skip it */
2358 if (end <= ex_ee_block) {
2359 ex--;
2360 ex_ee_block = le32_to_cpu(ex->ee_block);
2361 ex_ee_len = ext4_ext_get_actual_len(ex);
2362 continue;
2363 } else if (a != ex_ee_block &&
2364 b != ex_ee_block + ex_ee_len - 1) {
2365 /*
2366 * If this is a truncate, then this condition should
2367 * never happen because at least one of the end points
2368 * needs to be on the edge of the extent.
2369 */
2370 if (end == EXT_MAX_BLOCKS - 1) {
2371 ext_debug(" bad truncate %u:%u\n",
2372 start, end);
2373 block = 0;
2374 num = 0;
2375 err = -EIO;
2376 goto out;
2377 }
2378 /*
2379 * else this is a hole punch, so the extent needs to
2380 * be split since neither edge of the hole is on the
2381 * extent edge
2382 */
2383 else{
2384 map.m_pblk = ext4_ext_pblock(ex);
2385 map.m_lblk = ex_ee_block;
2386 map.m_len = b - ex_ee_block;
2387
2388 err = ext4_split_extent(handle,
2389 inode, path, &map, 0,
2390 EXT4_GET_BLOCKS_PUNCH_OUT_EXT |
2391 EXT4_GET_BLOCKS_PRE_IO);
2392
2393 if (err < 0)
2394 goto out;
2395
2396 ex_ee_len = ext4_ext_get_actual_len(ex);
2397
2398 b = ex_ee_block+ex_ee_len - 1 < end ?
2399 ex_ee_block+ex_ee_len - 1 : end;
2400
2401 /* Then remove tail of this extent */
2402 block = ex_ee_block;
2403 num = a - block;
2404 }
2405 } else if (a != ex_ee_block) {
2406 /* remove tail of the extent */
2407 block = ex_ee_block;
2408 num = a - block;
2409 } else if (b != ex_ee_block + ex_ee_len - 1) {
2410 /* remove head of the extent */
2411 block = b;
2412 num = ex_ee_block + ex_ee_len - b;
2413
2414 /*
2415 * If this is a truncate, this condition
2416 * should never happen
2417 */
2418 if (end == EXT_MAX_BLOCKS - 1) {
2419 ext_debug(" bad truncate %u:%u\n",
2420 start, end);
2421 err = -EIO;
2422 goto out;
2423 }
2424 } else {
2425 /* remove whole extent: excellent! */
2426 block = ex_ee_block;
2427 num = 0;
2428 if (a != ex_ee_block) {
2429 ext_debug(" bad truncate %u:%u\n",
2430 start, end);
2431 err = -EIO;
2432 goto out;
2433 }
2434
2435 if (b != ex_ee_block + ex_ee_len - 1) {
2436 ext_debug(" bad truncate %u:%u\n",
2437 start, end);
2438 err = -EIO;
2439 goto out;
2440 }
2441 }
2442
2443 /*
2444 * 3 for leaf, sb, and inode plus 2 (bmap and group
2445 * descriptor) for each block group; assume two block
2446 * groups plus ex_ee_len/blocks_per_block_group for
2447 * the worst case
2448 */
2449 credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
2450 if (ex == EXT_FIRST_EXTENT(eh)) {
2451 correct_index = 1;
2452 credits += (ext_depth(inode)) + 1;
2453 }
2454 credits += EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb);
2455
2456 err = ext4_ext_truncate_extend_restart(handle, inode, credits);
2457 if (err)
2458 goto out;
2459
2460 err = ext4_ext_get_access(handle, inode, path + depth);
2461 if (err)
2462 goto out;
2463
2464 err = ext4_remove_blocks(handle, inode, ex, partial_cluster,
2465 a, b);
2466 if (err)
2467 goto out;
2468
2469 if (num == 0) {
2470 /* this extent is removed; mark slot entirely unused */
2471 ext4_ext_store_pblock(ex, 0);
2472 } else if (block != ex_ee_block) {
2473 /*
2474 * If this was a head removal, then we need to update
2475 * the physical block since it is now at a different
2476 * location
2477 */
2478 ext4_ext_store_pblock(ex, ext4_ext_pblock(ex) + (b-a));
2479 }
2480
2481 ex->ee_block = cpu_to_le32(block);
2482 ex->ee_len = cpu_to_le16(num);
2483 /*
2484 * Do not mark uninitialized if all the blocks in the
2485 * extent have been removed.
2486 */
2487 if (uninitialized && num)
2488 ext4_ext_mark_uninitialized(ex);
2489
2490 err = ext4_ext_dirty(handle, inode, path + depth);
2491 if (err)
2492 goto out;
2493
2494 /*
2495 * If the extent was completely released,
2496 * we need to remove it from the leaf
2497 */
2498 if (num == 0) {
2499 if (end != EXT_MAX_BLOCKS - 1) {
2500 /*
2501 * For hole punching, we need to scoot all the
2502 * extents up when an extent is removed so that
2503 * we dont have blank extents in the middle
2504 */
2505 memmove(ex, ex+1, (EXT_LAST_EXTENT(eh) - ex) *
2506 sizeof(struct ext4_extent));
2507
2508 /* Now get rid of the one at the end */
2509 memset(EXT_LAST_EXTENT(eh), 0,
2510 sizeof(struct ext4_extent));
2511 }
2512 le16_add_cpu(&eh->eh_entries, -1);
2513 } else
2514 *partial_cluster = 0;
2515
2516 ext_debug("new extent: %u:%u:%llu\n", block, num,
2517 ext4_ext_pblock(ex));
2518 ex--;
2519 ex_ee_block = le32_to_cpu(ex->ee_block);
2520 ex_ee_len = ext4_ext_get_actual_len(ex);
2521 }
2522
2523 if (correct_index && eh->eh_entries)
2524 err = ext4_ext_correct_indexes(handle, inode, path);
2525
2526 /*
2527 * If there is still a entry in the leaf node, check to see if
2528 * it references the partial cluster. This is the only place
2529 * where it could; if it doesn't, we can free the cluster.
2530 */
2531 if (*partial_cluster && ex >= EXT_FIRST_EXTENT(eh) &&
2532 (EXT4_B2C(sbi, ext4_ext_pblock(ex) + ex_ee_len - 1) !=
2533 *partial_cluster)) {
2534 int flags = EXT4_FREE_BLOCKS_FORGET;
2535
2536 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2537 flags |= EXT4_FREE_BLOCKS_METADATA;
2538
2539 ext4_free_blocks(handle, inode, NULL,
2540 EXT4_C2B(sbi, *partial_cluster),
2541 sbi->s_cluster_ratio, flags);
2542 *partial_cluster = 0;
2543 }
2544
2545 /* if this leaf is free, then we should
2546 * remove it from index block above */
2547 if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
2548 err = ext4_ext_rm_idx(handle, inode, path + depth);
2549
2550out:
2551 return err;
2552}
2553
2554/*
2555 * ext4_ext_more_to_rm:
2556 * returns 1 if current index has to be freed (even partial)
2557 */
2558static int
2559ext4_ext_more_to_rm(struct ext4_ext_path *path)
2560{
2561 BUG_ON(path->p_idx == NULL);
2562
2563 if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
2564 return 0;
2565
2566 /*
2567 * if truncate on deeper level happened, it wasn't partial,
2568 * so we have to consider current index for truncation
2569 */
2570 if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
2571 return 0;
2572 return 1;
2573}
2574
2575static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
2576{
2577 struct super_block *sb = inode->i_sb;
2578 int depth = ext_depth(inode);
2579 struct ext4_ext_path *path;
2580 ext4_fsblk_t partial_cluster = 0;
2581 handle_t *handle;
2582 int i, err;
2583
2584 ext_debug("truncate since %u\n", start);
2585
2586 /* probably first extent we're gonna free will be last in block */
2587 handle = ext4_journal_start(inode, depth + 1);
2588 if (IS_ERR(handle))
2589 return PTR_ERR(handle);
2590
2591again:
2592 ext4_ext_invalidate_cache(inode);
2593
2594 trace_ext4_ext_remove_space(inode, start, depth);
2595
2596 /*
2597 * We start scanning from right side, freeing all the blocks
2598 * after i_size and walking into the tree depth-wise.
2599 */
2600 depth = ext_depth(inode);
2601 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1), GFP_NOFS);
2602 if (path == NULL) {
2603 ext4_journal_stop(handle);
2604 return -ENOMEM;
2605 }
2606 path[0].p_depth = depth;
2607 path[0].p_hdr = ext_inode_hdr(inode);
2608 if (ext4_ext_check(inode, path[0].p_hdr, depth)) {
2609 err = -EIO;
2610 goto out;
2611 }
2612 i = err = 0;
2613
2614 while (i >= 0 && err == 0) {
2615 if (i == depth) {
2616 /* this is leaf block */
2617 err = ext4_ext_rm_leaf(handle, inode, path,
2618 &partial_cluster, start,
2619 EXT_MAX_BLOCKS - 1);
2620 /* root level has p_bh == NULL, brelse() eats this */
2621 brelse(path[i].p_bh);
2622 path[i].p_bh = NULL;
2623 i--;
2624 continue;
2625 }
2626
2627 /* this is index block */
2628 if (!path[i].p_hdr) {
2629 ext_debug("initialize header\n");
2630 path[i].p_hdr = ext_block_hdr(path[i].p_bh);
2631 }
2632
2633 if (!path[i].p_idx) {
2634 /* this level hasn't been touched yet */
2635 path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
2636 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
2637 ext_debug("init index ptr: hdr 0x%p, num %d\n",
2638 path[i].p_hdr,
2639 le16_to_cpu(path[i].p_hdr->eh_entries));
2640 } else {
2641 /* we were already here, see at next index */
2642 path[i].p_idx--;
2643 }
2644
2645 ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
2646 i, EXT_FIRST_INDEX(path[i].p_hdr),
2647 path[i].p_idx);
2648 if (ext4_ext_more_to_rm(path + i)) {
2649 struct buffer_head *bh;
2650 /* go to the next level */
2651 ext_debug("move to level %d (block %llu)\n",
2652 i + 1, ext4_idx_pblock(path[i].p_idx));
2653 memset(path + i + 1, 0, sizeof(*path));
2654 bh = sb_bread(sb, ext4_idx_pblock(path[i].p_idx));
2655 if (!bh) {
2656 /* should we reset i_size? */
2657 err = -EIO;
2658 break;
2659 }
2660 if (WARN_ON(i + 1 > depth)) {
2661 err = -EIO;
2662 break;
2663 }
2664 if (ext4_ext_check(inode, ext_block_hdr(bh),
2665 depth - i - 1)) {
2666 err = -EIO;
2667 break;
2668 }
2669 path[i + 1].p_bh = bh;
2670
2671 /* save actual number of indexes since this
2672 * number is changed at the next iteration */
2673 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
2674 i++;
2675 } else {
2676 /* we finished processing this index, go up */
2677 if (path[i].p_hdr->eh_entries == 0 && i > 0) {
2678 /* index is empty, remove it;
2679 * handle must be already prepared by the
2680 * truncatei_leaf() */
2681 err = ext4_ext_rm_idx(handle, inode, path + i);
2682 }
2683 /* root level has p_bh == NULL, brelse() eats this */
2684 brelse(path[i].p_bh);
2685 path[i].p_bh = NULL;
2686 i--;
2687 ext_debug("return to level %d\n", i);
2688 }
2689 }
2690
2691 trace_ext4_ext_remove_space_done(inode, start, depth, partial_cluster,
2692 path->p_hdr->eh_entries);
2693
2694 /* If we still have something in the partial cluster and we have removed
2695 * even the first extent, then we should free the blocks in the partial
2696 * cluster as well. */
2697 if (partial_cluster && path->p_hdr->eh_entries == 0) {
2698 int flags = EXT4_FREE_BLOCKS_FORGET;
2699
2700 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2701 flags |= EXT4_FREE_BLOCKS_METADATA;
2702
2703 ext4_free_blocks(handle, inode, NULL,
2704 EXT4_C2B(EXT4_SB(sb), partial_cluster),
2705 EXT4_SB(sb)->s_cluster_ratio, flags);
2706 partial_cluster = 0;
2707 }
2708
2709 /* TODO: flexible tree reduction should be here */
2710 if (path->p_hdr->eh_entries == 0) {
2711 /*
2712 * truncate to zero freed all the tree,
2713 * so we need to correct eh_depth
2714 */
2715 err = ext4_ext_get_access(handle, inode, path);
2716 if (err == 0) {
2717 ext_inode_hdr(inode)->eh_depth = 0;
2718 ext_inode_hdr(inode)->eh_max =
2719 cpu_to_le16(ext4_ext_space_root(inode, 0));
2720 err = ext4_ext_dirty(handle, inode, path);
2721 }
2722 }
2723out:
2724 ext4_ext_drop_refs(path);
2725 kfree(path);
2726 if (err == -EAGAIN)
2727 goto again;
2728 ext4_journal_stop(handle);
2729
2730 return err;
2731}
2732
2733/*
2734 * called at mount time
2735 */
2736void ext4_ext_init(struct super_block *sb)
2737{
2738 /*
2739 * possible initialization would be here
2740 */
2741
2742 if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
2743#if defined(AGGRESSIVE_TEST) || defined(CHECK_BINSEARCH) || defined(EXTENTS_STATS)
2744 printk(KERN_INFO "EXT4-fs: file extents enabled");
2745#ifdef AGGRESSIVE_TEST
2746 printk(", aggressive tests");
2747#endif
2748#ifdef CHECK_BINSEARCH
2749 printk(", check binsearch");
2750#endif
2751#ifdef EXTENTS_STATS
2752 printk(", stats");
2753#endif
2754 printk("\n");
2755#endif
2756#ifdef EXTENTS_STATS
2757 spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
2758 EXT4_SB(sb)->s_ext_min = 1 << 30;
2759 EXT4_SB(sb)->s_ext_max = 0;
2760#endif
2761 }
2762}
2763
2764/*
2765 * called at umount time
2766 */
2767void ext4_ext_release(struct super_block *sb)
2768{
2769 if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
2770 return;
2771
2772#ifdef EXTENTS_STATS
2773 if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
2774 struct ext4_sb_info *sbi = EXT4_SB(sb);
2775 printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
2776 sbi->s_ext_blocks, sbi->s_ext_extents,
2777 sbi->s_ext_blocks / sbi->s_ext_extents);
2778 printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
2779 sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
2780 }
2781#endif
2782}
2783
2784/* FIXME!! we need to try to merge to left or right after zero-out */
2785static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
2786{
2787 ext4_fsblk_t ee_pblock;
2788 unsigned int ee_len;
2789 int ret;
2790
2791 ee_len = ext4_ext_get_actual_len(ex);
2792 ee_pblock = ext4_ext_pblock(ex);
2793
2794 ret = sb_issue_zeroout(inode->i_sb, ee_pblock, ee_len, GFP_NOFS);
2795 if (ret > 0)
2796 ret = 0;
2797
2798 return ret;
2799}
2800
2801/*
2802 * used by extent splitting.
