extents.c 75.9 KB
Newer Older
A
Alex Tomas 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
/*
 * Copyright (c) 2003-2006, Cluster File Systems, Inc, info@clusterfs.com
 * Written by Alex Tomas <alex@clusterfs.com>
 *
 * Architecture independence:
 *   Copyright (c) 2005, Bull S.A.
 *   Written by Pierre Peiffer <pierre.peiffer@bull.net>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public Licens
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-
 */

/*
 * Extents support for EXT4
 *
 * TODO:
 *   - ext4*_error() should be used in some situations
 *   - analyze all BUG()/BUG_ON(), use -EIO where appropriate
 *   - smart tree reduction
 */

#include <linux/module.h>
#include <linux/fs.h>
#include <linux/time.h>
35
#include <linux/jbd2.h>
A
Alex Tomas 已提交
36 37 38 39 40
#include <linux/highuid.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
#include <linux/string.h>
#include <linux/slab.h>
A
Amit Arora 已提交
41
#include <linux/falloc.h>
A
Alex Tomas 已提交
42
#include <asm/uaccess.h>
43 44
#include "ext4_jbd2.h"
#include "ext4_extents.h"
A
Alex Tomas 已提交
45 46


47 48 49 50
/*
 * ext_pblock:
 * combine low and high parts of physical block number into ext4_fsblk_t
 */
51
static ext4_fsblk_t ext_pblock(struct ext4_extent *ex)
52 53 54
{
	ext4_fsblk_t block;

55
	block = le32_to_cpu(ex->ee_start_lo);
56
	block |= ((ext4_fsblk_t) le16_to_cpu(ex->ee_start_hi) << 31) << 1;
57 58 59
	return block;
}

60 61 62 63
/*
 * idx_pblock:
 * combine low and high parts of a leaf physical block number into ext4_fsblk_t
 */
64
ext4_fsblk_t idx_pblock(struct ext4_extent_idx *ix)
65 66 67
{
	ext4_fsblk_t block;

68
	block = le32_to_cpu(ix->ei_leaf_lo);
69
	block |= ((ext4_fsblk_t) le16_to_cpu(ix->ei_leaf_hi) << 31) << 1;
70 71 72
	return block;
}

73 74 75 76 77
/*
 * ext4_ext_store_pblock:
 * stores a large physical block number into an extent struct,
 * breaking it into parts
 */
78
void ext4_ext_store_pblock(struct ext4_extent *ex, ext4_fsblk_t pb)
79
{
80
	ex->ee_start_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff));
81
	ex->ee_start_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
82 83
}

84 85 86 87 88
/*
 * ext4_idx_store_pblock:
 * stores a large physical block number into an index struct,
 * breaking it into parts
 */
89
static void ext4_idx_store_pblock(struct ext4_extent_idx *ix, ext4_fsblk_t pb)
90
{
91
	ix->ei_leaf_lo = cpu_to_le32((unsigned long) (pb & 0xffffffff));
92
	ix->ei_leaf_hi = cpu_to_le16((unsigned long) ((pb >> 31) >> 1) & 0xffff);
93 94
}

95
static int ext4_ext_journal_restart(handle_t *handle, int needed)
A
Alex Tomas 已提交
96 97 98 99
{
	int err;

	if (handle->h_buffer_credits > needed)
100 101
		return 0;
	err = ext4_journal_extend(handle, needed);
102
	if (err <= 0)
103 104
		return err;
	return ext4_journal_restart(handle, needed);
A
Alex Tomas 已提交
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
}

/*
 * could return:
 *  - EROFS
 *  - ENOMEM
 */
static int ext4_ext_get_access(handle_t *handle, struct inode *inode,
				struct ext4_ext_path *path)
{
	if (path->p_bh) {
		/* path points to block */
		return ext4_journal_get_write_access(handle, path->p_bh);
	}
	/* path points to leaf/index in inode body */
	/* we use in-core data, no need to protect them */
	return 0;
}

/*
 * could return:
 *  - EROFS
 *  - ENOMEM
 *  - EIO
 */
static int ext4_ext_dirty(handle_t *handle, struct inode *inode,
				struct ext4_ext_path *path)
{
	int err;
	if (path->p_bh) {
		/* path points to block */
		err = ext4_journal_dirty_metadata(handle, path->p_bh);
	} else {
		/* path points to leaf/index in inode body */
		err = ext4_mark_inode_dirty(handle, inode);
	}
	return err;
}

144
static ext4_fsblk_t ext4_ext_find_goal(struct inode *inode,
A
Alex Tomas 已提交
145
			      struct ext4_ext_path *path,
A
Aneesh Kumar K.V 已提交
146
			      ext4_lblk_t block)
A
Alex Tomas 已提交
147 148
{
	struct ext4_inode_info *ei = EXT4_I(inode);
149
	ext4_fsblk_t bg_start;
150
	ext4_fsblk_t last_block;
151
	ext4_grpblk_t colour;
A
Alex Tomas 已提交
152 153 154 155 156 157 158
	int depth;

	if (path) {
		struct ext4_extent *ex;
		depth = path->p_depth;

		/* try to predict block placement */
159 160
		ex = path[depth].p_ext;
		if (ex)
161
			return ext_pblock(ex)+(block-le32_to_cpu(ex->ee_block));
A
Alex Tomas 已提交
162

163 164
		/* it looks like index is empty;
		 * try to find starting block from index itself */
A
Alex Tomas 已提交
165 166 167 168 169 170 171
		if (path[depth].p_bh)
			return path[depth].p_bh->b_blocknr;
	}

	/* OK. use inode's group */
	bg_start = (ei->i_block_group * EXT4_BLOCKS_PER_GROUP(inode->i_sb)) +
		le32_to_cpu(EXT4_SB(inode->i_sb)->s_es->s_first_data_block);
172 173 174 175
	last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;

	if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
		colour = (current->pid % 16) *
A
Alex Tomas 已提交
176
			(EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
177 178
	else
		colour = (current->pid % 16) * ((last_block - bg_start) / 16);
A
Alex Tomas 已提交
179 180 181
	return bg_start + colour + block;
}

A
Aneesh Kumar K.V 已提交
182 183 184
/*
 * Allocation for a meta data block
 */
185
static ext4_fsblk_t
A
Aneesh Kumar K.V 已提交
186
ext4_ext_new_meta_block(handle_t *handle, struct inode *inode,
A
Alex Tomas 已提交
187 188 189
			struct ext4_ext_path *path,
			struct ext4_extent *ex, int *err)
{
190
	ext4_fsblk_t goal, newblock;
A
Alex Tomas 已提交
191 192

	goal = ext4_ext_find_goal(inode, path, le32_to_cpu(ex->ee_block));
193
	newblock = ext4_new_meta_block(handle, inode, goal, err);
A
Alex Tomas 已提交
194 195 196
	return newblock;
}

197
static int ext4_ext_space_block(struct inode *inode)
A
Alex Tomas 已提交
198 199 200 201 202
{
	int size;

	size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
			/ sizeof(struct ext4_extent);
203
#ifdef AGGRESSIVE_TEST
A
Alex Tomas 已提交
204 205 206 207 208 209
	if (size > 6)
		size = 6;
#endif
	return size;
}

210
static int ext4_ext_space_block_idx(struct inode *inode)
A
Alex Tomas 已提交
211 212 213 214 215
{
	int size;

	size = (inode->i_sb->s_blocksize - sizeof(struct ext4_extent_header))
			/ sizeof(struct ext4_extent_idx);
216
#ifdef AGGRESSIVE_TEST
A
Alex Tomas 已提交
217 218 219 220 221 222
	if (size > 5)
		size = 5;
#endif
	return size;
}

223
static int ext4_ext_space_root(struct inode *inode)
A
Alex Tomas 已提交
224 225 226 227 228 229
{
	int size;

	size = sizeof(EXT4_I(inode)->i_data);
	size -= sizeof(struct ext4_extent_header);
	size /= sizeof(struct ext4_extent);
230
#ifdef AGGRESSIVE_TEST
A
Alex Tomas 已提交
231 232 233 234 235 236
	if (size > 3)
		size = 3;
#endif
	return size;
}

237
static int ext4_ext_space_root_idx(struct inode *inode)
A
Alex Tomas 已提交
238 239 240 241 242 243
{
	int size;

	size = sizeof(EXT4_I(inode)->i_data);
	size -= sizeof(struct ext4_extent_header);
	size /= sizeof(struct ext4_extent_idx);
244
#ifdef AGGRESSIVE_TEST
A
Alex Tomas 已提交
245 246 247 248 249 250
	if (size > 4)
		size = 4;
#endif
	return size;
}

251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
/*
 * Calculate the number of metadata blocks needed
 * to allocate @blocks
 * Worse case is one block per extent
 */
int ext4_ext_calc_metadata_amount(struct inode *inode, int blocks)
{
	int lcap, icap, rcap, leafs, idxs, num;
	int newextents = blocks;

	rcap = ext4_ext_space_root_idx(inode);
	lcap = ext4_ext_space_block(inode);
	icap = ext4_ext_space_block_idx(inode);

	/* number of new leaf blocks needed */
	num = leafs = (newextents + lcap - 1) / lcap;

	/*
	 * Worse case, we need separate index block(s)
	 * to link all new leaf blocks
	 */
	idxs = (leafs + icap - 1) / icap;
	do {
		num += idxs;
		idxs = (idxs + icap - 1) / icap;
	} while (idxs > rcap);

	return num;
}

281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
static int
ext4_ext_max_entries(struct inode *inode, int depth)
{
	int max;

	if (depth == ext_depth(inode)) {
		if (depth == 0)
			max = ext4_ext_space_root(inode);
		else
			max = ext4_ext_space_root_idx(inode);
	} else {
		if (depth == 0)
			max = ext4_ext_space_block(inode);
		else
			max = ext4_ext_space_block_idx(inode);
	}

	return max;
}

static int __ext4_ext_check_header(const char *function, struct inode *inode,
					struct ext4_extent_header *eh,
					int depth)
{
	const char *error_msg;
	int max = 0;

	if (unlikely(eh->eh_magic != EXT4_EXT_MAGIC)) {
		error_msg = "invalid magic";
		goto corrupted;
	}
	if (unlikely(le16_to_cpu(eh->eh_depth) != depth)) {
		error_msg = "unexpected eh_depth";
		goto corrupted;
	}
	if (unlikely(eh->eh_max == 0)) {
		error_msg = "invalid eh_max";
		goto corrupted;
	}
	max = ext4_ext_max_entries(inode, depth);
	if (unlikely(le16_to_cpu(eh->eh_max) > max)) {
		error_msg = "too large eh_max";
		goto corrupted;
	}
	if (unlikely(le16_to_cpu(eh->eh_entries) > le16_to_cpu(eh->eh_max))) {
		error_msg = "invalid eh_entries";
		goto corrupted;
	}
	return 0;

corrupted:
	ext4_error(inode->i_sb, function,
			"bad header in inode #%lu: %s - magic %x, "
			"entries %u, max %u(%u), depth %u(%u)",
			inode->i_ino, error_msg, le16_to_cpu(eh->eh_magic),
			le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max),
			max, le16_to_cpu(eh->eh_depth), depth);

	return -EIO;
}

#define ext4_ext_check_header(inode, eh, depth)	\
343
	__ext4_ext_check_header(__func__, inode, eh, depth)
344

A
Alex Tomas 已提交
345 346 347 348 349 350 351 352
#ifdef EXT_DEBUG
static void ext4_ext_show_path(struct inode *inode, struct ext4_ext_path *path)
{
	int k, l = path->p_depth;

	ext_debug("path:");
	for (k = 0; k <= l; k++, path++) {
		if (path->p_idx) {
353
		  ext_debug("  %d->%llu", le32_to_cpu(path->p_idx->ei_block),
354
			    idx_pblock(path->p_idx));
A
Alex Tomas 已提交
355
		} else if (path->p_ext) {
356
			ext_debug("  %d:%d:%llu ",
A
Alex Tomas 已提交
357
				  le32_to_cpu(path->p_ext->ee_block),
A
Amit Arora 已提交
358
				  ext4_ext_get_actual_len(path->p_ext),
359
				  ext_pblock(path->p_ext));
A
Alex Tomas 已提交
360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379
		} else
			ext_debug("  []");
	}
	ext_debug("\n");
}

static void ext4_ext_show_leaf(struct inode *inode, struct ext4_ext_path *path)
{
	int depth = ext_depth(inode);
	struct ext4_extent_header *eh;
	struct ext4_extent *ex;
	int i;

	if (!path)
		return;

	eh = path[depth].p_hdr;
	ex = EXT_FIRST_EXTENT(eh);

	for (i = 0; i < le16_to_cpu(eh->eh_entries); i++, ex++) {
380
		ext_debug("%d:%d:%llu ", le32_to_cpu(ex->ee_block),
A
Amit Arora 已提交
381
			  ext4_ext_get_actual_len(ex), ext_pblock(ex));
A
Alex Tomas 已提交
382 383 384 385 386 387 388 389
	}
	ext_debug("\n");
}
#else
#define ext4_ext_show_path(inode,path)
#define ext4_ext_show_leaf(inode,path)
#endif

390
void ext4_ext_drop_refs(struct ext4_ext_path *path)
A
Alex Tomas 已提交
391 392 393 394 395 396 397 398 399 400 401 402
{
	int depth = path->p_depth;
	int i;

	for (i = 0; i <= depth; i++, path++)
		if (path->p_bh) {
			brelse(path->p_bh);
			path->p_bh = NULL;
		}
}

/*
403 404
 * ext4_ext_binsearch_idx:
 * binary search for the closest index of the given block
405
 * the header must be checked before calling this
A
Alex Tomas 已提交
406 407
 */
static void
A
Aneesh Kumar K.V 已提交
408 409
ext4_ext_binsearch_idx(struct inode *inode,
			struct ext4_ext_path *path, ext4_lblk_t block)
A
Alex Tomas 已提交
410 411 412 413 414
{
	struct ext4_extent_header *eh = path->p_hdr;
	struct ext4_extent_idx *r, *l, *m;


415
	ext_debug("binsearch for %u(idx):  ", block);
A
Alex Tomas 已提交
416 417

	l = EXT_FIRST_INDEX(eh) + 1;
D
Dmitry Monakhov 已提交
418
	r = EXT_LAST_INDEX(eh);
A
Alex Tomas 已提交
419 420 421 422 423 424
	while (l <= r) {
		m = l + (r - l) / 2;
		if (block < le32_to_cpu(m->ei_block))
			r = m - 1;
		else
			l = m + 1;
425 426 427
		ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ei_block),
				m, le32_to_cpu(m->ei_block),
				r, le32_to_cpu(r->ei_block));
A
Alex Tomas 已提交
428 429 430
	}

	path->p_idx = l - 1;
431
	ext_debug("  -> %d->%lld ", le32_to_cpu(path->p_idx->ei_block),
432
		  idx_pblock(path->p_idx));
A
Alex Tomas 已提交
433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449

