blob: 35f65cf4f318d72bce4e79995ddbc85d0e4bbfa3 [file] [log] [blame]
Alex Tomasa86c6182006-10-11 01:21:03 -07001/*
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
Alex Tomasa86c6182006-10-11 01:21:03 -070032#include <linux/fs.h>
33#include <linux/time.h>
Mingming Caocd02ff02007-10-16 18:38:25 -040034#include <linux/jbd2.h>
Alex Tomasa86c6182006-10-11 01:21:03 -070035#include <linux/highuid.h>
36#include <linux/pagemap.h>
37#include <linux/quotaops.h>
38#include <linux/string.h>
39#include <linux/slab.h>
Amit Aroraa2df2a62007-07-17 21:42:41 -040040#include <linux/falloc.h>
Alex Tomasa86c6182006-10-11 01:21:03 -070041#include <asm/uaccess.h>
Eric Sandeen6873fa02008-10-07 00:46:36 -040042#include <linux/fiemap.h>
Christoph Hellwig3dcf5452008-04-29 18:13:32 -040043#include "ext4_jbd2.h"
Theodore Ts'o4a092d72012-11-28 13:03:30 -050044#include "ext4_extents.h"
Tao Maf19d5872012-12-10 14:05:51 -050045#include "xattr.h"
Alex Tomasa86c6182006-10-11 01:21:03 -070046
Jiaying Zhang0562e0b2011-03-21 21:38:05 -040047#include <trace/events/ext4.h>
48
Lukas Czerner5f95d212012-03-19 23:03:19 -040049/*
50 * used by extent splitting.
51 */
52#define EXT4_EXT_MAY_ZEROOUT 0x1 /* safe to zeroout if split fails \
53 due to ENOSPC */
54#define EXT4_EXT_MARK_UNINIT1 0x2 /* mark first half uninitialized */
55#define EXT4_EXT_MARK_UNINIT2 0x4 /* mark second half uninitialized */
56
Dmitry Monakhovdee1f972012-10-10 01:04:58 -040057#define EXT4_EXT_DATA_VALID1 0x8 /* first half contains valid data */
58#define EXT4_EXT_DATA_VALID2 0x10 /* second half contains valid data */
59
Darrick J. Wong7ac59902012-04-29 18:37:10 -040060static __le32 ext4_extent_block_csum(struct inode *inode,
61 struct ext4_extent_header *eh)
62{
63 struct ext4_inode_info *ei = EXT4_I(inode);
64 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
65 __u32 csum;
66
67 csum = ext4_chksum(sbi, ei->i_csum_seed, (__u8 *)eh,
68 EXT4_EXTENT_TAIL_OFFSET(eh));
69 return cpu_to_le32(csum);
70}
71
72static int ext4_extent_block_csum_verify(struct inode *inode,
73 struct ext4_extent_header *eh)
74{
75 struct ext4_extent_tail *et;
76
77 if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
78 EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
79 return 1;
80
81 et = find_ext4_extent_tail(eh);
82 if (et->et_checksum != ext4_extent_block_csum(inode, eh))
83 return 0;
84 return 1;
85}
86
87static void ext4_extent_block_csum_set(struct inode *inode,
88 struct ext4_extent_header *eh)
89{
90 struct ext4_extent_tail *et;
91
92 if (!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
93 EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
94 return;
95
96 et = find_ext4_extent_tail(eh);
97 et->et_checksum = ext4_extent_block_csum(inode, eh);
98}
99
Allison Hendersond583fb82011-05-25 07:41:43 -0400100static int ext4_split_extent(handle_t *handle,
101 struct inode *inode,
102 struct ext4_ext_path *path,
103 struct ext4_map_blocks *map,
104 int split_flag,
105 int flags);
106
Lukas Czerner5f95d212012-03-19 23:03:19 -0400107static int ext4_split_extent_at(handle_t *handle,
108 struct inode *inode,
109 struct ext4_ext_path *path,
110 ext4_lblk_t split,
111 int split_flag,
112 int flags);
113
Lukas Czerner91dd8c12012-11-28 12:32:26 -0500114static int ext4_find_delayed_extent(struct inode *inode,
Zheng Liu69eb33d2013-02-18 00:31:07 -0500115 struct extent_status *newes);
Lukas Czerner91dd8c12012-11-28 12:32:26 -0500116
Jan Kara487caee2009-08-17 22:17:20 -0400117static int ext4_ext_truncate_extend_restart(handle_t *handle,
118 struct inode *inode,
119 int needed)
Alex Tomasa86c6182006-10-11 01:21:03 -0700120{
121 int err;
122
Frank Mayhar03901312009-01-07 00:06:22 -0500123 if (!ext4_handle_valid(handle))
124 return 0;
Alex Tomasa86c6182006-10-11 01:21:03 -0700125 if (handle->h_buffer_credits > needed)
Shen Feng9102e4f2008-07-11 19:27:31 -0400126 return 0;
127 err = ext4_journal_extend(handle, needed);
Theodore Ts'o0123c932008-08-01 20:57:54 -0400128 if (err <= 0)
Shen Feng9102e4f2008-07-11 19:27:31 -0400129 return err;
Jan Kara487caee2009-08-17 22:17:20 -0400130 err = ext4_truncate_restart_trans(handle, inode, needed);
Dmitry Monakhov0617b832010-05-17 01:00:00 -0400131 if (err == 0)
132 err = -EAGAIN;
Jan Kara487caee2009-08-17 22:17:20 -0400133
134 return err;
Alex Tomasa86c6182006-10-11 01:21:03 -0700135}
136
137/*
138 * could return:
139 * - EROFS
140 * - ENOMEM
141 */
142static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
143 struct ext4_ext_path *path)
144{
145 if (path->p_bh) {
146 /* path points to block */
147 return ext4_journal_get_write_access(handle, path->p_bh);
148 }
149 /* path points to leaf/index in inode body */
150 /* we use in-core data, no need to protect them */
151 return 0;
152}
153
154/*
155 * could return:
156 * - EROFS
157 * - ENOMEM
158 * - EIO
159 */
Darrick J. Wong26564972013-04-19 14:04:12 -0400160int __ext4_ext_dirty(const char *where, unsigned int line, handle_t *handle,
161 struct inode *inode, struct ext4_ext_path *path)
Alex Tomasa86c6182006-10-11 01:21:03 -0700162{
163 int err;
164 if (path->p_bh) {
Darrick J. Wong7ac59902012-04-29 18:37:10 -0400165 ext4_extent_block_csum_set(inode, ext_block_hdr(path->p_bh));
Alex Tomasa86c6182006-10-11 01:21:03 -0700166 /* path points to block */
Theodore Ts'o9ea7a0d2011-09-04 10:18:14 -0400167 err = __ext4_handle_dirty_metadata(where, line, handle,
168 inode, path->p_bh);
Alex Tomasa86c6182006-10-11 01:21:03 -0700169 } else {
170 /* path points to leaf/index in inode body */
171 err = ext4_mark_inode_dirty(handle, inode);
172 }
173 return err;
174}
175
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700176static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
Alex Tomasa86c6182006-10-11 01:21:03 -0700177 struct ext4_ext_path *path,
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -0500178 ext4_lblk_t block)
Alex Tomasa86c6182006-10-11 01:21:03 -0700179{
Alex Tomasa86c6182006-10-11 01:21:03 -0700180 if (path) {
Yongqiang Yang81fdbb42011-10-29 09:23:38 -0400181 int depth = path->p_depth;
Alex Tomasa86c6182006-10-11 01:21:03 -0700182 struct ext4_extent *ex;
Alex Tomasa86c6182006-10-11 01:21:03 -0700183
Kazuya Mioad4fb9c2011-01-10 12:12:28 -0500184 /*
185 * Try to predict block placement assuming that we are
186 * filling in a file which will eventually be
187 * non-sparse --- i.e., in the case of libbfd writing
188 * an ELF object sections out-of-order but in a way
189 * the eventually results in a contiguous object or
190 * executable file, or some database extending a table
191 * space file. However, this is actually somewhat
192 * non-ideal if we are writing a sparse file such as
193 * qemu or KVM writing a raw image file that is going
194 * to stay fairly sparse, since it will end up
195 * fragmenting the file system's free space. Maybe we
196 * should have some hueristics or some way to allow
197 * userspace to pass a hint to file system,
Tao Mab8d65682011-01-21 23:21:31 +0800198 * especially if the latter case turns out to be
Kazuya Mioad4fb9c2011-01-10 12:12:28 -0500199 * common.
200 */
Avantika Mathur7e028972006-12-06 20:41:33 -0800201 ex = path[depth].p_ext;
Kazuya Mioad4fb9c2011-01-10 12:12:28 -0500202 if (ex) {
203 ext4_fsblk_t ext_pblk = ext4_ext_pblock(ex);
204 ext4_lblk_t ext_block = le32_to_cpu(ex->ee_block);
205
206 if (block > ext_block)
207 return ext_pblk + (block - ext_block);
208 else
209 return ext_pblk - (ext_block - block);
210 }
Alex Tomasa86c6182006-10-11 01:21:03 -0700211
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700212 /* it looks like index is empty;
213 * try to find starting block from index itself */
Alex Tomasa86c6182006-10-11 01:21:03 -0700214 if (path[depth].p_bh)
215 return path[depth].p_bh->b_blocknr;
216 }
217
218 /* OK. use inode's group */
Eric Sandeenf86186b2011-06-28 10:01:31 -0400219 return ext4_inode_to_goal_block(inode);
Alex Tomasa86c6182006-10-11 01:21:03 -0700220}
221
Aneesh Kumar K.V654b4902008-07-11 19:27:31 -0400222/*
223 * Allocation for a meta data block
224 */
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700225static ext4_fsblk_t
Aneesh Kumar K.V654b4902008-07-11 19:27:31 -0400226ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
Alex Tomasa86c6182006-10-11 01:21:03 -0700227 struct ext4_ext_path *path,
Allison Henderson55f020d2011-05-25 07:41:26 -0400228 struct ext4_extent *ex, int *err, unsigned int flags)
Alex Tomasa86c6182006-10-11 01:21:03 -0700229{
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700230 ext4_fsblk_t goal, newblock;
Alex Tomasa86c6182006-10-11 01:21:03 -0700231
232 goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
Allison Henderson55f020d2011-05-25 07:41:26 -0400233 newblock = ext4_new_meta_blocks(handle, inode, goal, flags,
234 NULL, err);
Alex Tomasa86c6182006-10-11 01:21:03 -0700235 return newblock;
236}
237
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400238static inline int ext4_ext_space_block(struct inode *inode, int check)
Alex Tomasa86c6182006-10-11 01:21:03 -0700239{
240 int size;
241
242 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
243 / sizeof(struct ext4_extent);
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +0100244#ifdef AGGRESSIVE_TEST
Yongqiang Yang02dc62fb2011-10-29 09:29:11 -0400245 if (!check && size > 6)
246 size = 6;
Alex Tomasa86c6182006-10-11 01:21:03 -0700247#endif
248 return size;
249}
250
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400251static inline int ext4_ext_space_block_idx(struct inode *inode, int check)
Alex Tomasa86c6182006-10-11 01:21:03 -0700252{
253 int size;
254
255 size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
256 / sizeof(struct ext4_extent_idx);
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +0100257#ifdef AGGRESSIVE_TEST
Yongqiang Yang02dc62fb2011-10-29 09:29:11 -0400258 if (!check && size > 5)
259 size = 5;
Alex Tomasa86c6182006-10-11 01:21:03 -0700260#endif
261 return size;
262}
263
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400264static inline int ext4_ext_space_root(struct inode *inode, int check)
Alex Tomasa86c6182006-10-11 01:21:03 -0700265{
266 int size;
267
268 size = sizeof(EXT4_I(inode)->i_data);
269 size -= sizeof(struct ext4_extent_header);
270 size /= sizeof(struct ext4_extent);
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +0100271#ifdef AGGRESSIVE_TEST
Yongqiang Yang02dc62fb2011-10-29 09:29:11 -0400272 if (!check && size > 3)
273 size = 3;
Alex Tomasa86c6182006-10-11 01:21:03 -0700274#endif
275 return size;
276}
277
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400278static inline int ext4_ext_space_root_idx(struct inode *inode, int check)
Alex Tomasa86c6182006-10-11 01:21:03 -0700279{
280 int size;
281
282 size = sizeof(EXT4_I(inode)->i_data);
283 size -= sizeof(struct ext4_extent_header);
284 size /= sizeof(struct ext4_extent_idx);
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +0100285#ifdef AGGRESSIVE_TEST
Yongqiang Yang02dc62fb2011-10-29 09:29:11 -0400286 if (!check && size > 4)
287 size = 4;
Alex Tomasa86c6182006-10-11 01:21:03 -0700288#endif
289 return size;
290}
291
Mingming Caod2a17632008-07-14 17:52:37 -0400292/*
293 * Calculate the number of metadata blocks needed
294 * to allocate @blocks
295 * Worse case is one block per extent
296 */
Theodore Ts'o01f49d02011-01-10 12:13:03 -0500297int ext4_ext_calc_metadata_amount(struct inode *inode, ext4_lblk_t lblock)
Mingming Caod2a17632008-07-14 17:52:37 -0400298{
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500299 struct ext4_inode_info *ei = EXT4_I(inode);
Yongqiang Yang81fdbb42011-10-29 09:23:38 -0400300 int idxs;
Mingming Caod2a17632008-07-14 17:52:37 -0400301
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500302 idxs = ((inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
303 / sizeof(struct ext4_extent_idx));
Mingming Caod2a17632008-07-14 17:52:37 -0400304
305 /*
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500306 * If the new delayed allocation block is contiguous with the
307 * previous da block, it can share index blocks with the
308 * previous block, so we only need to allocate a new index
309 * block every idxs leaf blocks. At ldxs**2 blocks, we need
310 * an additional index block, and at ldxs**3 blocks, yet
311 * another index blocks.
Mingming Caod2a17632008-07-14 17:52:37 -0400312 */
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500313 if (ei->i_da_metadata_calc_len &&
314 ei->i_da_metadata_calc_last_lblock+1 == lblock) {
Yongqiang Yang81fdbb42011-10-29 09:23:38 -0400315 int num = 0;
316
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500317 if ((ei->i_da_metadata_calc_len % idxs) == 0)
318 num++;
319 if ((ei->i_da_metadata_calc_len % (idxs*idxs)) == 0)
320 num++;
321 if ((ei->i_da_metadata_calc_len % (idxs*idxs*idxs)) == 0) {
322 num++;
323 ei->i_da_metadata_calc_len = 0;
324 } else
325 ei->i_da_metadata_calc_len++;
326 ei->i_da_metadata_calc_last_lblock++;
327 return num;
328 }
Mingming Caod2a17632008-07-14 17:52:37 -0400329
Theodore Ts'o9d0be502010-01-01 02:41:30 -0500330 /*
331 * In the worst case we need a new set of index blocks at
332 * every level of the inode's extent tree.
333 */
334 ei->i_da_metadata_calc_len = 1;
335 ei->i_da_metadata_calc_last_lblock = lblock;
336 return ext_depth(inode) + 1;
Mingming Caod2a17632008-07-14 17:52:37 -0400337}
338
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400339static int
340ext4_ext_max_entries(struct inode *inode, int depth)
341{
342 int max;
343
344 if (depth == ext_depth(inode)) {
345 if (depth == 0)
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400346 max = ext4_ext_space_root(inode, 1);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400347 else
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400348 max = ext4_ext_space_root_idx(inode, 1);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400349 } else {
350 if (depth == 0)
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400351 max = ext4_ext_space_block(inode, 1);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400352 else
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400353 max = ext4_ext_space_block_idx(inode, 1);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400354 }
355
356 return max;
357}
358
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400359static int ext4_valid_extent(struct inode *inode, struct ext4_extent *ext)
360{
Theodore Ts'obf89d162010-10-27 21:30:14 -0400361 ext4_fsblk_t block = ext4_ext_pblock(ext);
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400362 int len = ext4_ext_get_actual_len(ext);
Theodore Ts'oe84a26c2009-04-22 20:52:25 -0400363
Theodore Ts'o31d4f3a2012-03-11 23:30:16 -0400364 if (len == 0)
365 return 0;
Theodore Ts'o6fd058f2009-05-17 15:38:01 -0400366 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, len);
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400367}
368
369static int ext4_valid_extent_idx(struct inode *inode,
370 struct ext4_extent_idx *ext_idx)
371{
Theodore Ts'obf89d162010-10-27 21:30:14 -0400372 ext4_fsblk_t block = ext4_idx_pblock(ext_idx);
Theodore Ts'oe84a26c2009-04-22 20:52:25 -0400373
Theodore Ts'o6fd058f2009-05-17 15:38:01 -0400374 return ext4_data_block_valid(EXT4_SB(inode->i_sb), block, 1);
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400375}
376
377static int ext4_valid_extent_entries(struct inode *inode,
378 struct ext4_extent_header *eh,
379 int depth)
380{
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400381 unsigned short entries;
382 if (eh->eh_entries == 0)
383 return 1;
384
385 entries = le16_to_cpu(eh->eh_entries);
386
387 if (depth == 0) {
388 /* leaf entries */
Yongqiang Yang81fdbb42011-10-29 09:23:38 -0400389 struct ext4_extent *ext = EXT_FIRST_EXTENT(eh);
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400390 while (entries) {
391 if (!ext4_valid_extent(inode, ext))
392 return 0;
393 ext++;
394 entries--;
395 }
396 } else {
Yongqiang Yang81fdbb42011-10-29 09:23:38 -0400397 struct ext4_extent_idx *ext_idx = EXT_FIRST_INDEX(eh);
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400398 while (entries) {
399 if (!ext4_valid_extent_idx(inode, ext_idx))
400 return 0;
401 ext_idx++;
402 entries--;
403 }
404 }
405 return 1;
406}
407
Theodore Ts'oc398eda2010-07-27 11:56:40 -0400408static int __ext4_ext_check(const char *function, unsigned int line,
409 struct inode *inode, struct ext4_extent_header *eh,
Theodore Ts'oc3491792013-08-16 21:21:41 -0400410 int depth, ext4_fsblk_t pblk)
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400411{
412 const char *error_msg;
413 int max = 0;
414
415 if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
416 error_msg = "invalid magic";
417 goto corrupted;
418 }
419 if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
420 error_msg = "unexpected eh_depth";
421 goto corrupted;
422 }
423 if (unlikely(eh->eh_max == 0)) {
424 error_msg = "invalid eh_max";
425 goto corrupted;
426 }
427 max = ext4_ext_max_entries(inode, depth);
428 if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
429 error_msg = "too large eh_max";
430 goto corrupted;
431 }
432 if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
433 error_msg = "invalid eh_entries";
434 goto corrupted;
435 }
Aneesh Kumar K.V56b19862009-03-12 09:51:20 -0400436 if (!ext4_valid_extent_entries(inode, eh, depth)) {
437 error_msg = "invalid extent entries";
438 goto corrupted;
439 }
Darrick J. Wong7ac59902012-04-29 18:37:10 -0400440 /* Verify checksum on non-root extent tree nodes */
441 if (ext_depth(inode) != depth &&
442 !ext4_extent_block_csum_verify(inode, eh)) {
443 error_msg = "extent tree corrupted";
444 goto corrupted;
445 }
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400446 return 0;
447
448corrupted:
Theodore Ts'oc398eda2010-07-27 11:56:40 -0400449 ext4_error_inode(inode, function, line, 0,
Theodore Ts'oc3491792013-08-16 21:21:41 -0400450 "pblk %llu bad header/extent: %s - magic %x, "
451 "entries %u, max %u(%u), depth %u(%u)",
452 (unsigned long long) pblk, error_msg,
453 le16_to_cpu(eh->eh_magic),
454 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
455 max, le16_to_cpu(eh->eh_depth), depth);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400456 return -EIO;
457}
458
Theodore Ts'oc3491792013-08-16 21:21:41 -0400459#define ext4_ext_check(inode, eh, depth, pblk) \
460 __ext4_ext_check(__func__, __LINE__, (inode), (eh), (depth), (pblk))
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400461
Aneesh Kumar K.V7a262f72009-03-27 16:39:58 -0400462int ext4_ext_check_inode(struct inode *inode)
463{
Theodore Ts'oc3491792013-08-16 21:21:41 -0400464 return ext4_ext_check(inode, ext_inode_hdr(inode), ext_depth(inode), 0);
Aneesh Kumar K.V7a262f72009-03-27 16:39:58 -0400465}
466
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400467static struct buffer_head *
468__read_extent_tree_block(const char *function, unsigned int line,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -0400469 struct inode *inode, ext4_fsblk_t pblk, int depth,
470 int flags)
Darrick J. Wongf8489122012-04-29 18:21:10 -0400471{
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400472 struct buffer_head *bh;
473 int err;
Darrick J. Wongf8489122012-04-29 18:21:10 -0400474
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400475 bh = sb_getblk(inode->i_sb, pblk);
476 if (unlikely(!bh))
477 return ERR_PTR(-ENOMEM);
478
479 if (!bh_uptodate_or_lock(bh)) {
480 trace_ext4_ext_load_extent(inode, pblk, _RET_IP_);
481 err = bh_submit_read(bh);
482 if (err < 0)
483 goto errout;
484 }
Theodore Ts'o7869a4a2013-08-16 22:05:14 -0400485 if (buffer_verified(bh) && !(flags & EXT4_EX_FORCE_CACHE))
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400486 return bh;
487 err = __ext4_ext_check(function, line, inode,
Theodore Ts'oc3491792013-08-16 21:21:41 -0400488 ext_block_hdr(bh), depth, pblk);
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400489 if (err)
490 goto errout;
Darrick J. Wongf8489122012-04-29 18:21:10 -0400491 set_buffer_verified(bh);
Theodore Ts'o107a7bd2013-08-16 21:23:41 -0400492 /*
493 * If this is a leaf block, cache all of its entries
494 */
495 if (!(flags & EXT4_EX_NOCACHE) && depth == 0) {
496 struct ext4_extent_header *eh = ext_block_hdr(bh);
497 struct ext4_extent *ex = EXT_FIRST_EXTENT(eh);
498 ext4_lblk_t prev = 0;
499 int i;
500
501 for (i = le16_to_cpu(eh->eh_entries); i > 0; i--, ex++) {
502 unsigned int status = EXTENT_STATUS_WRITTEN;
503 ext4_lblk_t lblk = le32_to_cpu(ex->ee_block);
504 int len = ext4_ext_get_actual_len(ex);
505
506 if (prev && (prev != lblk))
507 ext4_es_cache_extent(inode, prev,
508 lblk - prev, ~0,
509 EXTENT_STATUS_HOLE);
510
511 if (ext4_ext_is_uninitialized(ex))
512 status = EXTENT_STATUS_UNWRITTEN;
513 ext4_es_cache_extent(inode, lblk, len,
514 ext4_ext_pblock(ex), status);
515 prev = lblk + len;
516 }
517 }
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400518 return bh;
519errout:
520 put_bh(bh);
521 return ERR_PTR(err);
522
Darrick J. Wongf8489122012-04-29 18:21:10 -0400523}
524
Theodore Ts'o107a7bd2013-08-16 21:23:41 -0400525#define read_extent_tree_block(inode, pblk, depth, flags) \
526 __read_extent_tree_block(__func__, __LINE__, (inode), (pblk), \
527 (depth), (flags))
Darrick J. Wongf8489122012-04-29 18:21:10 -0400528
Theodore Ts'o7869a4a2013-08-16 22:05:14 -0400529/*
530 * This function is called to cache a file's extent information in the
531 * extent status tree
532 */
533int ext4_ext_precache(struct inode *inode)
534{
535 struct ext4_inode_info *ei = EXT4_I(inode);
536 struct ext4_ext_path *path = NULL;
537 struct buffer_head *bh;
538 int i = 0, depth, ret = 0;
539
540 if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
541 return 0; /* not an extent-mapped inode */
542
543 down_read(&ei->i_data_sem);
544 depth = ext_depth(inode);
545
546 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1),
547 GFP_NOFS);
548 if (path == NULL) {
549 up_read(&ei->i_data_sem);
550 return -ENOMEM;
551 }
552
553 /* Don't cache anything if there are no external extent blocks */
554 if (depth == 0)
555 goto out;
556 path[0].p_hdr = ext_inode_hdr(inode);
557 ret = ext4_ext_check(inode, path[0].p_hdr, depth, 0);
558 if (ret)
559 goto out;
560 path[0].p_idx = EXT_FIRST_INDEX(path[0].p_hdr);
561 while (i >= 0) {
562 /*
563 * If this is a leaf block or we've reached the end of
564 * the index block, go up
565 */
566 if ((i == depth) ||
567 path[i].p_idx > EXT_LAST_INDEX(path[i].p_hdr)) {
568 brelse(path[i].p_bh);
569 path[i].p_bh = NULL;
570 i--;
571 continue;
572 }
573 bh = read_extent_tree_block(inode,
574 ext4_idx_pblock(path[i].p_idx++),
575 depth - i - 1,
576 EXT4_EX_FORCE_CACHE);
577 if (IS_ERR(bh)) {
578 ret = PTR_ERR(bh);
579 break;
580 }
581 i++;
582 path[i].p_bh = bh;
583 path[i].p_hdr = ext_block_hdr(bh);
584 path[i].p_idx = EXT_FIRST_INDEX(path[i].p_hdr);
585 }
586 ext4_set_inode_state(inode, EXT4_STATE_EXT_PRECACHED);
587out:
588 up_read(&ei->i_data_sem);
589 ext4_ext_drop_refs(path);
590 kfree(path);
591 return ret;
592}
593
Alex Tomasa86c6182006-10-11 01:21:03 -0700594#ifdef EXT_DEBUG
595static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
596{
597 int k, l = path->p_depth;
598
599 ext_debug("path:");
600 for (k = 0; k <= l; k++, path++) {
601 if (path->p_idx) {
Mingming Cao2ae02102006-10-11 01:21:11 -0700602 ext_debug(" %d->%llu", le32_to_cpu(path->p_idx->ei_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -0400603 ext4_idx_pblock(path->p_idx));
Alex Tomasa86c6182006-10-11 01:21:03 -0700604 } else if (path->p_ext) {
Mingming553f9002009-09-18 13:34:55 -0400605 ext_debug(" %d:[%d]%d:%llu ",
Alex Tomasa86c6182006-10-11 01:21:03 -0700606 le32_to_cpu(path->p_ext->ee_block),
Mingming553f9002009-09-18 13:34:55 -0400607 ext4_ext_is_uninitialized(path->p_ext),
Amit Aroraa2df2a62007-07-17 21:42:41 -0400608 ext4_ext_get_actual_len(path->p_ext),
Theodore Ts'obf89d162010-10-27 21:30:14 -0400609 ext4_ext_pblock(path->p_ext));
Alex Tomasa86c6182006-10-11 01:21:03 -0700610 } else
611 ext_debug(" []");
612 }
613 ext_debug("\n");
614}
615
616static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
617{
618 int depth = ext_depth(inode);
619 struct ext4_extent_header *eh;
620 struct ext4_extent *ex;
621 int i;
622
623 if (!path)
624 return;
625
626 eh = path[depth].p_hdr;
627 ex = EXT_FIRST_EXTENT(eh);
628
Mingming553f9002009-09-18 13:34:55 -0400629 ext_debug("Displaying leaf extents for inode %lu\n", inode->i_ino);
630
Alex Tomasa86c6182006-10-11 01:21:03 -0700631 for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
Mingming553f9002009-09-18 13:34:55 -0400632 ext_debug("%d:[%d]%d:%llu ", le32_to_cpu(ex->ee_block),
633 ext4_ext_is_uninitialized(ex),
Theodore Ts'obf89d162010-10-27 21:30:14 -0400634 ext4_ext_get_actual_len(ex), ext4_ext_pblock(ex));
Alex Tomasa86c6182006-10-11 01:21:03 -0700635 }
636 ext_debug("\n");
637}
Yongqiang Yang1b16da72011-05-25 17:41:48 -0400638
639static void ext4_ext_show_move(struct inode *inode, struct ext4_ext_path *path,
640 ext4_fsblk_t newblock, int level)
641{
642 int depth = ext_depth(inode);
643 struct ext4_extent *ex;
644
645 if (depth != level) {
646 struct ext4_extent_idx *idx;
647 idx = path[level].p_idx;
648 while (idx <= EXT_MAX_INDEX(path[level].p_hdr)) {
649 ext_debug("%d: move %d:%llu in new index %llu\n", level,
650 le32_to_cpu(idx->ei_block),
651 ext4_idx_pblock(idx),
652 newblock);
653 idx++;
654 }
655
656 return;
657 }
658
659 ex = path[depth].p_ext;
660 while (ex <= EXT_MAX_EXTENT(path[depth].p_hdr)) {
661 ext_debug("move %d:%llu:[%d]%d in new leaf %llu\n",
662 le32_to_cpu(ex->ee_block),
663 ext4_ext_pblock(ex),
664 ext4_ext_is_uninitialized(ex),
665 ext4_ext_get_actual_len(ex),
666 newblock);
667 ex++;
668 }
669}
670
Alex Tomasa86c6182006-10-11 01:21:03 -0700671#else
Theodore Ts'oaf5bc922008-09-08 22:25:24 -0400672#define ext4_ext_show_path(inode, path)
673#define ext4_ext_show_leaf(inode, path)
Yongqiang Yang1b16da72011-05-25 17:41:48 -0400674#define ext4_ext_show_move(inode, path, newblock, level)
Alex Tomasa86c6182006-10-11 01:21:03 -0700675#endif
676
Aneesh Kumar K.Vb35905c2008-02-25 16:54:37 -0500677void ext4_ext_drop_refs(struct ext4_ext_path *path)
Alex Tomasa86c6182006-10-11 01:21:03 -0700678{
679 int depth = path->p_depth;
680 int i;
681
682 for (i = 0; i <= depth; i++, path++)
683 if (path->p_bh) {
684 brelse(path->p_bh);
685 path->p_bh = NULL;
686 }
687}
688
689/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700690 * ext4_ext_binsearch_idx:
691 * binary search for the closest index of the given block
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400692 * the header must be checked before calling this
Alex Tomasa86c6182006-10-11 01:21:03 -0700693 */
694static void
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -0500695ext4_ext_binsearch_idx(struct inode *inode,
