Commit | Line | Data |
---|---|---|
a86c6181 AT |
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> | |
cd02ff0b | 35 | #include <linux/jbd2.h> |
a86c6181 AT |
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> | |
a2df2a63 | 41 | #include <linux/falloc.h> |
a86c6181 | 42 | #include <asm/uaccess.h> |
3dcf5451 CH |
43 | #include "ext4_jbd2.h" |
44 | #include "ext4_extents.h" | |
a86c6181 AT |
45 | |
46 | ||
d0d856e8 RD |
47 | /* |
48 | * ext_pblock: | |
49 | * combine low and high parts of physical block number into ext4_fsblk_t | |
50 | */ | |
09b88252 | 51 | static ext4_fsblk_t ext_pblock(struct ext4_extent *ex) |
f65e6fba AT |
52 | { |
53 | ext4_fsblk_t block; | |
54 | ||
b377611d | 55 | block = le32_to_cpu(ex->ee_start_lo); |
9b8f1f01 | 56 | block |= ((ext4_fsblk_t) le16_to_cpu(ex->ee_start_hi) << 31) << 1; |
f65e6fba AT |
57 | return block; |
58 | } | |
59 | ||
d0d856e8 RD |
60 | /* |
61 | * idx_pblock: | |
62 | * combine low and high parts of a leaf physical block number into ext4_fsblk_t | |
63 | */ | |
c14c6fd5 | 64 | ext4_fsblk_t idx_pblock(struct ext4_extent_idx *ix) |
f65e6fba AT |
65 | { |
66 | ext4_fsblk_t block; | |
67 | ||
d8dd0b45 | 68 | block = le32_to_cpu(ix->ei_leaf_lo); |
9b8f1f01 | 69 | block |= ((ext4_fsblk_t) le16_to_cpu(ix->ei_leaf_hi) << 31) << 1; |
f65e6fba AT |
70 | return block; |
71 | } | |
72 | ||
d0d856e8 RD |
73 | /* |
74 | * ext4_ext_store_pblock: | |
75 | * stores a large physical block number into an extent struct, | |
76 | * breaking it into parts | |
77 | */ | |
c14c6fd5 | 78 | void ext4_ext_store_pblock(struct ext4_extent *ex, ext4_fsblk_t pb) |
f65e6fba | 79 | { |
b377611d | 80 | ex->ee_start_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff)); |
9b8f1f01 | 81 | ex->ee_start_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff); |
f65e6fba AT |
82 | } |
83 | ||
d0d856e8 RD |
84 | /* |
85 | * ext4_idx_store_pblock: | |
86 | * stores a large physical block number into an index struct, | |
87 | * breaking it into parts | |
88 | */ | |
09b88252 | 89 | static void ext4_idx_store_pblock(struct ext4_extent_idx *ix, ext4_fsblk_t pb) |
f65e6fba | 90 | { |
d8dd0b45 | 91 | ix->ei_leaf_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff)); |
9b8f1f01 | 92 | ix->ei_leaf_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff); |
f65e6fba AT |
93 | } |
94 | ||
9102e4fa | 95 | static int ext4_ext_journal_restart(handle_t *handle, int needed) |
a86c6181 AT |
96 | { |
97 | int err; | |
98 | ||
99 | if (handle->h_buffer_credits > needed) | |
9102e4fa SF |
100 | return 0; |
101 | err = ext4_journal_extend(handle, needed); | |
0123c939 | 102 | if (err <= 0) |
9102e4fa SF |
103 | return err; |
104 | return ext4_journal_restart(handle, needed); | |
a86c6181 AT |
105 | } |
106 | ||
107 | /* | |
108 | * could return: | |
109 | * - EROFS | |
110 | * - ENOMEM | |
111 | */ | |
112 | static int ext4_ext_get_access(handle_t *handle, struct inode *inode, | |
113 | struct ext4_ext_path *path) | |
114 | { | |
115 | if (path->p_bh) { | |
116 | /* path points to block */ | |
117 | return ext4_journal_get_write_access(handle, path->p_bh); | |
118 | } | |
119 | /* path points to leaf/index in inode body */ | |
120 | /* we use in-core data, no need to protect them */ | |
121 | return 0; | |
122 | } | |
123 | ||
124 | /* | |
125 | * could return: | |
126 | * - EROFS | |
127 | * - ENOMEM | |
128 | * - EIO | |
129 | */ | |
130 | static int ext4_ext_dirty(handle_t *handle, struct inode *inode, | |
131 | struct ext4_ext_path *path) | |
132 | { | |
133 | int err; | |
134 | if (path->p_bh) { | |
135 | /* path points to block */ | |
136 | err = ext4_journal_dirty_metadata(handle, path->p_bh); | |
137 | } else { | |
138 | /* path points to leaf/index in inode body */ | |
139 | err = ext4_mark_inode_dirty(handle, inode); | |
140 | } | |
141 | return err; | |
142 | } | |
143 | ||
f65e6fba | 144 | static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode, |
a86c6181 | 145 | struct ext4_ext_path *path, |
725d26d3 | 146 | ext4_lblk_t block) |
a86c6181 AT |
147 | { |
148 | struct ext4_inode_info *ei = EXT4_I(inode); | |
f65e6fba | 149 | ext4_fsblk_t bg_start; |
74d3487f | 150 | ext4_fsblk_t last_block; |
f65e6fba | 151 | ext4_grpblk_t colour; |
a86c6181 AT |
152 | int depth; |
153 | ||
154 | if (path) { | |
155 | struct ext4_extent *ex; | |
156 | depth = path->p_depth; | |
157 | ||
158 | /* try to predict block placement */ | |
7e028976 AM |
159 | ex = path[depth].p_ext; |
160 | if (ex) | |
f65e6fba | 161 | return ext_pblock(ex)+(block-le32_to_cpu(ex->ee_block)); |
a86c6181 | 162 | |
d0d856e8 RD |
163 | /* it looks like index is empty; |
164 | * try to find starting block from index itself */ | |
a86c6181 AT |
165 | if (path[depth].p_bh) |
166 | return path[depth].p_bh->b_blocknr; | |
167 | } | |
168 | ||
169 | /* OK. use inode's group */ | |
170 | bg_start = (ei->i_block_group * EXT4_BLOCKS_PER_GROUP(inode->i_sb)) + | |
171 | le32_to_cpu(EXT4_SB(inode->i_sb)->s_es->s_first_data_block); | |
74d3487f VC |
172 | last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1; |
173 | ||
174 | if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block) | |
175 | colour = (current->pid % 16) * | |
a86c6181 | 176 | (EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16); |
74d3487f VC |
177 | else |
178 | colour = (current->pid % 16) * ((last_block - bg_start) / 16); | |
a86c6181 AT |
179 | return bg_start + colour + block; |
180 | } | |
181 | ||
654b4908 AK |
182 | /* |
183 | * Allocation for a meta data block | |
184 | */ | |
f65e6fba | 185 | static ext4_fsblk_t |
654b4908 | 186 | ext4_ext_new_meta_block(handle_t *handle, struct inode *inode, |
a86c6181 AT |
187 | struct ext4_ext_path *path, |
188 | struct ext4_extent *ex, int *err) | |
189 | { | |
f65e6fba | 190 | ext4_fsblk_t goal, newblock; |
a86c6181 AT |
191 | |
192 | goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block)); | |
7061eba7 | 193 | newblock = ext4_new_meta_block(handle, inode, goal, err); |
a86c6181 AT |
194 | return newblock; |
195 | } | |
196 | ||
09b88252 | 197 | static int ext4_ext_space_block(struct inode *inode) |
a86c6181 AT |
198 | { |
199 | int size; | |
200 | ||
201 | size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header)) | |
202 | / sizeof(struct ext4_extent); | |
bbf2f9fb | 203 | #ifdef AGGRESSIVE_TEST |
a86c6181 AT |
204 | if (size > 6) |
205 | size = 6; | |
206 | #endif | |
207 | return size; | |
208 | } | |
209 | ||
09b88252 | 210 | static int ext4_ext_space_block_idx(struct inode *inode) |
a86c6181 AT |
211 | { |
212 | int size; | |
213 | ||
214 | size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header)) | |
215 | / sizeof(struct ext4_extent_idx); | |
bbf2f9fb | 216 | #ifdef AGGRESSIVE_TEST |
a86c6181 AT |
217 | if (size > 5) |
218 | size = 5; | |
219 | #endif | |
220 | return size; | |
221 | } | |
222 | ||
09b88252 | 223 | static int ext4_ext_space_root(struct inode *inode) |
a86c6181 AT |
224 | { |
225 | int size; | |
226 | ||
227 | size = sizeof(EXT4_I(inode)->i_data); | |
228 | size -= sizeof(struct ext4_extent_header); | |
229 | size /= sizeof(struct ext4_extent); | |
bbf2f9fb | 230 | #ifdef AGGRESSIVE_TEST |
a86c6181 AT |
231 | if (size > 3) |
232 | size = 3; | |
233 | #endif | |
234 | return size; | |
235 | } | |
236 | ||
09b88252 | 237 | static int ext4_ext_space_root_idx(struct inode *inode) |
a86c6181 AT |
238 | { |
239 | int size; | |
240 | ||
241 | size = sizeof(EXT4_I(inode)->i_data); | |
242 | size -= sizeof(struct ext4_extent_header); | |
243 | size /= sizeof(struct ext4_extent_idx); | |
bbf2f9fb | 244 | #ifdef AGGRESSIVE_TEST |
a86c6181 AT |
245 | if (size > 4) |
246 | size = 4; | |
247 | #endif | |
248 | return size; | |
249 | } | |
250 | ||
d2a17637 MC |
251 | /* |
252 | * Calculate the number of metadata blocks needed | |
253 | * to allocate @blocks | |
254 | * Worse case is one block per extent | |
255 | */ | |
256 | int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks) | |
257 | { | |
258 | int lcap, icap, rcap, leafs, idxs, num; | |
259 | int newextents = blocks; | |
260 | ||
261 | rcap = ext4_ext_space_root_idx(inode); | |
262 | lcap = ext4_ext_space_block(inode); | |
263 | icap = ext4_ext_space_block_idx(inode); | |
264 | ||
265 | /* number of new leaf blocks needed */ | |
266 | num = leafs = (newextents + lcap - 1) / lcap; | |
267 | ||
268 | /* | |
269 | * Worse case, we need separate index block(s) | |
270 | * to link all new leaf blocks | |
271 | */ | |
272 | idxs = (leafs + icap - 1) / icap; | |
273 | do { | |
274 | num += idxs; | |
275 | idxs = (idxs + icap - 1) / icap; | |
276 | } while (idxs > rcap); | |
277 | ||
278 | return num; | |
279 | } | |
280 | ||
c29c0ae7 AT |
281 | static int |
282 | ext4_ext_max_entries(struct inode *inode, int depth) | |
283 | { | |
284 | int max; | |
285 | ||
286 | if (depth == ext_depth(inode)) { | |
287 | if (depth == 0) | |
288 | max = ext4_ext_space_root(inode); | |
289 | else | |
290 | max = ext4_ext_space_root_idx(inode); | |
291 | } else { | |
292 | if (depth == 0) | |
293 | max = ext4_ext_space_block(inode); | |
294 | else | |
295 | max = ext4_ext_space_block_idx(inode); | |
296 | } | |
297 | ||
298 | return max; | |
299 | } | |
300 | ||
301 | static int __ext4_ext_check_header(const char *function, struct inode *inode, | |
302 | struct ext4_extent_header *eh, | |
303 | int depth) | |
304 | { | |
305 | const char *error_msg; | |
306 | int max = 0; | |
307 | ||
308 | if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) { | |
309 | error_msg = "invalid magic"; | |
310 | goto corrupted; | |
311 | } | |
312 | if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) { | |
313 | error_msg = "unexpected eh_depth"; | |
314 | goto corrupted; | |
315 | } | |
316 | if (unlikely(eh->eh_max == 0)) { | |
317 | error_msg = "invalid eh_max"; | |
318 | goto corrupted; | |
319 | } | |
320 | max = ext4_ext_max_entries(inode, depth); | |
321 | if (unlikely(le16_to_cpu(eh->eh_max) > max)) { | |
322 | error_msg = "too large eh_max"; | |
323 | goto corrupted; | |
324 | } | |
325 | if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) { | |
326 | error_msg = "invalid eh_entries"; | |
327 | goto corrupted; | |
328 | } | |
329 | return 0; | |
330 | ||
331 | corrupted: | |
332 | ext4_error(inode->i_sb, function, | |
333 | "bad header in inode #%lu: %s - magic %x, " | |
334 | "entries %u, max %u(%u), depth %u(%u)", | |
335 | inode->i_ino, error_msg, le16_to_cpu(eh->eh_magic), | |
336 | le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max), | |
337 | max, le16_to_cpu(eh->eh_depth), depth); | |
338 | ||
339 | return -EIO; | |
340 | } | |
341 | ||
342 | #define ext4_ext_check_header(inode, eh, depth) \ | |
46e665e9 | 343 | __ext4_ext_check_header(__func__, inode, eh, depth) |
c29c0ae7 | 344 | |
a86c6181 AT |
345 | #ifdef EXT_DEBUG |
346 | static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path) | |
347 | { | |
348 | int k, l = path->p_depth; | |
349 | ||
350 | ext_debug("path:"); | |
351 | for (k = 0; k <= l; k++, path++) { | |
352 | if (path->p_idx) { | |
2ae02107 | 353 | ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block), |
f65e6fba | 354 | idx_pblock(path->p_idx)); |
a86c6181 | 355 | } else if (path->p_ext) { |
2ae02107 | 356 | ext_debug(" %d:%d:%llu ", |
a86c6181 | 357 | le32_to_cpu(path->p_ext->ee_block), |
a2df2a63 | 358 | ext4_ext_get_actual_len(path->p_ext), |
f65e6fba | 359 | ext_pblock(path->p_ext)); |
a86c6181 AT |
360 | } else |
361 | ext_debug(" []"); | |
362 | } | |
363 | ext_debug("\n"); | |
364 | } | |
365 | ||
366 | static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path) | |
367 | { | |
368 | int depth = ext_depth(inode); | |
369 | struct ext4_extent_header *eh; | |
370 | struct ext4_extent *ex; | |
371 | int i; | |
372 | ||
373 | if (!path) | |
374 | return; | |
375 | ||
376 | eh = path[depth].p_hdr; | |
377 | ex = EXT_FIRST_EXTENT(eh); | |
378 | ||
379 | for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) { | |
2ae02107 | 380 | ext_debug("%d:%d:%llu ", le32_to_cpu(ex->ee_block), |
a2df2a63 | 381 | ext4_ext_get_actual_len(ex), ext_pblock(ex)); |
a86c6181 AT |
382 | } |
383 | ext_debug("\n"); | |
384 | } | |
385 | #else | |
af5bc92d TT |
386 | #define ext4_ext_show_path(inode, path) |
387 | #define ext4_ext_show_leaf(inode, path) | |
a86c6181 AT |
388 | #endif |
389 | ||
b35905c1 | 390 | void ext4_ext_drop_refs(struct ext4_ext_path *path) |
a86c6181 AT |
391 | { |
392 | int depth = path->p_depth; | |
393 | int i; | |
394 | ||
395 | for (i = 0; i <= depth; i++, path++) | |
396 | if (path->p_bh) { | |
397 | brelse(path->p_bh); | |
398 | path->p_bh = NULL; | |
399 | } | |
400 | } | |
401 | ||
402 | /* | |
d0d856e8 RD |
403 | * ext4_ext_binsearch_idx: |
404 | * binary search for the closest index of the given block | |
c29c0ae7 | 405 | * the header must be checked before calling this |
a86c6181 AT |
406 | */ |
407 | static void | |
725d26d3 AK |
408 | ext4_ext_binsearch_idx(struct inode *inode, |
409 | struct ext4_ext_path *path, ext4_lblk_t block) | |
a86c6181 AT |
410 | { |
411 | struct ext4_extent_header *eh = path->p_hdr; | |
412 | struct ext4_extent_idx *r, *l, *m; | |
413 | ||
a86c6181 | 414 | |
bba90743 | 415 | ext_debug("binsearch for %u(idx): ", block); |
a86c6181 AT |
416 | |
417 | l = EXT_FIRST_INDEX(eh) + 1; | |
e9f410b1 | 418 | r = EXT_LAST_INDEX(eh); |
a86c6181 AT |
419 | while (l <= r) { |
420 | m = l + (r - l) / 2; | |
421 | if (block < le32_to_cpu(m->ei_block)) | |
422 | r = m - 1; | |
423 | else | |
424 | l = m + 1; | |
26d535ed DM |
425 | ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block), |
426 | m, le32_to_cpu(m->ei_block), | |
427 | r, le32_to_cpu(r->ei_block)); | |
a86c6181 AT |
428 | } |
429 | ||
430 | path->p_idx = l - 1; | |
f65e6fba | 431 | ext_debug(" -> %d->%lld ", le32_to_cpu(path->p_idx->ei_block), |
26d535ed | 432 | idx_pblock(path->p_idx)); |