2803 */
2804#define EXT4_EXT_MAY_ZEROOUT 0x1 /* safe to zeroout if split fails \
2805 due to ENOSPC */
2806#define EXT4_EXT_MARK_UNINIT1 0x2 /* mark first half uninitialized */
2807#define EXT4_EXT_MARK_UNINIT2 0x4 /* mark second half uninitialized */
2808
2809/*
2810 * ext4_split_extent_at() splits an extent at given block.
2811 *
2812 * @handle: the journal handle
2813 * @inode: the file inode
2814 * @path: the path to the extent
2815 * @split: the logical block where the extent is splitted.
2816 * @split_flags: indicates if the extent could be zeroout if split fails, and
2817 * the states(init or uninit) of new extents.
2818 * @flags: flags used to insert new extent to extent tree.
2819 *
2820 *
2821 * Splits extent [a, b] into two extents [a, @split) and [@split, b], states
2822 * of which are deterimined by split_flag.
2823 *
2824 * There are two cases:
2825 * a> the extent are splitted into two extent.
2826 * b> split is not needed, and just mark the extent.
2827 *
2828 * return 0 on success.
2829 */
2830static int ext4_split_extent_at(handle_t *handle,
2831 struct inode *inode,
2832 struct ext4_ext_path *path,
2833 ext4_lblk_t split,
2834 int split_flag,
2835 int flags)
2836{
2837 ext4_fsblk_t newblock;
2838 ext4_lblk_t ee_block;
2839 struct ext4_extent *ex, newex, orig_ex;
2840 struct ext4_extent *ex2 = NULL;
2841 unsigned int ee_len, depth;
2842 int err = 0;
2843
2844 ext_debug("ext4_split_extents_at: inode %lu, logical"
2845 "block %llu\n", inode->i_ino, (unsigned long long)split);
2846
2847 ext4_ext_show_leaf(inode, path);
2848
2849 depth = ext_depth(inode);
2850 ex = path[depth].p_ext;
2851 ee_block = le32_to_cpu(ex->ee_block);
2852 ee_len = ext4_ext_get_actual_len(ex);
2853 newblock = split - ee_block + ext4_ext_pblock(ex);
2854
2855 BUG_ON(split < ee_block || split >= (ee_block + ee_len));
2856
2857 err = ext4_ext_get_access(handle, inode, path + depth);
2858 if (err)
2859 goto out;
2860
2861 if (split == ee_block) {
2862 /*
2863 * case b: block @split is the block that the extent begins with
2864 * then we just change the state of the extent, and splitting
2865 * is not needed.
2866 */
2867 if (split_flag & EXT4_EXT_MARK_UNINIT2)
2868 ext4_ext_mark_uninitialized(ex);
2869 else
2870 ext4_ext_mark_initialized(ex);
2871
2872 if (!(flags & EXT4_GET_BLOCKS_PRE_IO))
2873 ext4_ext_try_to_merge(inode, path, ex);
2874
2875 err = ext4_ext_dirty(handle, inode, path + depth);
2876 goto out;
2877 }
2878
2879 /* case a */
2880 memcpy(&orig_ex, ex, sizeof(orig_ex));
2881 ex->ee_len = cpu_to_le16(split - ee_block);
2882 if (split_flag & EXT4_EXT_MARK_UNINIT1)
2883 ext4_ext_mark_uninitialized(ex);
2884
2885 /*
2886 * path may lead to new leaf, not to original leaf any more
2887 * after ext4_ext_insert_extent() returns,
2888 */
2889 err = ext4_ext_dirty(handle, inode, path + depth);
2890 if (err)
2891 goto fix_extent_len;
2892
2893 ex2 = &newex;
2894 ex2->ee_block = cpu_to_le32(split);
2895 ex2->ee_len = cpu_to_le16(ee_len - (split - ee_block));
2896 ext4_ext_store_pblock(ex2, newblock);
2897 if (split_flag & EXT4_EXT_MARK_UNINIT2)
2898 ext4_ext_mark_uninitialized(ex2);
2899
2900 err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
2901 if (err == -ENOSPC && (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
2902 err = ext4_ext_zeroout(inode, &orig_ex);
2903 if (err)
2904 goto fix_extent_len;
2905 /* update the extent length and mark as initialized */
2906 ex->ee_len = cpu_to_le32(ee_len);
2907 ext4_ext_try_to_merge(inode, path, ex);
2908 err = ext4_ext_dirty(handle, inode, path + depth);
2909 goto out;
2910 } else if (err)
2911 goto fix_extent_len;
2912
2913out:
2914 ext4_ext_show_leaf(inode, path);
2915 return err;
2916
2917fix_extent_len:
2918 ex->ee_len = orig_ex.ee_len;
2919 ext4_ext_dirty(handle, inode, path + depth);
2920 return err;
2921}
2922
2923/*
2924 * ext4_split_extents() splits an extent and mark extent which is covered
2925 * by @map as split_flags indicates
2926 *
2927 * It may result in splitting the extent into multiple extents (upto three)
2928 * There are three possibilities:
2929 * a> There is no split required
2930 * b> Splits in two extents: Split is happening at either end of the extent
2931 * c> Splits in three extents: Somone is splitting in middle of the extent
2932 *
2933 */
2934static int ext4_split_extent(handle_t *handle,
2935 struct inode *inode,
2936 struct ext4_ext_path *path,
2937 struct ext4_map_blocks *map,
2938 int split_flag,
2939 int flags)
2940{
2941 ext4_lblk_t ee_block;
2942 struct ext4_extent *ex;
2943 unsigned int ee_len, depth;
2944 int err = 0;
2945 int uninitialized;
2946 int split_flag1, flags1;
2947
2948 depth = ext_depth(inode);
2949 ex = path[depth].p_ext;
2950 ee_block = le32_to_cpu(ex->ee_block);
2951 ee_len = ext4_ext_get_actual_len(ex);
2952 uninitialized = ext4_ext_is_uninitialized(ex);
2953
2954 if (map->m_lblk + map->m_len < ee_block + ee_len) {
2955 split_flag1 = split_flag & EXT4_EXT_MAY_ZEROOUT ?
2956 EXT4_EXT_MAY_ZEROOUT : 0;
2957 flags1 = flags | EXT4_GET_BLOCKS_PRE_IO;
2958 if (uninitialized)
2959 split_flag1 |= EXT4_EXT_MARK_UNINIT1 |
2960 EXT4_EXT_MARK_UNINIT2;
2961 err = ext4_split_extent_at(handle, inode, path,
2962 map->m_lblk + map->m_len, split_flag1, flags1);
2963 if (err)
2964 goto out;
2965 }
2966
2967 ext4_ext_drop_refs(path);
2968 path = ext4_ext_find_extent(inode, map->m_lblk, path);
2969 if (IS_ERR(path))
2970 return PTR_ERR(path);
2971
2972 if (map->m_lblk >= ee_block) {
2973 split_flag1 = split_flag & EXT4_EXT_MAY_ZEROOUT ?
2974 EXT4_EXT_MAY_ZEROOUT : 0;
2975 if (uninitialized)
2976 split_flag1 |= EXT4_EXT_MARK_UNINIT1;
2977 if (split_flag & EXT4_EXT_MARK_UNINIT2)
2978 split_flag1 |= EXT4_EXT_MARK_UNINIT2;
2979 err = ext4_split_extent_at(handle, inode, path,
2980 map->m_lblk, split_flag1, flags);
2981 if (err)
2982 goto out;
2983 }
2984
2985 ext4_ext_show_leaf(inode, path);
2986out:
2987 return err ? err : map->m_len;
2988}
2989
2990#define EXT4_EXT_ZERO_LEN 7
2991/*
2992 * This function is called by ext4_ext_map_blocks() if someone tries to write
2993 * to an uninitialized extent. It may result in splitting the uninitialized
2994 * extent into multiple extents (up to three - one initialized and two
2995 * uninitialized).
2996 * There are three possibilities:
2997 * a> There is no split required: Entire extent should be initialized
2998 * b> Splits in two extents: Write is happening at either end of the extent
2999 * c> Splits in three extents: Somone is writing in middle of the extent
3000 */
3001static int ext4_ext_convert_to_initialized(handle_t *handle,
3002 struct inode *inode,
3003 struct ext4_map_blocks *map,
3004 struct ext4_ext_path *path)
3005{
3006 struct ext4_map_blocks split_map;
3007 struct ext4_extent zero_ex;
3008 struct ext4_extent *ex;
3009 ext4_lblk_t ee_block, eof_block;
3010 unsigned int allocated, ee_len, depth;
3011 int err = 0;
3012 int split_flag = 0;
3013
3014 ext_debug("ext4_ext_convert_to_initialized: inode %lu, logical"
3015 "block %llu, max_blocks %u\n", inode->i_ino,
3016 (unsigned long long)map->m_lblk, map->m_len);
3017
3018 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3019 inode->i_sb->s_blocksize_bits;
3020 if (eof_block < map->m_lblk + map->m_len)
3021 eof_block = map->m_lblk + map->m_len;
3022
3023 depth = ext_depth(inode);
3024 ex = path[depth].p_ext;
3025 ee_block = le32_to_cpu(ex->ee_block);
3026 ee_len = ext4_ext_get_actual_len(ex);
3027 allocated = ee_len - (map->m_lblk - ee_block);
3028
3029 WARN_ON(map->m_lblk < ee_block);
3030 /*
3031 * It is safe to convert extent to initialized via explicit
3032 * zeroout only if extent is fully insde i_size or new_size.
