1// SPDX-License-Identifier: GPL-2.0
   2/*
   3 * Copyright (C) 2008 Oracle.  All rights reserved.
   4 */
   5
   6#include <linux/sched.h>
   7#include <linux/slab.h>
   8#include <linux/blkdev.h>
   9#include <linux/list_sort.h>
  10#include <linux/iversion.h>
 
  11#include "ctree.h"
  12#include "tree-log.h"
  13#include "disk-io.h"
  14#include "locking.h"
  15#include "print-tree.h"
  16#include "backref.h"
  17#include "compression.h"
  18#include "qgroup.h"
  19#include "inode-map.h"
 
 
  20
  21/* magic values for the inode_only field in btrfs_log_inode:
  22 *
  23 * LOG_INODE_ALL means to log everything
  24 * LOG_INODE_EXISTS means to log just enough to recreate the inode
  25 * during log replay
  26 */
  27#define LOG_INODE_ALL 0
  28#define LOG_INODE_EXISTS 1
  29#define LOG_OTHER_INODE 2
 
 
 
  30
  31/*
  32 * directory trouble cases
  33 *
  34 * 1) on rename or unlink, if the inode being unlinked isn't in the fsync
  35 * log, we must force a full commit before doing an fsync of the directory
  36 * where the unlink was done.
  37 * ---> record transid of last unlink/rename per directory
  38 *
  39 * mkdir foo/some_dir
  40 * normal commit
  41 * rename foo/some_dir foo2/some_dir
  42 * mkdir foo/some_dir
  43 * fsync foo/some_dir/some_file
  44 *
  45 * The fsync above will unlink the original some_dir without recording
  46 * it in its new location (foo2).  After a crash, some_dir will be gone
  47 * unless the fsync of some_file forces a full commit
  48 *
  49 * 2) we must log any new names for any file or dir that is in the fsync
  50 * log. ---> check inode while renaming/linking.
  51 *
  52 * 2a) we must log any new names for any file or dir during rename
  53 * when the directory they are being removed from was logged.
  54 * ---> check inode and old parent dir during rename
  55 *
  56 *  2a is actually the more important variant.  With the extra logging
  57 *  a crash might unlink the old name without recreating the new one
  58 *
  59 * 3) after a crash, we must go through any directories with a link count
  60 * of zero and redo the rm -rf
  61 *
  62 * mkdir f1/foo
  63 * normal commit
  64 * rm -rf f1/foo
  65 * fsync(f1)
  66 *
  67 * The directory f1 was fully removed from the FS, but fsync was never
  68 * called on f1, only its parent dir.  After a crash the rm -rf must
  69 * be replayed.  This must be able to recurse down the entire
  70 * directory tree.  The inode link count fixup code takes care of the
  71 * ugly details.
  72 */
  73
  74/*
  75 * stages for the tree walking.  The first
  76 * stage (0) is to only pin down the blocks we find
  77 * the second stage (1) is to make sure that all the inodes
  78 * we find in the log are created in the subvolume.
  79 *
  80 * The last stage is to deal with directories and links and extents
  81 * and all the other fun semantics
  82 */
  83#define LOG_WALK_PIN_ONLY 0
  84#define LOG_WALK_REPLAY_INODES 1
  85#define LOG_WALK_REPLAY_DIR_INDEX 2
  86#define LOG_WALK_REPLAY_ALL 3
 
 
  87
  88static int btrfs_log_inode(struct btrfs_trans_handle *trans,
  89			   struct btrfs_root *root, struct btrfs_inode *inode,
  90			   int inode_only,
  91			   const loff_t start,
  92			   const loff_t end,
  93			   struct btrfs_log_ctx *ctx);
  94static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
  95			     struct btrfs_root *root,
  96			     struct btrfs_path *path, u64 objectid);
  97static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
  98				       struct btrfs_root *root,
  99				       struct btrfs_root *log,
 100				       struct btrfs_path *path,
 101				       u64 dirid, int del_all);
 
 102
 103/*
 104 * tree logging is a special write ahead log used to make sure that
 105 * fsyncs and O_SYNCs can happen without doing full tree commits.
 106 *
 107 * Full tree commits are expensive because they require commonly
 108 * modified blocks to be recowed, creating many dirty pages in the
 109 * extent tree an 4x-6x higher write load than ext3.
 110 *
 111 * Instead of doing a tree commit on every fsync, we use the
 112 * key ranges and transaction ids to find items for a given file or directory
 113 * that have changed in this transaction.  Those items are copied into
 114 * a special tree (one per subvolume root), that tree is written to disk
 115 * and then the fsync is considered complete.
 116 *
 117 * After a crash, items are copied out of the log-tree back into the
 118 * subvolume tree.  Any file data extents found are recorded in the extent
 119 * allocation tree, and the log-tree freed.
 120 *
 121 * The log tree is read three times, once to pin down all the extents it is
 122 * using in ram and once, once to create all the inodes logged in the tree
 123 * and once to do all the other items.
 124 */
 125
 126/*
 127 * start a sub transaction and setup the log tree
 128 * this increments the log tree writer count to make the people
 129 * syncing the tree wait for us to finish
 130 */
 131static int start_log_trans(struct btrfs_trans_handle *trans,
 132			   struct btrfs_root *root,
 133			   struct btrfs_log_ctx *ctx)
 134{
 135	struct btrfs_fs_info *fs_info = root->fs_info;
 
 
 136	int ret = 0;
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 137
 138	mutex_lock(&root->log_mutex);
 139
 
 140	if (root->log_root) {
 141		if (btrfs_need_log_full_commit(fs_info, trans)) {
 
 
 142			ret = -EAGAIN;
 143			goto out;
 144		}
 145
 
 
 
 
 
 146		if (!root->log_start_pid) {
 147			clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
 148			root->log_start_pid = current->pid;
 149		} else if (root->log_start_pid != current->pid) {
 150			set_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
 151		}
 152	} else {
 153		mutex_lock(&fs_info->tree_log_mutex);
 154		if (!fs_info->log_root_tree)
 155			ret = btrfs_init_log_root_tree(trans, fs_info);
 156		mutex_unlock(&fs_info->tree_log_mutex);
 157		if (ret)
 
 
 
 158			goto out;
 
 159
 160		ret = btrfs_add_log_tree(trans, root);
 161		if (ret)
 162			goto out;
 163
 
 164		clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
 165		root->log_start_pid = current->pid;
 166	}
 167
 168	atomic_inc(&root->log_batch);
 169	atomic_inc(&root->log_writers);
 170	if (ctx) {
 171		int index = root->log_transid % 2;
 172		list_add_tail(&ctx->list, &root->log_ctxs[index]);
 173		ctx->log_transid = root->log_transid;
 174	}
 175
 176out:
 177	mutex_unlock(&root->log_mutex);
 178	return ret;
 179}
 180
 181/*
 182 * returns 0 if there was a log transaction running and we were able
 183 * to join, or returns -ENOENT if there were not transactions
 184 * in progress
 185 */
 186static int join_running_log_trans(struct btrfs_root *root)
 187{
 
 188	int ret = -ENOENT;
 189
 190	smp_mb();
 191	if (!root->log_root)
 192		return -ENOENT;
 193
 194	mutex_lock(&root->log_mutex);
 
 195	if (root->log_root) {
 
 
 196		ret = 0;
 
 
 
 
 197		atomic_inc(&root->log_writers);
 198	}
 199	mutex_unlock(&root->log_mutex);
 200	return ret;
 201}
 202
 203/*
 204 * This either makes the current running log transaction wait
 205 * until you call btrfs_end_log_trans() or it makes any future
 206 * log transactions wait until you call btrfs_end_log_trans()
 207 */
 208int btrfs_pin_log_trans(struct btrfs_root *root)
 209{
 210	int ret = -ENOENT;
 211
 212	mutex_lock(&root->log_mutex);
 213	atomic_inc(&root->log_writers);
 214	mutex_unlock(&root->log_mutex);
 215	return ret;
 216}
 217
 218/*
 219 * indicate we're done making changes to the log tree
 220 * and wake up anyone waiting to do a sync
 221 */
 222void btrfs_end_log_trans(struct btrfs_root *root)
 223{
 224	if (atomic_dec_and_test(&root->log_writers)) {
 225		/*
 226		 * Implicit memory barrier after atomic_dec_and_test
 227		 */
 228		if (waitqueue_active(&root->log_writer_wait))
 229			wake_up(&root->log_writer_wait);
 230	}
 231}
 232
 
 
 
 
 
 
 
 
 
 
 
 233
 234/*
 235 * the walk control struct is used to pass state down the chain when
 236 * processing the log tree.  The stage field tells us which part
 237 * of the log tree processing we are currently doing.  The others
 238 * are state fields used for that specific part
 239 */
 240struct walk_control {
 241	/* should we free the extent on disk when done?  This is used
 242	 * at transaction commit time while freeing a log tree
 243	 */
 244	int free;
 245
 246	/* should we write out the extent buffer?  This is used
 247	 * while flushing the log tree to disk during a sync
 248	 */
 249	int write;
 250
 251	/* should we wait for the extent buffer io to finish?  Also used
 252	 * while flushing the log tree to disk for a sync
 253	 */
 254	int wait;
 255
 256	/* pin only walk, we record which extents on disk belong to the
 257	 * log trees
 258	 */
 259	int pin;
 260
 261	/* what stage of the replay code we're currently in */
 262	int stage;
 263
 
 
 
 
 
 
 