#ifdef CHECK_BINSEARCH
	{
		struct ext4_extent_idx *chix, *ix;
		int k;

		chix = ix = EXT_FIRST_INDEX(eh);
		for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ix++) {
		  if (k != 0 &&
		      le32_to_cpu(ix->ei_block) <= le32_to_cpu(ix[-1].ei_block)) {
				printk("k=%d, ix=0x%p, first=0x%p\n", k,
					ix, EXT_FIRST_INDEX(eh));
				printk("%u <= %u\n",
				       le32_to_cpu(ix->ei_block),
				       le32_to_cpu(ix[-1].ei_block));
			}
			BUG_ON(k && le32_to_cpu(ix->ei_block)
D
Dave Kleikamp 已提交
450
					   <= le32_to_cpu(ix[-1].ei_block));
A
Alex Tomas 已提交
451 452 453 454 455 456 457 458 459 460 461
			if (block < le32_to_cpu(ix->ei_block))
				break;
			chix = ix;
		}
		BUG_ON(chix != path->p_idx);
	}
#endif

}

/*
462 463
 * ext4_ext_binsearch:
 * binary search for closest extent of the given block
464
 * the header must be checked before calling this
A
Alex Tomas 已提交
465 466
 */
static void
A
Aneesh Kumar K.V 已提交
467 468
ext4_ext_binsearch(struct inode *inode,
		struct ext4_ext_path *path, ext4_lblk_t block)
A
Alex Tomas 已提交
469 470 471 472 473 474
{
	struct ext4_extent_header *eh = path->p_hdr;
	struct ext4_extent *r, *l, *m;

	if (eh->eh_entries == 0) {
		/*
475 476
		 * this leaf is empty:
		 * we get such a leaf in split/add case
A
Alex Tomas 已提交
477 478 479 480
		 */
		return;
	}

481
	ext_debug("binsearch for %u:  ", block);
A
Alex Tomas 已提交
482 483

	l = EXT_FIRST_EXTENT(eh) + 1;
D
Dmitry Monakhov 已提交
484
	r = EXT_LAST_EXTENT(eh);
A
Alex Tomas 已提交
485 486 487 488 489 490 491

	while (l <= r) {
		m = l + (r - l) / 2;
		if (block < le32_to_cpu(m->ee_block))
			r = m - 1;
		else
			l = m + 1;
492 493 494
		ext_debug("%p(%u):%p(%u):%p(%u) ", l, le32_to_cpu(l->ee_block),
				m, le32_to_cpu(m->ee_block),
				r, le32_to_cpu(r->ee_block));
A
Alex Tomas 已提交
495 496 497
	}

	path->p_ext = l - 1;
498
	ext_debug("  -> %d:%llu:%d ",
D
Dave Kleikamp 已提交
499 500
			le32_to_cpu(path->p_ext->ee_block),
			ext_pblock(path->p_ext),
A
Amit Arora 已提交
501
			ext4_ext_get_actual_len(path->p_ext));
A
Alex Tomas 已提交
502 503 504 505 506 507 508 509 510

#ifdef CHECK_BINSEARCH
	{
		struct ext4_extent *chex, *ex;
		int k;

		chex = ex = EXT_FIRST_EXTENT(eh);
		for (k = 0; k < le16_to_cpu(eh->eh_entries); k++, ex++) {
			BUG_ON(k && le32_to_cpu(ex->ee_block)
D
Dave Kleikamp 已提交
511
					  <= le32_to_cpu(ex[-1].ee_block));
A
Alex Tomas 已提交
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
			if (block < le32_to_cpu(ex->ee_block))
				break;
			chex = ex;
		}
		BUG_ON(chex != path->p_ext);
	}
#endif

}

int ext4_ext_tree_init(handle_t *handle, struct inode *inode)
{
	struct ext4_extent_header *eh;

	eh = ext_inode_hdr(inode);
	eh->eh_depth = 0;
	eh->eh_entries = 0;
	eh->eh_magic = EXT4_EXT_MAGIC;
	eh->eh_max = cpu_to_le16(ext4_ext_space_root(inode));
	ext4_mark_inode_dirty(handle, inode);
	ext4_ext_invalidate_cache(inode);
	return 0;
}

struct ext4_ext_path *
A
Aneesh Kumar K.V 已提交
537 538
ext4_ext_find_extent(struct inode *inode, ext4_lblk_t block,
					struct ext4_ext_path *path)
A
Alex Tomas 已提交
539 540 541 542 543 544
{
	struct ext4_extent_header *eh;
	struct buffer_head *bh;
	short int depth, i, ppos = 0, alloc = 0;

	eh = ext_inode_hdr(inode);
545 546
	depth = ext_depth(inode);
	if (ext4_ext_check_header(inode, eh, depth))
A
Alex Tomas 已提交
547 548 549 550 551
		return ERR_PTR(-EIO);


	/* account possible depth increase */
	if (!path) {
A
Avantika Mathur 已提交
552
		path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 2),
A
Alex Tomas 已提交
553 554 555 556 557 558
				GFP_NOFS);
		if (!path)
			return ERR_PTR(-ENOMEM);
		alloc = 1;
	}
	path[0].p_hdr = eh;
559
	path[0].p_bh = NULL;
A
Alex Tomas 已提交
560

561
	i = depth;
A
Alex Tomas 已提交
562 563 564 565
	/* walk through the tree */
	while (i) {
		ext_debug("depth %d: num %d, max %d\n",
			  ppos, le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
566

A
Alex Tomas 已提交
567
		ext4_ext_binsearch_idx(inode, path + ppos, block);
568
		path[ppos].p_block = idx_pblock(path[ppos].p_idx);
A
Alex Tomas 已提交
569 570 571 572 573 574 575 576 577 578 579 580 581 582
		path[ppos].p_depth = i;
		path[ppos].p_ext = NULL;

		bh = sb_bread(inode->i_sb, path[ppos].p_block);
		if (!bh)
			goto err;

		eh = ext_block_hdr(bh);
		ppos++;
		BUG_ON(ppos > depth);
		path[ppos].p_bh = bh;
		path[ppos].p_hdr = eh;
		i--;

583
		if (ext4_ext_check_header(inode, eh, i))
A
Alex Tomas 已提交
584 585 586 587 588 589 590 591 592
			goto err;
	}

	path[ppos].p_depth = i;
	path[ppos].p_ext = NULL;
	path[ppos].p_idx = NULL;

	/* find extent */
	ext4_ext_binsearch(inode, path + ppos, block);
593 594 595
	/* if not an empty leaf */
	if (path[ppos].p_ext)
		path[ppos].p_block = ext_pblock(path[ppos].p_ext);
A
Alex Tomas 已提交
596 597 598 599 600 601 602 603 604 605 606 607 608

	ext4_ext_show_path(inode, path);

	return path;

err:
	ext4_ext_drop_refs(path);
	if (alloc)
		kfree(path);
	return ERR_PTR(-EIO);
}

/*
609 610 611
 * ext4_ext_insert_index:
 * insert new index [@logical;@ptr] into the block at @curp;
 * check where to insert: before @curp or after @curp
A
Alex Tomas 已提交
612 613 614
 */
static int ext4_ext_insert_index(handle_t *handle, struct inode *inode,
				struct ext4_ext_path *curp,
615
				int logical, ext4_fsblk_t ptr)
A
Alex Tomas 已提交
616 617 618 619
{
	struct ext4_extent_idx *ix;
	int len, err;

620 621
	err = ext4_ext_get_access(handle, inode, curp);
	if (err)
A
Alex Tomas 已提交
622 623 624 625 626 627 628 629 630
		return err;

	BUG_ON(logical == le32_to_cpu(curp->p_idx->ei_block));
	len = EXT_MAX_INDEX(curp->p_hdr) - curp->p_idx;
	if (logical > le32_to_cpu(curp->p_idx->ei_block)) {
		/* insert after */
		if (curp->p_idx != EXT_LAST_INDEX(curp->p_hdr)) {
			len = (len - 1) * sizeof(struct ext4_extent_idx);
			len = len < 0 ? 0 : len;
631
			ext_debug("insert new index %d after: %llu. "
A
Alex Tomas 已提交
632 633 634 635 636 637 638 639 640 641
					"move %d from 0x%p to 0x%p\n",
					logical, ptr, len,
					(curp->p_idx + 1), (curp->p_idx + 2));
			memmove(curp->p_idx + 2, curp->p_idx + 1, len);
		}
		ix = curp->p_idx + 1;
	} else {
		/* insert before */
		len = len * sizeof(struct ext4_extent_idx);
		len = len < 0 ? 0 : len;
642
		ext_debug("insert new index %d before: %llu. "
A
Alex Tomas 已提交
643 644 645 646 647 648 649 650
				"move %d from 0x%p to 0x%p\n",
				logical, ptr, len,
				curp->p_idx, (curp->p_idx + 1));
		memmove(curp->p_idx + 1, curp->p_idx, len);
		ix = curp->p_idx;
	}

	ix->ei_block = cpu_to_le32(logical);
651
	ext4_idx_store_pblock(ix, ptr);
M
Marcin Slusarz 已提交
652
	le16_add_cpu(&curp->p_hdr->eh_entries, 1);
A
Alex Tomas 已提交
653 654

	BUG_ON(le16_to_cpu(curp->p_hdr->eh_entries)
D
Dave Kleikamp 已提交
655
			     > le16_to_cpu(curp->p_hdr->eh_max));
A
Alex Tomas 已提交
656 657 658 659 660 661 662 663 664
	BUG_ON(ix > EXT_LAST_INDEX(curp->p_hdr));

	err = ext4_ext_dirty(handle, inode, curp);
	ext4_std_error(inode->i_sb, err);

	return err;
}

/*
665 666 667 668 669 670 671 672
 * ext4_ext_split:
 * inserts new subtree into the path, using free index entry
 * at depth @at:
 * - allocates all needed blocks (new leaf and all intermediate index blocks)
 * - makes decision where to split
 * - moves remaining extents and index entries (right to the split point)
 *   into the newly allocated blocks
 * - initializes subtree
A
Alex Tomas 已提交
673 674 675 676 677 678 679 680 681 682 683
 */
static int ext4_ext_split(handle_t *handle, struct inode *inode,
				struct ext4_ext_path *path,
				struct ext4_extent *newext, int at)
{
	struct buffer_head *bh = NULL;
	int depth = ext_depth(inode);
	struct ext4_extent_header *neh;
	struct ext4_extent_idx *fidx;
	struct ext4_extent *ex;
	int i = at, k, m, a;
684
	ext4_fsblk_t newblock, oldblock;
A
Alex Tomas 已提交
685
	__le32 border;
686
	ext4_fsblk_t *ablocks = NULL; /* array of allocated blocks */
A
Alex Tomas 已提交
687 688 689
	int err = 0;

	/* make decision: where to split? */
690
	/* FIXME: now decision is simplest: at current extent */
A
Alex Tomas 已提交
691

692
	/* if current leaf will be split, then we should use
A
Alex Tomas 已提交
693 694 695 696
	 * border from split point */
	BUG_ON(path[depth].p_ext > EXT_MAX_EXTENT(path[depth].p_hdr));
	if (path[depth].p_ext != EXT_MAX_EXTENT(path[depth].p_hdr)) {
		border = path[depth].p_ext[1].ee_block;
697
		ext_debug("leaf will be split."
A
Alex Tomas 已提交
698
				" next leaf starts at %d\n",
D
Dave Kleikamp 已提交
699
				  le32_to_cpu(border));
A
Alex Tomas 已提交
700 701 702 703
	} else {
		border = newext->ee_block;
		ext_debug("leaf will be added."
				" next leaf starts at %d\n",
D
Dave Kleikamp 已提交
704
				le32_to_cpu(border));
A
Alex Tomas 已提交
705 706 707
	}

	/*
708 709
	 * If error occurs, then we break processing
	 * and mark filesystem read-only. index won't
A
Alex Tomas 已提交
710
	 * be inserted and tree will be in consistent
711
	 * state. Next mount will repair buffers too.
A
Alex Tomas 已提交
712 713 714
	 */

	/*
715 716 717
	 * Get array to track all allocated blocks.
	 * We need this to handle errors and free blocks
	 * upon them.
A
Alex Tomas 已提交
718
	 */
A
Avantika Mathur 已提交
719
	ablocks = kzalloc(sizeof(ext4_fsblk_t) * depth, GFP_NOFS);
A
Alex Tomas 已提交
720 721 722 723 724 725
	if (!ablocks)
		return -ENOMEM;

	/* allocate all needed blocks */
	ext_debug("allocate %d blocks for indexes/leaf\n", depth - at);
	for (a = 0; a < depth - at; a++) {
A
Aneesh Kumar K.V 已提交
726 727
		newblock = ext4_ext_new_meta_block(handle, inode, path,
						   newext, &err);
A
Alex Tomas 已提交
728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
		if (newblock == 0)
			goto cleanup;
		ablocks[a] = newblock;
	}

	/* initialize new leaf */
	newblock = ablocks[--a];
	BUG_ON(newblock == 0);
	bh = sb_getblk(inode->i_sb, newblock);
	if (!bh) {
		err = -EIO;
		goto cleanup;
	}
	lock_buffer(bh);

743 744
	err = ext4_journal_get_create_access(handle, bh);
	if (err)
A
Alex Tomas 已提交
745 746 747 748 749 750 751 752 753
		goto cleanup;

	neh = ext_block_hdr(bh);
	neh->eh_entries = 0;
	neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
	neh->eh_magic = EXT4_EXT_MAGIC;
	neh->eh_depth = 0;
	ex = EXT_FIRST_EXTENT(neh);

754
	/* move remainder of path[depth] to the new leaf */
A
Alex Tomas 已提交
755 756 757 758 759 760 761
	BUG_ON(path[depth].p_hdr->eh_entries != path[depth].p_hdr->eh_max);
	/* start copy from next extent */
	/* TODO: we could do it by single memmove */
	m = 0;
	path[depth].p_ext++;
	while (path[depth].p_ext <=
			EXT_MAX_EXTENT(path[depth].p_hdr)) {
762
		ext_debug("move %d:%llu:%d in new leaf %llu\n",
D
Dave Kleikamp 已提交
763 764
				le32_to_cpu(path[depth].p_ext->ee_block),
				ext_pblock(path[depth].p_ext),
A
Amit Arora 已提交
765
				ext4_ext_get_actual_len(path[depth].p_ext),
A
Alex Tomas 已提交
766 767 768 769 770 771 772 773 774
				newblock);
		/*memmove(ex++, path[depth].p_ext++,
				sizeof(struct ext4_extent));
		neh->eh_entries++;*/
		path[depth].p_ext++;
		m++;
	}
	if (m) {
		memmove(ex, path[depth].p_ext-m, sizeof(struct ext4_extent)*m);
M
Marcin Slusarz 已提交
775
		le16_add_cpu(&neh->eh_entries, m);
A
Alex Tomas 已提交
776 777 778 779 780
	}

	set_buffer_uptodate(bh);
	unlock_buffer(bh);

781 782
	err = ext4_journal_dirty_metadata(handle, bh);
	if (err)
A
Alex Tomas 已提交
783 784 785 786 787 788
		goto cleanup;
	brelse(bh);
	bh = NULL;