3033 */
3034 split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
3035
3036 /* If extent has less than 2*EXT4_EXT_ZERO_LEN zerout directly */
3037 if (ee_len <= 2*EXT4_EXT_ZERO_LEN &&
3038 (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
3039 err = ext4_ext_zeroout(inode, ex);
3040 if (err)
3041 goto out;
3042
3043 err = ext4_ext_get_access(handle, inode, path + depth);
3044 if (err)
3045 goto out;
3046 ext4_ext_mark_initialized(ex);
3047 ext4_ext_try_to_merge(inode, path, ex);
3048 err = ext4_ext_dirty(handle, inode, path + depth);
3049 goto out;
3050 }
3051
3052 /*
3053 * four cases:
3054 * 1. split the extent into three extents.
3055 * 2. split the extent into two extents, zeroout the first half.
3056 * 3. split the extent into two extents, zeroout the second half.
3057 * 4. split the extent into two extents with out zeroout.
3058 */
3059 split_map.m_lblk = map->m_lblk;
3060 split_map.m_len = map->m_len;
3061
3062 if (allocated > map->m_len) {
3063 if (allocated <= EXT4_EXT_ZERO_LEN &&
3064 (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
3065 /* case 3 */
3066 zero_ex.ee_block =
3067 cpu_to_le32(map->m_lblk);
3068 zero_ex.ee_len = cpu_to_le16(allocated);
3069 ext4_ext_store_pblock(&zero_ex,
3070 ext4_ext_pblock(ex) + map->m_lblk - ee_block);
3071 err = ext4_ext_zeroout(inode, &zero_ex);
3072 if (err)
3073 goto out;
3074 split_map.m_lblk = map->m_lblk;
3075 split_map.m_len = allocated;
3076 } else if ((map->m_lblk - ee_block + map->m_len <
3077 EXT4_EXT_ZERO_LEN) &&
3078 (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
3079 /* case 2 */
3080 if (map->m_lblk != ee_block) {
3081 zero_ex.ee_block = ex->ee_block;
3082 zero_ex.ee_len = cpu_to_le16(map->m_lblk -
3083 ee_block);
3084 ext4_ext_store_pblock(&zero_ex,
3085 ext4_ext_pblock(ex));
3086 err = ext4_ext_zeroout(inode, &zero_ex);
3087 if (err)
3088 goto out;
3089 }
3090
3091 split_map.m_lblk = ee_block;
3092 split_map.m_len = map->m_lblk - ee_block + map->m_len;
3093 allocated = map->m_len;
3094 }
3095 }
3096
3097 allocated = ext4_split_extent(handle, inode, path,
3098 &split_map, split_flag, 0);
3099 if (allocated < 0)
3100 err = allocated;
3101
3102out:
3103 return err ? err : allocated;
3104}
3105
3106/*
3107 * This function is called by ext4_ext_map_blocks() from
3108 * ext4_get_blocks_dio_write() when DIO to write
3109 * to an uninitialized extent.
3110 *
3111 * Writing to an uninitialized extent may result in splitting the uninitialized
3112 * extent into multiple /initialized uninitialized extents (up to three)
3113 * There are three possibilities:
3114 * a> There is no split required: Entire extent should be uninitialized
3115 * b> Splits in two extents: Write is happening at either end of the extent
3116 * c> Splits in three extents: Somone is writing in middle of the extent
3117 *
3118 * One of more index blocks maybe needed if the extent tree grow after
3119 * the uninitialized extent split. To prevent ENOSPC occur at the IO
3120 * complete, we need to split the uninitialized extent before DIO submit
3121 * the IO. The uninitialized extent called at this time will be split
3122 * into three uninitialized extent(at most). After IO complete, the part
3123 * being filled will be convert to initialized by the end_io callback function
3124 * via ext4_convert_unwritten_extents().
3125 *
3126 * Returns the size of uninitialized extent to be written on success.
3127 */
3128static int ext4_split_unwritten_extents(handle_t *handle,
3129 struct inode *inode,
3130 struct ext4_map_blocks *map,
3131 struct ext4_ext_path *path,
3132 int flags)
3133{
3134 ext4_lblk_t eof_block;
3135 ext4_lblk_t ee_block;
3136 struct ext4_extent *ex;
3137 unsigned int ee_len;
3138 int split_flag = 0, depth;
3139
3140 ext_debug("ext4_split_unwritten_extents: inode %lu, logical"
3141 "block %llu, max_blocks %u\n", inode->i_ino,
3142 (unsigned long long)map->m_lblk, map->m_len);
3143
3144 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3145 inode->i_sb->s_blocksize_bits;
3146 if (eof_block < map->m_lblk + map->m_len)
3147 eof_block = map->m_lblk + map->m_len;
3148 /*
3149 * It is safe to convert extent to initialized via explicit
3150 * zeroout only if extent is fully insde i_size or new_size.
3151 */
3152 depth = ext_depth(inode);
3153 ex = path[depth].p_ext;
3154 ee_block = le32_to_cpu(ex->ee_block);
3155 ee_len = ext4_ext_get_actual_len(ex);
3156
3157 split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
3158 split_flag |= EXT4_EXT_MARK_UNINIT2;
3159
3160 flags |= EXT4_GET_BLOCKS_PRE_IO;
3161 return ext4_split_extent(handle, inode, path, map, split_flag, flags);
3162}
3163
3164static int ext4_convert_unwritten_extents_endio(handle_t *handle,
3165 struct inode *inode,
3166 struct ext4_ext_path *path)
3167{
3168 struct ext4_extent *ex;
3169 int depth;
3170 int err = 0;
3171
3172 depth = ext_depth(inode);
3173 ex = path[depth].p_ext;
3174
3175 ext_debug("ext4_convert_unwritten_extents_endio: inode %lu, logical"
3176 "block %llu, max_blocks %u\n", inode->i_ino,
3177 (unsigned long long)le32_to_cpu(ex->ee_block),
3178 ext4_ext_get_actual_len(ex));
3179
3180 err = ext4_ext_get_access(handle, inode, path + depth);
3181 if (err)
3182 goto out;
3183 /* first mark the extent as initialized */
3184 ext4_ext_mark_initialized(ex);
3185
3186 /* note: ext4_ext_correct_indexes() isn't needed here because
3187 * borders are not changed
3188 */
3189 ext4_ext_try_to_merge(inode, path, ex);
3190
3191 /* Mark modified extent as dirty */
3192 err = ext4_ext_dirty(handle, inode, path + depth);
3193out:
3194 ext4_ext_show_leaf(inode, path);
3195 return err;
3196}
3197
3198static void unmap_underlying_metadata_blocks(struct block_device *bdev,
3199 sector_t block, int count)
3200{
3201 int i;
3202 for (i = 0; i < count; i++)
3203 unmap_underlying_metadata(bdev, block + i);
3204}
3205
3206/*
3207 * Handle EOFBLOCKS_FL flag, clearing it if necessary
3208 */
3209static int check_eofblocks_fl(handle_t *handle, struct inode *inode,
3210 ext4_lblk_t lblk,
3211 struct ext4_ext_path *path,
3212 unsigned int len)
3213{
3214 int i, depth;
3215 struct ext4_extent_header *eh;
3216 struct ext4_extent *last_ex;
3217
3218 if (!ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS))
3219 return 0;
3220
3221 depth = ext_depth(inode);
3222 eh = path[depth].p_hdr;
3223
3224 if (unlikely(!eh->eh_entries)) {
3225 EXT4_ERROR_INODE(inode, "eh->eh_entries == 0 and "
3226 "EOFBLOCKS_FL set");
3227 return -EIO;
3228 }
3229 last_ex = EXT_LAST_EXTENT(eh);
3230 /*
3231 * We should clear the EOFBLOCKS_FL flag if we are writing the
3232 * last block in the last extent in the file. We test this by
3233 * first checking to see if the caller to
3234 * ext4_ext_get_blocks() was interested in the last block (or
3235 * a block beyond the last block) in the current extent. If
3236 * this turns out to be false, we can bail out from this
3237 * function immediately.
3238 */
3239 if (lblk + len < le32_to_cpu(last_ex->ee_block) +
3240 ext4_ext_get_actual_len(last_ex))
3241 return 0;
3242 /*
3243 * If the caller does appear to be planning to write at or
3244 * beyond the end of the current extent, we then test to see
3245 * if the current extent is the last extent in the file, by
3246 * checking to make sure it was reached via the rightmost node
3247 * at each level of the tree.
3248 */
3249 for (i = depth-1; i >= 0; i--)
3250 if (path[i].p_idx != EXT_LAST_INDEX(path[i].p_hdr))
3251 return 0;
3252 ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
3253 return ext4_mark_inode_dirty(handle, inode);
3254}
3255
3256/**
3257 * ext4_find_delalloc_range: find delayed allocated block in the given range.
3258 *
3259 * Goes through the buffer heads in the range [lblk_start, lblk_end] and returns
3260 * whether there are any buffers marked for delayed allocation. It returns '1'
3261 * on the first delalloc'ed buffer head found. If no buffer head in the given
3262 * range is marked for delalloc, it returns 0.
3263 * lblk_start should always be <= lblk_end.
3264 * search_hint_reverse is to indicate that searching in reverse from lblk_end to
3265 * lblk_start might be more efficient (i.e., we will likely hit the delalloc'ed
3266 * block sooner). This is useful when blocks are truncated sequentially from
3267 * lblk_start towards lblk_end.
3268 */
3269static int ext4_find_delalloc_range(struct inode *inode,
3270 ext4_lblk_t lblk_start,
3271 ext4_lblk_t lblk_end,
3272 int search_hint_reverse)
3273{
3274 struct address_space *mapping = inode->i_mapping;
3275 struct buffer_head *head, *bh = NULL;
3276 struct page *page;
3277 ext4_lblk_t i, pg_lblk;
3278 pgoff_t index;
3279
3280 /* reverse search wont work if fs block size is less than page size */
3281 if (inode->i_blkbits < PAGE_CACHE_SHIFT)
3282 search_hint_reverse = 0;
3283
3284 if (search_hint_reverse)
3285 i = lblk_end;
3286 else
3287 i = lblk_start;
3288
3289 index = i >> (PAGE_CACHE_SHIFT - inode->i_blkbits);
3290
3291 while ((i >= lblk_start) && (i <= lblk_end)) {
3292 page = find_get_page(mapping, index);
3293 if (!page)
3294 goto nextpage;
3295
3296 if (!page_has_buffers(page))
3297 goto nextpage;
3298
3299 head = page_buffers(page);
3300 if (!head)
3301 goto nextpage;
3302
3303 bh = head;
3304 pg_lblk = index << (PAGE_CACHE_SHIFT -
3305 inode->i_blkbits);
3306 do {
3307 if (unlikely(pg_lblk < lblk_start)) {
3308 /*
3309 * This is possible when fs block size is less
3310 * than page size and our cluster starts/ends in
3311 * middle of the page. So we need to skip the
3312 * initial few blocks till we reach the 'lblk'
3313 */
3314 pg_lblk++;
3315 continue;
3316 }
3317
3318 /* Check if the buffer is delayed allocated and that it
3319 * is not yet mapped. (when da-buffers are mapped during
3320 * their writeout, their da_mapped bit is set.)