a86c6181 AT |
433 | |
434 | #ifdef CHECK_BINSEARCH | |
435 | { | |
436 | struct ext4_extent_idx *chix, *ix; | |
437 | int k; | |
438 | ||
439 | chix = ix = EXT_FIRST_INDEX(eh); | |
440 | for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) { | |
441 | if (k != 0 && | |
442 | le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) { | |
4776004f TT |
443 | printk(KERN_DEBUG "k=%d, ix=0x%p, " |
444 | "first=0x%p\n", k, | |
445 | ix, EXT_FIRST_INDEX(eh)); | |
446 | printk(KERN_DEBUG "%u <= %u\n", | |
a86c6181 AT |
447 | le32_to_cpu(ix->ei_block), |
448 | le32_to_cpu(ix[-1].ei_block)); | |
449 | } | |
450 | BUG_ON(k && le32_to_cpu(ix->ei_block) | |
8c55e204 | 451 | <= le32_to_cpu(ix[-1].ei_block)); |
a86c6181 AT |
452 | if (block < le32_to_cpu(ix->ei_block)) |
453 | break; | |
454 | chix = ix; | |
455 | } | |
456 | BUG_ON(chix != path->p_idx); | |
457 | } | |
458 | #endif | |
459 | ||
460 | } | |
461 | ||
462 | /* | |
d0d856e8 RD |
463 | * ext4_ext_binsearch: |
464 | * binary search for closest extent of the given block | |
c29c0ae7 | 465 | * the header must be checked before calling this |
a86c6181 AT |
466 | */ |
467 | static void | |
725d26d3 AK |
468 | ext4_ext_binsearch(struct inode *inode, |
469 | struct ext4_ext_path *path, ext4_lblk_t block) | |
a86c6181 AT |
470 | { |
471 | struct ext4_extent_header *eh = path->p_hdr; | |
472 | struct ext4_extent *r, *l, *m; | |
473 | ||
a86c6181 AT |
474 | if (eh->eh_entries == 0) { |
475 | /* | |
d0d856e8 RD |
476 | * this leaf is empty: |
477 | * we get such a leaf in split/add case | |
a86c6181 AT |
478 | */ |
479 | return; | |
480 | } | |
481 | ||
bba90743 | 482 | ext_debug("binsearch for %u: ", block); |
a86c6181 AT |
483 | |
484 | l = EXT_FIRST_EXTENT(eh) + 1; | |
e9f410b1 | 485 | r = EXT_LAST_EXTENT(eh); |
a86c6181 AT |
486 | |
487 | while (l <= r) { | |
488 | m = l + (r - l) / 2; | |
489 | if (block < le32_to_cpu(m->ee_block)) | |
490 | r = m - 1; | |
491 | else | |
492 | l = m + 1; | |
26d535ed DM |
493 | ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block), |
494 | m, le32_to_cpu(m->ee_block), | |
495 | r, le32_to_cpu(r->ee_block)); | |
a86c6181 AT |
496 | } |
497 | ||
498 | path->p_ext = l - 1; | |
2ae02107 | 499 | ext_debug(" -> %d:%llu:%d ", |
8c55e204 DK |
500 | le32_to_cpu(path->p_ext->ee_block), |
501 | ext_pblock(path->p_ext), | |
a2df2a63 | 502 | ext4_ext_get_actual_len(path->p_ext)); |
a86c6181 AT |
503 | |
504 | #ifdef CHECK_BINSEARCH | |
505 | { | |
506 | struct ext4_extent *chex, *ex; | |
507 | int k; | |
508 | ||
509 | chex = ex = EXT_FIRST_EXTENT(eh); | |
510 | for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) { | |
511 | BUG_ON(k && le32_to_cpu(ex->ee_block) | |
8c55e204 | 512 | <= le32_to_cpu(ex[-1].ee_block)); |
a86c6181 AT |
513 | if (block < le32_to_cpu(ex->ee_block)) |
514 | break; | |
515 | chex = ex; | |
516 | } | |
517 | BUG_ON(chex != path->p_ext); | |
518 | } | |
519 | #endif | |
520 | ||
521 | } | |
522 | ||
523 | int ext4_ext_tree_init(handle_t *handle, struct inode *inode) | |
524 | { | |
525 | struct ext4_extent_header *eh; | |
526 | ||
527 | eh = ext_inode_hdr(inode); | |
528 | eh->eh_depth = 0; | |
529 | eh->eh_entries = 0; | |
530 | eh->eh_magic = EXT4_EXT_MAGIC; | |
531 | eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode)); | |
532 | ext4_mark_inode_dirty(handle, inode); | |
533 | ext4_ext_invalidate_cache(inode); | |
534 | return 0; | |
535 | } | |
536 | ||
537 | struct ext4_ext_path * | |
725d26d3 AK |
538 | ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block, |
539 | struct ext4_ext_path *path) | |
a86c6181 AT |
540 | { |
541 | struct ext4_extent_header *eh; | |
542 | struct buffer_head *bh; | |
543 | short int depth, i, ppos = 0, alloc = 0; | |
544 | ||
545 | eh = ext_inode_hdr(inode); | |
c29c0ae7 AT |
546 | depth = ext_depth(inode); |
547 | if (ext4_ext_check_header(inode, eh, depth)) | |
a86c6181 AT |
548 | return ERR_PTR(-EIO); |
549 | ||
a86c6181 AT |
550 | |
551 | /* account possible depth increase */ | |
552 | if (!path) { | |
5d4958f9 | 553 | path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2), |
a86c6181 AT |
554 | GFP_NOFS); |
555 | if (!path) | |
556 | return ERR_PTR(-ENOMEM); | |
557 | alloc = 1; | |
558 | } | |
a86c6181 | 559 | path[0].p_hdr = eh; |
1973adcb | 560 | path[0].p_bh = NULL; |
a86c6181 | 561 | |
c29c0ae7 | 562 | i = depth; |
a86c6181 AT |
563 | /* walk through the tree */ |
564 | while (i) { | |
565 | ext_debug("depth %d: num %d, max %d\n", | |
566 | ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max)); | |
c29c0ae7 | 567 | |
a86c6181 | 568 | ext4_ext_binsearch_idx(inode, path + ppos, block); |
f65e6fba | 569 | path[ppos].p_block = idx_pblock(path[ppos].p_idx); |
a86c6181 AT |
570 | path[ppos].p_depth = i; |
571 | path[ppos].p_ext = NULL; | |
572 | ||
573 | bh = sb_bread(inode->i_sb, path[ppos].p_block); | |
574 | if (!bh) | |
575 | goto err; | |
576 | ||
577 | eh = ext_block_hdr(bh); | |
578 | ppos++; | |
579 | BUG_ON(ppos > depth); | |
580 | path[ppos].p_bh = bh; | |
581 | path[ppos].p_hdr = eh; | |
582 | i--; | |
583 | ||
c29c0ae7 | 584 | if (ext4_ext_check_header(inode, eh, i)) |
a86c6181 AT |
585 | goto err; |
586 | } | |
587 | ||
588 | path[ppos].p_depth = i; | |
a86c6181 AT |
589 | path[ppos].p_ext = NULL; |
590 | path[ppos].p_idx = NULL; | |
591 | ||
a86c6181 AT |
592 | /* find extent */ |
593 | ext4_ext_binsearch(inode, path + ppos, block); | |
1973adcb SF |
594 | /* if not an empty leaf */ |
595 | if (path[ppos].p_ext) | |
596 | path[ppos].p_block = ext_pblock(path[ppos].p_ext); | |
a86c6181 AT |
597 | |
598 | ext4_ext_show_path(inode, path); | |
599 | ||
600 | return path; | |
601 | ||
602 | err: | |
603 | ext4_ext_drop_refs(path); | |
604 | if (alloc) | |
605 | kfree(path); | |
606 | return ERR_PTR(-EIO); | |
607 | } | |
608 | ||
609 | /* | |
d0d856e8 RD |
610 | * ext4_ext_insert_index: |
611 | * insert new index [@logical;@ptr] into the block at @curp; | |
612 | * check where to insert: before @curp or after @curp | |
a86c6181 AT |
613 | */ |
614 | static int ext4_ext_insert_index(handle_t *handle, struct inode *inode, | |
615 | struct ext4_ext_path *curp, | |
f65e6fba | 616 | int logical, ext4_fsblk_t ptr) |
a86c6181 AT |
617 | { |
618 | struct ext4_extent_idx *ix; | |
619 | int len, err; | |
620 | ||
7e028976 AM |
621 | err = ext4_ext_get_access(handle, inode, curp); |
622 | if (err) | |
a86c6181 AT |
623 | return err; |
624 | ||
625 | BUG_ON(logical == le32_to_cpu(curp->p_idx->ei_block)); | |
626 | len = EXT_MAX_INDEX(curp->p_hdr) - curp->p_idx; | |
627 | if (logical > le32_to_cpu(curp->p_idx->ei_block)) { | |
628 | /* insert after */ | |
629 | if (curp->p_idx != EXT_LAST_INDEX(curp->p_hdr)) { | |
630 | len = (len - 1) * sizeof(struct ext4_extent_idx); | |
631 | len = len < 0 ? 0 : len; | |
26d535ed | 632 | ext_debug("insert new index %d after: %llu. " |
a86c6181 AT |
633 | "move %d from 0x%p to 0x%p\n", |
634 | logical, ptr, len, | |
635 | (curp->p_idx + 1), (curp->p_idx + 2)); | |
636 | memmove(curp->p_idx + 2, curp->p_idx + 1, len); | |
637 | } | |
638 | ix = curp->p_idx + 1; | |
639 | } else { | |
640 | /* insert before */ | |
641 | len = len * sizeof(struct ext4_extent_idx); | |
642 | len = len < 0 ? 0 : len; | |
26d535ed | 643 | ext_debug("insert new index %d before: %llu. " |
a86c6181 AT |
644 | "move %d from 0x%p to 0x%p\n", |
645 | logical, ptr, len, | |
646 | curp->p_idx, (curp->p_idx + 1)); | |
647 | memmove(curp->p_idx + 1, curp->p_idx, len); | |
648 | ix = curp->p_idx; | |
649 | } | |
650 | ||
651 | ix->ei_block = cpu_to_le32(logical); | |
f65e6fba | 652 | ext4_idx_store_pblock(ix, ptr); |
e8546d06 | 653 | le16_add_cpu(&curp->p_hdr->eh_entries, 1); |
a86c6181 AT |
654 | |
655 | BUG_ON(le16_to_cpu(curp->p_hdr->eh_entries) | |
8c55e204 | 656 | > le16_to_cpu(curp->p_hdr->eh_max)); |
a86c6181 AT |
657 | BUG_ON(ix > EXT_LAST_INDEX(curp->p_hdr)); |
658 | ||
659 | err = ext4_ext_dirty(handle, inode, curp); | |
660 | ext4_std_error(inode->i_sb, err); | |
661 | ||
662 | return err; | |
663 | } | |
664 | ||
665 | /* | |
d0d856e8 RD |
666 | * ext4_ext_split: |
667 | * inserts new subtree into the path, using free index entry | |
668 | * at depth @at: | |
669 | * - allocates all needed blocks (new leaf and all intermediate index blocks) | |
670 | * - makes decision where to split | |
671 | * - moves remaining extents and index entries (right to the split point) | |
672 | * into the newly allocated blocks | |
673 | * - initializes subtree | |
a86c6181 AT |
674 | */ |
675 | static int ext4_ext_split(handle_t *handle, struct inode *inode, | |
676 | struct ext4_ext_path *path, | |
677 | struct ext4_extent *newext, int at) | |
678 | { | |
679 | struct buffer_head *bh = NULL; | |
680 | int depth = ext_depth(inode); | |
681 | struct ext4_extent_header *neh; | |
682 | struct ext4_extent_idx *fidx; | |
683 | struct ext4_extent *ex; | |
684 | int i = at, k, m, a; | |
f65e6fba | 685 | ext4_fsblk_t newblock, oldblock; |
a86c6181 | 686 | __le32 border; |
f65e6fba | 687 | ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */ |
a86c6181 AT |
688 | int err = 0; |
689 | ||
690 | /* make decision: where to split? */ | |
d0d856e8 | 691 | /* FIXME: now decision is simplest: at current extent */ |
a86c6181 | 692 | |
d0d856e8 | 693 | /* if current leaf will be split, then we should use |
a86c6181 AT |
694 | * border from split point */ |
695 | BUG_ON(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr)); | |
696 | if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) { | |
697 | border = path[depth].p_ext[1].ee_block; | |
d0d856e8 | 698 | ext_debug("leaf will be split." |
a86c6181 | 699 | " next leaf starts at %d\n", |
8c55e204 | 700 | le32_to_cpu(border)); |
a86c6181 AT |
701 | } else { |
702 | border = newext->ee_block; | |
703 | ext_debug("leaf will be added." | |
704 | " next leaf starts at %d\n", | |
8c55e204 | 705 | le32_to_cpu(border)); |
a86c6181 AT |
706 | } |
707 | ||
708 | /* | |
d0d856e8 RD |
709 | * If error occurs, then we break processing |
710 | * and mark filesystem read-only. index won't | |
a86c6181 | 711 | * be inserted and tree will be in consistent |
d0d856e8 | 712 | * state. Next mount will repair buffers too. |
a86c6181 AT |
713 | */ |
714 | ||
715 | /* | |
d0d856e8 RD |
716 | * Get array to track all allocated blocks. |
717 | * We need this to handle errors and free blocks | |
718 | * upon them. | |
a86c6181 | 719 | */ |
5d4958f9 | 720 | ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS); |
a86c6181 AT |
721 | if (!ablocks) |
722 | return -ENOMEM; | |
a86c6181 AT |
723 | |
724 | /* allocate all needed blocks */ | |
725 | ext_debug("allocate %d blocks for indexes/leaf\n", depth - at); | |
726 | for (a = 0; a < depth - at; a++) { | |
654b4908 AK |
727 | newblock = ext4_ext_new_meta_block(handle, inode, path, |
728 | newext, &err); | |
a86c6181 AT |
729 | if (newblock == 0) |
730 | goto cleanup; | |
731 | ablocks[a] = newblock; | |
732 | } | |
733 | ||
734 | /* initialize new leaf */ | |
735 | newblock = ablocks[--a]; | |
736 | BUG_ON(newblock == 0); | |
737 | bh = sb_getblk(inode->i_sb, newblock); | |
738 | if (!bh) { | |
739 | err = -EIO; | |
740 | goto cleanup; | |
741 | } | |
742 | lock_buffer(bh); | |
743 | ||
7e028976 AM |
744 | err = ext4_journal_get_create_access(handle, bh); |
745 | if (err) | |
a86c6181 AT |
746 | goto cleanup; |
747 | ||
748 | neh = ext_block_hdr(bh); | |
749 | neh->eh_entries = 0; | |
750 | neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode)); | |
751 | neh->eh_magic = EXT4_EXT_MAGIC; | |
752 | neh->eh_depth = 0; | |
753 | ex = EXT_FIRST_EXTENT(neh); | |
754 | ||
d0d856e8 | 755 | /* move remainder of path[depth] to the new leaf */ |
a86c6181 AT |
756 | BUG_ON(path[depth].p_hdr->eh_entries != path[depth].p_hdr->eh_max); |
757 | /* start copy from next extent */ | |
758 | /* TODO: we could do it by single memmove */ | |
759 | m = 0; | |
760 | path[depth].p_ext++; | |
761 | while (path[depth].p_ext <= | |
762 | EXT_MAX_EXTENT(path[depth].p_hdr)) { | |
2ae02107 | 763 | ext_debug("move %d:%llu:%d in new leaf %llu\n", |
8c55e204 DK |
764 | le32_to_cpu(path[depth].p_ext->ee_block), |
765 | ext_pblock(path[depth].p_ext), | |
a2df2a63 | 766 | ext4_ext_get_actual_len(path[depth].p_ext), |
a86c6181 AT |
767 | newblock); |
768 | /*memmove(ex++, path[depth].p_ext++, | |
769 | sizeof(struct ext4_extent)); | |
770 | neh->eh_entries++;*/ | |
771 | path[depth].p_ext++; | |
772 | m++; | |
773 | } | |
774 | if (m) { | |
775 | memmove(ex, path[depth].p_ext-m, sizeof(struct ext4_extent)*m); | |
e8546d06 | 776 | le16_add_cpu(&neh->eh_entries, m); |
a86c6181 AT |
777 | } |
778 | ||
779 | set_buffer_uptodate(bh); | |
780 | unlock_buffer(bh); | |
781 | ||
7e028976 AM |
782 | err = ext4_journal_dirty_metadata(handle, bh); |
783 | if (err) | |
a86c6181 AT |
784 | goto cleanup; |
785 | brelse(bh); | |
786 | bh = NULL; | |
787 | ||
788 | /* correct old leaf */ | |
789 | if (m) { | |
7e028976 AM |
790 | err = ext4_ext_get_access(handle, inode, path + depth); |
791 | if (err) | |
a86c6181 | 792 | goto cleanup; |
e8546d06 | 793 | le16_add_cpu(&path[depth].p_hdr->eh_entries, -m); |
7e028976 AM |
794 | err = ext4_ext_dirty(handle, inode, path + depth); |
795 | if (err) | |
a86c6181 AT |
796 | goto cleanup; |
797 | ||
798 | } | |
799 | ||
800 | /* create intermediate indexes */ | |
801 | k = depth - at - 1; | |
802 | BUG_ON(k < 0); | |
803 | if (k) | |
804 | ext_debug("create %d intermediate indices\n", k); | |
805 | /* insert new index into current index block */ | |
806 | /* current depth stored in i var */ | |
807 | i = depth - 1; | |
808 | while (k--) { | |
809 | oldblock = newblock; | |
810 | newblock = ablocks[--a]; | |
bba90743 | 811 | bh = sb_getblk(inode->i_sb, newblock); |
a86c6181 AT |
812 | if (!bh) { |
813 | err = -EIO; | |
814 | goto cleanup; | |
815 | } | |