696 struct ext4_ext_path *path, ext4_lblk_t block)
Alex Tomasa86c6182006-10-11 01:21:03 -0700697{
698 struct ext4_extent_header *eh = path->p_hdr;
699 struct ext4_extent_idx *r, *l, *m;
700
Alex Tomasa86c6182006-10-11 01:21:03 -0700701
Eric Sandeenbba90742008-01-28 23:58:27 -0500702 ext_debug("binsearch for %u(idx): ", block);
Alex Tomasa86c6182006-10-11 01:21:03 -0700703
704 l = EXT_FIRST_INDEX(eh) + 1;
Dmitry Monakhove9f410b2007-07-18 09:09:15 -0400705 r = EXT_LAST_INDEX(eh);
Alex Tomasa86c6182006-10-11 01:21:03 -0700706 while (l <= r) {
707 m = l + (r - l) / 2;
708 if (block < le32_to_cpu(m->ei_block))
709 r = m - 1;
710 else
711 l = m + 1;
Dmitry Monakhov26d535e2007-07-18 08:33:37 -0400712 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
713 m, le32_to_cpu(m->ei_block),
714 r, le32_to_cpu(r->ei_block));
Alex Tomasa86c6182006-10-11 01:21:03 -0700715 }
716
717 path->p_idx = l - 1;
Zheng Liu4a3c3a52012-05-28 17:55:16 -0400718 ext_debug(" -> %u->%lld ", le32_to_cpu(path->p_idx->ei_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -0400719 ext4_idx_pblock(path->p_idx));
Alex Tomasa86c6182006-10-11 01:21:03 -0700720
721#ifdef CHECK_BINSEARCH
722 {
723 struct ext4_extent_idx *chix, *ix;
724 int k;
725
726 chix = ix = EXT_FIRST_INDEX(eh);
727 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
728 if (k != 0 &&
729 le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
Theodore Ts'o4776004f2008-09-08 23:00:52 -0400730 printk(KERN_DEBUG "k=%d, ix=0x%p, "
731 "first=0x%p\n", k,
732 ix, EXT_FIRST_INDEX(eh));
733 printk(KERN_DEBUG "%u <= %u\n",
Alex Tomasa86c6182006-10-11 01:21:03 -0700734 le32_to_cpu(ix->ei_block),
735 le32_to_cpu(ix[-1].ei_block));
736 }
737 BUG_ON(k && le32_to_cpu(ix->ei_block)
Dave Kleikamp8c55e202007-05-24 13:04:54 -0400738 <= le32_to_cpu(ix[-1].ei_block));
Alex Tomasa86c6182006-10-11 01:21:03 -0700739 if (block < le32_to_cpu(ix->ei_block))
740 break;
741 chix = ix;
742 }
743 BUG_ON(chix != path->p_idx);
744 }
745#endif
746
747}
748
749/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700750 * ext4_ext_binsearch:
751 * binary search for closest extent of the given block
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400752 * the header must be checked before calling this
Alex Tomasa86c6182006-10-11 01:21:03 -0700753 */
754static void
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -0500755ext4_ext_binsearch(struct inode *inode,
756 struct ext4_ext_path *path, ext4_lblk_t block)
Alex Tomasa86c6182006-10-11 01:21:03 -0700757{
758 struct ext4_extent_header *eh = path->p_hdr;
759 struct ext4_extent *r, *l, *m;
760
Alex Tomasa86c6182006-10-11 01:21:03 -0700761 if (eh->eh_entries == 0) {
762 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700763 * this leaf is empty:
764 * we get such a leaf in split/add case
Alex Tomasa86c6182006-10-11 01:21:03 -0700765 */
766 return;
767 }
768
Eric Sandeenbba90742008-01-28 23:58:27 -0500769 ext_debug("binsearch for %u: ", block);
Alex Tomasa86c6182006-10-11 01:21:03 -0700770
771 l = EXT_FIRST_EXTENT(eh) + 1;
Dmitry Monakhove9f410b2007-07-18 09:09:15 -0400772 r = EXT_LAST_EXTENT(eh);
Alex Tomasa86c6182006-10-11 01:21:03 -0700773
774 while (l <= r) {
775 m = l + (r - l) / 2;
776 if (block < le32_to_cpu(m->ee_block))
777 r = m - 1;
778 else
779 l = m + 1;
Dmitry Monakhov26d535e2007-07-18 08:33:37 -0400780 ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
781 m, le32_to_cpu(m->ee_block),
782 r, le32_to_cpu(r->ee_block));
Alex Tomasa86c6182006-10-11 01:21:03 -0700783 }
784
785 path->p_ext = l - 1;
Mingming553f9002009-09-18 13:34:55 -0400786 ext_debug(" -> %d:%llu:[%d]%d ",
Dave Kleikamp8c55e202007-05-24 13:04:54 -0400787 le32_to_cpu(path->p_ext->ee_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -0400788 ext4_ext_pblock(path->p_ext),
Mingming553f9002009-09-18 13:34:55 -0400789 ext4_ext_is_uninitialized(path->p_ext),
Amit Aroraa2df2a62007-07-17 21:42:41 -0400790 ext4_ext_get_actual_len(path->p_ext));
Alex Tomasa86c6182006-10-11 01:21:03 -0700791
792#ifdef CHECK_BINSEARCH
793 {
794 struct ext4_extent *chex, *ex;
795 int k;
796
797 chex = ex = EXT_FIRST_EXTENT(eh);
798 for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
799 BUG_ON(k && le32_to_cpu(ex->ee_block)
Dave Kleikamp8c55e202007-05-24 13:04:54 -0400800 <= le32_to_cpu(ex[-1].ee_block));
Alex Tomasa86c6182006-10-11 01:21:03 -0700801 if (block < le32_to_cpu(ex->ee_block))
802 break;
803 chex = ex;
804 }
805 BUG_ON(chex != path->p_ext);
806 }
807#endif
808
809}
810
811int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
812{
813 struct ext4_extent_header *eh;
814
815 eh = ext_inode_hdr(inode);
816 eh->eh_depth = 0;
817 eh->eh_entries = 0;
818 eh->eh_magic = EXT4_EXT_MAGIC;
Theodore Ts'o55ad63b2009-08-28 10:40:33 -0400819 eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -0700820 ext4_mark_inode_dirty(handle, inode);
Alex Tomasa86c6182006-10-11 01:21:03 -0700821 return 0;
822}
823
824struct ext4_ext_path *
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -0500825ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -0400826 struct ext4_ext_path *path, int flags)
Alex Tomasa86c6182006-10-11 01:21:03 -0700827{
828 struct ext4_extent_header *eh;
829 struct buffer_head *bh;
830 short int depth, i, ppos = 0, alloc = 0;
Theodore Ts'o860d21e2013-01-12 16:19:36 -0500831 int ret;
Alex Tomasa86c6182006-10-11 01:21:03 -0700832
833 eh = ext_inode_hdr(inode);
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400834 depth = ext_depth(inode);
Alex Tomasa86c6182006-10-11 01:21:03 -0700835
836 /* account possible depth increase */
837 if (!path) {
Avantika Mathur5d4958f2006-12-06 20:41:35 -0800838 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
Alex Tomasa86c6182006-10-11 01:21:03 -0700839 GFP_NOFS);
840 if (!path)
841 return ERR_PTR(-ENOMEM);
842 alloc = 1;
843 }
Alex Tomasa86c6182006-10-11 01:21:03 -0700844 path[0].p_hdr = eh;
Shen Feng1973adc2008-07-11 19:27:31 -0400845 path[0].p_bh = NULL;
Alex Tomasa86c6182006-10-11 01:21:03 -0700846
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400847 i = depth;
Alex Tomasa86c6182006-10-11 01:21:03 -0700848 /* walk through the tree */
849 while (i) {
850 ext_debug("depth %d: num %d, max %d\n",
851 ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
Alex Tomasc29c0ae2007-07-18 09:19:09 -0400852
Alex Tomasa86c6182006-10-11 01:21:03 -0700853 ext4_ext_binsearch_idx(inode, path + ppos, block);
Theodore Ts'obf89d162010-10-27 21:30:14 -0400854 path[ppos].p_block = ext4_idx_pblock(path[ppos].p_idx);
Alex Tomasa86c6182006-10-11 01:21:03 -0700855 path[ppos].p_depth = i;
856 path[ppos].p_ext = NULL;
857
Theodore Ts'o107a7bd2013-08-16 21:23:41 -0400858 bh = read_extent_tree_block(inode, path[ppos].p_block, --i,
859 flags);
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400860 if (IS_ERR(bh)) {
861 ret = PTR_ERR(bh);
Alex Tomasa86c6182006-10-11 01:21:03 -0700862 goto err;
Theodore Ts'o860d21e2013-01-12 16:19:36 -0500863 }
Theodore Ts'o7d7ea892013-08-16 21:20:41 -0400864
Alex Tomasa86c6182006-10-11 01:21:03 -0700865 eh = ext_block_hdr(bh);
866 ppos++;
Frank Mayhar273df552010-03-02 11:46:09 -0500867 if (unlikely(ppos > depth)) {
868 put_bh(bh);
869 EXT4_ERROR_INODE(inode,
870 "ppos %d > depth %d", ppos, depth);
Theodore Ts'o860d21e2013-01-12 16:19:36 -0500871 ret = -EIO;
Frank Mayhar273df552010-03-02 11:46:09 -0500872 goto err;
873 }
Alex Tomasa86c6182006-10-11 01:21:03 -0700874 path[ppos].p_bh = bh;
875 path[ppos].p_hdr = eh;
Alex Tomasa86c6182006-10-11 01:21:03 -0700876 }
877
878 path[ppos].p_depth = i;
Alex Tomasa86c6182006-10-11 01:21:03 -0700879 path[ppos].p_ext = NULL;
880 path[ppos].p_idx = NULL;
881
Alex Tomasa86c6182006-10-11 01:21:03 -0700882 /* find extent */
883 ext4_ext_binsearch(inode, path + ppos, block);
Shen Feng1973adc2008-07-11 19:27:31 -0400884 /* if not an empty leaf */
885 if (path[ppos].p_ext)
Theodore Ts'obf89d162010-10-27 21:30:14 -0400886 path[ppos].p_block = ext4_ext_pblock(path[ppos].p_ext);
Alex Tomasa86c6182006-10-11 01:21:03 -0700887
888 ext4_ext_show_path(inode, path);
889
890 return path;
891
892err:
893 ext4_ext_drop_refs(path);
894 if (alloc)
895 kfree(path);
Theodore Ts'o860d21e2013-01-12 16:19:36 -0500896 return ERR_PTR(ret);
Alex Tomasa86c6182006-10-11 01:21:03 -0700897}
898
899/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700900 * ext4_ext_insert_index:
901 * insert new index [@logical;@ptr] into the block at @curp;
902 * check where to insert: before @curp or after @curp
Alex Tomasa86c6182006-10-11 01:21:03 -0700903 */
Theodore Ts'o1f109d52010-10-27 21:30:14 -0400904static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
905 struct ext4_ext_path *curp,
906 int logical, ext4_fsblk_t ptr)
Alex Tomasa86c6182006-10-11 01:21:03 -0700907{
908 struct ext4_extent_idx *ix;
909 int len, err;
910
Avantika Mathur7e028972006-12-06 20:41:33 -0800911 err = ext4_ext_get_access(handle, inode, curp);
912 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -0700913 return err;
914
Frank Mayhar273df552010-03-02 11:46:09 -0500915 if (unlikely(logical == le32_to_cpu(curp->p_idx->ei_block))) {
916 EXT4_ERROR_INODE(inode,
917 "logical %d == ei_block %d!",
918 logical, le32_to_cpu(curp->p_idx->ei_block));
919 return -EIO;
920 }
Robin Dongd4620312011-07-17 23:43:42 -0400921
922 if (unlikely(le16_to_cpu(curp->p_hdr->eh_entries)
923 >= le16_to_cpu(curp->p_hdr->eh_max))) {
924 EXT4_ERROR_INODE(inode,
925 "eh_entries %d >= eh_max %d!",
926 le16_to_cpu(curp->p_hdr->eh_entries),
927 le16_to_cpu(curp->p_hdr->eh_max));
928 return -EIO;
929 }
930
Alex Tomasa86c6182006-10-11 01:21:03 -0700931 if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
932 /* insert after */
Eric Gouriou80e675f2011-10-27 11:52:18 -0400933 ext_debug("insert new index %d after: %llu\n", logical, ptr);
Alex Tomasa86c6182006-10-11 01:21:03 -0700934 ix = curp->p_idx + 1;
935 } else {
936 /* insert before */
Eric Gouriou80e675f2011-10-27 11:52:18 -0400937 ext_debug("insert new index %d before: %llu\n", logical, ptr);
Alex Tomasa86c6182006-10-11 01:21:03 -0700938 ix = curp->p_idx;
939 }
940
Eric Gouriou80e675f2011-10-27 11:52:18 -0400941 len = EXT_LAST_INDEX(curp->p_hdr) - ix + 1;
942 BUG_ON(len < 0);
943 if (len > 0) {
944 ext_debug("insert new index %d: "
945 "move %d indices from 0x%p to 0x%p\n",
946 logical, len, ix, ix + 1);
947 memmove(ix + 1, ix, len * sizeof(struct ext4_extent_idx));
948 }
949
Tao Maf472e022011-10-17 10:13:46 -0400950 if (unlikely(ix > EXT_MAX_INDEX(curp->p_hdr))) {
951 EXT4_ERROR_INODE(inode, "ix > EXT_MAX_INDEX!");
952 return -EIO;
953 }
954
Alex Tomasa86c6182006-10-11 01:21:03 -0700955 ix->ei_block = cpu_to_le32(logical);
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700956 ext4_idx_store_pblock(ix, ptr);
Marcin Slusarze8546d02008-04-17 10:38:59 -0400957 le16_add_cpu(&curp->p_hdr->eh_entries, 1);
Alex Tomasa86c6182006-10-11 01:21:03 -0700958
Frank Mayhar273df552010-03-02 11:46:09 -0500959 if (unlikely(ix > EXT_LAST_INDEX(curp->p_hdr))) {
960 EXT4_ERROR_INODE(inode, "ix > EXT_LAST_INDEX!");
961 return -EIO;
962 }
Alex Tomasa86c6182006-10-11 01:21:03 -0700963
964 err = ext4_ext_dirty(handle, inode, curp);
965 ext4_std_error(inode->i_sb, err);
966
967 return err;
968}
969
970/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700971 * ext4_ext_split:
972 * inserts new subtree into the path, using free index entry
973 * at depth @at:
974 * - allocates all needed blocks (new leaf and all intermediate index blocks)
975 * - makes decision where to split
976 * - moves remaining extents and index entries (right to the split point)
977 * into the newly allocated blocks
978 * - initializes subtree
Alex Tomasa86c6182006-10-11 01:21:03 -0700979 */
980static int ext4_ext_split(handle_t *handle, struct inode *inode,
Allison Henderson55f020d2011-05-25 07:41:26 -0400981 unsigned int flags,
982 struct ext4_ext_path *path,
983 struct ext4_extent *newext, int at)
Alex Tomasa86c6182006-10-11 01:21:03 -0700984{
985 struct buffer_head *bh = NULL;
986 int depth = ext_depth(inode);
987 struct ext4_extent_header *neh;
988 struct ext4_extent_idx *fidx;
Alex Tomasa86c6182006-10-11 01:21:03 -0700989 int i = at, k, m, a;
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700990 ext4_fsblk_t newblock, oldblock;
Alex Tomasa86c6182006-10-11 01:21:03 -0700991 __le32 border;
Alex Tomasf65e6fb2006-10-11 01:21:05 -0700992 ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
Alex Tomasa86c6182006-10-11 01:21:03 -0700993 int err = 0;
994
995 /* make decision: where to split? */
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700996 /* FIXME: now decision is simplest: at current extent */
Alex Tomasa86c6182006-10-11 01:21:03 -0700997
Randy Dunlapd0d856e2006-10-11 01:21:07 -0700998 /* if current leaf will be split, then we should use
Alex Tomasa86c6182006-10-11 01:21:03 -0700999 * border from split point */
Frank Mayhar273df552010-03-02 11:46:09 -05001000 if (unlikely(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr))) {
1001 EXT4_ERROR_INODE(inode, "p_ext > EXT_MAX_EXTENT!");
1002 return -EIO;
1003 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001004 if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
1005 border = path[depth].p_ext[1].ee_block;
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001006 ext_debug("leaf will be split."
Alex Tomasa86c6182006-10-11 01:21:03 -07001007 " next leaf starts at %d\n",
Dave Kleikamp8c55e202007-05-24 13:04:54 -04001008 le32_to_cpu(border));
Alex Tomasa86c6182006-10-11 01:21:03 -07001009 } else {
1010 border = newext->ee_block;
1011 ext_debug("leaf will be added."
1012 " next leaf starts at %d\n",
Dave Kleikamp8c55e202007-05-24 13:04:54 -04001013 le32_to_cpu(border));
Alex Tomasa86c6182006-10-11 01:21:03 -07001014 }
1015
1016 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001017 * If error occurs, then we break processing
1018 * and mark filesystem read-only. index won't
Alex Tomasa86c6182006-10-11 01:21:03 -07001019 * be inserted and tree will be in consistent
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001020 * state. Next mount will repair buffers too.
Alex Tomasa86c6182006-10-11 01:21:03 -07001021 */
1022
1023 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001024 * Get array to track all allocated blocks.
1025 * We need this to handle errors and free blocks
1026 * upon them.
Alex Tomasa86c6182006-10-11 01:21:03 -07001027 */
Avantika Mathur5d4958f2006-12-06 20:41:35 -08001028 ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
Alex Tomasa86c6182006-10-11 01:21:03 -07001029 if (!ablocks)
1030 return -ENOMEM;
Alex Tomasa86c6182006-10-11 01:21:03 -07001031
1032 /* allocate all needed blocks */
1033 ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
1034 for (a = 0; a < depth - at; a++) {
Aneesh Kumar K.V654b4902008-07-11 19:27:31 -04001035 newblock = ext4_ext_new_meta_block(handle, inode, path,
Allison Henderson55f020d2011-05-25 07:41:26 -04001036 newext, &err, flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07001037 if (newblock == 0)
1038 goto cleanup;
1039 ablocks[a] = newblock;
1040 }
1041
1042 /* initialize new leaf */
1043 newblock = ablocks[--a];
Frank Mayhar273df552010-03-02 11:46:09 -05001044 if (unlikely(newblock == 0)) {
1045 EXT4_ERROR_INODE(inode, "newblock == 0!");
1046 err = -EIO;
1047 goto cleanup;
1048 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001049 bh = sb_getblk(inode->i_sb, newblock);
Wang Shilongaebf0242013-01-12 16:28:47 -05001050 if (unlikely(!bh)) {
Theodore Ts'o860d21e2013-01-12 16:19:36 -05001051 err = -ENOMEM;
Alex Tomasa86c6182006-10-11 01:21:03 -07001052 goto cleanup;
1053 }
1054 lock_buffer(bh);
1055
Avantika Mathur7e028972006-12-06 20:41:33 -08001056 err = ext4_journal_get_create_access(handle, bh);
1057 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001058 goto cleanup;
1059
1060 neh = ext_block_hdr(bh);
1061 neh->eh_entries = 0;
Theodore Ts'o55ad63b2009-08-28 10:40:33 -04001062 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -07001063 neh->eh_magic = EXT4_EXT_MAGIC;
1064 neh->eh_depth = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07001065
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001066 /* move remainder of path[depth] to the new leaf */
Frank Mayhar273df552010-03-02 11:46:09 -05001067 if (unlikely(path[depth].p_hdr->eh_entries !=
1068 path[depth].p_hdr->eh_max)) {
1069 EXT4_ERROR_INODE(inode, "eh_entries %d != eh_max %d!",
1070 path[depth].p_hdr->eh_entries,
1071 path[depth].p_hdr->eh_max);
1072 err = -EIO;
1073 goto cleanup;
1074 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001075 /* start copy from next extent */
Yongqiang Yang1b16da72011-05-25 17:41:48 -04001076 m = EXT_MAX_EXTENT(path[depth].p_hdr) - path[depth].p_ext++;
1077 ext4_ext_show_move(inode, path, newblock, depth);
Alex Tomasa86c6182006-10-11 01:21:03 -07001078 if (m) {
Yongqiang Yang1b16da72011-05-25 17:41:48 -04001079 struct ext4_extent *ex;
1080 ex = EXT_FIRST_EXTENT(neh);
1081 memmove(ex, path[depth].p_ext, sizeof(struct ext4_extent) * m);
Marcin Slusarze8546d02008-04-17 10:38:59 -04001082 le16_add_cpu(&neh->eh_entries, m);
Alex Tomasa86c6182006-10-11 01:21:03 -07001083 }
1084
Darrick J. Wong7ac59902012-04-29 18:37:10 -04001085 ext4_extent_block_csum_set(inode, neh);
Alex Tomasa86c6182006-10-11 01:21:03 -07001086 set_buffer_uptodate(bh);
1087 unlock_buffer(bh);
1088
Frank Mayhar03901312009-01-07 00:06:22 -05001089 err = ext4_handle_dirty_metadata(handle, inode, bh);
Avantika Mathur7e028972006-12-06 20:41:33 -08001090 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001091 goto cleanup;
1092 brelse(bh);
1093 bh = NULL;
1094
1095 /* correct old leaf */
1096 if (m) {
Avantika Mathur7e028972006-12-06 20:41:33 -08001097 err = ext4_ext_get_access(handle, inode, path + depth);
1098 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001099 goto cleanup;
Marcin Slusarze8546d02008-04-17 10:38:59 -04001100 le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
Avantika Mathur7e028972006-12-06 20:41:33 -08001101 err = ext4_ext_dirty(handle, inode, path + depth);
1102 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001103 goto cleanup;
1104
1105 }
1106
1107 /* create intermediate indexes */
1108 k = depth - at - 1;
Frank Mayhar273df552010-03-02 11:46:09 -05001109 if (unlikely(k < 0)) {
1110 EXT4_ERROR_INODE(inode, "k %d < 0!", k);
1111 err = -EIO;
1112 goto cleanup;
1113 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001114 if (k)
1115 ext_debug("create %d intermediate indices\n", k);
1116 /* insert new index into current index block */
1117 /* current depth stored in i var */
1118 i = depth - 1;
1119 while (k--) {
1120 oldblock = newblock;
1121 newblock = ablocks[--a];
Eric Sandeenbba90742008-01-28 23:58:27 -05001122 bh = sb_getblk(inode->i_sb, newblock);
Wang Shilongaebf0242013-01-12 16:28:47 -05001123 if (unlikely(!bh)) {
Theodore Ts'o860d21e2013-01-12 16:19:36 -05001124 err = -ENOMEM;
Alex Tomasa86c6182006-10-11 01:21:03 -07001125 goto cleanup;
1126 }
1127 lock_buffer(bh);
1128
Avantika Mathur7e028972006-12-06 20:41:33 -08001129 err = ext4_journal_get_create_access(handle, bh);
1130 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001131 goto cleanup;
1132
1133 neh = ext_block_hdr(bh);
1134 neh->eh_entries = cpu_to_le16(1);
1135 neh->eh_magic = EXT4_EXT_MAGIC;
Theodore Ts'o55ad63b2009-08-28 10:40:33 -04001136 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -07001137 neh->eh_depth = cpu_to_le16(depth - i);
1138 fidx = EXT_FIRST_INDEX(neh);
1139 fidx->ei_block = border;
Alex Tomasf65e6fb2006-10-11 01:21:05 -07001140 ext4_idx_store_pblock(fidx, oldblock);
Alex Tomasa86c6182006-10-11 01:21:03 -07001141
Eric Sandeenbba90742008-01-28 23:58:27 -05001142 ext_debug("int.index at %d (block %llu): %u -> %llu\n",
1143 i, newblock, le32_to_cpu(border), oldblock);
Alex Tomasa86c6182006-10-11 01:21:03 -07001144
Yongqiang Yang1b16da72011-05-25 17:41:48 -04001145 /* move remainder of path[i] to the new index block */
Frank Mayhar273df552010-03-02 11:46:09 -05001146 if (unlikely(EXT_MAX_INDEX(path[i].p_hdr) !=
1147 EXT_LAST_INDEX(path[i].p_hdr))) {
1148 EXT4_ERROR_INODE(inode,
1149 "EXT_MAX_INDEX != EXT_LAST_INDEX ee_block %d!",
1150 le32_to_cpu(path[i].p_ext->ee_block));
1151 err = -EIO;
1152 goto cleanup;
1153 }
Yongqiang Yang1b16da72011-05-25 17:41:48 -04001154 /* start copy indexes */
1155 m = EXT_MAX_INDEX(path[i].p_hdr) - path[i].p_idx++;
1156 ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
1157 EXT_MAX_INDEX(path[i].p_hdr));
1158 ext4_ext_show_move(inode, path, newblock, i);
Alex Tomasa86c6182006-10-11 01:21:03 -07001159 if (m) {
Yongqiang Yang1b16da72011-05-25 17:41:48 -04001160 memmove(++fidx, path[i].p_idx,
Alex Tomasa86c6182006-10-11 01:21:03 -07001161 sizeof(struct ext4_extent_idx) * m);
Marcin Slusarze8546d02008-04-17 10:38:59 -04001162 le16_add_cpu(&neh->eh_entries, m);
Alex Tomasa86c6182006-10-11 01:21:03 -07001163 }
Darrick J. Wong7ac59902012-04-29 18:37:10 -04001164 ext4_extent_block_csum_set(inode, neh);
Alex Tomasa86c6182006-10-11 01:21:03 -07001165 set_buffer_uptodate(bh);
1166 unlock_buffer(bh);
1167
Frank Mayhar03901312009-01-07 00:06:22 -05001168 err = ext4_handle_dirty_metadata(handle, inode, bh);
Avantika Mathur7e028972006-12-06 20:41:33 -08001169 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001170 goto cleanup;
1171 brelse(bh);
1172 bh = NULL;
1173
1174 /* correct old index */
1175 if (m) {
1176 err = ext4_ext_get_access(handle, inode, path + i);
1177 if (err)
1178 goto cleanup;
Marcin Slusarze8546d02008-04-17 10:38:59 -04001179 le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
Alex Tomasa86c6182006-10-11 01:21:03 -07001180 err = ext4_ext_dirty(handle, inode, path + i);
1181 if (err)
1182 goto cleanup;
1183 }
1184
1185 i--;
1186 }
1187
1188 /* insert new index */
Alex Tomasa86c6182006-10-11 01:21:03 -07001189 err = ext4_ext_insert_index(handle, inode, path + at,
1190 le32_to_cpu(border), newblock);
1191
1192cleanup:
1193 if (bh) {
1194 if (buffer_locked(bh))
1195 unlock_buffer(bh);
1196 brelse(bh);
1197 }
1198
1199 if (err) {
1200 /* free all allocated blocks in error case */
1201 for (i = 0; i < depth; i++) {
1202 if (!ablocks[i])
1203 continue;
Peter Huewe7dc57612011-02-21 21:01:42 -05001204 ext4_free_blocks(handle, inode, NULL, ablocks[i], 1,
Theodore Ts'oe6362602009-11-23 07:17:05 -05001205 EXT4_FREE_BLOCKS_METADATA);
Alex Tomasa86c6182006-10-11 01:21:03 -07001206 }
1207 }
1208 kfree(ablocks);
1209
1210 return err;
1211}
1212
1213/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001214 * ext4_ext_grow_indepth:
1215 * implements tree growing procedure:
1216 * - allocates new block
1217 * - moves top-level data (index block or leaf) into the new block
1218 * - initializes new top-level, creating index that points to the
1219 * just created block
Alex Tomasa86c6182006-10-11 01:21:03 -07001220 */
1221static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
Allison Henderson55f020d2011-05-25 07:41:26 -04001222 unsigned int flags,
Allison Henderson55f020d2011-05-25 07:41:26 -04001223 struct ext4_extent *newext)
Alex Tomasa86c6182006-10-11 01:21:03 -07001224{
Alex Tomasa86c6182006-10-11 01:21:03 -07001225 struct ext4_extent_header *neh;
Alex Tomasa86c6182006-10-11 01:21:03 -07001226 struct buffer_head *bh;
Alex Tomasf65e6fb2006-10-11 01:21:05 -07001227 ext4_fsblk_t newblock;
Alex Tomasa86c6182006-10-11 01:21:03 -07001228 int err = 0;
1229
Dmitry Monakhov1939dd82011-10-22 01:26:05 -04001230 newblock = ext4_ext_new_meta_block(handle, inode, NULL,
Allison Henderson55f020d2011-05-25 07:41:26 -04001231 newext, &err, flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07001232 if (newblock == 0)
1233 return err;
1234
1235 bh = sb_getblk(inode->i_sb, newblock);
Wang Shilongaebf0242013-01-12 16:28:47 -05001236 if (unlikely(!bh))
Theodore Ts'o860d21e2013-01-12 16:19:36 -05001237 return -ENOMEM;
Alex Tomasa86c6182006-10-11 01:21:03 -07001238 lock_buffer(bh);
1239
Avantika Mathur7e028972006-12-06 20:41:33 -08001240 err = ext4_journal_get_create_access(handle, bh);
1241 if (err) {
Alex Tomasa86c6182006-10-11 01:21:03 -07001242 unlock_buffer(bh);
1243 goto out;
1244 }
1245
1246 /* move top-level index/leaf into new block */
Dmitry Monakhov1939dd82011-10-22 01:26:05 -04001247 memmove(bh->b_data, EXT4_I(inode)->i_data,
1248 sizeof(EXT4_I(inode)->i_data));
Alex Tomasa86c6182006-10-11 01:21:03 -07001249
1250 /* set size of new block */
1251 neh = ext_block_hdr(bh);
1252 /* old root could have indexes or leaves
1253 * so calculate e_max right way */
1254 if (ext_depth(inode))
Theodore Ts'o55ad63b2009-08-28 10:40:33 -04001255 neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -07001256 else
Theodore Ts'o55ad63b2009-08-28 10:40:33 -04001257 neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -07001258 neh->eh_magic = EXT4_EXT_MAGIC;
Darrick J. Wong7ac59902012-04-29 18:37:10 -04001259 ext4_extent_block_csum_set(inode, neh);
Alex Tomasa86c6182006-10-11 01:21:03 -07001260 set_buffer_uptodate(bh);
1261 unlock_buffer(bh);
1262
Frank Mayhar03901312009-01-07 00:06:22 -05001263 err = ext4_handle_dirty_metadata(handle, inode, bh);
Avantika Mathur7e028972006-12-06 20:41:33 -08001264 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001265 goto out;
1266
Dmitry Monakhov1939dd82011-10-22 01:26:05 -04001267 /* Update top-level index: num,max,pointer */
Alex Tomasa86c6182006-10-11 01:21:03 -07001268 neh = ext_inode_hdr(inode);
Dmitry Monakhov1939dd82011-10-22 01:26:05 -04001269 neh->eh_entries = cpu_to_le16(1);
1270 ext4_idx_store_pblock(EXT_FIRST_INDEX(neh), newblock);
1271 if (neh->eh_depth == 0) {
1272 /* Root extent block becomes index block */
1273 neh->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode, 0));
1274 EXT_FIRST_INDEX(neh)->ei_block =
1275 EXT_FIRST_EXTENT(neh)->ee_block;
1276 }
Mingming Cao2ae02102006-10-11 01:21:11 -07001277 ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
Alex Tomasa86c6182006-10-11 01:21:03 -07001278 le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
Andi Kleen5a0790c2010-06-14 13:28:03 -04001279 le32_to_cpu(EXT_FIRST_INDEX(neh)->ei_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -04001280 ext4_idx_pblock(EXT_FIRST_INDEX(neh)));
Alex Tomasa86c6182006-10-11 01:21:03 -07001281
Wei Yongjunba39ebb2012-09-27 09:37:53 -04001282 le16_add_cpu(&neh->eh_depth, 1);
Dmitry Monakhov1939dd82011-10-22 01:26:05 -04001283 ext4_mark_inode_dirty(handle, inode);
Alex Tomasa86c6182006-10-11 01:21:03 -07001284out:
1285 brelse(bh);
1286
1287 return err;
1288}
1289
1290/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001291 * ext4_ext_create_new_leaf:
1292 * finds empty index and adds new leaf.
1293 * if no free index is found, then it requests in-depth growing.