	/* correct old leaf */
	if (m) {
789 790
		err = ext4_ext_get_access(handle, inode, path + depth);
		if (err)
A
Alex Tomas 已提交
791
			goto cleanup;
M
Marcin Slusarz 已提交
792
		le16_add_cpu(&path[depth].p_hdr->eh_entries, -m);
793 794
		err = ext4_ext_dirty(handle, inode, path + depth);
		if (err)
A
Alex Tomas 已提交
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809
			goto cleanup;

	}

	/* create intermediate indexes */
	k = depth - at - 1;
	BUG_ON(k < 0);
	if (k)
		ext_debug("create %d intermediate indices\n", k);
	/* insert new index into current index block */
	/* current depth stored in i var */
	i = depth - 1;
	while (k--) {
		oldblock = newblock;
		newblock = ablocks[--a];
810
		bh = sb_getblk(inode->i_sb, newblock);
A
Alex Tomas 已提交
811 812 813 814 815 816
		if (!bh) {
			err = -EIO;
			goto cleanup;
		}
		lock_buffer(bh);

817 818
		err = ext4_journal_get_create_access(handle, bh);
		if (err)
A
Alex Tomas 已提交
819 820 821 822 823 824 825 826 827
			goto cleanup;

		neh = ext_block_hdr(bh);
		neh->eh_entries = cpu_to_le16(1);
		neh->eh_magic = EXT4_EXT_MAGIC;
		neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
		neh->eh_depth = cpu_to_le16(depth - i);
		fidx = EXT_FIRST_INDEX(neh);
		fidx->ei_block = border;
828
		ext4_idx_store_pblock(fidx, oldblock);
A
Alex Tomas 已提交
829

830 831
		ext_debug("int.index at %d (block %llu): %u -> %llu\n",
				i, newblock, le32_to_cpu(border), oldblock);
A
Alex Tomas 已提交
832 833 834 835 836 837 838 839 840
		/* copy indexes */
		m = 0;
		path[i].p_idx++;

		ext_debug("cur 0x%p, last 0x%p\n", path[i].p_idx,
				EXT_MAX_INDEX(path[i].p_hdr));
		BUG_ON(EXT_MAX_INDEX(path[i].p_hdr) !=
				EXT_LAST_INDEX(path[i].p_hdr));
		while (path[i].p_idx <= EXT_MAX_INDEX(path[i].p_hdr)) {
841
			ext_debug("%d: move %d:%llu in new index %llu\n", i,
D
Dave Kleikamp 已提交
842 843 844
					le32_to_cpu(path[i].p_idx->ei_block),
					idx_pblock(path[i].p_idx),
					newblock);
A
Alex Tomas 已提交
845 846 847 848 849 850 851 852 853 854
			/*memmove(++fidx, path[i].p_idx++,
					sizeof(struct ext4_extent_idx));
			neh->eh_entries++;
			BUG_ON(neh->eh_entries > neh->eh_max);*/
			path[i].p_idx++;
			m++;
		}
		if (m) {
			memmove(++fidx, path[i].p_idx - m,
				sizeof(struct ext4_extent_idx) * m);
M
Marcin Slusarz 已提交
855
			le16_add_cpu(&neh->eh_entries, m);
A
Alex Tomas 已提交
856 857 858 859
		}
		set_buffer_uptodate(bh);
		unlock_buffer(bh);

860 861
		err = ext4_journal_dirty_metadata(handle, bh);
		if (err)
A
Alex Tomas 已提交
862 863 864 865 866 867 868 869 870
			goto cleanup;
		brelse(bh);
		bh = NULL;

		/* correct old index */
		if (m) {
			err = ext4_ext_get_access(handle, inode, path + i);
			if (err)
				goto cleanup;
M
Marcin Slusarz 已提交
871
			le16_add_cpu(&path[i].p_hdr->eh_entries, -m);
A
Alex Tomas 已提交
872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895
			err = ext4_ext_dirty(handle, inode, path + i);
			if (err)
				goto cleanup;
		}

		i--;
	}

	/* insert new index */
	err = ext4_ext_insert_index(handle, inode, path + at,
				    le32_to_cpu(border), newblock);

cleanup:
	if (bh) {
		if (buffer_locked(bh))
			unlock_buffer(bh);
		brelse(bh);
	}

	if (err) {
		/* free all allocated blocks in error case */
		for (i = 0; i < depth; i++) {
			if (!ablocks[i])
				continue;
896
			ext4_free_blocks(handle, inode, ablocks[i], 1, 1);
A
Alex Tomas 已提交
897 898 899 900 901 902 903 904
		}
	}
	kfree(ablocks);

	return err;
}

/*
905 906 907 908 909 910
 * ext4_ext_grow_indepth:
 * implements tree growing procedure:
 * - allocates new block
 * - moves top-level data (index block or leaf) into the new block
 * - initializes new top-level, creating index that points to the
 *   just created block
A
Alex Tomas 已提交
911 912 913 914 915 916 917 918 919
 */
static int ext4_ext_grow_indepth(handle_t *handle, struct inode *inode,
					struct ext4_ext_path *path,
					struct ext4_extent *newext)
{
	struct ext4_ext_path *curp = path;
	struct ext4_extent_header *neh;
	struct ext4_extent_idx *fidx;
	struct buffer_head *bh;
920
	ext4_fsblk_t newblock;
A
Alex Tomas 已提交
921 922
	int err = 0;

A
Aneesh Kumar K.V 已提交
923
	newblock = ext4_ext_new_meta_block(handle, inode, path, newext, &err);
A
Alex Tomas 已提交
924 925 926 927 928 929 930 931 932 933 934
	if (newblock == 0)
		return err;

	bh = sb_getblk(inode->i_sb, newblock);
	if (!bh) {
		err = -EIO;
		ext4_std_error(inode->i_sb, err);
		return err;
	}
	lock_buffer(bh);

935 936
	err = ext4_journal_get_create_access(handle, bh);
	if (err) {
A
Alex Tomas 已提交
937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955
		unlock_buffer(bh);
		goto out;
	}

	/* move top-level index/leaf into new block */
	memmove(bh->b_data, curp->p_hdr, sizeof(EXT4_I(inode)->i_data));

	/* set size of new block */
	neh = ext_block_hdr(bh);
	/* old root could have indexes or leaves
	 * so calculate e_max right way */
	if (ext_depth(inode))
	  neh->eh_max = cpu_to_le16(ext4_ext_space_block_idx(inode));
	else
	  neh->eh_max = cpu_to_le16(ext4_ext_space_block(inode));
	neh->eh_magic = EXT4_EXT_MAGIC;
	set_buffer_uptodate(bh);
	unlock_buffer(bh);

956 957
	err = ext4_journal_dirty_metadata(handle, bh);
	if (err)
A
Alex Tomas 已提交
958 959 960
		goto out;

	/* create index in new top-level index: num,max,pointer */
961 962
	err = ext4_ext_get_access(handle, inode, curp);
	if (err)
A
Alex Tomas 已提交
963 964 965 966 967 968
		goto out;

	curp->p_hdr->eh_magic = EXT4_EXT_MAGIC;
	curp->p_hdr->eh_max = cpu_to_le16(ext4_ext_space_root_idx(inode));
	curp->p_hdr->eh_entries = cpu_to_le16(1);
	curp->p_idx = EXT_FIRST_INDEX(curp->p_hdr);
D
Dmitry Monakhov 已提交
969 970 971 972 973 974 975

	if (path[0].p_hdr->eh_depth)
		curp->p_idx->ei_block =
			EXT_FIRST_INDEX(path[0].p_hdr)->ei_block;
	else
		curp->p_idx->ei_block =
			EXT_FIRST_EXTENT(path[0].p_hdr)->ee_block;
976
	ext4_idx_store_pblock(curp->p_idx, newblock);
A
Alex Tomas 已提交
977 978 979

	neh = ext_inode_hdr(inode);
	fidx = EXT_FIRST_INDEX(neh);
980
	ext_debug("new root: num %d(%d), lblock %d, ptr %llu\n",
A
Alex Tomas 已提交
981
		  le16_to_cpu(neh->eh_entries), le16_to_cpu(neh->eh_max),
982
		  le32_to_cpu(fidx->ei_block), idx_pblock(fidx));
A
Alex Tomas 已提交
983 984 985 986 987 988 989 990 991 992

	neh->eh_depth = cpu_to_le16(path->p_depth + 1);
	err = ext4_ext_dirty(handle, inode, curp);
out:
	brelse(bh);

	return err;
}

/*
993 994 995
 * ext4_ext_create_new_leaf:
 * finds empty index and adds new leaf.
 * if no free index is found, then it requests in-depth growing.
A
Alex Tomas 已提交
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
 */
static int ext4_ext_create_new_leaf(handle_t *handle, struct inode *inode,
					struct ext4_ext_path *path,
					struct ext4_extent *newext)
{
	struct ext4_ext_path *curp;
	int depth, i, err = 0;

repeat:
	i = depth = ext_depth(inode);

	/* walk up to the tree and look for free index entry */
	curp = path + depth;
	while (i > 0 && !EXT_HAS_FREE_INDEX(curp)) {
		i--;
		curp--;
	}

1014 1015
	/* we use already allocated block for index block,
	 * so subsequent data blocks should be contiguous */
A
Alex Tomas 已提交
1016 1017 1018 1019
	if (EXT_HAS_FREE_INDEX(curp)) {
		/* if we found index with free entry, then use that
		 * entry: create all needed subtree and add new leaf */
		err = ext4_ext_split(handle, inode, path, newext, i);
1020 1021
		if (err)
			goto out;
A
Alex Tomas 已提交
1022 1023 1024 1025

		/* refill path */
		ext4_ext_drop_refs(path);
		path = ext4_ext_find_extent(inode,
A
Aneesh Kumar K.V 已提交
1026 1027
				    (ext4_lblk_t)le32_to_cpu(newext->ee_block),
				    path);
A
Alex Tomas 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
		if (IS_ERR(path))
			err = PTR_ERR(path);
	} else {
		/* tree is full, time to grow in depth */
		err = ext4_ext_grow_indepth(handle, inode, path, newext);
		if (err)
			goto out;

		/* refill path */
		ext4_ext_drop_refs(path);
		path = ext4_ext_find_extent(inode,
A
Aneesh Kumar K.V 已提交
1039 1040
				   (ext4_lblk_t)le32_to_cpu(newext->ee_block),
				    path);
A
Alex Tomas 已提交
1041 1042 1043 1044 1045 1046
		if (IS_ERR(path)) {
			err = PTR_ERR(path);
			goto out;
		}

		/*
1047 1048
		 * only first (depth 0 -> 1) produces free space;
		 * in all other cases we have to split the grown tree
A
Alex Tomas 已提交
1049 1050 1051
		 */
		depth = ext_depth(inode);
		if (path[depth].p_hdr->eh_entries == path[depth].p_hdr->eh_max) {
1052
			/* now we need to split */
A
Alex Tomas 已提交
1053 1054 1055 1056 1057 1058 1059 1060
			goto repeat;
		}
	}

out:
	return err;
}

1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
/*
 * search the closest allocated block to the left for *logical
 * and returns it at @logical + it's physical address at @phys
 * if *logical is the smallest allocated block, the function
 * returns 0 at @phys
 * return value contains 0 (success) or error code
 */
int
ext4_ext_search_left(struct inode *inode, struct ext4_ext_path *path,
			ext4_lblk_t *logical, ext4_fsblk_t *phys)
{
	struct ext4_extent_idx *ix;
	struct ext4_extent *ex;
1074
	int depth, ee_len;
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087

	BUG_ON(path == NULL);
	depth = path->p_depth;
	*phys = 0;

	if (depth == 0 && path->p_ext == NULL)
		return 0;

	/* usually extent in the path covers blocks smaller
	 * then *logical, but it can be that extent is the
	 * first one in the file */

	ex = path[depth].p_ext;
1088
	ee_len = ext4_ext_get_actual_len(ex);
1089 1090 1091 1092 1093 1094 1095 1096 1097
	if (*logical < le32_to_cpu(ex->ee_block)) {
		BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex);
		while (--depth >= 0) {
			ix = path[depth].p_idx;
			BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr));
		}
		return 0;
	}

1098
	BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len));
1099

1100 1101
	*logical = le32_to_cpu(ex->ee_block) + ee_len - 1;
	*phys = ext_pblock(ex) + ee_len - 1;
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
	return 0;
}

/*
 * search the closest allocated block to the right for *logical
 * and returns it at @logical + it's physical address at @phys
 * if *logical is the smallest allocated block, the function
 * returns 0 at @phys
 * return value contains 0 (success) or error code
 */
int
ext4_ext_search_right(struct inode *inode, struct ext4_ext_path *path,
			ext4_lblk_t *logical, ext4_fsblk_t *phys)
{
	struct buffer_head *bh = NULL;
	struct ext4_extent_header *eh;
	struct ext4_extent_idx *ix;
	struct ext4_extent *ex;
	ext4_fsblk_t block;
1121
	int depth, ee_len;
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134

	BUG_ON(path == NULL);
	depth = path->p_depth;
	*phys = 0;

	if (depth == 0 && path->p_ext == NULL)
		return 0;

	/* usually extent in the path covers blocks smaller
	 * then *logical, but it can be that extent is the
	 * first one in the file */

	ex = path[depth].p_ext;
1135
	ee_len = ext4_ext_get_actual_len(ex);
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146
	if (*logical < le32_to_cpu(ex->ee_block)) {
		BUG_ON(EXT_FIRST_EXTENT(path[depth].p_hdr) != ex);
		while (--depth >= 0) {
			ix = path[depth].p_idx;
			BUG_ON(ix != EXT_FIRST_INDEX(path[depth].p_hdr));
		}
		*logical = le32_to_cpu(ex->ee_block);
		*phys = ext_pblock(ex);
		return 0;
	}

1147
	BUG_ON(*logical < (le32_to_cpu(ex->ee_block) + ee_len));
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204

	if (ex != EXT_LAST_EXTENT(path[depth].p_hdr)) {
		/* next allocated block in this leaf */
		ex++;
		*logical = le32_to_cpu(ex->ee_block);
		*phys = ext_pblock(ex);
		return 0;
	}

	/* go up and search for index to the right */
	while (--depth >= 0) {
		ix = path[depth].p_idx;
		if (ix != EXT_LAST_INDEX(path[depth].p_hdr))
			break;
	}

	if (depth < 0) {
		/* we've gone up to the root and
		 * found no index to the right */
		return 0;
	}

	/* we've found index to the right, let's
	 * follow it and find the closest allocated
	 * block to the right */
	ix++;
	block = idx_pblock(ix);
	while (++depth < path->p_depth) {
		bh = sb_bread(inode->i_sb, block);
		if (bh == NULL)
			return -EIO;
		eh = ext_block_hdr(bh);
		if (ext4_ext_check_header(inode, eh, depth)) {
			put_bh(bh);
			return -EIO;
		}
		ix = EXT_FIRST_INDEX(eh);
		block = idx_pblock(ix);
		put_bh(bh);
	}

	bh = sb_bread(inode->i_sb, block);
	if (bh == NULL)
		return -EIO;
	eh = ext_block_hdr(bh);
	if (ext4_ext_check_header(inode, eh, path->p_depth - depth)) {
		put_bh(bh);
		return -EIO;
	}
	ex = EXT_FIRST_EXTENT(eh);
	*logical = le32_to_cpu(ex->ee_block);
	*phys = ext_pblock(ex);
	put_bh(bh);
	return 0;