816 | lock_buffer(bh); | |
817 | ||
7e028976 AM |
818 | err = ext4_journal_get_create_access(handle, bh); |
819 | if (err) | |
a86c6181 AT |
820 | goto cleanup; |
821 | ||
822 | neh = ext_block_hdr(bh); | |
823 | neh->eh_entries = cpu_to_le16(1); | |
824 | neh->eh_magic = EXT4_EXT_MAGIC; | |
825 | neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode)); | |
826 | neh->eh_depth = cpu_to_le16(depth - i); | |
827 | fidx = EXT_FIRST_INDEX(neh); | |
828 | fidx->ei_block = border; | |
f65e6fba | 829 | ext4_idx_store_pblock(fidx, oldblock); |
a86c6181 | 830 | |
bba90743 ES |
831 | ext_debug("int.index at %d (block %llu): %u -> %llu\n", |
832 | i, newblock, le32_to_cpu(border), oldblock); | |
a86c6181 AT |
833 | /* copy indexes */ |
834 | m = 0; | |
835 | path[i].p_idx++; | |
836 | ||
837 | ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx, | |
838 | EXT_MAX_INDEX(path[i].p_hdr)); | |
839 | BUG_ON(EXT_MAX_INDEX(path[i].p_hdr) != | |
840 | EXT_LAST_INDEX(path[i].p_hdr)); | |
841 | while (path[i].p_idx <= EXT_MAX_INDEX(path[i].p_hdr)) { | |
26d535ed | 842 | ext_debug("%d: move %d:%llu in new index %llu\n", i, |
8c55e204 DK |
843 | le32_to_cpu(path[i].p_idx->ei_block), |
844 | idx_pblock(path[i].p_idx), | |
845 | newblock); | |
a86c6181 AT |
846 | /*memmove(++fidx, path[i].p_idx++, |
847 | sizeof(struct ext4_extent_idx)); | |
848 | neh->eh_entries++; | |
849 | BUG_ON(neh->eh_entries > neh->eh_max);*/ | |
850 | path[i].p_idx++; | |
851 | m++; | |
852 | } | |
853 | if (m) { | |
854 | memmove(++fidx, path[i].p_idx - m, | |
855 | sizeof(struct ext4_extent_idx) * m); | |
e8546d06 | 856 | le16_add_cpu(&neh->eh_entries, m); |
a86c6181 AT |
857 | } |
858 | set_buffer_uptodate(bh); | |
859 | unlock_buffer(bh); | |
860 | ||
7e028976 AM |
861 | err = ext4_journal_dirty_metadata(handle, bh); |
862 | if (err) | |
a86c6181 AT |
863 | goto cleanup; |
864 | brelse(bh); | |
865 | bh = NULL; | |
866 | ||
867 | /* correct old index */ | |
868 | if (m) { | |
869 | err = ext4_ext_get_access(handle, inode, path + i); | |
870 | if (err) | |
871 | goto cleanup; | |
e8546d06 | 872 | le16_add_cpu(&path[i].p_hdr->eh_entries, -m); |
a86c6181 AT |
873 | err = ext4_ext_dirty(handle, inode, path + i); |
874 | if (err) | |
875 | goto cleanup; | |
876 | } | |
877 | ||
878 | i--; | |
879 | } | |
880 | ||
881 | /* insert new index */ | |
a86c6181 AT |
882 | err = ext4_ext_insert_index(handle, inode, path + at, |
883 | le32_to_cpu(border), newblock); | |
884 | ||
885 | cleanup: | |
886 | if (bh) { | |
887 | if (buffer_locked(bh)) | |
888 | unlock_buffer(bh); | |
889 | brelse(bh); | |
890 | } | |
891 | ||
892 | if (err) { | |
893 | /* free all allocated blocks in error case */ | |
894 | for (i = 0; i < depth; i++) { | |
895 | if (!ablocks[i]) | |
896 | continue; | |
c9de560d | 897 | ext4_free_blocks(handle, inode, ablocks[i], 1, 1); |
a86c6181 AT |
898 | } |
899 | } | |
900 | kfree(ablocks); | |
901 | ||
902 | return err; | |
903 | } | |
904 | ||
905 | /* | |
d0d856e8 RD |
906 | * ext4_ext_grow_indepth: |
907 | * implements tree growing procedure: | |
908 | * - allocates new block | |
909 | * - moves top-level data (index block or leaf) into the new block | |
910 | * - initializes new top-level, creating index that points to the | |
911 | * just created block | |
a86c6181 AT |
912 | */ |
913 | static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode, | |
914 | struct ext4_ext_path *path, | |
915 | struct ext4_extent *newext) | |
916 | { | |
917 | struct ext4_ext_path *curp = path; | |
918 | struct ext4_extent_header *neh; | |
919 | struct ext4_extent_idx *fidx; | |
920 | struct buffer_head *bh; | |
f65e6fba | 921 | ext4_fsblk_t newblock; |
a86c6181 AT |
922 | int err = 0; |
923 | ||
654b4908 | 924 | newblock = ext4_ext_new_meta_block(handle, inode, path, newext, &err); |
a86c6181 AT |
925 | if (newblock == 0) |
926 | return err; | |
927 | ||
928 | bh = sb_getblk(inode->i_sb, newblock); | |
929 | if (!bh) { | |
930 | err = -EIO; | |
931 | ext4_std_error(inode->i_sb, err); | |
932 | return err; | |
933 | } | |
934 | lock_buffer(bh); | |
935 | ||
7e028976 AM |
936 | err = ext4_journal_get_create_access(handle, bh); |
937 | if (err) { | |
a86c6181 AT |
938 | unlock_buffer(bh); |
939 | goto out; | |
940 | } | |
941 | ||
942 | /* move top-level index/leaf into new block */ | |
943 | memmove(bh->b_data, curp->p_hdr, sizeof(EXT4_I(inode)->i_data)); | |
944 | ||
945 | /* set size of new block */ | |
946 | neh = ext_block_hdr(bh); | |
947 | /* old root could have indexes or leaves | |
948 | * so calculate e_max right way */ | |
949 | if (ext_depth(inode)) | |
950 | neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode)); | |
951 | else | |
952 | neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode)); | |
953 | neh->eh_magic = EXT4_EXT_MAGIC; | |
954 | set_buffer_uptodate(bh); | |
955 | unlock_buffer(bh); | |
956 | ||
7e028976 AM |
957 | err = ext4_journal_dirty_metadata(handle, bh); |
958 | if (err) | |
a86c6181 AT |
959 | goto out; |
960 | ||
961 | /* create index in new top-level index: num,max,pointer */ | |
7e028976 AM |
962 | err = ext4_ext_get_access(handle, inode, curp); |
963 | if (err) | |
a86c6181 AT |
964 | goto out; |
965 | ||
966 | curp->p_hdr->eh_magic = EXT4_EXT_MAGIC; | |
967 | curp->p_hdr->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode)); | |
968 | curp->p_hdr->eh_entries = cpu_to_le16(1); | |
969 | curp->p_idx = EXT_FIRST_INDEX(curp->p_hdr); | |
e9f410b1 DM |
970 | |
971 | if (path[0].p_hdr->eh_depth) | |
972 | curp->p_idx->ei_block = | |
973 | EXT_FIRST_INDEX(path[0].p_hdr)->ei_block; | |
974 | else | |
975 | curp->p_idx->ei_block = | |
976 | EXT_FIRST_EXTENT(path[0].p_hdr)->ee_block; | |
f65e6fba | 977 | ext4_idx_store_pblock(curp->p_idx, newblock); |
a86c6181 AT |
978 | |
979 | neh = ext_inode_hdr(inode); | |
980 | fidx = EXT_FIRST_INDEX(neh); | |
2ae02107 | 981 | ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n", |
a86c6181 | 982 | le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max), |
f65e6fba | 983 | le32_to_cpu(fidx->ei_block), idx_pblock(fidx)); |
a86c6181 AT |
984 | |
985 | neh->eh_depth = cpu_to_le16(path->p_depth + 1); | |
986 | err = ext4_ext_dirty(handle, inode, curp); | |
987 | out: | |
988 | brelse(bh); | |
989 | ||
990 | return err; | |
991 | } | |
992 | ||
993 | /* | |
d0d856e8 RD |
994 | * ext4_ext_create_new_leaf: |
995 | * finds empty index and adds new leaf. | |
996 | * if no free index is found, then it requests in-depth growing. | |
a86c6181 AT |
997 | */ |
998 | static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode, | |
999 | struct ext4_ext_path *path, | |
1000 | struct ext4_extent *newext) | |
1001 | { | |
1002 | struct ext4_ext_path *curp; | |
1003 | int depth, i, err = 0; | |
1004 | ||
1005 | repeat: | |
1006 | i = depth = ext_depth(inode); | |
1007 | ||
1008 | /* walk up to the tree and look for free index entry */ | |
1009 | curp = path + depth; | |
1010 | while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) { | |
1011 | i--; | |
1012 | curp--; | |
1013 | } | |
1014 | ||
d0d856e8 RD |
1015 | /* we use already allocated block for index block, |
1016 | * so subsequent data blocks should be contiguous */ | |
a86c6181 AT |
1017 | if (EXT_HAS_FREE_INDEX(curp)) { |
1018 | /* if we found index with free entry, then use that | |
1019 | * entry: create all needed subtree and add new leaf */ | |
1020 | err = ext4_ext_split(handle, inode, path, newext, i); | |
787e0981 SF |
1021 | if (err) |
1022 | goto out; | |
a86c6181 AT |
1023 | |
1024 | /* refill path */ | |
1025 | ext4_ext_drop_refs(path); | |
1026 | path = ext4_ext_find_extent(inode, | |
725d26d3 AK |
1027 | (ext4_lblk_t)le32_to_cpu(newext->ee_block), |
1028 | path); | |
a86c6181 AT |
1029 | if (IS_ERR(path)) |
1030 | err = PTR_ERR(path); | |
1031 | } else { | |
1032 | /* tree is full, time to grow in depth */ | |
1033 | err = ext4_ext_grow_indepth(handle, inode, path, newext); | |
1034 | if (err) | |
1035 | goto out; | |
1036 | ||
1037 | /* refill path */ | |
1038 | ext4_ext_drop_refs(path); | |
1039 | path = ext4_ext_find_extent(inode, | |
725d26d3 AK |
1040 | (ext4_lblk_t)le32_to_cpu(newext->ee_block), |
1041 | path); | |
a86c6181 AT |
1042 | if (IS_ERR(path)) { |
1043 | err = PTR_ERR(path); | |
1044 | goto out; | |
1045 | } | |
1046 | ||
1047 | /* | |
d0d856e8 RD |
1048 | * only first (depth 0 -> 1) produces free space; |
1049 | * in all other cases we have to split the grown tree | |
a86c6181 AT |
1050 | */ |
1051 | depth = ext_depth(inode); | |
1052 | if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) { | |
d0d856e8 | 1053 | /* now we need to split */ |
a86c6181 AT |
1054 | goto repeat; |
1055 | } | |
1056 | } | |
1057 | ||
1058 | out: | |
1059 | return err; | |
1060 | } | |
1061 | ||
1988b51e AT |
1062 | /* |
1063 | * search the closest allocated block to the left for *logical | |
1064 | * and returns it at @logical + it's physical address at @phys | |
1065 | * if *logical is the smallest allocated block, the function | |
1066 | * returns 0 at @phys | |
1067 | * return value contains 0 (success) or error code | |
1068 | */ | |
1069 | int | |
1070 | ext4_ext_search_left(struct inode *inode, struct ext4_ext_path *path, | |
1071 | ext4_lblk_t *logical, ext4_fsblk_t *phys) | |
1072 | { | |
1073 | struct ext4_extent_idx *ix; | |
1074 | struct ext4_extent *ex; | |
b939e376 | 1075 | int depth, ee_len; |
1988b51e AT |
1076 | |
1077 | BUG_ON(path == NULL); | |
1078 | depth = path->p_depth; | |
1079 | *phys = 0; | |
1080 | ||
1081 | if (depth == 0 && path->p_ext == NULL) | |
1082 | return 0; | |
1083 | ||
1084 | /* usually extent in the path covers blocks smaller | |
1085 | * then *logical, but it can be that extent is the | |
1086 | * first one in the file */ | |
1087 | ||
1088 | ex = path[depth].p_ext; | |
b939e376 | 1089 | ee_len = ext4_ext_get_actual_len(ex); |
1988b51e AT |
1090 | if (*logical < le32_to_cpu(ex->ee_block)) { |
1091 | BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex); | |
1092 | while (--depth >= 0) { | |
1093 | ix = path[depth].p_idx; | |
1094 | BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr)); | |
1095 | } | |
1096 | return 0; | |
1097 | } | |
1098 | ||
b939e376 | 1099 | BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len)); |
1988b51e | 1100 | |
b939e376 AK |
1101 | *logical = le32_to_cpu(ex->ee_block) + ee_len - 1; |
1102 | *phys = ext_pblock(ex) + ee_len - 1; | |
1988b51e AT |
1103 | return 0; |
1104 | } | |
1105 | ||
1106 | /* | |
1107 | * search the closest allocated block to the right for *logical | |
1108 | * and returns it at @logical + it's physical address at @phys | |
1109 | * if *logical is the smallest allocated block, the function | |
1110 | * returns 0 at @phys | |
1111 | * return value contains 0 (success) or error code | |
1112 | */ | |
1113 | int | |
1114 | ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path, | |
1115 | ext4_lblk_t *logical, ext4_fsblk_t *phys) | |
1116 | { | |
1117 | struct buffer_head *bh = NULL; | |
1118 | struct ext4_extent_header *eh; | |
1119 | struct ext4_extent_idx *ix; | |
1120 | struct ext4_extent *ex; | |
1121 | ext4_fsblk_t block; | |
b939e376 | 1122 | int depth, ee_len; |
1988b51e AT |
1123 | |
1124 | BUG_ON(path == NULL); | |
1125 | depth = path->p_depth; | |
1126 | *phys = 0; | |
1127 | ||
1128 | if (depth == 0 && path->p_ext == NULL) | |
1129 | return 0; | |
1130 | ||
1131 | /* usually extent in the path covers blocks smaller | |
1132 | * then *logical, but it can be that extent is the | |
1133 | * first one in the file */ | |
1134 | ||
1135 | ex = path[depth].p_ext; | |
b939e376 | 1136 | ee_len = ext4_ext_get_actual_len(ex); |
1988b51e AT |
1137 | if (*logical < le32_to_cpu(ex->ee_block)) { |
1138 | BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex); | |
1139 | while (--depth >= 0) { | |
1140 | ix = path[depth].p_idx; | |
1141 | BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr)); | |
1142 | } | |
1143 | *logical = le32_to_cpu(ex->ee_block); | |
1144 | *phys = ext_pblock(ex); | |
1145 | return 0; | |
1146 | } | |
1147 | ||
b939e376 | 1148 | BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len)); |
1988b51e AT |
1149 | |
1150 | if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) { | |
1151 | /* next allocated block in this leaf */ | |
1152 | ex++; | |
1153 | *logical = le32_to_cpu(ex->ee_block); | |
1154 | *phys = ext_pblock(ex); | |
1155 | return 0; | |
1156 | } | |
1157 | ||
1158 | /* go up and search for index to the right */ | |
1159 | while (--depth >= 0) { | |
1160 | ix = path[depth].p_idx; | |
1161 | if (ix != EXT_LAST_INDEX(path[depth].p_hdr)) | |
1162 | break; | |
1163 | } | |
1164 | ||
1165 | if (depth < 0) { | |
1166 | /* we've gone up to the root and | |
1167 | * found no index to the right */ | |
1168 | return 0; | |
1169 | } | |
1170 | ||
1171 | /* we've found index to the right, let's | |
1172 | * follow it and find the closest allocated | |
1173 | * block to the right */ | |
1174 | ix++; | |
1175 | block = idx_pblock(ix); | |
1176 | while (++depth < path->p_depth) { | |
1177 | bh = sb_bread(inode->i_sb, block); | |
1178 | if (bh == NULL) | |
1179 | return -EIO; | |
1180 | eh = ext_block_hdr(bh); | |
1181 | if (ext4_ext_check_header(inode, eh, depth)) { | |
1182 | put_bh(bh); | |
1183 | return -EIO; | |
1184 | } | |
1185 | ix = EXT_FIRST_INDEX(eh); | |
1186 | block = idx_pblock(ix); | |
1187 | put_bh(bh); | |
1188 | } | |
1189 | ||
1190 | bh = sb_bread(inode->i_sb, block); | |
1191 | if (bh == NULL) | |
1192 | return -EIO; | |
1193 | eh = ext_block_hdr(bh); | |
1194 | if (ext4_ext_check_header(inode, eh, path->p_depth - depth)) { | |
1195 | put_bh(bh); | |
1196 | return -EIO; | |
1197 | } | |
1198 | ex = EXT_FIRST_EXTENT(eh); | |
1199 | *logical = le32_to_cpu(ex->ee_block); | |
1200 | *phys = ext_pblock(ex); | |
1201 | put_bh(bh); | |
1202 | return 0; | |