3321 */
3322 if (buffer_delay(bh) && !buffer_da_mapped(bh)) {
3323 page_cache_release(page);
3324 trace_ext4_find_delalloc_range(inode,
3325 lblk_start, lblk_end,
3326 search_hint_reverse,
3327 1, i);
3328 return 1;
3329 }
3330 if (search_hint_reverse)
3331 i--;
3332 else
3333 i++;
3334 } while ((i >= lblk_start) && (i <= lblk_end) &&
3335 ((bh = bh->b_this_page) != head));
3336nextpage:
3337 if (page)
3338 page_cache_release(page);
3339 /*
3340 * Move to next page. 'i' will be the first lblk in the next
3341 * page.
3342 */
3343 if (search_hint_reverse)
3344 index--;
3345 else
3346 index++;
3347 i = index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
3348 }
3349
3350 trace_ext4_find_delalloc_range(inode, lblk_start, lblk_end,
3351 search_hint_reverse, 0, 0);
3352 return 0;
3353}
3354
3355int ext4_find_delalloc_cluster(struct inode *inode, ext4_lblk_t lblk,
3356 int search_hint_reverse)
3357{
3358 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3359 ext4_lblk_t lblk_start, lblk_end;
3360 lblk_start = lblk & (~(sbi->s_cluster_ratio - 1));
3361 lblk_end = lblk_start + sbi->s_cluster_ratio - 1;
3362
3363 return ext4_find_delalloc_range(inode, lblk_start, lblk_end,
3364 search_hint_reverse);
3365}
3366
3367/**
3368 * Determines how many complete clusters (out of those specified by the 'map')
3369 * are under delalloc and were reserved quota for.
3370 * This function is called when we are writing out the blocks that were
3371 * originally written with their allocation delayed, but then the space was
3372 * allocated using fallocate() before the delayed allocation could be resolved.
3373 * The cases to look for are:
3374 * ('=' indicated delayed allocated blocks
3375 * '-' indicates non-delayed allocated blocks)
3376 * (a) partial clusters towards beginning and/or end outside of allocated range
3377 * are not delalloc'ed.
3378 * Ex:
3379 * |----c---=|====c====|====c====|===-c----|
3380 * |++++++ allocated ++++++|
3381 * ==> 4 complete clusters in above example
3382 *
3383 * (b) partial cluster (outside of allocated range) towards either end is
3384 * marked for delayed allocation. In this case, we will exclude that
3385 * cluster.
3386 * Ex:
3387 * |----====c========|========c========|
3388 * |++++++ allocated ++++++|
3389 * ==> 1 complete clusters in above example
3390 *
3391 * Ex:
3392 * |================c================|
3393 * |++++++ allocated ++++++|
3394 * ==> 0 complete clusters in above example
3395 *
3396 * The ext4_da_update_reserve_space will be called only if we
3397 * determine here that there were some "entire" clusters that span
3398 * this 'allocated' range.
3399 * In the non-bigalloc case, this function will just end up returning num_blks
3400 * without ever calling ext4_find_delalloc_range.
3401 */
3402static unsigned int
3403get_reserved_cluster_alloc(struct inode *inode, ext4_lblk_t lblk_start,
3404 unsigned int num_blks)
3405{
3406 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3407 ext4_lblk_t alloc_cluster_start, alloc_cluster_end;
3408 ext4_lblk_t lblk_from, lblk_to, c_offset;
3409 unsigned int allocated_clusters = 0;
3410
3411 alloc_cluster_start = EXT4_B2C(sbi, lblk_start);
3412 alloc_cluster_end = EXT4_B2C(sbi, lblk_start + num_blks - 1);
3413
3414 /* max possible clusters for this allocation */
3415 allocated_clusters = alloc_cluster_end - alloc_cluster_start + 1;
3416
3417 trace_ext4_get_reserved_cluster_alloc(inode, lblk_start, num_blks);
3418
3419 /* Check towards left side */
3420 c_offset = lblk_start & (sbi->s_cluster_ratio - 1);
3421 if (c_offset) {
3422 lblk_from = lblk_start & (~(sbi->s_cluster_ratio - 1));
3423 lblk_to = lblk_from + c_offset - 1;
3424
3425 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to, 0))
3426 allocated_clusters--;
3427 }
3428
3429 /* Now check towards right. */
3430 c_offset = (lblk_start + num_blks) & (sbi->s_cluster_ratio - 1);
3431 if (allocated_clusters && c_offset) {
3432 lblk_from = lblk_start + num_blks;
3433 lblk_to = lblk_from + (sbi->s_cluster_ratio - c_offset) - 1;
3434
3435 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to, 0))
3436 allocated_clusters--;
3437 }
3438
3439 return allocated_clusters;
3440}
3441
3442static int
3443ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
3444 struct ext4_map_blocks *map,
3445 struct ext4_ext_path *path, int flags,
3446 unsigned int allocated, ext4_fsblk_t newblock)
3447{
3448 int ret = 0;
3449 int err = 0;
3450 ext4_io_end_t *io = EXT4_I(inode)->cur_aio_dio;
3451
3452 ext_debug("ext4_ext_handle_uninitialized_extents: inode %lu, logical"
3453 "block %llu, max_blocks %u, flags %d, allocated %u",
3454 inode->i_ino, (unsigned long long)map->m_lblk, map->m_len,
3455 flags, allocated);
3456 ext4_ext_show_leaf(inode, path);
3457
3458 trace_ext4_ext_handle_uninitialized_extents(inode, map, allocated,
3459 newblock);
3460
3461 /* get_block() before submit the IO, split the extent */
3462 if ((flags & EXT4_GET_BLOCKS_PRE_IO)) {
3463 ret = ext4_split_unwritten_extents(handle, inode, map,
3464 path, flags);
3465 /*
3466 * Flag the inode(non aio case) or end_io struct (aio case)
3467 * that this IO needs to conversion to written when IO is
3468 * completed
3469 */
3470 if (io && !(io->flag & EXT4_IO_END_UNWRITTEN)) {
3471 io->flag = EXT4_IO_END_UNWRITTEN;
3472 atomic_inc(&EXT4_I(inode)->i_aiodio_unwritten);
3473 } else
3474 ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
3475 if (ext4_should_dioread_nolock(inode))
3476 map->m_flags |= EXT4_MAP_UNINIT;
3477 goto out;
3478 }
3479 /* IO end_io complete, convert the filled extent to written */
3480 if ((flags & EXT4_GET_BLOCKS_CONVERT)) {
3481 ret = ext4_convert_unwritten_extents_endio(handle, inode,
3482 path);
3483 if (ret >= 0) {
3484 ext4_update_inode_fsync_trans(handle, inode, 1);
3485 err = check_eofblocks_fl(handle, inode, map->m_lblk,
3486 path, map->m_len);
3487 } else
3488 err = ret;
3489 goto out2;
3490 }
3491 /* buffered IO case */
3492 /*
3493 * repeat fallocate creation request
3494 * we already have an unwritten extent
3495 */
3496 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT)
3497 goto map_out;
3498
3499 /* buffered READ or buffered write_begin() lookup */
3500 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
3501 /*
3502 * We have blocks reserved already. We
3503 * return allocated blocks so that delalloc
3504 * won't do block reservation for us. But
3505 * the buffer head will be unmapped so that
3506 * a read from the block returns 0s.
3507 */
3508 map->m_flags |= EXT4_MAP_UNWRITTEN;
3509 goto out1;
3510 }
3511
3512 /* buffered write, writepage time, convert*/
3513 ret = ext4_ext_convert_to_initialized(handle, inode, map, path);
3514 if (ret >= 0) {
3515 ext4_update_inode_fsync_trans(handle, inode, 1);
3516 err = check_eofblocks_fl(handle, inode, map->m_lblk, path,
3517 map->m_len);
3518 if (err < 0)
3519 goto out2;
3520 }
3521
3522out:
3523 if (ret <= 0) {
3524 err = ret;
3525 goto out2;
3526 } else
3527 allocated = ret;
3528 map->m_flags |= EXT4_MAP_NEW;
3529 /*
3530 * if we allocated more blocks than requested
3531 * we need to make sure we unmap the extra block
3532 * allocated. The actual needed block will get
3533 * unmapped later when we find the buffer_head marked
3534 * new.
3535 */
3536 if (allocated > map->m_len) {
3537 unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
3538 newblock + map->m_len,
3539 allocated - map->m_len);
3540 allocated = map->m_len;
3541 }
3542
3543 /*
3544 * If we have done fallocate with the offset that is already
3545 * delayed allocated, we would have block reservation
3546 * and quota reservation done in the delayed write path.
3547 * But fallocate would have already updated quota and block
3548 * count for this offset. So cancel these reservation
3549 */
3550 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
3551 unsigned int reserved_clusters;
3552 reserved_clusters = get_reserved_cluster_alloc(inode,
3553 map->m_lblk, map->m_len);
3554 if (reserved_clusters)
3555 ext4_da_update_reserve_space(inode,
3556 reserved_clusters,
3557 0);
3558 }
3559
3560map_out:
3561 map->m_flags |= EXT4_MAP_MAPPED;
3562out1:
3563 if (allocated > map->m_len)
3564 allocated = map->m_len;
3565 ext4_ext_show_leaf(inode, path);
3566 map->m_pblk = newblock;
3567 map->m_len = allocated;
3568out2:
3569 if (path) {
3570 ext4_ext_drop_refs(path);
3571 kfree(path);
3572 }
3573 return err ? err : allocated;
3574}
3575
3576/*
3577 * get_implied_cluster_alloc - check to see if the requested
3578 * allocation (in the map structure) overlaps with a cluster already
3579 * allocated in an extent.
3580 * @sb The filesystem superblock structure
3581 * @map The requested lblk->pblk mapping
3582 * @ex The extent structure which might contain an implied
3583 * cluster allocation
3584 *
3585 * This function is called by ext4_ext_map_blocks() after we failed to
3586 * find blocks that were already in the inode's extent tree. Hence,
3587 * we know that the beginning of the requested region cannot overlap
3588 * the extent from the inode's extent tree. There are three cases we
3589 * want to catch. The first is this case:
3590 *
3591 * |--- cluster # N--|
3592 * |--- extent ---| |---- requested region ---|
3593 * |==========|
3594 *
3595 * The second case that we need to test for is this one:
3596 *
3597 * |--------- cluster # N ----------------|
3598 * |--- requested region --| |------- extent ----|
3599 * |=======================|
3600 *
3601 * The third case is when the requested region lies between two extents
3602 * within the same cluster:
3603 * |------------- cluster # N-------------|
3604 * |----- ex -----| |---- ex_right ----|
3605 * |------ requested region ------|
3606 * |================|
3607 *
3608 * In each of the above cases, we need to set the map->m_pblk and
3609 * map->m_len so it corresponds to the return the extent labelled as
3610 * "|====|" from cluster #N, since it is already in use for data in
3611 * cluster EXT4_B2C(sbi, map->m_lblk). We will then return 1 to
3612 * signal to ext4_ext_map_blocks() that map->m_pblk should be treated
3613 * as a new "allocated" block region. Otherwise, we will return 0 and
3614 * ext4_ext_map_blocks() will then allocate one or more new clusters
3615 * by calling ext4_mb_new_blocks().