 264	/* the root we are currently replaying */
 265	struct btrfs_root *replay_dest;
 266
 267	/* the trans handle for the current replay */
 268	struct btrfs_trans_handle *trans;
 269
 270	/* the function that gets used to process blocks we find in the
 271	 * tree.  Note the extent_buffer might not be up to date when it is
 272	 * passed in, and it must be checked or read if you need the data
 273	 * inside it
 274	 */
 275	int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
 276			    struct walk_control *wc, u64 gen, int level);
 277};
 278
 279/*
 280 * process_func used to pin down extents, write them or wait on them
 281 */
 282static int process_one_buffer(struct btrfs_root *log,
 283			      struct extent_buffer *eb,
 284			      struct walk_control *wc, u64 gen, int level)
 285{
 286	struct btrfs_fs_info *fs_info = log->fs_info;
 287	int ret = 0;
 288
 289	/*
 290	 * If this fs is mixed then we need to be able to process the leaves to
 291	 * pin down any logged extents, so we have to read the block.
 292	 */
 293	if (btrfs_fs_incompat(fs_info, MIXED_GROUPS)) {
 294		ret = btrfs_read_buffer(eb, gen, level, NULL);
 295		if (ret)
 296			return ret;
 297	}
 298
 299	if (wc->pin)
 300		ret = btrfs_pin_extent_for_log_replay(fs_info, eb->start,
 301						      eb->len);
 302
 303	if (!ret && btrfs_buffer_uptodate(eb, gen, 0)) {
 304		if (wc->pin && btrfs_header_level(eb) == 0)
 305			ret = btrfs_exclude_logged_extents(fs_info, eb);
 306		if (wc->write)
 307			btrfs_write_tree_block(eb);
 308		if (wc->wait)
 309			btrfs_wait_tree_block_writeback(eb);
 310	}
 311	return ret;
 312}
 313
 314/*
 315 * Item overwrite used by replay and tree logging.  eb, slot and key all refer
 316 * to the src data we are copying out.
 317 *
 318 * root is the tree we are copying into, and path is a scratch
 319 * path for use in this function (it should be released on entry and
 320 * will be released on exit).
 321 *
 322 * If the key is already in the destination tree the existing item is
 323 * overwritten.  If the existing item isn't big enough, it is extended.
 324 * If it is too large, it is truncated.
 325 *
 326 * If the key isn't in the destination yet, a new item is inserted.
 327 */
 328static noinline int overwrite_item(struct btrfs_trans_handle *trans,
 329				   struct btrfs_root *root,
 330				   struct btrfs_path *path,
 331				   struct extent_buffer *eb, int slot,
 332				   struct btrfs_key *key)
 333{
 334	struct btrfs_fs_info *fs_info = root->fs_info;
 335	int ret;
 336	u32 item_size;
 337	u64 saved_i_size = 0;
 338	int save_old_i_size = 0;
 339	unsigned long src_ptr;
 340	unsigned long dst_ptr;
 341	int overwrite_root = 0;
 342	bool inode_item = key->type == BTRFS_INODE_ITEM_KEY;
 343
 344	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
 345		overwrite_root = 1;
 346
 347	item_size = btrfs_item_size_nr(eb, slot);
 348	src_ptr = btrfs_item_ptr_offset(eb, slot);
 349
 350	/* look for the key in the destination tree */
 351	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
 352	if (ret < 0)
 353		return ret;
 354
 355	if (ret == 0) {
 356		char *src_copy;
 357		char *dst_copy;
 358		u32 dst_size = btrfs_item_size_nr(path->nodes[0],
 359						  path->slots[0]);
 360		if (dst_size != item_size)
 361			goto insert;
 362
 363		if (item_size == 0) {
 364			btrfs_release_path(path);
 365			return 0;
 366		}
 367		dst_copy = kmalloc(item_size, GFP_NOFS);
 368		src_copy = kmalloc(item_size, GFP_NOFS);
 369		if (!dst_copy || !src_copy) {
 370			btrfs_release_path(path);
 371			kfree(dst_copy);
 372			kfree(src_copy);
 373			return -ENOMEM;
 374		}
 375
 376		read_extent_buffer(eb, src_copy, src_ptr, item_size);
 377
 378		dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
 379		read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
 380				   item_size);
 381		ret = memcmp(dst_copy, src_copy, item_size);
 382
 383		kfree(dst_copy);
 384		kfree(src_copy);
 385		/*
 386		 * they have the same contents, just return, this saves
 387		 * us from cowing blocks in the destination tree and doing
 388		 * extra writes that may not have been done by a previous
 389		 * sync
 390		 */
 391		if (ret == 0) {
 392			btrfs_release_path(path);
 393			return 0;
 394		}
 395
 396		/*
 397		 * We need to load the old nbytes into the inode so when we
 398		 * replay the extents we've logged we get the right nbytes.
 399		 */
 400		if (inode_item) {
 401			struct btrfs_inode_item *item;
 402			u64 nbytes;
 403			u32 mode;
 404
 405			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
 406					      struct btrfs_inode_item);
 407			nbytes = btrfs_inode_nbytes(path->nodes[0], item);
 408			item = btrfs_item_ptr(eb, slot,
 409					      struct btrfs_inode_item);
 410			btrfs_set_inode_nbytes(eb, item, nbytes);
 411
 412			/*
 413			 * If this is a directory we need to reset the i_size to
 414			 * 0 so that we can set it up properly when replaying
 415			 * the rest of the items in this log.
 416			 */
 417			mode = btrfs_inode_mode(eb, item);
 418			if (S_ISDIR(mode))
 419				btrfs_set_inode_size(eb, item, 0);
 420		}
 421	} else if (inode_item) {
 422		struct btrfs_inode_item *item;
 423		u32 mode;
 424
 425		/*
 426		 * New inode, set nbytes to 0 so that the nbytes comes out
 427		 * properly when we replay the extents.
 428		 */
 429		item = btrfs_item_ptr(eb, slot, struct btrfs_inode_item);
 430		btrfs_set_inode_nbytes(eb, item, 0);
 431
 432		/*
 433		 * If this is a directory we need to reset the i_size to 0 so
 434		 * that we can set it up properly when replaying the rest of
 435		 * the items in this log.
 436		 */
 437		mode = btrfs_inode_mode(eb, item);
 438		if (S_ISDIR(mode))
 439			btrfs_set_inode_size(eb, item, 0);
 440	}
 441insert:
 442	btrfs_release_path(path);
 443	/* try to insert the key into the destination tree */
 444	path->skip_release_on_error = 1;
 445	ret = btrfs_insert_empty_item(trans, root, path,
 446				      key, item_size);
 447	path->skip_release_on_error = 0;
 448
 449	/* make sure any existing item is the correct size */
 450	if (ret == -EEXIST || ret == -EOVERFLOW) {
 451		u32 found_size;
 452		found_size = btrfs_item_size_nr(path->nodes[0],
 453						path->slots[0]);
 454		if (found_size > item_size)
 455			btrfs_truncate_item(fs_info, path, item_size, 1);
 456		else if (found_size < item_size)
 457			btrfs_extend_item(fs_info, path,
 458					  item_size - found_size);
 459	} else if (ret) {
 460		return ret;
 461	}
 462	dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
 463					path->slots[0]);
 464
 465	/* don't overwrite an existing inode if the generation number
 466	 * was logged as zero.  This is done when the tree logging code
 467	 * is just logging an inode to make sure it exists after recovery.
 468	 *
 469	 * Also, don't overwrite i_size on directories during replay.
 470	 * log replay inserts and removes directory items based on the
 471	 * state of the tree found in the subvolume, and i_size is modified
 472	 * as it goes
 473	 */
 474	if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
 475		struct btrfs_inode_item *src_item;
 476		struct btrfs_inode_item *dst_item;
 477
 478		src_item = (struct btrfs_inode_item *)src_ptr;
 479		dst_item = (struct btrfs_inode_item *)dst_ptr;
 480
 481		if (btrfs_inode_generation(eb, src_item) == 0) {
 482			struct extent_buffer *dst_eb = path->nodes[0];
 483			const u64 ino_size = btrfs_inode_size(eb, src_item);
 484
 485			/*
 486			 * For regular files an ino_size == 0 is used only when
 487			 * logging that an inode exists, as part of a directory
 488			 * fsync, and the inode wasn't fsynced before. In this
 489			 * case don't set the size of the inode in the fs/subvol
 490			 * tree, otherwise we would be throwing valid data away.
 491			 */
 492			if (S_ISREG(btrfs_inode_mode(eb, src_item)) &&
 493			    S_ISREG(btrfs_inode_mode(dst_eb, dst_item)) &&
 494			    ino_size != 0) {
 495				struct btrfs_map_token token;
 496
 497				btrfs_init_map_token(&token);
 498				btrfs_set_token_inode_size(dst_eb, dst_item,
 499							   ino_size, &token);
 500			}
 501			goto no_copy;
 502		}
 503
 504		if (overwrite_root &&
 505		    S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
 506		    S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
 507			save_old_i_size = 1;
 508			saved_i_size = btrfs_inode_size(path->nodes[0],
 509							dst_item);
 510		}
 511	}
 512
 513	copy_extent_buffer(path->nodes[0], eb, dst_ptr,
 514			   src_ptr, item_size);
 515
 516	if (save_old_i_size) {
 517		struct btrfs_inode_item *dst_item;
 518		dst_item = (struct btrfs_inode_item *)dst_ptr;
 519		btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
 520	}
 521
 522	/* make sure the generation is filled in */
 523	if (key->type == BTRFS_INODE_ITEM_KEY) {
 524		struct btrfs_inode_item *dst_item;
 525		dst_item = (struct btrfs_inode_item *)dst_ptr;
 526		if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
 527			btrfs_set_inode_generation(path->nodes[0], dst_item,
 528						   trans->transid);
 529		}
 530	}
 531no_copy:
 532	btrfs_mark_buffer_dirty(path->nodes[0]);
 533	btrfs_release_path(path);
 534	return 0;
 535}
 536
 537/*
 538 * simple helper to read an inode off the disk from a given root
 539 * This can only be called for subvolume roots and not for the log
 540 */
 541static noinline struct inode *read_one_inode(struct btrfs_root *root,
 542					     u64 objectid)
 543{
 544	struct btrfs_key key;
 545	struct inode *inode;
 546
 547	key.objectid = objectid;
 548	key.type = BTRFS_INODE_ITEM_KEY;
 549	key.offset = 0;
 550	inode = btrfs_iget(root->fs_info->sb, &key, root, NULL);
 551	if (IS_ERR(inode)) {
 552		inode = NULL;
 553	} else if (is_bad_inode(inode)) {
 554		iput(inode);
 555		inode = NULL;
 556	}
 557	return inode;
 558}
 559
 560/* replays a single extent in 'eb' at 'slot' with 'key' into the
 561 * subvolume 'root'.  path is released on entry and should be released
 562 * on exit.
 563 *
 564 * extents in the log tree have not been allocated out of the extent
 565 * tree yet.  So, this completes the allocation, taking a reference
 566 * as required if the extent already exists or creating a new extent
 567 * if it isn't in the extent allocation tree yet.
 568 *
 569 * The extent is inserted into the file, dropping any existing extents
 570 * from the file that overlap the new one.
 571 */
 572static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 573				      struct btrfs_root *root,
 574				      struct btrfs_path *path,
 575				      struct extent_buffer *eb, int slot,
 576				      struct btrfs_key *key)
 577{
 