1203 | ||
1204 | } | |
1205 | ||
a86c6181 | 1206 | /* |
d0d856e8 RD |
1207 | * ext4_ext_next_allocated_block: |
1208 | * returns allocated block in subsequent extent or EXT_MAX_BLOCK. | |
1209 | * NOTE: it considers block number from index entry as | |
1210 | * allocated block. Thus, index entries have to be consistent | |
1211 | * with leaves. | |
a86c6181 | 1212 | */ |
725d26d3 | 1213 | static ext4_lblk_t |
a86c6181 AT |
1214 | ext4_ext_next_allocated_block(struct ext4_ext_path *path) |
1215 | { | |
1216 | int depth; | |
1217 | ||
1218 | BUG_ON(path == NULL); | |
1219 | depth = path->p_depth; | |
1220 | ||
1221 | if (depth == 0 && path->p_ext == NULL) | |
1222 | return EXT_MAX_BLOCK; | |
1223 | ||
1224 | while (depth >= 0) { | |
1225 | if (depth == path->p_depth) { | |
1226 | /* leaf */ | |
1227 | if (path[depth].p_ext != | |
1228 | EXT_LAST_EXTENT(path[depth].p_hdr)) | |
1229 | return le32_to_cpu(path[depth].p_ext[1].ee_block); | |
1230 | } else { | |
1231 | /* index */ | |
1232 | if (path[depth].p_idx != | |
1233 | EXT_LAST_INDEX(path[depth].p_hdr)) | |
1234 | return le32_to_cpu(path[depth].p_idx[1].ei_block); | |
1235 | } | |
1236 | depth--; | |
1237 | } | |
1238 | ||
1239 | return EXT_MAX_BLOCK; | |
1240 | } | |
1241 | ||
1242 | /* | |
d0d856e8 | 1243 | * ext4_ext_next_leaf_block: |
a86c6181 AT |
1244 | * returns first allocated block from next leaf or EXT_MAX_BLOCK |
1245 | */ | |
725d26d3 | 1246 | static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode, |
63f57933 | 1247 | struct ext4_ext_path *path) |
a86c6181 AT |
1248 | { |
1249 | int depth; | |
1250 | ||
1251 | BUG_ON(path == NULL); | |
1252 | depth = path->p_depth; | |
1253 | ||
1254 | /* zero-tree has no leaf blocks at all */ | |
1255 | if (depth == 0) | |
1256 | return EXT_MAX_BLOCK; | |
1257 | ||
1258 | /* go to index block */ | |
1259 | depth--; | |
1260 | ||
1261 | while (depth >= 0) { | |
1262 | if (path[depth].p_idx != | |
1263 | EXT_LAST_INDEX(path[depth].p_hdr)) | |
725d26d3 AK |
1264 | return (ext4_lblk_t) |
1265 | le32_to_cpu(path[depth].p_idx[1].ei_block); | |
a86c6181 AT |
1266 | depth--; |
1267 | } | |
1268 | ||
1269 | return EXT_MAX_BLOCK; | |
1270 | } | |
1271 | ||
1272 | /* | |
d0d856e8 RD |
1273 | * ext4_ext_correct_indexes: |
1274 | * if leaf gets modified and modified extent is first in the leaf, | |
1275 | * then we have to correct all indexes above. | |
a86c6181 AT |
1276 | * TODO: do we need to correct tree in all cases? |
1277 | */ | |
1d03ec98 | 1278 | static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode, |
a86c6181 AT |
1279 | struct ext4_ext_path *path) |
1280 | { | |
1281 | struct ext4_extent_header *eh; | |
1282 | int depth = ext_depth(inode); | |
1283 | struct ext4_extent *ex; | |
1284 | __le32 border; | |
1285 | int k, err = 0; | |
1286 | ||
1287 | eh = path[depth].p_hdr; | |
1288 | ex = path[depth].p_ext; | |
1289 | BUG_ON(ex == NULL); | |
1290 | BUG_ON(eh == NULL); | |
1291 | ||
1292 | if (depth == 0) { | |
1293 | /* there is no tree at all */ | |
1294 | return 0; | |
1295 | } | |
1296 | ||
1297 | if (ex != EXT_FIRST_EXTENT(eh)) { | |
1298 | /* we correct tree if first leaf got modified only */ | |
1299 | return 0; | |
1300 | } | |
1301 | ||
1302 | /* | |
d0d856e8 | 1303 | * TODO: we need correction if border is smaller than current one |
a86c6181 AT |
1304 | */ |
1305 | k = depth - 1; | |
1306 | border = path[depth].p_ext->ee_block; | |
7e028976 AM |
1307 | err = ext4_ext_get_access(handle, inode, path + k); |
1308 | if (err) | |
a86c6181 AT |
1309 | return err; |
1310 | path[k].p_idx->ei_block = border; | |
7e028976 AM |
1311 | err = ext4_ext_dirty(handle, inode, path + k); |
1312 | if (err) | |
a86c6181 AT |
1313 | return err; |
1314 | ||
1315 | while (k--) { | |
1316 | /* change all left-side indexes */ | |
1317 | if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr)) | |
1318 | break; | |
7e028976 AM |
1319 | err = ext4_ext_get_access(handle, inode, path + k); |
1320 | if (err) | |
a86c6181 AT |
1321 | break; |
1322 | path[k].p_idx->ei_block = border; | |
7e028976 AM |
1323 | err = ext4_ext_dirty(handle, inode, path + k); |
1324 | if (err) | |
a86c6181 AT |
1325 | break; |
1326 | } | |
1327 | ||
1328 | return err; | |
1329 | } | |
1330 | ||
09b88252 | 1331 | static int |
a86c6181 AT |
1332 | ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1, |
1333 | struct ext4_extent *ex2) | |
1334 | { | |
749269fa | 1335 | unsigned short ext1_ee_len, ext2_ee_len, max_len; |
a2df2a63 AA |
1336 | |
1337 | /* | |
1338 | * Make sure that either both extents are uninitialized, or | |
1339 | * both are _not_. | |
1340 | */ | |
1341 | if (ext4_ext_is_uninitialized(ex1) ^ ext4_ext_is_uninitialized(ex2)) | |
1342 | return 0; | |
1343 | ||
749269fa AA |
1344 | if (ext4_ext_is_uninitialized(ex1)) |
1345 | max_len = EXT_UNINIT_MAX_LEN; | |
1346 | else | |
1347 | max_len = EXT_INIT_MAX_LEN; | |
1348 | ||
a2df2a63 AA |
1349 | ext1_ee_len = ext4_ext_get_actual_len(ex1); |
1350 | ext2_ee_len = ext4_ext_get_actual_len(ex2); | |
1351 | ||
1352 | if (le32_to_cpu(ex1->ee_block) + ext1_ee_len != | |
63f57933 | 1353 | le32_to_cpu(ex2->ee_block)) |
a86c6181 AT |
1354 | return 0; |
1355 | ||
471d4011 SB |
1356 | /* |
1357 | * To allow future support for preallocated extents to be added | |
1358 | * as an RO_COMPAT feature, refuse to merge to extents if | |
d0d856e8 | 1359 | * this can result in the top bit of ee_len being set. |
471d4011 | 1360 | */ |
749269fa | 1361 | if (ext1_ee_len + ext2_ee_len > max_len) |
471d4011 | 1362 | return 0; |
bbf2f9fb | 1363 | #ifdef AGGRESSIVE_TEST |
b939e376 | 1364 | if (ext1_ee_len >= 4) |
a86c6181 AT |
1365 | return 0; |
1366 | #endif | |
1367 | ||
a2df2a63 | 1368 | if (ext_pblock(ex1) + ext1_ee_len == ext_pblock(ex2)) |
a86c6181 AT |
1369 | return 1; |
1370 | return 0; | |
1371 | } | |
1372 | ||
56055d3a AA |
1373 | /* |
1374 | * This function tries to merge the "ex" extent to the next extent in the tree. | |
1375 | * It always tries to merge towards right. If you want to merge towards | |
1376 | * left, pass "ex - 1" as argument instead of "ex". | |
1377 | * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns | |
1378 | * 1 if they got merged. | |
1379 | */ | |
1380 | int ext4_ext_try_to_merge(struct inode *inode, | |
1381 | struct ext4_ext_path *path, | |
1382 | struct ext4_extent *ex) | |
1383 | { | |
1384 | struct ext4_extent_header *eh; | |
1385 | unsigned int depth, len; | |
1386 | int merge_done = 0; | |
1387 | int uninitialized = 0; | |
1388 | ||
1389 | depth = ext_depth(inode); | |
1390 | BUG_ON(path[depth].p_hdr == NULL); | |
1391 | eh = path[depth].p_hdr; | |
1392 | ||
1393 | while (ex < EXT_LAST_EXTENT(eh)) { | |
1394 | if (!ext4_can_extents_be_merged(inode, ex, ex + 1)) | |
1395 | break; | |
1396 | /* merge with next extent! */ | |
1397 | if (ext4_ext_is_uninitialized(ex)) | |
1398 | uninitialized = 1; | |
1399 | ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex) | |
1400 | + ext4_ext_get_actual_len(ex + 1)); | |
1401 | if (uninitialized) | |
1402 | ext4_ext_mark_uninitialized(ex); | |
1403 | ||
1404 | if (ex + 1 < EXT_LAST_EXTENT(eh)) { | |
1405 | len = (EXT_LAST_EXTENT(eh) - ex - 1) | |
1406 | * sizeof(struct ext4_extent); | |
1407 | memmove(ex + 1, ex + 2, len); | |
1408 | } | |
e8546d06 | 1409 | le16_add_cpu(&eh->eh_entries, -1); |
56055d3a AA |
1410 | merge_done = 1; |
1411 | WARN_ON(eh->eh_entries == 0); | |
1412 | if (!eh->eh_entries) | |
1413 | ext4_error(inode->i_sb, "ext4_ext_try_to_merge", | |
1414 | "inode#%lu, eh->eh_entries = 0!", inode->i_ino); | |
1415 | } | |
1416 | ||
1417 | return merge_done; | |
1418 | } | |
1419 | ||
25d14f98 AA |
1420 | /* |
1421 | * check if a portion of the "newext" extent overlaps with an | |
1422 | * existing extent. | |
1423 | * | |
1424 | * If there is an overlap discovered, it updates the length of the newext | |
1425 | * such that there will be no overlap, and then returns 1. | |
1426 | * If there is no overlap found, it returns 0. | |
1427 | */ | |
1428 | unsigned int ext4_ext_check_overlap(struct inode *inode, | |
1429 | struct ext4_extent *newext, | |
1430 | struct ext4_ext_path *path) | |
1431 | { | |
725d26d3 | 1432 | ext4_lblk_t b1, b2; |
25d14f98 AA |
1433 | unsigned int depth, len1; |
1434 | unsigned int ret = 0; | |
1435 | ||
1436 | b1 = le32_to_cpu(newext->ee_block); | |
a2df2a63 | 1437 | len1 = ext4_ext_get_actual_len(newext); |
25d14f98 AA |
1438 | depth = ext_depth(inode); |
1439 | if (!path[depth].p_ext) | |
1440 | goto out; | |
1441 | b2 = le32_to_cpu(path[depth].p_ext->ee_block); | |
1442 | ||
1443 | /* | |
1444 | * get the next allocated block if the extent in the path | |
2b2d6d01 | 1445 | * is before the requested block(s) |
25d14f98 AA |
1446 | */ |
1447 | if (b2 < b1) { | |
1448 | b2 = ext4_ext_next_allocated_block(path); | |
1449 | if (b2 == EXT_MAX_BLOCK) | |
1450 | goto out; | |
1451 | } | |
1452 | ||
725d26d3 | 1453 | /* check for wrap through zero on extent logical start block*/ |
25d14f98 AA |
1454 | if (b1 + len1 < b1) { |
1455 | len1 = EXT_MAX_BLOCK - b1; | |
1456 | newext->ee_len = cpu_to_le16(len1); | |
1457 | ret = 1; | |
1458 | } | |
1459 | ||
1460 | /* check for overlap */ | |
1461 | if (b1 + len1 > b2) { | |
1462 | newext->ee_len = cpu_to_le16(b2 - b1); | |
1463 | ret = 1; | |
1464 | } | |
1465 | out: | |
1466 | return ret; | |
1467 | } | |
1468 | ||
a86c6181 | 1469 | /* |
d0d856e8 RD |
1470 | * ext4_ext_insert_extent: |
1471 | * tries to merge requsted extent into the existing extent or | |
1472 | * inserts requested extent as new one into the tree, | |
1473 | * creating new leaf in the no-space case. | |
a86c6181 AT |
1474 | */ |
1475 | int ext4_ext_insert_extent(handle_t *handle, struct inode *inode, | |
1476 | struct ext4_ext_path *path, | |
1477 | struct ext4_extent *newext) | |
1478 | { | |
af5bc92d | 1479 | struct ext4_extent_header *eh; |
a86c6181 AT |
1480 | struct ext4_extent *ex, *fex; |
1481 | struct ext4_extent *nearex; /* nearest extent */ | |
1482 | struct ext4_ext_path *npath = NULL; | |
725d26d3 AK |
1483 | int depth, len, err; |
1484 | ext4_lblk_t next; | |
a2df2a63 | 1485 | unsigned uninitialized = 0; |
a86c6181 | 1486 | |
a2df2a63 | 1487 | BUG_ON(ext4_ext_get_actual_len(newext) == 0); |
a86c6181 AT |
1488 | depth = ext_depth(inode); |
1489 | ex = path[depth].p_ext; | |
1490 | BUG_ON(path[depth].p_hdr == NULL); | |
1491 | ||
1492 | /* try to insert block into found extent and return */ | |
1493 | if (ex && ext4_can_extents_be_merged(inode, ex, newext)) { | |
2ae02107 | 1494 | ext_debug("append %d block to %d:%d (from %llu)\n", |
a2df2a63 | 1495 | ext4_ext_get_actual_len(newext), |
a86c6181 | 1496 | le32_to_cpu(ex->ee_block), |
a2df2a63 | 1497 | ext4_ext_get_actual_len(ex), ext_pblock(ex)); |
7e028976 AM |
1498 | err = ext4_ext_get_access(handle, inode, path + depth); |
1499 | if (err) | |
a86c6181 | 1500 | return err; |
a2df2a63 AA |
1501 | |
1502 | /* | |
1503 | * ext4_can_extents_be_merged should have checked that either | |
1504 | * both extents are uninitialized, or both aren't. Thus we | |
1505 | * need to check only one of them here. | |
1506 | */ | |
1507 | if (ext4_ext_is_uninitialized(ex)) | |
1508 | uninitialized = 1; | |
1509 | ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex) | |
1510 | + ext4_ext_get_actual_len(newext)); | |
1511 | if (uninitialized) | |
1512 | ext4_ext_mark_uninitialized(ex); | |
a86c6181 AT |
1513 | eh = path[depth].p_hdr; |
1514 | nearex = ex; | |
1515 | goto merge; | |
1516 | } | |
1517 | ||
1518 | repeat: | |
1519 | depth = ext_depth(inode); | |
1520 | eh = path[depth].p_hdr; | |
1521 | if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) | |
1522 | goto has_space; | |
1523 | ||
1524 | /* probably next leaf has space for us? */ | |
1525 | fex = EXT_LAST_EXTENT(eh); | |
1526 | next = ext4_ext_next_leaf_block(inode, path); | |
1527 | if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block) | |
1528 | && next != EXT_MAX_BLOCK) { | |
1529 | ext_debug("next leaf block - %d\n", next); | |
1530 | BUG_ON(npath != NULL); | |
1531 | npath = ext4_ext_find_extent(inode, next, NULL); | |
1532 | if (IS_ERR(npath)) | |
1533 | return PTR_ERR(npath); | |
1534 | BUG_ON(npath->p_depth != path->p_depth); | |
1535 | eh = npath[depth].p_hdr; | |
1536 | if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) { | |
1537 | ext_debug("next leaf isnt full(%d)\n", | |
1538 | le16_to_cpu(eh->eh_entries)); | |
1539 | path = npath; | |
1540 | goto repeat; | |
1541 | } | |
1542 | ext_debug("next leaf has no free space(%d,%d)\n", | |
1543 | le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max)); | |
1544 | } | |
1545 | ||
1546 | /* | |
d0d856e8 RD |
1547 | * There is no free space in the found leaf. |
1548 | * We're gonna add a new leaf in the tree. | |
a86c6181 AT |
1549 | */ |
1550 | err = ext4_ext_create_new_leaf(handle, inode, path, newext); | |
1551 | if (err) | |
1552 | goto cleanup; | |
1553 | depth = ext_depth(inode); | |
1554 | eh = path[depth].p_hdr; | |
1555 | ||
1556 | has_space: | |
1557 | nearex = path[depth].p_ext; | |
1558 | ||
7e028976 AM |
1559 | err = ext4_ext_get_access(handle, inode, path + depth); |
1560 | if (err) | |
a86c6181 AT |
1561 | goto cleanup; |
1562 | ||
1563 | if (!nearex) { | |
1564 | /* there is no extent in this leaf, create first one */ | |
2ae02107 | 1565 | ext_debug("first extent in the leaf: %d:%llu:%d\n", |
8c55e204 DK |
1566 | le32_to_cpu(newext->ee_block), |
1567 | ext_pblock(newext), | |
a2df2a63 | 1568 | ext4_ext_get_actual_len(newext)); |
a86c6181 AT |
1569 | path[depth].p_ext = EXT_FIRST_EXTENT(eh); |
1570 | } else if (le32_to_cpu(newext->ee_block) | |
8c55e204 | 1571 | > le32_to_cpu(nearex->ee_block)) { |
a86c6181 AT |
1572 | /* BUG_ON(newext->ee_block == nearex->ee_block); */ |
1573 | if (nearex != EXT_LAST_EXTENT(eh)) { | |
1574 | len = EXT_MAX_EXTENT(eh) - nearex; | |