}

A
Alex Tomas 已提交
1205
/*
1206 1207 1208 1209 1210
 * ext4_ext_next_allocated_block:
 * returns allocated block in subsequent extent or EXT_MAX_BLOCK.
 * NOTE: it considers block number from index entry as
 * allocated block. Thus, index entries have to be consistent
 * with leaves.
A
Alex Tomas 已提交
1211
 */
A
Aneesh Kumar K.V 已提交
1212
static ext4_lblk_t
A
Alex Tomas 已提交
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
ext4_ext_next_allocated_block(struct ext4_ext_path *path)
{
	int depth;

	BUG_ON(path == NULL);
	depth = path->p_depth;

	if (depth == 0 && path->p_ext == NULL)
		return EXT_MAX_BLOCK;

	while (depth >= 0) {
		if (depth == path->p_depth) {
			/* leaf */
			if (path[depth].p_ext !=
					EXT_LAST_EXTENT(path[depth].p_hdr))
			  return le32_to_cpu(path[depth].p_ext[1].ee_block);
		} else {
			/* index */
			if (path[depth].p_idx !=
					EXT_LAST_INDEX(path[depth].p_hdr))
			  return le32_to_cpu(path[depth].p_idx[1].ei_block);
		}
		depth--;
	}

	return EXT_MAX_BLOCK;
}

/*
1242
 * ext4_ext_next_leaf_block:
A
Alex Tomas 已提交
1243 1244
 * returns first allocated block from next leaf or EXT_MAX_BLOCK
 */
A
Aneesh Kumar K.V 已提交
1245
static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode,
A
Andrew Morton 已提交
1246
					struct ext4_ext_path *path)
A
Alex Tomas 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
{
	int depth;

	BUG_ON(path == NULL);
	depth = path->p_depth;

	/* zero-tree has no leaf blocks at all */
	if (depth == 0)
		return EXT_MAX_BLOCK;

	/* go to index block */
	depth--;

	while (depth >= 0) {
		if (path[depth].p_idx !=
				EXT_LAST_INDEX(path[depth].p_hdr))
A
Aneesh Kumar K.V 已提交
1263 1264
			return (ext4_lblk_t)
				le32_to_cpu(path[depth].p_idx[1].ei_block);
A
Alex Tomas 已提交
1265 1266 1267 1268 1269 1270 1271
		depth--;
	}

	return EXT_MAX_BLOCK;
}

/*
1272 1273 1274
 * ext4_ext_correct_indexes:
 * if leaf gets modified and modified extent is first in the leaf,
 * then we have to correct all indexes above.
A
Alex Tomas 已提交
1275 1276
 * TODO: do we need to correct tree in all cases?
 */
A
Aneesh Kumar K.V 已提交
1277
static int ext4_ext_correct_indexes(handle_t *handle, struct inode *inode,
A
Alex Tomas 已提交
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
				struct ext4_ext_path *path)
{
	struct ext4_extent_header *eh;
	int depth = ext_depth(inode);
	struct ext4_extent *ex;
	__le32 border;
	int k, err = 0;

	eh = path[depth].p_hdr;
	ex = path[depth].p_ext;
	BUG_ON(ex == NULL);
	BUG_ON(eh == NULL);

	if (depth == 0) {
		/* there is no tree at all */
		return 0;
	}

	if (ex != EXT_FIRST_EXTENT(eh)) {
		/* we correct tree if first leaf got modified only */
		return 0;
	}

	/*
1302
	 * TODO: we need correction if border is smaller than current one
A
Alex Tomas 已提交
1303 1304 1305
	 */
	k = depth - 1;
	border = path[depth].p_ext->ee_block;
1306 1307
	err = ext4_ext_get_access(handle, inode, path + k);
	if (err)
A
Alex Tomas 已提交
1308 1309
		return err;
	path[k].p_idx->ei_block = border;
1310 1311
	err = ext4_ext_dirty(handle, inode, path + k);
	if (err)
A
Alex Tomas 已提交
1312 1313 1314 1315 1316 1317
		return err;

	while (k--) {
		/* change all left-side indexes */
		if (path[k+1].p_idx != EXT_FIRST_INDEX(path[k+1].p_hdr))
			break;
1318 1319
		err = ext4_ext_get_access(handle, inode, path + k);
		if (err)
A
Alex Tomas 已提交
1320 1321
			break;
		path[k].p_idx->ei_block = border;
1322 1323
		err = ext4_ext_dirty(handle, inode, path + k);
		if (err)
A
Alex Tomas 已提交
1324 1325 1326 1327 1328 1329
			break;
	}

	return err;
}

1330
static int
A
Alex Tomas 已提交
1331 1332 1333
ext4_can_extents_be_merged(struct inode *inode, struct ext4_extent *ex1,
				struct ext4_extent *ex2)
{
1334
	unsigned short ext1_ee_len, ext2_ee_len, max_len;
A
Amit Arora 已提交
1335 1336 1337 1338 1339 1340 1341 1342

	/*
	 * Make sure that either both extents are uninitialized, or
	 * both are _not_.
	 */
	if (ext4_ext_is_uninitialized(ex1) ^ ext4_ext_is_uninitialized(ex2))
		return 0;

1343 1344 1345 1346 1347
	if (ext4_ext_is_uninitialized(ex1))
		max_len = EXT_UNINIT_MAX_LEN;
	else
		max_len = EXT_INIT_MAX_LEN;

A
Amit Arora 已提交
1348 1349 1350 1351
	ext1_ee_len = ext4_ext_get_actual_len(ex1);
	ext2_ee_len = ext4_ext_get_actual_len(ex2);

	if (le32_to_cpu(ex1->ee_block) + ext1_ee_len !=
A
Andrew Morton 已提交
1352
			le32_to_cpu(ex2->ee_block))
A
Alex Tomas 已提交
1353 1354
		return 0;

1355 1356 1357
	/*
	 * To allow future support for preallocated extents to be added
	 * as an RO_COMPAT feature, refuse to merge to extents if
1358
	 * this can result in the top bit of ee_len being set.
1359
	 */
1360
	if (ext1_ee_len + ext2_ee_len > max_len)
1361
		return 0;
1362
#ifdef AGGRESSIVE_TEST
1363
	if (ext1_ee_len >= 4)
A
Alex Tomas 已提交
1364 1365 1366
		return 0;
#endif

A
Amit Arora 已提交
1367
	if (ext_pblock(ex1) + ext1_ee_len == ext_pblock(ex2))
A
Alex Tomas 已提交
1368 1369 1370 1371
		return 1;
	return 0;
}

1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
/*
 * This function tries to merge the "ex" extent to the next extent in the tree.
 * It always tries to merge towards right. If you want to merge towards
 * left, pass "ex - 1" as argument instead of "ex".
 * Returns 0 if the extents (ex and ex+1) were _not_ merged and returns
 * 1 if they got merged.
 */
int ext4_ext_try_to_merge(struct inode *inode,
			  struct ext4_ext_path *path,
			  struct ext4_extent *ex)
{
	struct ext4_extent_header *eh;
	unsigned int depth, len;
	int merge_done = 0;
	int uninitialized = 0;

	depth = ext_depth(inode);
	BUG_ON(path[depth].p_hdr == NULL);
	eh = path[depth].p_hdr;

	while (ex < EXT_LAST_EXTENT(eh)) {
		if (!ext4_can_extents_be_merged(inode, ex, ex + 1))
			break;
		/* merge with next extent! */
		if (ext4_ext_is_uninitialized(ex))
			uninitialized = 1;
		ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
				+ ext4_ext_get_actual_len(ex + 1));
		if (uninitialized)
			ext4_ext_mark_uninitialized(ex);

		if (ex + 1 < EXT_LAST_EXTENT(eh)) {
			len = (EXT_LAST_EXTENT(eh) - ex - 1)
				* sizeof(struct ext4_extent);
			memmove(ex + 1, ex + 2, len);
		}
M
Marcin Slusarz 已提交
1408
		le16_add_cpu(&eh->eh_entries, -1);
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418
		merge_done = 1;
		WARN_ON(eh->eh_entries == 0);
		if (!eh->eh_entries)
			ext4_error(inode->i_sb, "ext4_ext_try_to_merge",
			   "inode#%lu, eh->eh_entries = 0!", inode->i_ino);
	}

	return merge_done;
}

A
Amit Arora 已提交
1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430
/*
 * check if a portion of the "newext" extent overlaps with an
 * existing extent.
 *
 * If there is an overlap discovered, it updates the length of the newext
 * such that there will be no overlap, and then returns 1.
 * If there is no overlap found, it returns 0.
 */
unsigned int ext4_ext_check_overlap(struct inode *inode,
				    struct ext4_extent *newext,
				    struct ext4_ext_path *path)
{
A
Aneesh Kumar K.V 已提交
1431
	ext4_lblk_t b1, b2;
A
Amit Arora 已提交
1432 1433 1434 1435
	unsigned int depth, len1;
	unsigned int ret = 0;

	b1 = le32_to_cpu(newext->ee_block);
A
Amit Arora 已提交
1436
	len1 = ext4_ext_get_actual_len(newext);
A
Amit Arora 已提交
1437 1438 1439 1440 1441 1442 1443
	depth = ext_depth(inode);
	if (!path[depth].p_ext)
		goto out;
	b2 = le32_to_cpu(path[depth].p_ext->ee_block);

	/*
	 * get the next allocated block if the extent in the path
1444
	 * is before the requested block(s)
A
Amit Arora 已提交
1445 1446 1447 1448 1449 1450 1451
	 */
	if (b2 < b1) {
		b2 = ext4_ext_next_allocated_block(path);
		if (b2 == EXT_MAX_BLOCK)
			goto out;
	}

A
Aneesh Kumar K.V 已提交
1452
	/* check for wrap through zero on extent logical start block*/
A
Amit Arora 已提交
1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
	if (b1 + len1 < b1) {
		len1 = EXT_MAX_BLOCK - b1;
		newext->ee_len = cpu_to_le16(len1);
		ret = 1;
	}

	/* check for overlap */
	if (b1 + len1 > b2) {
		newext->ee_len = cpu_to_le16(b2 - b1);
		ret = 1;
	}
out:
	return ret;
}

A
Alex Tomas 已提交
1468
/*
1469 1470 1471 1472
 * ext4_ext_insert_extent:
 * tries to merge requsted extent into the existing extent or
 * inserts requested extent as new one into the tree,
 * creating new leaf in the no-space case.
A
Alex Tomas 已提交
1473 1474 1475 1476 1477 1478 1479 1480 1481
 */
int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
				struct ext4_ext_path *path,
				struct ext4_extent *newext)
{
	struct ext4_extent_header * eh;
	struct ext4_extent *ex, *fex;
	struct ext4_extent *nearex; /* nearest extent */
	struct ext4_ext_path *npath = NULL;
A
Aneesh Kumar K.V 已提交
1482 1483
	int depth, len, err;
	ext4_lblk_t next;
A
Amit Arora 已提交
1484
	unsigned uninitialized = 0;
A
Alex Tomas 已提交
1485

A
Amit Arora 已提交
1486
	BUG_ON(ext4_ext_get_actual_len(newext) == 0);
A
Alex Tomas 已提交
1487 1488 1489 1490 1491 1492
	depth = ext_depth(inode);
	ex = path[depth].p_ext;
	BUG_ON(path[depth].p_hdr == NULL);

	/* try to insert block into found extent and return */
	if (ex && ext4_can_extents_be_merged(inode, ex, newext)) {
1493
		ext_debug("append %d block to %d:%d (from %llu)\n",
A
Amit Arora 已提交
1494
				ext4_ext_get_actual_len(newext),
A
Alex Tomas 已提交
1495
				le32_to_cpu(ex->ee_block),
A
Amit Arora 已提交
1496
				ext4_ext_get_actual_len(ex), ext_pblock(ex));
1497 1498
		err = ext4_ext_get_access(handle, inode, path + depth);
		if (err)
A
Alex Tomas 已提交
1499
			return err;
A
Amit Arora 已提交
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511

		/*
		 * ext4_can_extents_be_merged should have checked that either
		 * both extents are uninitialized, or both aren't. Thus we
		 * need to check only one of them here.
		 */
		if (ext4_ext_is_uninitialized(ex))
			uninitialized = 1;
		ex->ee_len = cpu_to_le16(ext4_ext_get_actual_len(ex)
					+ ext4_ext_get_actual_len(newext));
		if (uninitialized)
			ext4_ext_mark_uninitialized(ex);
A
Alex Tomas 已提交
1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
		eh = path[depth].p_hdr;
		nearex = ex;
		goto merge;
	}

repeat:
	depth = ext_depth(inode);
	eh = path[depth].p_hdr;
	if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max))
		goto has_space;

	/* probably next leaf has space for us? */
	fex = EXT_LAST_EXTENT(eh);
	next = ext4_ext_next_leaf_block(inode, path);
	if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block)
	    && next != EXT_MAX_BLOCK) {
		ext_debug("next leaf block - %d\n", next);
		BUG_ON(npath != NULL);
		npath = ext4_ext_find_extent(inode, next, NULL);
		if (IS_ERR(npath))
			return PTR_ERR(npath);
		BUG_ON(npath->p_depth != path->p_depth);
		eh = npath[depth].p_hdr;
		if (le16_to_cpu(eh->eh_entries) < le16_to_cpu(eh->eh_max)) {
			ext_debug("next leaf isnt full(%d)\n",
				  le16_to_cpu(eh->eh_entries));
			path = npath;
			goto repeat;
		}
		ext_debug("next leaf has no free space(%d,%d)\n",
			  le16_to_cpu(eh->eh_entries), le16_to_cpu(eh->eh_max));
	}

	/*
1546 1547
	 * There is no free space in the found leaf.
	 * We're gonna add a new leaf in the tree.
A
Alex Tomas 已提交
1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
	 */
	err = ext4_ext_create_new_leaf(handle, inode, path, newext);
	if (err)
		goto cleanup;
	depth = ext_depth(inode);
	eh = path[depth].p_hdr;

has_space:
	nearex = path[depth].p_ext;

