tree-log.c 113.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
/*
 * Copyright (C) 2008 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 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
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/sched.h>
20
#include <linux/slab.h>
21
#include <linux/blkdev.h>
J
Josef Bacik 已提交
22
#include <linux/list_sort.h>
23 24 25 26 27
#include "ctree.h"
#include "transaction.h"
#include "disk-io.h"
#include "locking.h"
#include "print-tree.h"
M
Mark Fasheh 已提交
28
#include "backref.h"
29
#include "tree-log.h"
M
Mark Fasheh 已提交
30
#include "hash.h"
31 32 33 34 35 36 37 38 39 40

/* magic values for the inode_only field in btrfs_log_inode:
 *
 * LOG_INODE_ALL means to log everything
 * LOG_INODE_EXISTS means to log just enough to recreate the inode
 * during log replay
 */
#define LOG_INODE_ALL 0
#define LOG_INODE_EXISTS 1

41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
/*
 * directory trouble cases
 *
 * 1) on rename or unlink, if the inode being unlinked isn't in the fsync
 * log, we must force a full commit before doing an fsync of the directory
 * where the unlink was done.
 * ---> record transid of last unlink/rename per directory
 *
 * mkdir foo/some_dir
 * normal commit
 * rename foo/some_dir foo2/some_dir
 * mkdir foo/some_dir
 * fsync foo/some_dir/some_file
 *
 * The fsync above will unlink the original some_dir without recording
 * it in its new location (foo2).  After a crash, some_dir will be gone
 * unless the fsync of some_file forces a full commit
 *
 * 2) we must log any new names for any file or dir that is in the fsync
 * log. ---> check inode while renaming/linking.
 *
 * 2a) we must log any new names for any file or dir during rename
 * when the directory they are being removed from was logged.
 * ---> check inode and old parent dir during rename
 *
 *  2a is actually the more important variant.  With the extra logging
 *  a crash might unlink the old name without recreating the new one
 *
 * 3) after a crash, we must go through any directories with a link count
 * of zero and redo the rm -rf
 *
 * mkdir f1/foo
 * normal commit
 * rm -rf f1/foo
 * fsync(f1)
 *
 * The directory f1 was fully removed from the FS, but fsync was never
 * called on f1, only its parent dir.  After a crash the rm -rf must
 * be replayed.  This must be able to recurse down the entire
 * directory tree.  The inode link count fixup code takes care of the
 * ugly details.
 */

84 85 86 87 88 89 90 91 92 93 94
/*
 * stages for the tree walking.  The first
 * stage (0) is to only pin down the blocks we find
 * the second stage (1) is to make sure that all the inodes
 * we find in the log are created in the subvolume.
 *
 * The last stage is to deal with directories and links and extents
 * and all the other fun semantics
 */
#define LOG_WALK_PIN_ONLY 0
#define LOG_WALK_REPLAY_INODES 1
95 96
#define LOG_WALK_REPLAY_DIR_INDEX 2
#define LOG_WALK_REPLAY_ALL 3
97

98
static int btrfs_log_inode(struct btrfs_trans_handle *trans,
99 100
			     struct btrfs_root *root, struct inode *inode,
			     int inode_only);
101 102 103
static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
			     struct btrfs_root *root,
			     struct btrfs_path *path, u64 objectid);
104 105 106 107 108
static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
				       struct btrfs_root *root,
				       struct btrfs_root *log,
				       struct btrfs_path *path,
				       u64 dirid, int del_all);
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

/*
 * tree logging is a special write ahead log used to make sure that
 * fsyncs and O_SYNCs can happen without doing full tree commits.
 *
 * Full tree commits are expensive because they require commonly
 * modified blocks to be recowed, creating many dirty pages in the
 * extent tree an 4x-6x higher write load than ext3.
 *
 * Instead of doing a tree commit on every fsync, we use the
 * key ranges and transaction ids to find items for a given file or directory
 * that have changed in this transaction.  Those items are copied into
 * a special tree (one per subvolume root), that tree is written to disk
 * and then the fsync is considered complete.
 *
 * After a crash, items are copied out of the log-tree back into the
 * subvolume tree.  Any file data extents found are recorded in the extent
 * allocation tree, and the log-tree freed.
 *
 * The log tree is read three times, once to pin down all the extents it is
 * using in ram and once, once to create all the inodes logged in the tree
 * and once to do all the other items.
 */

/*
 * start a sub transaction and setup the log tree
 * this increments the log tree writer count to make the people
 * syncing the tree wait for us to finish
 */
static int start_log_trans(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root)
{
	int ret;
Y
Yan Zheng 已提交
142 143 144

	mutex_lock(&root->log_mutex);
	if (root->log_root) {
145 146 147 148 149 150 151
		if (!root->log_start_pid) {
			root->log_start_pid = current->pid;
			root->log_multiple_pids = false;
		} else if (root->log_start_pid != current->pid) {
			root->log_multiple_pids = true;
		}

M
Miao Xie 已提交
152
		atomic_inc(&root->log_batch);
Y
Yan Zheng 已提交
153 154 155 156
		atomic_inc(&root->log_writers);
		mutex_unlock(&root->log_mutex);
		return 0;
	}
157 158

	ret = 0;
159
	mutex_lock(&root->fs_info->tree_log_mutex);
160
	if (!root->fs_info->log_root_tree)
161
		ret = btrfs_init_log_root_tree(trans, root->fs_info);
162 163 164 165 166
	mutex_unlock(&root->fs_info->tree_log_mutex);
	if (ret)
		goto out;

	if (!root->log_root) {
167
		ret = btrfs_add_log_tree(trans, root);
168
		if (ret)
169
			goto out;
170
	}
171 172
	root->log_multiple_pids = false;
	root->log_start_pid = current->pid;
M
Miao Xie 已提交
173
	atomic_inc(&root->log_batch);
Y
Yan Zheng 已提交
174
	atomic_inc(&root->log_writers);
175
out:
Y
Yan Zheng 已提交
176
	mutex_unlock(&root->log_mutex);
177
	return ret;
178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
}

/*
 * returns 0 if there was a log transaction running and we were able
 * to join, or returns -ENOENT if there were not transactions
 * in progress
 */
static int join_running_log_trans(struct btrfs_root *root)
{
	int ret = -ENOENT;

	smp_mb();
	if (!root->log_root)
		return -ENOENT;

Y
Yan Zheng 已提交
193
	mutex_lock(&root->log_mutex);
194 195
	if (root->log_root) {
		ret = 0;
Y
Yan Zheng 已提交
196
		atomic_inc(&root->log_writers);
197
	}
Y
Yan Zheng 已提交
198
	mutex_unlock(&root->log_mutex);
199 200 201
	return ret;
}

202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
/*
 * This either makes the current running log transaction wait
 * until you call btrfs_end_log_trans() or it makes any future
 * log transactions wait until you call btrfs_end_log_trans()
 */
int btrfs_pin_log_trans(struct btrfs_root *root)
{
	int ret = -ENOENT;

	mutex_lock(&root->log_mutex);
	atomic_inc(&root->log_writers);
	mutex_unlock(&root->log_mutex);
	return ret;
}

217 218 219 220
/*
 * indicate we're done making changes to the log tree
 * and wake up anyone waiting to do a sync
 */
221
void btrfs_end_log_trans(struct btrfs_root *root)
222
{
Y
Yan Zheng 已提交
223 224 225 226 227
	if (atomic_dec_and_test(&root->log_writers)) {
		smp_mb();
		if (waitqueue_active(&root->log_writer_wait))
			wake_up(&root->log_writer_wait);
	}
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
}


/*
 * the walk control struct is used to pass state down the chain when
 * processing the log tree.  The stage field tells us which part
 * of the log tree processing we are currently doing.  The others
 * are state fields used for that specific part
 */
struct walk_control {
	/* should we free the extent on disk when done?  This is used
	 * at transaction commit time while freeing a log tree
	 */
	int free;

	/* should we write out the extent buffer?  This is used
	 * while flushing the log tree to disk during a sync
	 */
	int write;

	/* should we wait for the extent buffer io to finish?  Also used
	 * while flushing the log tree to disk for a sync
	 */
	int wait;

	/* pin only walk, we record which extents on disk belong to the
	 * log trees
	 */
	int pin;

	/* what stage of the replay code we're currently in */
	int stage;

	/* the root we are currently replaying */
	struct btrfs_root *replay_dest;

	/* the trans handle for the current replay */
	struct btrfs_trans_handle *trans;

	/* the function that gets used to process blocks we find in the
	 * tree.  Note the extent_buffer might not be up to date when it is
	 * passed in, and it must be checked or read if you need the data
	 * inside it
	 */
	int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
			    struct walk_control *wc, u64 gen);
};

/*
 * process_func used to pin down extents, write them or wait on them
 */
static int process_one_buffer(struct btrfs_root *log,
			      struct extent_buffer *eb,
			      struct walk_control *wc, u64 gen)
{
283 284
	int ret = 0;

285 286 287 288 289 290 291 292 293 294
	/*
	 * If this fs is mixed then we need to be able to process the leaves to
	 * pin down any logged extents, so we have to read the block.
	 */
	if (btrfs_fs_incompat(log->fs_info, MIXED_GROUPS)) {
		ret = btrfs_read_buffer(eb, gen);
		if (ret)
			return ret;
	}

J
Josef Bacik 已提交
295
	if (wc->pin)
296 297
		ret = btrfs_pin_extent_for_log_replay(log->fs_info->extent_root,
						      eb->start, eb->len);
298

299
	if (!ret && btrfs_buffer_uptodate(eb, gen, 0)) {
300 301
		if (wc->pin && btrfs_header_level(eb) == 0)
			ret = btrfs_exclude_logged_extents(log, eb);
302 303 304 305 306
		if (wc->write)
			btrfs_write_tree_block(eb);
		if (wc->wait)
			btrfs_wait_tree_block_writeback(eb);
	}
307
	return ret;
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
}

/*
 * Item overwrite used by replay and tree logging.  eb, slot and key all refer
 * to the src data we are copying out.
 *
 * root is the tree we are copying into, and path is a scratch
 * path for use in this function (it should be released on entry and
 * will be released on exit).
 *
 * If the key is already in the destination tree the existing item is
 * overwritten.  If the existing item isn't big enough, it is extended.
 * If it is too large, it is truncated.
 *
 * If the key isn't in the destination yet, a new item is inserted.
 */
static noinline int overwrite_item(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct btrfs_path *path,
				   struct extent_buffer *eb, int slot,
				   struct btrfs_key *key)
{
	int ret;
	u32 item_size;
	u64 saved_i_size = 0;
	int save_old_i_size = 0;
	unsigned long src_ptr;
	unsigned long dst_ptr;
	int overwrite_root = 0;
337
	bool inode_item = key->type == BTRFS_INODE_ITEM_KEY;
338 339 340 341 342 343 344 345 346

	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
		overwrite_root = 1;

	item_size = btrfs_item_size_nr(eb, slot);
	src_ptr = btrfs_item_ptr_offset(eb, slot);

	/* look for the key in the destination tree */
	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
347 348 349
	if (ret < 0)
		return ret;

350 351 352 353 354 355 356 357 358
	if (ret == 0) {
		char *src_copy;
		char *dst_copy;
		u32 dst_size = btrfs_item_size_nr(path->nodes[0],
						  path->slots[0]);
		if (dst_size != item_size)
			goto insert;

		if (item_size == 0) {
359
			btrfs_release_path(path);
360 361 362 363
			return 0;
		}
		dst_copy = kmalloc(item_size, GFP_NOFS);
		src_copy = kmalloc(item_size, GFP_NOFS);
364
		if (!dst_copy || !src_copy) {
365
			btrfs_release_path(path);
366 367 368 369
			kfree(dst_copy);
			kfree(src_copy);
			return -ENOMEM;
		}
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386

		read_extent_buffer(eb, src_copy, src_ptr, item_size);

		dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
		read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
				   item_size);
		ret = memcmp(dst_copy, src_copy, item_size);

		kfree(dst_copy);
		kfree(src_copy);
		/*
		 * they have the same contents, just return, this saves
		 * us from cowing blocks in the destination tree and doing
		 * extra writes that may not have been done by a previous
		 * sync
		 */
		if (ret == 0) {
387
			btrfs_release_path(path);
388 389 390
			return 0;
		}

391 392 393 394 395 396 397
		/*
		 * We need to load the old nbytes into the inode so when we
		 * replay the extents we've logged we get the right nbytes.
		 */
		if (inode_item) {
			struct btrfs_inode_item *item;
			u64 nbytes;
398
			u32 mode;
399 400 401 402 403 404 405

			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
					      struct btrfs_inode_item);
			nbytes = btrfs_inode_nbytes(path->nodes[0], item);
			item = btrfs_item_ptr(eb, slot,
					      struct btrfs_inode_item);
			btrfs_set_inode_nbytes(eb, item, nbytes);
406 407 408 409 410 411 412 413 414

			/*
			 * If this is a directory we need to reset the i_size to
			 * 0 so that we can set it up properly when replaying
			 * the rest of the items in this log.
			 */
			mode = btrfs_inode_mode(eb, item);
			if (S_ISDIR(mode))
				btrfs_set_inode_size(eb, item, 0);
415 416 417
		}
	} else if (inode_item) {
		struct btrfs_inode_item *item;
418
		u32 mode;
419 420 421 422 423 424 425

		/*
		 * New inode, set nbytes to 0 so that the nbytes comes out
		 * properly when we replay the extents.
		 */
		item = btrfs_item_ptr(eb, slot, struct btrfs_inode_item);
		btrfs_set_inode_nbytes(eb, item, 0);
426 427 428 429 430 431 432 433 434

		/*
		 * If this is a directory we need to reset the i_size to 0 so
		 * that we can set it up properly when replaying the rest of
		 * the items in this log.
		 */
		mode = btrfs_inode_mode(eb, item);
		if (S_ISDIR(mode))
			btrfs_set_inode_size(eb, item, 0);
435 436
	}
insert:
437
	btrfs_release_path(path);
438 439 440 441 442 443 444 445 446
	/* try to insert the key into the destination tree */
	ret = btrfs_insert_empty_item(trans, root, path,
				      key, item_size);

	/* make sure any existing item is the correct size */
	if (ret == -EEXIST) {
		u32 found_size;
		found_size = btrfs_item_size_nr(path->nodes[0],
						path->slots[0]);
447
		if (found_size > item_size)
448
			btrfs_truncate_item(root, path, item_size, 1);
449
		else if (found_size < item_size)
450
			btrfs_extend_item(root, path,
451
					  item_size - found_size);
452
	} else if (ret) {
453
		return ret;
454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
	}
	dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
					path->slots[0]);

	/* don't overwrite an existing inode if the generation number
	 * was logged as zero.  This is done when the tree logging code
	 * is just logging an inode to make sure it exists after recovery.
	 *
	 * Also, don't overwrite i_size on directories during replay.
	 * log replay inserts and removes directory items based on the
	 * state of the tree found in the subvolume, and i_size is modified
	 * as it goes
	 */
	if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
		struct btrfs_inode_item *src_item;
		struct btrfs_inode_item *dst_item;

		src_item = (struct btrfs_inode_item *)src_ptr;
		dst_item = (struct btrfs_inode_item *)dst_ptr;

		if (btrfs_inode_generation(eb, src_item) == 0)
			goto no_copy;

		if (overwrite_root &&
		    S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
		    S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
			save_old_i_size = 1;
			saved_i_size = btrfs_inode_size(path->nodes[0],
							dst_item);
		}
	}

	copy_extent_buffer(path->nodes[0], eb, dst_ptr,
			   src_ptr, item_size);

	if (save_old_i_size) {
		struct btrfs_inode_item *dst_item;
		dst_item = (struct btrfs_inode_item *)dst_ptr;
		btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
	}

	/* make sure the generation is filled in */
	if (key->type == BTRFS_INODE_ITEM_KEY) {
		struct btrfs_inode_item *dst_item;
		dst_item = (struct btrfs_inode_item *)dst_ptr;
		if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
			btrfs_set_inode_generation(path->nodes[0], dst_item,
						   trans->transid);
		}
	}
no_copy:
	btrfs_mark_buffer_dirty(path->nodes[0]);
506
	btrfs_release_path(path);
507 508 509 510 511 512 513 514 515 516
	return 0;
}

/*
 * simple helper to read an inode off the disk from a given root
 * This can only be called for subvolume roots and not for the log
 */
static noinline struct inode *read_one_inode(struct btrfs_root *root,
					     u64 objectid)
{
517
	struct btrfs_key key;
518 519
	struct inode *inode;

520 521 522
	key.objectid = objectid;
	key.type = BTRFS_INODE_ITEM_KEY;
	key.offset = 0;
523
	inode = btrfs_iget(root->fs_info->sb, &key, root, NULL);
524 525 526
	if (IS_ERR(inode)) {
		inode = NULL;
	} else if (is_bad_inode(inode)) {
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553
		iput(inode);
		inode = NULL;
	}
	return inode;
}

/* replays a single extent in 'eb' at 'slot' with 'key' into the
 * subvolume 'root'.  path is released on entry and should be released
 * on exit.
 *
 * extents in the log tree have not been allocated out of the extent
 * tree yet.  So, this completes the allocation, taking a reference
 * as required if the extent already exists or creating a new extent
 * if it isn't in the extent allocation tree yet.
 *
 * The extent is inserted into the file, dropping any existing extents
 * from the file that overlap the new one.
 */
static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_path *path,
				      struct extent_buffer *eb, int slot,
				      struct btrfs_key *key)
{
	int found_type;
	u64 extent_end;
	u64 start = key->offset;
554
	u64 nbytes = 0;
555 556 557 558 559 560 561 562
	struct btrfs_file_extent_item *item;
	struct inode *inode = NULL;
	unsigned long size;
	int ret = 0;

	item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
	found_type = btrfs_file_extent_type(eb, item);

Y
Yan Zheng 已提交
563
	if (found_type == BTRFS_FILE_EXTENT_REG ||
564 565 566 567 568 569 570 571 572 573 574
	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
		nbytes = btrfs_file_extent_num_bytes(eb, item);
		extent_end = start + nbytes;