Alex Tomasa86c6182006-10-11 01:21:03 -07001294 */
1295static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001296 unsigned int mb_flags,
1297 unsigned int gb_flags,
Allison Henderson55f020d2011-05-25 07:41:26 -04001298 struct ext4_ext_path *path,
1299 struct ext4_extent *newext)
Alex Tomasa86c6182006-10-11 01:21:03 -07001300{
1301 struct ext4_ext_path *curp;
1302 int depth, i, err = 0;
1303
1304repeat:
1305 i = depth = ext_depth(inode);
1306
1307 /* walk up to the tree and look for free index entry */
1308 curp = path + depth;
1309 while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
1310 i--;
1311 curp--;
1312 }
1313
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001314 /* we use already allocated block for index block,
1315 * so subsequent data blocks should be contiguous */
Alex Tomasa86c6182006-10-11 01:21:03 -07001316 if (EXT_HAS_FREE_INDEX(curp)) {
1317 /* if we found index with free entry, then use that
1318 * entry: create all needed subtree and add new leaf */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001319 err = ext4_ext_split(handle, inode, mb_flags, path, newext, i);
Shen Feng787e0982008-07-11 19:27:31 -04001320 if (err)
1321 goto out;
Alex Tomasa86c6182006-10-11 01:21:03 -07001322
1323 /* refill path */
1324 ext4_ext_drop_refs(path);
1325 path = ext4_ext_find_extent(inode,
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001326 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001327 path, gb_flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07001328 if (IS_ERR(path))
1329 err = PTR_ERR(path);
1330 } else {
1331 /* tree is full, time to grow in depth */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001332 err = ext4_ext_grow_indepth(handle, inode, mb_flags, newext);
Alex Tomasa86c6182006-10-11 01:21:03 -07001333 if (err)
1334 goto out;
1335
1336 /* refill path */
1337 ext4_ext_drop_refs(path);
1338 path = ext4_ext_find_extent(inode,
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001339 (ext4_lblk_t)le32_to_cpu(newext->ee_block),
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001340 path, gb_flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07001341 if (IS_ERR(path)) {
1342 err = PTR_ERR(path);
1343 goto out;
1344 }
1345
1346 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001347 * only first (depth 0 -> 1) produces free space;
1348 * in all other cases we have to split the grown tree
Alex Tomasa86c6182006-10-11 01:21:03 -07001349 */
1350 depth = ext_depth(inode);
1351 if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001352 /* now we need to split */
Alex Tomasa86c6182006-10-11 01:21:03 -07001353 goto repeat;
1354 }
1355 }
1356
1357out:
1358 return err;
1359}
1360
1361/*
Alex Tomas1988b512008-01-28 23:58:27 -05001362 * search the closest allocated block to the left for *logical
1363 * and returns it at @logical + it's physical address at @phys
1364 * if *logical is the smallest allocated block, the function
1365 * returns 0 at @phys
1366 * return value contains 0 (success) or error code
1367 */
Theodore Ts'o1f109d52010-10-27 21:30:14 -04001368static int ext4_ext_search_left(struct inode *inode,
1369 struct ext4_ext_path *path,
1370 ext4_lblk_t *logical, ext4_fsblk_t *phys)
Alex Tomas1988b512008-01-28 23:58:27 -05001371{
1372 struct ext4_extent_idx *ix;
1373 struct ext4_extent *ex;
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05001374 int depth, ee_len;
Alex Tomas1988b512008-01-28 23:58:27 -05001375
Frank Mayhar273df552010-03-02 11:46:09 -05001376 if (unlikely(path == NULL)) {
1377 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1378 return -EIO;
1379 }
Alex Tomas1988b512008-01-28 23:58:27 -05001380 depth = path->p_depth;
1381 *phys = 0;
1382
1383 if (depth == 0 && path->p_ext == NULL)
1384 return 0;
1385
1386 /* usually extent in the path covers blocks smaller
1387 * then *logical, but it can be that extent is the
1388 * first one in the file */
1389
1390 ex = path[depth].p_ext;
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05001391 ee_len = ext4_ext_get_actual_len(ex);
Alex Tomas1988b512008-01-28 23:58:27 -05001392 if (*logical < le32_to_cpu(ex->ee_block)) {
Frank Mayhar273df552010-03-02 11:46:09 -05001393 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1394 EXT4_ERROR_INODE(inode,
1395 "EXT_FIRST_EXTENT != ex *logical %d ee_block %d!",
1396 *logical, le32_to_cpu(ex->ee_block));
1397 return -EIO;
1398 }
Alex Tomas1988b512008-01-28 23:58:27 -05001399 while (--depth >= 0) {
1400 ix = path[depth].p_idx;
Frank Mayhar273df552010-03-02 11:46:09 -05001401 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1402 EXT4_ERROR_INODE(inode,
1403 "ix (%d) != EXT_FIRST_INDEX (%d) (depth %d)!",
Tao Ma6ee3b212011-10-08 16:08:34 -04001404 ix != NULL ? le32_to_cpu(ix->ei_block) : 0,
Frank Mayhar273df552010-03-02 11:46:09 -05001405 EXT_FIRST_INDEX(path[depth].p_hdr) != NULL ?
Tao Ma6ee3b212011-10-08 16:08:34 -04001406 le32_to_cpu(EXT_FIRST_INDEX(path[depth].p_hdr)->ei_block) : 0,
Frank Mayhar273df552010-03-02 11:46:09 -05001407 depth);
1408 return -EIO;
1409 }
Alex Tomas1988b512008-01-28 23:58:27 -05001410 }
1411 return 0;
1412 }
1413
Frank Mayhar273df552010-03-02 11:46:09 -05001414 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1415 EXT4_ERROR_INODE(inode,
1416 "logical %d < ee_block %d + ee_len %d!",
1417 *logical, le32_to_cpu(ex->ee_block), ee_len);
1418 return -EIO;
1419 }
Alex Tomas1988b512008-01-28 23:58:27 -05001420
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05001421 *logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
Theodore Ts'obf89d162010-10-27 21:30:14 -04001422 *phys = ext4_ext_pblock(ex) + ee_len - 1;
Alex Tomas1988b512008-01-28 23:58:27 -05001423 return 0;
1424}
1425
1426/*
1427 * search the closest allocated block to the right for *logical
1428 * and returns it at @logical + it's physical address at @phys
Tao Madf3ab172011-10-08 15:53:49 -04001429 * if *logical is the largest allocated block, the function
Alex Tomas1988b512008-01-28 23:58:27 -05001430 * returns 0 at @phys
1431 * return value contains 0 (success) or error code
1432 */
Theodore Ts'o1f109d52010-10-27 21:30:14 -04001433static int ext4_ext_search_right(struct inode *inode,
1434 struct ext4_ext_path *path,
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001435 ext4_lblk_t *logical, ext4_fsblk_t *phys,
1436 struct ext4_extent **ret_ex)
Alex Tomas1988b512008-01-28 23:58:27 -05001437{
1438 struct buffer_head *bh = NULL;
1439 struct ext4_extent_header *eh;
1440 struct ext4_extent_idx *ix;
1441 struct ext4_extent *ex;
1442 ext4_fsblk_t block;
Eric Sandeen395a87b2009-03-10 18:18:47 -04001443 int depth; /* Note, NOT eh_depth; depth from top of tree */
1444 int ee_len;
Alex Tomas1988b512008-01-28 23:58:27 -05001445
Frank Mayhar273df552010-03-02 11:46:09 -05001446 if (unlikely(path == NULL)) {
1447 EXT4_ERROR_INODE(inode, "path == NULL *logical %d!", *logical);
1448 return -EIO;
1449 }
Alex Tomas1988b512008-01-28 23:58:27 -05001450 depth = path->p_depth;
1451 *phys = 0;
1452
1453 if (depth == 0 && path->p_ext == NULL)
1454 return 0;
1455
1456 /* usually extent in the path covers blocks smaller
1457 * then *logical, but it can be that extent is the
1458 * first one in the file */
1459
1460 ex = path[depth].p_ext;
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05001461 ee_len = ext4_ext_get_actual_len(ex);
Alex Tomas1988b512008-01-28 23:58:27 -05001462 if (*logical < le32_to_cpu(ex->ee_block)) {
Frank Mayhar273df552010-03-02 11:46:09 -05001463 if (unlikely(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex)) {
1464 EXT4_ERROR_INODE(inode,
1465 "first_extent(path[%d].p_hdr) != ex",
1466 depth);
1467 return -EIO;
1468 }
Alex Tomas1988b512008-01-28 23:58:27 -05001469 while (--depth >= 0) {
1470 ix = path[depth].p_idx;
Frank Mayhar273df552010-03-02 11:46:09 -05001471 if (unlikely(ix != EXT_FIRST_INDEX(path[depth].p_hdr))) {
1472 EXT4_ERROR_INODE(inode,
1473 "ix != EXT_FIRST_INDEX *logical %d!",
1474 *logical);
1475 return -EIO;
1476 }
Alex Tomas1988b512008-01-28 23:58:27 -05001477 }
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001478 goto found_extent;
Alex Tomas1988b512008-01-28 23:58:27 -05001479 }
1480
Frank Mayhar273df552010-03-02 11:46:09 -05001481 if (unlikely(*logical < (le32_to_cpu(ex->ee_block) + ee_len))) {
1482 EXT4_ERROR_INODE(inode,
1483 "logical %d < ee_block %d + ee_len %d!",
1484 *logical, le32_to_cpu(ex->ee_block), ee_len);
1485 return -EIO;
1486 }
Alex Tomas1988b512008-01-28 23:58:27 -05001487
1488 if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
1489 /* next allocated block in this leaf */
1490 ex++;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001491 goto found_extent;
Alex Tomas1988b512008-01-28 23:58:27 -05001492 }
1493
1494 /* go up and search for index to the right */
1495 while (--depth >= 0) {
1496 ix = path[depth].p_idx;
1497 if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
Wu Fengguang25f1ee32008-11-25 17:24:23 -05001498 goto got_index;
Alex Tomas1988b512008-01-28 23:58:27 -05001499 }
1500
Wu Fengguang25f1ee32008-11-25 17:24:23 -05001501 /* we've gone up to the root and found no index to the right */
1502 return 0;
Alex Tomas1988b512008-01-28 23:58:27 -05001503
Wu Fengguang25f1ee32008-11-25 17:24:23 -05001504got_index:
Alex Tomas1988b512008-01-28 23:58:27 -05001505 /* we've found index to the right, let's
1506 * follow it and find the closest allocated
1507 * block to the right */
1508 ix++;
Theodore Ts'obf89d162010-10-27 21:30:14 -04001509 block = ext4_idx_pblock(ix);
Alex Tomas1988b512008-01-28 23:58:27 -05001510 while (++depth < path->p_depth) {
Eric Sandeen395a87b2009-03-10 18:18:47 -04001511 /* subtract from p_depth to get proper eh_depth */
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04001512 bh = read_extent_tree_block(inode, block,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001513 path->p_depth - depth, 0);
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04001514 if (IS_ERR(bh))
1515 return PTR_ERR(bh);
1516 eh = ext_block_hdr(bh);
Alex Tomas1988b512008-01-28 23:58:27 -05001517 ix = EXT_FIRST_INDEX(eh);
Theodore Ts'obf89d162010-10-27 21:30:14 -04001518 block = ext4_idx_pblock(ix);
Alex Tomas1988b512008-01-28 23:58:27 -05001519 put_bh(bh);
1520 }
1521
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001522 bh = read_extent_tree_block(inode, block, path->p_depth - depth, 0);
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04001523 if (IS_ERR(bh))
1524 return PTR_ERR(bh);
Alex Tomas1988b512008-01-28 23:58:27 -05001525 eh = ext_block_hdr(bh);
Alex Tomas1988b512008-01-28 23:58:27 -05001526 ex = EXT_FIRST_EXTENT(eh);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001527found_extent:
Alex Tomas1988b512008-01-28 23:58:27 -05001528 *logical = le32_to_cpu(ex->ee_block);
Theodore Ts'obf89d162010-10-27 21:30:14 -04001529 *phys = ext4_ext_pblock(ex);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001530 *ret_ex = ex;
1531 if (bh)
1532 put_bh(bh);
Alex Tomas1988b512008-01-28 23:58:27 -05001533 return 0;
Alex Tomas1988b512008-01-28 23:58:27 -05001534}
1535
1536/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001537 * ext4_ext_next_allocated_block:
Lukas Czernerf17722f2011-06-06 00:05:17 -04001538 * returns allocated block in subsequent extent or EXT_MAX_BLOCKS.
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001539 * NOTE: it considers block number from index entry as
1540 * allocated block. Thus, index entries have to be consistent
1541 * with leaves.
Alex Tomasa86c6182006-10-11 01:21:03 -07001542 */
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001543static ext4_lblk_t
Alex Tomasa86c6182006-10-11 01:21:03 -07001544ext4_ext_next_allocated_block(struct ext4_ext_path *path)
1545{
1546 int depth;
1547
1548 BUG_ON(path == NULL);
1549 depth = path->p_depth;
1550
1551 if (depth == 0 && path->p_ext == NULL)
Lukas Czernerf17722f2011-06-06 00:05:17 -04001552 return EXT_MAX_BLOCKS;
Alex Tomasa86c6182006-10-11 01:21:03 -07001553
1554 while (depth >= 0) {
1555 if (depth == path->p_depth) {
1556 /* leaf */
Curt Wohlgemuth6f8ff532011-10-26 04:38:59 -04001557 if (path[depth].p_ext &&
1558 path[depth].p_ext !=
Alex Tomasa86c6182006-10-11 01:21:03 -07001559 EXT_LAST_EXTENT(path[depth].p_hdr))
1560 return le32_to_cpu(path[depth].p_ext[1].ee_block);
1561 } else {
1562 /* index */
1563 if (path[depth].p_idx !=
1564 EXT_LAST_INDEX(path[depth].p_hdr))
1565 return le32_to_cpu(path[depth].p_idx[1].ei_block);
1566 }
1567 depth--;
1568 }
1569
Lukas Czernerf17722f2011-06-06 00:05:17 -04001570 return EXT_MAX_BLOCKS;
Alex Tomasa86c6182006-10-11 01:21:03 -07001571}
1572
1573/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001574 * ext4_ext_next_leaf_block:
Lukas Czernerf17722f2011-06-06 00:05:17 -04001575 * returns first allocated block from next leaf or EXT_MAX_BLOCKS
Alex Tomasa86c6182006-10-11 01:21:03 -07001576 */
Robin Dong57187892011-07-23 21:49:07 -04001577static ext4_lblk_t ext4_ext_next_leaf_block(struct ext4_ext_path *path)
Alex Tomasa86c6182006-10-11 01:21:03 -07001578{
1579 int depth;
1580
1581 BUG_ON(path == NULL);
1582 depth = path->p_depth;
1583
1584 /* zero-tree has no leaf blocks at all */
1585 if (depth == 0)
Lukas Czernerf17722f2011-06-06 00:05:17 -04001586 return EXT_MAX_BLOCKS;
Alex Tomasa86c6182006-10-11 01:21:03 -07001587
1588 /* go to index block */
1589 depth--;
1590
1591 while (depth >= 0) {
1592 if (path[depth].p_idx !=
1593 EXT_LAST_INDEX(path[depth].p_hdr))
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001594 return (ext4_lblk_t)
1595 le32_to_cpu(path[depth].p_idx[1].ei_block);
Alex Tomasa86c6182006-10-11 01:21:03 -07001596 depth--;
1597 }
1598
Lukas Czernerf17722f2011-06-06 00:05:17 -04001599 return EXT_MAX_BLOCKS;
Alex Tomasa86c6182006-10-11 01:21:03 -07001600}
1601
1602/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001603 * ext4_ext_correct_indexes:
1604 * if leaf gets modified and modified extent is first in the leaf,
1605 * then we have to correct all indexes above.
Alex Tomasa86c6182006-10-11 01:21:03 -07001606 * TODO: do we need to correct tree in all cases?
1607 */
Aneesh Kumar K.V1d03ec92008-01-28 23:58:27 -05001608static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
Alex Tomasa86c6182006-10-11 01:21:03 -07001609 struct ext4_ext_path *path)
1610{
1611 struct ext4_extent_header *eh;
1612 int depth = ext_depth(inode);
1613 struct ext4_extent *ex;
1614 __le32 border;
1615 int k, err = 0;
1616
1617 eh = path[depth].p_hdr;
1618 ex = path[depth].p_ext;
Frank Mayhar273df552010-03-02 11:46:09 -05001619
1620 if (unlikely(ex == NULL || eh == NULL)) {
1621 EXT4_ERROR_INODE(inode,
1622 "ex %p == NULL or eh %p == NULL", ex, eh);
1623 return -EIO;
1624 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001625
1626 if (depth == 0) {
1627 /* there is no tree at all */
1628 return 0;
1629 }
1630
1631 if (ex != EXT_FIRST_EXTENT(eh)) {
1632 /* we correct tree if first leaf got modified only */
1633 return 0;
1634 }
1635
1636 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001637 * TODO: we need correction if border is smaller than current one
Alex Tomasa86c6182006-10-11 01:21:03 -07001638 */
1639 k = depth - 1;
1640 border = path[depth].p_ext->ee_block;
Avantika Mathur7e028972006-12-06 20:41:33 -08001641 err = ext4_ext_get_access(handle, inode, path + k);
1642 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001643 return err;
1644 path[k].p_idx->ei_block = border;
Avantika Mathur7e028972006-12-06 20:41:33 -08001645 err = ext4_ext_dirty(handle, inode, path + k);
1646 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001647 return err;
1648
1649 while (k--) {
1650 /* change all left-side indexes */
1651 if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
1652 break;
Avantika Mathur7e028972006-12-06 20:41:33 -08001653 err = ext4_ext_get_access(handle, inode, path + k);
1654 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001655 break;
1656 path[k].p_idx->ei_block = border;
Avantika Mathur7e028972006-12-06 20:41:33 -08001657 err = ext4_ext_dirty(handle, inode, path + k);
1658 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07001659 break;
1660 }
1661
1662 return err;
1663}
1664
Akira Fujita748de672009-06-17 19:24:03 -04001665int
Alex Tomasa86c6182006-10-11 01:21:03 -07001666ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
1667 struct ext4_extent *ex2)
1668{
Eric Sandeenda0169b2013-11-04 09:58:26 -05001669 unsigned short ext1_ee_len, ext2_ee_len;
Amit Aroraa2df2a62007-07-17 21:42:41 -04001670
1671 /*
Dmitry Monakhovec22ba82013-03-04 00:36:06 -05001672 * Make sure that both extents are initialized. We don't merge
1673 * uninitialized extents so that we can be sure that end_io code has
1674 * the extent that was written properly split out and conversion to
1675 * initialized is trivial.
Amit Aroraa2df2a62007-07-17 21:42:41 -04001676 */
Dmitry Monakhovec22ba82013-03-04 00:36:06 -05001677 if (ext4_ext_is_uninitialized(ex1) || ext4_ext_is_uninitialized(ex2))
Amit Aroraa2df2a62007-07-17 21:42:41 -04001678 return 0;
1679
1680 ext1_ee_len = ext4_ext_get_actual_len(ex1);
1681 ext2_ee_len = ext4_ext_get_actual_len(ex2);
1682
1683 if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
Andrew Morton63f57932006-10-11 01:21:24 -07001684 le32_to_cpu(ex2->ee_block))
Alex Tomasa86c6182006-10-11 01:21:03 -07001685 return 0;
1686
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07001687 /*
1688 * To allow future support for preallocated extents to be added
1689 * as an RO_COMPAT feature, refuse to merge to extents if
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001690 * this can result in the top bit of ee_len being set.
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07001691 */
Eric Sandeenda0169b2013-11-04 09:58:26 -05001692 if (ext1_ee_len + ext2_ee_len > EXT_INIT_MAX_LEN)
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07001693 return 0;
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +01001694#ifdef AGGRESSIVE_TEST
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05001695 if (ext1_ee_len >= 4)
Alex Tomasa86c6182006-10-11 01:21:03 -07001696 return 0;
1697#endif
1698
Theodore Ts'obf89d162010-10-27 21:30:14 -04001699 if (ext4_ext_pblock(ex1) + ext1_ee_len == ext4_ext_pblock(ex2))
Alex Tomasa86c6182006-10-11 01:21:03 -07001700 return 1;
1701 return 0;
1702}
1703
1704/*
Amit Arora56055d32007-07-17 21:42:38 -04001705 * This function tries to merge the "ex" extent to the next extent in the tree.
1706 * It always tries to merge towards right. If you want to merge towards
1707 * left, pass "ex - 1" as argument instead of "ex".
1708 * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
1709 * 1 if they got merged.
1710 */
Yongqiang Yang197217a2011-05-03 11:45:29 -04001711static int ext4_ext_try_to_merge_right(struct inode *inode,
Theodore Ts'o1f109d52010-10-27 21:30:14 -04001712 struct ext4_ext_path *path,
1713 struct ext4_extent *ex)
Amit Arora56055d32007-07-17 21:42:38 -04001714{
1715 struct ext4_extent_header *eh;
1716 unsigned int depth, len;
1717 int merge_done = 0;
Amit Arora56055d32007-07-17 21:42:38 -04001718
1719 depth = ext_depth(inode);
1720 BUG_ON(path[depth].p_hdr == NULL);
1721 eh = path[depth].p_hdr;
1722
1723 while (ex < EXT_LAST_EXTENT(eh)) {
1724 if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
1725 break;
1726 /* merge with next extent! */
Amit Arora56055d32007-07-17 21:42:38 -04001727 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1728 + ext4_ext_get_actual_len(ex + 1));
Amit Arora56055d32007-07-17 21:42:38 -04001729
1730 if (ex + 1 < EXT_LAST_EXTENT(eh)) {
1731 len = (EXT_LAST_EXTENT(eh) - ex - 1)
1732 * sizeof(struct ext4_extent);
1733 memmove(ex + 1, ex + 2, len);
1734 }
Marcin Slusarze8546d02008-04-17 10:38:59 -04001735 le16_add_cpu(&eh->eh_entries, -1);
Amit Arora56055d32007-07-17 21:42:38 -04001736 merge_done = 1;
1737 WARN_ON(eh->eh_entries == 0);
1738 if (!eh->eh_entries)
Theodore Ts'o24676da2010-05-16 21:00:00 -04001739 EXT4_ERROR_INODE(inode, "eh->eh_entries = 0!");
Amit Arora56055d32007-07-17 21:42:38 -04001740 }
1741
1742 return merge_done;
1743}
1744
1745/*
Theodore Ts'oecb94f52012-08-17 09:44:17 -04001746 * This function does a very simple check to see if we can collapse
1747 * an extent tree with a single extent tree leaf block into the inode.
1748 */
1749static void ext4_ext_try_to_merge_up(handle_t *handle,
1750 struct inode *inode,
1751 struct ext4_ext_path *path)
1752{
1753 size_t s;
1754 unsigned max_root = ext4_ext_space_root(inode, 0);
1755 ext4_fsblk_t blk;
1756
1757 if ((path[0].p_depth != 1) ||
1758 (le16_to_cpu(path[0].p_hdr->eh_entries) != 1) ||
1759 (le16_to_cpu(path[1].p_hdr->eh_entries) > max_root))
1760 return;
1761
1762 /*
1763 * We need to modify the block allocation bitmap and the block
1764 * group descriptor to release the extent tree block. If we
1765 * can't get the journal credits, give up.
1766 */
1767 if (ext4_journal_extend(handle, 2))
1768 return;
1769
1770 /*
1771 * Copy the extent data up to the inode
1772 */
1773 blk = ext4_idx_pblock(path[0].p_idx);
1774 s = le16_to_cpu(path[1].p_hdr->eh_entries) *
1775 sizeof(struct ext4_extent_idx);
1776 s += sizeof(struct ext4_extent_header);
1777
1778 memcpy(path[0].p_hdr, path[1].p_hdr, s);
1779 path[0].p_depth = 0;
1780 path[0].p_ext = EXT_FIRST_EXTENT(path[0].p_hdr) +
1781 (path[1].p_ext - EXT_FIRST_EXTENT(path[1].p_hdr));
1782 path[0].p_hdr->eh_max = cpu_to_le16(max_root);
1783
1784 brelse(path[1].p_bh);
1785 ext4_free_blocks(handle, inode, NULL, blk, 1,
Jan Kara7d734532013-08-17 09:36:54 -04001786 EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET |
1787 EXT4_FREE_BLOCKS_RESERVE);
Theodore Ts'oecb94f52012-08-17 09:44:17 -04001788}
1789
1790/*
Yongqiang Yang197217a2011-05-03 11:45:29 -04001791 * This function tries to merge the @ex extent to neighbours in the tree.
1792 * return 1 if merge left else 0.
1793 */
Theodore Ts'oecb94f52012-08-17 09:44:17 -04001794static void ext4_ext_try_to_merge(handle_t *handle,
1795 struct inode *inode,
Yongqiang Yang197217a2011-05-03 11:45:29 -04001796 struct ext4_ext_path *path,
1797 struct ext4_extent *ex) {
1798 struct ext4_extent_header *eh;
1799 unsigned int depth;
1800 int merge_done = 0;
Yongqiang Yang197217a2011-05-03 11:45:29 -04001801
1802 depth = ext_depth(inode);
1803 BUG_ON(path[depth].p_hdr == NULL);
1804 eh = path[depth].p_hdr;
1805
1806 if (ex > EXT_FIRST_EXTENT(eh))
1807 merge_done = ext4_ext_try_to_merge_right(inode, path, ex - 1);
1808
1809 if (!merge_done)
Theodore Ts'oecb94f52012-08-17 09:44:17 -04001810 (void) ext4_ext_try_to_merge_right(inode, path, ex);
Yongqiang Yang197217a2011-05-03 11:45:29 -04001811
Theodore Ts'oecb94f52012-08-17 09:44:17 -04001812 ext4_ext_try_to_merge_up(handle, inode, path);
Yongqiang Yang197217a2011-05-03 11:45:29 -04001813}
1814
1815/*
Amit Arora25d14f92007-05-24 13:04:13 -04001816 * check if a portion of the "newext" extent overlaps with an
1817 * existing extent.
1818 *
1819 * If there is an overlap discovered, it updates the length of the newext
1820 * such that there will be no overlap, and then returns 1.
1821 * If there is no overlap found, it returns 0.
1822 */
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001823static unsigned int ext4_ext_check_overlap(struct ext4_sb_info *sbi,
1824 struct inode *inode,
Theodore Ts'o1f109d52010-10-27 21:30:14 -04001825 struct ext4_extent *newext,
1826 struct ext4_ext_path *path)
Amit Arora25d14f92007-05-24 13:04:13 -04001827{
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001828 ext4_lblk_t b1, b2;
Amit Arora25d14f92007-05-24 13:04:13 -04001829 unsigned int depth, len1;
1830 unsigned int ret = 0;
1831
1832 b1 = le32_to_cpu(newext->ee_block);
Amit Aroraa2df2a62007-07-17 21:42:41 -04001833 len1 = ext4_ext_get_actual_len(newext);
Amit Arora25d14f92007-05-24 13:04:13 -04001834 depth = ext_depth(inode);
1835 if (!path[depth].p_ext)
1836 goto out;
1837 b2 = le32_to_cpu(path[depth].p_ext->ee_block);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001838 b2 &= ~(sbi->s_cluster_ratio - 1);
Amit Arora25d14f92007-05-24 13:04:13 -04001839
1840 /*
1841 * get the next allocated block if the extent in the path
Theodore Ts'o2b2d6d02008-07-26 16:15:44 -04001842 * is before the requested block(s)
Amit Arora25d14f92007-05-24 13:04:13 -04001843 */
1844 if (b2 < b1) {
1845 b2 = ext4_ext_next_allocated_block(path);
Lukas Czernerf17722f2011-06-06 00:05:17 -04001846 if (b2 == EXT_MAX_BLOCKS)
Amit Arora25d14f92007-05-24 13:04:13 -04001847 goto out;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04001848 b2 &= ~(sbi->s_cluster_ratio - 1);
Amit Arora25d14f92007-05-24 13:04:13 -04001849 }
1850
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001851 /* check for wrap through zero on extent logical start block*/
Amit Arora25d14f92007-05-24 13:04:13 -04001852 if (b1 + len1 < b1) {
Lukas Czernerf17722f2011-06-06 00:05:17 -04001853 len1 = EXT_MAX_BLOCKS - b1;
Amit Arora25d14f92007-05-24 13:04:13 -04001854 newext->ee_len = cpu_to_le16(len1);
1855 ret = 1;
1856 }
1857
1858 /* check for overlap */
1859 if (b1 + len1 > b2) {
1860 newext->ee_len = cpu_to_le16(b2 - b1);
1861 ret = 1;
1862 }
1863out:
1864 return ret;
1865}
1866
1867/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001868 * ext4_ext_insert_extent:
1869 * tries to merge requsted extent into the existing extent or
1870 * inserts requested extent as new one into the tree,
1871 * creating new leaf in the no-space case.
Alex Tomasa86c6182006-10-11 01:21:03 -07001872 */
1873int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
1874 struct ext4_ext_path *path,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001875 struct ext4_extent *newext, int gb_flags)
Alex Tomasa86c6182006-10-11 01:21:03 -07001876{
Theodore Ts'oaf5bc922008-09-08 22:25:24 -04001877 struct ext4_extent_header *eh;
Alex Tomasa86c6182006-10-11 01:21:03 -07001878 struct ext4_extent *ex, *fex;
1879 struct ext4_extent *nearex; /* nearest extent */
1880 struct ext4_ext_path *npath = NULL;
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05001881 int depth, len, err;
1882 ext4_lblk_t next;
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001883 int mb_flags = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07001884
Frank Mayhar273df552010-03-02 11:46:09 -05001885 if (unlikely(ext4_ext_get_actual_len(newext) == 0)) {
1886 EXT4_ERROR_INODE(inode, "ext4_ext_get_actual_len(newext) == 0");
1887 return -EIO;
1888 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001889 depth = ext_depth(inode);
1890 ex = path[depth].p_ext;
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001891 eh = path[depth].p_hdr;
Frank Mayhar273df552010-03-02 11:46:09 -05001892 if (unlikely(path[depth].p_hdr == NULL)) {
1893 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
1894 return -EIO;
1895 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001896
1897 /* try to insert block into found extent and return */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001898 if (ex && !(gb_flags & EXT4_GET_BLOCKS_PRE_IO)) {
Amit Aroraa2df2a62007-07-17 21:42:41 -04001899
1900 /*
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001901 * Try to see whether we should rather test the extent on
1902 * right from ex, or from the left of ex. This is because
1903 * ext4_ext_find_extent() can return either extent on the
1904 * left, or on the right from the searched position. This
1905 * will make merging more effective.
Amit Aroraa2df2a62007-07-17 21:42:41 -04001906 */
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001907 if (ex < EXT_LAST_EXTENT(eh) &&
1908 (le32_to_cpu(ex->ee_block) +
1909 ext4_ext_get_actual_len(ex) <
1910 le32_to_cpu(newext->ee_block))) {
1911 ex += 1;
1912 goto prepend;
1913 } else if ((ex > EXT_FIRST_EXTENT(eh)) &&
1914 (le32_to_cpu(newext->ee_block) +
1915 ext4_ext_get_actual_len(newext) <
1916 le32_to_cpu(ex->ee_block)))
1917 ex -= 1;
1918
1919 /* Try to append newex to the ex */
1920 if (ext4_can_extents_be_merged(inode, ex, newext)) {
1921 ext_debug("append [%d]%d block to %u:[%d]%d"
1922 "(from %llu)\n",
1923 ext4_ext_is_uninitialized(newext),
1924 ext4_ext_get_actual_len(newext),
1925 le32_to_cpu(ex->ee_block),
1926 ext4_ext_is_uninitialized(ex),
1927 ext4_ext_get_actual_len(ex),
1928 ext4_ext_pblock(ex));
1929 err = ext4_ext_get_access(handle, inode,
1930 path + depth);
1931 if (err)
1932 return err;
1933
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001934 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
Amit Aroraa2df2a62007-07-17 21:42:41 -04001935 + ext4_ext_get_actual_len(newext));
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001936 eh = path[depth].p_hdr;
1937 nearex = ex;
1938 goto merge;
1939 }
1940
1941prepend:
1942 /* Try to prepend newex to the ex */
1943 if (ext4_can_extents_be_merged(inode, newext, ex)) {
1944 ext_debug("prepend %u[%d]%d block to %u:[%d]%d"
1945 "(from %llu)\n",
1946 le32_to_cpu(newext->ee_block),
1947 ext4_ext_is_uninitialized(newext),
1948 ext4_ext_get_actual_len(newext),
1949 le32_to_cpu(ex->ee_block),
1950 ext4_ext_is_uninitialized(ex),
1951 ext4_ext_get_actual_len(ex),
1952 ext4_ext_pblock(ex));
1953 err = ext4_ext_get_access(handle, inode,
1954 path + depth);
1955 if (err)
1956 return err;
1957
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001958 ex->ee_block = newext->ee_block;
1959 ext4_ext_store_pblock(ex, ext4_ext_pblock(newext));
1960 ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
1961 + ext4_ext_get_actual_len(newext));
Lukas Czernerbe8981b2013-04-03 23:33:28 -04001962 eh = path[depth].p_hdr;
1963 nearex = ex;
1964 goto merge;
1965 }
Alex Tomasa86c6182006-10-11 01:21:03 -07001966 }
1967
Alex Tomasa86c6182006-10-11 01:21:03 -07001968 depth = ext_depth(inode);
1969 eh = path[depth].p_hdr;
1970 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
1971 goto has_space;
1972
1973 /* probably next leaf has space for us? */
1974 fex = EXT_LAST_EXTENT(eh);
Robin Dong598dbdf2011-07-11 18:24:01 -04001975 next = EXT_MAX_BLOCKS;
1976 if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block))
Robin Dong57187892011-07-23 21:49:07 -04001977 next = ext4_ext_next_leaf_block(path);
Robin Dong598dbdf2011-07-11 18:24:01 -04001978 if (next != EXT_MAX_BLOCKS) {
Yongqiang Yang32de6752011-11-01 18:56:41 -04001979 ext_debug("next leaf block - %u\n", next);
Alex Tomasa86c6182006-10-11 01:21:03 -07001980 BUG_ON(npath != NULL);
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04001981 npath = ext4_ext_find_extent(inode, next, NULL, 0);
Alex Tomasa86c6182006-10-11 01:21:03 -07001982 if (IS_ERR(npath))
1983 return PTR_ERR(npath);
1984 BUG_ON(npath->p_depth != path->p_depth);
1985 eh = npath[depth].p_hdr;
1986 if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
Lucas De Marchi25985ed2011-03-30 22:57:33 -03001987 ext_debug("next leaf isn't full(%d)\n",
Alex Tomasa86c6182006-10-11 01:21:03 -07001988 le16_to_cpu(eh->eh_entries));
1989 path = npath;
Robin Dongffb505f2011-07-11 11:43:59 -04001990 goto has_space;
Alex Tomasa86c6182006-10-11 01:21:03 -07001991 }
1992 ext_debug("next leaf has no free space(%d,%d)\n",
1993 le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
1994 }
1995
1996 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07001997 * There is no free space in the found leaf.