3616 */
3617static int get_implied_cluster_alloc(struct super_block *sb,
3618 struct ext4_map_blocks *map,
3619 struct ext4_extent *ex,
3620 struct ext4_ext_path *path)
3621{
3622 struct ext4_sb_info *sbi = EXT4_SB(sb);
3623 ext4_lblk_t c_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
3624 ext4_lblk_t ex_cluster_start, ex_cluster_end;
3625 ext4_lblk_t rr_cluster_start, rr_cluster_end;
3626 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
3627 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
3628 unsigned short ee_len = ext4_ext_get_actual_len(ex);
3629
3630 /* The extent passed in that we are trying to match */
3631 ex_cluster_start = EXT4_B2C(sbi, ee_block);
3632 ex_cluster_end = EXT4_B2C(sbi, ee_block + ee_len - 1);
3633
3634 /* The requested region passed into ext4_map_blocks() */
3635 rr_cluster_start = EXT4_B2C(sbi, map->m_lblk);
3636 rr_cluster_end = EXT4_B2C(sbi, map->m_lblk + map->m_len - 1);
3637
3638 if ((rr_cluster_start == ex_cluster_end) ||
3639 (rr_cluster_start == ex_cluster_start)) {
3640 if (rr_cluster_start == ex_cluster_end)
3641 ee_start += ee_len - 1;
3642 map->m_pblk = (ee_start & ~(sbi->s_cluster_ratio - 1)) +
3643 c_offset;
3644 map->m_len = min(map->m_len,
3645 (unsigned) sbi->s_cluster_ratio - c_offset);
3646 /*
3647 * Check for and handle this case:
3648 *
3649 * |--------- cluster # N-------------|
3650 * |------- extent ----|
3651 * |--- requested region ---|
3652 * |===========|
3653 */
3654
3655 if (map->m_lblk < ee_block)
3656 map->m_len = min(map->m_len, ee_block - map->m_lblk);
3657
3658 /*
3659 * Check for the case where there is already another allocated
3660 * block to the right of 'ex' but before the end of the cluster.
3661 *
3662 * |------------- cluster # N-------------|
3663 * |----- ex -----| |---- ex_right ----|
3664 * |------ requested region ------|
3665 * |================|
3666 */
3667 if (map->m_lblk > ee_block) {
3668 ext4_lblk_t next = ext4_ext_next_allocated_block(path);
3669 map->m_len = min(map->m_len, next - map->m_lblk);
3670 }
3671
3672 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 1);
3673 return 1;
3674 }
3675
3676 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 0);
3677 return 0;
3678}
3679
3680
3681/*
3682 * Block allocation/map/preallocation routine for extents based files
3683 *
3684 *
3685 * Need to be called with
3686 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
3687 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
3688 *
3689 * return > 0, number of of blocks already mapped/allocated
3690 * if create == 0 and these are pre-allocated blocks
3691 * buffer head is unmapped
3692 * otherwise blocks are mapped
3693 *
3694 * return = 0, if plain look up failed (blocks have not been allocated)
3695 * buffer head is unmapped
3696 *
3697 * return < 0, error case.
3698 */
3699int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
3700 struct ext4_map_blocks *map, int flags)
3701{
3702 struct ext4_ext_path *path = NULL;
3703 struct ext4_extent newex, *ex, *ex2;
3704 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3705 ext4_fsblk_t newblock = 0;
3706 int free_on_err = 0, err = 0, depth, ret;
3707 unsigned int allocated = 0, offset = 0;
3708 unsigned int allocated_clusters = 0, reserved_clusters = 0;
3709 unsigned int punched_out = 0;
3710 unsigned int result = 0;
3711 struct ext4_allocation_request ar;
3712 ext4_io_end_t *io = EXT4_I(inode)->cur_aio_dio;
3713 ext4_lblk_t cluster_offset;
3714 struct ext4_map_blocks punch_map;
3715
3716 ext_debug("blocks %u/%u requested for inode %lu\n",
3717 map->m_lblk, map->m_len, inode->i_ino);
3718 trace_ext4_ext_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
3719
3720 /* check in cache */
3721 if (!(flags & EXT4_GET_BLOCKS_PUNCH_OUT_EXT) &&
3722 ext4_ext_in_cache(inode, map->m_lblk, &newex)) {
3723 if (!newex.ee_start_lo && !newex.ee_start_hi) {
3724 if ((sbi->s_cluster_ratio > 1) &&
3725 ext4_find_delalloc_cluster(inode, map->m_lblk, 0))
3726 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
3727
3728 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
3729 /*
3730 * block isn't allocated yet and
3731 * user doesn't want to allocate it
3732 */
3733 goto out2;
3734 }
3735 /* we should allocate requested block */
3736 } else {
3737 /* block is already allocated */
3738 if (sbi->s_cluster_ratio > 1)
3739 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
3740 newblock = map->m_lblk
3741 - le32_to_cpu(newex.ee_block)
3742 + ext4_ext_pblock(&newex);
3743 /* number of remaining blocks in the extent */
3744 allocated = ext4_ext_get_actual_len(&newex) -
3745 (map->m_lblk - le32_to_cpu(newex.ee_block));
3746 goto out;
3747 }
3748 }
3749
3750 /* find extent for this block */
3751 path = ext4_ext_find_extent(inode, map->m_lblk, NULL);
3752 if (IS_ERR(path)) {
3753 err = PTR_ERR(path);
3754 path = NULL;
3755 goto out2;
3756 }
3757
3758 depth = ext_depth(inode);
3759
3760 /*
3761 * consistent leaf must not be empty;
3762 * this situation is possible, though, _during_ tree modification;
3763 * this is why assert can't be put in ext4_ext_find_extent()
3764 */
3765 if (unlikely(path[depth].p_ext == NULL && depth != 0)) {
3766 EXT4_ERROR_INODE(inode, "bad extent address "
3767 "lblock: %lu, depth: %d pblock %lld",
3768 (unsigned long) map->m_lblk, depth,
3769 path[depth].p_block);
3770 err = -EIO;
3771 goto out2;
3772 }
3773
3774 ex = path[depth].p_ext;
3775 if (ex) {
3776 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
3777 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
3778 unsigned short ee_len;
3779
3780 /*
3781 * Uninitialized extents are treated as holes, except that
3782 * we split out initialized portions during a write.
3783 */
3784 ee_len = ext4_ext_get_actual_len(ex);
3785
3786 trace_ext4_ext_show_extent(inode, ee_block, ee_start, ee_len);
3787
3788 /* if found extent covers block, simply return it */
3789 if (in_range(map->m_lblk, ee_block, ee_len)) {
3790 ext4_fsblk_t partial_cluster = 0;
3791
3792 newblock = map->m_lblk - ee_block + ee_start;
3793 /* number of remaining blocks in the extent */
3794 allocated = ee_len - (map->m_lblk - ee_block);
3795 ext_debug("%u fit into %u:%d -> %llu\n", map->m_lblk,
3796 ee_block, ee_len, newblock);
3797
3798 if ((flags & EXT4_GET_BLOCKS_PUNCH_OUT_EXT) == 0) {
3799 /*
3800 * Do not put uninitialized extent
3801 * in the cache
3802 */
3803 if (!ext4_ext_is_uninitialized(ex)) {
3804 ext4_ext_put_in_cache(inode, ee_block,
3805 ee_len, ee_start);
3806 goto out;
3807 }
3808 ret = ext4_ext_handle_uninitialized_extents(
3809 handle, inode, map, path, flags,
3810 allocated, newblock);
3811 return ret;
3812 }
3813
3814 /*
3815 * Punch out the map length, but only to the
3816 * end of the extent
3817 */
3818 punched_out = allocated < map->m_len ?
3819 allocated : map->m_len;
3820
3821 /*
3822 * Sense extents need to be converted to
3823 * uninitialized, they must fit in an
3824 * uninitialized extent
3825 */
3826 if (punched_out > EXT_UNINIT_MAX_LEN)
3827 punched_out = EXT_UNINIT_MAX_LEN;
3828
3829 punch_map.m_lblk = map->m_lblk;
3830 punch_map.m_pblk = newblock;
3831 punch_map.m_len = punched_out;
3832 punch_map.m_flags = 0;
3833
3834 /* Check to see if the extent needs to be split */
3835 if (punch_map.m_len != ee_len ||
3836 punch_map.m_lblk != ee_block) {
3837
3838 ret = ext4_split_extent(handle, inode,
3839 path, &punch_map, 0,
3840 EXT4_GET_BLOCKS_PUNCH_OUT_EXT |
3841 EXT4_GET_BLOCKS_PRE_IO);
3842
3843 if (ret < 0) {
3844 err = ret;
3845 goto out2;
3846 }
3847 /*
3848 * find extent for the block at
3849 * the start of the hole
3850 */
3851 ext4_ext_drop_refs(path);
3852 kfree(path);
3853
3854 path = ext4_ext_find_extent(inode,
3855 map->m_lblk, NULL);
3856 if (IS_ERR(path)) {
3857 err = PTR_ERR(path);
3858 path = NULL;
3859 goto out2;
3860 }
3861
3862 depth = ext_depth(inode);
3863 ex = path[depth].p_ext;
3864 ee_len = ext4_ext_get_actual_len(ex);
3865 ee_block = le32_to_cpu(ex->ee_block);
3866 ee_start = ext4_ext_pblock(ex);
3867
3868 }
3869
3870 ext4_ext_mark_uninitialized(ex);
3871
3872 ext4_ext_invalidate_cache(inode);
3873
3874 err = ext4_ext_rm_leaf(handle, inode, path,
3875 &partial_cluster, map->m_lblk,
3876 map->m_lblk + punched_out);
3877
3878 if (!err && path->p_hdr->eh_entries == 0) {
3879 /*
3880 * Punch hole freed all of this sub tree,
3881 * so we need to correct eh_depth
3882 */
3883 err = ext4_ext_get_access(handle, inode, path);
3884 if (err == 0) {
3885 ext_inode_hdr(inode)->eh_depth = 0;
3886 ext_inode_hdr(inode)->eh_max =
3887 cpu_to_le16(ext4_ext_space_root(
3888 inode, 0));
3889
3890 err = ext4_ext_dirty(
3891 handle, inode, path);
3892 }
3893 }
3894
3895 goto out2;
3896 }
3897 }
3898
3899 if ((sbi->s_cluster_ratio > 1) &&
3900 ext4_find_delalloc_cluster(inode, map->m_lblk, 0))
3901 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
3902
3903 /*
3904 * requested block isn't allocated yet;
3905 * we couldn't try to create block if create flag is zero
3906 */
3907 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
3908 /*
3909 * put just found gap into cache to speed up
3910 * subsequent requests
3911 */
3912 ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
3913 goto out2;
3914 }
3915
3916 /*
3917 * Okay, we need to do block allocation.
3918 */
3919 map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
3920 newex.ee_block = cpu_to_le32(map->m_lblk);
3921 cluster_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
3922
3923 /*
3924 * If we are doing bigalloc, check to see if the extent returned
3925 * by ext4_ext_find_extent() implies a cluster we can use.
3926 */
3927 if (cluster_offset && ex &&
3928 get_implied_cluster_alloc(inode->i_sb, map, ex, path)) {
3929 ar.len = allocated = map->m_len;
3930 newblock = map->m_pblk;
3931 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
3932 goto got_allocated_blocks;
3933 }
3934
3935 /* find neighbour allocated blocks */
3936 ar.lleft = map->m_lblk;
3937 err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
3938 if (err)
3939 goto out2;
3940 ar.lright = map->m_lblk;
3941 ex2 = NULL;
3942 err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright, &ex2);
3943 if (err)
3944 goto out2;
3945
3946 /* Check if the extent after searching to the right implies a
3947 * cluster we can use. */
3948 if ((sbi->s_cluster_ratio > 1) && ex2 &&
3949 get_implied_cluster_alloc(inode->i_sb, map, ex2, path)) {
3950 ar.len = allocated = map->m_len;
3951 newblock = map->m_pblk;
3952 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
3953 goto got_allocated_blocks;
3954 }
3955
3956 /*
3957 * See if request is beyond maximum number of blocks we can have in
3958 * a single extent. For an initialized extent this limit is
3959 * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
3960 * EXT_UNINIT_MAX_LEN.