		/*
		 * We don't add to the inodes nbytes if we are prealloc or a
		 * hole.
		 */
		if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
			nbytes = 0;
	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
575
		size = btrfs_file_extent_inline_len(eb, slot, item);
576
		nbytes = btrfs_file_extent_ram_bytes(eb, item);
577
		extent_end = ALIGN(start + size, root->sectorsize);
578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
	} else {
		ret = 0;
		goto out;
	}

	inode = read_one_inode(root, key->objectid);
	if (!inode) {
		ret = -EIO;
		goto out;
	}

	/*
	 * first check to see if we already have this extent in the
	 * file.  This must be done before the btrfs_drop_extents run
	 * so we don't try to drop this extent.
	 */
L
Li Zefan 已提交
594
	ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
595 596
				       start, 0);

Y
Yan Zheng 已提交
597 598 599
	if (ret == 0 &&
	    (found_type == BTRFS_FILE_EXTENT_REG ||
	     found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618
		struct btrfs_file_extent_item cmp1;
		struct btrfs_file_extent_item cmp2;
		struct btrfs_file_extent_item *existing;
		struct extent_buffer *leaf;

		leaf = path->nodes[0];
		existing = btrfs_item_ptr(leaf, path->slots[0],
					  struct btrfs_file_extent_item);

		read_extent_buffer(eb, &cmp1, (unsigned long)item,
				   sizeof(cmp1));
		read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
				   sizeof(cmp2));

		/*
		 * we already have a pointer to this exact extent,
		 * we don't have to do anything
		 */
		if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
619
			btrfs_release_path(path);
620 621 622
			goto out;
		}
	}
623
	btrfs_release_path(path);
624 625

	/* drop any overlapping extents */
626
	ret = btrfs_drop_extents(trans, root, inode, start, extent_end, 1);
627 628
	if (ret)
		goto out;
629

Y
Yan Zheng 已提交
630 631
	if (found_type == BTRFS_FILE_EXTENT_REG ||
	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
632
		u64 offset;
Y
Yan Zheng 已提交
633 634 635 636 637
		unsigned long dest_offset;
		struct btrfs_key ins;

		ret = btrfs_insert_empty_item(trans, root, path, key,
					      sizeof(*item));
638 639
		if (ret)
			goto out;
Y
Yan Zheng 已提交
640 641 642 643 644 645 646 647
		dest_offset = btrfs_item_ptr_offset(path->nodes[0],
						    path->slots[0]);
		copy_extent_buffer(path->nodes[0], eb, dest_offset,
				(unsigned long)item,  sizeof(*item));

		ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
		ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
		ins.type = BTRFS_EXTENT_ITEM_KEY;
648
		offset = key->offset - btrfs_file_extent_offset(eb, item);
Y
Yan Zheng 已提交
649 650 651 652 653 654 655 656 657 658 659 660 661 662

		if (ins.objectid > 0) {
			u64 csum_start;
			u64 csum_end;
			LIST_HEAD(ordered_sums);
			/*
			 * is this extent already allocated in the extent
			 * allocation tree?  If so, just add a reference
			 */
			ret = btrfs_lookup_extent(root, ins.objectid,
						ins.offset);
			if (ret == 0) {
				ret = btrfs_inc_extent_ref(trans, root,
						ins.objectid, ins.offset,
663
						0, root->root_key.objectid,
A
Arne Jansen 已提交
664
						key->objectid, offset, 0);
665 666
				if (ret)
					goto out;
Y
Yan Zheng 已提交
667 668 669 670 671
			} else {
				/*
				 * insert the extent pointer in the extent
				 * allocation tree
				 */
672 673 674
				ret = btrfs_alloc_logged_file_extent(trans,
						root, root->root_key.objectid,
						key->objectid, offset, &ins);
675 676
				if (ret)
					goto out;
Y
Yan Zheng 已提交
677
			}
678
			btrfs_release_path(path);
Y
Yan Zheng 已提交
679 680 681 682 683 684 685 686 687 688 689 690 691

			if (btrfs_file_extent_compression(eb, item)) {
				csum_start = ins.objectid;
				csum_end = csum_start + ins.offset;
			} else {
				csum_start = ins.objectid +
					btrfs_file_extent_offset(eb, item);
				csum_end = csum_start +
					btrfs_file_extent_num_bytes(eb, item);
			}

			ret = btrfs_lookup_csums_range(root->log_root,
						csum_start, csum_end - 1,
A
Arne Jansen 已提交
692
						&ordered_sums, 0);
693 694
			if (ret)
				goto out;
Y
Yan Zheng 已提交
695 696 697 698 699
			while (!list_empty(&ordered_sums)) {
				struct btrfs_ordered_sum *sums;
				sums = list_entry(ordered_sums.next,
						struct btrfs_ordered_sum,
						list);
700 701
				if (!ret)
					ret = btrfs_csum_file_blocks(trans,
Y
Yan Zheng 已提交
702 703 704 705 706
						root->fs_info->csum_root,
						sums);
				list_del(&sums->list);
				kfree(sums);
			}
707 708
			if (ret)
				goto out;
Y
Yan Zheng 已提交
709
		} else {
710
			btrfs_release_path(path);
Y
Yan Zheng 已提交
711 712 713 714
		}
	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
		/* inline extents are easy, we just overwrite them */
		ret = overwrite_item(trans, root, path, eb, slot, key);
715 716
		if (ret)
			goto out;
Y
Yan Zheng 已提交
717
	}
718

719
	inode_add_bytes(inode, nbytes);
720
	ret = btrfs_update_inode(trans, root, inode);
721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752
out:
	if (inode)
		iput(inode);
	return ret;
}

/*
 * when cleaning up conflicts between the directory names in the
 * subvolume, directory names in the log and directory names in the
 * inode back references, we may have to unlink inodes from directories.
 *
 * This is a helper function to do the unlink of a specific directory
 * item
 */
static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_path *path,
				      struct inode *dir,
				      struct btrfs_dir_item *di)
{
	struct inode *inode;
	char *name;
	int name_len;
	struct extent_buffer *leaf;
	struct btrfs_key location;
	int ret;

	leaf = path->nodes[0];

	btrfs_dir_item_key_to_cpu(leaf, di, &location);
	name_len = btrfs_dir_name_len(leaf, di);
	name = kmalloc(name_len, GFP_NOFS);
753 754 755
	if (!name)
		return -ENOMEM;

756
	read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
757
	btrfs_release_path(path);
758 759

	inode = read_one_inode(root, location.objectid);
760
	if (!inode) {
761 762
		ret = -EIO;
		goto out;
763
	}
764

765
	ret = link_to_fixup_dir(trans, root, path, location.objectid);
766 767
	if (ret)
		goto out;
768

769
	ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
770 771
	if (ret)
		goto out;
772 773
	else
		ret = btrfs_run_delayed_items(trans, root);
774
out:
775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801
	kfree(name);
	iput(inode);
	return ret;
}

/*
 * helper function to see if a given name and sequence number found
 * in an inode back reference are already in a directory and correctly
 * point to this inode
 */
static noinline int inode_in_dir(struct btrfs_root *root,
				 struct btrfs_path *path,
				 u64 dirid, u64 objectid, u64 index,
				 const char *name, int name_len)
{
	struct btrfs_dir_item *di;
	struct btrfs_key location;
	int match = 0;

	di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
					 index, name, name_len, 0);
	if (di && !IS_ERR(di)) {
		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
		if (location.objectid != objectid)
			goto out;
	} else
		goto out;
802
	btrfs_release_path(path);
803 804 805 806 807 808 809 810 811 812

	di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
	if (di && !IS_ERR(di)) {
		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
		if (location.objectid != objectid)
			goto out;
	} else
		goto out;
	match = 1;
out:
813
	btrfs_release_path(path);
814 815 816 817 818 819 820 821 822 823 824 825 826 827 828
	return match;
}

/*
 * helper function to check a log tree for a named back reference in
 * an inode.  This is used to decide if a back reference that is
 * found in the subvolume conflicts with what we find in the log.
 *
 * inode backreferences may have multiple refs in a single item,
 * during replay we process one reference at a time, and we don't
 * want to delete valid links to a file from the subvolume if that
 * link is also in the log.
 */
static noinline int backref_in_log(struct btrfs_root *log,
				   struct btrfs_key *key,
M
Mark Fasheh 已提交
829
				   u64 ref_objectid,
830 831 832 833 834 835 836 837 838 839 840 841 842
				   char *name, int namelen)
{
	struct btrfs_path *path;
	struct btrfs_inode_ref *ref;
	unsigned long ptr;
	unsigned long ptr_end;
	unsigned long name_ptr;
	int found_name_len;
	int item_size;
	int ret;
	int match = 0;

	path = btrfs_alloc_path();
843 844 845
	if (!path)
		return -ENOMEM;

846 847 848 849 850
	ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
	if (ret != 0)
		goto out;

	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
M
Mark Fasheh 已提交
851 852 853 854 855 856 857 858 859 860

	if (key->type == BTRFS_INODE_EXTREF_KEY) {
		if (btrfs_find_name_in_ext_backref(path, ref_objectid,
						   name, namelen, NULL))
			match = 1;

		goto out;
	}

	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880
	ptr_end = ptr + item_size;
	while (ptr < ptr_end) {
		ref = (struct btrfs_inode_ref *)ptr;
		found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
		if (found_name_len == namelen) {
			name_ptr = (unsigned long)(ref + 1);
			ret = memcmp_extent_buffer(path->nodes[0], name,
						   name_ptr, namelen);
			if (ret == 0) {
				match = 1;
				goto out;
			}
		}
		ptr = (unsigned long)(ref + 1) + found_name_len;
	}
out:
	btrfs_free_path(path);
	return match;
}

881
static inline int __add_inode_ref(struct btrfs_trans_handle *trans,
882 883
				  struct btrfs_root *root,
				  struct btrfs_path *path,
884 885 886
				  struct btrfs_root *log_root,
				  struct inode *dir, struct inode *inode,
				  struct extent_buffer *eb,
M
Mark Fasheh 已提交
887 888 889
				  u64 inode_objectid, u64 parent_objectid,
				  u64 ref_index, char *name, int namelen,
				  int *search_done)
890
{
L
liubo 已提交
891
	int ret;
M
Mark Fasheh 已提交
892 893 894
	char *victim_name;
	int victim_name_len;
	struct extent_buffer *leaf;
895
	struct btrfs_dir_item *di;
M
Mark Fasheh 已提交
896 897
	struct btrfs_key search_key;
	struct btrfs_inode_extref *extref;
898

M
Mark Fasheh 已提交
899 900 901 902 903 904
again:
	/* Search old style refs */
	search_key.objectid = inode_objectid;
	search_key.type = BTRFS_INODE_REF_KEY;
	search_key.offset = parent_objectid;
	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
905 906 907 908
	if (ret == 0) {
		struct btrfs_inode_ref *victim_ref;
		unsigned long ptr;
		unsigned long ptr_end;
M
Mark Fasheh 已提交
909 910

		leaf = path->nodes[0];
911 912 913 914

		/* are we trying to overwrite a back ref for the root directory
		 * if so, just jump out, we're done
		 */
M
Mark Fasheh 已提交
915
		if (search_key.objectid == search_key.offset)
916
			return 1;
917 918 919 920 921 922 923

		/* check all the names in this back reference to see
		 * if they are in the log.  if so, we allow them to stay
		 * otherwise they must be unlinked as a conflict
		 */
		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
		ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
C
Chris Mason 已提交
924
		while (ptr < ptr_end) {
925 926 927 928
			victim_ref = (struct btrfs_inode_ref *)ptr;
			victim_name_len = btrfs_inode_ref_name_len(leaf,
								   victim_ref);
			victim_name = kmalloc(victim_name_len, GFP_NOFS);
929 930
			if (!victim_name)
				return -ENOMEM;
931 932 933 934 935

			read_extent_buffer(leaf, victim_name,
					   (unsigned long)(victim_ref + 1),
					   victim_name_len);

M
Mark Fasheh 已提交
936 937 938
			if (!backref_in_log(log_root, &search_key,
					    parent_objectid,
					    victim_name,
939
					    victim_name_len)) {
Z
Zach Brown 已提交
940
				inc_nlink(inode);
941
				btrfs_release_path(path);
942

943 944 945
				ret = btrfs_unlink_inode(trans, root, dir,
							 inode, victim_name,
							 victim_name_len);
M
Mark Fasheh 已提交
946
				kfree(victim_name);
947 948
				if (ret)
					return ret;
949 950 951
				ret = btrfs_run_delayed_items(trans, root);
				if (ret)
					return ret;
M
Mark Fasheh 已提交
952 953
				*search_done = 1;
				goto again;
954 955
			}
			kfree(victim_name);
M
Mark Fasheh 已提交
956

957 958 959
			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
		}

960 961 962 963
		/*
		 * NOTE: we have searched root tree and checked the
		 * coresponding ref, it does not need to check again.
		 */
964
		*search_done = 1;
965
	}
966
	btrfs_release_path(path);
967

M
Mark Fasheh 已提交
968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991
	/* Same search but for extended refs */
	extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
					   inode_objectid, parent_objectid, 0,
					   0);
	if (!IS_ERR_OR_NULL(extref)) {
		u32 item_size;
		u32 cur_offset = 0;
		unsigned long base;
		struct inode *victim_parent;

		leaf = path->nodes[0];

		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
		base = btrfs_item_ptr_offset(leaf, path->slots[0]);

		while (cur_offset < item_size) {
			extref = (struct btrfs_inode_extref *)base + cur_offset;

			victim_name_len = btrfs_inode_extref_name_len(leaf, extref);

			if (btrfs_inode_extref_parent(leaf, extref) != parent_objectid)
				goto next;

			victim_name = kmalloc(victim_name_len, GFP_NOFS);
992 993
			if (!victim_name)
				return -ENOMEM;
M
Mark Fasheh 已提交
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
			read_extent_buffer(leaf, victim_name, (unsigned long)&extref->name,
					   victim_name_len);

			search_key.objectid = inode_objectid;
			search_key.type = BTRFS_INODE_EXTREF_KEY;
			search_key.offset = btrfs_extref_hash(parent_objectid,
							      victim_name,
							      victim_name_len);
			ret = 0;
			if (!backref_in_log(log_root, &search_key,
					    parent_objectid, victim_name,
					    victim_name_len)) {
				ret = -ENOENT;
				victim_parent = read_one_inode(root,
							       parent_objectid);
				if (victim_parent) {
Z
Zach Brown 已提交
1010
					inc_nlink(inode);
M
Mark Fasheh 已提交
1011 1012 1013 1014 1015 1016 1017
					btrfs_release_path(path);

					ret = btrfs_unlink_inode(trans, root,
								 victim_parent,
								 inode,
								 victim_name,
								 victim_name_len);
1018 1019 1020
					if (!ret)
						ret = btrfs_run_delayed_items(
								  trans, root);
M
Mark Fasheh 已提交
1021 1022 1023
				}
				iput(victim_parent);
				kfree(victim_name);
1024 1025
				if (ret)
					return ret;
M
Mark Fasheh 已提交
1026 1027 1028 1029
				*search_done = 1;
				goto again;
			}
			kfree(victim_name);
1030 1031
			if (ret)
				return ret;
M
Mark Fasheh 已提交
1032 1033 1034 1035 1036 1037 1038
next:
			cur_offset += victim_name_len + sizeof(*extref);
		}
		*search_done = 1;
	}
	btrfs_release_path(path);

L
liubo 已提交
1039 1040
	/* look for a conflicting sequence number */
	di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
M
Mark Fasheh 已提交
1041
					 ref_index, name, namelen, 0);
L
liubo 已提交
1042 1043
	if (di && !IS_ERR(di)) {
		ret = drop_one_dir_item(trans, root, path, dir, di);
1044 1045
		if (ret)
			return ret;
L
liubo 已提交
1046 1047 1048 1049 1050 1051 1052 1053
	}
	btrfs_release_path(path);

	/* look for a conflicing name */
	di = btrfs_lookup_dir_item(trans, root, path, btrfs_ino(dir),
				   name, namelen, 0);
	if (di && !IS_ERR(di)) {
		ret = drop_one_dir_item(trans, root, path, dir, di);
1054 1055
		if (ret)
			return ret;
L
liubo 已提交
1056 1057 1058
	}
	btrfs_release_path(path);

1059 1060
	return 0;
}
1061

M
Mark Fasheh 已提交
1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103
static int extref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
			     u32 *namelen, char **name, u64 *index,
			     u64 *parent_objectid)
{
	struct btrfs_inode_extref *extref;

	extref = (struct btrfs_inode_extref *)ref_ptr;

	*namelen = btrfs_inode_extref_name_len(eb, extref);
	*name = kmalloc(*namelen, GFP_NOFS);
	if (*name == NULL)
		return -ENOMEM;

	read_extent_buffer(eb, *name, (unsigned long)&extref->name,
			   *namelen);

	*index = btrfs_inode_extref_index(eb, extref);
	if (parent_objectid)
		*parent_objectid = btrfs_inode_extref_parent(eb, extref);

	return 0;
}

static int ref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
			  u32 *namelen, char **name, u64 *index)
{
	struct btrfs_inode_ref *ref;

	ref = (struct btrfs_inode_ref *)ref_ptr;

	*namelen = btrfs_inode_ref_name_len(eb, ref);
	*name = kmalloc(*namelen, GFP_NOFS);
	if (*name == NULL)
		return -ENOMEM;

	read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);

	*index = btrfs_inode_ref_index(eb, ref);

	return 0;
}

1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
/*
 * replay one inode back reference item found in the log tree.
 * eb, slot and key refer to the buffer and key found in the log tree.
 * root is the destination we are replaying into, and path is for temp
 * use by this function.  (it should be released on return).
 */
static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root,
				  struct btrfs_root *log,
				  struct btrfs_path *path,
				  struct extent_buffer *eb, int slot,
				  struct btrfs_key *key)
{
1117 1118
	struct inode *dir = NULL;
	struct inode *inode = NULL;
1119 1120
	unsigned long ref_ptr;
	unsigned long ref_end;
1121
	char *name = NULL;
1122 1123 1124
	int namelen;
	int ret;
	int search_done = 0;
M
Mark Fasheh 已提交
1125 1126 1127
	int log_ref_ver = 0;
	u64 parent_objectid;
	u64 inode_objectid;
1128
	u64 ref_index = 0;
M
Mark Fasheh 已提交
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
	int ref_struct_size;

	ref_ptr = btrfs_item_ptr_offset(eb, slot);
	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);

	if (key->type == BTRFS_INODE_EXTREF_KEY) {
		struct btrfs_inode_extref *r;

		ref_struct_size = sizeof(struct btrfs_inode_extref);
		log_ref_ver = 1;
		r = (struct btrfs_inode_extref *)ref_ptr;
		parent_objectid = btrfs_inode_extref_parent(eb, r);
	} else {
		ref_struct_size = sizeof(struct btrfs_inode_ref);
		parent_objectid = key->offset;
	}
	inode_objectid = key->objectid;
1146