 578	struct btrfs_fs_info *fs_info = root->fs_info;
 579	int found_type;
 580	u64 extent_end;
 581	u64 start = key->offset;
 582	u64 nbytes = 0;
 583	struct btrfs_file_extent_item *item;
 584	struct inode *inode = NULL;
 585	unsigned long size;
 586	int ret = 0;
 587
 588	item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
 589	found_type = btrfs_file_extent_type(eb, item);
 590
 591	if (found_type == BTRFS_FILE_EXTENT_REG ||
 592	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
 593		nbytes = btrfs_file_extent_num_bytes(eb, item);
 594		extent_end = start + nbytes;
 595
 596		/*
 597		 * We don't add to the inodes nbytes if we are prealloc or a
 598		 * hole.
 599		 */
 600		if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
 601			nbytes = 0;
 602	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
 603		size = btrfs_file_extent_inline_len(eb, slot, item);
 604		nbytes = btrfs_file_extent_ram_bytes(eb, item);
 605		extent_end = ALIGN(start + size,
 606				   fs_info->sectorsize);
 607	} else {
 608		ret = 0;
 609		goto out;
 610	}
 611
 612	inode = read_one_inode(root, key->objectid);
 613	if (!inode) {
 614		ret = -EIO;
 615		goto out;
 616	}
 617
 618	/*
 619	 * first check to see if we already have this extent in the
 620	 * file.  This must be done before the btrfs_drop_extents run
 621	 * so we don't try to drop this extent.
 622	 */
 623	ret = btrfs_lookup_file_extent(trans, root, path,
 624			btrfs_ino(BTRFS_I(inode)), start, 0);
 625
 626	if (ret == 0 &&
 627	    (found_type == BTRFS_FILE_EXTENT_REG ||
 628	     found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
 629		struct btrfs_file_extent_item cmp1;
 630		struct btrfs_file_extent_item cmp2;
 631		struct btrfs_file_extent_item *existing;
 632		struct extent_buffer *leaf;
 633
 634		leaf = path->nodes[0];
 635		existing = btrfs_item_ptr(leaf, path->slots[0],
 636					  struct btrfs_file_extent_item);
 637
 638		read_extent_buffer(eb, &cmp1, (unsigned long)item,
 639				   sizeof(cmp1));
 640		read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
 641				   sizeof(cmp2));
 642
 643		/*
 644		 * we already have a pointer to this exact extent,
 645		 * we don't have to do anything
 646		 */
 647		if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
 648			btrfs_release_path(path);
 649			goto out;
 650		}
 651	}
 652	btrfs_release_path(path);
 653
 654	/* drop any overlapping extents */
 655	ret = btrfs_drop_extents(trans, root, inode, start, extent_end, 1);
 
 
 
 656	if (ret)
 657		goto out;
 658
 659	if (found_type == BTRFS_FILE_EXTENT_REG ||
 660	    found_type == BTRFS_FILE_EXTENT_PREALLOC) {
 661		u64 offset;
 662		unsigned long dest_offset;
 663		struct btrfs_key ins;
 664
 665		if (btrfs_file_extent_disk_bytenr(eb, item) == 0 &&
 666		    btrfs_fs_incompat(fs_info, NO_HOLES))
 667			goto update_inode;
 668
 669		ret = btrfs_insert_empty_item(trans, root, path, key,
 670					      sizeof(*item));
 671		if (ret)
 672			goto out;
 673		dest_offset = btrfs_item_ptr_offset(path->nodes[0],
 674						    path->slots[0]);
 675		copy_extent_buffer(path->nodes[0], eb, dest_offset,
 676				(unsigned long)item,  sizeof(*item));
 677
 678		ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
 679		ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
 680		ins.type = BTRFS_EXTENT_ITEM_KEY;
 681		offset = key->offset - btrfs_file_extent_offset(eb, item);
 682
 683		/*
 684		 * Manually record dirty extent, as here we did a shallow
 685		 * file extent item copy and skip normal backref update,
 686		 * but modifying extent tree all by ourselves.
 687		 * So need to manually record dirty extent for qgroup,
 688		 * as the owner of the file extent changed from log tree
 689		 * (doesn't affect qgroup) to fs/file tree(affects qgroup)
 690		 */
 691		ret = btrfs_qgroup_trace_extent(trans, fs_info,
 692				btrfs_file_extent_disk_bytenr(eb, item),
 693				btrfs_file_extent_disk_num_bytes(eb, item),
 694				GFP_NOFS);
 695		if (ret < 0)
 696			goto out;
 697
 698		if (ins.objectid > 0) {
 
 699			u64 csum_start;
 700			u64 csum_end;
 701			LIST_HEAD(ordered_sums);
 
 702			/*
 703			 * is this extent already allocated in the extent
 704			 * allocation tree?  If so, just add a reference
 705			 */
 706			ret = btrfs_lookup_data_extent(fs_info, ins.objectid,
 707						ins.offset);
 708			if (ret == 0) {
 709				ret = btrfs_inc_extent_ref(trans, root,
 710						ins.objectid, ins.offset,
 711						0, root->root_key.objectid,
 
 
 
 
 712						key->objectid, offset);
 
 713				if (ret)
 714					goto out;
 715			} else {
 716				/*
 717				 * insert the extent pointer in the extent
 718				 * allocation tree
 719				 */
 720				ret = btrfs_alloc_logged_file_extent(trans,
 721						fs_info,
 722						root->root_key.objectid,
 723						key->objectid, offset, &ins);
 724				if (ret)
 725					goto out;
 726			}
 727			btrfs_release_path(path);
 728
 729			if (btrfs_file_extent_compression(eb, item)) {
 730				csum_start = ins.objectid;
 731				csum_end = csum_start + ins.offset;
 732			} else {
 733				csum_start = ins.objectid +
 734					btrfs_file_extent_offset(eb, item);
 735				csum_end = csum_start +
 736					btrfs_file_extent_num_bytes(eb, item);
 737			}
 738
 739			ret = btrfs_lookup_csums_range(root->log_root,
 740						csum_start, csum_end - 1,
 741						&ordered_sums, 0);
 742			if (ret)
 743				goto out;
 744			/*
 745			 * Now delete all existing cums in the csum root that
 746			 * cover our range. We do this because we can have an
 747			 * extent that is completely referenced by one file
 748			 * extent item and partially referenced by another
 749			 * file extent item (like after using the clone or
 750			 * extent_same ioctls). In this case if we end up doing
 751			 * the replay of the one that partially references the
 752			 * extent first, and we do not do the csum deletion
 753			 * below, we can get 2 csum items in the csum tree that
 754			 * overlap each other. For example, imagine our log has
 755			 * the two following file extent items:
 756			 *
 757			 * key (257 EXTENT_DATA 409600)
 758			 *     extent data disk byte 12845056 nr 102400
 759			 *     extent data offset 20480 nr 20480 ram 102400
 760			 *
 761			 * key (257 EXTENT_DATA 819200)
 762			 *     extent data disk byte 12845056 nr 102400
 763			 *     extent data offset 0 nr 102400 ram 102400
 764			 *
 765			 * Where the second one fully references the 100K extent
 766			 * that starts at disk byte 12845056, and the log tree
 767			 * has a single csum item that covers the entire range
 768			 * of the extent:
 769			 *
 770			 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
 771			 *
 772			 * After the first file extent item is replayed, the
 773			 * csum tree gets the following csum item:
 774			 *
 775			 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
 776			 *
 777			 * Which covers the 20K sub-range starting at offset 20K
 778			 * of our extent. Now when we replay the second file
 779			 * extent item, if we do not delete existing csum items
 780			 * that cover any of its blocks, we end up getting two
 781			 * csum items in our csum tree that overlap each other:
 782			 *
 783			 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
 784			 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
 785			 *
 786			 * Which is a problem, because after this anyone trying
 787			 * to lookup up for the checksum of any block of our
 788			 * extent starting at an offset of 40K or higher, will
 789			 * end up looking at the second csum item only, which
 790			 * does not contain the checksum for any block starting
 791			 * at offset 40K or higher of our extent.
 792			 */
 793			while (!list_empty(&ordered_sums)) {
 794				struct btrfs_ordered_sum *sums;
 795				sums = list_entry(ordered_sums.next,
 796						struct btrfs_ordered_sum,
 797						list);
 798				if (!ret)
 799					ret = btrfs_del_csums(trans, fs_info,
 