3961 */
3962 if (map->m_len > EXT_INIT_MAX_LEN &&
3963 !(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
3964 map->m_len = EXT_INIT_MAX_LEN;
3965 else if (map->m_len > EXT_UNINIT_MAX_LEN &&
3966 (flags & EXT4_GET_BLOCKS_UNINIT_EXT))
3967 map->m_len = EXT_UNINIT_MAX_LEN;
3968
3969 /* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
3970 newex.ee_len = cpu_to_le16(map->m_len);
3971 err = ext4_ext_check_overlap(sbi, inode, &newex, path);
3972 if (err)
3973 allocated = ext4_ext_get_actual_len(&newex);
3974 else
3975 allocated = map->m_len;
3976
3977 /* allocate new block */
3978 ar.inode = inode;
3979 ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
3980 ar.logical = map->m_lblk;
3981 /*
3982 * We calculate the offset from the beginning of the cluster
3983 * for the logical block number, since when we allocate a
3984 * physical cluster, the physical block should start at the
3985 * same offset from the beginning of the cluster. This is
3986 * needed so that future calls to get_implied_cluster_alloc()
3987 * work correctly.
3988 */
3989 offset = map->m_lblk & (sbi->s_cluster_ratio - 1);
3990 ar.len = EXT4_NUM_B2C(sbi, offset+allocated);
3991 ar.goal -= offset;
3992 ar.logical -= offset;
3993 if (S_ISREG(inode->i_mode))
3994 ar.flags = EXT4_MB_HINT_DATA;
3995 else
3996 /* disable in-core preallocation for non-regular files */
3997 ar.flags = 0;
3998 if (flags & EXT4_GET_BLOCKS_NO_NORMALIZE)
3999 ar.flags |= EXT4_MB_HINT_NOPREALLOC;
4000 newblock = ext4_mb_new_blocks(handle, &ar, &err);
4001 if (!newblock)
4002 goto out2;
4003 ext_debug("allocate new block: goal %llu, found %llu/%u\n",
4004 ar.goal, newblock, allocated);
4005 free_on_err = 1;
4006 allocated_clusters = ar.len;
4007 ar.len = EXT4_C2B(sbi, ar.len) - offset;
4008 if (ar.len > allocated)
4009 ar.len = allocated;
4010
4011got_allocated_blocks:
4012 /* try to insert new extent into found leaf and return */
4013 ext4_ext_store_pblock(&newex, newblock + offset);
4014 newex.ee_len = cpu_to_le16(ar.len);
4015 /* Mark uninitialized */
4016 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT){
4017 ext4_ext_mark_uninitialized(&newex);
4018 /*
4019 * io_end structure was created for every IO write to an
4020 * uninitialized extent. To avoid unnecessary conversion,
4021 * here we flag the IO that really needs the conversion.
4022 * For non asycn direct IO case, flag the inode state
4023 * that we need to perform conversion when IO is done.
4024 */
4025 if ((flags & EXT4_GET_BLOCKS_PRE_IO)) {
4026 if (io && !(io->flag & EXT4_IO_END_UNWRITTEN)) {
4027 io->flag = EXT4_IO_END_UNWRITTEN;
4028 atomic_inc(&EXT4_I(inode)->i_aiodio_unwritten);
4029 } else
4030 ext4_set_inode_state(inode,
4031 EXT4_STATE_DIO_UNWRITTEN);
4032 }
4033 if (ext4_should_dioread_nolock(inode))
4034 map->m_flags |= EXT4_MAP_UNINIT;
4035 }
4036
4037 err = check_eofblocks_fl(handle, inode, map->m_lblk, path, ar.len);
4038 if (!err)
4039 err = ext4_ext_insert_extent(handle, inode, path,
4040 &newex, flags);
4041 if (err && free_on_err) {
4042 int fb_flags = flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE ?
4043 EXT4_FREE_BLOCKS_NO_QUOT_UPDATE : 0;
4044 /* free data blocks we just allocated */
4045 /* not a good idea to call discard here directly,
4046 * but otherwise we'd need to call it every free() */
4047 ext4_discard_preallocations(inode);
4048 ext4_free_blocks(handle, inode, NULL, ext4_ext_pblock(&newex),
4049 ext4_ext_get_actual_len(&newex), fb_flags);
4050 goto out2;
4051 }
4052
4053 /* previous routine could use block we allocated */
4054 newblock = ext4_ext_pblock(&newex);
4055 allocated = ext4_ext_get_actual_len(&newex);
4056 if (allocated > map->m_len)
4057 allocated = map->m_len;
4058 map->m_flags |= EXT4_MAP_NEW;
4059
4060 /*
4061 * Update reserved blocks/metadata blocks after successful
4062 * block allocation which had been deferred till now.
4063 */
4064 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
4065 /*
4066 * Check how many clusters we had reserved this allocted range.
4067 */
4068 reserved_clusters = get_reserved_cluster_alloc(inode,
4069 map->m_lblk, allocated);
4070 if (map->m_flags & EXT4_MAP_FROM_CLUSTER) {
4071 if (reserved_clusters) {
4072 /*
4073 * We have clusters reserved for this range.
4074 * But since we are not doing actual allocation
4075 * and are simply using blocks from previously
4076 * allocated cluster, we should release the
4077 * reservation and not claim quota.
4078 */
4079 ext4_da_update_reserve_space(inode,
4080 reserved_clusters, 0);
4081 }
4082 } else {
4083 BUG_ON(allocated_clusters < reserved_clusters);
4084 /* We will claim quota for all newly allocated blocks.*/
4085 ext4_da_update_reserve_space(inode, allocated_clusters,
4086 1);
4087 if (reserved_clusters < allocated_clusters) {
4088 struct ext4_inode_info *ei = EXT4_I(inode);
4089 int reservation = allocated_clusters -
4090 reserved_clusters;
4091 /*
4092 * It seems we claimed few clusters outside of
4093 * the range of this allocation. We should give
4094 * it back to the reservation pool. This can
4095 * happen in the following case:
4096 *
4097 * * Suppose s_cluster_ratio is 4 (i.e., each
4098 * cluster has 4 blocks. Thus, the clusters
4099 * are [0-3],[4-7],[8-11]...
4100 * * First comes delayed allocation write for
4101 * logical blocks 10 & 11. Since there were no
4102 * previous delayed allocated blocks in the
4103 * range [8-11], we would reserve 1 cluster
4104 * for this write.
4105 * * Next comes write for logical blocks 3 to 8.
4106 * In this case, we will reserve 2 clusters
4107 * (for [0-3] and [4-7]; and not for [8-11] as
4108 * that range has a delayed allocated blocks.
4109 * Thus total reserved clusters now becomes 3.
4110 * * Now, during the delayed allocation writeout
4111 * time, we will first write blocks [3-8] and
4112 * allocate 3 clusters for writing these
4113 * blocks. Also, we would claim all these
4114 * three clusters above.
4115 * * Now when we come here to writeout the
4116 * blocks [10-11], we would expect to claim
4117 * the reservation of 1 cluster we had made
4118 * (and we would claim it since there are no
4119 * more delayed allocated blocks in the range
4120 * [8-11]. But our reserved cluster count had
4121 * already gone to 0.
4122 *
4123 * Thus, at the step 4 above when we determine
4124 * that there are still some unwritten delayed
4125 * allocated blocks outside of our current
4126 * block range, we should increment the
4127 * reserved clusters count so that when the
4128 * remaining blocks finally gets written, we
4129 * could claim them.
4130 */
4131 dquot_reserve_block(inode,
4132 EXT4_C2B(sbi, reservation));
4133 spin_lock(&ei->i_block_reservation_lock);
4134 ei->i_reserved_data_blocks += reservation;
4135 spin_unlock(&ei->i_block_reservation_lock);
4136 }
4137 }
4138 }
4139
4140 /*
4141 * Cache the extent and update transaction to commit on fdatasync only
4142 * when it is _not_ an uninitialized extent.
4143 */
4144 if ((flags & EXT4_GET_BLOCKS_UNINIT_EXT) == 0) {
4145 ext4_ext_put_in_cache(inode, map->m_lblk, allocated, newblock);
4146 ext4_update_inode_fsync_trans(handle, inode, 1);
4147 } else
4148 ext4_update_inode_fsync_trans(handle, inode, 0);
4149out:
4150 if (allocated > map->m_len)
4151 allocated = map->m_len;
4152 ext4_ext_show_leaf(inode, path);
4153 map->m_flags |= EXT4_MAP_MAPPED;
4154 map->m_pblk = newblock;
4155 map->m_len = allocated;
4156out2:
4157 if (path) {
4158 ext4_ext_drop_refs(path);
4159 kfree(path);
4160 }
4161 trace_ext4_ext_map_blocks_exit(inode, map->m_lblk,
4162 newblock, map->m_len, err ? err : allocated);
4163
4164 result = (flags & EXT4_GET_BLOCKS_PUNCH_OUT_EXT) ?
4165 punched_out : allocated;
4166
4167 return err ? err : result;
4168}
4169
4170void ext4_ext_truncate(struct inode *inode)
4171{
4172 struct address_space *mapping = inode->i_mapping;
4173 struct super_block *sb = inode->i_sb;
4174 ext4_lblk_t last_block;
4175 handle_t *handle;
4176 loff_t page_len;
4177 int err = 0;
4178
4179 /*
4180 * finish any pending end_io work so we won't run the risk of
4181 * converting any truncated blocks to initialized later
4182 */
4183 ext4_flush_completed_IO(inode);
4184
4185 /*
4186 * probably first extent we're gonna free will be last in block
4187 */
4188 err = ext4_writepage_trans_blocks(inode);
4189 handle = ext4_journal_start(inode, err);
4190 if (IS_ERR(handle))
4191 return;
4192
4193 if (inode->i_size % PAGE_CACHE_SIZE != 0) {
4194 page_len = PAGE_CACHE_SIZE -
4195 (inode->i_size & (PAGE_CACHE_SIZE - 1));
4196
4197 err = ext4_discard_partial_page_buffers(handle,
4198 mapping, inode->i_size, page_len, 0);
4199
4200 if (err)
4201 goto out_stop;
4202 }
4203
4204 if (ext4_orphan_add(handle, inode))
4205 goto out_stop;
4206
4207 down_write(&EXT4_I(inode)->i_data_sem);
4208 ext4_ext_invalidate_cache(inode);
4209
4210 ext4_discard_preallocations(inode);
4211
4212 /*
4213 * TODO: optimization is possible here.
4214 * Probably we need not scan at all,
4215 * because page truncation is enough.
4216 */
4217
4218 /* we have to know where to truncate from in crash case */
4219 EXT4_I(inode)->i_disksize = inode->i_size;
4220 ext4_mark_inode_dirty(handle, inode);
4221
4222 last_block = (inode->i_size + sb->s_blocksize - 1)
4223 >> EXT4_BLOCK_SIZE_BITS(sb);
4224 err = ext4_ext_remove_space(inode, last_block);
4225
4226 /* In a multi-transaction truncate, we only make the final
4227 * transaction synchronous.
4228 */
4229 if (IS_SYNC(inode))
4230 ext4_handle_sync(handle);
4231
4232 up_write(&EXT4_I(inode)->i_data_sem);
4233
4234out_stop:
4235 /*
4236 * If this was a simple ftruncate() and the file will remain alive,
4237 * then we need to clear up the orphan record which we created above.
4238 * However, if this was a real unlink then we were called by
4239 * ext4_delete_inode(), and we allow that function to clean up the
4240 * orphan info for us.
4241 */
4242 if (inode->i_nlink)
4243 ext4_orphan_del(handle, inode);
4244
4245 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
4246 ext4_mark_inode_dirty(handle, inode);
4247 ext4_journal_stop(handle);
4248}
4249
4250static void ext4_falloc_update_inode(struct inode *inode,
4251 int mode, loff_t new_size, int update_ctime)
4252{
4253 struct timespec now;
4254
4255 if (update_ctime) {
4256 now = current_fs_time(inode->i_sb);
4257 if (!timespec_equal(&inode->i_ctime, &now))
4258 inode->i_ctime = now;
4259 }
4260 /*
4261 * Update only when preallocation was requested beyond
4262 * the file size.