1998 * We're gonna add a new leaf in the tree.
Alex Tomasa86c6182006-10-11 01:21:03 -07001999 */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002000 if (gb_flags & EXT4_GET_BLOCKS_METADATA_NOFAIL)
2001 mb_flags = EXT4_MB_USE_RESERVED;
2002 err = ext4_ext_create_new_leaf(handle, inode, mb_flags, gb_flags,
2003 path, newext);
Alex Tomasa86c6182006-10-11 01:21:03 -07002004 if (err)
2005 goto cleanup;
2006 depth = ext_depth(inode);
2007 eh = path[depth].p_hdr;
2008
2009has_space:
2010 nearex = path[depth].p_ext;
2011
Avantika Mathur7e028972006-12-06 20:41:33 -08002012 err = ext4_ext_get_access(handle, inode, path + depth);
2013 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07002014 goto cleanup;
2015
2016 if (!nearex) {
2017 /* there is no extent in this leaf, create first one */
Yongqiang Yang32de6752011-11-01 18:56:41 -04002018 ext_debug("first extent in the leaf: %u:%llu:[%d]%d\n",
Dave Kleikamp8c55e202007-05-24 13:04:54 -04002019 le32_to_cpu(newext->ee_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -04002020 ext4_ext_pblock(newext),
Mingming553f9002009-09-18 13:34:55 -04002021 ext4_ext_is_uninitialized(newext),
Amit Aroraa2df2a62007-07-17 21:42:41 -04002022 ext4_ext_get_actual_len(newext));
Eric Gouriou80e675f2011-10-27 11:52:18 -04002023 nearex = EXT_FIRST_EXTENT(eh);
2024 } else {
2025 if (le32_to_cpu(newext->ee_block)
Dave Kleikamp8c55e202007-05-24 13:04:54 -04002026 > le32_to_cpu(nearex->ee_block)) {
Eric Gouriou80e675f2011-10-27 11:52:18 -04002027 /* Insert after */
Yongqiang Yang32de6752011-11-01 18:56:41 -04002028 ext_debug("insert %u:%llu:[%d]%d before: "
2029 "nearest %p\n",
Dave Kleikamp8c55e202007-05-24 13:04:54 -04002030 le32_to_cpu(newext->ee_block),
Theodore Ts'obf89d162010-10-27 21:30:14 -04002031 ext4_ext_pblock(newext),
Mingming553f9002009-09-18 13:34:55 -04002032 ext4_ext_is_uninitialized(newext),
Amit Aroraa2df2a62007-07-17 21:42:41 -04002033 ext4_ext_get_actual_len(newext),
Eric Gouriou80e675f2011-10-27 11:52:18 -04002034 nearex);
2035 nearex++;
2036 } else {
2037 /* Insert before */
2038 BUG_ON(newext->ee_block == nearex->ee_block);
Yongqiang Yang32de6752011-11-01 18:56:41 -04002039 ext_debug("insert %u:%llu:[%d]%d after: "
2040 "nearest %p\n",
Eric Gouriou80e675f2011-10-27 11:52:18 -04002041 le32_to_cpu(newext->ee_block),
2042 ext4_ext_pblock(newext),
2043 ext4_ext_is_uninitialized(newext),
2044 ext4_ext_get_actual_len(newext),
2045 nearex);
Alex Tomasa86c6182006-10-11 01:21:03 -07002046 }
Eric Gouriou80e675f2011-10-27 11:52:18 -04002047 len = EXT_LAST_EXTENT(eh) - nearex + 1;
2048 if (len > 0) {
Yongqiang Yang32de6752011-11-01 18:56:41 -04002049 ext_debug("insert %u:%llu:[%d]%d: "
Eric Gouriou80e675f2011-10-27 11:52:18 -04002050 "move %d extents from 0x%p to 0x%p\n",
2051 le32_to_cpu(newext->ee_block),
2052 ext4_ext_pblock(newext),
2053 ext4_ext_is_uninitialized(newext),
2054 ext4_ext_get_actual_len(newext),
2055 len, nearex, nearex + 1);
2056 memmove(nearex + 1, nearex,
2057 len * sizeof(struct ext4_extent));
2058 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002059 }
2060
Marcin Slusarze8546d02008-04-17 10:38:59 -04002061 le16_add_cpu(&eh->eh_entries, 1);
Eric Gouriou80e675f2011-10-27 11:52:18 -04002062 path[depth].p_ext = nearex;
Alex Tomasa86c6182006-10-11 01:21:03 -07002063 nearex->ee_block = newext->ee_block;
Theodore Ts'obf89d162010-10-27 21:30:14 -04002064 ext4_ext_store_pblock(nearex, ext4_ext_pblock(newext));
Alex Tomasa86c6182006-10-11 01:21:03 -07002065 nearex->ee_len = newext->ee_len;
Alex Tomasa86c6182006-10-11 01:21:03 -07002066
2067merge:
HaiboLiue7bcf822012-07-09 16:29:28 -04002068 /* try to merge extents */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002069 if (!(gb_flags & EXT4_GET_BLOCKS_PRE_IO))
Theodore Ts'oecb94f52012-08-17 09:44:17 -04002070 ext4_ext_try_to_merge(handle, inode, path, nearex);
Alex Tomasa86c6182006-10-11 01:21:03 -07002071
Alex Tomasa86c6182006-10-11 01:21:03 -07002072
2073 /* time to correct all indexes above */
2074 err = ext4_ext_correct_indexes(handle, inode, path);
2075 if (err)
2076 goto cleanup;
2077
Theodore Ts'oecb94f52012-08-17 09:44:17 -04002078 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
Alex Tomasa86c6182006-10-11 01:21:03 -07002079
2080cleanup:
2081 if (npath) {
2082 ext4_ext_drop_refs(npath);
2083 kfree(npath);
2084 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002085 return err;
2086}
2087
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002088static int ext4_fill_fiemap_extents(struct inode *inode,
2089 ext4_lblk_t block, ext4_lblk_t num,
2090 struct fiemap_extent_info *fieinfo)
Eric Sandeen6873fa02008-10-07 00:46:36 -04002091{
2092 struct ext4_ext_path *path = NULL;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002093 struct ext4_extent *ex;
Zheng Liu69eb33d2013-02-18 00:31:07 -05002094 struct extent_status es;
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002095 ext4_lblk_t next, next_del, start = 0, end = 0;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002096 ext4_lblk_t last = block + num;
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002097 int exists, depth = 0, err = 0;
2098 unsigned int flags = 0;
2099 unsigned char blksize_bits = inode->i_sb->s_blocksize_bits;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002100
Lukas Czernerf17722f2011-06-06 00:05:17 -04002101 while (block < last && block != EXT_MAX_BLOCKS) {
Eric Sandeen6873fa02008-10-07 00:46:36 -04002102 num = last - block;
2103 /* find extent for this block */
Theodore Ts'ofab3a542009-12-09 21:30:02 -05002104 down_read(&EXT4_I(inode)->i_data_sem);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002105
2106 if (path && ext_depth(inode) != depth) {
2107 /* depth was changed. we have to realloc path */
2108 kfree(path);
2109 path = NULL;
2110 }
2111
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002112 path = ext4_ext_find_extent(inode, block, path, 0);
Eric Sandeen6873fa02008-10-07 00:46:36 -04002113 if (IS_ERR(path)) {
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002114 up_read(&EXT4_I(inode)->i_data_sem);
Eric Sandeen6873fa02008-10-07 00:46:36 -04002115 err = PTR_ERR(path);
2116 path = NULL;
2117 break;
2118 }
2119
2120 depth = ext_depth(inode);
Frank Mayhar273df552010-03-02 11:46:09 -05002121 if (unlikely(path[depth].p_hdr == NULL)) {
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002122 up_read(&EXT4_I(inode)->i_data_sem);
Frank Mayhar273df552010-03-02 11:46:09 -05002123 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
2124 err = -EIO;
2125 break;
2126 }
Eric Sandeen6873fa02008-10-07 00:46:36 -04002127 ex = path[depth].p_ext;
2128 next = ext4_ext_next_allocated_block(path);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002129 ext4_ext_drop_refs(path);
Eric Sandeen6873fa02008-10-07 00:46:36 -04002130
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002131 flags = 0;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002132 exists = 0;
2133 if (!ex) {
2134 /* there is no extent yet, so try to allocate
2135 * all requested space */
2136 start = block;
2137 end = block + num;
2138 } else if (le32_to_cpu(ex->ee_block) > block) {
2139 /* need to allocate space before found extent */
2140 start = block;
2141 end = le32_to_cpu(ex->ee_block);
2142 if (block + num < end)
2143 end = block + num;
2144 } else if (block >= le32_to_cpu(ex->ee_block)
2145 + ext4_ext_get_actual_len(ex)) {
2146 /* need to allocate space after found extent */
2147 start = block;
2148 end = block + num;
2149 if (end >= next)
2150 end = next;
2151 } else if (block >= le32_to_cpu(ex->ee_block)) {
2152 /*
2153 * some part of requested space is covered
2154 * by found extent
2155 */
2156 start = block;
2157 end = le32_to_cpu(ex->ee_block)
2158 + ext4_ext_get_actual_len(ex);
2159 if (block + num < end)
2160 end = block + num;
2161 exists = 1;
2162 } else {
2163 BUG();
2164 }
2165 BUG_ON(end <= start);
2166
2167 if (!exists) {
Zheng Liu69eb33d2013-02-18 00:31:07 -05002168 es.es_lblk = start;
2169 es.es_len = end - start;
2170 es.es_pblk = 0;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002171 } else {
Zheng Liu69eb33d2013-02-18 00:31:07 -05002172 es.es_lblk = le32_to_cpu(ex->ee_block);
2173 es.es_len = ext4_ext_get_actual_len(ex);
2174 es.es_pblk = ext4_ext_pblock(ex);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002175 if (ext4_ext_is_uninitialized(ex))
2176 flags |= FIEMAP_EXTENT_UNWRITTEN;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002177 }
2178
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002179 /*
Zheng Liu69eb33d2013-02-18 00:31:07 -05002180 * Find delayed extent and update es accordingly. We call
2181 * it even in !exists case to find out whether es is the
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002182 * last existing extent or not.
2183 */
Zheng Liu69eb33d2013-02-18 00:31:07 -05002184 next_del = ext4_find_delayed_extent(inode, &es);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002185 if (!exists && next_del) {
2186 exists = 1;
Jie Liu72dac952013-06-12 23:13:59 -04002187 flags |= (FIEMAP_EXTENT_DELALLOC |
2188 FIEMAP_EXTENT_UNKNOWN);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002189 }
2190 up_read(&EXT4_I(inode)->i_data_sem);
2191
Zheng Liu69eb33d2013-02-18 00:31:07 -05002192 if (unlikely(es.es_len == 0)) {
2193 EXT4_ERROR_INODE(inode, "es.es_len == 0");
Frank Mayhar273df552010-03-02 11:46:09 -05002194 err = -EIO;
2195 break;
2196 }
Eric Sandeen6873fa02008-10-07 00:46:36 -04002197
Zheng Liuf7fec032013-02-18 00:28:47 -05002198 /*
2199 * This is possible iff next == next_del == EXT_MAX_BLOCKS.
2200 * we need to check next == EXT_MAX_BLOCKS because it is
2201 * possible that an extent is with unwritten and delayed
2202 * status due to when an extent is delayed allocated and
2203 * is allocated by fallocate status tree will track both of
2204 * them in a extent.
2205 *
2206 * So we could return a unwritten and delayed extent, and
2207 * its block is equal to 'next'.
2208 */
2209 if (next == next_del && next == EXT_MAX_BLOCKS) {
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002210 flags |= FIEMAP_EXTENT_LAST;
2211 if (unlikely(next_del != EXT_MAX_BLOCKS ||
2212 next != EXT_MAX_BLOCKS)) {
2213 EXT4_ERROR_INODE(inode,
2214 "next extent == %u, next "
2215 "delalloc extent = %u",
2216 next, next_del);
2217 err = -EIO;
2218 break;
2219 }
Eric Sandeen6873fa02008-10-07 00:46:36 -04002220 }
2221
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002222 if (exists) {
2223 err = fiemap_fill_next_extent(fieinfo,
Zheng Liu69eb33d2013-02-18 00:31:07 -05002224 (__u64)es.es_lblk << blksize_bits,
2225 (__u64)es.es_pblk << blksize_bits,
2226 (__u64)es.es_len << blksize_bits,
Lukas Czerner91dd8c12012-11-28 12:32:26 -05002227 flags);
2228 if (err < 0)
2229 break;
2230 if (err == 1) {
2231 err = 0;
2232 break;
2233 }
Eric Sandeen6873fa02008-10-07 00:46:36 -04002234 }
2235
Zheng Liu69eb33d2013-02-18 00:31:07 -05002236 block = es.es_lblk + es.es_len;
Eric Sandeen6873fa02008-10-07 00:46:36 -04002237 }
2238
2239 if (path) {
2240 ext4_ext_drop_refs(path);
2241 kfree(path);
2242 }
2243
2244 return err;
2245}
2246
Alex Tomasa86c6182006-10-11 01:21:03 -07002247/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002248 * ext4_ext_put_gap_in_cache:
2249 * calculate boundaries of the gap that the requested block fits into
Alex Tomasa86c6182006-10-11 01:21:03 -07002250 * and cache this gap
2251 */
Avantika Mathur09b88252006-12-06 20:41:36 -08002252static void
Alex Tomasa86c6182006-10-11 01:21:03 -07002253ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002254 ext4_lblk_t block)
Alex Tomasa86c6182006-10-11 01:21:03 -07002255{
2256 int depth = ext_depth(inode);
Andi Shyti27b1b222013-08-28 14:00:00 -04002257 unsigned long len = 0;
2258 ext4_lblk_t lblock = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002259 struct ext4_extent *ex;
2260
2261 ex = path[depth].p_ext;
2262 if (ex == NULL) {
Zheng Liu69eb33d2013-02-18 00:31:07 -05002263 /*
2264 * there is no extent yet, so gap is [0;-] and we
2265 * don't cache it
2266 */
Alex Tomasa86c6182006-10-11 01:21:03 -07002267 ext_debug("cache gap(whole file):");
2268 } else if (block < le32_to_cpu(ex->ee_block)) {
2269 lblock = block;
2270 len = le32_to_cpu(ex->ee_block) - block;
Eric Sandeenbba90742008-01-28 23:58:27 -05002271 ext_debug("cache gap(before): %u [%u:%u]",
2272 block,
2273 le32_to_cpu(ex->ee_block),
2274 ext4_ext_get_actual_len(ex));
Zheng Liud100eef2013-02-18 00:29:59 -05002275 if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
2276 ext4_es_insert_extent(inode, lblock, len, ~0,
2277 EXTENT_STATUS_HOLE);
Alex Tomasa86c6182006-10-11 01:21:03 -07002278 } else if (block >= le32_to_cpu(ex->ee_block)
Amit Aroraa2df2a62007-07-17 21:42:41 -04002279 + ext4_ext_get_actual_len(ex)) {
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002280 ext4_lblk_t next;
Dave Kleikamp8c55e202007-05-24 13:04:54 -04002281 lblock = le32_to_cpu(ex->ee_block)
Amit Aroraa2df2a62007-07-17 21:42:41 -04002282 + ext4_ext_get_actual_len(ex);
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002283
2284 next = ext4_ext_next_allocated_block(path);
Eric Sandeenbba90742008-01-28 23:58:27 -05002285 ext_debug("cache gap(after): [%u:%u] %u",
2286 le32_to_cpu(ex->ee_block),
2287 ext4_ext_get_actual_len(ex),
2288 block);
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002289 BUG_ON(next == lblock);
2290 len = next - lblock;
Zheng Liud100eef2013-02-18 00:29:59 -05002291 if (!ext4_find_delalloc_range(inode, lblock, lblock + len - 1))
2292 ext4_es_insert_extent(inode, lblock, len, ~0,
2293 EXTENT_STATUS_HOLE);
Alex Tomasa86c6182006-10-11 01:21:03 -07002294 } else {
Alex Tomasa86c6182006-10-11 01:21:03 -07002295 BUG();
2296 }
2297
Eric Sandeenbba90742008-01-28 23:58:27 -05002298 ext_debug(" -> %u:%lu\n", lblock, len);
Alex Tomasa86c6182006-10-11 01:21:03 -07002299}
2300
2301/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002302 * ext4_ext_rm_idx:
2303 * removes index from the index block.
Alex Tomasa86c6182006-10-11 01:21:03 -07002304 */
Aneesh Kumar K.V1d03ec92008-01-28 23:58:27 -05002305static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
Forrest Liuc36575e2012-12-17 09:55:39 -05002306 struct ext4_ext_path *path, int depth)
Alex Tomasa86c6182006-10-11 01:21:03 -07002307{
Alex Tomasa86c6182006-10-11 01:21:03 -07002308 int err;
Alex Tomasf65e6fb2006-10-11 01:21:05 -07002309 ext4_fsblk_t leaf;
Alex Tomasa86c6182006-10-11 01:21:03 -07002310
2311 /* free index block */
Forrest Liuc36575e2012-12-17 09:55:39 -05002312 depth--;
2313 path = path + depth;
Theodore Ts'obf89d162010-10-27 21:30:14 -04002314 leaf = ext4_idx_pblock(path->p_idx);
Frank Mayhar273df552010-03-02 11:46:09 -05002315 if (unlikely(path->p_hdr->eh_entries == 0)) {
2316 EXT4_ERROR_INODE(inode, "path->p_hdr->eh_entries == 0");
2317 return -EIO;
2318 }
Avantika Mathur7e028972006-12-06 20:41:33 -08002319 err = ext4_ext_get_access(handle, inode, path);
2320 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07002321 return err;
Robin Dong0e1147b2011-07-27 21:29:33 -04002322
2323 if (path->p_idx != EXT_LAST_INDEX(path->p_hdr)) {
2324 int len = EXT_LAST_INDEX(path->p_hdr) - path->p_idx;
2325 len *= sizeof(struct ext4_extent_idx);
2326 memmove(path->p_idx, path->p_idx + 1, len);
2327 }
2328
Marcin Slusarze8546d02008-04-17 10:38:59 -04002329 le16_add_cpu(&path->p_hdr->eh_entries, -1);
Avantika Mathur7e028972006-12-06 20:41:33 -08002330 err = ext4_ext_dirty(handle, inode, path);
2331 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07002332 return err;
Mingming Cao2ae02102006-10-11 01:21:11 -07002333 ext_debug("index is empty, remove it, free block %llu\n", leaf);
Aditya Kalid8990242011-09-09 19:18:51 -04002334 trace_ext4_ext_rm_idx(inode, leaf);
2335
Peter Huewe7dc57612011-02-21 21:01:42 -05002336 ext4_free_blocks(handle, inode, NULL, leaf, 1,
Theodore Ts'oe6362602009-11-23 07:17:05 -05002337 EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET);
Forrest Liuc36575e2012-12-17 09:55:39 -05002338
2339 while (--depth >= 0) {
2340 if (path->p_idx != EXT_FIRST_INDEX(path->p_hdr))
2341 break;
2342 path--;
2343 err = ext4_ext_get_access(handle, inode, path);
2344 if (err)
2345 break;
2346 path->p_idx->ei_block = (path+1)->p_idx->ei_block;
2347 err = ext4_ext_dirty(handle, inode, path);
2348 if (err)
2349 break;
2350 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002351 return err;
2352}
2353
2354/*
Mingming Caoee12b632008-08-19 22:16:05 -04002355 * ext4_ext_calc_credits_for_single_extent:
2356 * This routine returns max. credits that needed to insert an extent
2357 * to the extent tree.
2358 * When pass the actual path, the caller should calculate credits
2359 * under i_data_sem.
Alex Tomasa86c6182006-10-11 01:21:03 -07002360 */
Mingming Cao525f4ed2008-08-19 22:15:58 -04002361int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
Alex Tomasa86c6182006-10-11 01:21:03 -07002362 struct ext4_ext_path *path)
2363{
Alex Tomasa86c6182006-10-11 01:21:03 -07002364 if (path) {
Mingming Caoee12b632008-08-19 22:16:05 -04002365 int depth = ext_depth(inode);
Mingming Caof3bd1f32008-08-19 22:16:03 -04002366 int ret = 0;
Mingming Caoee12b632008-08-19 22:16:05 -04002367
Alex Tomasa86c6182006-10-11 01:21:03 -07002368 /* probably there is space in leaf? */
Alex Tomasa86c6182006-10-11 01:21:03 -07002369 if (le16_to_cpu(path[depth].p_hdr->eh_entries)
Mingming Caoee12b632008-08-19 22:16:05 -04002370 < le16_to_cpu(path[depth].p_hdr->eh_max)) {
2371
2372 /*
2373 * There are some space in the leaf tree, no
2374 * need to account for leaf block credit
2375 *
2376 * bitmaps and block group descriptor blocks
Tao Madf3ab172011-10-08 15:53:49 -04002377 * and other metadata blocks still need to be
Mingming Caoee12b632008-08-19 22:16:05 -04002378 * accounted.
2379 */
Mingming Cao525f4ed2008-08-19 22:15:58 -04002380 /* 1 bitmap, 1 block group descriptor */
Mingming Caoee12b632008-08-19 22:16:05 -04002381 ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
Aneesh Kumar K.V5887e982009-07-05 23:12:04 -04002382 return ret;
Mingming Caoee12b632008-08-19 22:16:05 -04002383 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002384 }
2385
Mingming Cao525f4ed2008-08-19 22:15:58 -04002386 return ext4_chunk_trans_blocks(inode, nrblocks);
Mingming Caoee12b632008-08-19 22:16:05 -04002387}
Alex Tomasa86c6182006-10-11 01:21:03 -07002388
Mingming Caoee12b632008-08-19 22:16:05 -04002389/*
Jan Karafffb2732013-06-04 13:01:11 -04002390 * How many index/leaf blocks need to change/allocate to add @extents extents?
Mingming Caoee12b632008-08-19 22:16:05 -04002391 *
Jan Karafffb2732013-06-04 13:01:11 -04002392 * If we add a single extent, then in the worse case, each tree level
2393 * index/leaf need to be changed in case of the tree split.
Mingming Caoee12b632008-08-19 22:16:05 -04002394 *
Jan Karafffb2732013-06-04 13:01:11 -04002395 * If more extents are inserted, they could cause the whole tree split more
2396 * than once, but this is really rare.
Mingming Caoee12b632008-08-19 22:16:05 -04002397 */
Jan Karafffb2732013-06-04 13:01:11 -04002398int ext4_ext_index_trans_blocks(struct inode *inode, int extents)
Mingming Caoee12b632008-08-19 22:16:05 -04002399{
2400 int index;
Tao Maf19d5872012-12-10 14:05:51 -05002401 int depth;
2402
2403 /* If we are converting the inline data, only one is needed here. */
2404 if (ext4_has_inline_data(inode))
2405 return 1;
2406
2407 depth = ext_depth(inode);
Alex Tomasa86c6182006-10-11 01:21:03 -07002408
Jan Karafffb2732013-06-04 13:01:11 -04002409 if (extents <= 1)
Mingming Caoee12b632008-08-19 22:16:05 -04002410 index = depth * 2;
2411 else
2412 index = depth * 3;
Alex Tomasa86c6182006-10-11 01:21:03 -07002413
Mingming Caoee12b632008-08-19 22:16:05 -04002414 return index;
Alex Tomasa86c6182006-10-11 01:21:03 -07002415}
2416
Theodore Ts'o981250c2013-06-12 11:48:29 -04002417static inline int get_default_free_blocks_flags(struct inode *inode)
2418{
2419 if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
2420 return EXT4_FREE_BLOCKS_METADATA | EXT4_FREE_BLOCKS_FORGET;
2421 else if (ext4_should_journal_data(inode))
2422 return EXT4_FREE_BLOCKS_FORGET;
2423 return 0;
2424}
2425
Alex Tomasa86c6182006-10-11 01:21:03 -07002426static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002427 struct ext4_extent *ex,
Lukas Czernerd23142c2013-05-27 23:33:35 -04002428 long long *partial_cluster,
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002429 ext4_lblk_t from, ext4_lblk_t to)
Alex Tomasa86c6182006-10-11 01:21:03 -07002430{
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002431 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
Amit Aroraa2df2a62007-07-17 21:42:41 -04002432 unsigned short ee_len = ext4_ext_get_actual_len(ex);
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002433 ext4_fsblk_t pblk;
Theodore Ts'o981250c2013-06-12 11:48:29 -04002434 int flags = get_default_free_blocks_flags(inode);
Andrey Sidorov18888cf2012-09-19 14:14:53 -04002435
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002436 /*
2437 * For bigalloc file systems, we never free a partial cluster
2438 * at the beginning of the extent. Instead, we make a note
2439 * that we tried freeing the cluster, and check to see if we
2440 * need to free it on a subsequent call to ext4_remove_blocks,
2441 * or at the end of the ext4_truncate() operation.
2442 */
2443 flags |= EXT4_FREE_BLOCKS_NOFREE_FIRST_CLUSTER;
2444
Aditya Kalid8990242011-09-09 19:18:51 -04002445 trace_ext4_remove_blocks(inode, ex, from, to, *partial_cluster);
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002446 /*
2447 * If we have a partial cluster, and it's different from the
2448 * cluster of the last block, we need to explicitly free the
2449 * partial cluster here.
2450 */
2451 pblk = ext4_ext_pblock(ex) + ee_len - 1;
Lukas Czernerd23142c2013-05-27 23:33:35 -04002452 if ((*partial_cluster > 0) &&
2453 (EXT4_B2C(sbi, pblk) != *partial_cluster)) {
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002454 ext4_free_blocks(handle, inode, NULL,
2455 EXT4_C2B(sbi, *partial_cluster),
2456 sbi->s_cluster_ratio, flags);
2457 *partial_cluster = 0;
2458 }
2459
Alex Tomasa86c6182006-10-11 01:21:03 -07002460#ifdef EXTENTS_STATS
2461 {
2462 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
Alex Tomasa86c6182006-10-11 01:21:03 -07002463 spin_lock(&sbi->s_ext_stats_lock);
2464 sbi->s_ext_blocks += ee_len;
2465 sbi->s_ext_extents++;
2466 if (ee_len < sbi->s_ext_min)
2467 sbi->s_ext_min = ee_len;
2468 if (ee_len > sbi->s_ext_max)
2469 sbi->s_ext_max = ee_len;
2470 if (ext_depth(inode) > sbi->s_depth_max)
2471 sbi->s_depth_max = ext_depth(inode);
2472 spin_unlock(&sbi->s_ext_stats_lock);
2473 }
2474#endif
2475 if (from >= le32_to_cpu(ex->ee_block)
Amit Aroraa2df2a62007-07-17 21:42:41 -04002476 && to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
Alex Tomasa86c6182006-10-11 01:21:03 -07002477 /* tail removal */
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002478 ext4_lblk_t num;
Lukas Czernerd23142c2013-05-27 23:33:35 -04002479 unsigned int unaligned;
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002480
Amit Aroraa2df2a62007-07-17 21:42:41 -04002481 num = le32_to_cpu(ex->ee_block) + ee_len - from;
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002482 pblk = ext4_ext_pblock(ex) + ee_len - num;
Lukas Czernerd23142c2013-05-27 23:33:35 -04002483 /*
2484 * Usually we want to free partial cluster at the end of the
2485 * extent, except for the situation when the cluster is still
2486 * used by any other extent (partial_cluster is negative).
2487 */
2488 if (*partial_cluster < 0 &&
2489 -(*partial_cluster) == EXT4_B2C(sbi, pblk + num - 1))
2490 flags |= EXT4_FREE_BLOCKS_NOFREE_LAST_CLUSTER;
2491
2492 ext_debug("free last %u blocks starting %llu partial %lld\n",
2493 num, pblk, *partial_cluster);
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002494 ext4_free_blocks(handle, inode, NULL, pblk, num, flags);
2495 /*
2496 * If the block range to be freed didn't start at the
2497 * beginning of a cluster, and we removed the entire
Lukas Czernerd23142c2013-05-27 23:33:35 -04002498 * extent and the cluster is not used by any other extent,
2499 * save the partial cluster here, since we might need to
2500 * delete if we determine that the truncate operation has
2501 * removed all of the blocks in the cluster.
2502 *
2503 * On the other hand, if we did not manage to free the whole
2504 * extent, we have to mark the cluster as used (store negative
2505 * cluster number in partial_cluster).
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002506 */
Lukas Czernerd23142c2013-05-27 23:33:35 -04002507 unaligned = pblk & (sbi->s_cluster_ratio - 1);
2508 if (unaligned && (ee_len == num) &&
2509 (*partial_cluster != -((long long)EXT4_B2C(sbi, pblk))))
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002510 *partial_cluster = EXT4_B2C(sbi, pblk);
Lukas Czernerd23142c2013-05-27 23:33:35 -04002511 else if (unaligned)
2512 *partial_cluster = -((long long)EXT4_B2C(sbi, pblk));
2513 else if (*partial_cluster > 0)
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002514 *partial_cluster = 0;
Lukas Czerner78fb9cd2013-05-27 23:32:35 -04002515 } else
2516 ext4_error(sbi->s_sb, "strange request: removal(2) "
2517 "%u-%u from %u:%u\n",
2518 from, to, le32_to_cpu(ex->ee_block), ee_len);
Alex Tomasa86c6182006-10-11 01:21:03 -07002519 return 0;
2520}
2521
Allison Hendersond583fb82011-05-25 07:41:43 -04002522
2523/*
2524 * ext4_ext_rm_leaf() Removes the extents associated with the
2525 * blocks appearing between "start" and "end", and splits the extents
2526 * if "start" and "end" appear in the same extent
2527 *
2528 * @handle: The journal handle
2529 * @inode: The files inode
2530 * @path: The path to the leaf
Lukas Czernerd23142c2013-05-27 23:33:35 -04002531 * @partial_cluster: The cluster which we'll have to free if all extents
2532 * has been released from it. It gets negative in case
2533 * that the cluster is still used.
Allison Hendersond583fb82011-05-25 07:41:43 -04002534 * @start: The first block to remove
2535 * @end: The last block to remove
2536 */
Alex Tomasa86c6182006-10-11 01:21:03 -07002537static int
2538ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
Lukas Czernerd23142c2013-05-27 23:33:35 -04002539 struct ext4_ext_path *path,
2540 long long *partial_cluster,
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002541 ext4_lblk_t start, ext4_lblk_t end)
Alex Tomasa86c6182006-10-11 01:21:03 -07002542{
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002543 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
Alex Tomasa86c6182006-10-11 01:21:03 -07002544 int err = 0, correct_index = 0;
2545 int depth = ext_depth(inode), credits;
2546 struct ext4_extent_header *eh;
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002547 ext4_lblk_t a, b;
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05002548 unsigned num;
2549 ext4_lblk_t ex_ee_block;
Alex Tomasa86c6182006-10-11 01:21:03 -07002550 unsigned short ex_ee_len;
Amit Aroraa2df2a62007-07-17 21:42:41 -04002551 unsigned uninitialized = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002552 struct ext4_extent *ex;
Lukas Czernerd23142c2013-05-27 23:33:35 -04002553 ext4_fsblk_t pblk;
Alex Tomasa86c6182006-10-11 01:21:03 -07002554
Alex Tomasc29c0ae2007-07-18 09:19:09 -04002555 /* the header must be checked already in ext4_ext_remove_space() */
Lukas Czerner5f95d212012-03-19 23:03:19 -04002556 ext_debug("truncate since %u in leaf to %u\n", start, end);
Alex Tomasa86c6182006-10-11 01:21:03 -07002557 if (!path[depth].p_hdr)
2558 path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
2559 eh = path[depth].p_hdr;
Frank Mayhar273df552010-03-02 11:46:09 -05002560 if (unlikely(path[depth].p_hdr == NULL)) {
2561 EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
2562 return -EIO;
2563 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002564 /* find where to start removing */
Ashish Sangwan6ae06ff2013-07-01 08:12:41 -04002565 ex = path[depth].p_ext;
2566 if (!ex)
2567 ex = EXT_LAST_EXTENT(eh);
Alex Tomasa86c6182006-10-11 01:21:03 -07002568
2569 ex_ee_block = le32_to_cpu(ex->ee_block);
Amit Aroraa2df2a62007-07-17 21:42:41 -04002570 ex_ee_len = ext4_ext_get_actual_len(ex);
Alex Tomasa86c6182006-10-11 01:21:03 -07002571
Aditya Kalid8990242011-09-09 19:18:51 -04002572 trace_ext4_ext_rm_leaf(inode, start, ex, *partial_cluster);
2573
Alex Tomasa86c6182006-10-11 01:21:03 -07002574 while (ex >= EXT_FIRST_EXTENT(eh) &&
2575 ex_ee_block + ex_ee_len > start) {
Aneesh Kumar K.Va41f2072009-06-10 14:22:55 -04002576
2577 if (ext4_ext_is_uninitialized(ex))
2578 uninitialized = 1;
2579 else
2580 uninitialized = 0;
2581
Mingming553f9002009-09-18 13:34:55 -04002582 ext_debug("remove ext %u:[%d]%d\n", ex_ee_block,
2583 uninitialized, ex_ee_len);
Alex Tomasa86c6182006-10-11 01:21:03 -07002584 path[depth].p_ext = ex;
2585
2586 a = ex_ee_block > start ? ex_ee_block : start;
Allison Hendersond583fb82011-05-25 07:41:43 -04002587 b = ex_ee_block+ex_ee_len - 1 < end ?