 800							      sums->bytenr,
 801							      sums->len);
 802				if (!ret)
 803					ret = btrfs_csum_file_blocks(trans,
 804						fs_info->csum_root, sums);
 805				list_del(&sums->list);
 806				kfree(sums);
 807			}
 808			if (ret)
 809				goto out;
 810		} else {
 811			btrfs_release_path(path);
 812		}
 813	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
 814		/* inline extents are easy, we just overwrite them */
 815		ret = overwrite_item(trans, root, path, eb, slot, key);
 816		if (ret)
 817			goto out;
 818	}
 819
 820	inode_add_bytes(inode, nbytes);
 
 
 
 
 821update_inode:
 822	ret = btrfs_update_inode(trans, root, inode);
 
 823out:
 824	if (inode)
 825		iput(inode);
 826	return ret;
 827}
 828
 829/*
 830 * when cleaning up conflicts between the directory names in the
 831 * subvolume, directory names in the log and directory names in the
 832 * inode back references, we may have to unlink inodes from directories.
 833 *
 834 * This is a helper function to do the unlink of a specific directory
 835 * item
 836 */
 837static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
 838				      struct btrfs_root *root,
 839				      struct btrfs_path *path,
 840				      struct btrfs_inode *dir,
 841				      struct btrfs_dir_item *di)
 842{
 843	struct inode *inode;
 844	char *name;
 845	int name_len;
 846	struct extent_buffer *leaf;
 847	struct btrfs_key location;
 848	int ret;
 849
 850	leaf = path->nodes[0];
 851
 852	btrfs_dir_item_key_to_cpu(leaf, di, &location);
 853	name_len = btrfs_dir_name_len(leaf, di);
 854	name = kmalloc(name_len, GFP_NOFS);
 855	if (!name)
 856		return -ENOMEM;
 857
 858	read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
 859	btrfs_release_path(path);
 860
 861	inode = read_one_inode(root, location.objectid);
 862	if (!inode) {
 863		ret = -EIO;
 864		goto out;
 865	}
 866
 867	ret = link_to_fixup_dir(trans, root, path, location.objectid);
 868	if (ret)
 869		goto out;
 870
 871	ret = btrfs_unlink_inode(trans, root, dir, BTRFS_I(inode), name,
 872			name_len);
 873	if (ret)
 874		goto out;
 875	else
 876		ret = btrfs_run_delayed_items(trans);
 877out:
 878	kfree(name);
 879	iput(inode);
 880	return ret;
 881}
 882
 883/*
 884 * helper function to see if a given name and sequence number found
 885 * in an inode back reference are already in a directory and correctly
 886 * point to this inode
 
 
 887 */
 888static noinline int inode_in_dir(struct btrfs_root *root,
 889				 struct btrfs_path *path,
 890				 u64 dirid, u64 objectid, u64 index,
 891				 const char *name, int name_len)
 892{
 893	struct btrfs_dir_item *di;
 894	struct btrfs_key location;
 895	int match = 0;
 896
 897	di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
 898					 index, name, name_len, 0);
 899	if (di && !IS_ERR(di)) {
 
 
 
 
 900		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
 901		if (location.objectid != objectid)
 902			goto out;
 903	} else
 904		goto out;
 905	btrfs_release_path(path);
 906
 
 907	di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
 908	if (di && !IS_ERR(di)) {
 909		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
 910		if (location.objectid != objectid)
 911			goto out;
 912	} else
 913		goto out;
 914	match = 1;
 
 
 
 
 915out:
 916	btrfs_release_path(path);
 917	return match;
 918}
 919
 920/*
 921 * helper function to check a log tree for a named back reference in
 922 * an inode.  This is used to decide if a back reference that is
 923 * found in the subvolume conflicts with what we find in the log.
 924 *
 925 * inode backreferences may have multiple refs in a single item,
 926 * during replay we process one reference at a time, and we don't
 927 * want to delete valid links to a file from the subvolume if that
 928 * link is also in the log.
 929 */
 930static noinline int backref_in_log(struct btrfs_root *log,
 931				   struct btrfs_key *key,
 932				   u64 ref_objectid,
 933				   const char *name, int namelen)
 934{
 935	struct btrfs_path *path;
 936	struct btrfs_inode_ref *ref;
 937	unsigned long ptr;
 938	unsigned long ptr_end;
 939	unsigned long name_ptr;
 940	int found_name_len;
 941	int item_size;
 942	int ret;
 943	int match = 0;
 944
 945	path = btrfs_alloc_path();
 946	if (!path)
 947		return -ENOMEM;
 948
 949	ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
 950	if (ret != 0)
 951		goto out;
 952
 953	ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
 954
 955	if (key->type == BTRFS_INODE_EXTREF_KEY) {
 956		if (btrfs_find_name_in_ext_backref(path->nodes[0],
 957						   path->slots[0],
 958						   ref_objectid,
 959						   name, namelen, NULL))
 960			match = 1;
 961
 962		goto out;
 963	}
 964
 965	item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
 966	ptr_end = ptr + item_size;
 967	while (ptr < ptr_end) {
 968		ref = (struct btrfs_inode_ref *)ptr;
 969		found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
 970		if (found_name_len == namelen) {
 971			name_ptr = (unsigned long)(ref + 1);
 972			ret = memcmp_extent_buffer(path->nodes[0], name,
 973						   name_ptr, namelen);
 974			if (ret == 0) {
 975				match = 1;
 976				goto out;
 977			}
 978		}
 979		ptr = (unsigned long)(ref + 1) + found_name_len;
 980	}
 981out:
 982	btrfs_free_path(path);
 983	return match;
 984}
 985
 986static inline int __add_inode_ref(struct btrfs_trans_handle *trans,
 987				  struct btrfs_root *root,
 988				  struct btrfs_path *path,
 989				  struct btrfs_root *log_root,
 990				  struct btrfs_inode *dir,
 991				  struct btrfs_inode *inode,
 992				  u64 inode_objectid, u64 parent_objectid,
 993				  u64 ref_index, char *name, int namelen,
 994				  int *search_done)
 995{
 996	int ret;
 997	char *victim_name;
 998	int victim_name_len;
 999	struct extent_buffer *leaf;
1000	struct btrfs_dir_item *di;
1001	struct btrfs_key search_key;
1002	struct btrfs_inode_extref *extref;
1003
1004again:
1005	/* Search old style refs */
1006	search_key.objectid = inode_objectid;
1007	search_key.type = BTRFS_INODE_REF_KEY;
1008	search_key.offset = parent_objectid;
1009	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
1010	if (ret == 0) {
1011		struct btrfs_inode_ref *victim_ref;
1012		unsigned long ptr;
1013		unsigned long ptr_end;
1014
1015		leaf = path->nodes[0];
1016
1017		/* are we trying to overwrite a back ref for the root directory
1018		 * if so, just jump out, we're done
1019		 */
1020		if (search_key.objectid == search_key.offset)
1021			return 1;
1022
1023		/* check all the names in this back reference to see
1024		 * if they are in the log.  if so, we allow them to stay
1025		 * otherwise they must be unlinked as a conflict
1026		 */
1027		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1028		ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
1029		while (ptr < ptr_end) {
1030			victim_ref = (struct btrfs_inode_ref *)ptr;
1031			victim_name_len = btrfs_inode_ref_name_len(leaf,
1032								   victim_ref);
1033			victim_name = kmalloc(victim_name_len, GFP_NOFS);
1034			if (!victim_name)
1035				return -ENOMEM;
1036
1037			read_extent_buffer(leaf, victim_name,
1038					   (unsigned long)(victim_ref + 1),
1039					   victim_name_len);
1040
1041			if (!backref_in_log(log_root, &search_key,
1042					    parent_objectid,
1043					    victim_name,
1044					    victim_name_len)) {
 
 
 
1045				inc_nlink(&inode->vfs_inode);
1046				btrfs_release_path(path);
1047
1048				ret = btrfs_unlink_inode(trans, root, dir, inode,
1049						victim_name, victim_name_len);
1050				kfree(victim_name);
1051				if (ret)
1052					return ret;
1053				ret = btrfs_run_delayed_items(trans);
1054				if (ret)
1055					return ret;
1056				*search_done = 1;
1057				goto again;
1058			}
1059			kfree(victim_name);
1060
1061			ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
1062		}
1063
1064		/*
1065		 * NOTE: we have searched root tree and checked the
1066		 * corresponding ref, it does not need to check again.
1067		 */
1068		*search_done = 1;
1069	}
1070	btrfs_release_path(path);
1071
1072	/* Same search but for extended refs */
1073	extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
1074					   inode_objectid, parent_objectid, 0,
1075					   0);
1076	if (!IS_ERR_OR_NULL(extref)) {
1077		u32 item_size;
1078		u32 cur_offset = 0;
1079		unsigned long base;
1080		struct inode *victim_parent;
1081
1082		leaf = path->nodes[0];
1083
1084		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1085		base = btrfs_item_ptr_offset(leaf, path->slots[0]);
1086
1087		while (cur_offset < item_size) {
1088			extref = (struct btrfs_inode_extref *)(base + cur_offset);
1089
1090			victim_name_len = btrfs_inode_extref_name_len(leaf, extref);
1091
1092			if (btrfs_inode_extref_parent(leaf, extref) != parent_objectid)
1093				goto next;
1094
1095			victim_name = kmalloc(victim_name_len, GFP_NOFS);
1096			if (!victim_name)
1097				return -ENOMEM;
1098			read_extent_buffer(leaf, victim_name, (unsigned long)&extref->name,
1099					   victim_name_len);
1100
1101			search_key.objectid = inode_objectid;
1102			search_key.type = BTRFS_INODE_EXTREF_KEY;
1103			search_key.offset = btrfs_extref_hash(parent_objectid,
1104							      victim_name,
1105							      victim_name_len);
1106			ret = 0;
1107			if (!backref_in_log(log_root, &search_key,
1108					    parent_objectid, victim_name,
1109					    victim_name_len)) {
 