4263 */
4264 if (!(mode & FALLOC_FL_KEEP_SIZE)) {
4265 if (new_size > i_size_read(inode))
4266 i_size_write(inode, new_size);
4267 if (new_size > EXT4_I(inode)->i_disksize)
4268 ext4_update_i_disksize(inode, new_size);
4269 } else {
4270 /*
4271 * Mark that we allocate beyond EOF so the subsequent truncate
4272 * can proceed even if the new size is the same as i_size.
4273 */
4274 if (new_size > i_size_read(inode))
4275 ext4_set_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
4276 }
4277
4278}
4279
4280/*
4281 * preallocate space for a file. This implements ext4's fallocate file
4282 * operation, which gets called from sys_fallocate system call.
4283 * For block-mapped files, posix_fallocate should fall back to the method
4284 * of writing zeroes to the required new blocks (the same behavior which is
4285 * expected for file systems which do not support fallocate() system call).
4286 */
4287long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
4288{
4289 struct inode *inode = file->f_path.dentry->d_inode;
4290 handle_t *handle;
4291 loff_t new_size;
4292 unsigned int max_blocks;
4293 int ret = 0;
4294 int ret2 = 0;
4295 int retries = 0;
4296 struct ext4_map_blocks map;
4297 unsigned int credits, blkbits = inode->i_blkbits;
4298
4299 /*
4300 * currently supporting (pre)allocate mode for extent-based
4301 * files _only_
4302 */
4303 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
4304 return -EOPNOTSUPP;
4305
4306 /* Return error if mode is not supported */
4307 if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
4308 return -EOPNOTSUPP;
4309
4310 if (mode & FALLOC_FL_PUNCH_HOLE)
4311 return ext4_punch_hole(file, offset, len);
4312
4313 trace_ext4_fallocate_enter(inode, offset, len, mode);
4314 map.m_lblk = offset >> blkbits;
4315 /*
4316 * We can't just convert len to max_blocks because
4317 * If blocksize = 4096 offset = 3072 and len = 2048
4318 */
4319 max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
4320 - map.m_lblk;
4321 /*
4322 * credits to insert 1 extent into extent tree
4323 */
4324 credits = ext4_chunk_trans_blocks(inode, max_blocks);
4325 mutex_lock(&inode->i_mutex);
4326 ret = inode_newsize_ok(inode, (len + offset));
4327 if (ret) {
4328 mutex_unlock(&inode->i_mutex);
4329 trace_ext4_fallocate_exit(inode, offset, max_blocks, ret);
4330 return ret;
4331 }
4332retry:
4333 while (ret >= 0 && ret < max_blocks) {
4334 map.m_lblk = map.m_lblk + ret;
4335 map.m_len = max_blocks = max_blocks - ret;
4336 handle = ext4_journal_start(inode, credits);
4337 if (IS_ERR(handle)) {
4338 ret = PTR_ERR(handle);
4339 break;
4340 }
4341 ret = ext4_map_blocks(handle, inode, &map,
4342 EXT4_GET_BLOCKS_CREATE_UNINIT_EXT |
4343 EXT4_GET_BLOCKS_NO_NORMALIZE);
4344 if (ret <= 0) {
4345#ifdef EXT4FS_DEBUG
4346 WARN_ON(ret <= 0);
4347 printk(KERN_ERR "%s: ext4_ext_map_blocks "
4348 "returned error inode#%lu, block=%u, "
4349 "max_blocks=%u", __func__,
4350 inode->i_ino, map.m_lblk, max_blocks);
4351#endif
4352 ext4_mark_inode_dirty(handle, inode);
4353 ret2 = ext4_journal_stop(handle);
4354 break;
4355 }
4356 if ((map.m_lblk + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
4357 blkbits) >> blkbits))
4358 new_size = offset + len;
4359 else
4360 new_size = ((loff_t) map.m_lblk + ret) << blkbits;
4361
4362 ext4_falloc_update_inode(inode, mode, new_size,
4363 (map.m_flags & EXT4_MAP_NEW));
4364 ext4_mark_inode_dirty(handle, inode);
4365 ret2 = ext4_journal_stop(handle);
4366 if (ret2)
4367 break;
4368 }
4369 if (ret == -ENOSPC &&
4370 ext4_should_retry_alloc(inode->i_sb, &retries)) {
4371 ret = 0;
4372 goto retry;
4373 }
4374 mutex_unlock(&inode->i_mutex);
4375 trace_ext4_fallocate_exit(inode, offset, max_blocks,
4376 ret > 0 ? ret2 : ret);
4377 return ret > 0 ? ret2 : ret;
4378}
4379
4380/*
4381 * This function convert a range of blocks to written extents
4382 * The caller of this function will pass the start offset and the size.
4383 * all unwritten extents within this range will be converted to
4384 * written extents.
4385 *
4386 * This function is called from the direct IO end io call back
4387 * function, to convert the fallocated extents after IO is completed.
4388 * Returns 0 on success.
4389 */
4390int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset,
4391 ssize_t len)
4392{
4393 handle_t *handle;
4394 unsigned int max_blocks;
4395 int ret = 0;
4396 int ret2 = 0;
4397 struct ext4_map_blocks map;
4398 unsigned int credits, blkbits = inode->i_blkbits;
4399
4400 map.m_lblk = offset >> blkbits;
4401 /*
4402 * We can't just convert len to max_blocks because
4403 * If blocksize = 4096 offset = 3072 and len = 2048
4404 */
4405 max_blocks = ((EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) -
4406 map.m_lblk);
4407 /*
4408 * credits to insert 1 extent into extent tree
4409 */
4410 credits = ext4_chunk_trans_blocks(inode, max_blocks);
4411 while (ret >= 0 && ret < max_blocks) {
4412 map.m_lblk += ret;
4413 map.m_len = (max_blocks -= ret);
4414 handle = ext4_journal_start(inode, credits);
4415 if (IS_ERR(handle)) {
4416 ret = PTR_ERR(handle);
4417 break;
4418 }
4419 ret = ext4_map_blocks(handle, inode, &map,
4420 EXT4_GET_BLOCKS_IO_CONVERT_EXT);
4421 if (ret <= 0) {
4422 WARN_ON(ret <= 0);
4423 printk(KERN_ERR "%s: ext4_ext_map_blocks "
4424 "returned error inode#%lu, block=%u, "
4425 "max_blocks=%u", __func__,
4426 inode->i_ino, map.m_lblk, map.m_len);
4427 }
4428 ext4_mark_inode_dirty(handle, inode);
4429 ret2 = ext4_journal_stop(handle);
4430 if (ret <= 0 || ret2 )
4431 break;
4432 }
4433 return ret > 0 ? ret2 : ret;
4434}
4435
4436/*
4437 * Callback function called for each extent to gather FIEMAP information.
4438 */
4439static int ext4_ext_fiemap_cb(struct inode *inode, ext4_lblk_t next,
4440 struct ext4_ext_cache *newex, struct ext4_extent *ex,
4441 void *data)
4442{
4443 __u64 logical;
4444 __u64 physical;
4445 __u64 length;
4446 __u32 flags = 0;
4447 int ret = 0;
4448 struct fiemap_extent_info *fieinfo = data;
4449 unsigned char blksize_bits;
4450
4451 blksize_bits = inode->i_sb->s_blocksize_bits;
4452 logical = (__u64)newex->ec_block << blksize_bits;
4453
4454 if (newex->ec_start == 0) {
4455 /*
4456 * No extent in extent-tree contains block @newex->ec_start,
4457 * then the block may stay in 1)a hole or 2)delayed-extent.
4458 *
4459 * Holes or delayed-extents are processed as follows.
4460 * 1. lookup dirty pages with specified range in pagecache.
4461 * If no page is got, then there is no delayed-extent and
4462 * return with EXT_CONTINUE.
4463 * 2. find the 1st mapped buffer,
4464 * 3. check if the mapped buffer is both in the request range
4465 * and a delayed buffer. If not, there is no delayed-extent,
4466 * then return.
4467 * 4. a delayed-extent is found, the extent will be collected.
4468 */
4469 ext4_lblk_t end = 0;
4470 pgoff_t last_offset;
4471 pgoff_t offset;
4472 pgoff_t index;
4473 pgoff_t start_index = 0;
4474 struct page **pages = NULL;
4475 struct buffer_head *bh = NULL;
4476 struct buffer_head *head = NULL;
4477 unsigned int nr_pages = PAGE_SIZE / sizeof(struct page *);
4478
4479 pages = kmalloc(PAGE_SIZE, GFP_KERNEL);
4480 if (pages == NULL)
4481 return -ENOMEM;
4482
4483 offset = logical >> PAGE_SHIFT;
4484repeat:
4485 last_offset = offset;
4486 head = NULL;
4487 ret = find_get_pages_tag(inode->i_mapping, &offset,
4488 PAGECACHE_TAG_DIRTY, nr_pages, pages);
4489
4490 if (!(flags & FIEMAP_EXTENT_DELALLOC)) {
4491 /* First time, try to find a mapped buffer. */
4492 if (ret == 0) {
4493out:
4494 for (index = 0; index < ret; index++)
4495 page_cache_release(pages[index]);
4496 /* just a hole. */
4497 kfree(pages);
4498 return EXT_CONTINUE;
4499 }
4500 index = 0;
4501
4502next_page:
4503 /* Try to find the 1st mapped buffer. */
4504 end = ((__u64)pages[index]->index << PAGE_SHIFT) >>
4505 blksize_bits;
4506 if (!page_has_buffers(pages[index]))
4507 goto out;
4508 head = page_buffers(pages[index]);
4509 if (!head)
4510 goto out;
4511
4512 index++;
4513 bh = head;
4514 do {
4515 if (end >= newex->ec_block +
4516 newex->ec_len)
4517 /* The buffer is out of
4518 * the request range.
4519 */
4520 goto out;
4521
4522 if (buffer_mapped(bh) &&
4523 end >= newex->ec_block) {
4524 start_index = index - 1;
4525 /* get the 1st mapped buffer. */
4526 goto found_mapped_buffer;
4527 }
4528
4529 bh = bh->b_this_page;
4530 end++;
4531 } while (bh != head);
4532
4533 /* No mapped buffer in the range found in this page,
4534 * We need to look up next page.
4535 */
4536 if (index >= ret) {
4537 /* There is no page left, but we need to limit
4538 * newex->ec_len.
4539 */
4540 newex->ec_len = end - newex->ec_block;
4541 goto out;
4542 }
4543 goto next_page;
4544 } else {
4545 /*Find contiguous delayed buffers. */
4546 if (ret > 0 && pages[0]->index == last_offset)
4547 head = page_buffers(pages[0]);
4548 bh = head;
4549 index = 1;
4550 start_index = 0;
4551 }
4552
4553found_mapped_buffer:
4554 if (bh != NULL && buffer_delay(bh)) {
4555 /* 1st or contiguous delayed buffer found. */
4556 if (!(flags & FIEMAP_EXTENT_DELALLOC)) {
4557 /*
4558 * 1st delayed buffer found, record
4559 * the start of extent.