1147 1148 1149 1150 1151 1152
	/*
	 * it is possible that we didn't log all the parent directories
	 * for a given inode.  If we don't find the dir, just don't
	 * copy the back ref in.  The link count fixup code will take
	 * care of the rest
	 */
M
Mark Fasheh 已提交
1153
	dir = read_one_inode(root, parent_objectid);
1154 1155 1156 1157
	if (!dir) {
		ret = -ENOENT;
		goto out;
	}
1158

M
Mark Fasheh 已提交
1159
	inode = read_one_inode(root, inode_objectid);
1160
	if (!inode) {
1161 1162
		ret = -EIO;
		goto out;
1163 1164 1165
	}

	while (ref_ptr < ref_end) {
M
Mark Fasheh 已提交
1166 1167 1168 1169 1170 1171 1172 1173 1174
		if (log_ref_ver) {
			ret = extref_get_fields(eb, ref_ptr, &namelen, &name,
						&ref_index, &parent_objectid);
			/*
			 * parent object can change from one array
			 * item to another.
			 */
			if (!dir)
				dir = read_one_inode(root, parent_objectid);
1175 1176 1177 1178
			if (!dir) {
				ret = -ENOENT;
				goto out;
			}
M
Mark Fasheh 已提交
1179 1180 1181 1182 1183
		} else {
			ret = ref_get_fields(eb, ref_ptr, &namelen, &name,
					     &ref_index);
		}
		if (ret)
1184
			goto out;
1185 1186 1187

		/* if we already have a perfect match, we're done */
		if (!inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
M
Mark Fasheh 已提交
1188
				  ref_index, name, namelen)) {
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
			/*
			 * look for a conflicting back reference in the
			 * metadata. if we find one we have to unlink that name
			 * of the file before we add our new link.  Later on, we
			 * overwrite any existing back reference, and we don't
			 * want to create dangling pointers in the directory.
			 */

			if (!search_done) {
				ret = __add_inode_ref(trans, root, path, log,
M
Mark Fasheh 已提交
1199 1200 1201 1202
						      dir, inode, eb,
						      inode_objectid,
						      parent_objectid,
						      ref_index, name, namelen,
1203
						      &search_done);
1204 1205 1206
				if (ret) {
					if (ret == 1)
						ret = 0;
1207 1208
					goto out;
				}
1209 1210 1211 1212
			}

			/* insert our name */
			ret = btrfs_add_link(trans, dir, inode, name, namelen,
M
Mark Fasheh 已提交
1213
					     0, ref_index);
1214 1215
			if (ret)
				goto out;
1216 1217 1218 1219

			btrfs_update_inode(trans, root, inode);
		}

M
Mark Fasheh 已提交
1220
		ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
1221
		kfree(name);
1222
		name = NULL;
M
Mark Fasheh 已提交
1223 1224 1225 1226
		if (log_ref_ver) {
			iput(dir);
			dir = NULL;
		}
1227
	}
1228 1229 1230

	/* finally write the back reference in the inode */
	ret = overwrite_item(trans, root, path, eb, slot, key);
1231
out:
1232
	btrfs_release_path(path);
1233
	kfree(name);
1234 1235
	iput(dir);
	iput(inode);
1236
	return ret;
1237 1238
}

1239 1240 1241 1242
static int insert_orphan_item(struct btrfs_trans_handle *trans,
			      struct btrfs_root *root, u64 offset)
{
	int ret;
1243 1244
	ret = btrfs_find_item(root, NULL, BTRFS_ORPHAN_OBJECTID,
			offset, BTRFS_ORPHAN_ITEM_KEY, NULL);
1245 1246 1247 1248 1249
	if (ret > 0)
		ret = btrfs_insert_orphan_item(trans, root, offset);
	return ret;
}

M
Mark Fasheh 已提交
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
static int count_inode_extrefs(struct btrfs_root *root,
			       struct inode *inode, struct btrfs_path *path)
{
	int ret = 0;
	int name_len;
	unsigned int nlink = 0;
	u32 item_size;
	u32 cur_offset = 0;
	u64 inode_objectid = btrfs_ino(inode);
	u64 offset = 0;
	unsigned long ptr;
	struct btrfs_inode_extref *extref;
	struct extent_buffer *leaf;

	while (1) {
		ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
					    &extref, &offset);
		if (ret)
			break;
1269

M
Mark Fasheh 已提交
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
		leaf = path->nodes[0];
		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);

		while (cur_offset < item_size) {
			extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
			name_len = btrfs_inode_extref_name_len(leaf, extref);

			nlink++;

			cur_offset += name_len + sizeof(*extref);
		}

		offset++;
		btrfs_release_path(path);
	}
	btrfs_release_path(path);

	if (ret < 0)
		return ret;
	return nlink;
}

static int count_inode_refs(struct btrfs_root *root,
			       struct inode *inode, struct btrfs_path *path)
1295 1296 1297
{
	int ret;
	struct btrfs_key key;
M
Mark Fasheh 已提交
1298
	unsigned int nlink = 0;
1299 1300 1301
	unsigned long ptr;
	unsigned long ptr_end;
	int name_len;
L
Li Zefan 已提交
1302
	u64 ino = btrfs_ino(inode);
1303

L
Li Zefan 已提交
1304
	key.objectid = ino;
1305 1306 1307
	key.type = BTRFS_INODE_REF_KEY;
	key.offset = (u64)-1;

C
Chris Mason 已提交
1308
	while (1) {
1309 1310 1311 1312 1313 1314 1315 1316
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		if (ret < 0)
			break;
		if (ret > 0) {
			if (path->slots[0] == 0)
				break;
			path->slots[0]--;
		}
1317
process_slot:
1318 1319
		btrfs_item_key_to_cpu(path->nodes[0], &key,
				      path->slots[0]);
L
Li Zefan 已提交
1320
		if (key.objectid != ino ||
1321 1322 1323 1324 1325
		    key.type != BTRFS_INODE_REF_KEY)
			break;
		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
		ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
						   path->slots[0]);
C
Chris Mason 已提交
1326
		while (ptr < ptr_end) {
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
			struct btrfs_inode_ref *ref;

			ref = (struct btrfs_inode_ref *)ptr;
			name_len = btrfs_inode_ref_name_len(path->nodes[0],
							    ref);
			ptr = (unsigned long)(ref + 1) + name_len;
			nlink++;
		}

		if (key.offset == 0)
			break;
1338 1339 1340 1341
		if (path->slots[0] > 0) {
			path->slots[0]--;
			goto process_slot;
		}
1342
		key.offset--;
1343
		btrfs_release_path(path);
1344
	}
1345
	btrfs_release_path(path);
M
Mark Fasheh 已提交
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389

	return nlink;
}

/*
 * There are a few corners where the link count of the file can't
 * be properly maintained during replay.  So, instead of adding
 * lots of complexity to the log code, we just scan the backrefs
 * for any file that has been through replay.
 *
 * The scan will update the link count on the inode to reflect the
 * number of back refs found.  If it goes down to zero, the iput
 * will free the inode.
 */
static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
					   struct btrfs_root *root,
					   struct inode *inode)
{
	struct btrfs_path *path;
	int ret;
	u64 nlink = 0;
	u64 ino = btrfs_ino(inode);

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	ret = count_inode_refs(root, inode, path);
	if (ret < 0)
		goto out;

	nlink = ret;

	ret = count_inode_extrefs(root, inode, path);
	if (ret == -ENOENT)
		ret = 0;

	if (ret < 0)
		goto out;

	nlink += ret;

	ret = 0;

1390
	if (nlink != inode->i_nlink) {
M
Miklos Szeredi 已提交
1391
		set_nlink(inode, nlink);
1392 1393
		btrfs_update_inode(trans, root, inode);
	}
1394
	BTRFS_I(inode)->index_cnt = (u64)-1;
1395

1396 1397 1398
	if (inode->i_nlink == 0) {
		if (S_ISDIR(inode->i_mode)) {
			ret = replay_dir_deletes(trans, root, NULL, path,
L
Li Zefan 已提交
1399
						 ino, 1);
1400 1401
			if (ret)
				goto out;
1402
		}
L
Li Zefan 已提交
1403
		ret = insert_orphan_item(trans, root, ino);
1404 1405
	}

M
Mark Fasheh 已提交
1406 1407 1408
out:
	btrfs_free_path(path);
	return ret;
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
}

static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
					    struct btrfs_root *root,
					    struct btrfs_path *path)
{
	int ret;
	struct btrfs_key key;
	struct inode *inode;

	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
	key.type = BTRFS_ORPHAN_ITEM_KEY;
	key.offset = (u64)-1;
C
Chris Mason 已提交
1422
	while (1) {
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
		if (ret < 0)
			break;

		if (ret == 1) {
			if (path->slots[0] == 0)
				break;
			path->slots[0]--;
		}

		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
		if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
		    key.type != BTRFS_ORPHAN_ITEM_KEY)
			break;

		ret = btrfs_del_item(trans, root, path);
1439 1440
		if (ret)
			goto out;
1441

1442
		btrfs_release_path(path);
1443
		inode = read_one_inode(root, key.offset);
1444 1445
		if (!inode)
			return -EIO;
1446 1447 1448

		ret = fixup_inode_link_count(trans, root, inode);
		iput(inode);
1449 1450
		if (ret)
			goto out;
1451

1452 1453 1454 1455 1456 1457
		/*
		 * fixup on a directory may create new entries,
		 * make sure we always look for the highset possible
		 * offset
		 */
		key.offset = (u64)-1;
1458
	}
1459 1460
	ret = 0;
out:
1461
	btrfs_release_path(path);
1462
	return ret;
1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
}


/*
 * record a given inode in the fixup dir so we can check its link
 * count when replay is done.  The link count is incremented here
 * so the inode won't go away until we check it
 */
static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_path *path,
				      u64 objectid)
{
	struct btrfs_key key;
	int ret = 0;
	struct inode *inode;

	inode = read_one_inode(root, objectid);
1481 1482
	if (!inode)
		return -EIO;
1483 1484 1485 1486 1487 1488 1489

	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
	btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
	key.offset = objectid;

	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);

1490
	btrfs_release_path(path);
1491
	if (ret == 0) {
1492 1493 1494
		if (!inode->i_nlink)
			set_nlink(inode, 1);
		else
Z
Zach Brown 已提交
1495
			inc_nlink(inode);
1496
		ret = btrfs_update_inode(trans, root, inode);
1497 1498 1499
	} else if (ret == -EEXIST) {
		ret = 0;
	} else {
1500
		BUG(); /* Logic Error */
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531
	}
	iput(inode);

	return ret;
}

/*
 * when replaying the log for a directory, we only insert names
 * for inodes that actually exist.  This means an fsync on a directory
 * does not implicitly fsync all the new files in it
 */
static noinline int insert_one_name(struct btrfs_trans_handle *trans,
				    struct btrfs_root *root,
				    struct btrfs_path *path,
				    u64 dirid, u64 index,
				    char *name, int name_len, u8 type,
				    struct btrfs_key *location)
{
	struct inode *inode;
	struct inode *dir;
	int ret;

	inode = read_one_inode(root, location->objectid);
	if (!inode)
		return -ENOENT;

	dir = read_one_inode(root, dirid);
	if (!dir) {
		iput(inode);
		return -EIO;
	}
1532

1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
	ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);

	/* FIXME, put inode into FIXUP list */

	iput(inode);
	iput(dir);
	return ret;
}

/*
 * take a single entry in a log directory item and replay it into
 * the subvolume.
 *
 * if a conflicting item exists in the subdirectory already,
 * the inode it points to is unlinked and put into the link count
 * fix up tree.
 *
 * If a name from the log points to a file or directory that does
 * not exist in the FS, it is skipped.  fsyncs on directories
 * do not force down inodes inside that directory, just changes to the
 * names or unlinks in a directory.
 */
static noinline int replay_one_name(struct btrfs_trans_handle *trans,
				    struct btrfs_root *root,
				    struct btrfs_path *path,
				    struct extent_buffer *eb,
				    struct btrfs_dir_item *di,
				    struct btrfs_key *key)
{
	char *name;
	int name_len;
	struct btrfs_dir_item *dst_di;
	struct btrfs_key found_key;
	struct btrfs_key log_key;
	struct inode *dir;
	u8 log_type;
C
Chris Mason 已提交
1569
	int exists;
1570
	int ret = 0;
1571
	bool update_size = (key->type == BTRFS_DIR_INDEX_KEY);
1572 1573

	dir = read_one_inode(root, key->objectid);
1574 1575
	if (!dir)
		return -EIO;
1576 1577 1578

	name_len = btrfs_dir_name_len(eb, di);
	name = kmalloc(name_len, GFP_NOFS);
1579 1580 1581 1582
	if (!name) {
		ret = -ENOMEM;
		goto out;
	}
1583

1584 1585 1586 1587 1588
	log_type = btrfs_dir_type(eb, di);
	read_extent_buffer(eb, name, (unsigned long)(di + 1),
		   name_len);

	btrfs_dir_item_key_to_cpu(eb, di, &log_key);
C
Chris Mason 已提交
1589 1590 1591 1592 1593
	exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
	if (exists == 0)
		exists = 1;
	else
		exists = 0;
1594
	btrfs_release_path(path);
C
Chris Mason 已提交
1595

1596 1597 1598
	if (key->type == BTRFS_DIR_ITEM_KEY) {
		dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
				       name, name_len, 1);
C
Chris Mason 已提交
1599
	} else if (key->type == BTRFS_DIR_INDEX_KEY) {
1600 1601 1602 1603 1604
		dst_di = btrfs_lookup_dir_index_item(trans, root, path,
						     key->objectid,
						     key->offset, name,
						     name_len, 1);
	} else {
1605 1606 1607
		/* Corruption */
		ret = -EINVAL;
		goto out;
1608
	}
1609
	if (IS_ERR_OR_NULL(dst_di)) {
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
		/* we need a sequence number to insert, so we only
		 * do inserts for the BTRFS_DIR_INDEX_KEY types
		 */
		if (key->type != BTRFS_DIR_INDEX_KEY)
			goto out;
		goto insert;
	}

	btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
	/* the existing item matches the logged item */
	if (found_key.objectid == log_key.objectid &&
	    found_key.type == log_key.type &&
	    found_key.offset == log_key.offset &&
	    btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
		goto out;
	}

	/*
	 * don't drop the conflicting directory entry if the inode
	 * for the new entry doesn't exist
	 */
C
Chris Mason 已提交
1631
	if (!exists)
1632 1633 1634
		goto out;

	ret = drop_one_dir_item(trans, root, path, dir, dst_di);
1635 1636
	if (ret)
		goto out;
1637 1638 1639 1640

	if (key->type == BTRFS_DIR_INDEX_KEY)
		goto insert;
out:
1641
	btrfs_release_path(path);
1642 1643 1644 1645
	if (!ret && update_size) {
		btrfs_i_size_write(dir, dir->i_size + name_len * 2);
		ret = btrfs_update_inode(trans, root, dir);
	}
1646 1647
	kfree(name);
	iput(dir);
1648
	return ret;
1649 1650

insert:
1651
	btrfs_release_path(path);
1652 1653
	ret = insert_one_name(trans, root, path, key->objectid, key->offset,
			      name, name_len, log_type, &log_key);
1654 1655
	if (ret && ret != -ENOENT)
		goto out;
1656
	update_size = false;
1657
	ret = 0;
1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
	goto out;
}

/*
 * find all the names in a directory item and reconcile them into
 * the subvolume.  Only BTRFS_DIR_ITEM_KEY types will have more than
 * one name in a directory item, but the same code gets used for
 * both directory index types
 */
static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
					struct btrfs_root *root,
					struct btrfs_path *path,
					struct extent_buffer *eb, int slot,
					struct btrfs_key *key)
{
	int ret;
	u32 item_size = btrfs_item_size_nr(eb, slot);
	struct btrfs_dir_item *di;
	int name_len;
	unsigned long ptr;
	unsigned long ptr_end;

	ptr = btrfs_item_ptr_offset(eb, slot);
	ptr_end = ptr + item_size;
C
Chris Mason 已提交
1682
	while (ptr < ptr_end) {
1683
		di = (struct btrfs_dir_item *)ptr;
1684 1685
		if (verify_dir_item(root, eb, di))
			return -EIO;
1686 1687
		name_len = btrfs_dir_name_len(eb, di);
		ret = replay_one_name(trans, root, path, eb, di, key);
1688 1689
		if (ret)
			return ret;
1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774
		ptr = (unsigned long)(di + 1);
		ptr += name_len;
	}
	return 0;
}

/*
 * directory replay has two parts.  There are the standard directory
 * items in the log copied from the subvolume, and range items
 * created in the log while the subvolume was logged.
 *
 * The range items tell us which parts of the key space the log
 * is authoritative for.  During replay, if a key in the subvolume
 * directory is in a logged range item, but not actually in the log
 * that means it was deleted from the directory before the fsync
 * and should be removed.
 */
static noinline int find_dir_range(struct btrfs_root *root,
				   struct btrfs_path *path,
				   u64 dirid, int key_type,
				   u64 *start_ret, u64 *end_ret)
{
	struct btrfs_key key;
	u64 found_end;
	struct btrfs_dir_log_item *item;
	int ret;
	int nritems;

	if (*start_ret == (u64)-1)
		return 1;

	key.objectid = dirid;
	key.type = key_type;
	key.offset = *start_ret;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto out;
	if (ret > 0) {
		if (path->slots[0] == 0)
			goto out;
		path->slots[0]--;
	}
	if (ret != 0)
		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);

	if (key.type != key_type || key.objectid != dirid) {
		ret = 1;
		goto next;
	}
	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
			      struct btrfs_dir_log_item);
	found_end = btrfs_dir_log_end(path->nodes[0], item);

	if (*start_ret >= key.offset && *start_ret <= found_end) {
		ret = 0;
		*start_ret = key.offset;
		*end_ret = found_end;
		goto out;
	}
	ret = 1;
next:
	/* check the next slot in the tree to see if it is a valid item */
	nritems = btrfs_header_nritems(path->nodes[0]);
	if (path->slots[0] >= nritems) {
		ret = btrfs_next_leaf(root, path);
		if (ret)
			goto out;
	} else {
		path->slots[0]++;
	}