1558 1559
	err = ext4_ext_get_access(handle, inode, path + depth);
	if (err)
A
Alex Tomas 已提交
1560 1561 1562 1563
		goto cleanup;

	if (!nearex) {
		/* there is no extent in this leaf, create first one */
1564
		ext_debug("first extent in the leaf: %d:%llu:%d\n",
D
Dave Kleikamp 已提交
1565 1566
				le32_to_cpu(newext->ee_block),
				ext_pblock(newext),
A
Amit Arora 已提交
1567
				ext4_ext_get_actual_len(newext));
A
Alex Tomas 已提交
1568 1569
		path[depth].p_ext = EXT_FIRST_EXTENT(eh);
	} else if (le32_to_cpu(newext->ee_block)
D
Dave Kleikamp 已提交
1570
			   > le32_to_cpu(nearex->ee_block)) {
A
Alex Tomas 已提交
1571 1572 1573 1574 1575
/*		BUG_ON(newext->ee_block == nearex->ee_block); */
		if (nearex != EXT_LAST_EXTENT(eh)) {
			len = EXT_MAX_EXTENT(eh) - nearex;
			len = (len - 1) * sizeof(struct ext4_extent);
			len = len < 0 ? 0 : len;
1576
			ext_debug("insert %d:%llu:%d after: nearest 0x%p, "
A
Alex Tomas 已提交
1577
					"move %d from 0x%p to 0x%p\n",
D
Dave Kleikamp 已提交
1578 1579
					le32_to_cpu(newext->ee_block),
					ext_pblock(newext),
A
Amit Arora 已提交
1580
					ext4_ext_get_actual_len(newext),
A
Alex Tomas 已提交
1581 1582 1583 1584 1585 1586 1587 1588
					nearex, len, nearex + 1, nearex + 2);
			memmove(nearex + 2, nearex + 1, len);
		}
		path[depth].p_ext = nearex + 1;
	} else {
		BUG_ON(newext->ee_block == nearex->ee_block);
		len = (EXT_MAX_EXTENT(eh) - nearex) * sizeof(struct ext4_extent);
		len = len < 0 ? 0 : len;
1589
		ext_debug("insert %d:%llu:%d before: nearest 0x%p, "
A
Alex Tomas 已提交
1590 1591
				"move %d from 0x%p to 0x%p\n",
				le32_to_cpu(newext->ee_block),
1592
				ext_pblock(newext),
A
Amit Arora 已提交
1593
				ext4_ext_get_actual_len(newext),
A
Alex Tomas 已提交
1594 1595 1596 1597 1598
				nearex, len, nearex + 1, nearex + 2);
		memmove(nearex + 1, nearex, len);
		path[depth].p_ext = nearex;
	}

M
Marcin Slusarz 已提交
1599
	le16_add_cpu(&eh->eh_entries, 1);
A
Alex Tomas 已提交
1600 1601
	nearex = path[depth].p_ext;
	nearex->ee_block = newext->ee_block;
1602
	ext4_ext_store_pblock(nearex, ext_pblock(newext));
A
Alex Tomas 已提交
1603 1604 1605 1606
	nearex->ee_len = newext->ee_len;

merge:
	/* try to merge extents to the right */
1607
	ext4_ext_try_to_merge(inode, path, nearex);
A
Alex Tomas 已提交
1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627

	/* try to merge extents to the left */

	/* time to correct all indexes above */
	err = ext4_ext_correct_indexes(handle, inode, path);
	if (err)
		goto cleanup;

	err = ext4_ext_dirty(handle, inode, path + depth);

cleanup:
	if (npath) {
		ext4_ext_drop_refs(npath);
		kfree(npath);
	}
	ext4_ext_tree_changed(inode);
	ext4_ext_invalidate_cache(inode);
	return err;
}

1628
static void
A
Aneesh Kumar K.V 已提交
1629
ext4_ext_put_in_cache(struct inode *inode, ext4_lblk_t block,
1630
			__u32 len, ext4_fsblk_t start, int type)
A
Alex Tomas 已提交
1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
{
	struct ext4_ext_cache *cex;
	BUG_ON(len == 0);
	cex = &EXT4_I(inode)->i_cached_extent;
	cex->ec_type = type;
	cex->ec_block = block;
	cex->ec_len = len;
	cex->ec_start = start;
}

/*
1642 1643
 * ext4_ext_put_gap_in_cache:
 * calculate boundaries of the gap that the requested block fits into
A
Alex Tomas 已提交
1644 1645
 * and cache this gap
 */
1646
static void
A
Alex Tomas 已提交
1647
ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path,
A
Aneesh Kumar K.V 已提交
1648
				ext4_lblk_t block)
A
Alex Tomas 已提交
1649 1650
{
	int depth = ext_depth(inode);
A
Aneesh Kumar K.V 已提交
1651 1652
	unsigned long len;
	ext4_lblk_t lblock;
A
Alex Tomas 已提交
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
	struct ext4_extent *ex;

	ex = path[depth].p_ext;
	if (ex == NULL) {
		/* there is no extent yet, so gap is [0;-] */
		lblock = 0;
		len = EXT_MAX_BLOCK;
		ext_debug("cache gap(whole file):");
	} else if (block < le32_to_cpu(ex->ee_block)) {
		lblock = block;
		len = le32_to_cpu(ex->ee_block) - block;
1664 1665 1666 1667
		ext_debug("cache gap(before): %u [%u:%u]",
				block,
				le32_to_cpu(ex->ee_block),
				 ext4_ext_get_actual_len(ex));
A
Alex Tomas 已提交
1668
	} else if (block >= le32_to_cpu(ex->ee_block)
A
Amit Arora 已提交
1669
			+ ext4_ext_get_actual_len(ex)) {
A
Aneesh Kumar K.V 已提交
1670
		ext4_lblk_t next;
D
Dave Kleikamp 已提交
1671
		lblock = le32_to_cpu(ex->ee_block)
A
Amit Arora 已提交
1672
			+ ext4_ext_get_actual_len(ex);
A
Aneesh Kumar K.V 已提交
1673 1674

		next = ext4_ext_next_allocated_block(path);
1675 1676 1677 1678
		ext_debug("cache gap(after): [%u:%u] %u",
				le32_to_cpu(ex->ee_block),
				ext4_ext_get_actual_len(ex),
				block);
A
Aneesh Kumar K.V 已提交
1679 1680
		BUG_ON(next == lblock);
		len = next - lblock;
A
Alex Tomas 已提交
1681 1682 1683 1684 1685
	} else {
		lblock = len = 0;
		BUG();
	}

1686
	ext_debug(" -> %u:%lu\n", lblock, len);
A
Alex Tomas 已提交
1687 1688 1689
	ext4_ext_put_in_cache(inode, lblock, len, 0, EXT4_EXT_CACHE_GAP);
}

1690
static int
A
Aneesh Kumar K.V 已提交
1691
ext4_ext_in_cache(struct inode *inode, ext4_lblk_t block,
A
Alex Tomas 已提交
1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704
			struct ext4_extent *ex)
{
	struct ext4_ext_cache *cex;

	cex = &EXT4_I(inode)->i_cached_extent;

	/* has cache valid data? */
	if (cex->ec_type == EXT4_EXT_CACHE_NO)
		return EXT4_EXT_CACHE_NO;

	BUG_ON(cex->ec_type != EXT4_EXT_CACHE_GAP &&
			cex->ec_type != EXT4_EXT_CACHE_EXTENT);
	if (block >= cex->ec_block && block < cex->ec_block + cex->ec_len) {
D
Dave Kleikamp 已提交
1705
		ex->ee_block = cpu_to_le32(cex->ec_block);
1706
		ext4_ext_store_pblock(ex, cex->ec_start);
D
Dave Kleikamp 已提交
1707
		ex->ee_len = cpu_to_le16(cex->ec_len);
1708 1709 1710
		ext_debug("%u cached by %u:%u:%llu\n",
				block,
				cex->ec_block, cex->ec_len, cex->ec_start);
A
Alex Tomas 已提交
1711 1712 1713 1714 1715 1716 1717 1718
		return cex->ec_type;
	}

	/* not in cache */
	return EXT4_EXT_CACHE_NO;
}

/*
1719 1720 1721 1722
 * ext4_ext_rm_idx:
 * removes index from the index block.
 * It's used in truncate case only, thus all requests are for
 * last index in the block only.
A
Alex Tomas 已提交
1723
 */
A
Aneesh Kumar K.V 已提交
1724
static int ext4_ext_rm_idx(handle_t *handle, struct inode *inode,
A
Alex Tomas 已提交
1725 1726 1727 1728
			struct ext4_ext_path *path)
{
	struct buffer_head *bh;
	int err;
1729
	ext4_fsblk_t leaf;
A
Alex Tomas 已提交
1730 1731 1732

	/* free index block */
	path--;
1733
	leaf = idx_pblock(path->p_idx);
A
Alex Tomas 已提交
1734
	BUG_ON(path->p_hdr->eh_entries == 0);
1735 1736
	err = ext4_ext_get_access(handle, inode, path);
	if (err)
A
Alex Tomas 已提交
1737
		return err;
M
Marcin Slusarz 已提交
1738
	le16_add_cpu(&path->p_hdr->eh_entries, -1);
1739 1740
	err = ext4_ext_dirty(handle, inode, path);
	if (err)
A
Alex Tomas 已提交
1741
		return err;
1742
	ext_debug("index is empty, remove it, free block %llu\n", leaf);
A
Alex Tomas 已提交
1743 1744
	bh = sb_find_get_block(inode->i_sb, leaf);
	ext4_forget(handle, 1, inode, bh, leaf);
1745
	ext4_free_blocks(handle, inode, leaf, 1, 1);
A
Alex Tomas 已提交
1746 1747 1748 1749
	return err;
}

/*
1750 1751 1752 1753 1754
 * ext4_ext_calc_credits_for_single_extent:
 * This routine returns max. credits that needed to insert an extent
 * to the extent tree.
 * When pass the actual path, the caller should calculate credits
 * under i_data_sem.
A
Alex Tomas 已提交
1755
 */
1756
int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int num,
A
Alex Tomas 已提交
1757 1758 1759
						struct ext4_ext_path *path)
{
	if (path) {
1760
		int depth = ext_depth(inode);
1761
		int ret = 0;
1762

A
Alex Tomas 已提交
1763 1764
		/* probably there is space in leaf? */
		if (le16_to_cpu(path[depth].p_hdr->eh_entries)
1765
				< le16_to_cpu(path[depth].p_hdr->eh_max)) {
A
Alex Tomas 已提交
1766

1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778
			/*
			 *  There are some space in the leaf tree, no
			 *  need to account for leaf block credit
			 *
			 *  bitmaps and block group descriptor blocks
			 *  and other metadat blocks still need to be
			 *  accounted.
			 */
			/* 1 one bitmap, 1 block group descriptor */
			ret = 2 + EXT4_META_TRANS_BLOCKS(inode->i_sb);
		}
	}
A
Alex Tomas 已提交
1779

1780
	return ext4_chunk_trans_blocks(inode, num);
1781
}
A
Alex Tomas 已提交
1782

1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797
/*
 * How many index/leaf blocks need to change/allocate to modify nrblocks?
 *
 * if nrblocks are fit in a single extent (chunk flag is 1), then
 * in the worse case, each tree level index/leaf need to be changed
 * if the tree split due to insert a new extent, then the old tree
 * index/leaf need to be updated too
 *
 * If the nrblocks are discontiguous, they could cause
 * the whole tree split more than once, but this is really rare.
 */
int ext4_ext_index_trans_blocks(struct inode *inode, int num, int chunk)
{
	int index;
	int depth = ext_depth(inode);
A
Alex Tomas 已提交
1798

1799 1800 1801 1802
	if (chunk)
		index = depth * 2;
	else
		index = depth * 3;
A
Alex Tomas 已提交
1803

1804
	return index;
A
Alex Tomas 已提交
1805 1806 1807 1808
}

static int ext4_remove_blocks(handle_t *handle, struct inode *inode,
				struct ext4_extent *ex,
A
Aneesh Kumar K.V 已提交
1809
				ext4_lblk_t from, ext4_lblk_t to)
A
Alex Tomas 已提交
1810 1811
{
	struct buffer_head *bh;
A
Amit Arora 已提交
1812
	unsigned short ee_len =  ext4_ext_get_actual_len(ex);
1813
	int i, metadata = 0;
A
Alex Tomas 已提交
1814

1815 1816
	if (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode))
		metadata = 1;
A
Alex Tomas 已提交
1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
#ifdef EXTENTS_STATS
	{
		struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
		spin_lock(&sbi->s_ext_stats_lock);
		sbi->s_ext_blocks += ee_len;
		sbi->s_ext_extents++;
		if (ee_len < sbi->s_ext_min)
			sbi->s_ext_min = ee_len;
		if (ee_len > sbi->s_ext_max)
			sbi->s_ext_max = ee_len;
		if (ext_depth(inode) > sbi->s_depth_max)
			sbi->s_depth_max = ext_depth(inode);
		spin_unlock(&sbi->s_ext_stats_lock);
	}
#endif
	if (from >= le32_to_cpu(ex->ee_block)
A
Amit Arora 已提交
1833
	    && to == le32_to_cpu(ex->ee_block) + ee_len - 1) {
A
Alex Tomas 已提交
1834
		/* tail removal */
A
Aneesh Kumar K.V 已提交
1835
		ext4_lblk_t num;
1836
		ext4_fsblk_t start;
A
Aneesh Kumar K.V 已提交
1837

A
Amit Arora 已提交
1838 1839
		num = le32_to_cpu(ex->ee_block) + ee_len - from;
		start = ext_pblock(ex) + ee_len - num;
A
Aneesh Kumar K.V 已提交
1840
		ext_debug("free last %u blocks starting %llu\n", num, start);
A
Alex Tomas 已提交
1841 1842 1843 1844
		for (i = 0; i < num; i++) {
			bh = sb_find_get_block(inode->i_sb, start + i);
			ext4_forget(handle, 0, inode, bh, start + i);
		}
1845
		ext4_free_blocks(handle, inode, start, num, metadata);
A
Alex Tomas 已提交
1846
	} else if (from == le32_to_cpu(ex->ee_block)
A
Amit Arora 已提交
1847
		   && to <= le32_to_cpu(ex->ee_block) + ee_len - 1) {
A
Aneesh Kumar K.V 已提交
1848
		printk(KERN_INFO "strange request: removal %u-%u from %u:%u\n",
A
Amit Arora 已提交
1849
			from, to, le32_to_cpu(ex->ee_block), ee_len);
A
Alex Tomas 已提交
1850
	} else {
A
Aneesh Kumar K.V 已提交
1851 1852 1853
		printk(KERN_INFO "strange request: removal(2) "
				"%u-%u from %u:%u\n",
				from, to, le32_to_cpu(ex->ee_block), ee_len);
A
Alex Tomas 已提交
1854 1855 1856 1857 1858 1859
	}
	return 0;
}

static int
ext4_ext_rm_leaf(handle_t *handle, struct inode *inode,
A
Aneesh Kumar K.V 已提交
1860
		struct ext4_ext_path *path, ext4_lblk_t start)
A
Alex Tomas 已提交
1861 1862 1863 1864
{
	int err = 0, correct_index = 0;
	int depth = ext_depth(inode), credits;
	struct ext4_extent_header *eh;
A
Aneesh Kumar K.V 已提交
1865 1866 1867
	ext4_lblk_t a, b, block;
	unsigned num;
	ext4_lblk_t ex_ee_block;
A
Alex Tomas 已提交
1868
	unsigned short ex_ee_len;
A
Amit Arora 已提交
1869
	unsigned uninitialized = 0;
A
Alex Tomas 已提交
1870 1871
	struct ext4_extent *ex;