4560 */
4561 flags |= FIEMAP_EXTENT_DELALLOC;
4562 newex->ec_block = end;
4563 logical = (__u64)end << blksize_bits;
4564 }
4565 /* Find contiguous delayed buffers. */
4566 do {
4567 if (!buffer_delay(bh))
4568 goto found_delayed_extent;
4569 bh = bh->b_this_page;
4570 end++;
4571 } while (bh != head);
4572
4573 for (; index < ret; index++) {
4574 if (!page_has_buffers(pages[index])) {
4575 bh = NULL;
4576 break;
4577 }
4578 head = page_buffers(pages[index]);
4579 if (!head) {
4580 bh = NULL;
4581 break;
4582 }
4583
4584 if (pages[index]->index !=
4585 pages[start_index]->index + index
4586 - start_index) {
4587 /* Blocks are not contiguous. */
4588 bh = NULL;
4589 break;
4590 }
4591 bh = head;
4592 do {
4593 if (!buffer_delay(bh))
4594 /* Delayed-extent ends. */
4595 goto found_delayed_extent;
4596 bh = bh->b_this_page;
4597 end++;
4598 } while (bh != head);
4599 }
4600 } else if (!(flags & FIEMAP_EXTENT_DELALLOC))
4601 /* a hole found. */
4602 goto out;
4603
4604found_delayed_extent:
4605 newex->ec_len = min(end - newex->ec_block,
4606 (ext4_lblk_t)EXT_INIT_MAX_LEN);
4607 if (ret == nr_pages && bh != NULL &&
4608 newex->ec_len < EXT_INIT_MAX_LEN &&
4609 buffer_delay(bh)) {
4610 /* Have not collected an extent and continue. */
4611 for (index = 0; index < ret; index++)
4612 page_cache_release(pages[index]);
4613 goto repeat;
4614 }
4615
4616 for (index = 0; index < ret; index++)
4617 page_cache_release(pages[index]);
4618 kfree(pages);
4619 }
4620
4621 physical = (__u64)newex->ec_start << blksize_bits;
4622 length = (__u64)newex->ec_len << blksize_bits;
4623
4624 if (ex && ext4_ext_is_uninitialized(ex))
4625 flags |= FIEMAP_EXTENT_UNWRITTEN;
4626
4627 if (next == EXT_MAX_BLOCKS)
4628 flags |= FIEMAP_EXTENT_LAST;
4629
4630 ret = fiemap_fill_next_extent(fieinfo, logical, physical,
4631 length, flags);
4632 if (ret < 0)
4633 return ret;
4634 if (ret == 1)
4635 return EXT_BREAK;
4636 return EXT_CONTINUE;
4637}
4638/* fiemap flags we can handle specified here */
4639#define EXT4_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
4640
4641static int ext4_xattr_fiemap(struct inode *inode,
4642 struct fiemap_extent_info *fieinfo)
4643{
4644 __u64 physical = 0;
4645 __u64 length;
4646 __u32 flags = FIEMAP_EXTENT_LAST;
4647 int blockbits = inode->i_sb->s_blocksize_bits;
4648 int error = 0;
4649
4650 /* in-inode? */
4651 if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
4652 struct ext4_iloc iloc;
4653 int offset; /* offset of xattr in inode */
4654
4655 error = ext4_get_inode_loc(inode, &iloc);
4656 if (error)
4657 return error;
4658 physical = iloc.bh->b_blocknr << blockbits;
4659 offset = EXT4_GOOD_OLD_INODE_SIZE +
4660 EXT4_I(inode)->i_extra_isize;
4661 physical += offset;
4662 length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
4663 flags |= FIEMAP_EXTENT_DATA_INLINE;
4664 brelse(iloc.bh);
4665 } else { /* external block */
4666 physical = EXT4_I(inode)->i_file_acl << blockbits;
4667 length = inode->i_sb->s_blocksize;
4668 }
4669
4670 if (physical)
4671 error = fiemap_fill_next_extent(fieinfo, 0, physical,
4672 length, flags);
4673 return (error < 0 ? error : 0);
4674}
4675
4676/*
4677 * ext4_ext_punch_hole
4678 *
4679 * Punches a hole of "length" bytes in a file starting
4680 * at byte "offset"
4681 *
4682 * @inode: The inode of the file to punch a hole in
4683 * @offset: The starting byte offset of the hole
4684 * @length: The length of the hole
4685 *
4686 * Returns the number of blocks removed or negative on err
4687 */
4688int ext4_ext_punch_hole(struct file *file, loff_t offset, loff_t length)
4689{
4690 struct inode *inode = file->f_path.dentry->d_inode;
4691 struct super_block *sb = inode->i_sb;
4692 struct ext4_ext_cache cache_ex;
4693 ext4_lblk_t first_block, last_block, num_blocks, iblock, max_blocks;
4694 struct address_space *mapping = inode->i_mapping;
4695 struct ext4_map_blocks map;
4696 handle_t *handle;
4697 loff_t first_page, last_page, page_len;
4698 loff_t first_page_offset, last_page_offset;
4699 int ret, credits, blocks_released, err = 0;
4700
4701 /* No need to punch hole beyond i_size */
4702 if (offset >= inode->i_size)
4703 return 0;
4704
4705 /*
4706 * If the hole extends beyond i_size, set the hole
4707 * to end after the page that contains i_size
4708 */
4709 if (offset + length > inode->i_size) {
4710 length = inode->i_size +
4711 PAGE_CACHE_SIZE - (inode->i_size & (PAGE_CACHE_SIZE - 1)) -
4712 offset;
4713 }
4714
4715 first_block = (offset + sb->s_blocksize - 1) >>
4716 EXT4_BLOCK_SIZE_BITS(sb);
4717 last_block = (offset + length) >> EXT4_BLOCK_SIZE_BITS(sb);
4718
4719 first_page = (offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
4720 last_page = (offset + length) >> PAGE_CACHE_SHIFT;
4721
4722 first_page_offset = first_page << PAGE_CACHE_SHIFT;
4723 last_page_offset = last_page << PAGE_CACHE_SHIFT;
4724
4725 /*
4726 * Write out all dirty pages to avoid race conditions
4727 * Then release them.
4728 */
4729 if (mapping->nrpages && mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) {
4730 err = filemap_write_and_wait_range(mapping,
4731 offset, offset + length - 1);
4732
4733 if (err)
4734 return err;
4735 }
4736
4737 /* Now release the pages */
4738 if (last_page_offset > first_page_offset) {
4739 truncate_inode_pages_range(mapping, first_page_offset,
4740 last_page_offset-1);
4741 }
4742
4743 /* finish any pending end_io work */
4744 ext4_flush_completed_IO(inode);
4745
4746 credits = ext4_writepage_trans_blocks(inode);
4747 handle = ext4_journal_start(inode, credits);
4748 if (IS_ERR(handle))
4749 return PTR_ERR(handle);
4750
4751 err = ext4_orphan_add(handle, inode);
4752 if (err)
4753 goto out;
4754
4755 /*
4756 * Now we need to zero out the non-page-aligned data in the
4757 * pages at the start and tail of the hole, and unmap the buffer
4758 * heads for the block aligned regions of the page that were
4759 * completely zeroed.
4760 */
4761 if (first_page > last_page) {
4762 /*
4763 * If the file space being truncated is contained within a page
4764 * just zero out and unmap the middle of that page
4765 */
4766 err = ext4_discard_partial_page_buffers(handle,
4767 mapping, offset, length, 0);
4768
4769 if (err)
4770 goto out;
4771 } else {
4772 /*
4773 * zero out and unmap the partial page that contains
4774 * the start of the hole
4775 */
4776 page_len = first_page_offset - offset;
4777 if (page_len > 0) {
4778 err = ext4_discard_partial_page_buffers(handle, mapping,
4779 offset, page_len, 0);
4780 if (err)
4781 goto out;
4782 }
4783
4784 /*
4785 * zero out and unmap the partial page that contains
4786 * the end of the hole
4787 */
4788 page_len = offset + length - last_page_offset;
4789 if (page_len > 0) {
4790 err = ext4_discard_partial_page_buffers(handle, mapping,
4791 last_page_offset, page_len, 0);
4792 if (err)
4793 goto out;
4794 }
4795 }
4796
4797
4798 /*
4799 * If i_size is contained in the last page, we need to
4800 * unmap and zero the partial page after i_size
4801 */
4802 if (inode->i_size >> PAGE_CACHE_SHIFT == last_page &&
4803 inode->i_size % PAGE_CACHE_SIZE != 0) {
4804
4805 page_len = PAGE_CACHE_SIZE -
4806 (inode->i_size & (PAGE_CACHE_SIZE - 1));
4807
4808 if (page_len > 0) {
4809 err = ext4_discard_partial_page_buffers(handle,
4810 mapping, inode->i_size, page_len, 0);
4811
4812 if (err)
4813 goto out;
4814 }
4815 }
4816
4817 /* If there are no blocks to remove, return now */
4818 if (first_block >= last_block)
4819 goto out;
4820
4821 down_write(&EXT4_I(inode)->i_data_sem);
4822 ext4_ext_invalidate_cache(inode);
4823 ext4_discard_preallocations(inode);
4824
4825 /*
4826 * Loop over all the blocks and identify blocks
4827 * that need to be punched out
4828 */
4829 iblock = first_block;
4830 blocks_released = 0;
4831 while (iblock < last_block) {
4832 max_blocks = last_block - iblock;
4833 num_blocks = 1;
4834 memset(&map, 0, sizeof(map));
4835 map.m_lblk = iblock;
4836 map.m_len = max_blocks;
4837 ret = ext4_ext_map_blocks(handle, inode, &map,
4838 EXT4_GET_BLOCKS_PUNCH_OUT_EXT);
4839
4840 if (ret > 0) {
4841 blocks_released += ret;
4842 num_blocks = ret;
4843 } else if (ret == 0) {
4844 /*
4845 * If map blocks could not find the block,
4846 * then it is in a hole. If the hole was
4847 * not already cached, then map blocks should
4848 * put it in the cache. So we can get the hole
4849 * out of the cache
4850 */
4851 memset(&cache_ex, 0, sizeof(cache_ex));
4852 if ((ext4_ext_check_cache(inode, iblock, &cache_ex)) &&
4853 !cache_ex.ec_start) {
4854
4855 /* The hole is cached */
4856 num_blocks = cache_ex.ec_block +
4857 cache_ex.ec_len - iblock;
4858
4859 } else {
4860 /* The block could not be identified */
4861 err = -EIO;
4862 break;
4863 }
4864 } else {
4865 /* Map blocks error */
4866 err = ret;
4867 break;
4868 }
4869
4870 if (num_blocks == 0) {
4871 /* This condition should never happen */
4872 ext_debug("Block lookup failed");
4873 err = -EIO;
4874 break;
4875 }
4876
4877 iblock += num_blocks;
4878 }
4879
4880 if (blocks_released > 0) {
4881 ext4_ext_invalidate_cache(inode);
4882 ext4_discard_preallocations(inode);
4883 }
4884
4885 if (IS_SYNC(inode))
4886 ext4_handle_sync(handle);
4887
4888 up_write(&EXT4_I(inode)->i_data_sem);
4889
4890out:
4891 ext4_orphan_del(handle, inode);
4892 inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
4893 ext4_mark_inode_dirty(handle, inode);
4894 ext4_journal_stop(handle);
4895 return err;
4896}
4897int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
4898 __u64 start, __u64 len)
4899{
4900 ext4_lblk_t start_blk;
4901 int error = 0;
4902
4903 /* fallback to generic here if not in extents fmt */
4904 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
4905 return generic_block_fiemap(inode, fieinfo, start, len,
4906 ext4_get_block);
4907
4908 if (fiemap_check_flags(fieinfo, EXT4_FIEMAP_FLAGS))
4909 return -EBADR;
4910
4911 if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
4912 error = ext4_xattr_fiemap(inode, fieinfo);
4913 } else {
4914 ext4_lblk_t len_blks;
4915 __u64 last_blk;
4916
4917 start_blk = start >> inode->i_sb->s_blocksize_bits;
4918 last_blk = (start + len - 1) >> inode->i_sb->s_blocksize_bits;
4919 if (last_blk >= EXT_MAX_BLOCKS)
4920 last_blk = EXT_MAX_BLOCKS-1;
4921 len_blks = ((ext4_lblk_t) last_blk) - start_blk + 1;
4922
4923 /*
4924 * Walk the extent tree gathering extent information.
4925 * ext4_ext_fiemap_cb will push extents back to user.
4926 */
4927 error = ext4_ext_walk_space(inode, start_blk, len_blks,
4928 ext4_ext_fiemap_cb, fieinfo);
4929 }
4930
4931 return error;
4932}