2588 ex_ee_block+ex_ee_len - 1 : end;
Alex Tomasa86c6182006-10-11 01:21:03 -07002589
2590 ext_debug(" border %u:%u\n", a, b);
2591
Allison Hendersond583fb82011-05-25 07:41:43 -04002592 /* If this extent is beyond the end of the hole, skip it */
Lukas Czerner5f95d212012-03-19 23:03:19 -04002593 if (end < ex_ee_block) {
Lukas Czernerd23142c2013-05-27 23:33:35 -04002594 /*
2595 * We're going to skip this extent and move to another,
2596 * so if this extent is not cluster aligned we have
2597 * to mark the current cluster as used to avoid
2598 * accidentally freeing it later on
2599 */
2600 pblk = ext4_ext_pblock(ex);
2601 if (pblk & (sbi->s_cluster_ratio - 1))
2602 *partial_cluster =
2603 -((long long)EXT4_B2C(sbi, pblk));
Allison Hendersond583fb82011-05-25 07:41:43 -04002604 ex--;
2605 ex_ee_block = le32_to_cpu(ex->ee_block);
2606 ex_ee_len = ext4_ext_get_actual_len(ex);
2607 continue;
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002608 } else if (b != ex_ee_block + ex_ee_len - 1) {
Lukas Czernerdc1841d2012-03-19 23:07:43 -04002609 EXT4_ERROR_INODE(inode,
2610 "can not handle truncate %u:%u "
2611 "on extent %u:%u",
2612 start, end, ex_ee_block,
2613 ex_ee_block + ex_ee_len - 1);
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002614 err = -EIO;
2615 goto out;
Alex Tomasa86c6182006-10-11 01:21:03 -07002616 } else if (a != ex_ee_block) {
2617 /* remove tail of the extent */
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002618 num = a - ex_ee_block;
Alex Tomasa86c6182006-10-11 01:21:03 -07002619 } else {
2620 /* remove whole extent: excellent! */
Alex Tomasa86c6182006-10-11 01:21:03 -07002621 num = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002622 }
Theodore Ts'o34071da2008-08-01 21:59:19 -04002623 /*
2624 * 3 for leaf, sb, and inode plus 2 (bmap and group
2625 * descriptor) for each block group; assume two block
2626 * groups plus ex_ee_len/blocks_per_block_group for
2627 * the worst case
2628 */
2629 credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
Alex Tomasa86c6182006-10-11 01:21:03 -07002630 if (ex == EXT_FIRST_EXTENT(eh)) {
2631 correct_index = 1;
2632 credits += (ext_depth(inode)) + 1;
2633 }
Dmitry Monakhov5aca07e2009-12-08 22:42:15 -05002634 credits += EXT4_MAXQUOTAS_TRANS_BLOCKS(inode->i_sb);
Alex Tomasa86c6182006-10-11 01:21:03 -07002635
Jan Kara487caee2009-08-17 22:17:20 -04002636 err = ext4_ext_truncate_extend_restart(handle, inode, credits);
Shen Feng9102e4f2008-07-11 19:27:31 -04002637 if (err)
Alex Tomasa86c6182006-10-11 01:21:03 -07002638 goto out;
Alex Tomasa86c6182006-10-11 01:21:03 -07002639
2640 err = ext4_ext_get_access(handle, inode, path + depth);
2641 if (err)
2642 goto out;
2643
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002644 err = ext4_remove_blocks(handle, inode, ex, partial_cluster,
2645 a, b);
Alex Tomasa86c6182006-10-11 01:21:03 -07002646 if (err)
2647 goto out;
2648
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002649 if (num == 0)
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002650 /* this extent is removed; mark slot entirely unused */
Alex Tomasf65e6fb2006-10-11 01:21:05 -07002651 ext4_ext_store_pblock(ex, 0);
Alex Tomasa86c6182006-10-11 01:21:03 -07002652
Alex Tomasa86c6182006-10-11 01:21:03 -07002653 ex->ee_len = cpu_to_le16(num);
Amit Arora749269f2007-07-18 09:02:56 -04002654 /*
2655 * Do not mark uninitialized if all the blocks in the
2656 * extent have been removed.
2657 */
2658 if (uninitialized && num)
Amit Aroraa2df2a62007-07-17 21:42:41 -04002659 ext4_ext_mark_uninitialized(ex);
Allison Hendersond583fb82011-05-25 07:41:43 -04002660 /*
2661 * If the extent was completely released,
2662 * we need to remove it from the leaf
2663 */
2664 if (num == 0) {
Lukas Czernerf17722f2011-06-06 00:05:17 -04002665 if (end != EXT_MAX_BLOCKS - 1) {
Allison Hendersond583fb82011-05-25 07:41:43 -04002666 /*
2667 * For hole punching, we need to scoot all the
2668 * extents up when an extent is removed so that
2669 * we dont have blank extents in the middle
2670 */
2671 memmove(ex, ex+1, (EXT_LAST_EXTENT(eh) - ex) *
2672 sizeof(struct ext4_extent));
2673
2674 /* Now get rid of the one at the end */
2675 memset(EXT_LAST_EXTENT(eh), 0,
2676 sizeof(struct ext4_extent));
2677 }
2678 le16_add_cpu(&eh->eh_entries, -1);
Lukas Czernerd23142c2013-05-27 23:33:35 -04002679 } else if (*partial_cluster > 0)
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002680 *partial_cluster = 0;
Allison Hendersond583fb82011-05-25 07:41:43 -04002681
Dmitry Monakhov750c9c42011-10-25 05:35:05 -04002682 err = ext4_ext_dirty(handle, inode, path + depth);
2683 if (err)
2684 goto out;
2685
Yongqiang Yangbf52c6f2011-11-01 18:59:26 -04002686 ext_debug("new extent: %u:%u:%llu\n", ex_ee_block, num,
Theodore Ts'obf89d162010-10-27 21:30:14 -04002687 ext4_ext_pblock(ex));
Alex Tomasa86c6182006-10-11 01:21:03 -07002688 ex--;
2689 ex_ee_block = le32_to_cpu(ex->ee_block);
Amit Aroraa2df2a62007-07-17 21:42:41 -04002690 ex_ee_len = ext4_ext_get_actual_len(ex);
Alex Tomasa86c6182006-10-11 01:21:03 -07002691 }
2692
2693 if (correct_index && eh->eh_entries)
2694 err = ext4_ext_correct_indexes(handle, inode, path);
2695
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002696 /*
Lukas Czernerd23142c2013-05-27 23:33:35 -04002697 * Free the partial cluster only if the current extent does not
2698 * reference it. Otherwise we might free used cluster.
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002699 */
Lukas Czernerd23142c2013-05-27 23:33:35 -04002700 if (*partial_cluster > 0 &&
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002701 (EXT4_B2C(sbi, ext4_ext_pblock(ex) + ex_ee_len - 1) !=
2702 *partial_cluster)) {
Theodore Ts'o981250c2013-06-12 11:48:29 -04002703 int flags = get_default_free_blocks_flags(inode);
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002704
2705 ext4_free_blocks(handle, inode, NULL,
2706 EXT4_C2B(sbi, *partial_cluster),
2707 sbi->s_cluster_ratio, flags);
2708 *partial_cluster = 0;
2709 }
2710
Alex Tomasa86c6182006-10-11 01:21:03 -07002711 /* if this leaf is free, then we should
2712 * remove it from index block above */
2713 if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
Forrest Liuc36575e2012-12-17 09:55:39 -05002714 err = ext4_ext_rm_idx(handle, inode, path, depth);
Alex Tomasa86c6182006-10-11 01:21:03 -07002715
2716out:
2717 return err;
2718}
2719
2720/*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002721 * ext4_ext_more_to_rm:
2722 * returns 1 if current index has to be freed (even partial)
Alex Tomasa86c6182006-10-11 01:21:03 -07002723 */
Avantika Mathur09b88252006-12-06 20:41:36 -08002724static int
Alex Tomasa86c6182006-10-11 01:21:03 -07002725ext4_ext_more_to_rm(struct ext4_ext_path *path)
2726{
2727 BUG_ON(path->p_idx == NULL);
2728
2729 if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
2730 return 0;
2731
2732 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002733 * if truncate on deeper level happened, it wasn't partial,
Alex Tomasa86c6182006-10-11 01:21:03 -07002734 * so we have to consider current index for truncation
2735 */
2736 if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
2737 return 0;
2738 return 1;
2739}
2740
Theodore Ts'o26a4c0c2013-04-03 12:45:17 -04002741int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start,
2742 ext4_lblk_t end)
Alex Tomasa86c6182006-10-11 01:21:03 -07002743{
2744 struct super_block *sb = inode->i_sb;
2745 int depth = ext_depth(inode);
Ashish Sangwan968dee72012-07-22 22:49:08 -04002746 struct ext4_ext_path *path = NULL;
Lukas Czernerd23142c2013-05-27 23:33:35 -04002747 long long partial_cluster = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002748 handle_t *handle;
Dmitry Monakhov6f2080e2012-09-30 23:03:50 -04002749 int i = 0, err = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002750
Lukas Czerner5f95d212012-03-19 23:03:19 -04002751 ext_debug("truncate since %u to %u\n", start, end);
Alex Tomasa86c6182006-10-11 01:21:03 -07002752
2753 /* probably first extent we're gonna free will be last in block */
Theodore Ts'o9924a922013-02-08 21:59:22 -05002754 handle = ext4_journal_start(inode, EXT4_HT_TRUNCATE, depth + 1);
Alex Tomasa86c6182006-10-11 01:21:03 -07002755 if (IS_ERR(handle))
2756 return PTR_ERR(handle);
2757
Dmitry Monakhov0617b832010-05-17 01:00:00 -04002758again:
Lukas Czerner61801322013-05-27 23:32:35 -04002759 trace_ext4_ext_remove_space(inode, start, end, depth);
Aditya Kalid8990242011-09-09 19:18:51 -04002760
Alex Tomasa86c6182006-10-11 01:21:03 -07002761 /*
Lukas Czerner5f95d212012-03-19 23:03:19 -04002762 * Check if we are removing extents inside the extent tree. If that
2763 * is the case, we are going to punch a hole inside the extent tree
2764 * so we have to check whether we need to split the extent covering
2765 * the last block to remove so we can easily remove the part of it
2766 * in ext4_ext_rm_leaf().
2767 */
2768 if (end < EXT_MAX_BLOCKS - 1) {
2769 struct ext4_extent *ex;
2770 ext4_lblk_t ee_block;
2771
2772 /* find extent for this block */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002773 path = ext4_ext_find_extent(inode, end, NULL, EXT4_EX_NOCACHE);
Lukas Czerner5f95d212012-03-19 23:03:19 -04002774 if (IS_ERR(path)) {
2775 ext4_journal_stop(handle);
2776 return PTR_ERR(path);
2777 }
2778 depth = ext_depth(inode);
Dmitry Monakhov6f2080e2012-09-30 23:03:50 -04002779 /* Leaf not may not exist only if inode has no blocks at all */
Lukas Czerner5f95d212012-03-19 23:03:19 -04002780 ex = path[depth].p_ext;
Ashish Sangwan968dee72012-07-22 22:49:08 -04002781 if (!ex) {
Dmitry Monakhov6f2080e2012-09-30 23:03:50 -04002782 if (depth) {
2783 EXT4_ERROR_INODE(inode,
2784 "path[%d].p_hdr == NULL",
2785 depth);
2786 err = -EIO;
2787 }
2788 goto out;
Ashish Sangwan968dee72012-07-22 22:49:08 -04002789 }
Lukas Czerner5f95d212012-03-19 23:03:19 -04002790
2791 ee_block = le32_to_cpu(ex->ee_block);
2792
2793 /*
2794 * See if the last block is inside the extent, if so split
2795 * the extent at 'end' block so we can easily remove the
2796 * tail of the first part of the split extent in
2797 * ext4_ext_rm_leaf().
2798 */
2799 if (end >= ee_block &&
2800 end < ee_block + ext4_ext_get_actual_len(ex) - 1) {
2801 int split_flag = 0;
2802
2803 if (ext4_ext_is_uninitialized(ex))
2804 split_flag = EXT4_EXT_MARK_UNINIT1 |
2805 EXT4_EXT_MARK_UNINIT2;
2806
2807 /*
2808 * Split the extent in two so that 'end' is the last
Lukas Czerner27dd4382013-04-09 22:11:22 -04002809 * block in the first new extent. Also we should not
2810 * fail removing space due to ENOSPC so try to use
2811 * reserved block if that happens.
Lukas Czerner5f95d212012-03-19 23:03:19 -04002812 */
2813 err = ext4_split_extent_at(handle, inode, path,
Lukas Czerner27dd4382013-04-09 22:11:22 -04002814 end + 1, split_flag,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002815 EXT4_EX_NOCACHE |
Lukas Czerner27dd4382013-04-09 22:11:22 -04002816 EXT4_GET_BLOCKS_PRE_IO |
2817 EXT4_GET_BLOCKS_METADATA_NOFAIL);
Lukas Czerner5f95d212012-03-19 23:03:19 -04002818
2819 if (err < 0)
2820 goto out;
2821 }
Lukas Czerner5f95d212012-03-19 23:03:19 -04002822 }
Lukas Czerner5f95d212012-03-19 23:03:19 -04002823 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002824 * We start scanning from right side, freeing all the blocks
2825 * after i_size and walking into the tree depth-wise.
Alex Tomasa86c6182006-10-11 01:21:03 -07002826 */
Dmitry Monakhov0617b832010-05-17 01:00:00 -04002827 depth = ext_depth(inode);
Ashish Sangwan968dee72012-07-22 22:49:08 -04002828 if (path) {
2829 int k = i = depth;
2830 while (--k > 0)
2831 path[k].p_block =
2832 le16_to_cpu(path[k].p_hdr->eh_entries)+1;
2833 } else {
2834 path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1),
2835 GFP_NOFS);
2836 if (path == NULL) {
2837 ext4_journal_stop(handle);
2838 return -ENOMEM;
2839 }
2840 path[0].p_depth = depth;
2841 path[0].p_hdr = ext_inode_hdr(inode);
Theodore Ts'o89a4e482012-08-17 08:54:52 -04002842 i = 0;
Lukas Czerner5f95d212012-03-19 23:03:19 -04002843
Theodore Ts'oc3491792013-08-16 21:21:41 -04002844 if (ext4_ext_check(inode, path[0].p_hdr, depth, 0)) {
Ashish Sangwan968dee72012-07-22 22:49:08 -04002845 err = -EIO;
2846 goto out;
2847 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002848 }
Ashish Sangwan968dee72012-07-22 22:49:08 -04002849 err = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07002850
2851 while (i >= 0 && err == 0) {
2852 if (i == depth) {
2853 /* this is leaf block */
Allison Hendersond583fb82011-05-25 07:41:43 -04002854 err = ext4_ext_rm_leaf(handle, inode, path,
Theodore Ts'o0aa06002011-09-09 18:54:51 -04002855 &partial_cluster, start,
Lukas Czerner5f95d212012-03-19 23:03:19 -04002856 end);
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002857 /* root level has p_bh == NULL, brelse() eats this */
Alex Tomasa86c6182006-10-11 01:21:03 -07002858 brelse(path[i].p_bh);
2859 path[i].p_bh = NULL;
2860 i--;
2861 continue;
2862 }
2863
2864 /* this is index block */
2865 if (!path[i].p_hdr) {
2866 ext_debug("initialize header\n");
2867 path[i].p_hdr = ext_block_hdr(path[i].p_bh);
Alex Tomasa86c6182006-10-11 01:21:03 -07002868 }
2869
Alex Tomasa86c6182006-10-11 01:21:03 -07002870 if (!path[i].p_idx) {
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002871 /* this level hasn't been touched yet */
Alex Tomasa86c6182006-10-11 01:21:03 -07002872 path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
2873 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
2874 ext_debug("init index ptr: hdr 0x%p, num %d\n",
2875 path[i].p_hdr,
2876 le16_to_cpu(path[i].p_hdr->eh_entries));
2877 } else {
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002878 /* we were already here, see at next index */
Alex Tomasa86c6182006-10-11 01:21:03 -07002879 path[i].p_idx--;
2880 }
2881
2882 ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
2883 i, EXT_FIRST_INDEX(path[i].p_hdr),
2884 path[i].p_idx);
2885 if (ext4_ext_more_to_rm(path + i)) {
Alex Tomasc29c0ae2007-07-18 09:19:09 -04002886 struct buffer_head *bh;
Alex Tomasa86c6182006-10-11 01:21:03 -07002887 /* go to the next level */
Mingming Cao2ae02102006-10-11 01:21:11 -07002888 ext_debug("move to level %d (block %llu)\n",
Theodore Ts'obf89d162010-10-27 21:30:14 -04002889 i + 1, ext4_idx_pblock(path[i].p_idx));
Alex Tomasa86c6182006-10-11 01:21:03 -07002890 memset(path + i + 1, 0, sizeof(*path));
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04002891 bh = read_extent_tree_block(inode,
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04002892 ext4_idx_pblock(path[i].p_idx), depth - i - 1,
2893 EXT4_EX_NOCACHE);
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04002894 if (IS_ERR(bh)) {
Alex Tomasa86c6182006-10-11 01:21:03 -07002895 /* should we reset i_size? */
Theodore Ts'o7d7ea892013-08-16 21:20:41 -04002896 err = PTR_ERR(bh);
Alex Tomasa86c6182006-10-11 01:21:03 -07002897 break;
2898 }
Theodore Ts'o76828c82013-07-15 12:27:47 -04002899 /* Yield here to deal with large extent trees.
2900 * Should be a no-op if we did IO above. */
2901 cond_resched();
Alex Tomasc29c0ae2007-07-18 09:19:09 -04002902 if (WARN_ON(i + 1 > depth)) {
2903 err = -EIO;
2904 break;
2905 }
Alex Tomasc29c0ae2007-07-18 09:19:09 -04002906 path[i + 1].p_bh = bh;
Alex Tomasa86c6182006-10-11 01:21:03 -07002907
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002908 /* save actual number of indexes since this
2909 * number is changed at the next iteration */
Alex Tomasa86c6182006-10-11 01:21:03 -07002910 path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
2911 i++;
2912 } else {
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002913 /* we finished processing this index, go up */
Alex Tomasa86c6182006-10-11 01:21:03 -07002914 if (path[i].p_hdr->eh_entries == 0 && i > 0) {
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002915 /* index is empty, remove it;
Alex Tomasa86c6182006-10-11 01:21:03 -07002916 * handle must be already prepared by the
2917 * truncatei_leaf() */
Forrest Liuc36575e2012-12-17 09:55:39 -05002918 err = ext4_ext_rm_idx(handle, inode, path, i);
Alex Tomasa86c6182006-10-11 01:21:03 -07002919 }
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002920 /* root level has p_bh == NULL, brelse() eats this */
Alex Tomasa86c6182006-10-11 01:21:03 -07002921 brelse(path[i].p_bh);
2922 path[i].p_bh = NULL;
2923 i--;
2924 ext_debug("return to level %d\n", i);
2925 }
2926 }
2927
Lukas Czerner61801322013-05-27 23:32:35 -04002928 trace_ext4_ext_remove_space_done(inode, start, end, depth,
2929 partial_cluster, path->p_hdr->eh_entries);
Aditya Kalid8990242011-09-09 19:18:51 -04002930
Aditya Kali7b415bf2011-09-09 19:04:51 -04002931 /* If we still have something in the partial cluster and we have removed
2932 * even the first extent, then we should free the blocks in the partial
2933 * cluster as well. */
Lukas Czernerd23142c2013-05-27 23:33:35 -04002934 if (partial_cluster > 0 && path->p_hdr->eh_entries == 0) {
Theodore Ts'o981250c2013-06-12 11:48:29 -04002935 int flags = get_default_free_blocks_flags(inode);
Aditya Kali7b415bf2011-09-09 19:04:51 -04002936
2937 ext4_free_blocks(handle, inode, NULL,
2938 EXT4_C2B(EXT4_SB(sb), partial_cluster),
2939 EXT4_SB(sb)->s_cluster_ratio, flags);
2940 partial_cluster = 0;
2941 }
2942
Alex Tomasa86c6182006-10-11 01:21:03 -07002943 /* TODO: flexible tree reduction should be here */
2944 if (path->p_hdr->eh_entries == 0) {
2945 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07002946 * truncate to zero freed all the tree,
2947 * so we need to correct eh_depth
Alex Tomasa86c6182006-10-11 01:21:03 -07002948 */
2949 err = ext4_ext_get_access(handle, inode, path);
2950 if (err == 0) {
2951 ext_inode_hdr(inode)->eh_depth = 0;
2952 ext_inode_hdr(inode)->eh_max =
Theodore Ts'o55ad63b2009-08-28 10:40:33 -04002953 cpu_to_le16(ext4_ext_space_root(inode, 0));
Alex Tomasa86c6182006-10-11 01:21:03 -07002954 err = ext4_ext_dirty(handle, inode, path);
2955 }
2956 }
2957out:
Alex Tomasa86c6182006-10-11 01:21:03 -07002958 ext4_ext_drop_refs(path);
2959 kfree(path);
Ashish Sangwan968dee72012-07-22 22:49:08 -04002960 if (err == -EAGAIN) {
2961 path = NULL;
Dmitry Monakhov0617b832010-05-17 01:00:00 -04002962 goto again;
Ashish Sangwan968dee72012-07-22 22:49:08 -04002963 }
Alex Tomasa86c6182006-10-11 01:21:03 -07002964 ext4_journal_stop(handle);
2965
2966 return err;
2967}
2968
2969/*
2970 * called at mount time
2971 */
2972void ext4_ext_init(struct super_block *sb)
2973{
2974 /*
2975 * possible initialization would be here
2976 */
2977
Theodore Ts'o83982b62009-01-06 14:53:16 -05002978 if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS)) {
Theodore Ts'o90576c02009-09-29 15:51:30 -04002979#if defined(AGGRESSIVE_TEST) || defined(CHECK_BINSEARCH) || defined(EXTENTS_STATS)
Theodore Ts'o92b97812012-03-19 23:41:49 -04002980 printk(KERN_INFO "EXT4-fs: file extents enabled"
Robert P. J. Daybbf2f9f2007-02-17 19:20:16 +01002981#ifdef AGGRESSIVE_TEST
Theodore Ts'o92b97812012-03-19 23:41:49 -04002982 ", aggressive tests"
Alex Tomasa86c6182006-10-11 01:21:03 -07002983#endif
2984#ifdef CHECK_BINSEARCH
Theodore Ts'o92b97812012-03-19 23:41:49 -04002985 ", check binsearch"
Alex Tomasa86c6182006-10-11 01:21:03 -07002986#endif
2987#ifdef EXTENTS_STATS
Theodore Ts'o92b97812012-03-19 23:41:49 -04002988 ", stats"
Alex Tomasa86c6182006-10-11 01:21:03 -07002989#endif
Theodore Ts'o92b97812012-03-19 23:41:49 -04002990 "\n");
Theodore Ts'o90576c02009-09-29 15:51:30 -04002991#endif
Alex Tomasa86c6182006-10-11 01:21:03 -07002992#ifdef EXTENTS_STATS
2993 spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
2994 EXT4_SB(sb)->s_ext_min = 1 << 30;
2995 EXT4_SB(sb)->s_ext_max = 0;
2996#endif
2997 }
2998}
2999
3000/*
3001 * called at umount time
3002 */
3003void ext4_ext_release(struct super_block *sb)
3004{
Theodore Ts'o83982b62009-01-06 14:53:16 -05003005 if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_EXTENTS))
Alex Tomasa86c6182006-10-11 01:21:03 -07003006 return;
3007
3008#ifdef EXTENTS_STATS
3009 if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
3010 struct ext4_sb_info *sbi = EXT4_SB(sb);
3011 printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
3012 sbi->s_ext_blocks, sbi->s_ext_extents,
3013 sbi->s_ext_blocks / sbi->s_ext_extents);
3014 printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
3015 sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
3016 }
3017#endif
3018}
3019
Zheng Liud7b2a002013-08-28 14:47:06 -04003020static int ext4_zeroout_es(struct inode *inode, struct ext4_extent *ex)
3021{
3022 ext4_lblk_t ee_block;
3023 ext4_fsblk_t ee_pblock;
3024 unsigned int ee_len;
3025
3026 ee_block = le32_to_cpu(ex->ee_block);
3027 ee_len = ext4_ext_get_actual_len(ex);
3028 ee_pblock = ext4_ext_pblock(ex);
3029
3030 if (ee_len == 0)
3031 return 0;
3032
3033 return ext4_es_insert_extent(inode, ee_block, ee_len, ee_pblock,
3034 EXTENT_STATUS_WRITTEN);
3035}
3036
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003037/* FIXME!! we need to try to merge to left or right after zero-out */
3038static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
3039{
Lukas Czerner24075182010-10-27 21:30:06 -04003040 ext4_fsblk_t ee_pblock;
3041 unsigned int ee_len;
Jing Zhangb7203032010-05-12 00:00:00 -04003042 int ret;
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003043
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003044 ee_len = ext4_ext_get_actual_len(ex);
Theodore Ts'obf89d162010-10-27 21:30:14 -04003045 ee_pblock = ext4_ext_pblock(ex);
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003046
Theodore Ts'oa107e5a2010-10-27 23:44:47 -04003047 ret = sb_issue_zeroout(inode->i_sb, ee_pblock, ee_len, GFP_NOFS);
Lukas Czerner24075182010-10-27 21:30:06 -04003048 if (ret > 0)
3049 ret = 0;
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003050
Lukas Czerner24075182010-10-27 21:30:06 -04003051 return ret;
Aneesh Kumar K.V093a0882008-04-29 08:11:12 -04003052}
3053
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003054/*
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003055 * ext4_split_extent_at() splits an extent at given block.
3056 *
3057 * @handle: the journal handle
3058 * @inode: the file inode
3059 * @path: the path to the extent
3060 * @split: the logical block where the extent is splitted.
3061 * @split_flags: indicates if the extent could be zeroout if split fails, and
3062 * the states(init or uninit) of new extents.
3063 * @flags: flags used to insert new extent to extent tree.
3064 *
3065 *
3066 * Splits extent [a, b] into two extents [a, @split) and [@split, b], states
3067 * of which are deterimined by split_flag.
3068 *
3069 * There are two cases:
3070 * a> the extent are splitted into two extent.
3071 * b> split is not needed, and just mark the extent.
3072 *
3073 * return 0 on success.
3074 */
3075static int ext4_split_extent_at(handle_t *handle,
3076 struct inode *inode,
3077 struct ext4_ext_path *path,
3078 ext4_lblk_t split,
3079 int split_flag,
3080 int flags)
3081{
3082 ext4_fsblk_t newblock;
3083 ext4_lblk_t ee_block;
Zheng Liuadb23552013-03-10 21:13:05 -04003084 struct ext4_extent *ex, newex, orig_ex, zero_ex;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003085 struct ext4_extent *ex2 = NULL;
3086 unsigned int ee_len, depth;
3087 int err = 0;
3088
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003089 BUG_ON((split_flag & (EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2)) ==
3090 (EXT4_EXT_DATA_VALID1 | EXT4_EXT_DATA_VALID2));
3091
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003092 ext_debug("ext4_split_extents_at: inode %lu, logical"
3093 "block %llu\n", inode->i_ino, (unsigned long long)split);
3094
3095 ext4_ext_show_leaf(inode, path);
3096
3097 depth = ext_depth(inode);
3098 ex = path[depth].p_ext;
3099 ee_block = le32_to_cpu(ex->ee_block);
3100 ee_len = ext4_ext_get_actual_len(ex);
3101 newblock = split - ee_block + ext4_ext_pblock(ex);
3102
3103 BUG_ON(split < ee_block || split >= (ee_block + ee_len));
Dmitry Monakhov357b66f2013-03-04 00:34:34 -05003104 BUG_ON(!ext4_ext_is_uninitialized(ex) &&
3105 split_flag & (EXT4_EXT_MAY_ZEROOUT |
3106 EXT4_EXT_MARK_UNINIT1 |
3107 EXT4_EXT_MARK_UNINIT2));
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003108
3109 err = ext4_ext_get_access(handle, inode, path + depth);
3110 if (err)
3111 goto out;
3112
3113 if (split == ee_block) {
3114 /*
3115 * case b: block @split is the block that the extent begins with
3116 * then we just change the state of the extent, and splitting
3117 * is not needed.