	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);

	if (key.type != key_type || key.objectid != dirid) {
		ret = 1;
		goto out;
	}
	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
			      struct btrfs_dir_log_item);
	found_end = btrfs_dir_log_end(path->nodes[0], item);
	*start_ret = key.offset;
	*end_ret = found_end;
	ret = 0;
out:
1775
	btrfs_release_path(path);
1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
	return ret;
}

/*
 * this looks for a given directory item in the log.  If the directory
 * item is not in the log, the item is removed and the inode it points
 * to is unlinked
 */
static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
				      struct btrfs_root *root,
				      struct btrfs_root *log,
				      struct btrfs_path *path,
				      struct btrfs_path *log_path,
				      struct inode *dir,
				      struct btrfs_key *dir_key)
{
	int ret;
	struct extent_buffer *eb;
	int slot;
	u32 item_size;
	struct btrfs_dir_item *di;
	struct btrfs_dir_item *log_di;
	int name_len;
	unsigned long ptr;
	unsigned long ptr_end;
	char *name;
	struct inode *inode;
	struct btrfs_key location;

again:
	eb = path->nodes[0];
	slot = path->slots[0];
	item_size = btrfs_item_size_nr(eb, slot);
	ptr = btrfs_item_ptr_offset(eb, slot);
	ptr_end = ptr + item_size;
C
Chris Mason 已提交
1811
	while (ptr < ptr_end) {
1812
		di = (struct btrfs_dir_item *)ptr;
1813 1814 1815 1816 1817
		if (verify_dir_item(root, eb, di)) {
			ret = -EIO;
			goto out;
		}

1818 1819 1820 1821 1822 1823 1824 1825 1826
		name_len = btrfs_dir_name_len(eb, di);
		name = kmalloc(name_len, GFP_NOFS);
		if (!name) {
			ret = -ENOMEM;
			goto out;
		}
		read_extent_buffer(eb, name, (unsigned long)(di + 1),
				  name_len);
		log_di = NULL;
1827
		if (log && dir_key->type == BTRFS_DIR_ITEM_KEY) {
1828 1829 1830
			log_di = btrfs_lookup_dir_item(trans, log, log_path,
						       dir_key->objectid,
						       name, name_len, 0);
1831
		} else if (log && dir_key->type == BTRFS_DIR_INDEX_KEY) {
1832 1833 1834 1835 1836 1837
			log_di = btrfs_lookup_dir_index_item(trans, log,
						     log_path,
						     dir_key->objectid,
						     dir_key->offset,
						     name, name_len, 0);
		}
1838
		if (!log_di || (IS_ERR(log_di) && PTR_ERR(log_di) == -ENOENT)) {
1839
			btrfs_dir_item_key_to_cpu(eb, di, &location);
1840 1841
			btrfs_release_path(path);
			btrfs_release_path(log_path);
1842
			inode = read_one_inode(root, location.objectid);
1843 1844 1845 1846
			if (!inode) {
				kfree(name);
				return -EIO;
			}
1847 1848 1849

			ret = link_to_fixup_dir(trans, root,
						path, location.objectid);
1850 1851 1852 1853 1854 1855
			if (ret) {
				kfree(name);
				iput(inode);
				goto out;
			}

Z
Zach Brown 已提交
1856
			inc_nlink(inode);
1857 1858
			ret = btrfs_unlink_inode(trans, root, dir, inode,
						 name, name_len);
1859
			if (!ret)
1860
				ret = btrfs_run_delayed_items(trans, root);
1861 1862
			kfree(name);
			iput(inode);
1863 1864
			if (ret)
				goto out;
1865 1866 1867 1868 1869 1870 1871 1872 1873 1874

			/* there might still be more names under this key
			 * check and repeat if required
			 */
			ret = btrfs_search_slot(NULL, root, dir_key, path,
						0, 0);
			if (ret == 0)
				goto again;
			ret = 0;
			goto out;
1875 1876 1877
		} else if (IS_ERR(log_di)) {
			kfree(name);
			return PTR_ERR(log_di);
1878
		}
1879
		btrfs_release_path(log_path);
1880 1881 1882 1883 1884 1885 1886
		kfree(name);

		ptr = (unsigned long)(di + 1);
		ptr += name_len;
	}
	ret = 0;
out:
1887 1888
	btrfs_release_path(path);
	btrfs_release_path(log_path);
1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905
	return ret;
}

/*
 * deletion replay happens before we copy any new directory items
 * out of the log or out of backreferences from inodes.  It
 * scans the log to find ranges of keys that log is authoritative for,
 * and then scans the directory to find items in those ranges that are
 * not present in the log.
 *
 * Anything we don't find in the log is unlinked and removed from the
 * directory.
 */
static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
				       struct btrfs_root *root,
				       struct btrfs_root *log,
				       struct btrfs_path *path,
1906
				       u64 dirid, int del_all)
1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934
{
	u64 range_start;
	u64 range_end;
	int key_type = BTRFS_DIR_LOG_ITEM_KEY;
	int ret = 0;
	struct btrfs_key dir_key;
	struct btrfs_key found_key;
	struct btrfs_path *log_path;
	struct inode *dir;

	dir_key.objectid = dirid;
	dir_key.type = BTRFS_DIR_ITEM_KEY;
	log_path = btrfs_alloc_path();
	if (!log_path)
		return -ENOMEM;

	dir = read_one_inode(root, dirid);
	/* it isn't an error if the inode isn't there, that can happen
	 * because we replay the deletes before we copy in the inode item
	 * from the log
	 */
	if (!dir) {
		btrfs_free_path(log_path);
		return 0;
	}
again:
	range_start = 0;
	range_end = 0;
C
Chris Mason 已提交
1935
	while (1) {
1936 1937 1938 1939 1940 1941 1942 1943
		if (del_all)
			range_end = (u64)-1;
		else {
			ret = find_dir_range(log, path, dirid, key_type,
					     &range_start, &range_end);
			if (ret != 0)
				break;
		}
1944 1945

		dir_key.offset = range_start;
C
Chris Mason 已提交
1946
		while (1) {
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968
			int nritems;
			ret = btrfs_search_slot(NULL, root, &dir_key, path,
						0, 0);
			if (ret < 0)
				goto out;

			nritems = btrfs_header_nritems(path->nodes[0]);
			if (path->slots[0] >= nritems) {
				ret = btrfs_next_leaf(root, path);
				if (ret)
					break;
			}
			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
					      path->slots[0]);
			if (found_key.objectid != dirid ||
			    found_key.type != dir_key.type)
				goto next_type;

			if (found_key.offset > range_end)
				break;

			ret = check_item_in_log(trans, root, log, path,
1969 1970
						log_path, dir,
						&found_key);
1971 1972
			if (ret)
				goto out;
1973 1974 1975 1976
			if (found_key.offset == (u64)-1)
				break;
			dir_key.offset = found_key.offset + 1;
		}
1977
		btrfs_release_path(path);
1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
		if (range_end == (u64)-1)
			break;
		range_start = range_end + 1;
	}

next_type:
	ret = 0;
	if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
		key_type = BTRFS_DIR_LOG_INDEX_KEY;
		dir_key.type = BTRFS_DIR_INDEX_KEY;
1988
		btrfs_release_path(path);
1989 1990 1991
		goto again;
	}
out:
1992
	btrfs_release_path(path);
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
	btrfs_free_path(log_path);
	iput(dir);
	return ret;
}

/*
 * the process_func used to replay items from the log tree.  This
 * gets called in two different stages.  The first stage just looks
 * for inodes and makes sure they are all copied into the subvolume.
 *
 * The second stage copies all the other item types from the log into
 * the subvolume.  The two stage approach is slower, but gets rid of
 * lots of complexity around inodes referencing other inodes that exist
 * only in the log (references come from either directory items or inode
 * back refs).
 */
static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
			     struct walk_control *wc, u64 gen)
{
	int nritems;
	struct btrfs_path *path;
	struct btrfs_root *root = wc->replay_dest;
	struct btrfs_key key;
	int level;
	int i;
	int ret;

2020 2021 2022
	ret = btrfs_read_buffer(eb, gen);
	if (ret)
		return ret;
2023 2024 2025 2026 2027 2028 2029

	level = btrfs_header_level(eb);

	if (level != 0)
		return 0;

	path = btrfs_alloc_path();
2030 2031
	if (!path)
		return -ENOMEM;
2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047

	nritems = btrfs_header_nritems(eb);
	for (i = 0; i < nritems; i++) {
		btrfs_item_key_to_cpu(eb, &key, i);

		/* inode keys are done during the first stage */
		if (key.type == BTRFS_INODE_ITEM_KEY &&
		    wc->stage == LOG_WALK_REPLAY_INODES) {
			struct btrfs_inode_item *inode_item;
			u32 mode;

			inode_item = btrfs_item_ptr(eb, i,
					    struct btrfs_inode_item);
			mode = btrfs_inode_mode(eb, inode_item);
			if (S_ISDIR(mode)) {
				ret = replay_dir_deletes(wc->trans,
2048
					 root, log, path, key.objectid, 0);
2049 2050
				if (ret)
					break;
2051 2052 2053
			}
			ret = overwrite_item(wc->trans, root, path,
					     eb, i, &key);
2054 2055
			if (ret)
				break;
2056

2057 2058 2059
			/* for regular files, make sure corresponding
			 * orhpan item exist. extents past the new EOF
			 * will be truncated later by orphan cleanup.
2060 2061
			 */
			if (S_ISREG(mode)) {
2062 2063
				ret = insert_orphan_item(wc->trans, root,
							 key.objectid);
2064 2065
				if (ret)
					break;
2066
			}
2067

2068 2069
			ret = link_to_fixup_dir(wc->trans, root,
						path, key.objectid);
2070 2071
			if (ret)
				break;
2072
		}
2073 2074 2075 2076 2077 2078 2079 2080 2081

		if (key.type == BTRFS_DIR_INDEX_KEY &&
		    wc->stage == LOG_WALK_REPLAY_DIR_INDEX) {
			ret = replay_one_dir_item(wc->trans, root, path,
						  eb, i, &key);
			if (ret)
				break;
		}

2082 2083 2084 2085 2086 2087 2088
		if (wc->stage < LOG_WALK_REPLAY_ALL)
			continue;

		/* these keys are simply copied */
		if (key.type == BTRFS_XATTR_ITEM_KEY) {
			ret = overwrite_item(wc->trans, root, path,
					     eb, i, &key);
2089 2090
			if (ret)
				break;
2091 2092
		} else if (key.type == BTRFS_INODE_REF_KEY ||
			   key.type == BTRFS_INODE_EXTREF_KEY) {
M
Mark Fasheh 已提交
2093 2094
			ret = add_inode_ref(wc->trans, root, log, path,
					    eb, i, &key);
2095 2096 2097
			if (ret && ret != -ENOENT)
				break;
			ret = 0;
2098 2099 2100
		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
			ret = replay_one_extent(wc->trans, root, path,
						eb, i, &key);
2101 2102
			if (ret)
				break;
2103
		} else if (key.type == BTRFS_DIR_ITEM_KEY) {
2104 2105
			ret = replay_one_dir_item(wc->trans, root, path,
						  eb, i, &key);
2106 2107
			if (ret)
				break;
2108 2109 2110
		}
	}
	btrfs_free_path(path);
2111
	return ret;
2112 2113
}

C
Chris Mason 已提交
2114
static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
				   struct btrfs_root *root,
				   struct btrfs_path *path, int *level,
				   struct walk_control *wc)
{
	u64 root_owner;
	u64 bytenr;
	u64 ptr_gen;
	struct extent_buffer *next;
	struct extent_buffer *cur;
	struct extent_buffer *parent;
	u32 blocksize;
	int ret = 0;

	WARN_ON(*level < 0);
	WARN_ON(*level >= BTRFS_MAX_LEVEL);

C
Chris Mason 已提交
2131
	while (*level > 0) {
2132 2133 2134 2135
		WARN_ON(*level < 0);
		WARN_ON(*level >= BTRFS_MAX_LEVEL);
		cur = path->nodes[*level];

2136
		WARN_ON(btrfs_header_level(cur) != *level);
2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149

		if (path->slots[*level] >=
		    btrfs_header_nritems(cur))
			break;

		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
		blocksize = btrfs_level_size(root, *level - 1);

		parent = path->nodes[*level];
		root_owner = btrfs_header_owner(parent);

		next = btrfs_find_create_tree_block(root, bytenr, blocksize);
2150 2151
		if (!next)
			return -ENOMEM;
2152 2153

		if (*level == 1) {
2154
			ret = wc->process_func(root, next, wc, ptr_gen);
2155 2156
			if (ret) {
				free_extent_buffer(next);
2157
				return ret;
2158
			}
2159

2160 2161
			path->slots[*level]++;
			if (wc->free) {
2162 2163 2164 2165 2166
				ret = btrfs_read_buffer(next, ptr_gen);
				if (ret) {
					free_extent_buffer(next);
					return ret;
				}
2167

2168 2169 2170 2171 2172 2173 2174
				if (trans) {
					btrfs_tree_lock(next);
					btrfs_set_lock_blocking(next);
					clean_tree_block(trans, root, next);
					btrfs_wait_tree_block_writeback(next);
					btrfs_tree_unlock(next);
				}
2175 2176 2177

				WARN_ON(root_owner !=
					BTRFS_TREE_LOG_OBJECTID);
2178
				ret = btrfs_free_and_pin_reserved_extent(root,
2179
							 bytenr, blocksize);
2180 2181 2182 2183
				if (ret) {
					free_extent_buffer(next);
					return ret;
				}
2184 2185 2186 2187
			}
			free_extent_buffer(next);
			continue;
		}
2188 2189 2190 2191 2192
		ret = btrfs_read_buffer(next, ptr_gen);
		if (ret) {
			free_extent_buffer(next);
			return ret;
		}
2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204

		WARN_ON(*level <= 0);
		if (path->nodes[*level-1])
			free_extent_buffer(path->nodes[*level-1]);
		path->nodes[*level-1] = next;
		*level = btrfs_header_level(next);
		path->slots[*level] = 0;
		cond_resched();
	}
	WARN_ON(*level < 0);
	WARN_ON(*level >= BTRFS_MAX_LEVEL);

2205
	path->slots[*level] = btrfs_header_nritems(path->nodes[*level]);
2206 2207 2208 2209 2210

	cond_resched();
	return 0;
}

C
Chris Mason 已提交
2211
static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
2212 2213 2214 2215 2216 2217 2218 2219 2220
				 struct btrfs_root *root,
				 struct btrfs_path *path, int *level,
				 struct walk_control *wc)
{
	u64 root_owner;
	int i;
	int slot;
	int ret;

C
Chris Mason 已提交
2221
	for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
2222
		slot = path->slots[i];
2223
		if (slot + 1 < btrfs_header_nritems(path->nodes[i])) {
2224 2225 2226 2227 2228
			path->slots[i]++;
			*level = i;
			WARN_ON(*level == 0);
			return 0;
		} else {
Z
Zheng Yan 已提交
2229 2230 2231 2232 2233 2234 2235
			struct extent_buffer *parent;
			if (path->nodes[*level] == root->node)
				parent = path->nodes[*level];
			else
				parent = path->nodes[*level + 1];

			root_owner = btrfs_header_owner(parent);
2236
			ret = wc->process_func(root, path->nodes[*level], wc,
2237
				 btrfs_header_generation(path->nodes[*level]));
2238 2239 2240
			if (ret)
				return ret;

2241 2242 2243 2244 2245
			if (wc->free) {
				struct extent_buffer *next;

				next = path->nodes[*level];

2246 2247 2248 2249 2250 2251 2252
				if (trans) {
					btrfs_tree_lock(next);
					btrfs_set_lock_blocking(next);
					clean_tree_block(trans, root, next);
					btrfs_wait_tree_block_writeback(next);
					btrfs_tree_unlock(next);
				}
2253 2254

				WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
2255
				ret = btrfs_free_and_pin_reserved_extent(root,
2256
						path->nodes[*level]->start,
2257
						path->nodes[*level]->len);
2258 2259
				if (ret)
					return ret;
2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
			}
			free_extent_buffer(path->nodes[*level]);
			path->nodes[*level] = NULL;
			*level = i + 1;
		}
	}
	return 1;
}

/*
 * drop the reference count on the tree rooted at 'snap'.  This traverses
 * the tree freeing any blocks that have a ref count of zero after being
 * decremented.
 */
static int walk_log_tree(struct btrfs_trans_handle *trans,
			 struct btrfs_root *log, struct walk_control *wc)
{
	int ret = 0;
	int wret;
	int level;
	struct btrfs_path *path;
	int orig_level;

	path = btrfs_alloc_path();
T
Tsutomu Itoh 已提交
2284 2285
	if (!path)
		return -ENOMEM;
2286 2287 2288 2289 2290 2291 2292

	level = btrfs_header_level(log->node);
	orig_level = level;
	path->nodes[level] = log->node;
	extent_buffer_get(log->node);
	path->slots[level] = 0;

C
Chris Mason 已提交
2293
	while (1) {
2294 2295 2296
		wret = walk_down_log_tree(trans, log, path, &level, wc);
		if (wret > 0)
			break;
2297
		if (wret < 0) {
2298
			ret = wret;
2299 2300
			goto out;
		}
2301 2302 2303 2304

		wret = walk_up_log_tree(trans, log, path, &level, wc);
		if (wret > 0)
			break;
2305
		if (wret < 0) {
2306
			ret = wret;
2307 2308
			goto out;
		}
2309 2310 2311 2312
	}

	/* was the root node processed? if not, catch it here */
	if (path->nodes[orig_level]) {
2313
		ret = wc->process_func(log, path->nodes[orig_level], wc,
2314
			 btrfs_header_generation(path->nodes[orig_level]));
2315 2316
		if (ret)
			goto out;
2317 2318 2319 2320 2321
		if (wc->free) {
			struct extent_buffer *next;

			next = path->nodes[orig_level];

2322 2323 2324 2325 2326 2327 2328
			if (trans) {
				btrfs_tree_lock(next);
				btrfs_set_lock_blocking(next);
				clean_tree_block(trans, log, next);
				btrfs_wait_tree_block_writeback(next);
				btrfs_tree_unlock(next);
			}
2329 2330 2331