 
1110				ret = -ENOENT;
1111				victim_parent = read_one_inode(root,
1112						parent_objectid);
1113				if (victim_parent) {
1114					inc_nlink(&inode->vfs_inode);
1115					btrfs_release_path(path);
1116
1117					ret = btrfs_unlink_inode(trans, root,
1118							BTRFS_I(victim_parent),
1119							inode,
1120							victim_name,
1121							victim_name_len);
1122					if (!ret)
1123						ret = btrfs_run_delayed_items(
1124								  trans);
1125				}
1126				iput(victim_parent);
1127				kfree(victim_name);
1128				if (ret)
1129					return ret;
1130				*search_done = 1;
1131				goto again;
1132			}
1133			kfree(victim_name);
1134next:
1135			cur_offset += victim_name_len + sizeof(*extref);
1136		}
1137		*search_done = 1;
1138	}
1139	btrfs_release_path(path);
1140
1141	/* look for a conflicting sequence number */
1142	di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
1143					 ref_index, name, namelen, 0);
1144	if (di && !IS_ERR(di)) {
 
 
 
1145		ret = drop_one_dir_item(trans, root, path, dir, di);
1146		if (ret)
1147			return ret;
1148	}
1149	btrfs_release_path(path);
1150
1151	/* look for a conflicing name */
1152	di = btrfs_lookup_dir_item(trans, root, path, btrfs_ino(dir),
1153				   name, namelen, 0);
1154	if (di && !IS_ERR(di)) {
 
 
1155		ret = drop_one_dir_item(trans, root, path, dir, di);
1156		if (ret)
1157			return ret;
1158	}
1159	btrfs_release_path(path);
1160
1161	return 0;
1162}
1163
1164static int extref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
1165			     u32 *namelen, char **name, u64 *index,
1166			     u64 *parent_objectid)
1167{
1168	struct btrfs_inode_extref *extref;
1169
1170	extref = (struct btrfs_inode_extref *)ref_ptr;
1171
1172	*namelen = btrfs_inode_extref_name_len(eb, extref);
1173	*name = kmalloc(*namelen, GFP_NOFS);
1174	if (*name == NULL)
1175		return -ENOMEM;
1176
1177	read_extent_buffer(eb, *name, (unsigned long)&extref->name,
1178			   *namelen);
1179
1180	if (index)
1181		*index = btrfs_inode_extref_index(eb, extref);
1182	if (parent_objectid)
1183		*parent_objectid = btrfs_inode_extref_parent(eb, extref);
1184
1185	return 0;
1186}
1187
1188static int ref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
1189			  u32 *namelen, char **name, u64 *index)
1190{
1191	struct btrfs_inode_ref *ref;
1192
1193	ref = (struct btrfs_inode_ref *)ref_ptr;
1194
1195	*namelen = btrfs_inode_ref_name_len(eb, ref);
1196	*name = kmalloc(*namelen, GFP_NOFS);
1197	if (*name == NULL)
1198		return -ENOMEM;
1199
1200	read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);
1201
1202	if (index)
1203		*index = btrfs_inode_ref_index(eb, ref);
1204
1205	return 0;
1206}
1207
1208/*
1209 * Take an inode reference item from the log tree and iterate all names from the
1210 * inode reference item in the subvolume tree with the same key (if it exists).
1211 * For any name that is not in the inode reference item from the log tree, do a
1212 * proper unlink of that name (that is, remove its entry from the inode
1213 * reference item and both dir index keys).
1214 */
1215static int unlink_old_inode_refs(struct btrfs_trans_handle *trans,
1216				 struct btrfs_root *root,
1217				 struct btrfs_path *path,
1218				 struct btrfs_inode *inode,
1219				 struct extent_buffer *log_eb,
1220				 int log_slot,
1221				 struct btrfs_key *key)
1222{
1223	int ret;
1224	unsigned long ref_ptr;
1225	unsigned long ref_end;
1226	struct extent_buffer *eb;
1227
1228again:
1229	btrfs_release_path(path);
1230	ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
1231	if (ret > 0) {
1232		ret = 0;
1233		goto out;
1234	}
1235	if (ret < 0)
1236		goto out;
1237
1238	eb = path->nodes[0];
1239	ref_ptr = btrfs_item_ptr_offset(eb, path->slots[0]);
1240	ref_end = ref_ptr + btrfs_item_size_nr(eb, path->slots[0]);
1241	while (ref_ptr < ref_end) {
1242		char *name = NULL;
1243		int namelen;
1244		u64 parent_id;
1245
1246		if (key->type == BTRFS_INODE_EXTREF_KEY) {
1247			ret = extref_get_fields(eb, ref_ptr, &namelen, &name,
1248						NULL, &parent_id);
1249		} else {
1250			parent_id = key->offset;
1251			ret = ref_get_fields(eb, ref_ptr, &namelen, &name,
1252					     NULL);
1253		}
1254		if (ret)
1255			goto out;
1256
1257		if (key->type == BTRFS_INODE_EXTREF_KEY)
1258			ret = btrfs_find_name_in_ext_backref(log_eb, log_slot,
1259							     parent_id, name,
1260							     namelen, NULL);
1261		else
1262			ret = btrfs_find_name_in_backref(log_eb, log_slot, name,
1263							 namelen, NULL);
1264
1265		if (!ret) {
1266			struct inode *dir;
1267
1268			btrfs_release_path(path);
1269			dir = read_one_inode(root, parent_id);
1270			if (!dir) {
1271				ret = -ENOENT;
1272				kfree(name);
1273				goto out;
1274			}
1275			ret = btrfs_unlink_inode(trans, root, BTRFS_I(dir),
1276						 inode, name, namelen);
1277			kfree(name);
1278			iput(dir);
1279			if (ret)
1280				goto out;
1281			goto again;
1282		}
1283
1284		kfree(name);
1285		ref_ptr += namelen;
1286		if (key->type == BTRFS_INODE_EXTREF_KEY)
1287			ref_ptr += sizeof(struct btrfs_inode_extref);
1288		else
1289			ref_ptr += sizeof(struct btrfs_inode_ref);
1290	}
1291	ret = 0;
1292 out:
1293	btrfs_release_path(path);
1294	return ret;
1295}
1296
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1297/*
1298 * replay one inode back reference item found in the log tree.
1299 * eb, slot and key refer to the buffer and key found in the log tree.
1300 * root is the destination we are replaying into, and path is for temp
1301 * use by this function.  (it should be released on return).
1302 */
1303static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
1304				  struct btrfs_root *root,
1305				  struct btrfs_root *log,
1306				  struct btrfs_path *path,
1307				  struct extent_buffer *eb, int slot,
1308				  struct btrfs_key *key)
1309{
1310	struct inode *dir = NULL;
1311	struct inode *inode = NULL;
1312	unsigned long ref_ptr;
1313	unsigned long ref_end;
1314	char *name = NULL;
1315	int namelen;
1316	int ret;
1317	int search_done = 0;
1318	int log_ref_ver = 0;
1319	u64 parent_objectid;
1320	u64 inode_objectid;
1321	u64 ref_index = 0;
1322	int ref_struct_size;
1323
1324	ref_ptr = btrfs_item_ptr_offset(eb, slot);
1325	ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
1326
1327	if (key->type == BTRFS_INODE_EXTREF_KEY) {
1328		struct btrfs_inode_extref *r;
1329
1330		ref_struct_size = sizeof(struct btrfs_inode_extref);
1331		log_ref_ver = 1;
1332		r = (struct btrfs_inode_extref *)ref_ptr;
1333		parent_objectid = btrfs_inode_extref_parent(eb, r);
1334	} else {
1335		ref_struct_size = sizeof(struct btrfs_inode_ref);
1336		parent_objectid = key->offset;
1337	}
1338	inode_objectid = key->objectid;
1339
1340	/*
1341	 * it is possible that we didn't log all the parent directories
1342	 * for a given inode.  If we don't find the dir, just don't
1343	 * copy the back ref in.  The link count fixup code will take
1344	 * care of the rest
1345	 */
1346	dir = read_one_inode(root, parent_objectid);
1347	if (!dir) {
1348		ret = -ENOENT;
1349		goto out;
1350	}
1351
1352	inode = read_one_inode(root, inode_objectid);
1353	if (!inode) {
1354		ret = -EIO;
1355		goto out;
1356	}
1357
1358	while (ref_ptr < ref_end) {
1359		if (log_ref_ver) {
1360			ret = extref_get_fields(eb, ref_ptr, &namelen, &name,
1361						&ref_index, &parent_objectid);
1362			/*
1363			 * parent object can change from one array
1364			 * item to another.
1365			 */
1366			if (!dir)
1367				dir = read_one_inode(root, parent_objectid);
1368			if (!dir) {
1369				ret = -ENOENT;
1370				goto out;
1371			}
1372		} else {
1373			ret = ref_get_fields(eb, ref_ptr, &namelen, &name,
1374					     &ref_index);
1375		}
1376		if (ret)
1377			goto out;
1378
1379		/* if we already have a perfect match, we're done */
1380		if (!inode_in_dir(root, path, btrfs_ino(BTRFS_I(dir)),
1381					btrfs_ino(BTRFS_I(inode)), ref_index,
1382					name, namelen)) {
 