3118 */
3119 if (split_flag & EXT4_EXT_MARK_UNINIT2)
3120 ext4_ext_mark_uninitialized(ex);
3121 else
3122 ext4_ext_mark_initialized(ex);
3123
3124 if (!(flags & EXT4_GET_BLOCKS_PRE_IO))
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003125 ext4_ext_try_to_merge(handle, inode, path, ex);
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003126
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003127 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003128 goto out;
3129 }
3130
3131 /* case a */
3132 memcpy(&orig_ex, ex, sizeof(orig_ex));
3133 ex->ee_len = cpu_to_le16(split - ee_block);
3134 if (split_flag & EXT4_EXT_MARK_UNINIT1)
3135 ext4_ext_mark_uninitialized(ex);
3136
3137 /*
3138 * path may lead to new leaf, not to original leaf any more
3139 * after ext4_ext_insert_extent() returns,
3140 */
3141 err = ext4_ext_dirty(handle, inode, path + depth);
3142 if (err)
3143 goto fix_extent_len;
3144
3145 ex2 = &newex;
3146 ex2->ee_block = cpu_to_le32(split);
3147 ex2->ee_len = cpu_to_le16(ee_len - (split - ee_block));
3148 ext4_ext_store_pblock(ex2, newblock);
3149 if (split_flag & EXT4_EXT_MARK_UNINIT2)
3150 ext4_ext_mark_uninitialized(ex2);
3151
3152 err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
3153 if (err == -ENOSPC && (EXT4_EXT_MAY_ZEROOUT & split_flag)) {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003154 if (split_flag & (EXT4_EXT_DATA_VALID1|EXT4_EXT_DATA_VALID2)) {
Zheng Liuadb23552013-03-10 21:13:05 -04003155 if (split_flag & EXT4_EXT_DATA_VALID1) {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003156 err = ext4_ext_zeroout(inode, ex2);
Zheng Liuadb23552013-03-10 21:13:05 -04003157 zero_ex.ee_block = ex2->ee_block;
Zheng Liu8cde7ad2013-04-03 12:27:18 -04003158 zero_ex.ee_len = cpu_to_le16(
3159 ext4_ext_get_actual_len(ex2));
Zheng Liuadb23552013-03-10 21:13:05 -04003160 ext4_ext_store_pblock(&zero_ex,
3161 ext4_ext_pblock(ex2));
3162 } else {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003163 err = ext4_ext_zeroout(inode, ex);
Zheng Liuadb23552013-03-10 21:13:05 -04003164 zero_ex.ee_block = ex->ee_block;
Zheng Liu8cde7ad2013-04-03 12:27:18 -04003165 zero_ex.ee_len = cpu_to_le16(
3166 ext4_ext_get_actual_len(ex));
Zheng Liuadb23552013-03-10 21:13:05 -04003167 ext4_ext_store_pblock(&zero_ex,
3168 ext4_ext_pblock(ex));
3169 }
3170 } else {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003171 err = ext4_ext_zeroout(inode, &orig_ex);
Zheng Liuadb23552013-03-10 21:13:05 -04003172 zero_ex.ee_block = orig_ex.ee_block;
Zheng Liu8cde7ad2013-04-03 12:27:18 -04003173 zero_ex.ee_len = cpu_to_le16(
3174 ext4_ext_get_actual_len(&orig_ex));
Zheng Liuadb23552013-03-10 21:13:05 -04003175 ext4_ext_store_pblock(&zero_ex,
3176 ext4_ext_pblock(&orig_ex));
3177 }
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003178
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003179 if (err)
3180 goto fix_extent_len;
3181 /* update the extent length and mark as initialized */
Al Viroaf1584f2012-04-12 20:32:25 -04003182 ex->ee_len = cpu_to_le16(ee_len);
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003183 ext4_ext_try_to_merge(handle, inode, path, ex);
3184 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
Zheng Liuadb23552013-03-10 21:13:05 -04003185 if (err)
3186 goto fix_extent_len;
3187
3188 /* update extent status tree */
Zheng Liud7b2a002013-08-28 14:47:06 -04003189 err = ext4_zeroout_es(inode, &zero_ex);
Zheng Liuadb23552013-03-10 21:13:05 -04003190
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003191 goto out;
3192 } else if (err)
3193 goto fix_extent_len;
3194
3195out:
3196 ext4_ext_show_leaf(inode, path);
3197 return err;
3198
3199fix_extent_len:
3200 ex->ee_len = orig_ex.ee_len;
3201 ext4_ext_dirty(handle, inode, path + depth);
3202 return err;
3203}
3204
3205/*
3206 * ext4_split_extents() splits an extent and mark extent which is covered
3207 * by @map as split_flags indicates
3208 *
Anatol Pomozov70261f52013-08-28 14:40:12 -04003209 * It may result in splitting the extent into multiple extents (up to three)
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003210 * There are three possibilities:
3211 * a> There is no split required
3212 * b> Splits in two extents: Split is happening at either end of the extent
3213 * c> Splits in three extents: Somone is splitting in middle of the extent
3214 *
3215 */
3216static int ext4_split_extent(handle_t *handle,
3217 struct inode *inode,
3218 struct ext4_ext_path *path,
3219 struct ext4_map_blocks *map,
3220 int split_flag,
3221 int flags)
3222{
3223 ext4_lblk_t ee_block;
3224 struct ext4_extent *ex;
3225 unsigned int ee_len, depth;
3226 int err = 0;
3227 int uninitialized;
3228 int split_flag1, flags1;
Zheng Liu3a225672013-03-10 21:20:23 -04003229 int allocated = map->m_len;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003230
3231 depth = ext_depth(inode);
3232 ex = path[depth].p_ext;
3233 ee_block = le32_to_cpu(ex->ee_block);
3234 ee_len = ext4_ext_get_actual_len(ex);
3235 uninitialized = ext4_ext_is_uninitialized(ex);
3236
3237 if (map->m_lblk + map->m_len < ee_block + ee_len) {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003238 split_flag1 = split_flag & EXT4_EXT_MAY_ZEROOUT;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003239 flags1 = flags | EXT4_GET_BLOCKS_PRE_IO;
3240 if (uninitialized)
3241 split_flag1 |= EXT4_EXT_MARK_UNINIT1 |
3242 EXT4_EXT_MARK_UNINIT2;
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003243 if (split_flag & EXT4_EXT_DATA_VALID2)
3244 split_flag1 |= EXT4_EXT_DATA_VALID1;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003245 err = ext4_split_extent_at(handle, inode, path,
3246 map->m_lblk + map->m_len, split_flag1, flags1);
Yongqiang Yang93917412011-05-22 20:49:12 -04003247 if (err)
3248 goto out;
Zheng Liu3a225672013-03-10 21:20:23 -04003249 } else {
3250 allocated = ee_len - (map->m_lblk - ee_block);
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003251 }
Dmitry Monakhov357b66f2013-03-04 00:34:34 -05003252 /*
3253 * Update path is required because previous ext4_split_extent_at() may
3254 * result in split of original leaf or extent zeroout.
3255 */
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003256 ext4_ext_drop_refs(path);
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04003257 path = ext4_ext_find_extent(inode, map->m_lblk, path, 0);
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003258 if (IS_ERR(path))
3259 return PTR_ERR(path);
Dmitry Monakhov357b66f2013-03-04 00:34:34 -05003260 depth = ext_depth(inode);
3261 ex = path[depth].p_ext;
3262 uninitialized = ext4_ext_is_uninitialized(ex);
3263 split_flag1 = 0;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003264
3265 if (map->m_lblk >= ee_block) {
Dmitry Monakhov357b66f2013-03-04 00:34:34 -05003266 split_flag1 = split_flag & EXT4_EXT_DATA_VALID2;
3267 if (uninitialized) {
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003268 split_flag1 |= EXT4_EXT_MARK_UNINIT1;
Dmitry Monakhov357b66f2013-03-04 00:34:34 -05003269 split_flag1 |= split_flag & (EXT4_EXT_MAY_ZEROOUT |
3270 EXT4_EXT_MARK_UNINIT2);
3271 }
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003272 err = ext4_split_extent_at(handle, inode, path,
3273 map->m_lblk, split_flag1, flags);
3274 if (err)
3275 goto out;
3276 }
3277
3278 ext4_ext_show_leaf(inode, path);
3279out:
Zheng Liu3a225672013-03-10 21:20:23 -04003280 return err ? err : allocated;
Yongqiang Yang47ea3bb2011-05-03 12:23:07 -04003281}
3282
Amit Arora56055d32007-07-17 21:42:38 -04003283/*
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003284 * This function is called by ext4_ext_map_blocks() if someone tries to write
Amit Arora56055d32007-07-17 21:42:38 -04003285 * to an uninitialized extent. It may result in splitting the uninitialized
Lucas De Marchi25985ed2011-03-30 22:57:33 -03003286 * extent into multiple extents (up to three - one initialized and two
Amit Arora56055d32007-07-17 21:42:38 -04003287 * uninitialized).
3288 * There are three possibilities:
3289 * a> There is no split required: Entire extent should be initialized
3290 * b> Splits in two extents: Write is happening at either end of the extent
3291 * c> Splits in three extents: Somone is writing in middle of the extent
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003292 *
3293 * Pre-conditions:
3294 * - The extent pointed to by 'path' is uninitialized.
3295 * - The extent pointed to by 'path' contains a superset
3296 * of the logical span [map->m_lblk, map->m_lblk + map->m_len).
3297 *
3298 * Post-conditions on success:
3299 * - the returned value is the number of blocks beyond map->l_lblk
3300 * that are allocated and initialized.
3301 * It is guaranteed to be >= map->m_len.
Amit Arora56055d32007-07-17 21:42:38 -04003302 */
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05003303static int ext4_ext_convert_to_initialized(handle_t *handle,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003304 struct inode *inode,
3305 struct ext4_map_blocks *map,
Lukas Czerner27dd4382013-04-09 22:11:22 -04003306 struct ext4_ext_path *path,
3307 int flags)
Amit Arora56055d32007-07-17 21:42:38 -04003308{
Zheng Liu67a5da52012-08-17 09:54:17 -04003309 struct ext4_sb_info *sbi;
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003310 struct ext4_extent_header *eh;
Yongqiang Yang667eff32011-05-03 12:25:07 -04003311 struct ext4_map_blocks split_map;
3312 struct ext4_extent zero_ex;
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003313 struct ext4_extent *ex, *abut_ex;
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003314 ext4_lblk_t ee_block, eof_block;
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003315 unsigned int ee_len, depth, map_len = map->m_len;
3316 int allocated = 0, max_zeroout = 0;
Amit Arora56055d32007-07-17 21:42:38 -04003317 int err = 0;
Yongqiang Yang667eff32011-05-03 12:25:07 -04003318 int split_flag = 0;
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003319
3320 ext_debug("ext4_ext_convert_to_initialized: inode %lu, logical"
3321 "block %llu, max_blocks %u\n", inode->i_ino,
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003322 (unsigned long long)map->m_lblk, map_len);
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003323
Zheng Liu67a5da52012-08-17 09:54:17 -04003324 sbi = EXT4_SB(inode->i_sb);
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003325 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3326 inode->i_sb->s_blocksize_bits;
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003327 if (eof_block < map->m_lblk + map_len)
3328 eof_block = map->m_lblk + map_len;
Amit Arora56055d32007-07-17 21:42:38 -04003329
3330 depth = ext_depth(inode);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003331 eh = path[depth].p_hdr;
Amit Arora56055d32007-07-17 21:42:38 -04003332 ex = path[depth].p_ext;
3333 ee_block = le32_to_cpu(ex->ee_block);
3334 ee_len = ext4_ext_get_actual_len(ex);
Zheng Liuadb23552013-03-10 21:13:05 -04003335 zero_ex.ee_len = 0;
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003336
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003337 trace_ext4_ext_convert_to_initialized_enter(inode, map, ex);
3338
3339 /* Pre-conditions */
3340 BUG_ON(!ext4_ext_is_uninitialized(ex));
3341 BUG_ON(!in_range(map->m_lblk, ee_block, ee_len));
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003342
3343 /*
3344 * Attempt to transfer newly initialized blocks from the currently
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003345 * uninitialized extent to its neighbor. This is much cheaper
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003346 * than an insertion followed by a merge as those involve costly
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003347 * memmove() calls. Transferring to the left is the common case in
3348 * steady state for workloads doing fallocate(FALLOC_FL_KEEP_SIZE)
3349 * followed by append writes.
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003350 *
3351 * Limitations of the current logic:
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003352 * - L1: we do not deal with writes covering the whole extent.
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003353 * This would require removing the extent if the transfer
3354 * is possible.
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003355 * - L2: we only attempt to merge with an extent stored in the
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003356 * same extent tree node.
3357 */
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003358 if ((map->m_lblk == ee_block) &&
3359 /* See if we can merge left */
3360 (map_len < ee_len) && /*L1*/
3361 (ex > EXT_FIRST_EXTENT(eh))) { /*L2*/
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003362 ext4_lblk_t prev_lblk;
3363 ext4_fsblk_t prev_pblk, ee_pblk;
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003364 unsigned int prev_len;
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003365
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003366 abut_ex = ex - 1;
3367 prev_lblk = le32_to_cpu(abut_ex->ee_block);
3368 prev_len = ext4_ext_get_actual_len(abut_ex);
3369 prev_pblk = ext4_ext_pblock(abut_ex);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003370 ee_pblk = ext4_ext_pblock(ex);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003371
3372 /*
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003373 * A transfer of blocks from 'ex' to 'abut_ex' is allowed
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003374 * upon those conditions:
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003375 * - C1: abut_ex is initialized,
3376 * - C2: abut_ex is logically abutting ex,
3377 * - C3: abut_ex is physically abutting ex,
3378 * - C4: abut_ex can receive the additional blocks without
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003379 * overflowing the (initialized) length limit.
3380 */
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003381 if ((!ext4_ext_is_uninitialized(abut_ex)) && /*C1*/
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003382 ((prev_lblk + prev_len) == ee_block) && /*C2*/
3383 ((prev_pblk + prev_len) == ee_pblk) && /*C3*/
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003384 (prev_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003385 err = ext4_ext_get_access(handle, inode, path + depth);
3386 if (err)
3387 goto out;
3388
3389 trace_ext4_ext_convert_to_initialized_fastpath(inode,
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003390 map, ex, abut_ex);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003391
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003392 /* Shift the start of ex by 'map_len' blocks */
3393 ex->ee_block = cpu_to_le32(ee_block + map_len);
3394 ext4_ext_store_pblock(ex, ee_pblk + map_len);
3395 ex->ee_len = cpu_to_le16(ee_len - map_len);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003396 ext4_ext_mark_uninitialized(ex); /* Restore the flag */
3397
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003398 /* Extend abut_ex by 'map_len' blocks */
3399 abut_ex->ee_len = cpu_to_le16(prev_len + map_len);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003400
3401 /* Result: number of initialized blocks past m_lblk */
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003402 allocated = map_len;
3403 }
3404 } else if (((map->m_lblk + map_len) == (ee_block + ee_len)) &&
3405 (map_len < ee_len) && /*L1*/
3406 ex < EXT_LAST_EXTENT(eh)) { /*L2*/
3407 /* See if we can merge right */
3408 ext4_lblk_t next_lblk;
3409 ext4_fsblk_t next_pblk, ee_pblk;
3410 unsigned int next_len;
3411
3412 abut_ex = ex + 1;
3413 next_lblk = le32_to_cpu(abut_ex->ee_block);
3414 next_len = ext4_ext_get_actual_len(abut_ex);
3415 next_pblk = ext4_ext_pblock(abut_ex);
3416 ee_pblk = ext4_ext_pblock(ex);
3417
3418 /*
3419 * A transfer of blocks from 'ex' to 'abut_ex' is allowed
3420 * upon those conditions:
3421 * - C1: abut_ex is initialized,
3422 * - C2: abut_ex is logically abutting ex,
3423 * - C3: abut_ex is physically abutting ex,
3424 * - C4: abut_ex can receive the additional blocks without
3425 * overflowing the (initialized) length limit.
3426 */
3427 if ((!ext4_ext_is_uninitialized(abut_ex)) && /*C1*/
3428 ((map->m_lblk + map_len) == next_lblk) && /*C2*/
3429 ((ee_pblk + ee_len) == next_pblk) && /*C3*/
3430 (next_len < (EXT_INIT_MAX_LEN - map_len))) { /*C4*/
3431 err = ext4_ext_get_access(handle, inode, path + depth);
3432 if (err)
3433 goto out;
3434
3435 trace_ext4_ext_convert_to_initialized_fastpath(inode,
3436 map, ex, abut_ex);
3437
3438 /* Shift the start of abut_ex by 'map_len' blocks */
3439 abut_ex->ee_block = cpu_to_le32(next_lblk - map_len);
3440 ext4_ext_store_pblock(abut_ex, next_pblk - map_len);
3441 ex->ee_len = cpu_to_le16(ee_len - map_len);
3442 ext4_ext_mark_uninitialized(ex); /* Restore the flag */
3443
3444 /* Extend abut_ex by 'map_len' blocks */
3445 abut_ex->ee_len = cpu_to_le16(next_len + map_len);
3446
3447 /* Result: number of initialized blocks past m_lblk */
3448 allocated = map_len;
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003449 }
3450 }
Lukas Czernerbc2d9db2013-04-03 23:33:27 -04003451 if (allocated) {
3452 /* Mark the block containing both extents as dirty */
3453 ext4_ext_dirty(handle, inode, path + depth);
3454
3455 /* Update path to point to the right extent */
3456 path[depth].p_ext = abut_ex;
3457 goto out;
3458 } else
3459 allocated = ee_len - (map->m_lblk - ee_block);
Eric Gouriou6f91bc52011-10-27 11:43:23 -04003460
Yongqiang Yang667eff32011-05-03 12:25:07 -04003461 WARN_ON(map->m_lblk < ee_block);
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003462 /*
3463 * It is safe to convert extent to initialized via explicit
3464 * zeroout only if extent is fully insde i_size or new_size.
3465 */
Yongqiang Yang667eff32011-05-03 12:25:07 -04003466 split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003467
Zheng Liu67a5da52012-08-17 09:54:17 -04003468 if (EXT4_EXT_MAY_ZEROOUT & split_flag)
3469 max_zeroout = sbi->s_extent_max_zeroout_kb >>
Lukas Czerner4f42f802013-03-12 12:40:04 -04003470 (inode->i_sb->s_blocksize_bits - 10);
Zheng Liu67a5da52012-08-17 09:54:17 -04003471
3472 /* If extent is less than s_max_zeroout_kb, zeroout directly */
3473 if (max_zeroout && (ee_len <= max_zeroout)) {
Yongqiang Yang667eff32011-05-03 12:25:07 -04003474 err = ext4_ext_zeroout(inode, ex);
Aneesh Kumar K.V3977c962008-04-17 10:38:59 -04003475 if (err)
Aneesh Kumar K.Vd03856b2008-08-02 18:51:32 -04003476 goto out;
Zheng Liuadb23552013-03-10 21:13:05 -04003477 zero_ex.ee_block = ex->ee_block;
Zheng Liu8cde7ad2013-04-03 12:27:18 -04003478 zero_ex.ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex));
Zheng Liuadb23552013-03-10 21:13:05 -04003479 ext4_ext_store_pblock(&zero_ex, ext4_ext_pblock(ex));
Aneesh Kumar K.Vd03856b2008-08-02 18:51:32 -04003480
3481 err = ext4_ext_get_access(handle, inode, path + depth);
3482 if (err)
3483 goto out;
Yongqiang Yang667eff32011-05-03 12:25:07 -04003484 ext4_ext_mark_initialized(ex);
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003485 ext4_ext_try_to_merge(handle, inode, path, ex);
3486 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
Yongqiang Yang667eff32011-05-03 12:25:07 -04003487 goto out;
Amit Arora56055d32007-07-17 21:42:38 -04003488 }
Yongqiang Yang667eff32011-05-03 12:25:07 -04003489
Amit Arora56055d32007-07-17 21:42:38 -04003490 /*
Yongqiang Yang667eff32011-05-03 12:25:07 -04003491 * four cases:
3492 * 1. split the extent into three extents.
3493 * 2. split the extent into two extents, zeroout the first half.
3494 * 3. split the extent into two extents, zeroout the second half.
3495 * 4. split the extent into two extents with out zeroout.
Amit Arora56055d32007-07-17 21:42:38 -04003496 */
Yongqiang Yang667eff32011-05-03 12:25:07 -04003497 split_map.m_lblk = map->m_lblk;
3498 split_map.m_len = map->m_len;
3499
Zheng Liu67a5da52012-08-17 09:54:17 -04003500 if (max_zeroout && (allocated > map->m_len)) {
3501 if (allocated <= max_zeroout) {
Yongqiang Yang667eff32011-05-03 12:25:07 -04003502 /* case 3 */
3503 zero_ex.ee_block =
Allison Henderson9b940f82011-05-16 10:11:09 -04003504 cpu_to_le32(map->m_lblk);
3505 zero_ex.ee_len = cpu_to_le16(allocated);
Yongqiang Yang667eff32011-05-03 12:25:07 -04003506 ext4_ext_store_pblock(&zero_ex,
3507 ext4_ext_pblock(ex) + map->m_lblk - ee_block);
3508 err = ext4_ext_zeroout(inode, &zero_ex);
Amit Arora56055d32007-07-17 21:42:38 -04003509 if (err)
3510 goto out;
Yongqiang Yang667eff32011-05-03 12:25:07 -04003511 split_map.m_lblk = map->m_lblk;
3512 split_map.m_len = allocated;
Zheng Liu67a5da52012-08-17 09:54:17 -04003513 } else if (map->m_lblk - ee_block + map->m_len < max_zeroout) {
Yongqiang Yang667eff32011-05-03 12:25:07 -04003514 /* case 2 */
3515 if (map->m_lblk != ee_block) {
3516 zero_ex.ee_block = ex->ee_block;
3517 zero_ex.ee_len = cpu_to_le16(map->m_lblk -
3518 ee_block);
3519 ext4_ext_store_pblock(&zero_ex,
3520 ext4_ext_pblock(ex));
3521 err = ext4_ext_zeroout(inode, &zero_ex);
3522 if (err)
3523 goto out;
3524 }
3525
Yongqiang Yang667eff32011-05-03 12:25:07 -04003526 split_map.m_lblk = ee_block;
Allison Henderson9b940f82011-05-16 10:11:09 -04003527 split_map.m_len = map->m_lblk - ee_block + map->m_len;
3528 allocated = map->m_len;
Amit Arora56055d32007-07-17 21:42:38 -04003529 }
3530 }
Yongqiang Yang667eff32011-05-03 12:25:07 -04003531
3532 allocated = ext4_split_extent(handle, inode, path,
Lukas Czerner27dd4382013-04-09 22:11:22 -04003533 &split_map, split_flag, flags);
Yongqiang Yang667eff32011-05-03 12:25:07 -04003534 if (allocated < 0)
3535 err = allocated;
3536
Amit Arora56055d32007-07-17 21:42:38 -04003537out:
Zheng Liuadb23552013-03-10 21:13:05 -04003538 /* If we have gotten a failure, don't zero out status tree */
3539 if (!err)
Zheng Liud7b2a002013-08-28 14:47:06 -04003540 err = ext4_zeroout_es(inode, &zero_ex);
Amit Arora56055d32007-07-17 21:42:38 -04003541 return err ? err : allocated;
3542}
3543
Aneesh Kumar K.Vc278bfe2008-01-28 23:58:27 -05003544/*
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003545 * This function is called by ext4_ext_map_blocks() from
Mingming Cao00314622009-09-28 15:49:08 -04003546 * ext4_get_blocks_dio_write() when DIO to write
3547 * to an uninitialized extent.
3548 *
Paul Bollefd018fe2011-02-15 00:05:43 +01003549 * Writing to an uninitialized extent may result in splitting the uninitialized
Wang Sheng-Hui30cb27d2012-08-18 22:38:07 -04003550 * extent into multiple initialized/uninitialized extents (up to three)
Mingming Cao00314622009-09-28 15:49:08 -04003551 * There are three possibilities:
3552 * a> There is no split required: Entire extent should be uninitialized
3553 * b> Splits in two extents: Write is happening at either end of the extent
3554 * c> Splits in three extents: Somone is writing in middle of the extent
3555 *
3556 * One of more index blocks maybe needed if the extent tree grow after
Uwe Kleine-Königb5950762010-11-01 15:38:34 -04003557 * the uninitialized extent split. To prevent ENOSPC occur at the IO
Mingming Cao00314622009-09-28 15:49:08 -04003558 * complete, we need to split the uninitialized extent before DIO submit
Uwe Kleine-König421f91d2010-06-11 12:17:00 +02003559 * the IO. The uninitialized extent called at this time will be split
Mingming Cao00314622009-09-28 15:49:08 -04003560 * into three uninitialized extent(at most). After IO complete, the part
3561 * being filled will be convert to initialized by the end_io callback function
3562 * via ext4_convert_unwritten_extents().
Mingmingba230c32009-11-06 04:01:23 -05003563 *
3564 * Returns the size of uninitialized extent to be written on success.
Mingming Cao00314622009-09-28 15:49:08 -04003565 */
3566static int ext4_split_unwritten_extents(handle_t *handle,
3567 struct inode *inode,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003568 struct ext4_map_blocks *map,
Mingming Cao00314622009-09-28 15:49:08 -04003569 struct ext4_ext_path *path,
Mingming Cao00314622009-09-28 15:49:08 -04003570 int flags)
3571{
Yongqiang Yang667eff32011-05-03 12:25:07 -04003572 ext4_lblk_t eof_block;
3573 ext4_lblk_t ee_block;
3574 struct ext4_extent *ex;
3575 unsigned int ee_len;
3576 int split_flag = 0, depth;
Mingming Cao00314622009-09-28 15:49:08 -04003577
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003578 ext_debug("ext4_split_unwritten_extents: inode %lu, logical"
3579 "block %llu, max_blocks %u\n", inode->i_ino,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003580 (unsigned long long)map->m_lblk, map->m_len);
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003581
3582 eof_block = (inode->i_size + inode->i_sb->s_blocksize - 1) >>
3583 inode->i_sb->s_blocksize_bits;
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003584 if (eof_block < map->m_lblk + map->m_len)
3585 eof_block = map->m_lblk + map->m_len;
Mingming Cao00314622009-09-28 15:49:08 -04003586 /*
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003587 * It is safe to convert extent to initialized via explicit
3588 * zeroout only if extent is fully insde i_size or new_size.
3589 */
Yongqiang Yang667eff32011-05-03 12:25:07 -04003590 depth = ext_depth(inode);
3591 ex = path[depth].p_ext;
3592 ee_block = le32_to_cpu(ex->ee_block);
3593 ee_len = ext4_ext_get_actual_len(ex);
Dmitry Monakhov21ca0872010-05-16 06:00:00 -04003594
Yongqiang Yang667eff32011-05-03 12:25:07 -04003595 split_flag |= ee_block + ee_len <= eof_block ? EXT4_EXT_MAY_ZEROOUT : 0;
3596 split_flag |= EXT4_EXT_MARK_UNINIT2;
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003597 if (flags & EXT4_GET_BLOCKS_CONVERT)
3598 split_flag |= EXT4_EXT_DATA_VALID2;
Yongqiang Yang667eff32011-05-03 12:25:07 -04003599 flags |= EXT4_GET_BLOCKS_PRE_IO;
3600 return ext4_split_extent(handle, inode, path, map, split_flag, flags);
Mingming Cao00314622009-09-28 15:49:08 -04003601}
Yongqiang Yang197217a2011-05-03 11:45:29 -04003602
Jiaying Zhangc7064ef2010-03-02 13:28:44 -05003603static int ext4_convert_unwritten_extents_endio(handle_t *handle,
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003604 struct inode *inode,
3605 struct ext4_map_blocks *map,
3606 struct ext4_ext_path *path)
Mingming Cao00314622009-09-28 15:49:08 -04003607{
3608 struct ext4_extent *ex;
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003609 ext4_lblk_t ee_block;
3610 unsigned int ee_len;
Mingming Cao00314622009-09-28 15:49:08 -04003611 int depth;
3612 int err = 0;
Mingming Cao00314622009-09-28 15:49:08 -04003613
3614 depth = ext_depth(inode);
Mingming Cao00314622009-09-28 15:49:08 -04003615 ex = path[depth].p_ext;
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003616 ee_block = le32_to_cpu(ex->ee_block);
3617 ee_len = ext4_ext_get_actual_len(ex);
Mingming Cao00314622009-09-28 15:49:08 -04003618
Yongqiang Yang197217a2011-05-03 11:45:29 -04003619 ext_debug("ext4_convert_unwritten_extents_endio: inode %lu, logical"
3620 "block %llu, max_blocks %u\n", inode->i_ino,
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003621 (unsigned long long)ee_block, ee_len);
3622
Dmitry Monakhovff95ec22013-03-04 00:41:05 -05003623 /* If extent is larger than requested it is a clear sign that we still
3624 * have some extent state machine issues left. So extent_split is still
3625 * required.
3626 * TODO: Once all related issues will be fixed this situation should be
3627 * illegal.
3628 */
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003629 if (ee_block != map->m_lblk || ee_len > map->m_len) {
Dmitry Monakhovff95ec22013-03-04 00:41:05 -05003630#ifdef EXT4_DEBUG
3631 ext4_warning("Inode (%ld) finished: extent logical block %llu,"
3632 " len %u; IO logical block %llu, len %u\n",
3633 inode->i_ino, (unsigned long long)ee_block, ee_len,
3634 (unsigned long long)map->m_lblk, map->m_len);
3635#endif
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003636 err = ext4_split_unwritten_extents(handle, inode, map, path,
3637 EXT4_GET_BLOCKS_CONVERT);
3638 if (err < 0)
3639 goto out;
3640 ext4_ext_drop_refs(path);
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04003641 path = ext4_ext_find_extent(inode, map->m_lblk, path, 0);
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003642 if (IS_ERR(path)) {
3643 err = PTR_ERR(path);
3644 goto out;
3645 }
3646 depth = ext_depth(inode);
3647 ex = path[depth].p_ext;
3648 }
Yongqiang Yang197217a2011-05-03 11:45:29 -04003649
Mingming Cao00314622009-09-28 15:49:08 -04003650 err = ext4_ext_get_access(handle, inode, path + depth);
3651 if (err)
3652 goto out;
3653 /* first mark the extent as initialized */
3654 ext4_ext_mark_initialized(ex);
3655
Yongqiang Yang197217a2011-05-03 11:45:29 -04003656 /* note: ext4_ext_correct_indexes() isn't needed here because
3657 * borders are not changed
Mingming Cao00314622009-09-28 15:49:08 -04003658 */
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003659 ext4_ext_try_to_merge(handle, inode, path, ex);
Yongqiang Yang197217a2011-05-03 11:45:29 -04003660
Mingming Cao00314622009-09-28 15:49:08 -04003661 /* Mark modified extent as dirty */
Theodore Ts'oecb94f52012-08-17 09:44:17 -04003662 err = ext4_ext_dirty(handle, inode, path + path->p_depth);
Mingming Cao00314622009-09-28 15:49:08 -04003663out:
3664 ext4_ext_show_leaf(inode, path);
3665 return err;
3666}
3667
Aneesh Kumar K.V515f41c2009-12-29 23:39:06 -05003668static void unmap_underlying_metadata_blocks(struct block_device *bdev,
3669 sector_t block, int count)
3670{
3671 int i;
3672 for (i = 0; i < count; i++)
3673 unmap_underlying_metadata(bdev, block + i);
3674}
3675
Theodore Ts'o58590b02010-10-27 21:23:12 -04003676/*
3677 * Handle EOFBLOCKS_FL flag, clearing it if necessary
3678 */
3679static int check_eofblocks_fl(handle_t *handle, struct inode *inode,
Eric Sandeend002ebf2011-01-10 13:03:35 -05003680 ext4_lblk_t lblk,
Theodore Ts'o58590b02010-10-27 21:23:12 -04003681 struct ext4_ext_path *path,
3682 unsigned int len)
3683{
3684 int i, depth;
3685 struct ext4_extent_header *eh;
Sergey Senozhatsky65922cb2011-03-23 14:08:27 -04003686 struct ext4_extent *last_ex;
Theodore Ts'o58590b02010-10-27 21:23:12 -04003687
3688 if (!ext4_test_inode_flag(inode, EXT4_INODE_EOFBLOCKS))
3689 return 0;
3690
3691 depth = ext_depth(inode);
3692 eh = path[depth].p_hdr;
Theodore Ts'o58590b02010-10-27 21:23:12 -04003693
Lukas Czernerafcff5d2012-03-21 21:47:55 -04003694 /*
3695 * We're going to remove EOFBLOCKS_FL entirely in future so we
3696 * do not care for this case anymore. Simply remove the flag
3697 * if there are no extents.
3698 */
3699 if (unlikely(!eh->eh_entries))
3700 goto out;
Theodore Ts'o58590b02010-10-27 21:23:12 -04003701 last_ex = EXT_LAST_EXTENT(eh);
3702 /*
3703 * We should clear the EOFBLOCKS_FL flag if we are writing the
3704 * last block in the last extent in the file. We test this by
3705 * first checking to see if the caller to
3706 * ext4_ext_get_blocks() was interested in the last block (or
3707 * a block beyond the last block) in the current extent. If
3708 * this turns out to be false, we can bail out from this
3709 * function immediately.
3710 */
Eric Sandeend002ebf2011-01-10 13:03:35 -05003711 if (lblk + len < le32_to_cpu(last_ex->ee_block) +
Theodore Ts'o58590b02010-10-27 21:23:12 -04003712 ext4_ext_get_actual_len(last_ex))
3713 return 0;
3714 /*
3715 * If the caller does appear to be planning to write at or
3716 * beyond the end of the current extent, we then test to see
3717 * if the current extent is the last extent in the file, by
3718 * checking to make sure it was reached via the rightmost node
3719 * at each level of the tree.
3720 */
3721 for (i = depth-1; i >= 0; i--)
3722 if (path[i].p_idx != EXT_LAST_INDEX(path[i].p_hdr))
3723 return 0;
Lukas Czernerafcff5d2012-03-21 21:47:55 -04003724out:
Theodore Ts'o58590b02010-10-27 21:23:12 -04003725 ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
3726 return ext4_mark_inode_dirty(handle, inode);
3727}
3728
Aditya Kali7b415bf2011-09-09 19:04:51 -04003729/**
3730 * ext4_find_delalloc_range: find delayed allocated block in the given range.
3731 *
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003732 * Return 1 if there is a delalloc block in the range, otherwise 0.