			WARN_ON(log->root_key.objectid !=
				BTRFS_TREE_LOG_OBJECTID);
2332
			ret = btrfs_free_and_pin_reserved_extent(log, next->start,
2333
							 next->len);
2334 2335
			if (ret)
				goto out;
2336 2337 2338
		}
	}

2339
out:
2340 2341 2342 2343
	btrfs_free_path(path);
	return ret;
}

Y
Yan Zheng 已提交
2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363
/*
 * helper function to update the item for a given subvolumes log root
 * in the tree of log roots
 */
static int update_log_root(struct btrfs_trans_handle *trans,
			   struct btrfs_root *log)
{
	int ret;

	if (log->log_transid == 1) {
		/* insert root item on the first sync */
		ret = btrfs_insert_root(trans, log->fs_info->log_root_tree,
				&log->root_key, &log->root_item);
	} else {
		ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
				&log->root_key, &log->root_item);
	}
	return ret;
}

2364 2365
static int wait_log_commit(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root, unsigned long transid)
2366 2367
{
	DEFINE_WAIT(wait);
Y
Yan Zheng 已提交
2368
	int index = transid % 2;
2369
	int ret = 0;
2370

Y
Yan Zheng 已提交
2371 2372 2373 2374 2375
	/*
	 * we only allow two pending log transactions at a time,
	 * so we know that if ours is more than 2 older than the
	 * current transaction, we're done
	 */
2376
	do {
2377 2378 2379 2380 2381 2382
		if (ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) ==
		    trans->transid) {
			ret = -EAGAIN;
			break;
		}

Y
Yan Zheng 已提交
2383 2384 2385
		prepare_to_wait(&root->log_commit_wait[index],
				&wait, TASK_UNINTERRUPTIBLE);
		mutex_unlock(&root->log_mutex);
2386

2387
		if (root->log_transid < transid + 2 &&
Y
Yan Zheng 已提交
2388 2389
		    atomic_read(&root->log_commit[index]))
			schedule();
2390

Y
Yan Zheng 已提交
2391 2392
		finish_wait(&root->log_commit_wait[index], &wait);
		mutex_lock(&root->log_mutex);
2393
	} while (root->log_transid < transid + 2 &&
Y
Yan Zheng 已提交
2394
		 atomic_read(&root->log_commit[index]));
2395 2396

	return ret;
Y
Yan Zheng 已提交
2397 2398
}

2399 2400
static void wait_for_writer(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root)
Y
Yan Zheng 已提交
2401 2402
{
	DEFINE_WAIT(wait);
2403
	while (ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) !=
2404
	       trans->transid && atomic_read(&root->log_writers)) {
Y
Yan Zheng 已提交
2405 2406 2407
		prepare_to_wait(&root->log_writer_wait,
				&wait, TASK_UNINTERRUPTIBLE);
		mutex_unlock(&root->log_mutex);
2408
		if (ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) !=
2409
		    trans->transid && atomic_read(&root->log_writers))
2410
			schedule();
Y
Yan Zheng 已提交
2411 2412 2413
		mutex_lock(&root->log_mutex);
		finish_wait(&root->log_writer_wait, &wait);
	}
2414 2415 2416 2417 2418
}

/*
 * btrfs_sync_log does sends a given tree log down to the disk and
 * updates the super blocks to record it.  When this call is done,
2419 2420 2421 2422 2423 2424 2425 2426
 * you know that any inodes previously logged are safely on disk only
 * if it returns 0.
 *
 * Any other return value means you need to call btrfs_commit_transaction.
 * Some of the edge cases for fsyncing directories that have had unlinks
 * or renames done in the past mean that sometimes the only safe
 * fsync is to commit the whole FS.  When btrfs_sync_log returns -EAGAIN,
 * that has happened.
2427 2428 2429 2430
 */
int btrfs_sync_log(struct btrfs_trans_handle *trans,
		   struct btrfs_root *root)
{
Y
Yan Zheng 已提交
2431 2432
	int index1;
	int index2;
2433
	int mark;
2434 2435
	int ret;
	struct btrfs_root *log = root->log_root;
Y
Yan Zheng 已提交
2436
	struct btrfs_root *log_root_tree = root->fs_info->log_root_tree;
2437
	unsigned long log_transid = 0;
2438
	struct blk_plug plug;
2439

Y
Yan Zheng 已提交
2440
	mutex_lock(&root->log_mutex);
2441
	log_transid = root->log_transid;
Y
Yan Zheng 已提交
2442 2443
	index1 = root->log_transid % 2;
	if (atomic_read(&root->log_commit[index1])) {
2444
		ret = wait_log_commit(trans, root, root->log_transid);
Y
Yan Zheng 已提交
2445
		mutex_unlock(&root->log_mutex);
2446
		return ret;
2447
	}
Y
Yan Zheng 已提交
2448 2449 2450 2451
	atomic_set(&root->log_commit[index1], 1);

	/* wait for previous tree log sync to complete */
	if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
2452
		wait_log_commit(trans, root, root->log_transid - 1);
2453

2454
	while (1) {
M
Miao Xie 已提交
2455
		int batch = atomic_read(&root->log_batch);
2456 2457
		/* when we're on an ssd, just kick the log commit out */
		if (!btrfs_test_opt(root, SSD) && root->log_multiple_pids) {
2458 2459 2460 2461
			mutex_unlock(&root->log_mutex);
			schedule_timeout_uninterruptible(1);
			mutex_lock(&root->log_mutex);
		}
2462
		wait_for_writer(trans, root);
M
Miao Xie 已提交
2463
		if (batch == atomic_read(&root->log_batch))
2464 2465 2466
			break;
	}

2467
	/* bail out if we need to do a full commit */
2468 2469
	if (ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) ==
	    trans->transid) {
2470
		ret = -EAGAIN;
2471
		btrfs_free_logged_extents(log, log_transid);
2472 2473 2474 2475
		mutex_unlock(&root->log_mutex);
		goto out;
	}

2476 2477 2478 2479 2480
	if (log_transid % 2 == 0)
		mark = EXTENT_DIRTY;
	else
		mark = EXTENT_NEW;

2481 2482 2483
	/* we start IO on  all the marked extents here, but we don't actually
	 * wait for them until later.
	 */
2484
	blk_start_plug(&plug);
2485
	ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark);
2486
	if (ret) {
2487
		blk_finish_plug(&plug);
2488
		btrfs_abort_transaction(trans, root, ret);
2489
		btrfs_free_logged_extents(log, log_transid);
2490 2491 2492
		mutex_unlock(&root->log_mutex);
		goto out;
	}
Y
Yan Zheng 已提交
2493

2494
	btrfs_set_root_node(&log->root_item, log->node);
Y
Yan Zheng 已提交
2495 2496 2497

	root->log_transid++;
	log->log_transid = root->log_transid;
2498
	root->log_start_pid = 0;
Y
Yan Zheng 已提交
2499
	/*
2500 2501 2502
	 * IO has been started, blocks of the log tree have WRITTEN flag set
	 * in their headers. new modifications of the log will be written to
	 * new positions. so it's safe to allow log writers to go in.
Y
Yan Zheng 已提交
2503 2504 2505 2506
	 */
	mutex_unlock(&root->log_mutex);

	mutex_lock(&log_root_tree->log_mutex);
M
Miao Xie 已提交
2507
	atomic_inc(&log_root_tree->log_batch);
Y
Yan Zheng 已提交
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519
	atomic_inc(&log_root_tree->log_writers);
	mutex_unlock(&log_root_tree->log_mutex);

	ret = update_log_root(trans, log);

	mutex_lock(&log_root_tree->log_mutex);
	if (atomic_dec_and_test(&log_root_tree->log_writers)) {
		smp_mb();
		if (waitqueue_active(&log_root_tree->log_writer_wait))
			wake_up(&log_root_tree->log_writer_wait);
	}

2520
	if (ret) {
2521
		blk_finish_plug(&plug);
2522 2523 2524 2525 2526
		if (ret != -ENOSPC) {
			btrfs_abort_transaction(trans, root, ret);
			mutex_unlock(&log_root_tree->log_mutex);
			goto out;
		}
2527 2528
		ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) =
								trans->transid;
2529
		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2530
		btrfs_free_logged_extents(log, log_transid);
2531 2532 2533 2534 2535
		mutex_unlock(&log_root_tree->log_mutex);
		ret = -EAGAIN;
		goto out;
	}

Y
Yan Zheng 已提交
2536 2537
	index2 = log_root_tree->log_transid % 2;
	if (atomic_read(&log_root_tree->log_commit[index2])) {
2538
		blk_finish_plug(&plug);
2539
		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2540 2541
		ret = wait_log_commit(trans, log_root_tree,
				      log_root_tree->log_transid);
2542
		btrfs_free_logged_extents(log, log_transid);
Y
Yan Zheng 已提交
2543 2544 2545 2546 2547
		mutex_unlock(&log_root_tree->log_mutex);
		goto out;
	}
	atomic_set(&log_root_tree->log_commit[index2], 1);

2548 2549 2550 2551 2552 2553
	if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2])) {
		wait_log_commit(trans, log_root_tree,
				log_root_tree->log_transid - 1);
	}

	wait_for_writer(trans, log_root_tree);
Y
Yan Zheng 已提交
2554

2555 2556 2557 2558
	/*
	 * now that we've moved on to the tree of log tree roots,
	 * check the full commit flag again
	 */
2559 2560
	if (ACCESS_ONCE(root->fs_info->last_trans_log_full_commit) ==
	    trans->transid) {
2561
		blk_finish_plug(&plug);
2562
		btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2563
		btrfs_free_logged_extents(log, log_transid);
2564 2565 2566 2567
		mutex_unlock(&log_root_tree->log_mutex);
		ret = -EAGAIN;
		goto out_wake_log_root;
	}
Y
Yan Zheng 已提交
2568

2569 2570 2571 2572
	ret = btrfs_write_marked_extents(log_root_tree,
					 &log_root_tree->dirty_log_pages,
					 EXTENT_DIRTY | EXTENT_NEW);
	blk_finish_plug(&plug);
2573 2574
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
2575
		btrfs_free_logged_extents(log, log_transid);
2576 2577 2578
		mutex_unlock(&log_root_tree->log_mutex);
		goto out_wake_log_root;
	}
2579
	btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2580 2581 2582
	btrfs_wait_marked_extents(log_root_tree,
				  &log_root_tree->dirty_log_pages,
				  EXTENT_NEW | EXTENT_DIRTY);
2583
	btrfs_wait_logged_extents(log, log_transid);
2584

2585
	btrfs_set_super_log_root(root->fs_info->super_for_commit,
Y
Yan Zheng 已提交
2586
				log_root_tree->node->start);
2587
	btrfs_set_super_log_root_level(root->fs_info->super_for_commit,
Y
Yan Zheng 已提交
2588
				btrfs_header_level(log_root_tree->node));
2589

Y
Yan Zheng 已提交
2590 2591 2592 2593 2594 2595 2596 2597 2598 2599
	log_root_tree->log_transid++;
	mutex_unlock(&log_root_tree->log_mutex);

	/*
	 * nobody else is going to jump in and write the the ctree
	 * super here because the log_commit atomic below is protecting
	 * us.  We must be called with a transaction handle pinning
	 * the running transaction open, so a full commit can't hop
	 * in and cause problems either.
	 */
2600 2601 2602 2603 2604
	ret = write_ctree_super(trans, root->fs_info->tree_root, 1);
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto out_wake_log_root;
	}
Y
Yan Zheng 已提交
2605

2606 2607 2608 2609 2610
	mutex_lock(&root->log_mutex);
	if (root->last_log_commit < log_transid)
		root->last_log_commit = log_transid;
	mutex_unlock(&root->log_mutex);

2611
out_wake_log_root:
Y
Yan Zheng 已提交
2612 2613 2614 2615
	atomic_set(&log_root_tree->log_commit[index2], 0);
	smp_mb();
	if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
		wake_up(&log_root_tree->log_commit_wait[index2]);
2616
out:
Y
Yan Zheng 已提交
2617 2618 2619 2620
	atomic_set(&root->log_commit[index1], 0);
	smp_mb();
	if (waitqueue_active(&root->log_commit_wait[index1]))
		wake_up(&root->log_commit_wait[index1]);
2621
	return ret;
2622 2623
}

2624 2625
static void free_log_tree(struct btrfs_trans_handle *trans,
			  struct btrfs_root *log)
2626 2627
{
	int ret;
2628 2629
	u64 start;
	u64 end;
2630 2631 2632 2633 2634
	struct walk_control wc = {
		.free = 1,
		.process_func = process_one_buffer
	};

2635 2636 2637 2638
	ret = walk_log_tree(trans, log, &wc);
	/* I don't think this can happen but just in case */
	if (ret)
		btrfs_abort_transaction(trans, log, ret);
2639

C
Chris Mason 已提交
2640
	while (1) {
2641
		ret = find_first_extent_bit(&log->dirty_log_pages,
2642 2643
				0, &start, &end, EXTENT_DIRTY | EXTENT_NEW,
				NULL);
2644 2645 2646
		if (ret)
			break;

2647 2648
		clear_extent_bits(&log->dirty_log_pages, start, end,
				  EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS);
2649 2650
	}

2651 2652 2653 2654 2655 2656 2657 2658
	/*
	 * We may have short-circuited the log tree with the full commit logic
	 * and left ordered extents on our list, so clear these out to keep us
	 * from leaking inodes and memory.
	 */
	btrfs_free_logged_extents(log, 0);
	btrfs_free_logged_extents(log, 1);

Y
Yan Zheng 已提交
2659 2660
	free_extent_buffer(log->node);
	kfree(log);
2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682
}

/*
 * free all the extents used by the tree log.  This should be called
 * at commit time of the full transaction
 */
int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
{
	if (root->log_root) {
		free_log_tree(trans, root->log_root);
		root->log_root = NULL;
	}
	return 0;
}

int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
			     struct btrfs_fs_info *fs_info)
{
	if (fs_info->log_root_tree) {
		free_log_tree(trans, fs_info->log_root_tree);
		fs_info->log_root_tree = NULL;
	}
2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715
	return 0;
}

/*
 * If both a file and directory are logged, and unlinks or renames are
 * mixed in, we have a few interesting corners:
 *
 * create file X in dir Y
 * link file X to X.link in dir Y
 * fsync file X
 * unlink file X but leave X.link
 * fsync dir Y
 *
 * After a crash we would expect only X.link to exist.  But file X
 * didn't get fsync'd again so the log has back refs for X and X.link.
 *
 * We solve this by removing directory entries and inode backrefs from the
 * log when a file that was logged in the current transaction is
 * unlinked.  Any later fsync will include the updated log entries, and
 * we'll be able to reconstruct the proper directory items from backrefs.
 *
 * This optimizations allows us to avoid relogging the entire inode
 * or the entire directory.
 */
int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root,
				 const char *name, int name_len,
				 struct inode *dir, u64 index)
{
	struct btrfs_root *log;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
	int ret;
2716
	int err = 0;
2717
	int bytes_del = 0;
L
Li Zefan 已提交
2718
	u64 dir_ino = btrfs_ino(dir);
2719

2720 2721 2722
	if (BTRFS_I(dir)->logged_trans < trans->transid)
		return 0;

2723 2724 2725 2726 2727 2728 2729 2730
	ret = join_running_log_trans(root);
	if (ret)
		return 0;

	mutex_lock(&BTRFS_I(dir)->log_mutex);

	log = root->log_root;
	path = btrfs_alloc_path();
2731 2732 2733 2734
	if (!path) {
		err = -ENOMEM;
		goto out_unlock;
	}
2735

L
Li Zefan 已提交
2736
	di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
2737
				   name, name_len, -1);
2738 2739 2740 2741 2742
	if (IS_ERR(di)) {
		err = PTR_ERR(di);
		goto fail;
	}
	if (di) {
2743 2744
		ret = btrfs_delete_one_dir_name(trans, log, path, di);
		bytes_del += name_len;
2745 2746 2747 2748
		if (ret) {
			err = ret;
			goto fail;
		}
2749
	}
2750
	btrfs_release_path(path);
L
Li Zefan 已提交
2751
	di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
2752
					 index, name, name_len, -1);
2753 2754 2755 2756 2757
	if (IS_ERR(di)) {
		err = PTR_ERR(di);
		goto fail;
	}
	if (di) {
2758 2759
		ret = btrfs_delete_one_dir_name(trans, log, path, di);
		bytes_del += name_len;
2760 2761 2762 2763
		if (ret) {
			err = ret;
			goto fail;
		}
2764 2765 2766 2767 2768 2769 2770 2771
	}

	/* update the directory size in the log to reflect the names
	 * we have removed
	 */
	if (bytes_del) {
		struct btrfs_key key;

L
Li Zefan 已提交
2772
		key.objectid = dir_ino;
2773 2774
		key.offset = 0;
		key.type = BTRFS_INODE_ITEM_KEY;
2775
		btrfs_release_path(path);
2776 2777

		ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
2778 2779 2780 2781
		if (ret < 0) {
			err = ret;
			goto fail;
		}
2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796
		if (ret == 0) {
			struct btrfs_inode_item *item;
			u64 i_size;

			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
					      struct btrfs_inode_item);
			i_size = btrfs_inode_size(path->nodes[0], item);
			if (i_size > bytes_del)
				i_size -= bytes_del;
			else
				i_size = 0;
			btrfs_set_inode_size(path->nodes[0], item, i_size);
			btrfs_mark_buffer_dirty(path->nodes[0]);
		} else
			ret = 0;
2797
		btrfs_release_path(path);
2798
	}
2799
fail:
2800
	btrfs_free_path(path);
2801
out_unlock:
2802
	mutex_unlock(&BTRFS_I(dir)->log_mutex);
2803 2804 2805
	if (ret == -ENOSPC) {
		root->fs_info->last_trans_log_full_commit = trans->transid;
		ret = 0;
2806 2807 2808
	} else if (ret < 0)
		btrfs_abort_transaction(trans, root, ret);

2809
	btrfs_end_log_trans(root);
2810

2811
	return err;
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823
}

/* see comments for btrfs_del_dir_entries_in_log */
int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root,
			       const char *name, int name_len,
			       struct inode *inode, u64 dirid)
{
	struct btrfs_root *log;
	u64 index;
	int ret;