 
1383			/*
1384			 * look for a conflicting back reference in the
1385			 * metadata. if we find one we have to unlink that name
1386			 * of the file before we add our new link.  Later on, we
1387			 * overwrite any existing back reference, and we don't
1388			 * want to create dangling pointers in the directory.
1389			 */
1390
1391			if (!search_done) {
1392				ret = __add_inode_ref(trans, root, path, log,
1393						      BTRFS_I(dir),
1394						      BTRFS_I(inode),
1395						      inode_objectid,
1396						      parent_objectid,
1397						      ref_index, name, namelen,
1398						      &search_done);
1399				if (ret) {
1400					if (ret == 1)
1401						ret = 0;
1402					goto out;
1403				}
1404			}
1405
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1406			/* insert our name */
1407			ret = btrfs_add_link(trans, BTRFS_I(dir),
1408					BTRFS_I(inode),
1409					name, namelen, 0, ref_index);
1410			if (ret)
1411				goto out;
1412
1413			btrfs_update_inode(trans, root, inode);
 
 
1414		}
 
1415
1416		ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
1417		kfree(name);
1418		name = NULL;
1419		if (log_ref_ver) {
1420			iput(dir);
1421			dir = NULL;
1422		}
1423	}
1424
1425	/*
1426	 * Before we overwrite the inode reference item in the subvolume tree
1427	 * with the item from the log tree, we must unlink all names from the
1428	 * parent directory that are in the subvolume's tree inode reference
1429	 * item, otherwise we end up with an inconsistent subvolume tree where
1430	 * dir index entries exist for a name but there is no inode reference
1431	 * item with the same name.
1432	 */
1433	ret = unlink_old_inode_refs(trans, root, path, BTRFS_I(inode), eb, slot,
1434				    key);
1435	if (ret)
1436		goto out;
1437
1438	/* finally write the back reference in the inode */
1439	ret = overwrite_item(trans, root, path, eb, slot, key);
1440out:
1441	btrfs_release_path(path);
1442	kfree(name);
1443	iput(dir);
1444	iput(inode);
1445	return ret;
1446}
1447
1448static int insert_orphan_item(struct btrfs_trans_handle *trans,
1449			      struct btrfs_root *root, u64 ino)
1450{
1451	int ret;
1452
1453	ret = btrfs_insert_orphan_item(trans, root, ino);
1454	if (ret == -EEXIST)
1455		ret = 0;
1456
1457	return ret;
1458}
1459
1460static int count_inode_extrefs(struct btrfs_root *root,
1461		struct btrfs_inode *inode, struct btrfs_path *path)
1462{
1463	int ret = 0;
1464	int name_len;
1465	unsigned int nlink = 0;
1466	u32 item_size;
1467	u32 cur_offset = 0;
1468	u64 inode_objectid = btrfs_ino(inode);
1469	u64 offset = 0;
1470	unsigned long ptr;
1471	struct btrfs_inode_extref *extref;
1472	struct extent_buffer *leaf;
1473
1474	while (1) {
1475		ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
1476					    &extref, &offset);
1477		if (ret)
1478			break;
1479
1480		leaf = path->nodes[0];
1481		item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1482		ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1483		cur_offset = 0;
1484
1485		while (cur_offset < item_size) {
1486			extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
1487			name_len = btrfs_inode_extref_name_len(leaf, extref);
1488
1489			nlink++;
1490
1491			cur_offset += name_len + sizeof(*extref);
1492		}
1493
1494		offset++;
1495		btrfs_release_path(path);
1496	}
1497	btrfs_release_path(path);
1498
1499	if (ret < 0 && ret != -ENOENT)
1500		return ret;
1501	return nlink;
1502}
1503
1504static int count_inode_refs(struct btrfs_root *root,
1505			struct btrfs_inode *inode, struct btrfs_path *path)
1506{
1507	int ret;
1508	struct btrfs_key key;
1509	unsigned int nlink = 0;
1510	unsigned long ptr;
1511	unsigned long ptr_end;
1512	int name_len;
1513	u64 ino = btrfs_ino(inode);
1514
1515	key.objectid = ino;
1516	key.type = BTRFS_INODE_REF_KEY;
1517	key.offset = (u64)-1;
1518
1519	while (1) {
1520		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1521		if (ret < 0)
1522			break;
1523		if (ret > 0) {
1524			if (path->slots[0] == 0)
1525				break;
1526			path->slots[0]--;
1527		}
1528process_slot:
1529		btrfs_item_key_to_cpu(path->nodes[0], &key,
1530				      path->slots[0]);
1531		if (key.objectid != ino ||
1532		    key.type != BTRFS_INODE_REF_KEY)
1533			break;
1534		ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
1535		ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
1536						   path->slots[0]);
1537		while (ptr < ptr_end) {
1538			struct btrfs_inode_ref *ref;
1539
1540			ref = (struct btrfs_inode_ref *)ptr;
1541			name_len = btrfs_inode_ref_name_len(path->nodes[0],
1542							    ref);
1543			ptr = (unsigned long)(ref + 1) + name_len;
1544			nlink++;
1545		}
1546
1547		if (key.offset == 0)
1548			break;
1549		if (path->slots[0] > 0) {
1550			path->slots[0]--;
1551			goto process_slot;
1552		}
1553		key.offset--;
1554		btrfs_release_path(path);
1555	}
1556	btrfs_release_path(path);
1557
1558	return nlink;
1559}
1560
1561/*
1562 * There are a few corners where the link count of the file can't
1563 * be properly maintained during replay.  So, instead of adding
1564 * lots of complexity to the log code, we just scan the backrefs
1565 * for any file that has been through replay.
1566 *
1567 * The scan will update the link count on the inode to reflect the
1568 * number of back refs found.  If it goes down to zero, the iput
1569 * will free the inode.
1570 */
1571static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
1572					   struct btrfs_root *root,
1573					   struct inode *inode)
1574{
1575	struct btrfs_path *path;
1576	int ret;
1577	u64 nlink = 0;
1578	u64 ino = btrfs_ino(BTRFS_I(inode));
1579
1580	path = btrfs_alloc_path();
1581	if (!path)
1582		return -ENOMEM;
1583
1584	ret = count_inode_refs(root, BTRFS_I(inode), path);
1585	if (ret < 0)
1586		goto out;
1587
1588	nlink = ret;
1589
1590	ret = count_inode_extrefs(root, BTRFS_I(inode), path);
1591	if (ret < 0)
1592		goto out;
1593
1594	nlink += ret;
1595
1596	ret = 0;
1597
1598	if (nlink != inode->i_nlink) {
1599		set_nlink(inode, nlink);
1600		btrfs_update_inode(trans, root, inode);
 
 
1601	}
1602	BTRFS_I(inode)->index_cnt = (u64)-1;
1603
1604	if (inode->i_nlink == 0) {
1605		if (S_ISDIR(inode->i_mode)) {
1606			ret = replay_dir_deletes(trans, root, NULL, path,
1607						 ino, 1);
1608			if (ret)
1609				goto out;
1610		}
1611		ret = insert_orphan_item(trans, root, ino);
 
 
1612	}
1613
1614out:
1615	btrfs_free_path(path);
1616	return ret;
1617}
1618
1619static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
1620					    struct btrfs_root *root,
1621					    struct btrfs_path *path)
1622{
1623	int ret;
1624	struct btrfs_key key;
1625	struct inode *inode;
1626
1627	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1628	key.type = BTRFS_ORPHAN_ITEM_KEY;
1629	key.offset = (u64)-1;
1630	while (1) {
1631		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1632		if (ret < 0)
1633			break;
1634
1635		if (ret == 1) {
 
1636			if (path->slots[0] == 0)
1637				break;
1638			path->slots[0]--;
1639		}
1640
1641		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1642		if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
1643		    key.type != BTRFS_ORPHAN_ITEM_KEY)
1644			break;
1645
1646		ret = btrfs_del_item(trans, root, path);
1647		if (ret)
1648			goto out;
1649
1650		btrfs_release_path(path);
1651		inode = read_one_inode(root, key.offset);
1652		if (!inode)
1653			return -EIO;
 