1575 | len = (len - 1) * sizeof(struct ext4_extent); | |
1576 | len = len < 0 ? 0 : len; | |
2ae02107 | 1577 | ext_debug("insert %d:%llu:%d after: nearest 0x%p, " |
a86c6181 | 1578 | "move %d from 0x%p to 0x%p\n", |
8c55e204 DK |
1579 | le32_to_cpu(newext->ee_block), |
1580 | ext_pblock(newext), | |
a2df2a63 | 1581 | ext4_ext_get_actual_len(newext), |
a86c6181 AT |
1582 | nearex, len, nearex + 1, nearex + 2); |
1583 | memmove(nearex + 2, nearex + 1, len); | |
1584 | } | |
1585 | path[depth].p_ext = nearex + 1; | |
1586 | } else { | |
1587 | BUG_ON(newext->ee_block == nearex->ee_block); | |
1588 | len = (EXT_MAX_EXTENT(eh) - nearex) * sizeof(struct ext4_extent); | |
1589 | len = len < 0 ? 0 : len; | |
2ae02107 | 1590 | ext_debug("insert %d:%llu:%d before: nearest 0x%p, " |
a86c6181 AT |
1591 | "move %d from 0x%p to 0x%p\n", |
1592 | le32_to_cpu(newext->ee_block), | |
f65e6fba | 1593 | ext_pblock(newext), |
a2df2a63 | 1594 | ext4_ext_get_actual_len(newext), |
a86c6181 AT |
1595 | nearex, len, nearex + 1, nearex + 2); |
1596 | memmove(nearex + 1, nearex, len); | |
1597 | path[depth].p_ext = nearex; | |
1598 | } | |
1599 | ||
e8546d06 | 1600 | le16_add_cpu(&eh->eh_entries, 1); |
a86c6181 AT |
1601 | nearex = path[depth].p_ext; |
1602 | nearex->ee_block = newext->ee_block; | |
b377611d | 1603 | ext4_ext_store_pblock(nearex, ext_pblock(newext)); |
a86c6181 | 1604 | nearex->ee_len = newext->ee_len; |
a86c6181 AT |
1605 | |
1606 | merge: | |
1607 | /* try to merge extents to the right */ | |
56055d3a | 1608 | ext4_ext_try_to_merge(inode, path, nearex); |
a86c6181 AT |
1609 | |
1610 | /* try to merge extents to the left */ | |
1611 | ||
1612 | /* time to correct all indexes above */ | |
1613 | err = ext4_ext_correct_indexes(handle, inode, path); | |
1614 | if (err) | |
1615 | goto cleanup; | |
1616 | ||
1617 | err = ext4_ext_dirty(handle, inode, path + depth); | |
1618 | ||
1619 | cleanup: | |
1620 | if (npath) { | |
1621 | ext4_ext_drop_refs(npath); | |
1622 | kfree(npath); | |
1623 | } | |
1624 | ext4_ext_tree_changed(inode); | |
1625 | ext4_ext_invalidate_cache(inode); | |
1626 | return err; | |
1627 | } | |
1628 | ||
09b88252 | 1629 | static void |
725d26d3 | 1630 | ext4_ext_put_in_cache(struct inode *inode, ext4_lblk_t block, |
dd54567a | 1631 | __u32 len, ext4_fsblk_t start, int type) |
a86c6181 AT |
1632 | { |
1633 | struct ext4_ext_cache *cex; | |
1634 | BUG_ON(len == 0); | |
1635 | cex = &EXT4_I(inode)->i_cached_extent; | |
1636 | cex->ec_type = type; | |
1637 | cex->ec_block = block; | |
1638 | cex->ec_len = len; | |
1639 | cex->ec_start = start; | |
1640 | } | |
1641 | ||
1642 | /* | |
d0d856e8 RD |
1643 | * ext4_ext_put_gap_in_cache: |
1644 | * calculate boundaries of the gap that the requested block fits into | |
a86c6181 AT |
1645 | * and cache this gap |
1646 | */ | |
09b88252 | 1647 | static void |
a86c6181 | 1648 | ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path, |
725d26d3 | 1649 | ext4_lblk_t block) |
a86c6181 AT |
1650 | { |
1651 | int depth = ext_depth(inode); | |
725d26d3 AK |
1652 | unsigned long len; |
1653 | ext4_lblk_t lblock; | |
a86c6181 AT |
1654 | struct ext4_extent *ex; |
1655 | ||
1656 | ex = path[depth].p_ext; | |
1657 | if (ex == NULL) { | |
1658 | /* there is no extent yet, so gap is [0;-] */ | |
1659 | lblock = 0; | |
1660 | len = EXT_MAX_BLOCK; | |
1661 | ext_debug("cache gap(whole file):"); | |
1662 | } else if (block < le32_to_cpu(ex->ee_block)) { | |
1663 | lblock = block; | |
1664 | len = le32_to_cpu(ex->ee_block) - block; | |
bba90743 ES |
1665 | ext_debug("cache gap(before): %u [%u:%u]", |
1666 | block, | |
1667 | le32_to_cpu(ex->ee_block), | |
1668 | ext4_ext_get_actual_len(ex)); | |
a86c6181 | 1669 | } else if (block >= le32_to_cpu(ex->ee_block) |
a2df2a63 | 1670 | + ext4_ext_get_actual_len(ex)) { |
725d26d3 | 1671 | ext4_lblk_t next; |
8c55e204 | 1672 | lblock = le32_to_cpu(ex->ee_block) |
a2df2a63 | 1673 | + ext4_ext_get_actual_len(ex); |
725d26d3 AK |
1674 | |
1675 | next = ext4_ext_next_allocated_block(path); | |
bba90743 ES |
1676 | ext_debug("cache gap(after): [%u:%u] %u", |
1677 | le32_to_cpu(ex->ee_block), | |
1678 | ext4_ext_get_actual_len(ex), | |
1679 | block); | |
725d26d3 AK |
1680 | BUG_ON(next == lblock); |
1681 | len = next - lblock; | |
a86c6181 AT |
1682 | } else { |
1683 | lblock = len = 0; | |
1684 | BUG(); | |
1685 | } | |
1686 | ||
bba90743 | 1687 | ext_debug(" -> %u:%lu\n", lblock, len); |
a86c6181 AT |
1688 | ext4_ext_put_in_cache(inode, lblock, len, 0, EXT4_EXT_CACHE_GAP); |
1689 | } | |
1690 | ||
09b88252 | 1691 | static int |
725d26d3 | 1692 | ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block, |
a86c6181 AT |
1693 | struct ext4_extent *ex) |
1694 | { | |
1695 | struct ext4_ext_cache *cex; | |
1696 | ||
1697 | cex = &EXT4_I(inode)->i_cached_extent; | |
1698 | ||
1699 | /* has cache valid data? */ | |
1700 | if (cex->ec_type == EXT4_EXT_CACHE_NO) | |
1701 | return EXT4_EXT_CACHE_NO; | |
1702 | ||
1703 | BUG_ON(cex->ec_type != EXT4_EXT_CACHE_GAP && | |
1704 | cex->ec_type != EXT4_EXT_CACHE_EXTENT); | |
1705 | if (block >= cex->ec_block && block < cex->ec_block + cex->ec_len) { | |
8c55e204 | 1706 | ex->ee_block = cpu_to_le32(cex->ec_block); |
f65e6fba | 1707 | ext4_ext_store_pblock(ex, cex->ec_start); |
8c55e204 | 1708 | ex->ee_len = cpu_to_le16(cex->ec_len); |
bba90743 ES |
1709 | ext_debug("%u cached by %u:%u:%llu\n", |
1710 | block, | |
1711 | cex->ec_block, cex->ec_len, cex->ec_start); | |
a86c6181 AT |
1712 | return cex->ec_type; |
1713 | } | |
1714 | ||
1715 | /* not in cache */ | |
1716 | return EXT4_EXT_CACHE_NO; | |
1717 | } | |
1718 | ||
1719 | /* | |
d0d856e8 RD |
1720 | * ext4_ext_rm_idx: |
1721 | * removes index from the index block. | |
1722 | * It's used in truncate case only, thus all requests are for | |
1723 | * last index in the block only. | |
a86c6181 | 1724 | */ |
1d03ec98 | 1725 | static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode, |
a86c6181 AT |
1726 | struct ext4_ext_path *path) |
1727 | { | |
1728 | struct buffer_head *bh; | |
1729 | int err; | |
f65e6fba | 1730 | ext4_fsblk_t leaf; |
a86c6181 AT |
1731 | |
1732 | /* free index block */ | |
1733 | path--; | |
f65e6fba | 1734 | leaf = idx_pblock(path->p_idx); |
a86c6181 | 1735 | BUG_ON(path->p_hdr->eh_entries == 0); |
7e028976 AM |
1736 | err = ext4_ext_get_access(handle, inode, path); |
1737 | if (err) | |
a86c6181 | 1738 | return err; |
e8546d06 | 1739 | le16_add_cpu(&path->p_hdr->eh_entries, -1); |
7e028976 AM |
1740 | err = ext4_ext_dirty(handle, inode, path); |
1741 | if (err) | |
a86c6181 | 1742 | return err; |
2ae02107 | 1743 | ext_debug("index is empty, remove it, free block %llu\n", leaf); |
a86c6181 AT |
1744 | bh = sb_find_get_block(inode->i_sb, leaf); |
1745 | ext4_forget(handle, 1, inode, bh, leaf); | |
c9de560d | 1746 | ext4_free_blocks(handle, inode, leaf, 1, 1); |
a86c6181 AT |
1747 | return err; |
1748 | } | |
1749 | ||
1750 | /* | |
ee12b630 MC |
1751 | * ext4_ext_calc_credits_for_single_extent: |
1752 | * This routine returns max. credits that needed to insert an extent | |
1753 | * to the extent tree. | |
1754 | * When pass the actual path, the caller should calculate credits | |
1755 | * under i_data_sem. | |
a86c6181 | 1756 | */ |
525f4ed8 | 1757 | int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks, |
a86c6181 AT |
1758 | struct ext4_ext_path *path) |
1759 | { | |
a86c6181 | 1760 | if (path) { |
ee12b630 | 1761 | int depth = ext_depth(inode); |
f3bd1f3f | 1762 | int ret = 0; |
ee12b630 | 1763 | |
a86c6181 | 1764 | /* probably there is space in leaf? */ |
a86c6181 | 1765 | if (le16_to_cpu(path[depth].p_hdr->eh_entries) |
ee12b630 | 1766 | < le16_to_cpu(path[depth].p_hdr->eh_max)) { |
a86c6181 | 1767 | |
ee12b630 MC |
1768 | /* |
1769 | * There are some space in the leaf tree, no | |
1770 | * need to account for leaf block credit | |
1771 | * | |
1772 | * bitmaps and block group descriptor blocks | |
1773 | * and other metadat blocks still need to be | |
1774 | * accounted. | |
1775 | */ | |
525f4ed8 | 1776 | /* 1 bitmap, 1 block group descriptor */ |
ee12b630 MC |
1777 | ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb); |
1778 | } | |
1779 | } | |
a86c6181 | 1780 | |
525f4ed8 | 1781 | return ext4_chunk_trans_blocks(inode, nrblocks); |
ee12b630 | 1782 | } |
a86c6181 | 1783 | |
ee12b630 MC |
1784 | /* |
1785 | * How many index/leaf blocks need to change/allocate to modify nrblocks? | |
1786 | * | |
1787 | * if nrblocks are fit in a single extent (chunk flag is 1), then | |
1788 | * in the worse case, each tree level index/leaf need to be changed | |
1789 | * if the tree split due to insert a new extent, then the old tree | |
1790 | * index/leaf need to be updated too | |
1791 | * | |
1792 | * If the nrblocks are discontiguous, they could cause | |
1793 | * the whole tree split more than once, but this is really rare. | |
1794 | */ | |
525f4ed8 | 1795 | int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks, int chunk) |
ee12b630 MC |
1796 | { |
1797 | int index; | |
1798 | int depth = ext_depth(inode); | |
a86c6181 | 1799 | |
ee12b630 MC |
1800 | if (chunk) |
1801 | index = depth * 2; | |
1802 | else | |
1803 | index = depth * 3; | |
a86c6181 | 1804 | |
ee12b630 | 1805 | return index; |
a86c6181 AT |
1806 | } |
1807 | ||
1808 | static int ext4_remove_blocks(handle_t *handle, struct inode *inode, | |
1809 | struct ext4_extent *ex, | |
725d26d3 | 1810 | ext4_lblk_t from, ext4_lblk_t to) |
a86c6181 AT |
1811 | { |
1812 | struct buffer_head *bh; | |
a2df2a63 | 1813 | unsigned short ee_len = ext4_ext_get_actual_len(ex); |
c9de560d | 1814 | int i, metadata = 0; |
a86c6181 | 1815 | |
c9de560d AT |
1816 | if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode)) |
1817 | metadata = 1; | |
a86c6181 AT |
1818 | #ifdef EXTENTS_STATS |
1819 | { | |
1820 | struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); | |
a86c6181 AT |
1821 | spin_lock(&sbi->s_ext_stats_lock); |
1822 | sbi->s_ext_blocks += ee_len; | |
1823 | sbi->s_ext_extents++; | |
1824 | if (ee_len < sbi->s_ext_min) | |
1825 | sbi->s_ext_min = ee_len; | |
1826 | if (ee_len > sbi->s_ext_max) | |
1827 | sbi->s_ext_max = ee_len; | |
1828 | if (ext_depth(inode) > sbi->s_depth_max) | |
1829 | sbi->s_depth_max = ext_depth(inode); | |
1830 | spin_unlock(&sbi->s_ext_stats_lock); | |
1831 | } | |
1832 | #endif | |
1833 | if (from >= le32_to_cpu(ex->ee_block) | |
a2df2a63 | 1834 | && to == le32_to_cpu(ex->ee_block) + ee_len - 1) { |
a86c6181 | 1835 | /* tail removal */ |
725d26d3 | 1836 | ext4_lblk_t num; |
f65e6fba | 1837 | ext4_fsblk_t start; |
725d26d3 | 1838 | |
a2df2a63 AA |
1839 | num = le32_to_cpu(ex->ee_block) + ee_len - from; |
1840 | start = ext_pblock(ex) + ee_len - num; | |
725d26d3 | 1841 | ext_debug("free last %u blocks starting %llu\n", num, start); |
a86c6181 AT |
1842 | for (i = 0; i < num; i++) { |
1843 | bh = sb_find_get_block(inode->i_sb, start + i); | |
1844 | ext4_forget(handle, 0, inode, bh, start + i); | |
1845 | } | |
c9de560d | 1846 | ext4_free_blocks(handle, inode, start, num, metadata); |
a86c6181 | 1847 | } else if (from == le32_to_cpu(ex->ee_block) |
a2df2a63 | 1848 | && to <= le32_to_cpu(ex->ee_block) + ee_len - 1) { |
725d26d3 | 1849 | printk(KERN_INFO "strange request: removal %u-%u from %u:%u\n", |
a2df2a63 | 1850 | from, to, le32_to_cpu(ex->ee_block), ee_len); |
a86c6181 | 1851 | } else { |
725d26d3 AK |
1852 | printk(KERN_INFO "strange request: removal(2) " |
1853 | "%u-%u from %u:%u\n", | |
1854 | from, to, le32_to_cpu(ex->ee_block), ee_len); | |
a86c6181 AT |
1855 | } |
1856 | return 0; | |
1857 | } | |
1858 | ||
1859 | static int | |
1860 | ext4_ext_rm_leaf(handle_t *handle, struct inode *inode, | |
725d26d3 | 1861 | struct ext4_ext_path *path, ext4_lblk_t start) |
a86c6181 AT |
1862 | { |
1863 | int err = 0, correct_index = 0; | |
1864 | int depth = ext_depth(inode), credits; | |
1865 | struct ext4_extent_header *eh; | |
725d26d3 AK |
1866 | ext4_lblk_t a, b, block; |
1867 | unsigned num; | |
1868 | ext4_lblk_t ex_ee_block; | |
a86c6181 | 1869 | unsigned short ex_ee_len; |
a2df2a63 | 1870 | unsigned uninitialized = 0; |
a86c6181 AT |
1871 | struct ext4_extent *ex; |
1872 | ||
c29c0ae7 | 1873 | /* the header must be checked already in ext4_ext_remove_space() */ |
725d26d3 | 1874 | ext_debug("truncate since %u in leaf\n", start); |
a86c6181 AT |
1875 | if (!path[depth].p_hdr) |
1876 | path[depth].p_hdr = ext_block_hdr(path[depth].p_bh); | |
1877 | eh = path[depth].p_hdr; | |
1878 | BUG_ON(eh == NULL); | |
a86c6181 AT |
1879 | |
1880 | /* find where to start removing */ | |
1881 | ex = EXT_LAST_EXTENT(eh); | |
1882 | ||
1883 | ex_ee_block = le32_to_cpu(ex->ee_block); | |
a2df2a63 AA |
1884 | if (ext4_ext_is_uninitialized(ex)) |
1885 | uninitialized = 1; | |
1886 | ex_ee_len = ext4_ext_get_actual_len(ex); | |
a86c6181 AT |
1887 | |
1888 | while (ex >= EXT_FIRST_EXTENT(eh) && | |
1889 | ex_ee_block + ex_ee_len > start) { | |
1890 | ext_debug("remove ext %lu:%u\n", ex_ee_block, ex_ee_len); | |
1891 | path[depth].p_ext = ex; | |
1892 | ||
1893 | a = ex_ee_block > start ? ex_ee_block : start; | |
1894 | b = ex_ee_block + ex_ee_len - 1 < EXT_MAX_BLOCK ? | |
1895 | ex_ee_block + ex_ee_len - 1 : EXT_MAX_BLOCK; | |
1896 | ||
1897 | ext_debug(" border %u:%u\n", a, b); | |
1898 | ||
1899 | if (a != ex_ee_block && b != ex_ee_block + ex_ee_len - 1) { | |
1900 | block = 0; | |
1901 | num = 0; | |
1902 | BUG(); | |
1903 | } else if (a != ex_ee_block) { | |
1904 | /* remove tail of the extent */ | |
1905 | block = ex_ee_block; | |
1906 | num = a - block; | |
1907 | } else if (b != ex_ee_block + ex_ee_len - 1) { | |
1908 | /* remove head of the extent */ | |
1909 | block = a; | |
1910 | num = b - a; | |
1911 | /* there is no "make a hole" API yet */ | |
1912 | BUG(); | |
1913 | } else { | |
1914 | /* remove whole extent: excellent! */ | |