2824 2825 2826
	if (BTRFS_I(inode)->logged_trans < trans->transid)
		return 0;

2827 2828 2829 2830 2831 2832
	ret = join_running_log_trans(root);
	if (ret)
		return 0;
	log = root->log_root;
	mutex_lock(&BTRFS_I(inode)->log_mutex);

L
Li Zefan 已提交
2833
	ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
2834 2835
				  dirid, &index);
	mutex_unlock(&BTRFS_I(inode)->log_mutex);
2836 2837 2838
	if (ret == -ENOSPC) {
		root->fs_info->last_trans_log_full_commit = trans->transid;
		ret = 0;
2839 2840
	} else if (ret < 0 && ret != -ENOENT)
		btrfs_abort_transaction(trans, root, ret);
2841
	btrfs_end_log_trans(root);
2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867

	return ret;
}

/*
 * creates a range item in the log for 'dirid'.  first_offset and
 * last_offset tell us which parts of the key space the log should
 * be considered authoritative for.
 */
static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
				       struct btrfs_root *log,
				       struct btrfs_path *path,
				       int key_type, u64 dirid,
				       u64 first_offset, u64 last_offset)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_dir_log_item *item;

	key.objectid = dirid;
	key.offset = first_offset;
	if (key_type == BTRFS_DIR_ITEM_KEY)
		key.type = BTRFS_DIR_LOG_ITEM_KEY;
	else
		key.type = BTRFS_DIR_LOG_INDEX_KEY;
	ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
2868 2869
	if (ret)
		return ret;
2870 2871 2872 2873 2874

	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
			      struct btrfs_dir_log_item);
	btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
	btrfs_mark_buffer_dirty(path->nodes[0]);
2875
	btrfs_release_path(path);
2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892
	return 0;
}

/*
 * log all the items included in the current transaction for a given
 * directory.  This also creates the range items in the log tree required
 * to replay anything deleted before the fsync
 */
static noinline int log_dir_items(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root, struct inode *inode,
			  struct btrfs_path *path,
			  struct btrfs_path *dst_path, int key_type,
			  u64 min_offset, u64 *last_offset_ret)
{
	struct btrfs_key min_key;
	struct btrfs_root *log = root->log_root;
	struct extent_buffer *src;
2893
	int err = 0;
2894 2895 2896 2897 2898
	int ret;
	int i;
	int nritems;
	u64 first_offset = min_offset;
	u64 last_offset = (u64)-1;
L
Li Zefan 已提交
2899
	u64 ino = btrfs_ino(inode);
2900 2901 2902

	log = root->log_root;

L
Li Zefan 已提交
2903
	min_key.objectid = ino;
2904 2905 2906 2907 2908
	min_key.type = key_type;
	min_key.offset = min_offset;

	path->keep_locks = 1;

2909
	ret = btrfs_search_forward(root, &min_key, path, trans->transid);
2910 2911 2912 2913 2914

	/*
	 * we didn't find anything from this transaction, see if there
	 * is anything at all
	 */
L
Li Zefan 已提交
2915 2916
	if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
		min_key.objectid = ino;
2917 2918
		min_key.type = key_type;
		min_key.offset = (u64)-1;
2919
		btrfs_release_path(path);
2920 2921
		ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
		if (ret < 0) {
2922
			btrfs_release_path(path);
2923 2924
			return ret;
		}
L
Li Zefan 已提交
2925
		ret = btrfs_previous_item(root, path, ino, key_type);
2926 2927 2928 2929 2930 2931 2932 2933 2934 2935

		/* if ret == 0 there are items for this type,
		 * create a range to tell us the last key of this type.
		 * otherwise, there are no items in this directory after
		 * *min_offset, and we create a range to indicate that.
		 */
		if (ret == 0) {
			struct btrfs_key tmp;
			btrfs_item_key_to_cpu(path->nodes[0], &tmp,
					      path->slots[0]);
C
Chris Mason 已提交
2936
			if (key_type == tmp.type)
2937 2938 2939 2940 2941 2942
				first_offset = max(min_offset, tmp.offset) + 1;
		}
		goto done;
	}

	/* go backward to find any previous key */
L
Li Zefan 已提交
2943
	ret = btrfs_previous_item(root, path, ino, key_type);
2944 2945 2946 2947 2948 2949 2950 2951
	if (ret == 0) {
		struct btrfs_key tmp;
		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
		if (key_type == tmp.type) {
			first_offset = tmp.offset;
			ret = overwrite_item(trans, log, dst_path,
					     path->nodes[0], path->slots[0],
					     &tmp);
2952 2953 2954 2955
			if (ret) {
				err = ret;
				goto done;
			}
2956 2957
		}
	}
2958
	btrfs_release_path(path);
2959 2960 2961

	/* find the first key from this transaction again */
	ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
2962
	if (WARN_ON(ret != 0))
2963 2964 2965 2966 2967 2968
		goto done;

	/*
	 * we have a block from this transaction, log every item in it
	 * from our directory
	 */
C
Chris Mason 已提交
2969
	while (1) {
2970 2971 2972 2973 2974 2975
		struct btrfs_key tmp;
		src = path->nodes[0];
		nritems = btrfs_header_nritems(src);
		for (i = path->slots[0]; i < nritems; i++) {
			btrfs_item_key_to_cpu(src, &min_key, i);

L
Li Zefan 已提交
2976
			if (min_key.objectid != ino || min_key.type != key_type)
2977 2978 2979
				goto done;
			ret = overwrite_item(trans, log, dst_path, src, i,
					     &min_key);
2980 2981 2982 2983
			if (ret) {
				err = ret;
				goto done;
			}
2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996
		}
		path->slots[0] = nritems;

		/*
		 * look ahead to the next item and see if it is also
		 * from this directory and from this transaction
		 */
		ret = btrfs_next_leaf(root, path);
		if (ret == 1) {
			last_offset = (u64)-1;
			goto done;
		}
		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
L
Li Zefan 已提交
2997
		if (tmp.objectid != ino || tmp.type != key_type) {
2998 2999 3000 3001 3002 3003 3004
			last_offset = (u64)-1;
			goto done;
		}
		if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
			ret = overwrite_item(trans, log, dst_path,
					     path->nodes[0], path->slots[0],
					     &tmp);
3005 3006 3007 3008
			if (ret)
				err = ret;
			else
				last_offset = tmp.offset;
3009 3010 3011 3012
			goto done;
		}
	}
done:
3013 3014
	btrfs_release_path(path);
	btrfs_release_path(dst_path);
3015

3016 3017 3018 3019 3020 3021 3022
	if (err == 0) {
		*last_offset_ret = last_offset;
		/*
		 * insert the log range keys to indicate where the log
		 * is valid
		 */
		ret = insert_dir_log_key(trans, log, path, key_type,
L
Li Zefan 已提交
3023
					 ino, first_offset, last_offset);
3024 3025 3026 3027
		if (ret)
			err = ret;
	}
	return err;
3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054
}

/*
 * logging directories is very similar to logging inodes, We find all the items
 * from the current transaction and write them to the log.
 *
 * The recovery code scans the directory in the subvolume, and if it finds a
 * key in the range logged that is not present in the log tree, then it means
 * that dir entry was unlinked during the transaction.
 *
 * In order for that scan to work, we must include one key smaller than
 * the smallest logged by this transaction and one key larger than the largest
 * key logged by this transaction.
 */
static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root, struct inode *inode,
			  struct btrfs_path *path,
			  struct btrfs_path *dst_path)
{
	u64 min_key;
	u64 max_key;
	int ret;
	int key_type = BTRFS_DIR_ITEM_KEY;

again:
	min_key = 0;
	max_key = 0;
C
Chris Mason 已提交
3055
	while (1) {
3056 3057 3058
		ret = log_dir_items(trans, root, inode, path,
				    dst_path, key_type, min_key,
				    &max_key);
3059 3060
		if (ret)
			return ret;
3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086
		if (max_key == (u64)-1)
			break;
		min_key = max_key + 1;
	}

	if (key_type == BTRFS_DIR_ITEM_KEY) {
		key_type = BTRFS_DIR_INDEX_KEY;
		goto again;
	}
	return 0;
}

/*
 * a helper function to drop items from the log before we relog an
 * inode.  max_key_type indicates the highest item type to remove.
 * This cannot be run for file data extents because it does not
 * free the extents they point to.
 */
static int drop_objectid_items(struct btrfs_trans_handle *trans,
				  struct btrfs_root *log,
				  struct btrfs_path *path,
				  u64 objectid, int max_key_type)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_key found_key;
3087
	int start_slot;
3088 3089 3090 3091 3092

	key.objectid = objectid;
	key.type = max_key_type;
	key.offset = (u64)-1;

C
Chris Mason 已提交
3093
	while (1) {
3094
		ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
3095
		BUG_ON(ret == 0); /* Logic error */
3096
		if (ret < 0)
3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108
			break;

		if (path->slots[0] == 0)
			break;

		path->slots[0]--;
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);

		if (found_key.objectid != objectid)
			break;

3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120
		found_key.offset = 0;
		found_key.type = 0;
		ret = btrfs_bin_search(path->nodes[0], &found_key, 0,
				       &start_slot);

		ret = btrfs_del_items(trans, log, path, start_slot,
				      path->slots[0] - start_slot + 1);
		/*
		 * If start slot isn't 0 then we don't need to re-search, we've
		 * found the last guy with the objectid in this tree.
		 */
		if (ret || start_slot != 0)
3121
			break;
3122
		btrfs_release_path(path);
3123
	}
3124
	btrfs_release_path(path);
3125 3126
	if (ret > 0)
		ret = 0;
3127
	return ret;
3128 3129
}

3130 3131 3132 3133 3134
static void fill_inode_item(struct btrfs_trans_handle *trans,
			    struct extent_buffer *leaf,
			    struct btrfs_inode_item *item,
			    struct inode *inode, int log_inode_only)
{
3135 3136 3137
	struct btrfs_map_token token;

	btrfs_init_map_token(&token);
3138 3139 3140 3141 3142 3143 3144

	if (log_inode_only) {
		/* set the generation to zero so the recover code
		 * can tell the difference between an logging
		 * just to say 'this inode exists' and a logging
		 * to say 'update this inode with these values'
		 */
3145 3146
		btrfs_set_token_inode_generation(leaf, item, 0, &token);
		btrfs_set_token_inode_size(leaf, item, 0, &token);
3147
	} else {
3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181
		btrfs_set_token_inode_generation(leaf, item,
						 BTRFS_I(inode)->generation,
						 &token);
		btrfs_set_token_inode_size(leaf, item, inode->i_size, &token);
	}

	btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token);
	btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token);
	btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token);
	btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token);

	btrfs_set_token_timespec_sec(leaf, btrfs_inode_atime(item),
				     inode->i_atime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, btrfs_inode_atime(item),
				      inode->i_atime.tv_nsec, &token);

	btrfs_set_token_timespec_sec(leaf, btrfs_inode_mtime(item),
				     inode->i_mtime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, btrfs_inode_mtime(item),
				      inode->i_mtime.tv_nsec, &token);

	btrfs_set_token_timespec_sec(leaf, btrfs_inode_ctime(item),
				     inode->i_ctime.tv_sec, &token);
	btrfs_set_token_timespec_nsec(leaf, btrfs_inode_ctime(item),
				      inode->i_ctime.tv_nsec, &token);

	btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode),
				     &token);

	btrfs_set_token_inode_sequence(leaf, item, inode->i_version, &token);
	btrfs_set_token_inode_transid(leaf, item, trans->transid, &token);
	btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token);
	btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token);
	btrfs_set_token_inode_block_group(leaf, item, 0, &token);
3182 3183
}

3184 3185 3186 3187 3188 3189 3190
static int log_inode_item(struct btrfs_trans_handle *trans,
			  struct btrfs_root *log, struct btrfs_path *path,
			  struct inode *inode)
{
	struct btrfs_inode_item *inode_item;
	int ret;

3191 3192
	ret = btrfs_insert_empty_item(trans, log, path,
				      &BTRFS_I(inode)->location,
3193 3194 3195 3196 3197 3198 3199 3200 3201 3202
				      sizeof(*inode_item));
	if (ret && ret != -EEXIST)
		return ret;
	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
				    struct btrfs_inode_item);
	fill_inode_item(trans, path->nodes[0], inode_item, inode, 0);
	btrfs_release_path(path);
	return 0;
}

3203
static noinline int copy_items(struct btrfs_trans_handle *trans,
3204
			       struct inode *inode,
3205
			       struct btrfs_path *dst_path,
3206
			       struct btrfs_path *src_path, u64 *last_extent,
3207 3208 3209 3210
			       int start_slot, int nr, int inode_only)
{
	unsigned long src_offset;
	unsigned long dst_offset;
3211
	struct btrfs_root *log = BTRFS_I(inode)->root->log_root;
3212 3213
	struct btrfs_file_extent_item *extent;
	struct btrfs_inode_item *inode_item;
3214 3215
	struct extent_buffer *src = src_path->nodes[0];
	struct btrfs_key first_key, last_key, key;
3216 3217 3218 3219 3220
	int ret;
	struct btrfs_key *ins_keys;
	u32 *ins_sizes;
	char *ins_data;
	int i;
3221
	struct list_head ordered_sums;
3222
	int skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
3223 3224 3225
	bool has_extents = false;
	bool need_find_last_extent = (*last_extent == 0);
	bool done = false;
3226 3227

	INIT_LIST_HEAD(&ordered_sums);
3228 3229 3230

	ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
			   nr * sizeof(u32), GFP_NOFS);
3231 3232 3233
	if (!ins_data)
		return -ENOMEM;

3234 3235
	first_key.objectid = (u64)-1;

3236 3237 3238 3239 3240 3241 3242 3243 3244
	ins_sizes = (u32 *)ins_data;
	ins_keys = (struct btrfs_key *)(ins_data + nr * sizeof(u32));

	for (i = 0; i < nr; i++) {
		ins_sizes[i] = btrfs_item_size_nr(src, i + start_slot);
		btrfs_item_key_to_cpu(src, ins_keys + i, i + start_slot);
	}
	ret = btrfs_insert_empty_items(trans, log, dst_path,
				       ins_keys, ins_sizes, nr);
3245 3246 3247 3248
	if (ret) {
		kfree(ins_data);
		return ret;
	}
3249

3250
	for (i = 0; i < nr; i++, dst_path->slots[0]++) {
3251 3252 3253 3254 3255
		dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
						   dst_path->slots[0]);

		src_offset = btrfs_item_ptr_offset(src, start_slot + i);

3256 3257 3258
		if ((i == (nr - 1)))
			last_key = ins_keys[i];

3259
		if (ins_keys[i].type == BTRFS_INODE_ITEM_KEY) {
3260 3261 3262
			inode_item = btrfs_item_ptr(dst_path->nodes[0],
						    dst_path->slots[0],
						    struct btrfs_inode_item);
3263 3264 3265 3266 3267
			fill_inode_item(trans, dst_path->nodes[0], inode_item,
					inode, inode_only == LOG_INODE_EXISTS);
		} else {
			copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
					   src_offset, ins_sizes[i]);
3268
		}
3269

3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284
		/*
		 * We set need_find_last_extent here in case we know we were
		 * processing other items and then walk into the first extent in
		 * the inode.  If we don't hit an extent then nothing changes,
		 * we'll do the last search the next time around.
		 */
		if (ins_keys[i].type == BTRFS_EXTENT_DATA_KEY) {
			has_extents = true;
			if (need_find_last_extent &&
			    first_key.objectid == (u64)-1)
				first_key = ins_keys[i];
		} else {
			need_find_last_extent = false;
		}

3285 3286 3287 3288
		/* take a reference on file data extents so that truncates
		 * or deletes of this inode don't have to relog the inode
		 * again
		 */
3289 3290
		if (btrfs_key_type(ins_keys + i) == BTRFS_EXTENT_DATA_KEY &&
		    !skip_csum) {
3291 3292 3293 3294
			int found_type;
			extent = btrfs_item_ptr(src, start_slot + i,
						struct btrfs_file_extent_item);

3295 3296 3297
			if (btrfs_file_extent_generation(src, extent) < trans->transid)
				continue;

3298
			found_type = btrfs_file_extent_type(src, extent);
3299
			if (found_type == BTRFS_FILE_EXTENT_REG) {
3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310
				u64 ds, dl, cs, cl;
				ds = btrfs_file_extent_disk_bytenr(src,
								extent);
				/* ds == 0 is a hole */
				if (ds == 0)
					continue;

				dl = btrfs_file_extent_disk_num_bytes(src,
								extent);
				cs = btrfs_file_extent_offset(src, extent);
				cl = btrfs_file_extent_num_bytes(src,
3311
								extent);
3312 3313 3314 3315 3316
				if (btrfs_file_extent_compression(src,
								  extent)) {
					cs = 0;
					cl = dl;
				}
3317 3318 3319 3320

				ret = btrfs_lookup_csums_range(
						log->fs_info->csum_root,
						ds + cs, ds + cs + cl - 1,
A
Arne Jansen 已提交
3321
						&ordered_sums, 0);
3322 3323 3324 3325 3326
				if (ret) {
					btrfs_release_path(dst_path);
					kfree(ins_data);
					return ret;
				}
3327 3328 3329 3330 3331
			}
		}
	}

	btrfs_mark_buffer_dirty(dst_path->nodes[0]);
3332
	btrfs_release_path(dst_path);
3333
	kfree(ins_data);
3334 3335 3336 3337 3338

	/*
	 * we have to do this after the loop above to avoid changing the
	 * log tree while trying to change the log tree.
	 */
3339
	ret = 0;
C
Chris Mason 已提交
3340
	while (!list_empty(&ordered_sums)) {
3341 3342 3343
		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
						   struct btrfs_ordered_sum,
						   list);
3344 3345
		if (!ret)
			ret = btrfs_csum_file_blocks(trans, log, sums);
3346 3347 3348
		list_del(&sums->list);
		kfree(sums);
	}
3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377

	if (!has_extents)
		return ret;