1872
	/* the header must be checked already in ext4_ext_remove_space() */
A
Aneesh Kumar K.V 已提交
1873
	ext_debug("truncate since %u in leaf\n", start);
A
Alex Tomas 已提交
1874 1875 1876 1877 1878 1879 1880 1881 1882
	if (!path[depth].p_hdr)
		path[depth].p_hdr = ext_block_hdr(path[depth].p_bh);
	eh = path[depth].p_hdr;
	BUG_ON(eh == NULL);

	/* find where to start removing */
	ex = EXT_LAST_EXTENT(eh);

	ex_ee_block = le32_to_cpu(ex->ee_block);
A
Amit Arora 已提交
1883 1884 1885
	if (ext4_ext_is_uninitialized(ex))
		uninitialized = 1;
	ex_ee_len = ext4_ext_get_actual_len(ex);
A
Alex Tomas 已提交
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919

	while (ex >= EXT_FIRST_EXTENT(eh) &&
			ex_ee_block + ex_ee_len > start) {
		ext_debug("remove ext %lu:%u\n", ex_ee_block, ex_ee_len);
		path[depth].p_ext = ex;

		a = ex_ee_block > start ? ex_ee_block : start;
		b = ex_ee_block + ex_ee_len - 1 < EXT_MAX_BLOCK ?
			ex_ee_block + ex_ee_len - 1 : EXT_MAX_BLOCK;

		ext_debug("  border %u:%u\n", a, b);

		if (a != ex_ee_block && b != ex_ee_block + ex_ee_len - 1) {
			block = 0;
			num = 0;
			BUG();
		} else if (a != ex_ee_block) {
			/* remove tail of the extent */
			block = ex_ee_block;
			num = a - block;
		} else if (b != ex_ee_block + ex_ee_len - 1) {
			/* remove head of the extent */
			block = a;
			num = b - a;
			/* there is no "make a hole" API yet */
			BUG();
		} else {
			/* remove whole extent: excellent! */
			block = ex_ee_block;
			num = 0;
			BUG_ON(a != ex_ee_block);
			BUG_ON(b != ex_ee_block + ex_ee_len - 1);
		}

1920 1921 1922 1923 1924 1925 1926
		/*
		 * 3 for leaf, sb, and inode plus 2 (bmap and group
		 * descriptor) for each block group; assume two block
		 * groups plus ex_ee_len/blocks_per_block_group for
		 * the worst case
		 */
		credits = 7 + 2*(ex_ee_len/EXT4_BLOCKS_PER_GROUP(inode->i_sb));
A
Alex Tomas 已提交
1927 1928 1929 1930 1931 1932
		if (ex == EXT_FIRST_EXTENT(eh)) {
			correct_index = 1;
			credits += (ext_depth(inode)) + 1;
		}
		credits += 2 * EXT4_QUOTA_TRANS_BLOCKS(inode->i_sb);

1933 1934
		err = ext4_ext_journal_restart(handle, credits);
		if (err)
A
Alex Tomas 已提交
1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
			goto out;

		err = ext4_ext_get_access(handle, inode, path + depth);
		if (err)
			goto out;

		err = ext4_remove_blocks(handle, inode, ex, a, b);
		if (err)
			goto out;

		if (num == 0) {
1946
			/* this extent is removed; mark slot entirely unused */
1947
			ext4_ext_store_pblock(ex, 0);
M
Marcin Slusarz 已提交
1948
			le16_add_cpu(&eh->eh_entries, -1);
A
Alex Tomas 已提交
1949 1950 1951 1952
		}

		ex->ee_block = cpu_to_le32(block);
		ex->ee_len = cpu_to_le16(num);
1953 1954 1955 1956 1957
		/*
		 * Do not mark uninitialized if all the blocks in the
		 * extent have been removed.
		 */
		if (uninitialized && num)
A
Amit Arora 已提交
1958
			ext4_ext_mark_uninitialized(ex);
A
Alex Tomas 已提交
1959 1960 1961 1962 1963

		err = ext4_ext_dirty(handle, inode, path + depth);
		if (err)
			goto out;

1964
		ext_debug("new extent: %u:%u:%llu\n", block, num,
1965
				ext_pblock(ex));
A
Alex Tomas 已提交
1966 1967
		ex--;
		ex_ee_block = le32_to_cpu(ex->ee_block);
A
Amit Arora 已提交
1968
		ex_ee_len = ext4_ext_get_actual_len(ex);
A
Alex Tomas 已提交
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983
	}

	if (correct_index && eh->eh_entries)
		err = ext4_ext_correct_indexes(handle, inode, path);

	/* if this leaf is free, then we should
	 * remove it from index block above */
	if (err == 0 && eh->eh_entries == 0 && path[depth].p_bh != NULL)
		err = ext4_ext_rm_idx(handle, inode, path + depth);

out:
	return err;
}

/*
1984 1985
 * ext4_ext_more_to_rm:
 * returns 1 if current index has to be freed (even partial)
A
Alex Tomas 已提交
1986
 */
1987
static int
A
Alex Tomas 已提交
1988 1989 1990 1991 1992 1993 1994 1995
ext4_ext_more_to_rm(struct ext4_ext_path *path)
{
	BUG_ON(path->p_idx == NULL);

	if (path->p_idx < EXT_FIRST_INDEX(path->p_hdr))
		return 0;

	/*
1996
	 * if truncate on deeper level happened, it wasn't partial,
A
Alex Tomas 已提交
1997 1998 1999 2000 2001 2002 2003
	 * so we have to consider current index for truncation
	 */
	if (le16_to_cpu(path->p_hdr->eh_entries) == path->p_block)
		return 0;
	return 1;
}

A
Aneesh Kumar K.V 已提交
2004
static int ext4_ext_remove_space(struct inode *inode, ext4_lblk_t start)
A
Alex Tomas 已提交
2005 2006 2007 2008 2009 2010 2011
{
	struct super_block *sb = inode->i_sb;
	int depth = ext_depth(inode);
	struct ext4_ext_path *path;
	handle_t *handle;
	int i = 0, err = 0;

A
Aneesh Kumar K.V 已提交
2012
	ext_debug("truncate since %u\n", start);
A
Alex Tomas 已提交
2013 2014 2015 2016 2017 2018 2019 2020 2021

	/* probably first extent we're gonna free will be last in block */
	handle = ext4_journal_start(inode, depth + 1);
	if (IS_ERR(handle))
		return PTR_ERR(handle);

	ext4_ext_invalidate_cache(inode);

	/*
2022 2023
	 * We start scanning from right side, freeing all the blocks
	 * after i_size and walking into the tree depth-wise.
A
Alex Tomas 已提交
2024
	 */
2025
	path = kzalloc(sizeof(struct ext4_ext_path) * (depth + 1), GFP_NOFS);
A
Alex Tomas 已提交
2026 2027 2028 2029 2030
	if (path == NULL) {
		ext4_journal_stop(handle);
		return -ENOMEM;
	}
	path[0].p_hdr = ext_inode_hdr(inode);
2031
	if (ext4_ext_check_header(inode, path[0].p_hdr, depth)) {
A
Alex Tomas 已提交
2032 2033 2034 2035 2036 2037 2038 2039 2040
		err = -EIO;
		goto out;
	}
	path[0].p_depth = depth;

	while (i >= 0 && err == 0) {
		if (i == depth) {
			/* this is leaf block */
			err = ext4_ext_rm_leaf(handle, inode, path, start);
2041
			/* root level has p_bh == NULL, brelse() eats this */
A
Alex Tomas 已提交
2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054
			brelse(path[i].p_bh);
			path[i].p_bh = NULL;
			i--;
			continue;
		}

		/* this is index block */
		if (!path[i].p_hdr) {
			ext_debug("initialize header\n");
			path[i].p_hdr = ext_block_hdr(path[i].p_bh);
		}

		if (!path[i].p_idx) {
2055
			/* this level hasn't been touched yet */
A
Alex Tomas 已提交
2056 2057 2058 2059 2060 2061
			path[i].p_idx = EXT_LAST_INDEX(path[i].p_hdr);
			path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries)+1;
			ext_debug("init index ptr: hdr 0x%p, num %d\n",
				  path[i].p_hdr,
				  le16_to_cpu(path[i].p_hdr->eh_entries));
		} else {
2062
			/* we were already here, see at next index */
A
Alex Tomas 已提交
2063 2064 2065 2066 2067 2068 2069
			path[i].p_idx--;
		}

		ext_debug("level %d - index, first 0x%p, cur 0x%p\n",
				i, EXT_FIRST_INDEX(path[i].p_hdr),
				path[i].p_idx);
		if (ext4_ext_more_to_rm(path + i)) {
2070
			struct buffer_head *bh;
A
Alex Tomas 已提交
2071
			/* go to the next level */
2072
			ext_debug("move to level %d (block %llu)\n",
2073
				  i + 1, idx_pblock(path[i].p_idx));
A
Alex Tomas 已提交
2074
			memset(path + i + 1, 0, sizeof(*path));
2075 2076
			bh = sb_bread(sb, idx_pblock(path[i].p_idx));
			if (!bh) {
A
Alex Tomas 已提交
2077 2078 2079 2080
				/* should we reset i_size? */
				err = -EIO;
				break;
			}
2081 2082 2083 2084 2085 2086 2087 2088 2089 2090
			if (WARN_ON(i + 1 > depth)) {
				err = -EIO;
				break;
			}
			if (ext4_ext_check_header(inode, ext_block_hdr(bh),
							depth - i - 1)) {
				err = -EIO;
				break;
			}
			path[i + 1].p_bh = bh;
A
Alex Tomas 已提交
2091

2092 2093
			/* save actual number of indexes since this
			 * number is changed at the next iteration */
A
Alex Tomas 已提交
2094 2095 2096
			path[i].p_block = le16_to_cpu(path[i].p_hdr->eh_entries);
			i++;
		} else {
2097
			/* we finished processing this index, go up */
A
Alex Tomas 已提交
2098
			if (path[i].p_hdr->eh_entries == 0 && i > 0) {
2099
				/* index is empty, remove it;
A
Alex Tomas 已提交
2100 2101 2102 2103
				 * handle must be already prepared by the
				 * truncatei_leaf() */
				err = ext4_ext_rm_idx(handle, inode, path + i);
			}
2104
			/* root level has p_bh == NULL, brelse() eats this */
A
Alex Tomas 已提交
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114
			brelse(path[i].p_bh);
			path[i].p_bh = NULL;
			i--;
			ext_debug("return to level %d\n", i);
		}
	}

	/* TODO: flexible tree reduction should be here */
	if (path->p_hdr->eh_entries == 0) {
		/*
2115 2116
		 * truncate to zero freed all the tree,
		 * so we need to correct eh_depth
A
Alex Tomas 已提交
2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145
		 */
		err = ext4_ext_get_access(handle, inode, path);
		if (err == 0) {
			ext_inode_hdr(inode)->eh_depth = 0;
			ext_inode_hdr(inode)->eh_max =
				cpu_to_le16(ext4_ext_space_root(inode));
			err = ext4_ext_dirty(handle, inode, path);
		}
	}
out:
	ext4_ext_tree_changed(inode);
	ext4_ext_drop_refs(path);
	kfree(path);
	ext4_journal_stop(handle);

	return err;
}

/*
 * called at mount time
 */
void ext4_ext_init(struct super_block *sb)
{
	/*
	 * possible initialization would be here
	 */

	if (test_opt(sb, EXTENTS)) {
		printk("EXT4-fs: file extents enabled");
2146 2147
#ifdef AGGRESSIVE_TEST
		printk(", aggressive tests");
A
Alex Tomas 已提交
2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
#endif
#ifdef CHECK_BINSEARCH
		printk(", check binsearch");
#endif
#ifdef EXTENTS_STATS
		printk(", stats");
#endif
		printk("\n");
#ifdef EXTENTS_STATS
		spin_lock_init(&EXT4_SB(sb)->s_ext_stats_lock);
		EXT4_SB(sb)->s_ext_min = 1 << 30;
		EXT4_SB(sb)->s_ext_max = 0;
#endif
	}
}

/*
 * called at umount time
 */
void ext4_ext_release(struct super_block *sb)
{
	if (!test_opt(sb, EXTENTS))
		return;

#ifdef EXTENTS_STATS
	if (EXT4_SB(sb)->s_ext_blocks && EXT4_SB(sb)->s_ext_extents) {
		struct ext4_sb_info *sbi = EXT4_SB(sb);
		printk(KERN_ERR "EXT4-fs: %lu blocks in %lu extents (%lu ave)\n",
			sbi->s_ext_blocks, sbi->s_ext_extents,
			sbi->s_ext_blocks / sbi->s_ext_extents);
		printk(KERN_ERR "EXT4-fs: extents: %lu min, %lu max, max depth %lu\n",
			sbi->s_ext_min, sbi->s_ext_max, sbi->s_depth_max);
	}
#endif
}

2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
static void bi_complete(struct bio *bio, int error)
{
	complete((struct completion *)bio->bi_private);
}

/* FIXME!! we need to try to merge to left or right after zero-out  */
static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
{
	int ret = -EIO;
	struct bio *bio;
	int blkbits, blocksize;
	sector_t ee_pblock;
	struct completion event;
	unsigned int ee_len, len, done, offset;


	blkbits   = inode->i_blkbits;
	blocksize = inode->i_sb->s_blocksize;
	ee_len    = ext4_ext_get_actual_len(ex);
	ee_pblock = ext_pblock(ex);

	/* convert ee_pblock to 512 byte sectors */
	ee_pblock = ee_pblock << (blkbits - 9);

	while (ee_len > 0) {

		if (ee_len > BIO_MAX_PAGES)
			len = BIO_MAX_PAGES;
		else
			len = ee_len;

		bio = bio_alloc(GFP_NOIO, len);
		if (!bio)
			return -ENOMEM;
		bio->bi_sector = ee_pblock;
		bio->bi_bdev   = inode->i_sb->s_bdev;

		done = 0;
		offset = 0;
		while (done < len) {
			ret = bio_add_page(bio, ZERO_PAGE(0),
							blocksize, offset);
			if (ret != blocksize) {
				/*
				 * We can't add any more pages because of
				 * hardware limitations.  Start a new bio.
				 */
				break;
			}
			done++;
			offset += blocksize;
			if (offset >= PAGE_CACHE_SIZE)
				offset = 0;
		}

		init_completion(&event);
		bio->bi_private = &event;
		bio->bi_end_io = bi_complete;
		submit_bio(WRITE, bio);
		wait_for_completion(&event);

		if (test_bit(BIO_UPTODATE, &bio->bi_flags))
			ret = 0;
		else {
			ret = -EIO;
			break;
		}
		bio_put(bio);
		ee_len    -= done;
		ee_pblock += done  << (blkbits - 9);
	}
	return ret;
}