1915 | block = ex_ee_block; | |
1916 | num = 0; | |
1917 | BUG_ON(a != ex_ee_block); | |
1918 | BUG_ON(b != ex_ee_block + ex_ee_len - 1); | |
1919 | } | |
1920 | ||
34071da7 TT |
1921 | /* |
1922 | * 3 for leaf, sb, and inode plus 2 (bmap and group | |
1923 | * descriptor) for each block group; assume two block | |
1924 | * groups plus ex_ee_len/blocks_per_block_group for | |
1925 | * the worst case | |
1926 | */ | |
1927 | credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb)); | |
a86c6181 AT |
1928 | if (ex == EXT_FIRST_EXTENT(eh)) { |
1929 | correct_index = 1; | |
1930 | credits += (ext_depth(inode)) + 1; | |
1931 | } | |
a86c6181 | 1932 | credits += 2 * EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb); |
a86c6181 | 1933 | |
9102e4fa SF |
1934 | err = ext4_ext_journal_restart(handle, credits); |
1935 | if (err) | |
a86c6181 | 1936 | goto out; |
a86c6181 AT |
1937 | |
1938 | err = ext4_ext_get_access(handle, inode, path + depth); | |
1939 | if (err) | |
1940 | goto out; | |
1941 | ||
1942 | err = ext4_remove_blocks(handle, inode, ex, a, b); | |
1943 | if (err) | |
1944 | goto out; | |
1945 | ||
1946 | if (num == 0) { | |
d0d856e8 | 1947 | /* this extent is removed; mark slot entirely unused */ |
f65e6fba | 1948 | ext4_ext_store_pblock(ex, 0); |
e8546d06 | 1949 | le16_add_cpu(&eh->eh_entries, -1); |
a86c6181 AT |
1950 | } |
1951 | ||
1952 | ex->ee_block = cpu_to_le32(block); | |
1953 | ex->ee_len = cpu_to_le16(num); | |
749269fa AA |
1954 | /* |
1955 | * Do not mark uninitialized if all the blocks in the | |
1956 | * extent have been removed. | |
1957 | */ | |
1958 | if (uninitialized && num) | |
a2df2a63 | 1959 | ext4_ext_mark_uninitialized(ex); |
a86c6181 AT |
1960 | |
1961 | err = ext4_ext_dirty(handle, inode, path + depth); | |
1962 | if (err) | |
1963 | goto out; | |
1964 | ||
2ae02107 | 1965 | ext_debug("new extent: %u:%u:%llu\n", block, num, |
f65e6fba | 1966 | ext_pblock(ex)); |
a86c6181 AT |
1967 | ex--; |
1968 | ex_ee_block = le32_to_cpu(ex->ee_block); | |
a2df2a63 | 1969 | ex_ee_len = ext4_ext_get_actual_len(ex); |
a86c6181 AT |
1970 | } |
1971 | ||
1972 | if (correct_index && eh->eh_entries) | |
1973 | err = ext4_ext_correct_indexes(handle, inode, path); | |
1974 | ||
1975 | /* if this leaf is free, then we should | |
1976 | * remove it from index block above */ | |
1977 | if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL) | |
1978 | err = ext4_ext_rm_idx(handle, inode, path + depth); | |
1979 | ||
1980 | out: | |
1981 | return err; | |
1982 | } | |
1983 | ||
1984 | /* | |
d0d856e8 RD |
1985 | * ext4_ext_more_to_rm: |
1986 | * returns 1 if current index has to be freed (even partial) | |
a86c6181 | 1987 | */ |
09b88252 | 1988 | static int |
a86c6181 AT |
1989 | ext4_ext_more_to_rm(struct ext4_ext_path *path) |
1990 | { | |
1991 | BUG_ON(path->p_idx == NULL); | |
1992 | ||
1993 | if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr)) | |
1994 | return 0; | |
1995 | ||
1996 | /* | |
d0d856e8 | 1997 | * if truncate on deeper level happened, it wasn't partial, |
a86c6181 AT |
1998 | * so we have to consider current index for truncation |
1999 | */ | |
2000 | if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block) | |
2001 | return 0; | |
2002 | return 1; | |
2003 | } | |
2004 | ||
1d03ec98 | 2005 | static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start) |
a86c6181 AT |
2006 | { |
2007 | struct super_block *sb = inode->i_sb; | |
2008 | int depth = ext_depth(inode); | |
2009 | struct ext4_ext_path *path; | |
2010 | handle_t *handle; | |
2011 | int i = 0, err = 0; | |
2012 | ||
725d26d3 | 2013 | ext_debug("truncate since %u\n", start); |
a86c6181 AT |
2014 | |
2015 | /* probably first extent we're gonna free will be last in block */ | |
2016 | handle = ext4_journal_start(inode, depth + 1); | |
2017 | if (IS_ERR(handle)) | |
2018 | return PTR_ERR(handle); | |
2019 | ||
2020 | ext4_ext_invalidate_cache(inode); | |
2021 | ||
2022 | /* | |
d0d856e8 RD |
2023 | * We start scanning from right side, freeing all the blocks |
2024 | * after i_size and walking into the tree depth-wise. | |
a86c6181 | 2025 | */ |
216553c4 | 2026 | path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1), GFP_NOFS); |
a86c6181 AT |
2027 | if (path == NULL) { |
2028 | ext4_journal_stop(handle); | |
2029 | return -ENOMEM; | |
2030 | } | |
a86c6181 | 2031 | path[0].p_hdr = ext_inode_hdr(inode); |
c29c0ae7 | 2032 | if (ext4_ext_check_header(inode, path[0].p_hdr, depth)) { |
a86c6181 AT |
2033 | err = -EIO; |
2034 | goto out; | |
2035 | } | |
2036 | path[0].p_depth = depth; | |
2037 | ||
2038 | while (i >= 0 && err == 0) { | |
2039 | if (i == depth) { | |
2040 | /* this is leaf block */ | |
2041 | err = ext4_ext_rm_leaf(handle, inode, path, start); | |
d0d856e8 | 2042 | /* root level has p_bh == NULL, brelse() eats this */ |
a86c6181 AT |
2043 | brelse(path[i].p_bh); |
2044 | path[i].p_bh = NULL; | |
2045 | i--; | |
2046 | continue; | |
2047 | } | |
2048 | ||
2049 | /* this is index block */ | |
2050 | if (!path[i].p_hdr) { | |
2051 | ext_debug("initialize header\n"); | |
2052 | path[i].p_hdr = ext_block_hdr(path[i].p_bh); | |
a86c6181 AT |
2053 | } |
2054 | ||
a86c6181 | 2055 | if (!path[i].p_idx) { |
d0d856e8 | 2056 | /* this level hasn't been touched yet */ |
a86c6181 AT |
2057 | path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr); |
2058 | path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1; | |
2059 | ext_debug("init index ptr: hdr 0x%p, num %d\n", | |
2060 | path[i].p_hdr, | |
2061 | le16_to_cpu(path[i].p_hdr->eh_entries)); | |
2062 | } else { | |
d0d856e8 | 2063 | /* we were already here, see at next index */ |
a86c6181 AT |
2064 | path[i].p_idx--; |
2065 | } | |
2066 | ||
2067 | ext_debug("level %d - index, first 0x%p, cur 0x%p\n", | |
2068 | i, EXT_FIRST_INDEX(path[i].p_hdr), | |
2069 | path[i].p_idx); | |
2070 | if (ext4_ext_more_to_rm(path + i)) { | |
c29c0ae7 | 2071 | struct buffer_head *bh; |
a86c6181 | 2072 | /* go to the next level */ |
2ae02107 | 2073 | ext_debug("move to level %d (block %llu)\n", |
f65e6fba | 2074 | i + 1, idx_pblock(path[i].p_idx)); |
a86c6181 | 2075 | memset(path + i + 1, 0, sizeof(*path)); |
c29c0ae7 AT |
2076 | bh = sb_bread(sb, idx_pblock(path[i].p_idx)); |
2077 | if (!bh) { | |
a86c6181 AT |
2078 | /* should we reset i_size? */ |
2079 | err = -EIO; | |
2080 | break; | |
2081 | } | |
c29c0ae7 AT |
2082 | if (WARN_ON(i + 1 > depth)) { |
2083 | err = -EIO; | |
2084 | break; | |
2085 | } | |
2086 | if (ext4_ext_check_header(inode, ext_block_hdr(bh), | |
2087 | depth - i - 1)) { | |
2088 | err = -EIO; | |
2089 | break; | |
2090 | } | |
2091 | path[i + 1].p_bh = bh; | |
a86c6181 | 2092 | |
d0d856e8 RD |
2093 | /* save actual number of indexes since this |
2094 | * number is changed at the next iteration */ | |
a86c6181 AT |
2095 | path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries); |
2096 | i++; | |
2097 | } else { | |
d0d856e8 | 2098 | /* we finished processing this index, go up */ |
a86c6181 | 2099 | if (path[i].p_hdr->eh_entries == 0 && i > 0) { |
d0d856e8 | 2100 | /* index is empty, remove it; |
a86c6181 AT |
2101 | * handle must be already prepared by the |
2102 | * truncatei_leaf() */ | |
2103 | err = ext4_ext_rm_idx(handle, inode, path + i); | |
2104 | } | |
d0d856e8 | 2105 | /* root level has p_bh == NULL, brelse() eats this */ |
a86c6181 AT |
2106 | brelse(path[i].p_bh); |
2107 | path[i].p_bh = NULL; | |
2108 | i--; | |
2109 | ext_debug("return to level %d\n", i); | |
2110 | } | |
2111 | } | |
2112 | ||
2113 | /* TODO: flexible tree reduction should be here */ | |
2114 | if (path->p_hdr->eh_entries == 0) { | |
2115 | /* | |
d0d856e8 RD |
2116 | * truncate to zero freed all the tree, |
2117 | * so we need to correct eh_depth | |
a86c6181 AT |
2118 | */ |
2119 | err = ext4_ext_get_access(handle, inode, path); | |
2120 | if (err == 0) { | |
2121 | ext_inode_hdr(inode)->eh_depth = 0; | |
2122 | ext_inode_hdr(inode)->eh_max = | |
2123 | cpu_to_le16(ext4_ext_space_root(inode)); | |
2124 | err = ext4_ext_dirty(handle, inode, path); | |
2125 | } | |
2126 | } | |
2127 | out: | |
2128 | ext4_ext_tree_changed(inode); | |
2129 | ext4_ext_drop_refs(path); | |
2130 | kfree(path); | |
2131 | ext4_journal_stop(handle); | |
2132 | ||
2133 | return err; | |
2134 | } | |
2135 | ||
2136 | /* | |
2137 | * called at mount time | |
2138 | */ | |
2139 | void ext4_ext_init(struct super_block *sb) | |
2140 | { | |
2141 | /* | |
2142 | * possible initialization would be here | |
2143 | */ | |
2144 | ||
2145 | if (test_opt(sb, EXTENTS)) { | |
4776004f | 2146 | printk(KERN_INFO "EXT4-fs: file extents enabled"); |
bbf2f9fb RD |
2147 | #ifdef AGGRESSIVE_TEST |
2148 | printk(", aggressive tests"); | |
a86c6181 AT |
2149 | #endif |
2150 | #ifdef CHECK_BINSEARCH | |
2151 | printk(", check binsearch"); | |
2152 | #endif | |
2153 | #ifdef EXTENTS_STATS | |
2154 | printk(", stats"); | |
2155 | #endif | |
2156 | printk("\n"); | |
2157 | #ifdef EXTENTS_STATS | |
2158 | spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock); | |
2159 | EXT4_SB(sb)->s_ext_min = 1 << 30; | |
2160 | EXT4_SB(sb)->s_ext_max = 0; | |
2161 | #endif | |
2162 | } | |
2163 | } | |
2164 | ||
2165 | /* | |
2166 | * called at umount time | |
2167 | */ | |
2168 | void ext4_ext_release(struct super_block *sb) | |
2169 | { | |
2170 | if (!test_opt(sb, EXTENTS)) | |
2171 | return; | |
2172 | ||
2173 | #ifdef EXTENTS_STATS | |
2174 | if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) { | |
2175 | struct ext4_sb_info *sbi = EXT4_SB(sb); | |
2176 | printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n", | |
2177 | sbi->s_ext_blocks, sbi->s_ext_extents, | |
2178 | sbi->s_ext_blocks / sbi->s_ext_extents); | |
2179 | printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n", | |
2180 | sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max); | |
2181 | } | |
2182 | #endif | |
2183 | } | |
2184 | ||
093a088b AK |
2185 | static void bi_complete(struct bio *bio, int error) |
2186 | { | |
2187 | complete((struct completion *)bio->bi_private); | |
2188 | } | |
2189 | ||
2190 | /* FIXME!! we need to try to merge to left or right after zero-out */ | |
2191 | static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex) | |
2192 | { | |
2193 | int ret = -EIO; | |
2194 | struct bio *bio; | |
2195 | int blkbits, blocksize; | |
2196 | sector_t ee_pblock; | |
2197 | struct completion event; | |
2198 | unsigned int ee_len, len, done, offset; | |
2199 | ||
2200 | ||
2201 | blkbits = inode->i_blkbits; | |
2202 | blocksize = inode->i_sb->s_blocksize; | |
2203 | ee_len = ext4_ext_get_actual_len(ex); | |
2204 | ee_pblock = ext_pblock(ex); | |
2205 | ||
2206 | /* convert ee_pblock to 512 byte sectors */ | |
2207 | ee_pblock = ee_pblock << (blkbits - 9); | |
2208 | ||
2209 | while (ee_len > 0) { | |
2210 | ||
2211 | if (ee_len > BIO_MAX_PAGES) | |
2212 | len = BIO_MAX_PAGES; | |
2213 | else | |
2214 | len = ee_len; | |
2215 | ||
2216 | bio = bio_alloc(GFP_NOIO, len); | |
2217 | if (!bio) | |
2218 | return -ENOMEM; | |
2219 | bio->bi_sector = ee_pblock; | |
2220 | bio->bi_bdev = inode->i_sb->s_bdev; | |
2221 | ||
2222 | done = 0; | |
2223 | offset = 0; | |
2224 | while (done < len) { | |
2225 | ret = bio_add_page(bio, ZERO_PAGE(0), | |
2226 | blocksize, offset); | |
2227 | if (ret != blocksize) { | |
2228 | /* | |
2229 | * We can't add any more pages because of | |
2230 | * hardware limitations. Start a new bio. | |
2231 | */ | |
2232 | break; | |
2233 | } | |
2234 | done++; | |
2235 | offset += blocksize; | |
2236 | if (offset >= PAGE_CACHE_SIZE) | |
2237 | offset = 0; | |
2238 | } | |
2239 | ||
2240 | init_completion(&event); | |
2241 | bio->bi_private = &event; | |
2242 | bio->bi_end_io = bi_complete; | |
2243 | submit_bio(WRITE, bio); | |
2244 | wait_for_completion(&event); | |
2245 | ||
2246 | if (test_bit(BIO_UPTODATE, &bio->bi_flags)) | |
2247 | ret = 0; | |
2248 | else { | |
2249 | ret = -EIO; | |
2250 | break; | |
2251 | } | |
2252 | bio_put(bio); | |
2253 | ee_len -= done; | |
2254 | ee_pblock += done << (blkbits - 9); | |
2255 | } | |
2256 | return ret; | |
2257 | } | |
2258 | ||
3977c965 AK |
2259 | #define EXT4_EXT_ZERO_LEN 7 |
2260 | ||
56055d3a AA |
2261 | /* |
2262 | * This function is called by ext4_ext_get_blocks() if someone tries to write | |
2263 | * to an uninitialized extent. It may result in splitting the uninitialized | |
2264 | * extent into multiple extents (upto three - one initialized and two | |
2265 | * uninitialized). | |
2266 | * There are three possibilities: | |
2267 | * a> There is no split required: Entire extent should be initialized | |
2268 | * b> Splits in two extents: Write is happening at either end of the extent | |
2269 | * c> Splits in three extents: Somone is writing in middle of the extent | |
2270 | */ | |
725d26d3 AK |
2271 | static int ext4_ext_convert_to_initialized(handle_t *handle, |
2272 | struct inode *inode, | |
2273 | struct ext4_ext_path *path, | |
2274 | ext4_lblk_t iblock, | |
2275 | unsigned long max_blocks) | |
56055d3a | 2276 | { |
95c3889c | 2277 | struct ext4_extent *ex, newex, orig_ex; |
56055d3a AA |
2278 | struct ext4_extent *ex1 = NULL; |
2279 | struct ext4_extent *ex2 = NULL; | |
2280 | struct ext4_extent *ex3 = NULL; | |
2281 | struct ext4_extent_header *eh; | |
725d26d3 AK |
2282 | ext4_lblk_t ee_block; |
2283 | unsigned int allocated, ee_len, depth; | |
56055d3a AA |
2284 | ext4_fsblk_t newblock; |
2285 | int err = 0; | |
2286 | int ret = 0; | |
2287 | ||
2288 | depth = ext_depth(inode); | |
2289 | eh = path[depth].p_hdr; | |
2290 | ex = path[depth].p_ext; | |
2291 | ee_block = le32_to_cpu(ex->ee_block); | |
2292 | ee_len = ext4_ext_get_actual_len(ex); | |
2293 | allocated = ee_len - (iblock - ee_block); | |
2294 | newblock = iblock - ee_block + ext_pblock(ex); | |
2295 | ex2 = ex; | |