Aditya Kali7b415bf2011-09-09 19:04:51 -04003733 */
Zheng Liuf7fec032013-02-18 00:28:47 -05003734int ext4_find_delalloc_range(struct inode *inode,
3735 ext4_lblk_t lblk_start,
3736 ext4_lblk_t lblk_end)
Aditya Kali7b415bf2011-09-09 19:04:51 -04003737{
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003738 struct extent_status es;
Aditya Kali7b415bf2011-09-09 19:04:51 -04003739
Yan, Zhenge30b5dc2013-05-03 02:15:52 -04003740 ext4_es_find_delayed_extent_range(inode, lblk_start, lblk_end, &es);
Zheng Liu06b0c882013-02-18 00:26:51 -05003741 if (es.es_len == 0)
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003742 return 0; /* there is no delay extent in this tree */
Zheng Liu06b0c882013-02-18 00:26:51 -05003743 else if (es.es_lblk <= lblk_start &&
3744 lblk_start < es.es_lblk + es.es_len)
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003745 return 1;
Zheng Liu06b0c882013-02-18 00:26:51 -05003746 else if (lblk_start <= es.es_lblk && es.es_lblk <= lblk_end)
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003747 return 1;
Aditya Kali7b415bf2011-09-09 19:04:51 -04003748 else
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003749 return 0;
Aditya Kali7b415bf2011-09-09 19:04:51 -04003750}
3751
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003752int ext4_find_delalloc_cluster(struct inode *inode, ext4_lblk_t lblk)
Aditya Kali7b415bf2011-09-09 19:04:51 -04003753{
3754 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3755 ext4_lblk_t lblk_start, lblk_end;
3756 lblk_start = lblk & (~(sbi->s_cluster_ratio - 1));
3757 lblk_end = lblk_start + sbi->s_cluster_ratio - 1;
3758
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003759 return ext4_find_delalloc_range(inode, lblk_start, lblk_end);
Aditya Kali7b415bf2011-09-09 19:04:51 -04003760}
3761
3762/**
3763 * Determines how many complete clusters (out of those specified by the 'map')
3764 * are under delalloc and were reserved quota for.
3765 * This function is called when we are writing out the blocks that were
3766 * originally written with their allocation delayed, but then the space was
3767 * allocated using fallocate() before the delayed allocation could be resolved.
3768 * The cases to look for are:
3769 * ('=' indicated delayed allocated blocks
3770 * '-' indicates non-delayed allocated blocks)
3771 * (a) partial clusters towards beginning and/or end outside of allocated range
3772 * are not delalloc'ed.
3773 * Ex:
3774 * |----c---=|====c====|====c====|===-c----|
3775 * |++++++ allocated ++++++|
3776 * ==> 4 complete clusters in above example
3777 *
3778 * (b) partial cluster (outside of allocated range) towards either end is
3779 * marked for delayed allocation. In this case, we will exclude that
3780 * cluster.
3781 * Ex:
3782 * |----====c========|========c========|
3783 * |++++++ allocated ++++++|
3784 * ==> 1 complete clusters in above example
3785 *
3786 * Ex:
3787 * |================c================|
3788 * |++++++ allocated ++++++|
3789 * ==> 0 complete clusters in above example
3790 *
3791 * The ext4_da_update_reserve_space will be called only if we
3792 * determine here that there were some "entire" clusters that span
3793 * this 'allocated' range.
3794 * In the non-bigalloc case, this function will just end up returning num_blks
3795 * without ever calling ext4_find_delalloc_range.
3796 */
3797static unsigned int
3798get_reserved_cluster_alloc(struct inode *inode, ext4_lblk_t lblk_start,
3799 unsigned int num_blks)
3800{
3801 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
3802 ext4_lblk_t alloc_cluster_start, alloc_cluster_end;
3803 ext4_lblk_t lblk_from, lblk_to, c_offset;
3804 unsigned int allocated_clusters = 0;
3805
3806 alloc_cluster_start = EXT4_B2C(sbi, lblk_start);
3807 alloc_cluster_end = EXT4_B2C(sbi, lblk_start + num_blks - 1);
3808
3809 /* max possible clusters for this allocation */
3810 allocated_clusters = alloc_cluster_end - alloc_cluster_start + 1;
3811
Aditya Kalid8990242011-09-09 19:18:51 -04003812 trace_ext4_get_reserved_cluster_alloc(inode, lblk_start, num_blks);
3813
Aditya Kali7b415bf2011-09-09 19:04:51 -04003814 /* Check towards left side */
3815 c_offset = lblk_start & (sbi->s_cluster_ratio - 1);
3816 if (c_offset) {
3817 lblk_from = lblk_start & (~(sbi->s_cluster_ratio - 1));
3818 lblk_to = lblk_from + c_offset - 1;
3819
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003820 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
Aditya Kali7b415bf2011-09-09 19:04:51 -04003821 allocated_clusters--;
3822 }
3823
3824 /* Now check towards right. */
3825 c_offset = (lblk_start + num_blks) & (sbi->s_cluster_ratio - 1);
3826 if (allocated_clusters && c_offset) {
3827 lblk_from = lblk_start + num_blks;
3828 lblk_to = lblk_from + (sbi->s_cluster_ratio - c_offset) - 1;
3829
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05003830 if (ext4_find_delalloc_range(inode, lblk_from, lblk_to))
Aditya Kali7b415bf2011-09-09 19:04:51 -04003831 allocated_clusters--;
3832 }
3833
3834 return allocated_clusters;
3835}
3836
Mingming Cao00314622009-09-28 15:49:08 -04003837static int
3838ext4_ext_handle_uninitialized_extents(handle_t *handle, struct inode *inode,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003839 struct ext4_map_blocks *map,
Mingming Cao00314622009-09-28 15:49:08 -04003840 struct ext4_ext_path *path, int flags,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003841 unsigned int allocated, ext4_fsblk_t newblock)
Mingming Cao00314622009-09-28 15:49:08 -04003842{
3843 int ret = 0;
3844 int err = 0;
Dmitry Monakhovf45ee3a2012-09-28 23:21:09 -04003845 ext4_io_end_t *io = ext4_inode_aio(inode);
Mingming Cao00314622009-09-28 15:49:08 -04003846
Zheng Liu88635ca2011-12-28 19:00:25 -05003847 ext_debug("ext4_ext_handle_uninitialized_extents: inode %lu, logical "
3848 "block %llu, max_blocks %u, flags %x, allocated %u\n",
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003849 inode->i_ino, (unsigned long long)map->m_lblk, map->m_len,
Mingming Cao00314622009-09-28 15:49:08 -04003850 flags, allocated);
3851 ext4_ext_show_leaf(inode, path);
3852
Lukas Czerner27dd4382013-04-09 22:11:22 -04003853 /*
3854 * When writing into uninitialized space, we should not fail to
3855 * allocate metadata blocks for the new extent block if needed.
3856 */
3857 flags |= EXT4_GET_BLOCKS_METADATA_NOFAIL;
3858
Zheng Liub5645532012-11-08 14:33:43 -05003859 trace_ext4_ext_handle_uninitialized_extents(inode, map, flags,
3860 allocated, newblock);
Aditya Kalid8990242011-09-09 19:18:51 -04003861
Jiaying Zhangc7064ef2010-03-02 13:28:44 -05003862 /* get_block() before submit the IO, split the extent */
Jiaying Zhang744692d2010-03-04 16:14:02 -05003863 if ((flags & EXT4_GET_BLOCKS_PRE_IO)) {
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003864 ret = ext4_split_unwritten_extents(handle, inode, map,
3865 path, flags);
Dmitry Monakhov82e54222012-09-28 23:36:25 -04003866 if (ret <= 0)
3867 goto out;
Mingming5f524952009-11-10 10:48:04 -05003868 /*
3869 * Flag the inode(non aio case) or end_io struct (aio case)
Lucas De Marchi25985ed2011-03-30 22:57:33 -03003870 * that this IO needs to conversion to written when IO is
Mingming5f524952009-11-10 10:48:04 -05003871 * completed
3872 */
Tao Ma0edeb712011-10-31 17:30:44 -04003873 if (io)
3874 ext4_set_io_unwritten_flag(inode, io);
3875 else
Theodore Ts'o19f5fb72010-01-24 14:34:07 -05003876 ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
Zheng Liua25a4e12013-02-18 00:28:04 -05003877 map->m_flags |= EXT4_MAP_UNWRITTEN;
Jiaying Zhang744692d2010-03-04 16:14:02 -05003878 if (ext4_should_dioread_nolock(inode))
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003879 map->m_flags |= EXT4_MAP_UNINIT;
Mingming Cao00314622009-09-28 15:49:08 -04003880 goto out;
3881 }
Jiaying Zhangc7064ef2010-03-02 13:28:44 -05003882 /* IO end_io complete, convert the filled extent to written */
Jiaying Zhang744692d2010-03-04 16:14:02 -05003883 if ((flags & EXT4_GET_BLOCKS_CONVERT)) {
Dmitry Monakhovdee1f972012-10-10 01:04:58 -04003884 ret = ext4_convert_unwritten_extents_endio(handle, inode, map,
Mingming Cao00314622009-09-28 15:49:08 -04003885 path);
Theodore Ts'o58590b02010-10-27 21:23:12 -04003886 if (ret >= 0) {
Jan Karab436b9b2009-12-08 23:51:10 -05003887 ext4_update_inode_fsync_trans(handle, inode, 1);
Eric Sandeend002ebf2011-01-10 13:03:35 -05003888 err = check_eofblocks_fl(handle, inode, map->m_lblk,
3889 path, map->m_len);
Theodore Ts'o58590b02010-10-27 21:23:12 -04003890 } else
3891 err = ret;
Zheng Liucdee7842013-03-10 21:08:52 -04003892 map->m_flags |= EXT4_MAP_MAPPED;
3893 if (allocated > map->m_len)
3894 allocated = map->m_len;
3895 map->m_len = allocated;
Mingming Cao00314622009-09-28 15:49:08 -04003896 goto out2;
3897 }
3898 /* buffered IO case */
3899 /*
3900 * repeat fallocate creation request
3901 * we already have an unwritten extent
3902 */
Zheng Liua25a4e12013-02-18 00:28:04 -05003903 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT) {
3904 map->m_flags |= EXT4_MAP_UNWRITTEN;
Mingming Cao00314622009-09-28 15:49:08 -04003905 goto map_out;
Zheng Liua25a4e12013-02-18 00:28:04 -05003906 }
Mingming Cao00314622009-09-28 15:49:08 -04003907
3908 /* buffered READ or buffered write_begin() lookup */
3909 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
3910 /*
3911 * We have blocks reserved already. We
3912 * return allocated blocks so that delalloc
3913 * won't do block reservation for us. But
3914 * the buffer head will be unmapped so that
3915 * a read from the block returns 0s.
3916 */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003917 map->m_flags |= EXT4_MAP_UNWRITTEN;
Mingming Cao00314622009-09-28 15:49:08 -04003918 goto out1;
3919 }
3920
3921 /* buffered write, writepage time, convert*/
Lukas Czerner27dd4382013-04-09 22:11:22 -04003922 ret = ext4_ext_convert_to_initialized(handle, inode, map, path, flags);
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04003923 if (ret >= 0)
Jan Karab436b9b2009-12-08 23:51:10 -05003924 ext4_update_inode_fsync_trans(handle, inode, 1);
Mingming Cao00314622009-09-28 15:49:08 -04003925out:
3926 if (ret <= 0) {
3927 err = ret;
3928 goto out2;
3929 } else
3930 allocated = ret;
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003931 map->m_flags |= EXT4_MAP_NEW;
Aneesh Kumar K.V515f41c2009-12-29 23:39:06 -05003932 /*
3933 * if we allocated more blocks than requested
3934 * we need to make sure we unmap the extra block
3935 * allocated. The actual needed block will get
3936 * unmapped later when we find the buffer_head marked
3937 * new.
3938 */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003939 if (allocated > map->m_len) {
Aneesh Kumar K.V515f41c2009-12-29 23:39:06 -05003940 unmap_underlying_metadata_blocks(inode->i_sb->s_bdev,
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003941 newblock + map->m_len,
3942 allocated - map->m_len);
3943 allocated = map->m_len;
Aneesh Kumar K.V515f41c2009-12-29 23:39:06 -05003944 }
Zheng Liu3a225672013-03-10 21:20:23 -04003945 map->m_len = allocated;
Aneesh Kumar K.V5f634d02010-01-25 04:00:31 -05003946
3947 /*
3948 * If we have done fallocate with the offset that is already
3949 * delayed allocated, we would have block reservation
3950 * and quota reservation done in the delayed write path.
3951 * But fallocate would have already updated quota and block
3952 * count for this offset. So cancel these reservation
3953 */
Aditya Kali7b415bf2011-09-09 19:04:51 -04003954 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
3955 unsigned int reserved_clusters;
3956 reserved_clusters = get_reserved_cluster_alloc(inode,
3957 map->m_lblk, map->m_len);
3958 if (reserved_clusters)
3959 ext4_da_update_reserve_space(inode,
3960 reserved_clusters,
3961 0);
3962 }
Aneesh Kumar K.V5f634d02010-01-25 04:00:31 -05003963
Mingming Cao00314622009-09-28 15:49:08 -04003964map_out:
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003965 map->m_flags |= EXT4_MAP_MAPPED;
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04003966 if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0) {
3967 err = check_eofblocks_fl(handle, inode, map->m_lblk, path,
3968 map->m_len);
3969 if (err < 0)
3970 goto out2;
3971 }
Mingming Cao00314622009-09-28 15:49:08 -04003972out1:
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003973 if (allocated > map->m_len)
3974 allocated = map->m_len;
Mingming Cao00314622009-09-28 15:49:08 -04003975 ext4_ext_show_leaf(inode, path);
Theodore Ts'oe35fd662010-05-16 19:00:00 -04003976 map->m_pblk = newblock;
3977 map->m_len = allocated;
Mingming Cao00314622009-09-28 15:49:08 -04003978out2:
3979 if (path) {
3980 ext4_ext_drop_refs(path);
3981 kfree(path);
3982 }
3983 return err ? err : allocated;
3984}
Theodore Ts'o58590b02010-10-27 21:23:12 -04003985
Mingming Cao00314622009-09-28 15:49:08 -04003986/*
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04003987 * get_implied_cluster_alloc - check to see if the requested
3988 * allocation (in the map structure) overlaps with a cluster already
3989 * allocated in an extent.
Aditya Kalid8990242011-09-09 19:18:51 -04003990 * @sb The filesystem superblock structure
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04003991 * @map The requested lblk->pblk mapping
3992 * @ex The extent structure which might contain an implied
3993 * cluster allocation
3994 *
3995 * This function is called by ext4_ext_map_blocks() after we failed to
3996 * find blocks that were already in the inode's extent tree. Hence,
3997 * we know that the beginning of the requested region cannot overlap
3998 * the extent from the inode's extent tree. There are three cases we
3999 * want to catch. The first is this case:
4000 *
4001 * |--- cluster # N--|
4002 * |--- extent ---| |---- requested region ---|
4003 * |==========|
4004 *
4005 * The second case that we need to test for is this one:
4006 *
4007 * |--------- cluster # N ----------------|
4008 * |--- requested region --| |------- extent ----|
4009 * |=======================|
4010 *
4011 * The third case is when the requested region lies between two extents
4012 * within the same cluster:
4013 * |------------- cluster # N-------------|
4014 * |----- ex -----| |---- ex_right ----|
4015 * |------ requested region ------|
4016 * |================|
4017 *
4018 * In each of the above cases, we need to set the map->m_pblk and
4019 * map->m_len so it corresponds to the return the extent labelled as
4020 * "|====|" from cluster #N, since it is already in use for data in
4021 * cluster EXT4_B2C(sbi, map->m_lblk). We will then return 1 to
4022 * signal to ext4_ext_map_blocks() that map->m_pblk should be treated
4023 * as a new "allocated" block region. Otherwise, we will return 0 and
4024 * ext4_ext_map_blocks() will then allocate one or more new clusters
4025 * by calling ext4_mb_new_blocks().
4026 */
Aditya Kalid8990242011-09-09 19:18:51 -04004027static int get_implied_cluster_alloc(struct super_block *sb,
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004028 struct ext4_map_blocks *map,
4029 struct ext4_extent *ex,
4030 struct ext4_ext_path *path)
4031{
Aditya Kalid8990242011-09-09 19:18:51 -04004032 struct ext4_sb_info *sbi = EXT4_SB(sb);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004033 ext4_lblk_t c_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
4034 ext4_lblk_t ex_cluster_start, ex_cluster_end;
Curt Wohlgemuth14d7f3e2011-12-18 17:39:02 -05004035 ext4_lblk_t rr_cluster_start;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004036 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
4037 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
4038 unsigned short ee_len = ext4_ext_get_actual_len(ex);
4039
4040 /* The extent passed in that we are trying to match */
4041 ex_cluster_start = EXT4_B2C(sbi, ee_block);
4042 ex_cluster_end = EXT4_B2C(sbi, ee_block + ee_len - 1);
4043
4044 /* The requested region passed into ext4_map_blocks() */
4045 rr_cluster_start = EXT4_B2C(sbi, map->m_lblk);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004046
4047 if ((rr_cluster_start == ex_cluster_end) ||
4048 (rr_cluster_start == ex_cluster_start)) {
4049 if (rr_cluster_start == ex_cluster_end)
4050 ee_start += ee_len - 1;
4051 map->m_pblk = (ee_start & ~(sbi->s_cluster_ratio - 1)) +
4052 c_offset;
4053 map->m_len = min(map->m_len,
4054 (unsigned) sbi->s_cluster_ratio - c_offset);
4055 /*
4056 * Check for and handle this case:
4057 *
4058 * |--------- cluster # N-------------|
4059 * |------- extent ----|
4060 * |--- requested region ---|
4061 * |===========|
4062 */
4063
4064 if (map->m_lblk < ee_block)
4065 map->m_len = min(map->m_len, ee_block - map->m_lblk);
4066
4067 /*
4068 * Check for the case where there is already another allocated
4069 * block to the right of 'ex' but before the end of the cluster.
4070 *
4071 * |------------- cluster # N-------------|
4072 * |----- ex -----| |---- ex_right ----|
4073 * |------ requested region ------|
4074 * |================|
4075 */
4076 if (map->m_lblk > ee_block) {
4077 ext4_lblk_t next = ext4_ext_next_allocated_block(path);
4078 map->m_len = min(map->m_len, next - map->m_lblk);
4079 }
Aditya Kalid8990242011-09-09 19:18:51 -04004080
4081 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 1);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004082 return 1;
4083 }
Aditya Kalid8990242011-09-09 19:18:51 -04004084
4085 trace_ext4_get_implied_cluster_alloc_exit(sb, map, 0);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004086 return 0;
4087}
4088
4089
4090/*
Mingming Caof5ab0d12008-02-25 15:29:55 -05004091 * Block allocation/map/preallocation routine for extents based files
4092 *
4093 *
Aneesh Kumar K.Vc278bfe2008-01-28 23:58:27 -05004094 * Need to be called with
Aneesh Kumar K.V0e855ac2008-01-28 23:58:26 -05004095 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
4096 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
Mingming Caof5ab0d12008-02-25 15:29:55 -05004097 *
4098 * return > 0, number of of blocks already mapped/allocated
4099 * if create == 0 and these are pre-allocated blocks
4100 * buffer head is unmapped
4101 * otherwise blocks are mapped
4102 *
4103 * return = 0, if plain look up failed (blocks have not been allocated)
4104 * buffer head is unmapped
4105 *
4106 * return < 0, error case.
Aneesh Kumar K.Vc278bfe2008-01-28 23:58:27 -05004107 */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004108int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
4109 struct ext4_map_blocks *map, int flags)
Alex Tomasa86c6182006-10-11 01:21:03 -07004110{
4111 struct ext4_ext_path *path = NULL;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004112 struct ext4_extent newex, *ex, *ex2;
4113 struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
Jiaying Zhang0562e0b2011-03-21 21:38:05 -04004114 ext4_fsblk_t newblock = 0;
Zheng Liu37794732012-11-08 14:47:52 -05004115 int free_on_err = 0, err = 0, depth;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004116 unsigned int allocated = 0, offset = 0;
Yongqiang Yang81fdbb42011-10-29 09:23:38 -04004117 unsigned int allocated_clusters = 0;
Alex Tomasc9de5602008-01-29 00:19:52 -05004118 struct ext4_allocation_request ar;
Dmitry Monakhovf45ee3a2012-09-28 23:21:09 -04004119 ext4_io_end_t *io = ext4_inode_aio(inode);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004120 ext4_lblk_t cluster_offset;
Dmitry Monakhov82e54222012-09-28 23:36:25 -04004121 int set_unwritten = 0;
Alex Tomasa86c6182006-10-11 01:21:03 -07004122
Mingming84fe3be2009-09-01 08:44:37 -04004123 ext_debug("blocks %u/%u requested for inode %lu\n",
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004124 map->m_lblk, map->m_len, inode->i_ino);
Jiaying Zhang0562e0b2011-03-21 21:38:05 -04004125 trace_ext4_ext_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07004126
Alex Tomasa86c6182006-10-11 01:21:03 -07004127 /* find extent for this block */
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04004128 path = ext4_ext_find_extent(inode, map->m_lblk, NULL, 0);
Alex Tomasa86c6182006-10-11 01:21:03 -07004129 if (IS_ERR(path)) {
4130 err = PTR_ERR(path);
4131 path = NULL;
4132 goto out2;
4133 }
4134
4135 depth = ext_depth(inode);
4136
4137 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07004138 * consistent leaf must not be empty;
4139 * this situation is possible, though, _during_ tree modification;
Alex Tomasa86c6182006-10-11 01:21:03 -07004140 * this is why assert can't be put in ext4_ext_find_extent()
4141 */
Frank Mayhar273df552010-03-02 11:46:09 -05004142 if (unlikely(path[depth].p_ext == NULL && depth != 0)) {
4143 EXT4_ERROR_INODE(inode, "bad extent address "
Theodore Ts'of70f3622010-05-16 23:00:00 -04004144 "lblock: %lu, depth: %d pblock %lld",
4145 (unsigned long) map->m_lblk, depth,
4146 path[depth].p_block);
Surbhi Palande034fb4c2009-12-14 09:53:52 -05004147 err = -EIO;
4148 goto out2;
4149 }
Alex Tomasa86c6182006-10-11 01:21:03 -07004150
Avantika Mathur7e028972006-12-06 20:41:33 -08004151 ex = path[depth].p_ext;
4152 if (ex) {
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05004153 ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
Theodore Ts'obf89d162010-10-27 21:30:14 -04004154 ext4_fsblk_t ee_start = ext4_ext_pblock(ex);
Amit Aroraa2df2a62007-07-17 21:42:41 -04004155 unsigned short ee_len;
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07004156
4157 /*
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07004158 * Uninitialized extents are treated as holes, except that
Amit Arora56055d32007-07-17 21:42:38 -04004159 * we split out initialized portions during a write.
Suparna Bhattacharya471d4012006-10-11 01:21:06 -07004160 */
Amit Aroraa2df2a62007-07-17 21:42:41 -04004161 ee_len = ext4_ext_get_actual_len(ex);
Aditya Kalid8990242011-09-09 19:18:51 -04004162
4163 trace_ext4_ext_show_extent(inode, ee_block, ee_start, ee_len);
4164
Randy Dunlapd0d856e2006-10-11 01:21:07 -07004165 /* if found extent covers block, simply return it */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004166 if (in_range(map->m_lblk, ee_block, ee_len)) {
4167 newblock = map->m_lblk - ee_block + ee_start;
Randy Dunlapd0d856e2006-10-11 01:21:07 -07004168 /* number of remaining blocks in the extent */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004169 allocated = ee_len - (map->m_lblk - ee_block);
4170 ext_debug("%u fit into %u:%d -> %llu\n", map->m_lblk,
4171 ee_block, ee_len, newblock);
Amit Arora56055d32007-07-17 21:42:38 -04004172
Zheng Liu69eb33d2013-02-18 00:31:07 -05004173 if (!ext4_ext_is_uninitialized(ex))
Lukas Czerner78771912012-03-19 23:05:43 -04004174 goto out;
Zheng Liu69eb33d2013-02-18 00:31:07 -05004175
Zheng Liu37794732012-11-08 14:47:52 -05004176 allocated = ext4_ext_handle_uninitialized_extents(
Lukas Czerner78771912012-03-19 23:05:43 -04004177 handle, inode, map, path, flags,
4178 allocated, newblock);
Zheng Liu37794732012-11-08 14:47:52 -05004179 goto out3;
Alex Tomasa86c6182006-10-11 01:21:03 -07004180 }
4181 }
4182
Aditya Kali7b415bf2011-09-09 19:04:51 -04004183 if ((sbi->s_cluster_ratio > 1) &&
Zheng Liu7d1b1fb2012-11-08 21:57:35 -05004184 ext4_find_delalloc_cluster(inode, map->m_lblk))
Aditya Kali7b415bf2011-09-09 19:04:51 -04004185 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
4186
Alex Tomasa86c6182006-10-11 01:21:03 -07004187 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07004188 * requested block isn't allocated yet;
Alex Tomasa86c6182006-10-11 01:21:03 -07004189 * we couldn't try to create block if create flag is zero
4190 */
Theodore Ts'oc2177052009-05-14 00:58:52 -04004191 if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) {
Amit Arora56055d32007-07-17 21:42:38 -04004192 /*
4193 * put just found gap into cache to speed up
4194 * subsequent requests
4195 */
Zheng Liud100eef2013-02-18 00:29:59 -05004196 if ((flags & EXT4_GET_BLOCKS_NO_PUT_HOLE) == 0)
4197 ext4_ext_put_gap_in_cache(inode, path, map->m_lblk);
Alex Tomasa86c6182006-10-11 01:21:03 -07004198 goto out2;
4199 }
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004200
Alex Tomasa86c6182006-10-11 01:21:03 -07004201 /*
Theodore Ts'oc2ea3fd2008-10-10 09:40:52 -04004202 * Okay, we need to do block allocation.
Andrew Morton63f57932006-10-11 01:21:24 -07004203 */
Aditya Kali7b415bf2011-09-09 19:04:51 -04004204 map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004205 newex.ee_block = cpu_to_le32(map->m_lblk);
4206 cluster_offset = map->m_lblk & (sbi->s_cluster_ratio-1);
4207
4208 /*
4209 * If we are doing bigalloc, check to see if the extent returned
4210 * by ext4_ext_find_extent() implies a cluster we can use.
4211 */
4212 if (cluster_offset && ex &&
Aditya Kalid8990242011-09-09 19:18:51 -04004213 get_implied_cluster_alloc(inode->i_sb, map, ex, path)) {
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004214 ar.len = allocated = map->m_len;
4215 newblock = map->m_pblk;
Aditya Kali7b415bf2011-09-09 19:04:51 -04004216 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004217 goto got_allocated_blocks;
4218 }
Alex Tomasa86c6182006-10-11 01:21:03 -07004219
Alex Tomasc9de5602008-01-29 00:19:52 -05004220 /* find neighbour allocated blocks */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004221 ar.lleft = map->m_lblk;
Alex Tomasc9de5602008-01-29 00:19:52 -05004222 err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
4223 if (err)
4224 goto out2;
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004225 ar.lright = map->m_lblk;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004226 ex2 = NULL;
4227 err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright, &ex2);
Alex Tomasc9de5602008-01-29 00:19:52 -05004228 if (err)
4229 goto out2;
Amit Arora25d14f92007-05-24 13:04:13 -04004230
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004231 /* Check if the extent after searching to the right implies a
4232 * cluster we can use. */
4233 if ((sbi->s_cluster_ratio > 1) && ex2 &&
Aditya Kalid8990242011-09-09 19:18:51 -04004234 get_implied_cluster_alloc(inode->i_sb, map, ex2, path)) {
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004235 ar.len = allocated = map->m_len;
4236 newblock = map->m_pblk;
Aditya Kali7b415bf2011-09-09 19:04:51 -04004237 map->m_flags |= EXT4_MAP_FROM_CLUSTER;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004238 goto got_allocated_blocks;
4239 }
4240
Amit Arora749269f2007-07-18 09:02:56 -04004241 /*
4242 * See if request is beyond maximum number of blocks we can have in
4243 * a single extent. For an initialized extent this limit is
4244 * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
4245 * EXT_UNINIT_MAX_LEN.
4246 */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004247 if (map->m_len > EXT_INIT_MAX_LEN &&
Theodore Ts'oc2177052009-05-14 00:58:52 -04004248 !(flags & EXT4_GET_BLOCKS_UNINIT_EXT))
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004249 map->m_len = EXT_INIT_MAX_LEN;
4250 else if (map->m_len > EXT_UNINIT_MAX_LEN &&
Theodore Ts'oc2177052009-05-14 00:58:52 -04004251 (flags & EXT4_GET_BLOCKS_UNINIT_EXT))
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004252 map->m_len = EXT_UNINIT_MAX_LEN;
Amit Arora749269f2007-07-18 09:02:56 -04004253
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004254 /* Check if we can really insert (m_lblk)::(m_lblk + m_len) extent */
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004255 newex.ee_len = cpu_to_le16(map->m_len);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004256 err = ext4_ext_check_overlap(sbi, inode, &newex, path);
Amit Arora25d14f92007-05-24 13:04:13 -04004257 if (err)
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05004258 allocated = ext4_ext_get_actual_len(&newex);
Amit Arora25d14f92007-05-24 13:04:13 -04004259 else
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004260 allocated = map->m_len;
Alex Tomasc9de5602008-01-29 00:19:52 -05004261
4262 /* allocate new block */
4263 ar.inode = inode;
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004264 ar.goal = ext4_ext_find_goal(inode, path, map->m_lblk);
4265 ar.logical = map->m_lblk;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004266 /*
4267 * We calculate the offset from the beginning of the cluster
4268 * for the logical block number, since when we allocate a
4269 * physical cluster, the physical block should start at the
4270 * same offset from the beginning of the cluster. This is
4271 * needed so that future calls to get_implied_cluster_alloc()
4272 * work correctly.
4273 */
4274 offset = map->m_lblk & (sbi->s_cluster_ratio - 1);
4275 ar.len = EXT4_NUM_B2C(sbi, offset+allocated);
4276 ar.goal -= offset;
4277 ar.logical -= offset;
Alex Tomasc9de5602008-01-29 00:19:52 -05004278 if (S_ISREG(inode->i_mode))
4279 ar.flags = EXT4_MB_HINT_DATA;
4280 else
4281 /* disable in-core preallocation for non-regular files */
4282 ar.flags = 0;
Vivek Haldar556b27a2011-05-25 07:41:54 -04004283 if (flags & EXT4_GET_BLOCKS_NO_NORMALIZE)
4284 ar.flags |= EXT4_MB_HINT_NOPREALLOC;
Alex Tomasc9de5602008-01-29 00:19:52 -05004285 newblock = ext4_mb_new_blocks(handle, &ar, &err);
Alex Tomasa86c6182006-10-11 01:21:03 -07004286 if (!newblock)
4287 goto out2;
Mingming84fe3be2009-09-01 08:44:37 -04004288 ext_debug("allocate new block: goal %llu, found %llu/%u\n",
Theodore Ts'o498e5f22008-11-05 00:14:04 -05004289 ar.goal, newblock, allocated);
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004290 free_on_err = 1;
Aditya Kali7b415bf2011-09-09 19:04:51 -04004291 allocated_clusters = ar.len;
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004292 ar.len = EXT4_C2B(sbi, ar.len) - offset;
4293 if (ar.len > allocated)
4294 ar.len = allocated;
Alex Tomasa86c6182006-10-11 01:21:03 -07004295
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004296got_allocated_blocks:
Alex Tomasa86c6182006-10-11 01:21:03 -07004297 /* try to insert new extent into found leaf and return */
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004298 ext4_ext_store_pblock(&newex, newblock + offset);
Alex Tomasc9de5602008-01-29 00:19:52 -05004299 newex.ee_len = cpu_to_le16(ar.len);
Mingming Cao8d5d02e2009-09-28 15:48:29 -04004300 /* Mark uninitialized */
4301 if (flags & EXT4_GET_BLOCKS_UNINIT_EXT){
Amit Aroraa2df2a62007-07-17 21:42:41 -04004302 ext4_ext_mark_uninitialized(&newex);
Zheng Liua25a4e12013-02-18 00:28:04 -05004303 map->m_flags |= EXT4_MAP_UNWRITTEN;
Mingming Cao8d5d02e2009-09-28 15:48:29 -04004304 /*
Jiaying Zhang744692d2010-03-04 16:14:02 -05004305 * io_end structure was created for every IO write to an
Lucas De Marchi25985ed2011-03-30 22:57:33 -03004306 * uninitialized extent. To avoid unnecessary conversion,
Jiaying Zhang744692d2010-03-04 16:14:02 -05004307 * here we flag the IO that really needs the conversion.
Mingming5f524952009-11-10 10:48:04 -05004308 * For non asycn direct IO case, flag the inode state
Lucas De Marchi25985ed2011-03-30 22:57:33 -03004309 * that we need to perform conversion when IO is done.