2258 2259
#define EXT4_EXT_ZERO_LEN 7

2260 2261 2262 2263 2264 2265 2266 2267 2268 2269
/*
 * This function is called by ext4_ext_get_blocks() if someone tries to write
 * to an uninitialized extent. It may result in splitting the uninitialized
 * extent into multiple extents (upto three - one initialized and two
 * uninitialized).
 * There are three possibilities:
 *   a> There is no split required: Entire extent should be initialized
 *   b> Splits in two extents: Write is happening at either end of the extent
 *   c> Splits in three extents: Somone is writing in middle of the extent
 */
A
Aneesh Kumar K.V 已提交
2270 2271 2272 2273 2274
static int ext4_ext_convert_to_initialized(handle_t *handle,
						struct inode *inode,
						struct ext4_ext_path *path,
						ext4_lblk_t iblock,
						unsigned long max_blocks)
2275
{
2276
	struct ext4_extent *ex, newex, orig_ex;
2277 2278 2279 2280
	struct ext4_extent *ex1 = NULL;
	struct ext4_extent *ex2 = NULL;
	struct ext4_extent *ex3 = NULL;
	struct ext4_extent_header *eh;
A
Aneesh Kumar K.V 已提交
2281 2282
	ext4_lblk_t ee_block;
	unsigned int allocated, ee_len, depth;
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294
	ext4_fsblk_t newblock;
	int err = 0;
	int ret = 0;

	depth = ext_depth(inode);
	eh = path[depth].p_hdr;
	ex = path[depth].p_ext;
	ee_block = le32_to_cpu(ex->ee_block);
	ee_len = ext4_ext_get_actual_len(ex);
	allocated = ee_len - (iblock - ee_block);
	newblock = iblock - ee_block + ext_pblock(ex);
	ex2 = ex;
2295 2296 2297
	orig_ex.ee_block = ex->ee_block;
	orig_ex.ee_len   = cpu_to_le16(ee_len);
	ext4_ext_store_pblock(&orig_ex, ext_pblock(ex));
2298

2299 2300 2301
	err = ext4_ext_get_access(handle, inode, path + depth);
	if (err)
		goto out;
2302 2303 2304 2305 2306 2307 2308 2309 2310 2311
	/* If extent has less than 2*EXT4_EXT_ZERO_LEN zerout directly */
	if (ee_len <= 2*EXT4_EXT_ZERO_LEN) {
		err =  ext4_ext_zeroout(inode, &orig_ex);
		if (err)
			goto fix_extent_len;
		/* update the extent length and mark as initialized */
		ex->ee_block = orig_ex.ee_block;
		ex->ee_len   = orig_ex.ee_len;
		ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
		ext4_ext_dirty(handle, inode, path + depth);
2312 2313
		/* zeroed the full extent */
		return allocated;
2314
	}
2315

2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332
	/* ex1: ee_block to iblock - 1 : uninitialized */
	if (iblock > ee_block) {
		ex1 = ex;
		ex1->ee_len = cpu_to_le16(iblock - ee_block);
		ext4_ext_mark_uninitialized(ex1);
		ex2 = &newex;
	}
	/*
	 * for sanity, update the length of the ex2 extent before
	 * we insert ex3, if ex1 is NULL. This is to avoid temporary
	 * overlap of blocks.
	 */
	if (!ex1 && allocated > max_blocks)
		ex2->ee_len = cpu_to_le16(max_blocks);
	/* ex3: to ee_block + ee_len : uninitialised */
	if (allocated > max_blocks) {
		unsigned int newdepth;
2333 2334
		/* If extent has less than EXT4_EXT_ZERO_LEN zerout directly */
		if (allocated <= EXT4_EXT_ZERO_LEN) {
2335 2336 2337 2338
			/*
			 * iblock == ee_block is handled by the zerouout
			 * at the beginning.
			 * Mark first half uninitialized.
2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360
			 * Mark second half initialized and zero out the
			 * initialized extent
			 */
			ex->ee_block = orig_ex.ee_block;
			ex->ee_len   = cpu_to_le16(ee_len - allocated);
			ext4_ext_mark_uninitialized(ex);
			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
			ext4_ext_dirty(handle, inode, path + depth);

			ex3 = &newex;
			ex3->ee_block = cpu_to_le32(iblock);
			ext4_ext_store_pblock(ex3, newblock);
			ex3->ee_len = cpu_to_le16(allocated);
			err = ext4_ext_insert_extent(handle, inode, path, ex3);
			if (err == -ENOSPC) {
				err =  ext4_ext_zeroout(inode, &orig_ex);
				if (err)
					goto fix_extent_len;
				ex->ee_block = orig_ex.ee_block;
				ex->ee_len   = orig_ex.ee_len;
				ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
				ext4_ext_dirty(handle, inode, path + depth);
2361
				/* blocks available from iblock */
2362
				return allocated;
2363 2364 2365 2366

			} else if (err)
				goto fix_extent_len;

2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388
			/*
			 * We need to zero out the second half because
			 * an fallocate request can update file size and
			 * converting the second half to initialized extent
			 * implies that we can leak some junk data to user
			 * space.
			 */
			err =  ext4_ext_zeroout(inode, ex3);
			if (err) {
				/*
				 * We should actually mark the
				 * second half as uninit and return error
				 * Insert would have changed the extent
				 */
				depth = ext_depth(inode);
				ext4_ext_drop_refs(path);
				path = ext4_ext_find_extent(inode,
								iblock, path);
				if (IS_ERR(path)) {
					err = PTR_ERR(path);
					return err;
				}
2389
				/* get the second half extent details */
2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400
				ex = path[depth].p_ext;
				err = ext4_ext_get_access(handle, inode,
								path + depth);
				if (err)
					return err;
				ext4_ext_mark_uninitialized(ex);
				ext4_ext_dirty(handle, inode, path + depth);
				return err;
			}

			/* zeroed the second half */
2401 2402
			return allocated;
		}
2403 2404 2405 2406 2407 2408
		ex3 = &newex;
		ex3->ee_block = cpu_to_le32(iblock + max_blocks);
		ext4_ext_store_pblock(ex3, newblock + max_blocks);
		ex3->ee_len = cpu_to_le16(allocated - max_blocks);
		ext4_ext_mark_uninitialized(ex3);
		err = ext4_ext_insert_extent(handle, inode, path, ex3);
2409 2410 2411 2412 2413
		if (err == -ENOSPC) {
			err =  ext4_ext_zeroout(inode, &orig_ex);
			if (err)
				goto fix_extent_len;
			/* update the extent length and mark as initialized */
2414 2415 2416 2417
			ex->ee_block = orig_ex.ee_block;
			ex->ee_len   = orig_ex.ee_len;
			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
			ext4_ext_dirty(handle, inode, path + depth);
2418
			/* zeroed the full extent */
2419
			/* blocks available from iblock */
2420
			return allocated;
2421 2422 2423

		} else if (err)
			goto fix_extent_len;
2424 2425 2426 2427 2428
		/*
		 * The depth, and hence eh & ex might change
		 * as part of the insert above.
		 */
		newdepth = ext_depth(inode);
2429 2430 2431 2432 2433 2434
		/*
		 * update the extent length after successfull insert of the
		 * split extent
		 */
		orig_ex.ee_len = cpu_to_le16(ee_len -
						ext4_ext_get_actual_len(ex3));
2435 2436 2437 2438 2439 2440
		depth = newdepth;
		ext4_ext_drop_refs(path);
		path = ext4_ext_find_extent(inode, iblock, path);
		if (IS_ERR(path)) {
			err = PTR_ERR(path);
			goto out;
2441
		}
2442 2443 2444 2445 2446 2447 2448 2449 2450
		eh = path[depth].p_hdr;
		ex = path[depth].p_ext;
		if (ex2 != &newex)
			ex2 = ex;

		err = ext4_ext_get_access(handle, inode, path + depth);
		if (err)
			goto out;

2451
		allocated = max_blocks;
2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466

		/* If extent has less than EXT4_EXT_ZERO_LEN and we are trying
		 * to insert a extent in the middle zerout directly
		 * otherwise give the extent a chance to merge to left
		 */
		if (le16_to_cpu(orig_ex.ee_len) <= EXT4_EXT_ZERO_LEN &&
							iblock != ee_block) {
			err =  ext4_ext_zeroout(inode, &orig_ex);
			if (err)
				goto fix_extent_len;
			/* update the extent length and mark as initialized */
			ex->ee_block = orig_ex.ee_block;
			ex->ee_len   = orig_ex.ee_len;
			ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
			ext4_ext_dirty(handle, inode, path + depth);
2467
			/* zero out the first half */
2468
			/* blocks available from iblock */
2469
			return allocated;
2470
		}
2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526
	}
	/*
	 * If there was a change of depth as part of the
	 * insertion of ex3 above, we need to update the length
	 * of the ex1 extent again here
	 */
	if (ex1 && ex1 != ex) {
		ex1 = ex;
		ex1->ee_len = cpu_to_le16(iblock - ee_block);
		ext4_ext_mark_uninitialized(ex1);
		ex2 = &newex;
	}
	/* ex2: iblock to iblock + maxblocks-1 : initialised */
	ex2->ee_block = cpu_to_le32(iblock);
	ext4_ext_store_pblock(ex2, newblock);
	ex2->ee_len = cpu_to_le16(allocated);
	if (ex2 != ex)
		goto insert;
	/*
	 * New (initialized) extent starts from the first block
	 * in the current extent. i.e., ex2 == ex
	 * We have to see if it can be merged with the extent
	 * on the left.
	 */
	if (ex2 > EXT_FIRST_EXTENT(eh)) {
		/*
		 * To merge left, pass "ex2 - 1" to try_to_merge(),
		 * since it merges towards right _only_.
		 */
		ret = ext4_ext_try_to_merge(inode, path, ex2 - 1);
		if (ret) {
			err = ext4_ext_correct_indexes(handle, inode, path);
			if (err)
				goto out;
			depth = ext_depth(inode);
			ex2--;
		}
	}
	/*
	 * Try to Merge towards right. This might be required
	 * only when the whole extent is being written to.
	 * i.e. ex2 == ex and ex3 == NULL.
	 */
	if (!ex3) {
		ret = ext4_ext_try_to_merge(inode, path, ex2);
		if (ret) {
			err = ext4_ext_correct_indexes(handle, inode, path);
			if (err)
				goto out;
		}
	}
	/* Mark modified extent as dirty */
	err = ext4_ext_dirty(handle, inode, path + depth);
	goto out;
insert:
	err = ext4_ext_insert_extent(handle, inode, path, &newex);
2527 2528 2529 2530 2531
	if (err == -ENOSPC) {
		err =  ext4_ext_zeroout(inode, &orig_ex);
		if (err)
			goto fix_extent_len;
		/* update the extent length and mark as initialized */
2532 2533 2534 2535
		ex->ee_block = orig_ex.ee_block;
		ex->ee_len   = orig_ex.ee_len;
		ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
		ext4_ext_dirty(handle, inode, path + depth);
2536 2537
		/* zero out the first half */
		return allocated;
2538 2539
	} else if (err)
		goto fix_extent_len;
2540 2541
out:
	return err ? err : allocated;
2542 2543 2544 2545 2546 2547 2548 2549

fix_extent_len:
	ex->ee_block = orig_ex.ee_block;
	ex->ee_len   = orig_ex.ee_len;
	ext4_ext_store_pblock(ex, ext_pblock(&orig_ex));
	ext4_ext_mark_uninitialized(ex);
	ext4_ext_dirty(handle, inode, path + depth);
	return err;
2550 2551
}

2552
/*
2553 2554 2555
 * Block allocation/map/preallocation routine for extents based files
 *
 *
2556
 * Need to be called with
2557 2558
 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
2559 2560 2561 2562 2563 2564 2565 2566 2567 2568
 *
 * return > 0, number of of blocks already mapped/allocated
 *          if create == 0 and these are pre-allocated blocks
 *          	buffer head is unmapped
 *          otherwise blocks are mapped
 *
 * return = 0, if plain look up failed (blocks have not been allocated)
 *          buffer head is unmapped
 *
 * return < 0, error case.
2569
 */
2570
int ext4_ext_get_blocks(handle_t *handle, struct inode *inode,
A
Aneesh Kumar K.V 已提交
2571
			ext4_lblk_t iblock,
A
Alex Tomas 已提交
2572 2573 2574 2575
			unsigned long max_blocks, struct buffer_head *bh_result,
			int create, int extend_disksize)
{
	struct ext4_ext_path *path = NULL;
2576
	struct ext4_extent_header *eh;
A
Alex Tomas 已提交
2577
	struct ext4_extent newex, *ex;
2578
	ext4_fsblk_t goal, newblock;
2579
	int err = 0, depth, ret;
A
Alex Tomas 已提交
2580
	unsigned long allocated = 0;
2581
	struct ext4_allocation_request ar;
2582
	loff_t disksize;
A
Alex Tomas 已提交
2583 2584

	__clear_bit(BH_New, &bh_result->b_state);
2585 2586
	ext_debug("blocks %u/%lu requested for inode %u\n",
			iblock, max_blocks, inode->i_ino);
A
Alex Tomas 已提交
2587 2588

	/* check in cache */
2589 2590
	goal = ext4_ext_in_cache(inode, iblock, &newex);
	if (goal) {
A
Alex Tomas 已提交
2591 2592
		if (goal == EXT4_EXT_CACHE_GAP) {
			if (!create) {
2593 2594 2595 2596
				/*
				 * block isn't allocated yet and
				 * user doesn't want to allocate it
				 */
A
Alex Tomas 已提交
2597 2598 2599 2600 2601
				goto out2;
			}
			/* we should allocate requested block */
		} else if (goal == EXT4_EXT_CACHE_EXTENT) {
			/* block is already allocated */
D
Dave Kleikamp 已提交
2602 2603 2604
			newblock = iblock
				   - le32_to_cpu(newex.ee_block)
				   + ext_pblock(&newex);
2605
			/* number of remaining blocks in the extent */
2606
			allocated = ext4_ext_get_actual_len(&newex) -
A
Alex Tomas 已提交
2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624
					(iblock - le32_to_cpu(newex.ee_block));
			goto out;
		} else {
			BUG();
		}
	}

	/* find extent for this block */
	path = ext4_ext_find_extent(inode, iblock, NULL);
	if (IS_ERR(path)) {
		err = PTR_ERR(path);
		path = NULL;
		goto out2;
	}

	depth = ext_depth(inode);

	/*
2625 2626
	 * consistent leaf must not be empty;
	 * this situation is possible, though, _during_ tree modification;
A
Alex Tomas 已提交
2627 2628 2629
	 * this is why assert can't be put in ext4_ext_find_extent()
	 */
	BUG_ON(path[depth].p_ext == NULL && depth != 0);
2630
	eh = path[depth].p_hdr;
A
Alex Tomas 已提交
2631

2632 2633
	ex = path[depth].p_ext;
	if (ex) {
A
Aneesh Kumar K.V 已提交
2634
		ext4_lblk_t ee_block = le32_to_cpu(ex->ee_block);
2635
		ext4_fsblk_t ee_start = ext_pblock(ex);
A
Amit Arora 已提交
2636
		unsigned short ee_len;
2637 2638 2639