95c3889c AK |
2296 | orig_ex.ee_block = ex->ee_block; |
2297 | orig_ex.ee_len = cpu_to_le16(ee_len); | |
2298 | ext4_ext_store_pblock(&orig_ex, ext_pblock(ex)); | |
56055d3a | 2299 | |
9df5643a AK |
2300 | err = ext4_ext_get_access(handle, inode, path + depth); |
2301 | if (err) | |
2302 | goto out; | |
3977c965 AK |
2303 | /* If extent has less than 2*EXT4_EXT_ZERO_LEN zerout directly */ |
2304 | if (ee_len <= 2*EXT4_EXT_ZERO_LEN) { | |
2305 | err = ext4_ext_zeroout(inode, &orig_ex); | |
2306 | if (err) | |
2307 | goto fix_extent_len; | |
2308 | /* update the extent length and mark as initialized */ | |
2309 | ex->ee_block = orig_ex.ee_block; | |
2310 | ex->ee_len = orig_ex.ee_len; | |
2311 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
2312 | ext4_ext_dirty(handle, inode, path + depth); | |
161e7b7c AK |
2313 | /* zeroed the full extent */ |
2314 | return allocated; | |
3977c965 | 2315 | } |
9df5643a | 2316 | |
56055d3a AA |
2317 | /* ex1: ee_block to iblock - 1 : uninitialized */ |
2318 | if (iblock > ee_block) { | |
2319 | ex1 = ex; | |
2320 | ex1->ee_len = cpu_to_le16(iblock - ee_block); | |
2321 | ext4_ext_mark_uninitialized(ex1); | |
2322 | ex2 = &newex; | |
2323 | } | |
2324 | /* | |
2325 | * for sanity, update the length of the ex2 extent before | |
2326 | * we insert ex3, if ex1 is NULL. This is to avoid temporary | |
2327 | * overlap of blocks. | |
2328 | */ | |
2329 | if (!ex1 && allocated > max_blocks) | |
2330 | ex2->ee_len = cpu_to_le16(max_blocks); | |
2331 | /* ex3: to ee_block + ee_len : uninitialised */ | |
2332 | if (allocated > max_blocks) { | |
2333 | unsigned int newdepth; | |
3977c965 AK |
2334 | /* If extent has less than EXT4_EXT_ZERO_LEN zerout directly */ |
2335 | if (allocated <= EXT4_EXT_ZERO_LEN) { | |
d03856bd AK |
2336 | /* |
2337 | * iblock == ee_block is handled by the zerouout | |
2338 | * at the beginning. | |
2339 | * Mark first half uninitialized. | |
3977c965 AK |
2340 | * Mark second half initialized and zero out the |
2341 | * initialized extent | |
2342 | */ | |
2343 | ex->ee_block = orig_ex.ee_block; | |
2344 | ex->ee_len = cpu_to_le16(ee_len - allocated); | |
2345 | ext4_ext_mark_uninitialized(ex); | |
2346 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
2347 | ext4_ext_dirty(handle, inode, path + depth); | |
2348 | ||
2349 | ex3 = &newex; | |
2350 | ex3->ee_block = cpu_to_le32(iblock); | |
2351 | ext4_ext_store_pblock(ex3, newblock); | |
2352 | ex3->ee_len = cpu_to_le16(allocated); | |
2353 | err = ext4_ext_insert_extent(handle, inode, path, ex3); | |
2354 | if (err == -ENOSPC) { | |
2355 | err = ext4_ext_zeroout(inode, &orig_ex); | |
2356 | if (err) | |
2357 | goto fix_extent_len; | |
2358 | ex->ee_block = orig_ex.ee_block; | |
2359 | ex->ee_len = orig_ex.ee_len; | |
2360 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
2361 | ext4_ext_dirty(handle, inode, path + depth); | |
d03856bd | 2362 | /* blocks available from iblock */ |
161e7b7c | 2363 | return allocated; |
3977c965 AK |
2364 | |
2365 | } else if (err) | |
2366 | goto fix_extent_len; | |
2367 | ||
161e7b7c AK |
2368 | /* |
2369 | * We need to zero out the second half because | |
2370 | * an fallocate request can update file size and | |
2371 | * converting the second half to initialized extent | |
2372 | * implies that we can leak some junk data to user | |
2373 | * space. | |
2374 | */ | |
2375 | err = ext4_ext_zeroout(inode, ex3); | |
2376 | if (err) { | |
2377 | /* | |
2378 | * We should actually mark the | |
2379 | * second half as uninit and return error | |
2380 | * Insert would have changed the extent | |
2381 | */ | |
2382 | depth = ext_depth(inode); | |
2383 | ext4_ext_drop_refs(path); | |
2384 | path = ext4_ext_find_extent(inode, | |
2385 | iblock, path); | |
2386 | if (IS_ERR(path)) { | |
2387 | err = PTR_ERR(path); | |
2388 | return err; | |
2389 | } | |
d03856bd | 2390 | /* get the second half extent details */ |
161e7b7c AK |
2391 | ex = path[depth].p_ext; |
2392 | err = ext4_ext_get_access(handle, inode, | |
2393 | path + depth); | |
2394 | if (err) | |
2395 | return err; | |
2396 | ext4_ext_mark_uninitialized(ex); | |
2397 | ext4_ext_dirty(handle, inode, path + depth); | |
2398 | return err; | |
2399 | } | |
2400 | ||
2401 | /* zeroed the second half */ | |
3977c965 AK |
2402 | return allocated; |
2403 | } | |
56055d3a AA |
2404 | ex3 = &newex; |
2405 | ex3->ee_block = cpu_to_le32(iblock + max_blocks); | |
2406 | ext4_ext_store_pblock(ex3, newblock + max_blocks); | |
2407 | ex3->ee_len = cpu_to_le16(allocated - max_blocks); | |
2408 | ext4_ext_mark_uninitialized(ex3); | |
2409 | err = ext4_ext_insert_extent(handle, inode, path, ex3); | |
093a088b AK |
2410 | if (err == -ENOSPC) { |
2411 | err = ext4_ext_zeroout(inode, &orig_ex); | |
2412 | if (err) | |
2413 | goto fix_extent_len; | |
2414 | /* update the extent length and mark as initialized */ | |
95c3889c AK |
2415 | ex->ee_block = orig_ex.ee_block; |
2416 | ex->ee_len = orig_ex.ee_len; | |
2417 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
95c3889c | 2418 | ext4_ext_dirty(handle, inode, path + depth); |
161e7b7c | 2419 | /* zeroed the full extent */ |
d03856bd | 2420 | /* blocks available from iblock */ |
161e7b7c | 2421 | return allocated; |
093a088b AK |
2422 | |
2423 | } else if (err) | |
2424 | goto fix_extent_len; | |
56055d3a AA |
2425 | /* |
2426 | * The depth, and hence eh & ex might change | |
2427 | * as part of the insert above. | |
2428 | */ | |
2429 | newdepth = ext_depth(inode); | |
95c3889c AK |
2430 | /* |
2431 | * update the extent length after successfull insert of the | |
2432 | * split extent | |
2433 | */ | |
2434 | orig_ex.ee_len = cpu_to_le16(ee_len - | |
2435 | ext4_ext_get_actual_len(ex3)); | |
d03856bd AK |
2436 | depth = newdepth; |
2437 | ext4_ext_drop_refs(path); | |
2438 | path = ext4_ext_find_extent(inode, iblock, path); | |
2439 | if (IS_ERR(path)) { | |
2440 | err = PTR_ERR(path); | |
2441 | goto out; | |
56055d3a | 2442 | } |
d03856bd AK |
2443 | eh = path[depth].p_hdr; |
2444 | ex = path[depth].p_ext; | |
2445 | if (ex2 != &newex) | |
2446 | ex2 = ex; | |
2447 | ||
2448 | err = ext4_ext_get_access(handle, inode, path + depth); | |
2449 | if (err) | |
2450 | goto out; | |
2451 | ||
56055d3a | 2452 | allocated = max_blocks; |
3977c965 AK |
2453 | |
2454 | /* If extent has less than EXT4_EXT_ZERO_LEN and we are trying | |
2455 | * to insert a extent in the middle zerout directly | |
2456 | * otherwise give the extent a chance to merge to left | |
2457 | */ | |
2458 | if (le16_to_cpu(orig_ex.ee_len) <= EXT4_EXT_ZERO_LEN && | |
2459 | iblock != ee_block) { | |
2460 | err = ext4_ext_zeroout(inode, &orig_ex); | |
2461 | if (err) | |
2462 | goto fix_extent_len; | |
2463 | /* update the extent length and mark as initialized */ | |
2464 | ex->ee_block = orig_ex.ee_block; | |
2465 | ex->ee_len = orig_ex.ee_len; | |
2466 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
2467 | ext4_ext_dirty(handle, inode, path + depth); | |
161e7b7c | 2468 | /* zero out the first half */ |
d03856bd | 2469 | /* blocks available from iblock */ |
161e7b7c | 2470 | return allocated; |
3977c965 | 2471 | } |
56055d3a AA |
2472 | } |
2473 | /* | |
2474 | * If there was a change of depth as part of the | |
2475 | * insertion of ex3 above, we need to update the length | |
2476 | * of the ex1 extent again here | |
2477 | */ | |
2478 | if (ex1 && ex1 != ex) { | |
2479 | ex1 = ex; | |
2480 | ex1->ee_len = cpu_to_le16(iblock - ee_block); | |
2481 | ext4_ext_mark_uninitialized(ex1); | |
2482 | ex2 = &newex; | |
2483 | } | |
2484 | /* ex2: iblock to iblock + maxblocks-1 : initialised */ | |
2485 | ex2->ee_block = cpu_to_le32(iblock); | |
56055d3a AA |
2486 | ext4_ext_store_pblock(ex2, newblock); |
2487 | ex2->ee_len = cpu_to_le16(allocated); | |
2488 | if (ex2 != ex) | |
2489 | goto insert; | |
56055d3a AA |
2490 | /* |
2491 | * New (initialized) extent starts from the first block | |
2492 | * in the current extent. i.e., ex2 == ex | |
2493 | * We have to see if it can be merged with the extent | |
2494 | * on the left. | |
2495 | */ | |
2496 | if (ex2 > EXT_FIRST_EXTENT(eh)) { | |
2497 | /* | |
2498 | * To merge left, pass "ex2 - 1" to try_to_merge(), | |
2499 | * since it merges towards right _only_. | |
2500 | */ | |
2501 | ret = ext4_ext_try_to_merge(inode, path, ex2 - 1); | |
2502 | if (ret) { | |
2503 | err = ext4_ext_correct_indexes(handle, inode, path); | |
2504 | if (err) | |
2505 | goto out; | |
2506 | depth = ext_depth(inode); | |
2507 | ex2--; | |
2508 | } | |
2509 | } | |
2510 | /* | |
2511 | * Try to Merge towards right. This might be required | |
2512 | * only when the whole extent is being written to. | |
2513 | * i.e. ex2 == ex and ex3 == NULL. | |
2514 | */ | |
2515 | if (!ex3) { | |
2516 | ret = ext4_ext_try_to_merge(inode, path, ex2); | |
2517 | if (ret) { | |
2518 | err = ext4_ext_correct_indexes(handle, inode, path); | |
2519 | if (err) | |
2520 | goto out; | |
2521 | } | |
2522 | } | |
2523 | /* Mark modified extent as dirty */ | |
2524 | err = ext4_ext_dirty(handle, inode, path + depth); | |
2525 | goto out; | |
2526 | insert: | |
2527 | err = ext4_ext_insert_extent(handle, inode, path, &newex); | |
093a088b AK |
2528 | if (err == -ENOSPC) { |
2529 | err = ext4_ext_zeroout(inode, &orig_ex); | |
2530 | if (err) | |
2531 | goto fix_extent_len; | |
2532 | /* update the extent length and mark as initialized */ | |
95c3889c AK |
2533 | ex->ee_block = orig_ex.ee_block; |
2534 | ex->ee_len = orig_ex.ee_len; | |
2535 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
95c3889c | 2536 | ext4_ext_dirty(handle, inode, path + depth); |
161e7b7c AK |
2537 | /* zero out the first half */ |
2538 | return allocated; | |
093a088b AK |
2539 | } else if (err) |
2540 | goto fix_extent_len; | |
56055d3a AA |
2541 | out: |
2542 | return err ? err : allocated; | |
093a088b AK |
2543 | |
2544 | fix_extent_len: | |
2545 | ex->ee_block = orig_ex.ee_block; | |
2546 | ex->ee_len = orig_ex.ee_len; | |
2547 | ext4_ext_store_pblock(ex, ext_pblock(&orig_ex)); | |
2548 | ext4_ext_mark_uninitialized(ex); | |
2549 | ext4_ext_dirty(handle, inode, path + depth); | |
2550 | return err; | |
56055d3a AA |
2551 | } |
2552 | ||
c278bfec | 2553 | /* |
f5ab0d1f MC |
2554 | * Block allocation/map/preallocation routine for extents based files |
2555 | * | |
2556 | * | |
c278bfec | 2557 | * Need to be called with |
0e855ac8 AK |
2558 | * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block |
2559 | * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem) | |
f5ab0d1f MC |
2560 | * |
2561 | * return > 0, number of of blocks already mapped/allocated | |
2562 | * if create == 0 and these are pre-allocated blocks | |
2563 | * buffer head is unmapped | |
2564 | * otherwise blocks are mapped | |
2565 | * | |
2566 | * return = 0, if plain look up failed (blocks have not been allocated) | |
2567 | * buffer head is unmapped | |
2568 | * | |
2569 | * return < 0, error case. | |
c278bfec | 2570 | */ |
f65e6fba | 2571 | int ext4_ext_get_blocks(handle_t *handle, struct inode *inode, |
725d26d3 | 2572 | ext4_lblk_t iblock, |
a86c6181 AT |
2573 | unsigned long max_blocks, struct buffer_head *bh_result, |
2574 | int create, int extend_disksize) | |
2575 | { | |
2576 | struct ext4_ext_path *path = NULL; | |
56055d3a | 2577 | struct ext4_extent_header *eh; |
a86c6181 | 2578 | struct ext4_extent newex, *ex; |
f65e6fba | 2579 | ext4_fsblk_t goal, newblock; |
56055d3a | 2580 | int err = 0, depth, ret; |
a86c6181 | 2581 | unsigned long allocated = 0; |
c9de560d | 2582 | struct ext4_allocation_request ar; |
61628a3f | 2583 | loff_t disksize; |
a86c6181 AT |
2584 | |
2585 | __clear_bit(BH_New, &bh_result->b_state); | |
bba90743 ES |
2586 | ext_debug("blocks %u/%lu requested for inode %u\n", |
2587 | iblock, max_blocks, inode->i_ino); | |
a86c6181 AT |
2588 | |
2589 | /* check in cache */ | |
7e028976 AM |
2590 | goal = ext4_ext_in_cache(inode, iblock, &newex); |
2591 | if (goal) { | |
a86c6181 AT |
2592 | if (goal == EXT4_EXT_CACHE_GAP) { |
2593 | if (!create) { | |
56055d3a AA |
2594 | /* |
2595 | * block isn't allocated yet and | |
2596 | * user doesn't want to allocate it | |
2597 | */ | |
a86c6181 AT |
2598 | goto out2; |
2599 | } | |
2600 | /* we should allocate requested block */ | |
2601 | } else if (goal == EXT4_EXT_CACHE_EXTENT) { | |
2602 | /* block is already allocated */ | |
8c55e204 DK |
2603 | newblock = iblock |
2604 | - le32_to_cpu(newex.ee_block) | |
2605 | + ext_pblock(&newex); | |
d0d856e8 | 2606 | /* number of remaining blocks in the extent */ |
b939e376 | 2607 | allocated = ext4_ext_get_actual_len(&newex) - |
a86c6181 AT |
2608 | (iblock - le32_to_cpu(newex.ee_block)); |
2609 | goto out; | |
2610 | } else { | |
2611 | BUG(); | |
2612 | } | |
2613 | } | |
2614 | ||
2615 | /* find extent for this block */ | |
2616 | path = ext4_ext_find_extent(inode, iblock, NULL); | |
2617 | if (IS_ERR(path)) { | |
2618 | err = PTR_ERR(path); | |
2619 | path = NULL; | |
2620 | goto out2; | |
2621 | } | |
2622 | ||
2623 | depth = ext_depth(inode); | |
2624 | ||
2625 | /* | |
d0d856e8 RD |
2626 | * consistent leaf must not be empty; |
2627 | * this situation is possible, though, _during_ tree modification; | |
a86c6181 AT |
2628 | * this is why assert can't be put in ext4_ext_find_extent() |
2629 | */ | |
2630 | BUG_ON(path[depth].p_ext == NULL && depth != 0); | |
56055d3a | 2631 | eh = path[depth].p_hdr; |
a86c6181 | 2632 | |
7e028976 AM |
2633 | ex = path[depth].p_ext; |
2634 | if (ex) { | |
725d26d3 | 2635 | ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block); |
f65e6fba | 2636 | ext4_fsblk_t ee_start = ext_pblock(ex); |
a2df2a63 | 2637 | unsigned short ee_len; |
471d4011 SB |
2638 | |
2639 | /* | |
471d4011 | 2640 | * Uninitialized extents are treated as holes, except that |
56055d3a | 2641 | * we split out initialized portions during a write. |