	/*
	 * Because we use btrfs_search_forward we could skip leaves that were
	 * not modified and then assume *last_extent is valid when it really
	 * isn't.  So back up to the previous leaf and read the end of the last
	 * extent before we go and fill in holes.
	 */
	if (need_find_last_extent) {
		u64 len;

		ret = btrfs_prev_leaf(BTRFS_I(inode)->root, src_path);
		if (ret < 0)
			return ret;
		if (ret)
			goto fill_holes;
		if (src_path->slots[0])
			src_path->slots[0]--;
		src = src_path->nodes[0];
		btrfs_item_key_to_cpu(src, &key, src_path->slots[0]);
		if (key.objectid != btrfs_ino(inode) ||
		    key.type != BTRFS_EXTENT_DATA_KEY)
			goto fill_holes;
		extent = btrfs_item_ptr(src, src_path->slots[0],
					struct btrfs_file_extent_item);
		if (btrfs_file_extent_type(src, extent) ==
		    BTRFS_FILE_EXTENT_INLINE) {
3378 3379 3380
			len = btrfs_file_extent_inline_len(src,
							   src_path->slots[0],
							   extent);
3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443
			*last_extent = ALIGN(key.offset + len,
					     log->sectorsize);
		} else {
			len = btrfs_file_extent_num_bytes(src, extent);
			*last_extent = key.offset + len;
		}
	}
fill_holes:
	/* So we did prev_leaf, now we need to move to the next leaf, but a few
	 * things could have happened
	 *
	 * 1) A merge could have happened, so we could currently be on a leaf
	 * that holds what we were copying in the first place.
	 * 2) A split could have happened, and now not all of the items we want
	 * are on the same leaf.
	 *
	 * So we need to adjust how we search for holes, we need to drop the
	 * path and re-search for the first extent key we found, and then walk
	 * forward until we hit the last one we copied.
	 */
	if (need_find_last_extent) {
		/* btrfs_prev_leaf could return 1 without releasing the path */
		btrfs_release_path(src_path);
		ret = btrfs_search_slot(NULL, BTRFS_I(inode)->root, &first_key,
					src_path, 0, 0);
		if (ret < 0)
			return ret;
		ASSERT(ret == 0);
		src = src_path->nodes[0];
		i = src_path->slots[0];
	} else {
		i = start_slot;
	}

	/*
	 * Ok so here we need to go through and fill in any holes we may have
	 * to make sure that holes are punched for those areas in case they had
	 * extents previously.
	 */
	while (!done) {
		u64 offset, len;
		u64 extent_end;

		if (i >= btrfs_header_nritems(src_path->nodes[0])) {
			ret = btrfs_next_leaf(BTRFS_I(inode)->root, src_path);
			if (ret < 0)
				return ret;
			ASSERT(ret == 0);
			src = src_path->nodes[0];
			i = 0;
		}

		btrfs_item_key_to_cpu(src, &key, i);
		if (!btrfs_comp_cpu_keys(&key, &last_key))
			done = true;
		if (key.objectid != btrfs_ino(inode) ||
		    key.type != BTRFS_EXTENT_DATA_KEY) {
			i++;
			continue;
		}
		extent = btrfs_item_ptr(src, i, struct btrfs_file_extent_item);
		if (btrfs_file_extent_type(src, extent) ==
		    BTRFS_FILE_EXTENT_INLINE) {
3444
			len = btrfs_file_extent_inline_len(src, i, extent);
3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470
			extent_end = ALIGN(key.offset + len, log->sectorsize);
		} else {
			len = btrfs_file_extent_num_bytes(src, extent);
			extent_end = key.offset + len;
		}
		i++;

		if (*last_extent == key.offset) {
			*last_extent = extent_end;
			continue;
		}
		offset = *last_extent;
		len = key.offset - *last_extent;
		ret = btrfs_insert_file_extent(trans, log, btrfs_ino(inode),
					       offset, 0, 0, len, 0, len, 0,
					       0, 0);
		if (ret)
			break;
		*last_extent = offset + len;
	}
	/*
	 * Need to let the callers know we dropped the path so they should
	 * re-search.
	 */
	if (!ret && need_find_last_extent)
		ret = 1;
3471
	return ret;
3472 3473
}

J
Josef Bacik 已提交
3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489
static int extent_cmp(void *priv, struct list_head *a, struct list_head *b)
{
	struct extent_map *em1, *em2;

	em1 = list_entry(a, struct extent_map, list);
	em2 = list_entry(b, struct extent_map, list);

	if (em1->start < em2->start)
		return -1;
	else if (em1->start > em2->start)
		return 1;
	return 0;
}

static int log_one_extent(struct btrfs_trans_handle *trans,
			  struct inode *inode, struct btrfs_root *root,
3490 3491
			  struct extent_map *em, struct btrfs_path *path,
			  struct list_head *logged_list)
J
Josef Bacik 已提交
3492 3493
{
	struct btrfs_root *log = root->log_root;
3494 3495
	struct btrfs_file_extent_item *fi;
	struct extent_buffer *leaf;
3496
	struct btrfs_ordered_extent *ordered;
3497
	struct list_head ordered_sums;
3498
	struct btrfs_map_token token;
J
Josef Bacik 已提交
3499
	struct btrfs_key key;
3500 3501 3502 3503
	u64 mod_start = em->mod_start;
	u64 mod_len = em->mod_len;
	u64 csum_offset;
	u64 csum_len;
3504 3505
	u64 extent_offset = em->start - em->orig_start;
	u64 block_len;
J
Josef Bacik 已提交
3506
	int ret;
3507
	bool skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
3508
	int extent_inserted = 0;
J
Josef Bacik 已提交
3509

3510
	INIT_LIST_HEAD(&ordered_sums);
3511
	btrfs_init_map_token(&token);
3512

3513 3514 3515
	ret = __btrfs_drop_extents(trans, log, inode, path, em->start,
				   em->start + em->len, NULL, 0, 1,
				   sizeof(*fi), &extent_inserted);
J
Josef Bacik 已提交
3516
	if (ret)
3517
		return ret;
3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528

	if (!extent_inserted) {
		key.objectid = btrfs_ino(inode);
		key.type = BTRFS_EXTENT_DATA_KEY;
		key.offset = em->start;

		ret = btrfs_insert_empty_item(trans, log, path, &key,
					      sizeof(*fi));
		if (ret)
			return ret;
	}
3529 3530 3531
	leaf = path->nodes[0];
	fi = btrfs_item_ptr(leaf, path->slots[0],
			    struct btrfs_file_extent_item);
3532

3533 3534
	btrfs_set_token_file_extent_generation(leaf, fi, em->generation,
					       &token);
3535 3536
	if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) {
		skip_csum = true;
3537 3538 3539
		btrfs_set_token_file_extent_type(leaf, fi,
						 BTRFS_FILE_EXTENT_PREALLOC,
						 &token);
3540
	} else {
3541 3542 3543
		btrfs_set_token_file_extent_type(leaf, fi,
						 BTRFS_FILE_EXTENT_REG,
						 &token);
3544
		if (em->block_start == EXTENT_MAP_HOLE)
3545 3546 3547 3548 3549
			skip_csum = true;
	}

	block_len = max(em->block_len, em->orig_block_len);
	if (em->compress_type != BTRFS_COMPRESS_NONE) {
3550 3551 3552 3553 3554
		btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
							em->block_start,
							&token);
		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
							   &token);
3555
	} else if (em->block_start < EXTENT_MAP_LAST_BYTE) {
3556 3557 3558 3559 3560
		btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
							em->block_start -
							extent_offset, &token);
		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
							   &token);
3561
	} else {
3562 3563 3564 3565 3566 3567 3568 3569 3570
		btrfs_set_token_file_extent_disk_bytenr(leaf, fi, 0, &token);
		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, 0,
							   &token);
	}

	btrfs_set_token_file_extent_offset(leaf, fi,
					   em->start - em->orig_start,
					   &token);
	btrfs_set_token_file_extent_num_bytes(leaf, fi, em->len, &token);
J
Josef Bacik 已提交
3571
	btrfs_set_token_file_extent_ram_bytes(leaf, fi, em->ram_bytes, &token);
3572 3573 3574 3575
	btrfs_set_token_file_extent_compression(leaf, fi, em->compress_type,
						&token);
	btrfs_set_token_file_extent_encryption(leaf, fi, 0, &token);
	btrfs_set_token_file_extent_other_encoding(leaf, fi, 0, &token);
3576
	btrfs_mark_buffer_dirty(leaf);
3577

3578 3579 3580 3581
	btrfs_release_path(path);
	if (ret) {
		return ret;
	}
3582

3583 3584
	if (skip_csum)
		return 0;
J
Josef Bacik 已提交
3585

3586 3587 3588 3589
	/*
	 * First check and see if our csums are on our outstanding ordered
	 * extents.
	 */
3590
	list_for_each_entry(ordered, logged_list, log_list) {
3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638
		struct btrfs_ordered_sum *sum;

		if (!mod_len)
			break;

		if (ordered->file_offset + ordered->len <= mod_start ||
		    mod_start + mod_len <= ordered->file_offset)
			continue;

		/*
		 * We are going to copy all the csums on this ordered extent, so
		 * go ahead and adjust mod_start and mod_len in case this
		 * ordered extent has already been logged.
		 */
		if (ordered->file_offset > mod_start) {
			if (ordered->file_offset + ordered->len >=
			    mod_start + mod_len)
				mod_len = ordered->file_offset - mod_start;
			/*
			 * If we have this case
			 *
			 * |--------- logged extent ---------|
			 *       |----- ordered extent ----|
			 *
			 * Just don't mess with mod_start and mod_len, we'll
			 * just end up logging more csums than we need and it
			 * will be ok.
			 */
		} else {
			if (ordered->file_offset + ordered->len <
			    mod_start + mod_len) {
				mod_len = (mod_start + mod_len) -
					(ordered->file_offset + ordered->len);
				mod_start = ordered->file_offset +
					ordered->len;
			} else {
				mod_len = 0;
			}
		}

		/*
		 * To keep us from looping for the above case of an ordered
		 * extent that falls inside of the logged extent.
		 */
		if (test_and_set_bit(BTRFS_ORDERED_LOGGED_CSUM,
				     &ordered->flags))
			continue;

3639 3640 3641 3642 3643
		if (ordered->csum_bytes_left) {
			btrfs_start_ordered_extent(inode, ordered, 0);
			wait_event(ordered->wait,
				   ordered->csum_bytes_left == 0);
		}
3644 3645 3646

		list_for_each_entry(sum, &ordered->list, list) {
			ret = btrfs_csum_file_blocks(trans, log, sum);
3647
			if (ret)
3648 3649 3650 3651 3652 3653 3654 3655 3656
				goto unlocked;
		}

	}
unlocked:

	if (!mod_len || ret)
		return ret;

3657 3658 3659 3660 3661 3662 3663
	if (em->compress_type) {
		csum_offset = 0;
		csum_len = block_len;
	} else {
		csum_offset = mod_start - em->start;
		csum_len = mod_len;
	}
3664

3665 3666 3667 3668 3669 3670 3671
	/* block start is already adjusted for the file extent offset. */
	ret = btrfs_lookup_csums_range(log->fs_info->csum_root,
				       em->block_start + csum_offset,
				       em->block_start + csum_offset +
				       csum_len - 1, &ordered_sums, 0);
	if (ret)
		return ret;
J
Josef Bacik 已提交
3672

3673 3674 3675 3676 3677 3678 3679 3680
	while (!list_empty(&ordered_sums)) {
		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
						   struct btrfs_ordered_sum,
						   list);
		if (!ret)
			ret = btrfs_csum_file_blocks(trans, log, sums);
		list_del(&sums->list);
		kfree(sums);
J
Josef Bacik 已提交
3681 3682
	}

3683
	return ret;
J
Josef Bacik 已提交
3684 3685 3686 3687 3688
}

static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root,
				     struct inode *inode,
3689 3690
				     struct btrfs_path *path,
				     struct list_head *logged_list)
J
Josef Bacik 已提交
3691 3692 3693 3694 3695 3696
{
	struct extent_map *em, *n;
	struct list_head extents;
	struct extent_map_tree *tree = &BTRFS_I(inode)->extent_tree;
	u64 test_gen;
	int ret = 0;
3697
	int num = 0;
J
Josef Bacik 已提交
3698 3699 3700 3701 3702 3703 3704 3705

	INIT_LIST_HEAD(&extents);

	write_lock(&tree->lock);
	test_gen = root->fs_info->last_trans_committed;

	list_for_each_entry_safe(em, n, &tree->modified_extents, list) {
		list_del_init(&em->list);
3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718

		/*
		 * Just an arbitrary number, this can be really CPU intensive
		 * once we start getting a lot of extents, and really once we
		 * have a bunch of extents we just want to commit since it will
		 * be faster.
		 */
		if (++num > 32768) {
			list_del_init(&tree->modified_extents);
			ret = -EFBIG;
			goto process;
		}

J
Josef Bacik 已提交
3719 3720
		if (em->generation <= test_gen)
			continue;
3721 3722 3723
		/* Need a ref to keep it from getting evicted from cache */
		atomic_inc(&em->refs);
		set_bit(EXTENT_FLAG_LOGGING, &em->flags);
J
Josef Bacik 已提交
3724
		list_add_tail(&em->list, &extents);
3725
		num++;
J
Josef Bacik 已提交
3726 3727 3728 3729
	}

	list_sort(NULL, &extents, extent_cmp);

3730
process:
J
Josef Bacik 已提交
3731 3732 3733 3734 3735 3736 3737 3738 3739
	while (!list_empty(&extents)) {
		em = list_entry(extents.next, struct extent_map, list);

		list_del_init(&em->list);

		/*
		 * If we had an error we just need to delete everybody from our
		 * private list.
		 */
3740
		if (ret) {
3741
			clear_em_logging(tree, em);
3742
			free_extent_map(em);
J
Josef Bacik 已提交
3743
			continue;
3744 3745 3746
		}

		write_unlock(&tree->lock);
J
Josef Bacik 已提交
3747

3748
		ret = log_one_extent(trans, inode, root, em, path, logged_list);
3749
		write_lock(&tree->lock);
3750 3751
		clear_em_logging(tree, em);
		free_extent_map(em);
J
Josef Bacik 已提交
3752
	}
3753 3754
	WARN_ON(!list_empty(&extents));
	write_unlock(&tree->lock);
J
Josef Bacik 已提交
3755 3756 3757 3758 3759

	btrfs_release_path(path);
	return ret;
}

3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773
/* log a single inode in the tree log.
 * At least one parent directory for this inode must exist in the tree
 * or be logged already.
 *
 * Any items from this inode changed by the current transaction are copied
 * to the log tree.  An extra reference is taken on any extents in this
 * file, allowing us to avoid a whole pile of corner cases around logging
 * blocks that have been removed from the tree.
 *
 * See LOG_INODE_ALL and related defines for a description of what inode_only
 * does.
 *
 * This handles both files and directories.
 */
3774
static int btrfs_log_inode(struct btrfs_trans_handle *trans,
3775 3776 3777 3778 3779 3780 3781 3782
			     struct btrfs_root *root, struct inode *inode,
			     int inode_only)
{
	struct btrfs_path *path;
	struct btrfs_path *dst_path;
	struct btrfs_key min_key;
	struct btrfs_key max_key;
	struct btrfs_root *log = root->log_root;
3783
	struct extent_buffer *src = NULL;
3784
	LIST_HEAD(logged_list);
3785
	u64 last_extent = 0;
3786
	int err = 0;
3787
	int ret;
3788
	int nritems;
3789 3790
	int ins_start_slot = 0;
	int ins_nr;
J
Josef Bacik 已提交
3791
	bool fast_search = false;
L
Li Zefan 已提交
3792
	u64 ino = btrfs_ino(inode);
3793 3794

	path = btrfs_alloc_path();
3795 3796
	if (!path)
		return -ENOMEM;
3797
	dst_path = btrfs_alloc_path();
3798 3799 3800 3801
	if (!dst_path) {
		btrfs_free_path(path);
		return -ENOMEM;
	}
3802

L
Li Zefan 已提交
3803
	min_key.objectid = ino;
3804 3805 3806
	min_key.type = BTRFS_INODE_ITEM_KEY;
	min_key.offset = 0;

L
Li Zefan 已提交
3807
	max_key.objectid = ino;
3808 3809


J
Josef Bacik 已提交
3810
	/* today the code can only do partial logging of directories */
3811 3812 3813 3814
	if (S_ISDIR(inode->i_mode) ||
	    (!test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
		       &BTRFS_I(inode)->runtime_flags) &&
	     inode_only == LOG_INODE_EXISTS))
3815 3816 3817 3818 3819
		max_key.type = BTRFS_XATTR_ITEM_KEY;
	else
		max_key.type = (u8)-1;
	max_key.offset = (u64)-1;

3820 3821 3822 3823 3824 3825 3826 3827 3828
	/* Only run delayed items if we are a dir or a new file */
	if (S_ISDIR(inode->i_mode) ||
	    BTRFS_I(inode)->generation > root->fs_info->last_trans_committed) {
		ret = btrfs_commit_inode_delayed_items(trans, inode);
		if (ret) {
			btrfs_free_path(path);
			btrfs_free_path(dst_path);
			return ret;
		}
3829 3830
	}

3831 3832
	mutex_lock(&BTRFS_I(inode)->log_mutex);

3833
	btrfs_get_logged_extents(inode, &logged_list);
3834

3835 3836 3837 3838 3839 3840 3841 3842 3843
	/*
	 * a brute force approach to making sure we get the most uptodate
	 * copies of everything.
	 */
	if (S_ISDIR(inode->i_mode)) {
		int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;

		if (inode_only == LOG_INODE_EXISTS)
			max_key_type = BTRFS_XATTR_ITEM_KEY;
L
Li Zefan 已提交
3844
		ret = drop_objectid_items(trans, log, path, ino, max_key_type);
3845
	} else {
J
Josef Bacik 已提交
3846 3847
		if (test_and_clear_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
				       &BTRFS_I(inode)->runtime_flags)) {
3848 3849
			clear_bit(BTRFS_INODE_COPY_EVERYTHING,
				  &BTRFS_I(inode)->runtime_flags);
J
Josef Bacik 已提交
3850 3851
			ret = btrfs_truncate_inode_items(trans, log,
							 inode, 0, 0);
3852
		} else if (test_and_clear_bit(BTRFS_INODE_COPY_EVERYTHING,
3853 3854
					      &BTRFS_I(inode)->runtime_flags) ||
			   inode_only == LOG_INODE_EXISTS) {
3855 3856
			if (inode_only == LOG_INODE_ALL)
				fast_search = true;
3857
			max_key.type = BTRFS_XATTR_ITEM_KEY;
J
Josef Bacik 已提交
3858
			ret = drop_objectid_items(trans, log, path, ino,
3859
						  max_key.type);
3860 3861 3862 3863 3864 3865 3866 3867 3868
		} else {
			if (inode_only == LOG_INODE_ALL)
				fast_search = true;
			ret = log_inode_item(trans, log, dst_path, inode);
			if (ret) {
				err = ret;
				goto out_unlock;
			}
			goto log_extents;
J
Josef Bacik 已提交
3869
		}
3870