Mingming Cao8d5d02e2009-09-28 15:48:29 -04004310 */
Dmitry Monakhov82e54222012-09-28 23:36:25 -04004311 if ((flags & EXT4_GET_BLOCKS_PRE_IO))
4312 set_unwritten = 1;
Jiaying Zhang744692d2010-03-04 16:14:02 -05004313 if (ext4_should_dioread_nolock(inode))
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004314 map->m_flags |= EXT4_MAP_UNINIT;
Mingming Cao8d5d02e2009-09-28 15:48:29 -04004315 }
Jiaying Zhangc8d46e42010-02-24 09:52:53 -05004316
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04004317 err = 0;
4318 if ((flags & EXT4_GET_BLOCKS_KEEP_SIZE) == 0)
4319 err = check_eofblocks_fl(handle, inode, map->m_lblk,
4320 path, ar.len);
Jiaying Zhang575a1d42011-07-10 20:07:25 -04004321 if (!err)
4322 err = ext4_ext_insert_extent(handle, inode, path,
4323 &newex, flags);
Dmitry Monakhov82e54222012-09-28 23:36:25 -04004324
4325 if (!err && set_unwritten) {
4326 if (io)
4327 ext4_set_io_unwritten_flag(inode, io);
4328 else
4329 ext4_set_inode_state(inode,
4330 EXT4_STATE_DIO_UNWRITTEN);
4331 }
4332
Theodore Ts'o4d33b1e2011-09-09 18:52:51 -04004333 if (err && free_on_err) {
Maxim Patlasov7132de72011-07-10 19:37:48 -04004334 int fb_flags = flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE ?
4335 EXT4_FREE_BLOCKS_NO_QUOT_UPDATE : 0;
Alex Tomas315054f2007-05-24 13:04:25 -04004336 /* free data blocks we just allocated */
Alex Tomasc9de5602008-01-29 00:19:52 -05004337 /* not a good idea to call discard here directly,
4338 * but otherwise we'd need to call it every free() */
Theodore Ts'oc2ea3fd2008-10-10 09:40:52 -04004339 ext4_discard_preallocations(inode);
Theodore Ts'oc8e15132013-07-15 00:09:37 -04004340 ext4_free_blocks(handle, inode, NULL, newblock,
4341 EXT4_C2B(sbi, allocated_clusters), fb_flags);
Alex Tomasa86c6182006-10-11 01:21:03 -07004342 goto out2;
Alex Tomas315054f2007-05-24 13:04:25 -04004343 }
Alex Tomasa86c6182006-10-11 01:21:03 -07004344
Alex Tomasa86c6182006-10-11 01:21:03 -07004345 /* previous routine could use block we allocated */
Theodore Ts'obf89d162010-10-27 21:30:14 -04004346 newblock = ext4_ext_pblock(&newex);
Aneesh Kumar K.Vb939e372008-01-28 23:58:27 -05004347 allocated = ext4_ext_get_actual_len(&newex);
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004348 if (allocated > map->m_len)
4349 allocated = map->m_len;
4350 map->m_flags |= EXT4_MAP_NEW;
Alex Tomasa86c6182006-10-11 01:21:03 -07004351
Jan Karab436b9b2009-12-08 23:51:10 -05004352 /*
Aneesh Kumar K.V5f634d02010-01-25 04:00:31 -05004353 * Update reserved blocks/metadata blocks after successful
4354 * block allocation which had been deferred till now.
4355 */
Aditya Kali7b415bf2011-09-09 19:04:51 -04004356 if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
Yongqiang Yang81fdbb42011-10-29 09:23:38 -04004357 unsigned int reserved_clusters;
Aditya Kali7b415bf2011-09-09 19:04:51 -04004358 /*
Yongqiang Yang81fdbb42011-10-29 09:23:38 -04004359 * Check how many clusters we had reserved this allocated range
Aditya Kali7b415bf2011-09-09 19:04:51 -04004360 */
4361 reserved_clusters = get_reserved_cluster_alloc(inode,
4362 map->m_lblk, allocated);
4363 if (map->m_flags & EXT4_MAP_FROM_CLUSTER) {
4364 if (reserved_clusters) {
4365 /*
4366 * We have clusters reserved for this range.
4367 * But since we are not doing actual allocation
4368 * and are simply using blocks from previously
4369 * allocated cluster, we should release the
4370 * reservation and not claim quota.
4371 */
4372 ext4_da_update_reserve_space(inode,
4373 reserved_clusters, 0);
4374 }
4375 } else {
4376 BUG_ON(allocated_clusters < reserved_clusters);
Aditya Kali7b415bf2011-09-09 19:04:51 -04004377 if (reserved_clusters < allocated_clusters) {
Aditya Kali5356f2612011-09-09 19:20:51 -04004378 struct ext4_inode_info *ei = EXT4_I(inode);
Aditya Kali7b415bf2011-09-09 19:04:51 -04004379 int reservation = allocated_clusters -
4380 reserved_clusters;
4381 /*
4382 * It seems we claimed few clusters outside of
4383 * the range of this allocation. We should give
4384 * it back to the reservation pool. This can
4385 * happen in the following case:
4386 *
4387 * * Suppose s_cluster_ratio is 4 (i.e., each
4388 * cluster has 4 blocks. Thus, the clusters
4389 * are [0-3],[4-7],[8-11]...
4390 * * First comes delayed allocation write for
4391 * logical blocks 10 & 11. Since there were no
4392 * previous delayed allocated blocks in the
4393 * range [8-11], we would reserve 1 cluster
4394 * for this write.
4395 * * Next comes write for logical blocks 3 to 8.
4396 * In this case, we will reserve 2 clusters
4397 * (for [0-3] and [4-7]; and not for [8-11] as
4398 * that range has a delayed allocated blocks.
4399 * Thus total reserved clusters now becomes 3.
4400 * * Now, during the delayed allocation writeout
4401 * time, we will first write blocks [3-8] and
4402 * allocate 3 clusters for writing these
4403 * blocks. Also, we would claim all these
4404 * three clusters above.
4405 * * Now when we come here to writeout the
4406 * blocks [10-11], we would expect to claim
4407 * the reservation of 1 cluster we had made
4408 * (and we would claim it since there are no
4409 * more delayed allocated blocks in the range
4410 * [8-11]. But our reserved cluster count had
4411 * already gone to 0.
4412 *
4413 * Thus, at the step 4 above when we determine
4414 * that there are still some unwritten delayed
4415 * allocated blocks outside of our current
4416 * block range, we should increment the
4417 * reserved clusters count so that when the
4418 * remaining blocks finally gets written, we
4419 * could claim them.
4420 */
Aditya Kali5356f2612011-09-09 19:20:51 -04004421 dquot_reserve_block(inode,
4422 EXT4_C2B(sbi, reservation));
4423 spin_lock(&ei->i_block_reservation_lock);
4424 ei->i_reserved_data_blocks += reservation;
4425 spin_unlock(&ei->i_block_reservation_lock);
Aditya Kali7b415bf2011-09-09 19:04:51 -04004426 }
Lukas Czerner232ec872013-03-10 22:46:30 -04004427 /*
4428 * We will claim quota for all newly allocated blocks.
4429 * We're updating the reserved space *after* the
4430 * correction above so we do not accidentally free
4431 * all the metadata reservation because we might
4432 * actually need it later on.
4433 */
4434 ext4_da_update_reserve_space(inode, allocated_clusters,
4435 1);
Aditya Kali7b415bf2011-09-09 19:04:51 -04004436 }
4437 }
Aneesh Kumar K.V5f634d02010-01-25 04:00:31 -05004438
4439 /*
Jan Karab436b9b2009-12-08 23:51:10 -05004440 * Cache the extent and update transaction to commit on fdatasync only
4441 * when it is _not_ an uninitialized extent.
4442 */
Zheng Liu69eb33d2013-02-18 00:31:07 -05004443 if ((flags & EXT4_GET_BLOCKS_UNINIT_EXT) == 0)
Jan Karab436b9b2009-12-08 23:51:10 -05004444 ext4_update_inode_fsync_trans(handle, inode, 1);
Zheng Liu69eb33d2013-02-18 00:31:07 -05004445 else
Jan Karab436b9b2009-12-08 23:51:10 -05004446 ext4_update_inode_fsync_trans(handle, inode, 0);
Alex Tomasa86c6182006-10-11 01:21:03 -07004447out:
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004448 if (allocated > map->m_len)
4449 allocated = map->m_len;
Alex Tomasa86c6182006-10-11 01:21:03 -07004450 ext4_ext_show_leaf(inode, path);
Theodore Ts'oe35fd662010-05-16 19:00:00 -04004451 map->m_flags |= EXT4_MAP_MAPPED;
4452 map->m_pblk = newblock;
4453 map->m_len = allocated;
Alex Tomasa86c6182006-10-11 01:21:03 -07004454out2:
4455 if (path) {
4456 ext4_ext_drop_refs(path);
4457 kfree(path);
4458 }
Allison Hendersone8613042011-05-25 07:41:46 -04004459
Zheng Liu37794732012-11-08 14:47:52 -05004460out3:
Theodore Ts'o63b99962013-07-16 10:28:47 -04004461 trace_ext4_ext_map_blocks_exit(inode, flags, map,
4462 err ? err : allocated);
4463 ext4_es_lru_add(inode);
Lukas Czerner78771912012-03-19 23:05:43 -04004464 return err ? err : allocated;
Alex Tomasa86c6182006-10-11 01:21:03 -07004465}
4466
Theodore Ts'o819c4922013-04-03 12:47:17 -04004467void ext4_ext_truncate(handle_t *handle, struct inode *inode)
Alex Tomasa86c6182006-10-11 01:21:03 -07004468{
Alex Tomasa86c6182006-10-11 01:21:03 -07004469 struct super_block *sb = inode->i_sb;
Aneesh Kumar K.V725d26d2008-01-28 23:58:27 -05004470 ext4_lblk_t last_block;
Alex Tomasa86c6182006-10-11 01:21:03 -07004471 int err = 0;
4472
4473 /*
Randy Dunlapd0d856e2006-10-11 01:21:07 -07004474 * TODO: optimization is possible here.
4475 * Probably we need not scan at all,
4476 * because page truncation is enough.
Alex Tomasa86c6182006-10-11 01:21:03 -07004477 */
Alex Tomasa86c6182006-10-11 01:21:03 -07004478
4479 /* we have to know where to truncate from in crash case */
4480 EXT4_I(inode)->i_disksize = inode->i_size;
4481 ext4_mark_inode_dirty(handle, inode);
4482
4483 last_block = (inode->i_size + sb->s_blocksize - 1)
4484 >> EXT4_BLOCK_SIZE_BITS(sb);
Theodore Ts'o8acd5e92013-07-15 00:09:19 -04004485retry:
Zheng Liu51865fd2012-11-08 21:57:32 -05004486 err = ext4_es_remove_extent(inode, last_block,
4487 EXT_MAX_BLOCKS - last_block);
Theodore Ts'o94eec0f2013-07-29 12:12:56 -04004488 if (err == -ENOMEM) {
Theodore Ts'o8acd5e92013-07-15 00:09:19 -04004489 cond_resched();
4490 congestion_wait(BLK_RW_ASYNC, HZ/50);
4491 goto retry;
4492 }
4493 if (err) {
4494 ext4_std_error(inode->i_sb, err);
4495 return;
4496 }
Lukas Czerner5f95d212012-03-19 23:03:19 -04004497 err = ext4_ext_remove_space(inode, last_block, EXT_MAX_BLOCKS - 1);
Theodore Ts'o8acd5e92013-07-15 00:09:19 -04004498 ext4_std_error(inode->i_sb, err);
Alex Tomasa86c6182006-10-11 01:21:03 -07004499}
4500
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004501static void ext4_falloc_update_inode(struct inode *inode,
4502 int mode, loff_t new_size, int update_ctime)
4503{
4504 struct timespec now;
4505
4506 if (update_ctime) {
4507 now = current_fs_time(inode->i_sb);
4508 if (!timespec_equal(&inode->i_ctime, &now))
4509 inode->i_ctime = now;
4510 }
4511 /*
4512 * Update only when preallocation was requested beyond
4513 * the file size.
4514 */
Aneesh Kumar K.Vcf17fea2008-09-13 13:06:18 -04004515 if (!(mode & FALLOC_FL_KEEP_SIZE)) {
4516 if (new_size > i_size_read(inode))
4517 i_size_write(inode, new_size);
4518 if (new_size > EXT4_I(inode)->i_disksize)
4519 ext4_update_i_disksize(inode, new_size);
Jiaying Zhangc8d46e42010-02-24 09:52:53 -05004520 } else {
4521 /*
4522 * Mark that we allocate beyond EOF so the subsequent truncate
4523 * can proceed even if the new size is the same as i_size.
4524 */
4525 if (new_size > i_size_read(inode))
Dmitry Monakhov12e9b892010-05-16 22:00:00 -04004526 ext4_set_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004527 }
4528
4529}
4530
Amit Aroraa2df2a62007-07-17 21:42:41 -04004531/*
Christoph Hellwig2fe17c12011-01-14 13:07:43 +01004532 * preallocate space for a file. This implements ext4's fallocate file
Amit Aroraa2df2a62007-07-17 21:42:41 -04004533 * operation, which gets called from sys_fallocate system call.
4534 * For block-mapped files, posix_fallocate should fall back to the method
4535 * of writing zeroes to the required new blocks (the same behavior which is
4536 * expected for file systems which do not support fallocate() system call).
4537 */
Christoph Hellwig2fe17c12011-01-14 13:07:43 +01004538long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
Amit Aroraa2df2a62007-07-17 21:42:41 -04004539{
Al Viro496ad9a2013-01-23 17:07:38 -05004540 struct inode *inode = file_inode(file);
Amit Aroraa2df2a62007-07-17 21:42:41 -04004541 handle_t *handle;
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004542 loff_t new_size;
Theodore Ts'o498e5f22008-11-05 00:14:04 -05004543 unsigned int max_blocks;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004544 int ret = 0;
4545 int ret2 = 0;
4546 int retries = 0;
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04004547 int flags;
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004548 struct ext4_map_blocks map;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004549 unsigned int credits, blkbits = inode->i_blkbits;
4550
Allison Hendersona4bb6b62011-05-25 07:41:50 -04004551 /* Return error if mode is not supported */
4552 if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
4553 return -EOPNOTSUPP;
4554
4555 if (mode & FALLOC_FL_PUNCH_HOLE)
Ashish Sangwanaeb28172013-07-01 08:12:38 -04004556 return ext4_punch_hole(inode, offset, len);
Allison Hendersona4bb6b62011-05-25 07:41:50 -04004557
Tao Ma0c8d4142012-12-10 14:06:03 -05004558 ret = ext4_convert_inline_data(inode);
4559 if (ret)
4560 return ret;
4561
Zheng Liu8bad6fc2013-01-28 09:21:37 -05004562 /*
4563 * currently supporting (pre)allocate mode for extent-based
4564 * files _only_
4565 */
4566 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
4567 return -EOPNOTSUPP;
4568
Jiaying Zhang0562e0b2011-03-21 21:38:05 -04004569 trace_ext4_fallocate_enter(inode, offset, len, mode);
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004570 map.m_lblk = offset >> blkbits;
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004571 /*
4572 * We can't just convert len to max_blocks because
4573 * If blocksize = 4096 offset = 3072 and len = 2048
4574 */
Amit Aroraa2df2a62007-07-17 21:42:41 -04004575 max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004576 - map.m_lblk;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004577 /*
Mingming Caof3bd1f32008-08-19 22:16:03 -04004578 * credits to insert 1 extent into extent tree
Amit Aroraa2df2a62007-07-17 21:42:41 -04004579 */
Mingming Caof3bd1f32008-08-19 22:16:03 -04004580 credits = ext4_chunk_trans_blocks(inode, max_blocks);
Aneesh Kumar K.V55bd7252008-02-15 12:47:21 -05004581 mutex_lock(&inode->i_mutex);
Nikanth Karthikesan6d19c422010-05-16 14:00:00 -04004582 ret = inode_newsize_ok(inode, (len + offset));
4583 if (ret) {
4584 mutex_unlock(&inode->i_mutex);
Jiaying Zhang0562e0b2011-03-21 21:38:05 -04004585 trace_ext4_fallocate_exit(inode, offset, max_blocks, ret);
Nikanth Karthikesan6d19c422010-05-16 14:00:00 -04004586 return ret;
4587 }
Greg Harm3c6fe772011-10-31 18:41:47 -04004588 flags = EXT4_GET_BLOCKS_CREATE_UNINIT_EXT;
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04004589 if (mode & FALLOC_FL_KEEP_SIZE)
4590 flags |= EXT4_GET_BLOCKS_KEEP_SIZE;
Greg Harm3c6fe772011-10-31 18:41:47 -04004591 /*
4592 * Don't normalize the request if it can fit in one extent so
4593 * that it doesn't get unnecessarily split into multiple
4594 * extents.
4595 */
4596 if (len <= EXT_UNINIT_MAX_LEN << blkbits)
4597 flags |= EXT4_GET_BLOCKS_NO_NORMALIZE;
Dmitry Monakhov60d46162012-10-05 11:32:02 -04004598
Amit Aroraa2df2a62007-07-17 21:42:41 -04004599retry:
4600 while (ret >= 0 && ret < max_blocks) {
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004601 map.m_lblk = map.m_lblk + ret;
4602 map.m_len = max_blocks = max_blocks - ret;
Theodore Ts'o9924a922013-02-08 21:59:22 -05004603 handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS,
4604 credits);
Amit Aroraa2df2a62007-07-17 21:42:41 -04004605 if (IS_ERR(handle)) {
4606 ret = PTR_ERR(handle);
4607 break;
4608 }
Dmitry Monakhova4e5d882011-10-25 08:15:12 -04004609 ret = ext4_map_blocks(handle, inode, &map, flags);
Aneesh Kumar K.V221879c2008-01-28 23:58:27 -05004610 if (ret <= 0) {
Aneesh Kumar K.V2c986152008-02-25 15:41:35 -05004611#ifdef EXT4FS_DEBUG
Lukas Czernerb06acd32013-01-28 21:21:12 -05004612 ext4_warning(inode->i_sb,
4613 "inode #%lu: block %u: len %u: "
4614 "ext4_ext_map_blocks returned %d",
4615 inode->i_ino, map.m_lblk,
4616 map.m_len, ret);
Aneesh Kumar K.V2c986152008-02-25 15:41:35 -05004617#endif
Amit Aroraa2df2a62007-07-17 21:42:41 -04004618 ext4_mark_inode_dirty(handle, inode);
4619 ret2 = ext4_journal_stop(handle);
4620 break;
4621 }
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004622 if ((map.m_lblk + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004623 blkbits) >> blkbits))
4624 new_size = offset + len;
4625 else
Utako Kusaka29ae07b2011-07-27 22:11:20 -04004626 new_size = ((loff_t) map.m_lblk + ret) << blkbits;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004627
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004628 ext4_falloc_update_inode(inode, mode, new_size,
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004629 (map.m_flags & EXT4_MAP_NEW));
Amit Aroraa2df2a62007-07-17 21:42:41 -04004630 ext4_mark_inode_dirty(handle, inode);
Zheng Liuf4e95b32012-06-30 19:12:57 -04004631 if ((file->f_flags & O_SYNC) && ret >= max_blocks)
4632 ext4_handle_sync(handle);
Amit Aroraa2df2a62007-07-17 21:42:41 -04004633 ret2 = ext4_journal_stop(handle);
4634 if (ret2)
4635 break;
4636 }
Aneesh Kumar K.Vfd287842008-04-29 08:11:12 -04004637 if (ret == -ENOSPC &&
4638 ext4_should_retry_alloc(inode->i_sb, &retries)) {
4639 ret = 0;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004640 goto retry;
Amit Aroraa2df2a62007-07-17 21:42:41 -04004641 }
Aneesh Kumar K.V55bd7252008-02-15 12:47:21 -05004642 mutex_unlock(&inode->i_mutex);
Jiaying Zhang0562e0b2011-03-21 21:38:05 -04004643 trace_ext4_fallocate_exit(inode, offset, max_blocks,
4644 ret > 0 ? ret2 : ret);
Amit Aroraa2df2a62007-07-17 21:42:41 -04004645 return ret > 0 ? ret2 : ret;
4646}
Eric Sandeen6873fa02008-10-07 00:46:36 -04004647
4648/*
Mingming Cao00314622009-09-28 15:49:08 -04004649 * This function convert a range of blocks to written extents
4650 * The caller of this function will pass the start offset and the size.
4651 * all unwritten extents within this range will be converted to
4652 * written extents.
4653 *
4654 * This function is called from the direct IO end io call back
4655 * function, to convert the fallocated extents after IO is completed.
Mingming109f5562009-11-10 10:48:08 -05004656 * Returns 0 on success.
Mingming Cao00314622009-09-28 15:49:08 -04004657 */
Jan Kara6b523df2013-06-04 13:21:11 -04004658int ext4_convert_unwritten_extents(handle_t *handle, struct inode *inode,
4659 loff_t offset, ssize_t len)
Mingming Cao00314622009-09-28 15:49:08 -04004660{
Mingming Cao00314622009-09-28 15:49:08 -04004661 unsigned int max_blocks;
4662 int ret = 0;
4663 int ret2 = 0;
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004664 struct ext4_map_blocks map;
Mingming Cao00314622009-09-28 15:49:08 -04004665 unsigned int credits, blkbits = inode->i_blkbits;
4666
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004667 map.m_lblk = offset >> blkbits;
Mingming Cao00314622009-09-28 15:49:08 -04004668 /*
4669 * We can't just convert len to max_blocks because
4670 * If blocksize = 4096 offset = 3072 and len = 2048
4671 */
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004672 max_blocks = ((EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits) -
4673 map.m_lblk);
Mingming Cao00314622009-09-28 15:49:08 -04004674 /*
Jan Kara6b523df2013-06-04 13:21:11 -04004675 * This is somewhat ugly but the idea is clear: When transaction is
4676 * reserved, everything goes into it. Otherwise we rather start several
4677 * smaller transactions for conversion of each extent separately.
Mingming Cao00314622009-09-28 15:49:08 -04004678 */
Jan Kara6b523df2013-06-04 13:21:11 -04004679 if (handle) {
4680 handle = ext4_journal_start_reserved(handle,
4681 EXT4_HT_EXT_CONVERT);
4682 if (IS_ERR(handle))
4683 return PTR_ERR(handle);
4684 credits = 0;
4685 } else {
4686 /*
4687 * credits to insert 1 extent into extent tree
4688 */
4689 credits = ext4_chunk_trans_blocks(inode, max_blocks);
4690 }
Mingming Cao00314622009-09-28 15:49:08 -04004691 while (ret >= 0 && ret < max_blocks) {
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004692 map.m_lblk += ret;
4693 map.m_len = (max_blocks -= ret);
Jan Kara6b523df2013-06-04 13:21:11 -04004694 if (credits) {
4695 handle = ext4_journal_start(inode, EXT4_HT_MAP_BLOCKS,
4696 credits);
4697 if (IS_ERR(handle)) {
4698 ret = PTR_ERR(handle);
4699 break;
4700 }
Mingming Cao00314622009-09-28 15:49:08 -04004701 }
Theodore Ts'o2ed88682010-05-16 20:00:00 -04004702 ret = ext4_map_blocks(handle, inode, &map,
Jiaying Zhangc7064ef2010-03-02 13:28:44 -05004703 EXT4_GET_BLOCKS_IO_CONVERT_EXT);
Lukas Czernerb06acd32013-01-28 21:21:12 -05004704 if (ret <= 0)
4705 ext4_warning(inode->i_sb,
4706 "inode #%lu: block %u: len %u: "
4707 "ext4_ext_map_blocks returned %d",
4708 inode->i_ino, map.m_lblk,
4709 map.m_len, ret);
Mingming Cao00314622009-09-28 15:49:08 -04004710 ext4_mark_inode_dirty(handle, inode);
Jan Kara6b523df2013-06-04 13:21:11 -04004711 if (credits)
4712 ret2 = ext4_journal_stop(handle);
4713 if (ret <= 0 || ret2)
Mingming Cao00314622009-09-28 15:49:08 -04004714 break;
4715 }
Jan Kara6b523df2013-06-04 13:21:11 -04004716 if (!credits)
4717 ret2 = ext4_journal_stop(handle);
Mingming Cao00314622009-09-28 15:49:08 -04004718 return ret > 0 ? ret2 : ret;
4719}
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004720
Mingming Cao00314622009-09-28 15:49:08 -04004721/*
Zheng Liu69eb33d2013-02-18 00:31:07 -05004722 * If newes is not existing extent (newes->ec_pblk equals zero) find
4723 * delayed extent at start of newes and update newes accordingly and
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004724 * return start of the next delayed extent.
4725 *
Zheng Liu69eb33d2013-02-18 00:31:07 -05004726 * If newes is existing extent (newes->ec_pblk is not equal zero)
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004727 * return start of next delayed extent or EXT_MAX_BLOCKS if no delayed
Zheng Liu69eb33d2013-02-18 00:31:07 -05004728 * extent found. Leave newes unmodified.
Eric Sandeen6873fa02008-10-07 00:46:36 -04004729 */
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004730static int ext4_find_delayed_extent(struct inode *inode,
Zheng Liu69eb33d2013-02-18 00:31:07 -05004731 struct extent_status *newes)
Eric Sandeen6873fa02008-10-07 00:46:36 -04004732{
Zheng Liub3aff3e2012-11-08 21:57:37 -05004733 struct extent_status es;
Zheng Liube401362013-02-18 00:27:26 -05004734 ext4_lblk_t block, next_del;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004735
Zheng Liu69eb33d2013-02-18 00:31:07 -05004736 if (newes->es_pblk == 0) {
Yan, Zhenge30b5dc2013-05-03 02:15:52 -04004737 ext4_es_find_delayed_extent_range(inode, newes->es_lblk,
4738 newes->es_lblk + newes->es_len - 1, &es);
4739
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004740 /*
Zheng Liu69eb33d2013-02-18 00:31:07 -05004741 * No extent in extent-tree contains block @newes->es_pblk,
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004742 * then the block may stay in 1)a hole or 2)delayed-extent.
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004743 */
Zheng Liu06b0c882013-02-18 00:26:51 -05004744 if (es.es_len == 0)
Zheng Liub3aff3e2012-11-08 21:57:37 -05004745 /* A hole found. */
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004746 return 0;
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004747
Zheng Liu69eb33d2013-02-18 00:31:07 -05004748 if (es.es_lblk > newes->es_lblk) {
Zheng Liub3aff3e2012-11-08 21:57:37 -05004749 /* A hole found. */
Zheng Liu69eb33d2013-02-18 00:31:07 -05004750 newes->es_len = min(es.es_lblk - newes->es_lblk,
4751 newes->es_len);
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004752 return 0;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004753 }
Yongqiang Yang6d9c85e2011-02-27 17:25:47 -05004754
Zheng Liu69eb33d2013-02-18 00:31:07 -05004755 newes->es_len = es.es_lblk + es.es_len - newes->es_lblk;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004756 }
4757
Zheng Liu69eb33d2013-02-18 00:31:07 -05004758 block = newes->es_lblk + newes->es_len;
Yan, Zhenge30b5dc2013-05-03 02:15:52 -04004759 ext4_es_find_delayed_extent_range(inode, block, EXT_MAX_BLOCKS, &es);
Zheng Liube401362013-02-18 00:27:26 -05004760 if (es.es_len == 0)
4761 next_del = EXT_MAX_BLOCKS;
4762 else
4763 next_del = es.es_lblk;
4764
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004765 return next_del;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004766}
Eric Sandeen6873fa02008-10-07 00:46:36 -04004767/* fiemap flags we can handle specified here */
4768#define EXT4_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR)
4769
Aneesh Kumar K.V3a06d772008-11-22 15:04:59 -05004770static int ext4_xattr_fiemap(struct inode *inode,
4771 struct fiemap_extent_info *fieinfo)
Eric Sandeen6873fa02008-10-07 00:46:36 -04004772{
4773 __u64 physical = 0;
4774 __u64 length;
4775 __u32 flags = FIEMAP_EXTENT_LAST;
4776 int blockbits = inode->i_sb->s_blocksize_bits;
4777 int error = 0;
4778
4779 /* in-inode? */
Theodore Ts'o19f5fb72010-01-24 14:34:07 -05004780 if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
Eric Sandeen6873fa02008-10-07 00:46:36 -04004781 struct ext4_iloc iloc;
4782 int offset; /* offset of xattr in inode */
4783
4784 error = ext4_get_inode_loc(inode, &iloc);
4785 if (error)
4786 return error;
Jan Karaa60697f2013-05-31 19:38:56 -04004787 physical = (__u64)iloc.bh->b_blocknr << blockbits;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004788 offset = EXT4_GOOD_OLD_INODE_SIZE +
4789 EXT4_I(inode)->i_extra_isize;
4790 physical += offset;
4791 length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
4792 flags |= FIEMAP_EXTENT_DATA_INLINE;
Curt Wohlgemuthfd2dd9f2010-04-03 17:44:16 -04004793 brelse(iloc.bh);
Eric Sandeen6873fa02008-10-07 00:46:36 -04004794 } else { /* external block */
Jan Karaa60697f2013-05-31 19:38:56 -04004795 physical = (__u64)EXT4_I(inode)->i_file_acl << blockbits;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004796 length = inode->i_sb->s_blocksize;
4797 }
4798
4799 if (physical)
4800 error = fiemap_fill_next_extent(fieinfo, 0, physical,
4801 length, flags);
4802 return (error < 0 ? error : 0);
4803}
4804
4805int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
4806 __u64 start, __u64 len)
4807{
4808 ext4_lblk_t start_blk;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004809 int error = 0;
4810
Tao Ma94191982012-12-10 14:06:02 -05004811 if (ext4_has_inline_data(inode)) {
4812 int has_inline = 1;
4813
4814 error = ext4_inline_data_fiemap(inode, fieinfo, &has_inline);
4815
4816 if (has_inline)
4817 return error;
4818 }
4819
Theodore Ts'o7869a4a2013-08-16 22:05:14 -04004820 if (fieinfo->fi_flags & FIEMAP_FLAG_CACHE) {
4821 error = ext4_ext_precache(inode);
4822 if (error)
4823 return error;
4824 }
4825
Eric Sandeen6873fa02008-10-07 00:46:36 -04004826 /* fallback to generic here if not in extents fmt */
Dmitry Monakhov12e9b892010-05-16 22:00:00 -04004827 if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
Eric Sandeen6873fa02008-10-07 00:46:36 -04004828 return generic_block_fiemap(inode, fieinfo, start, len,
4829 ext4_get_block);
4830
4831 if (fiemap_check_flags(fieinfo, EXT4_FIEMAP_FLAGS))
4832 return -EBADR;
4833
4834 if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
4835 error = ext4_xattr_fiemap(inode, fieinfo);
4836 } else {
Leonard Michlmayraca92ff2010-03-04 17:07:28 -05004837 ext4_lblk_t len_blks;
4838 __u64 last_blk;
4839
Eric Sandeen6873fa02008-10-07 00:46:36 -04004840 start_blk = start >> inode->i_sb->s_blocksize_bits;
Leonard Michlmayraca92ff2010-03-04 17:07:28 -05004841 last_blk = (start + len - 1) >> inode->i_sb->s_blocksize_bits;
Lukas Czernerf17722f2011-06-06 00:05:17 -04004842 if (last_blk >= EXT_MAX_BLOCKS)
4843 last_blk = EXT_MAX_BLOCKS-1;
Leonard Michlmayraca92ff2010-03-04 17:07:28 -05004844 len_blks = ((ext4_lblk_t) last_blk) - start_blk + 1;
Eric Sandeen6873fa02008-10-07 00:46:36 -04004845
4846 /*
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004847 * Walk the extent tree gathering extent information
4848 * and pushing extents back to the user.
Eric Sandeen6873fa02008-10-07 00:46:36 -04004849 */
Lukas Czerner91dd8c12012-11-28 12:32:26 -05004850 error = ext4_fill_fiemap_extents(inode, start_blk,
4851 len_blks, fieinfo);
Eric Sandeen6873fa02008-10-07 00:46:36 -04004852 }
Theodore Ts'o107a7bd2013-08-16 21:23:41 -04004853 ext4_es_lru_add(inode);
Eric Sandeen6873fa02008-10-07 00:46:36 -04004854 return error;
4855}