 
1654
1655		ret = fixup_inode_link_count(trans, root, inode);
1656		iput(inode);
1657		if (ret)
1658			goto out;
1659
1660		/*
1661		 * fixup on a directory may create new entries,
1662		 * make sure we always look for the highset possible
1663		 * offset
1664		 */
1665		key.offset = (u64)-1;
1666	}
1667	ret = 0;
1668out:
1669	btrfs_release_path(path);
1670	return ret;
1671}
1672
1673
1674/*
1675 * record a given inode in the fixup dir so we can check its link
1676 * count when replay is done.  The link count is incremented here
1677 * so the inode won't go away until we check it
1678 */
1679static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
1680				      struct btrfs_root *root,
1681				      struct btrfs_path *path,
1682				      u64 objectid)
1683{
1684	struct btrfs_key key;
1685	int ret = 0;
1686	struct inode *inode;
1687
1688	inode = read_one_inode(root, objectid);
1689	if (!inode)
1690		return -EIO;
1691
1692	key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1693	key.type = BTRFS_ORPHAN_ITEM_KEY;
1694	key.offset = objectid;
1695
1696	ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
1697
1698	btrfs_release_path(path);
1699	if (ret == 0) {
1700		if (!inode->i_nlink)
1701			set_nlink(inode, 1);
1702		else
1703			inc_nlink(inode);
1704		ret = btrfs_update_inode(trans, root, inode);
1705	} else if (ret == -EEXIST) {
1706		ret = 0;
1707	} else {
1708		BUG(); /* Logic Error */
1709	}
1710	iput(inode);
1711
1712	return ret;
1713}
1714
1715/*
1716 * when replaying the log for a directory, we only insert names
1717 * for inodes that actually exist.  This means an fsync on a directory
1718 * does not implicitly fsync all the new files in it
1719 */
1720static noinline int insert_one_name(struct btrfs_trans_handle *trans,
1721				    struct btrfs_root *root,
1722				    u64 dirid, u64 index,
1723				    char *name, int name_len,
1724				    struct btrfs_key *location)
1725{
1726	struct inode *inode;
1727	struct inode *dir;
1728	int ret;
1729
1730	inode = read_one_inode(root, location->objectid);
1731	if (!inode)
1732		return -ENOENT;
1733
1734	dir = read_one_inode(root, dirid);
1735	if (!dir) {
1736		iput(inode);
1737		return -EIO;
1738	}
1739
1740	ret = btrfs_add_link(trans, BTRFS_I(dir), BTRFS_I(inode), name,
1741			name_len, 1, index);
1742
1743	/* FIXME, put inode into FIXUP list */
1744
1745	iput(inode);
1746	iput(dir);
1747	return ret;
1748}
1749
1750/*
1751 * Return true if an inode reference exists in the log for the given name,
1752 * inode and parent inode.
1753 */
1754static bool name_in_log_ref(struct btrfs_root *log_root,
1755			    const char *name, const int name_len,
1756			    const u64 dirid, const u64 ino)
1757{
1758	struct btrfs_key search_key;
1759
1760	search_key.objectid = ino;
1761	search_key.type = BTRFS_INODE_REF_KEY;
1762	search_key.offset = dirid;
1763	if (backref_in_log(log_root, &search_key, dirid, name, name_len))
1764		return true;
1765
1766	search_key.type = BTRFS_INODE_EXTREF_KEY;
1767	search_key.offset = btrfs_extref_hash(dirid, name, name_len);
1768	if (backref_in_log(log_root, &search_key, dirid, name, name_len))
1769		return true;
1770
1771	return false;
1772}
1773
1774/*
1775 * take a single entry in a log directory item and replay it into
1776 * the subvolume.
1777 *
1778 * if a conflicting item exists in the subdirectory already,
1779 * the inode it points to is unlinked and put into the link count
1780 * fix up tree.
1781 *
1782 * If a name from the log points to a file or directory that does
1783 * not exist in the FS, it is skipped.  fsyncs on directories
1784 * do not force down inodes inside that directory, just changes to the
1785 * names or unlinks in a directory.
1786 *
1787 * Returns < 0 on error, 0 if the name wasn't replayed (dentry points to a
1788 * non-existing inode) and 1 if the name was replayed.
1789 */
1790static noinline int replay_one_name(struct btrfs_trans_handle *trans,
1791				    struct btrfs_root *root,
1792				    struct btrfs_path *path,
1793				    struct extent_buffer *eb,
1794				    struct btrfs_dir_item *di,
1795				    struct btrfs_key *key)
1796{
1797	char *name;
1798	int name_len;
1799	struct btrfs_dir_item *dst_di;
1800	struct btrfs_key found_key;
1801	struct btrfs_key log_key;
1802	struct inode *dir;
1803	u8 log_type;
1804	int exists;
1805	int ret = 0;
1806	bool update_size = (key->type == BTRFS_DIR_INDEX_KEY);
1807	bool name_added = false;
1808
1809	dir = read_one_inode(root, key->objectid);
1810	if (!dir)
1811		return -EIO;
1812
1813	name_len = btrfs_dir_name_len(eb, di);
1814	name = kmalloc(name_len, GFP_NOFS);
1815	if (!name) {
1816		ret = -ENOMEM;
1817		goto out;
1818	}
1819
1820	log_type = btrfs_dir_type(eb, di);
1821	read_extent_buffer(eb, name, (unsigned long)(di + 1),
1822		   name_len);
1823
1824	btrfs_dir_item_key_to_cpu(eb, di, &log_key);
1825	exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
1826	if (exists == 0)
1827		exists = 1;
1828	else
1829		exists = 0;
1830	btrfs_release_path(path);
 
 
 
 
1831
1832	if (key->type == BTRFS_DIR_ITEM_KEY) {
1833		dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
1834				       name, name_len, 1);
1835	} else if (key->type == BTRFS_DIR_INDEX_KEY) {
1836		dst_di = btrfs_lookup_dir_index_item(trans, root, path,
1837						     key->objectid,
1838						     key->offset, name,
1839						     name_len, 1);
1840	} else {
1841		/* Corruption */
1842		ret = -EINVAL;
1843		goto out;
1844	}
1845	if (IS_ERR_OR_NULL(dst_di)) {
 
 
 
 
 
 
 