3871
	}
3872 3873 3874 3875
	if (ret) {
		err = ret;
		goto out_unlock;
	}
3876 3877
	path->keep_locks = 1;

C
Chris Mason 已提交
3878
	while (1) {
3879
		ins_nr = 0;
3880
		ret = btrfs_search_forward(root, &min_key,
3881
					   path, trans->transid);
3882 3883
		if (ret != 0)
			break;
3884
again:
3885
		/* note, ins_nr might be > 0 here, cleanup outside the loop */
L
Li Zefan 已提交
3886
		if (min_key.objectid != ino)
3887 3888 3889
			break;
		if (min_key.type > max_key.type)
			break;
3890

3891
		src = path->nodes[0];
3892 3893 3894 3895 3896 3897 3898
		if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
			ins_nr++;
			goto next_slot;
		} else if (!ins_nr) {
			ins_start_slot = path->slots[0];
			ins_nr = 1;
			goto next_slot;
3899 3900
		}

3901 3902 3903
		ret = copy_items(trans, inode, dst_path, path, &last_extent,
				 ins_start_slot, ins_nr, inode_only);
		if (ret < 0) {
3904 3905
			err = ret;
			goto out_unlock;
3906 3907 3908 3909
		} if (ret) {
			ins_nr = 0;
			btrfs_release_path(path);
			continue;
3910
		}
3911 3912 3913
		ins_nr = 1;
		ins_start_slot = path->slots[0];
next_slot:
3914

3915 3916 3917 3918 3919 3920 3921
		nritems = btrfs_header_nritems(path->nodes[0]);
		path->slots[0]++;
		if (path->slots[0] < nritems) {
			btrfs_item_key_to_cpu(path->nodes[0], &min_key,
					      path->slots[0]);
			goto again;
		}
3922
		if (ins_nr) {
3923 3924
			ret = copy_items(trans, inode, dst_path, path,
					 &last_extent, ins_start_slot,
3925
					 ins_nr, inode_only);
3926
			if (ret < 0) {
3927 3928 3929
				err = ret;
				goto out_unlock;
			}
3930
			ret = 0;
3931 3932
			ins_nr = 0;
		}
3933
		btrfs_release_path(path);
3934

3935
		if (min_key.offset < (u64)-1) {
3936
			min_key.offset++;
3937
		} else if (min_key.type < max_key.type) {
3938
			min_key.type++;
3939 3940
			min_key.offset = 0;
		} else {
3941
			break;
3942
		}
3943
	}
3944
	if (ins_nr) {
3945 3946 3947
		ret = copy_items(trans, inode, dst_path, path, &last_extent,
				 ins_start_slot, ins_nr, inode_only);
		if (ret < 0) {
3948 3949 3950
			err = ret;
			goto out_unlock;
		}
3951
		ret = 0;
3952 3953
		ins_nr = 0;
	}
J
Josef Bacik 已提交
3954

3955
log_extents:
3956 3957
	btrfs_release_path(path);
	btrfs_release_path(dst_path);
J
Josef Bacik 已提交
3958
	if (fast_search) {
3959 3960
		ret = btrfs_log_changed_extents(trans, root, inode, dst_path,
						&logged_list);
J
Josef Bacik 已提交
3961 3962 3963 3964
		if (ret) {
			err = ret;
			goto out_unlock;
		}
3965
	} else if (inode_only == LOG_INODE_ALL) {
3966 3967 3968
		struct extent_map_tree *tree = &BTRFS_I(inode)->extent_tree;
		struct extent_map *em, *n;

3969
		write_lock(&tree->lock);
3970 3971
		list_for_each_entry_safe(em, n, &tree->modified_extents, list)
			list_del_init(&em->list);
3972
		write_unlock(&tree->lock);
J
Josef Bacik 已提交
3973 3974
	}

3975
	if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
3976
		ret = log_directory_changes(trans, root, inode, path, dst_path);
3977 3978 3979 3980
		if (ret) {
			err = ret;
			goto out_unlock;
		}
3981
	}
3982
	BTRFS_I(inode)->logged_trans = trans->transid;
3983
	BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->last_sub_trans;
3984
out_unlock:
3985 3986 3987 3988
	if (unlikely(err))
		btrfs_put_logged_extents(&logged_list);
	else
		btrfs_submit_logged_extents(&logged_list, log);
3989 3990 3991 3992
	mutex_unlock(&BTRFS_I(inode)->log_mutex);

	btrfs_free_path(path);
	btrfs_free_path(dst_path);
3993
	return err;
3994 3995
}

3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006
/*
 * follow the dentry parent pointers up the chain and see if any
 * of the directories in it require a full commit before they can
 * be logged.  Returns zero if nothing special needs to be done or 1 if
 * a full commit is required.
 */
static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
					       struct inode *inode,
					       struct dentry *parent,
					       struct super_block *sb,
					       u64 last_committed)
4007
{
4008 4009
	int ret = 0;
	struct btrfs_root *root;
4010
	struct dentry *old_parent = NULL;
4011
	struct inode *orig_inode = inode;
4012

4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023
	/*
	 * for regular files, if its inode is already on disk, we don't
	 * have to worry about the parents at all.  This is because
	 * we can use the last_unlink_trans field to record renames
	 * and other fun in this file.
	 */
	if (S_ISREG(inode->i_mode) &&
	    BTRFS_I(inode)->generation <= last_committed &&
	    BTRFS_I(inode)->last_unlink_trans <= last_committed)
			goto out;

4024 4025 4026 4027 4028 4029 4030
	if (!S_ISDIR(inode->i_mode)) {
		if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
			goto out;
		inode = parent->d_inode;
	}

	while (1) {
4031 4032 4033 4034 4035 4036 4037 4038
		/*
		 * If we are logging a directory then we start with our inode,
		 * not our parents inode, so we need to skipp setting the
		 * logged_trans so that further down in the log code we don't
		 * think this inode has already been logged.
		 */
		if (inode != orig_inode)
			BTRFS_I(inode)->logged_trans = trans->transid;
4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056
		smp_mb();

		if (BTRFS_I(inode)->last_unlink_trans > last_committed) {
			root = BTRFS_I(inode)->root;

			/*
			 * make sure any commits to the log are forced
			 * to be full commits
			 */
			root->fs_info->last_trans_log_full_commit =
				trans->transid;
			ret = 1;
			break;
		}

		if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
			break;

4057
		if (IS_ROOT(parent))
4058 4059
			break;

4060 4061 4062
		parent = dget_parent(parent);
		dput(old_parent);
		old_parent = parent;
4063 4064 4065
		inode = parent->d_inode;

	}
4066
	dput(old_parent);
4067
out:
4068 4069 4070 4071 4072 4073 4074 4075 4076
	return ret;
}

/*
 * helper function around btrfs_log_inode to make sure newly created
 * parent directories also end up in the log.  A minimal inode and backref
 * only logging is done of any parent directories that are older than
 * the last committed transaction
 */
4077 4078 4079
static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
			    	  struct btrfs_root *root, struct inode *inode,
			    	  struct dentry *parent, int exists_only)
4080
{
4081
	int inode_only = exists_only ? LOG_INODE_EXISTS : LOG_INODE_ALL;
4082
	struct super_block *sb;
4083
	struct dentry *old_parent = NULL;
4084 4085 4086 4087 4088
	int ret = 0;
	u64 last_committed = root->fs_info->last_trans_committed;

	sb = inode->i_sb;

S
Sage Weil 已提交
4089 4090 4091 4092 4093
	if (btrfs_test_opt(root, NOTREELOG)) {
		ret = 1;
		goto end_no_trans;
	}

4094 4095 4096 4097 4098 4099
	if (root->fs_info->last_trans_log_full_commit >
	    root->fs_info->last_trans_committed) {
		ret = 1;
		goto end_no_trans;
	}

4100 4101 4102 4103 4104 4105
	if (root != BTRFS_I(inode)->root ||
	    btrfs_root_refs(&root->root_item) == 0) {
		ret = 1;
		goto end_no_trans;
	}

4106 4107 4108 4109
	ret = check_parent_dirs_for_sync(trans, inode, parent,
					 sb, last_committed);
	if (ret)
		goto end_no_trans;
4110

4111
	if (btrfs_inode_in_log(inode, trans->transid)) {
4112 4113 4114 4115
		ret = BTRFS_NO_LOG_SYNC;
		goto end_no_trans;
	}

4116 4117
	ret = start_log_trans(trans, root);
	if (ret)
4118
		goto end_no_trans;
4119

4120
	ret = btrfs_log_inode(trans, root, inode, inode_only);
4121 4122
	if (ret)
		goto end_trans;
4123

4124 4125 4126 4127 4128 4129 4130 4131
	/*
	 * for regular files, if its inode is already on disk, we don't
	 * have to worry about the parents at all.  This is because
	 * we can use the last_unlink_trans field to record renames
	 * and other fun in this file.
	 */
	if (S_ISREG(inode->i_mode) &&
	    BTRFS_I(inode)->generation <= last_committed &&
4132 4133 4134 4135
	    BTRFS_I(inode)->last_unlink_trans <= last_committed) {
		ret = 0;
		goto end_trans;
	}
4136 4137

	inode_only = LOG_INODE_EXISTS;
4138 4139
	while (1) {
		if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
4140 4141
			break;

4142
		inode = parent->d_inode;
4143 4144 4145
		if (root != BTRFS_I(inode)->root)
			break;

4146 4147 4148
		if (BTRFS_I(inode)->generation >
		    root->fs_info->last_trans_committed) {
			ret = btrfs_log_inode(trans, root, inode, inode_only);
4149 4150
			if (ret)
				goto end_trans;
4151
		}
4152
		if (IS_ROOT(parent))
4153
			break;
4154

4155 4156 4157
		parent = dget_parent(parent);
		dput(old_parent);
		old_parent = parent;
4158
	}
4159
	ret = 0;
4160
end_trans:
4161
	dput(old_parent);
4162 4163 4164 4165
	if (ret < 0) {
		root->fs_info->last_trans_log_full_commit = trans->transid;
		ret = 1;
	}
4166 4167 4168
	btrfs_end_log_trans(root);
end_no_trans:
	return ret;
4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179
}

/*
 * it is not safe to log dentry if the chunk root has added new
 * chunks.  This returns 0 if the dentry was logged, and 1 otherwise.
 * If this returns 1, you must commit the transaction to safely get your
 * data on disk.
 */
int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root, struct dentry *dentry)
{
4180 4181 4182 4183 4184 4185 4186
	struct dentry *parent = dget_parent(dentry);
	int ret;

	ret = btrfs_log_inode_parent(trans, root, dentry->d_inode, parent, 0);
	dput(parent);

	return ret;
4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208
}

/*
 * should be called during mount to recover any replay any log trees
 * from the FS
 */
int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_trans_handle *trans;
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_key tmp_key;
	struct btrfs_root *log;
	struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
	struct walk_control wc = {
		.process_func = process_one_buffer,
		.stage = 0,
	};

	path = btrfs_alloc_path();
T
Tsutomu Itoh 已提交
4209 4210 4211 4212
	if (!path)
		return -ENOMEM;

	fs_info->log_root_recovering = 1;
4213

4214
	trans = btrfs_start_transaction(fs_info->tree_root, 0);
4215 4216 4217 4218
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto error;
	}
4219 4220 4221 4222

	wc.trans = trans;
	wc.pin = 1;

T
Tsutomu Itoh 已提交
4223
	ret = walk_log_tree(trans, log_root_tree, &wc);
4224 4225 4226 4227 4228
	if (ret) {
		btrfs_error(fs_info, ret, "Failed to pin buffers while "
			    "recovering log root tree.");
		goto error;
	}
4229 4230 4231 4232 4233 4234

again:
	key.objectid = BTRFS_TREE_LOG_OBJECTID;
	key.offset = (u64)-1;
	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);

C
Chris Mason 已提交
4235
	while (1) {
4236
		ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
4237 4238 4239 4240 4241 4242

		if (ret < 0) {
			btrfs_error(fs_info, ret,
				    "Couldn't find tree log root.");
			goto error;
		}
4243 4244 4245 4246 4247 4248 4249
		if (ret > 0) {
			if (path->slots[0] == 0)
				break;
			path->slots[0]--;
		}
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);
4250
		btrfs_release_path(path);
4251 4252 4253
		if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
			break;

4254
		log = btrfs_read_fs_root(log_root_tree, &found_key);
4255 4256 4257 4258 4259 4260
		if (IS_ERR(log)) {
			ret = PTR_ERR(log);
			btrfs_error(fs_info, ret,
				    "Couldn't read tree log root.");
			goto error;
		}
4261 4262 4263 4264 4265 4266

		tmp_key.objectid = found_key.offset;
		tmp_key.type = BTRFS_ROOT_ITEM_KEY;
		tmp_key.offset = (u64)-1;

		wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
4267 4268
		if (IS_ERR(wc.replay_dest)) {
			ret = PTR_ERR(wc.replay_dest);
4269 4270 4271
			free_extent_buffer(log->node);
			free_extent_buffer(log->commit_root);
			kfree(log);
4272 4273 4274 4275
			btrfs_error(fs_info, ret, "Couldn't read target root "
				    "for tree log recovery.");
			goto error;
		}
4276

Y
Yan Zheng 已提交
4277
		wc.replay_dest->log_root = log;
4278
		btrfs_record_root_in_trans(trans, wc.replay_dest);
4279 4280
		ret = walk_log_tree(trans, log, &wc);

4281
		if (!ret && wc.stage == LOG_WALK_REPLAY_ALL) {
4282 4283 4284 4285 4286
			ret = fixup_inode_link_counts(trans, wc.replay_dest,
						      path);
		}

		key.offset = found_key.offset - 1;
Y
Yan Zheng 已提交
4287
		wc.replay_dest->log_root = NULL;
4288
		free_extent_buffer(log->node);
4289
		free_extent_buffer(log->commit_root);
4290 4291
		kfree(log);

4292 4293 4294
		if (ret)
			goto error;

4295 4296 4297
		if (found_key.offset == 0)
			break;
	}
4298
	btrfs_release_path(path);
4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314

	/* step one is to pin it all, step two is to replay just inodes */
	if (wc.pin) {
		wc.pin = 0;
		wc.process_func = replay_one_buffer;
		wc.stage = LOG_WALK_REPLAY_INODES;
		goto again;
	}
	/* step three is to replay everything */
	if (wc.stage < LOG_WALK_REPLAY_ALL) {
		wc.stage++;
		goto again;
	}

	btrfs_free_path(path);

4315 4316 4317 4318 4319
	/* step 4: commit the transaction, which also unpins the blocks */
	ret = btrfs_commit_transaction(trans, fs_info->tree_root);
	if (ret)
		return ret;

4320 4321 4322 4323
	free_extent_buffer(log_root_tree->node);
	log_root_tree->log_root = NULL;
	fs_info->log_root_recovering = 0;
	kfree(log_root_tree);
4324

4325
	return 0;
4326
error:
4327 4328
	if (wc.trans)
		btrfs_end_transaction(wc.trans, fs_info->tree_root);
4329 4330
	btrfs_free_path(path);
	return ret;
4331
}
4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344

/*
 * there are some corner cases where we want to force a full
 * commit instead of allowing a directory to be logged.
 *
 * They revolve around files there were unlinked from the directory, and
 * this function updates the parent directory so that a full commit is
 * properly done if it is fsync'd later after the unlinks are done.
 */
void btrfs_record_unlink_dir(struct btrfs_trans_handle *trans,
			     struct inode *dir, struct inode *inode,
			     int for_rename)
{
4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357
	/*
	 * when we're logging a file, if it hasn't been renamed
	 * or unlinked, and its inode is fully committed on disk,
	 * we don't have to worry about walking up the directory chain
	 * to log its parents.
	 *
	 * So, we use the last_unlink_trans field to put this transid
	 * into the file.  When the file is logged we check it and
	 * don't log the parents if the file is fully on disk.
	 */
	if (S_ISREG(inode->i_mode))
		BTRFS_I(inode)->last_unlink_trans = trans->transid;

4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402
	/*
	 * if this directory was already logged any new
	 * names for this file/dir will get recorded
	 */
	smp_mb();
	if (BTRFS_I(dir)->logged_trans == trans->transid)
		return;

	/*
	 * if the inode we're about to unlink was logged,
	 * the log will be properly updated for any new names
	 */
	if (BTRFS_I(inode)->logged_trans == trans->transid)
		return;

	/*
	 * when renaming files across directories, if the directory
	 * there we're unlinking from gets fsync'd later on, there's
	 * no way to find the destination directory later and fsync it
	 * properly.  So, we have to be conservative and force commits
	 * so the new name gets discovered.
	 */
	if (for_rename)
		goto record;

	/* we can safely do the unlink without any special recording */
	return;

record:
	BTRFS_I(dir)->last_unlink_trans = trans->transid;
}

/*
 * Call this after adding a new name for a file and it will properly
 * update the log to reflect the new name.
 *
 * It will return zero if all goes well, and it will return 1 if a
 * full transaction commit is required.
 */
int btrfs_log_new_name(struct btrfs_trans_handle *trans,
			struct inode *inode, struct inode *old_dir,
			struct dentry *parent)
{
	struct btrfs_root * root = BTRFS_I(inode)->root;

4403 4404 4405 4406 4407 4408 4409
	/*
	 * this will force the logging code to walk the dentry chain
	 * up for the file
	 */
	if (S_ISREG(inode->i_mode))
		BTRFS_I(inode)->last_unlink_trans = trans->transid;

4410 4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422
	/*
	 * if this inode hasn't been logged and directory we're renaming it
	 * from hasn't been logged, we don't need to log it
	 */
	if (BTRFS_I(inode)->logged_trans <=
	    root->fs_info->last_trans_committed &&
	    (!old_dir || BTRFS_I(old_dir)->logged_trans <=
		    root->fs_info->last_trans_committed))
		return 0;

	return btrfs_log_inode_parent(trans, root, inode, parent, 1);
}