471d4011 | 2642 | */ |
a2df2a63 | 2643 | ee_len = ext4_ext_get_actual_len(ex); |
d0d856e8 | 2644 | /* if found extent covers block, simply return it */ |
8c55e204 | 2645 | if (iblock >= ee_block && iblock < ee_block + ee_len) { |
a86c6181 | 2646 | newblock = iblock - ee_block + ee_start; |
d0d856e8 | 2647 | /* number of remaining blocks in the extent */ |
a86c6181 | 2648 | allocated = ee_len - (iblock - ee_block); |
bba90743 | 2649 | ext_debug("%u fit into %lu:%d -> %llu\n", iblock, |
a86c6181 | 2650 | ee_block, ee_len, newblock); |
56055d3a | 2651 | |
a2df2a63 | 2652 | /* Do not put uninitialized extent in the cache */ |
56055d3a | 2653 | if (!ext4_ext_is_uninitialized(ex)) { |
a2df2a63 AA |
2654 | ext4_ext_put_in_cache(inode, ee_block, |
2655 | ee_len, ee_start, | |
2656 | EXT4_EXT_CACHE_EXTENT); | |
56055d3a AA |
2657 | goto out; |
2658 | } | |
2659 | if (create == EXT4_CREATE_UNINITIALIZED_EXT) | |
2660 | goto out; | |
e067ba00 AK |
2661 | if (!create) { |
2662 | /* | |
2663 | * We have blocks reserved already. We | |
2664 | * return allocated blocks so that delalloc | |
2665 | * won't do block reservation for us. But | |
2666 | * the buffer head will be unmapped so that | |
2667 | * a read from the block returns 0s. | |
2668 | */ | |
2669 | if (allocated > max_blocks) | |
2670 | allocated = max_blocks; | |
953e622b | 2671 | set_buffer_unwritten(bh_result); |
56055d3a | 2672 | goto out2; |
e067ba00 | 2673 | } |
56055d3a AA |
2674 | |
2675 | ret = ext4_ext_convert_to_initialized(handle, inode, | |
2676 | path, iblock, | |
2677 | max_blocks); | |
dbf9d7da DM |
2678 | if (ret <= 0) { |
2679 | err = ret; | |
56055d3a | 2680 | goto out2; |
dbf9d7da | 2681 | } else |
56055d3a AA |
2682 | allocated = ret; |
2683 | goto outnew; | |
a86c6181 AT |
2684 | } |
2685 | } | |
2686 | ||
2687 | /* | |
d0d856e8 | 2688 | * requested block isn't allocated yet; |
a86c6181 AT |
2689 | * we couldn't try to create block if create flag is zero |
2690 | */ | |
2691 | if (!create) { | |
56055d3a AA |
2692 | /* |
2693 | * put just found gap into cache to speed up | |
2694 | * subsequent requests | |
2695 | */ | |
a86c6181 AT |
2696 | ext4_ext_put_gap_in_cache(inode, path, iblock); |
2697 | goto out2; | |
2698 | } | |
2699 | /* | |
63f57933 AM |
2700 | * Okay, we need to do block allocation. Lazily initialize the block |
2701 | * allocation info here if necessary. | |
2702 | */ | |
a86c6181 AT |
2703 | if (S_ISREG(inode->i_mode) && (!EXT4_I(inode)->i_block_alloc_info)) |
2704 | ext4_init_block_alloc_info(inode); | |
2705 | ||
c9de560d AT |
2706 | /* find neighbour allocated blocks */ |
2707 | ar.lleft = iblock; | |
2708 | err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft); | |
2709 | if (err) | |
2710 | goto out2; | |
2711 | ar.lright = iblock; | |
2712 | err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright); | |
2713 | if (err) | |
2714 | goto out2; | |
25d14f98 | 2715 | |
749269fa AA |
2716 | /* |
2717 | * See if request is beyond maximum number of blocks we can have in | |
2718 | * a single extent. For an initialized extent this limit is | |
2719 | * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is | |
2720 | * EXT_UNINIT_MAX_LEN. | |
2721 | */ | |
2722 | if (max_blocks > EXT_INIT_MAX_LEN && | |
2723 | create != EXT4_CREATE_UNINITIALIZED_EXT) | |
2724 | max_blocks = EXT_INIT_MAX_LEN; | |
2725 | else if (max_blocks > EXT_UNINIT_MAX_LEN && | |
2726 | create == EXT4_CREATE_UNINITIALIZED_EXT) | |
2727 | max_blocks = EXT_UNINIT_MAX_LEN; | |
2728 | ||
25d14f98 AA |
2729 | /* Check if we can really insert (iblock)::(iblock+max_blocks) extent */ |
2730 | newex.ee_block = cpu_to_le32(iblock); | |
2731 | newex.ee_len = cpu_to_le16(max_blocks); | |
2732 | err = ext4_ext_check_overlap(inode, &newex, path); | |
2733 | if (err) | |
b939e376 | 2734 | allocated = ext4_ext_get_actual_len(&newex); |
25d14f98 AA |
2735 | else |
2736 | allocated = max_blocks; | |
c9de560d AT |
2737 | |
2738 | /* allocate new block */ | |
2739 | ar.inode = inode; | |
2740 | ar.goal = ext4_ext_find_goal(inode, path, iblock); | |
2741 | ar.logical = iblock; | |
2742 | ar.len = allocated; | |
2743 | if (S_ISREG(inode->i_mode)) | |
2744 | ar.flags = EXT4_MB_HINT_DATA; | |
2745 | else | |
2746 | /* disable in-core preallocation for non-regular files */ | |
2747 | ar.flags = 0; | |
2748 | newblock = ext4_mb_new_blocks(handle, &ar, &err); | |
a86c6181 AT |
2749 | if (!newblock) |
2750 | goto out2; | |
2ae02107 | 2751 | ext_debug("allocate new block: goal %llu, found %llu/%lu\n", |
a86c6181 AT |
2752 | goal, newblock, allocated); |
2753 | ||
2754 | /* try to insert new extent into found leaf and return */ | |
f65e6fba | 2755 | ext4_ext_store_pblock(&newex, newblock); |
c9de560d | 2756 | newex.ee_len = cpu_to_le16(ar.len); |
a2df2a63 AA |
2757 | if (create == EXT4_CREATE_UNINITIALIZED_EXT) /* Mark uninitialized */ |
2758 | ext4_ext_mark_uninitialized(&newex); | |
a86c6181 | 2759 | err = ext4_ext_insert_extent(handle, inode, path, &newex); |
315054f0 AT |
2760 | if (err) { |
2761 | /* free data blocks we just allocated */ | |
c9de560d AT |
2762 | /* not a good idea to call discard here directly, |
2763 | * but otherwise we'd need to call it every free() */ | |
2764 | ext4_mb_discard_inode_preallocations(inode); | |
315054f0 | 2765 | ext4_free_blocks(handle, inode, ext_pblock(&newex), |
b939e376 | 2766 | ext4_ext_get_actual_len(&newex), 0); |
a86c6181 | 2767 | goto out2; |
315054f0 | 2768 | } |
a86c6181 | 2769 | |
a86c6181 | 2770 | /* previous routine could use block we allocated */ |
f65e6fba | 2771 | newblock = ext_pblock(&newex); |
b939e376 | 2772 | allocated = ext4_ext_get_actual_len(&newex); |
56055d3a | 2773 | outnew: |
61628a3f MC |
2774 | if (extend_disksize) { |
2775 | disksize = ((loff_t) iblock + ar.len) << inode->i_blkbits; | |
2776 | if (disksize > i_size_read(inode)) | |
2777 | disksize = i_size_read(inode); | |
2778 | if (disksize > EXT4_I(inode)->i_disksize) | |
2779 | EXT4_I(inode)->i_disksize = disksize; | |
2780 | } | |
a379cd1d | 2781 | |
953e622b | 2782 | set_buffer_new(bh_result); |
a86c6181 | 2783 | |
a2df2a63 AA |
2784 | /* Cache only when it is _not_ an uninitialized extent */ |
2785 | if (create != EXT4_CREATE_UNINITIALIZED_EXT) | |
2786 | ext4_ext_put_in_cache(inode, iblock, allocated, newblock, | |
2787 | EXT4_EXT_CACHE_EXTENT); | |
a86c6181 AT |
2788 | out: |
2789 | if (allocated > max_blocks) | |
2790 | allocated = max_blocks; | |
2791 | ext4_ext_show_leaf(inode, path); | |
953e622b | 2792 | set_buffer_mapped(bh_result); |
a86c6181 AT |
2793 | bh_result->b_bdev = inode->i_sb->s_bdev; |
2794 | bh_result->b_blocknr = newblock; | |
2795 | out2: | |
2796 | if (path) { | |
2797 | ext4_ext_drop_refs(path); | |
2798 | kfree(path); | |
2799 | } | |
a86c6181 AT |
2800 | return err ? err : allocated; |
2801 | } | |
2802 | ||
cf108bca | 2803 | void ext4_ext_truncate(struct inode *inode) |
a86c6181 AT |
2804 | { |
2805 | struct address_space *mapping = inode->i_mapping; | |
2806 | struct super_block *sb = inode->i_sb; | |
725d26d3 | 2807 | ext4_lblk_t last_block; |
a86c6181 AT |
2808 | handle_t *handle; |
2809 | int err = 0; | |
2810 | ||
2811 | /* | |
2812 | * probably first extent we're gonna free will be last in block | |
2813 | */ | |
f3bd1f3f | 2814 | err = ext4_writepage_trans_blocks(inode); |
a86c6181 | 2815 | handle = ext4_journal_start(inode, err); |
cf108bca | 2816 | if (IS_ERR(handle)) |
a86c6181 | 2817 | return; |
a86c6181 | 2818 | |
cf108bca JK |
2819 | if (inode->i_size & (sb->s_blocksize - 1)) |
2820 | ext4_block_truncate_page(handle, mapping, inode->i_size); | |
a86c6181 | 2821 | |
9ddfc3dc JK |
2822 | if (ext4_orphan_add(handle, inode)) |
2823 | goto out_stop; | |
2824 | ||
0e855ac8 | 2825 | down_write(&EXT4_I(inode)->i_data_sem); |
a86c6181 AT |
2826 | ext4_ext_invalidate_cache(inode); |
2827 | ||
88aa3cff | 2828 | ext4_discard_reservation(inode); |
c9de560d | 2829 | |
a86c6181 | 2830 | /* |
d0d856e8 RD |
2831 | * TODO: optimization is possible here. |
2832 | * Probably we need not scan at all, | |
2833 | * because page truncation is enough. | |
a86c6181 | 2834 | */ |
a86c6181 AT |
2835 | |
2836 | /* we have to know where to truncate from in crash case */ | |
2837 | EXT4_I(inode)->i_disksize = inode->i_size; | |
2838 | ext4_mark_inode_dirty(handle, inode); | |
2839 | ||
2840 | last_block = (inode->i_size + sb->s_blocksize - 1) | |
2841 | >> EXT4_BLOCK_SIZE_BITS(sb); | |
2842 | err = ext4_ext_remove_space(inode, last_block); | |
2843 | ||
2844 | /* In a multi-transaction truncate, we only make the final | |
56055d3a AA |
2845 | * transaction synchronous. |
2846 | */ | |
a86c6181 AT |
2847 | if (IS_SYNC(inode)) |
2848 | handle->h_sync = 1; | |
2849 | ||
2850 | out_stop: | |
9ddfc3dc | 2851 | up_write(&EXT4_I(inode)->i_data_sem); |
a86c6181 | 2852 | /* |
d0d856e8 | 2853 | * If this was a simple ftruncate() and the file will remain alive, |
a86c6181 AT |
2854 | * then we need to clear up the orphan record which we created above. |
2855 | * However, if this was a real unlink then we were called by | |
2856 | * ext4_delete_inode(), and we allow that function to clean up the | |
2857 | * orphan info for us. | |
2858 | */ | |
2859 | if (inode->i_nlink) | |
2860 | ext4_orphan_del(handle, inode); | |
2861 | ||
ef737728 SR |
2862 | inode->i_mtime = inode->i_ctime = ext4_current_time(inode); |
2863 | ext4_mark_inode_dirty(handle, inode); | |
a86c6181 AT |
2864 | ext4_journal_stop(handle); |
2865 | } | |
2866 | ||
fd28784a AK |
2867 | static void ext4_falloc_update_inode(struct inode *inode, |
2868 | int mode, loff_t new_size, int update_ctime) | |
2869 | { | |
2870 | struct timespec now; | |
2871 | ||
2872 | if (update_ctime) { | |
2873 | now = current_fs_time(inode->i_sb); | |
2874 | if (!timespec_equal(&inode->i_ctime, &now)) | |
2875 | inode->i_ctime = now; | |
2876 | } | |
2877 | /* | |
2878 | * Update only when preallocation was requested beyond | |
2879 | * the file size. | |
2880 | */ | |
cf17fea6 AK |
2881 | if (!(mode & FALLOC_FL_KEEP_SIZE)) { |
2882 | if (new_size > i_size_read(inode)) | |
2883 | i_size_write(inode, new_size); | |
2884 | if (new_size > EXT4_I(inode)->i_disksize) | |
2885 | ext4_update_i_disksize(inode, new_size); | |
fd28784a AK |
2886 | } |
2887 | ||
2888 | } | |
2889 | ||
a2df2a63 AA |
2890 | /* |
2891 | * preallocate space for a file. This implements ext4's fallocate inode | |
2892 | * operation, which gets called from sys_fallocate system call. | |
2893 | * For block-mapped files, posix_fallocate should fall back to the method | |
2894 | * of writing zeroes to the required new blocks (the same behavior which is | |
2895 | * expected for file systems which do not support fallocate() system call). | |
2896 | */ | |
2897 | long ext4_fallocate(struct inode *inode, int mode, loff_t offset, loff_t len) | |
2898 | { | |
2899 | handle_t *handle; | |
725d26d3 | 2900 | ext4_lblk_t block; |
fd28784a | 2901 | loff_t new_size; |
725d26d3 | 2902 | unsigned long max_blocks; |
a2df2a63 AA |
2903 | int ret = 0; |
2904 | int ret2 = 0; | |
2905 | int retries = 0; | |
2906 | struct buffer_head map_bh; | |
2907 | unsigned int credits, blkbits = inode->i_blkbits; | |
2908 | ||
2909 | /* | |
2910 | * currently supporting (pre)allocate mode for extent-based | |
2911 | * files _only_ | |
2912 | */ | |
2913 | if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)) | |
2914 | return -EOPNOTSUPP; | |
2915 | ||
2916 | /* preallocation to directories is currently not supported */ | |
2917 | if (S_ISDIR(inode->i_mode)) | |
2918 | return -ENODEV; | |
2919 | ||
2920 | block = offset >> blkbits; | |
fd28784a AK |
2921 | /* |
2922 | * We can't just convert len to max_blocks because | |
2923 | * If blocksize = 4096 offset = 3072 and len = 2048 | |
2924 | */ | |
a2df2a63 | 2925 | max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) |
fd28784a | 2926 | - block; |
a2df2a63 | 2927 | /* |
f3bd1f3f | 2928 | * credits to insert 1 extent into extent tree |
a2df2a63 | 2929 | */ |
f3bd1f3f | 2930 | credits = ext4_chunk_trans_blocks(inode, max_blocks); |
55bd725a | 2931 | mutex_lock(&inode->i_mutex); |
a2df2a63 AA |
2932 | retry: |
2933 | while (ret >= 0 && ret < max_blocks) { | |
2934 | block = block + ret; | |
2935 | max_blocks = max_blocks - ret; | |
2936 | handle = ext4_journal_start(inode, credits); | |
2937 | if (IS_ERR(handle)) { | |
2938 | ret = PTR_ERR(handle); | |
2939 | break; | |
2940 | } | |
55bd725a | 2941 | ret = ext4_get_blocks_wrap(handle, inode, block, |
a2df2a63 | 2942 | max_blocks, &map_bh, |
d2a17637 | 2943 | EXT4_CREATE_UNINITIALIZED_EXT, 0, 0); |
221879c9 | 2944 | if (ret <= 0) { |
2c98615d AK |
2945 | #ifdef EXT4FS_DEBUG |
2946 | WARN_ON(ret <= 0); | |
2947 | printk(KERN_ERR "%s: ext4_ext_get_blocks " | |
2948 | "returned error inode#%lu, block=%u, " | |
2949 | "max_blocks=%lu", __func__, | |
221879c9 | 2950 | inode->i_ino, block, max_blocks); |
2c98615d | 2951 | #endif |
a2df2a63 AA |
2952 | ext4_mark_inode_dirty(handle, inode); |
2953 | ret2 = ext4_journal_stop(handle); | |
2954 | break; | |
2955 | } | |
fd28784a AK |
2956 | if ((block + ret) >= (EXT4_BLOCK_ALIGN(offset + len, |
2957 | blkbits) >> blkbits)) | |
2958 | new_size = offset + len; | |
2959 | else | |
2960 | new_size = (block + ret) << blkbits; | |
a2df2a63 | 2961 | |
fd28784a AK |
2962 | ext4_falloc_update_inode(inode, mode, new_size, |
2963 | buffer_new(&map_bh)); | |
a2df2a63 AA |
2964 | ext4_mark_inode_dirty(handle, inode); |
2965 | ret2 = ext4_journal_stop(handle); | |
2966 | if (ret2) | |
2967 | break; | |
2968 | } | |
fd28784a AK |
2969 | if (ret == -ENOSPC && |
2970 | ext4_should_retry_alloc(inode->i_sb, &retries)) { | |
2971 | ret = 0; | |
a2df2a63 | 2972 | goto retry; |
a2df2a63 | 2973 | } |
55bd725a | 2974 | mutex_unlock(&inode->i_mutex); |
a2df2a63 AA |
2975 | return ret > 0 ? ret2 : ret; |
2976 | } |