		/*
		 * Uninitialized extents are treated as holes, except that
2640
		 * we split out initialized portions during a write.
2641
		 */
A
Amit Arora 已提交
2642
		ee_len = ext4_ext_get_actual_len(ex);
2643
		/* if found extent covers block, simply return it */
D
Dave Kleikamp 已提交
2644
		if (iblock >= ee_block && iblock < ee_block + ee_len) {
A
Alex Tomas 已提交
2645
			newblock = iblock - ee_block + ee_start;
2646
			/* number of remaining blocks in the extent */
A
Alex Tomas 已提交
2647
			allocated = ee_len - (iblock - ee_block);
2648
			ext_debug("%u fit into %lu:%d -> %llu\n", iblock,
A
Alex Tomas 已提交
2649
					ee_block, ee_len, newblock);
2650

A
Amit Arora 已提交
2651
			/* Do not put uninitialized extent in the cache */
2652
			if (!ext4_ext_is_uninitialized(ex)) {
A
Amit Arora 已提交
2653 2654 2655
				ext4_ext_put_in_cache(inode, ee_block,
							ee_len, ee_start,
							EXT4_EXT_CACHE_EXTENT);
2656 2657 2658 2659
				goto out;
			}
			if (create == EXT4_CREATE_UNINITIALIZED_EXT)
				goto out;
2660 2661 2662 2663 2664 2665 2666 2667 2668 2669
			if (!create) {
				/*
				 * We have blocks reserved already.  We
				 * return allocated blocks so that delalloc
				 * won't do block reservation for us.  But
				 * the buffer head will be unmapped so that
				 * a read from the block returns 0s.
				 */
				if (allocated > max_blocks)
					allocated = max_blocks;
2670
				set_buffer_unwritten(bh_result);
2671
				goto out2;
2672
			}
2673 2674 2675 2676

			ret = ext4_ext_convert_to_initialized(handle, inode,
								path, iblock,
								max_blocks);
2677 2678
			if (ret <= 0) {
				err = ret;
2679
				goto out2;
2680
			} else
2681 2682
				allocated = ret;
			goto outnew;
A
Alex Tomas 已提交
2683 2684 2685 2686
		}
	}

	/*
2687
	 * requested block isn't allocated yet;
A
Alex Tomas 已提交
2688 2689 2690
	 * we couldn't try to create block if create flag is zero
	 */
	if (!create) {
2691 2692 2693 2694
		/*
		 * put just found gap into cache to speed up
		 * subsequent requests
		 */
A
Alex Tomas 已提交
2695 2696 2697 2698
		ext4_ext_put_gap_in_cache(inode, path, iblock);
		goto out2;
	}
	/*
A
Andrew Morton 已提交
2699 2700 2701
	 * Okay, we need to do block allocation.  Lazily initialize the block
	 * allocation info here if necessary.
	 */
A
Alex Tomas 已提交
2702 2703 2704
	if (S_ISREG(inode->i_mode) && (!EXT4_I(inode)->i_block_alloc_info))
		ext4_init_block_alloc_info(inode);

2705 2706 2707 2708 2709 2710 2711 2712 2713
	/* find neighbour allocated blocks */
	ar.lleft = iblock;
	err = ext4_ext_search_left(inode, path, &ar.lleft, &ar.pleft);
	if (err)
		goto out2;
	ar.lright = iblock;
	err = ext4_ext_search_right(inode, path, &ar.lright, &ar.pright);
	if (err)
		goto out2;
A
Amit Arora 已提交
2714

2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727
	/*
	 * See if request is beyond maximum number of blocks we can have in
	 * a single extent. For an initialized extent this limit is
	 * EXT_INIT_MAX_LEN and for an uninitialized extent this limit is
	 * EXT_UNINIT_MAX_LEN.
	 */
	if (max_blocks > EXT_INIT_MAX_LEN &&
	    create != EXT4_CREATE_UNINITIALIZED_EXT)
		max_blocks = EXT_INIT_MAX_LEN;
	else if (max_blocks > EXT_UNINIT_MAX_LEN &&
		 create == EXT4_CREATE_UNINITIALIZED_EXT)
		max_blocks = EXT_UNINIT_MAX_LEN;

A
Amit Arora 已提交
2728 2729 2730 2731 2732
	/* Check if we can really insert (iblock)::(iblock+max_blocks) extent */
	newex.ee_block = cpu_to_le32(iblock);
	newex.ee_len = cpu_to_le16(max_blocks);
	err = ext4_ext_check_overlap(inode, &newex, path);
	if (err)
2733
		allocated = ext4_ext_get_actual_len(&newex);
A
Amit Arora 已提交
2734 2735
	else
		allocated = max_blocks;
2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747

	/* allocate new block */
	ar.inode = inode;
	ar.goal = ext4_ext_find_goal(inode, path, iblock);
	ar.logical = iblock;
	ar.len = allocated;
	if (S_ISREG(inode->i_mode))
		ar.flags = EXT4_MB_HINT_DATA;
	else
		/* disable in-core preallocation for non-regular files */
		ar.flags = 0;
	newblock = ext4_mb_new_blocks(handle, &ar, &err);
A
Alex Tomas 已提交
2748 2749
	if (!newblock)
		goto out2;
2750
	ext_debug("allocate new block: goal %llu, found %llu/%lu\n",
A
Alex Tomas 已提交
2751 2752 2753
			goal, newblock, allocated);

	/* try to insert new extent into found leaf and return */
2754
	ext4_ext_store_pblock(&newex, newblock);
2755
	newex.ee_len = cpu_to_le16(ar.len);
A
Amit Arora 已提交
2756 2757
	if (create == EXT4_CREATE_UNINITIALIZED_EXT)  /* Mark uninitialized */
		ext4_ext_mark_uninitialized(&newex);
A
Alex Tomas 已提交
2758
	err = ext4_ext_insert_extent(handle, inode, path, &newex);
2759 2760
	if (err) {
		/* free data blocks we just allocated */
2761 2762 2763
		/* not a good idea to call discard here directly,
		 * but otherwise we'd need to call it every free() */
		ext4_mb_discard_inode_preallocations(inode);
2764
		ext4_free_blocks(handle, inode, ext_pblock(&newex),
2765
					ext4_ext_get_actual_len(&newex), 0);
A
Alex Tomas 已提交
2766
		goto out2;
2767
	}
A
Alex Tomas 已提交
2768 2769

	/* previous routine could use block we allocated */
2770
	newblock = ext_pblock(&newex);
2771
	allocated = ext4_ext_get_actual_len(&newex);
2772
outnew:
2773 2774 2775 2776 2777 2778 2779
	if (extend_disksize) {
		disksize = ((loff_t) iblock + ar.len) << inode->i_blkbits;
		if (disksize > i_size_read(inode))
			disksize = i_size_read(inode);
		if (disksize > EXT4_I(inode)->i_disksize)
			EXT4_I(inode)->i_disksize = disksize;
	}
2780

2781
	set_buffer_new(bh_result);
A
Alex Tomas 已提交
2782

A
Amit Arora 已提交
2783 2784 2785 2786
	/* Cache only when it is _not_ an uninitialized extent */
	if (create != EXT4_CREATE_UNINITIALIZED_EXT)
		ext4_ext_put_in_cache(inode, iblock, allocated, newblock,
						EXT4_EXT_CACHE_EXTENT);
A
Alex Tomas 已提交
2787 2788 2789 2790
out:
	if (allocated > max_blocks)
		allocated = max_blocks;
	ext4_ext_show_leaf(inode, path);
2791
	set_buffer_mapped(bh_result);
A
Alex Tomas 已提交
2792 2793 2794 2795 2796 2797 2798 2799 2800 2801
	bh_result->b_bdev = inode->i_sb->s_bdev;
	bh_result->b_blocknr = newblock;
out2:
	if (path) {
		ext4_ext_drop_refs(path);
		kfree(path);
	}
	return err ? err : allocated;
}

2802
void ext4_ext_truncate(struct inode *inode)
A
Alex Tomas 已提交
2803 2804 2805
{
	struct address_space *mapping = inode->i_mapping;
	struct super_block *sb = inode->i_sb;
A
Aneesh Kumar K.V 已提交
2806
	ext4_lblk_t last_block;
A
Alex Tomas 已提交
2807 2808 2809 2810 2811 2812
	handle_t *handle;
	int err = 0;

	/*
	 * probably first extent we're gonna free will be last in block
	 */
2813
	err = ext4_writepage_trans_blocks(inode);
A
Alex Tomas 已提交
2814
	handle = ext4_journal_start(inode, err);
2815
	if (IS_ERR(handle))
A
Alex Tomas 已提交
2816 2817
		return;

2818 2819
	if (inode->i_size & (sb->s_blocksize - 1))
		ext4_block_truncate_page(handle, mapping, inode->i_size);
A
Alex Tomas 已提交
2820

2821 2822 2823
	if (ext4_orphan_add(handle, inode))
		goto out_stop;

2824
	down_write(&EXT4_I(inode)->i_data_sem);
A
Alex Tomas 已提交
2825 2826
	ext4_ext_invalidate_cache(inode);

2827
	ext4_discard_reservation(inode);
2828

A
Alex Tomas 已提交
2829
	/*
2830 2831 2832
	 * TODO: optimization is possible here.
	 * Probably we need not scan at all,
	 * because page truncation is enough.
A
Alex Tomas 已提交
2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843
	 */

	/* we have to know where to truncate from in crash case */
	EXT4_I(inode)->i_disksize = inode->i_size;
	ext4_mark_inode_dirty(handle, inode);

	last_block = (inode->i_size + sb->s_blocksize - 1)
			>> EXT4_BLOCK_SIZE_BITS(sb);
	err = ext4_ext_remove_space(inode, last_block);

	/* In a multi-transaction truncate, we only make the final
2844 2845
	 * transaction synchronous.
	 */
A
Alex Tomas 已提交
2846 2847 2848 2849
	if (IS_SYNC(inode))
		handle->h_sync = 1;

out_stop:
2850
	up_write(&EXT4_I(inode)->i_data_sem);
A
Alex Tomas 已提交
2851
	/*
2852
	 * If this was a simple ftruncate() and the file will remain alive,
A
Alex Tomas 已提交
2853 2854 2855 2856 2857 2858 2859 2860
	 * then we need to clear up the orphan record which we created above.
	 * However, if this was a real unlink then we were called by
	 * ext4_delete_inode(), and we allow that function to clean up the
	 * orphan info for us.
	 */
	if (inode->i_nlink)
		ext4_orphan_del(handle, inode);

2861 2862
	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
	ext4_mark_inode_dirty(handle, inode);
A
Alex Tomas 已提交
2863 2864 2865
	ext4_journal_stop(handle);
}

2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887
static void ext4_falloc_update_inode(struct inode *inode,
				int mode, loff_t new_size, int update_ctime)
{
	struct timespec now;

	if (update_ctime) {
		now = current_fs_time(inode->i_sb);
		if (!timespec_equal(&inode->i_ctime, &now))
			inode->i_ctime = now;
	}
	/*
	 * Update only when preallocation was requested beyond
	 * the file size.
	 */
	if (!(mode & FALLOC_FL_KEEP_SIZE) &&
				new_size > i_size_read(inode)) {
		i_size_write(inode, new_size);
		EXT4_I(inode)->i_disksize = new_size;
	}

}

A
Amit Arora 已提交
2888 2889 2890 2891 2892 2893 2894 2895 2896 2897
/*
 * preallocate space for a file. This implements ext4's fallocate inode
 * operation, which gets called from sys_fallocate system call.
 * For block-mapped files, posix_fallocate should fall back to the method
 * of writing zeroes to the required new blocks (the same behavior which is
 * expected for file systems which do not support fallocate() system call).
 */
long ext4_fallocate(struct inode *inode, int mode, loff_t offset, loff_t len)
{
	handle_t *handle;
A
Aneesh Kumar K.V 已提交
2898
	ext4_lblk_t block;
2899
	loff_t new_size;
A
Aneesh Kumar K.V 已提交
2900
	unsigned long max_blocks;
A
Amit Arora 已提交
2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918
	int ret = 0;
	int ret2 = 0;
	int retries = 0;
	struct buffer_head map_bh;
	unsigned int credits, blkbits = inode->i_blkbits;

	/*
	 * currently supporting (pre)allocate mode for extent-based
	 * files _only_
	 */
	if (!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL))
		return -EOPNOTSUPP;

	/* preallocation to directories is currently not supported */
	if (S_ISDIR(inode->i_mode))
		return -ENODEV;

	block = offset >> blkbits;
2919 2920 2921 2922
	/*
	 * We can't just convert len to max_blocks because
	 * If blocksize = 4096 offset = 3072 and len = 2048
	 */
A
Amit Arora 已提交
2923
	max_blocks = (EXT4_BLOCK_ALIGN(len + offset, blkbits) >> blkbits)
2924
							- block;
A
Amit Arora 已提交
2925
	/*
2926
	 * credits to insert 1 extent into extent tree
A
Amit Arora 已提交
2927
	 */
2928
	credits = ext4_chunk_trans_blocks(inode, max_blocks);
2929
	mutex_lock(&inode->i_mutex);
A
Amit Arora 已提交
2930 2931 2932 2933 2934 2935 2936 2937 2938
retry:
	while (ret >= 0 && ret < max_blocks) {
		block = block + ret;
		max_blocks = max_blocks - ret;
		handle = ext4_journal_start(inode, credits);
		if (IS_ERR(handle)) {
			ret = PTR_ERR(handle);
			break;
		}
2939
		ret = ext4_get_blocks_wrap(handle, inode, block,
A
Amit Arora 已提交
2940
					  max_blocks, &map_bh,
2941
					  EXT4_CREATE_UNINITIALIZED_EXT, 0, 0);
2942
		if (ret <= 0) {
2943 2944 2945 2946 2947
#ifdef EXT4FS_DEBUG
			WARN_ON(ret <= 0);
			printk(KERN_ERR "%s: ext4_ext_get_blocks "
				    "returned error inode#%lu, block=%u, "
				    "max_blocks=%lu", __func__,
2948
				    inode->i_ino, block, max_blocks);
2949
#endif
A
Amit Arora 已提交
2950 2951 2952 2953
			ext4_mark_inode_dirty(handle, inode);
			ret2 = ext4_journal_stop(handle);
			break;
		}
2954 2955 2956 2957 2958
		if ((block + ret) >= (EXT4_BLOCK_ALIGN(offset + len,
						blkbits) >> blkbits))
			new_size = offset + len;
		else
			new_size = (block + ret) << blkbits;
A
Amit Arora 已提交
2959

2960 2961
		ext4_falloc_update_inode(inode, mode, new_size,
						buffer_new(&map_bh));
A
Amit Arora 已提交
2962 2963 2964 2965 2966
		ext4_mark_inode_dirty(handle, inode);
		ret2 = ext4_journal_stop(handle);
		if (ret2)
			break;
	}
2967 2968 2969
	if (ret == -ENOSPC &&
			ext4_should_retry_alloc(inode->i_sb, &retries)) {
		ret = 0;
A
Amit Arora 已提交
2970 2971
		goto retry;
	}
2972
	mutex_unlock(&inode->i_mutex);
A
Amit Arora 已提交
2973 2974
	return ret > 0 ? ret2 : ret;
}