1846		/* we need a sequence number to insert, so we only
1847		 * do inserts for the BTRFS_DIR_INDEX_KEY types
1848		 */
1849		if (key->type != BTRFS_DIR_INDEX_KEY)
1850			goto out;
1851		goto insert;
1852	}
1853
1854	btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
1855	/* the existing item matches the logged item */
1856	if (found_key.objectid == log_key.objectid &&
1857	    found_key.type == log_key.type &&
1858	    found_key.offset == log_key.offset &&
1859	    btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
1860		update_size = false;
1861		goto out;
1862	}
1863
1864	/*
1865	 * don't drop the conflicting directory entry if the inode
1866	 * for the new entry doesn't exist
1867	 */
1868	if (!exists)
1869		goto out;
1870
1871	ret = drop_one_dir_item(trans, root, path, BTRFS_I(dir), dst_di);
1872	if (ret)
1873		goto out;
1874
1875	if (key->type == BTRFS_DIR_INDEX_KEY)
1876		goto insert;
1877out:
1878	btrfs_release_path(path);
1879	if (!ret && update_size) {
1880		btrfs_i_size_write(BTRFS_I(dir), dir->i_size + name_len * 2);
1881		ret = btrfs_update_inode(trans, root, dir);
1882	}
1883	kfree(name);
1884	iput(dir);
1885	if (!ret && name_added)
1886		ret = 1;
1887	return ret;
1888
1889insert:
1890	if (name_in_log_ref(root->log_root, name, name_len,
1891			    key->objectid, log_key.objectid)) {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1892		/* The dentry will be added later. */
1893		ret = 0;
1894		update_size = false;
1895		goto out;
1896	}
1897	btrfs_release_path(path);
1898	ret = insert_one_name(trans, root, key->objectid, key->offset,
1899			      name, name_len, &log_key);
1900	if (ret && ret != -ENOENT && ret != -EEXIST)
1901		goto out;
1902	if (!ret)
1903		name_added = true;
1904	update_size = false;
1905	ret = 0;
1906	goto out;
1907}
1908
1909/*
1910 * find all the names in a directory item and reconcile them into
1911 * the subvolume.  Only BTRFS_DIR_ITEM_KEY types will have more than
1912 * one name in a directory item, but the same code gets used for
1913 * both directory index types
1914 */
1915static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
1916					struct btrfs_root *root,
1917					struct btrfs_path *path,
1918					struct extent_buffer *eb, int slot,
1919					struct btrfs_key *key)
1920{
1921	int ret = 0;
1922	u32 item_size = btrfs_item_size_nr(eb, slot);
1923	struct btrfs_dir_item *di;
1924	int name_len;
1925	unsigned long ptr;
1926	unsigned long ptr_end;
1927	struct btrfs_path *fixup_path = NULL;
1928
1929	ptr = btrfs_item_ptr_offset(eb, slot);
1930	ptr_end = ptr + item_size;
1931	while (ptr < ptr_end) {
1932		di = (struct btrfs_dir_item *)ptr;
1933		name_len = btrfs_dir_name_len(eb, di);
1934		ret = replay_one_name(trans, root, path, eb, di, key);
1935		if (ret < 0)
1936			break;
1937		ptr = (unsigned long)(di + 1);
1938		ptr += name_len;
1939
1940		/*
1941		 * If this entry refers to a non-directory (directories can not
1942		 * have a link count > 1) and it was added in the transaction
1943		 * that was not committed, make sure we fixup the link count of
1944		 * the inode it the entry points to. Otherwise something like
1945		 * the following would result in a directory pointing to an
1946		 * inode with a wrong link that does not account for this dir
1947		 * entry:
1948		 *
1949		 * mkdir testdir
1950		 * touch testdir/foo
1951		 * touch testdir/bar
1952		 * sync
1953		 *
1954		 * ln testdir/bar testdir/bar_link
1955		 * ln testdir/foo testdir/foo_link
1956		 * xfs_io -c "fsync" testdir/bar
1957		 *
1958		 * <power failure>
1959		 *
1960		 * mount fs, log replay happens
1961		 *
1962		 * File foo would remain with a link count of 1 when it has two
1963		 * entries pointing to it in the directory testdir. This would
1964		 * make it impossible to ever delete the parent directory has
1965		 * it would result in stale dentries that can never be deleted.
1966		 */
1967		if (ret == 1 && btrfs_dir_type(eb, di) != BTRFS_FT_DIR) {
1968			struct btrfs_key di_key;
1969
1970			if (!fixup_path) {
1971				fixup_path = btrfs_alloc_path();
1972				if (!fixup_path) {
1973					ret = -ENOMEM;
1974					break;
1975				}
1976			}
1977
1978			btrfs_dir_item_key_to_cpu(eb, di, &di_key);
1979			ret = link_to_fixup_dir(trans, root, fixup_path,
1980						di_key.objectid);
1981			if (ret)
1982				break;
1983		}
1984		ret = 0;
1985	}
1986	btrfs_free_path(fixup_path);
1987	return ret;
1988}
1989
1990/*
1991 * directory replay has two parts.  There are the standard directory
1992 * items in the log copied from the subvolume, and range items
1993 * created in the log while the subvolume was logged.
1994 *
1995 * The range items tell us which parts of the key space the log
1996 * is authoritative for.  During replay, if a key in the subvolume
1997 * directory is in a logged range item, but not actually in the log
1998 * that means it was deleted from the directory before the fsync
1999 * and should be removed.
2000 */
2001static noinline int find_dir_range(struct btrfs_root *root,
2002				   struct btrfs_path *path,
2003				   u64 dirid, int key_type,
2004				   u64 *start_ret, u64 *end_ret)
2005{
2006	struct btrfs_key key;
2007	u64 found_end;
2008	struct btrfs_dir_log_item *item;
2009	int ret;
2010	int nritems;
2011
2012	if (*start_ret == (u64)-1)
2013		return 1;
2014
2015	key.objectid = dirid;
2016	key.type = key_type;
2017	key.offset = *start_ret;
2018
2019	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
2020	if (ret < 0)
2021		goto out;
2022	if (ret > 0) {
2023		if (path->slots[0] == 0)
2024			goto out;
2025		path->slots[0]--;
2026	}
2027	if (ret != 0)
2028		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
2029
2030	if (key.type != key_type || key.objectid != dirid) {
2031		ret = 1;
2032		goto next;
2033	}
2034	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
2035			      struct btrfs_dir_log_item);
2036	found_end = btrfs_dir_log_end(path->nodes[0], item);
2037
2038	if (*start_ret >= key.offset && *start_ret <= found_end) {
2039		ret = 0;
2040		*start_ret = key.offset;
2041		*end_ret = found_end;
2042		goto out;
2043	}
2044	ret = 1;
2045next:
2046	/* check the next slot in the tree to see if it is a valid item */
2047	nritems = btrfs_header_nritems(path->nodes[0]);
2048	path->slots[0]++;
2049	if (path->slots[0] >= nritems) {
2050		ret = btrfs_next_leaf(root, path);
2051		if (ret)
2052			goto out;
2053	}
2054
2055	btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
2056
2057	if (key.type != key_type || key.objectid != dirid) {
2058		ret = 1;
2059		goto out;
2060	}
2061	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
2062			      struct btrfs_dir_log_item);
2063	found_end = btrfs_dir_log_end(path->nodes[0], item);
2064	*start_ret = key.offset;
2065	*end_ret = found_end;
2066	ret = 0;
2067out:
2068	btrfs_release_path(path);
2069	return ret;
2070}
2071
2072/*
2073 * this looks for a given directory item in the log.  If the directory
2074 * item is not in the log, the item is removed and the inode it points
2075 * to is unlinked
2076 */
2077static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
2078				      struct btrfs_root *root,
2079				      struct btrfs_root *log,
2080				      struct btrfs_path *path,
2081				      struct btrfs_path *log_path,
2082				      struct inode *dir,
2083				      struct btrfs_key *dir_key)
2084{
2085	int ret;
2086	struct extent_buffer *eb;
2087	int slot;
2088	u32 item_size;
2089	struct btrfs_dir_item *di;
2090	struct btrfs_dir_item *log_di;
2091	int name_len;
2092	unsigned long ptr;
2093	unsigned long ptr_end;
2094	char *name;
2095	struct inode *inode;
2096	struct btrfs_key location;
2097
2098again:
2099	eb = path->nodes[0];
2100	slot = path->slots[0];
2101	item_size = btrfs_item_size_nr(eb, slot);
2102	ptr = btrfs_item_ptr_offset(eb, slot);
2103	ptr_end = ptr + item_size;
2104	while (ptr < ptr_end) {
2105		di = (struct btrfs_dir_item *)ptr;
2106		name_len = btrfs_dir_name_len(eb, di);
2107		name = kmalloc(name_len, GFP_NOFS);
2108		if (!name) {
2109			ret = -ENOMEM;
2110			goto out;
2111		}
2112		read_extent_buffer(eb, name, (unsigned long)(di + 1),
2113				  name_len);
2114		log_di = NULL;
2115		if (log && dir_key->type == BTRFS_DIR_ITEM_KEY) {
2116			log_di = btrfs_lookup_dir_item(trans, log, log_path,
2117						       dir_key->objectid,
2118						       name, name_len, 0);
2119		} else if (log && dir_key->type == BTRFS_DIR_INDEX_KEY) {
2120			log_di = btrfs_lookup_dir_index_item(trans, log,
2121						     log_path,
2122						     dir_key->objectid,
2123						     dir_key->offset,
2124						     name, name_len, 0);
2125		}
2126		if (!log_di || (IS_ERR(log_di) && PTR_ERR(log_di) == -ENOENT)) {
2127			btrfs_dir_item_key_to_cpu(eb, di, &location);
2128			btrfs_release_path(path);
2129			btrfs_release_path(log_path);
2130			inode = read_one_inode(root, location.objectid);
2131			if (!inode) {
2132				kfree(name);
2133				return -EIO;
2134			}
2135
2136			ret = link_to_fixup_dir(trans, root,
2137						path, location.objectid);
2138			if (ret) {
2139				kfree(name);
2140				iput(inode);
2141				goto out;
2142			}
2143
2144			inc_nlink(inode);
2145			ret = btrfs_unlink_inode(trans, root, BTRFS_I(dir),
2146					BTRFS_I(inode), name, name_len);
2147			if (!ret)
2148				ret = btrfs_run_delayed_items(trans);
2149			kfree(name);
2150			iput(inode);
2151			if (ret)
2152				goto out;
2153
2154			/* there might still be more names under this key
2155			 * check and repeat if required
2156			 */
2157			ret = btrfs_search_slot(NULL, root, dir_key, path,
2158						0, 0);
2159			if (ret == 0)
2160				goto again;
2161			ret = 0;
2162			goto out;
2163		} else if (IS_ERR(log_di)) {
2164			kfree(name);
2165			return PTR_ERR(log_di);
2166		}
2167		btrfs_release_path(log_path);
2168		kfree(name);
2169
2170		ptr = (unsigned long)(di + 1);
2171		ptr += name_len;
2172	}
2173	ret = 0;
2174out:
2175	btrfs_release_path(path);
2176	btrfs_release_path(log_path);
2177	return ret;
2178}
2179
2180static int replay_xattr_deletes(struct btrfs_trans_handle *trans,
2181			      struct btrfs_root *root,
2182			      struct btrfs_root *log,
2183			      struct btrfs_path *path,
2184			      const u64 ino)
2185{
2186	struct btrfs_key search_key;
2187	struct btrfs_path *log_path;
2188	int i;
2189	int nritems;
2190	int ret;
2191
2192	log_path = btrfs_alloc_path();
2193	if (!log_path)
2194		return -ENOMEM;
2195
2196	search_key.objectid = ino;
2197	search_key.type = BTRFS_XATTR_ITEM_KEY;
2198	search_key.offset = 0;
2199again:
2200	ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
2201	if (ret < 0)
2202		goto out;
2203process_leaf:
2204	nritems = btrfs_header_nritems(path->nodes[0]);
2205	for (i = path->slots[0]; i < nritems; i++) {
2206		struct btrfs_key key;
2207		struct btrfs_dir_item *di;
2208		struct btrfs_dir_item *log_di;
2209		u32 total_size;
2210		u32 cur;
2211
2212		btrfs_item_key_to_cpu(path->nodes[0], &key, i);
2213		if (key.objectid != ino || key.type != BTRFS_XATTR_ITEM_KEY) {
2214			ret = 0;
2215			goto out;
2216		}
2217
2218		di = btrfs_item_ptr(path->nodes[0], i, struct btrfs_dir_item);
2219		total_size = btrfs_item_size_nr(path->nodes[0], i);
2220		cur = 0;
2221		while (cur < total_size) {
2222			u16 name_len = btrfs_dir_name_len(path->nodes[0], di);
2223			u16 data_len = btrfs_dir_data_len(path->nodes[0], di);
2224			u32 this_len = sizeof(*di) + name_len + data_len;
2225			char *name;
2226
2227			name = kmalloc(name_len, GFP_NOFS);
2228			if (!name) {
2229				ret = -ENOMEM;
2230				goto out;
2231			}
2232			read_extent_buffer(path->nodes[0], name,
2233					   (unsigned long)(di + 1), name_len);
2234
2235			log_di = btrfs_lookup_xattr(NULL, log, log_path, ino,
2236						    name, name_len, 0);
2237			btrfs_release_path(log_path);
2238			if (!log_di) {
2239				/* Doesn't exist in log tree, so delete it. */
2240				btrfs_release_path(path);
2241				di = btrfs_lookup_xattr(trans, root, path, ino,
2242							name, name_len, -1);
2243				kfree(name);
2244				if (IS_ERR(di)) {
2245					ret = PTR_ERR(di);
2246					goto out;
2247				}
2248				ASSERT(di);
2249				ret = btrfs_delete_one_dir_name(trans, root,
2250								path, di);
2251				if (ret)
2252					goto out;
2253				btrfs_release_path(path);
2254				search_key = key;
2255				goto again;
2256			}
2257			kfree(name);
2258			if (IS_ERR(log_di)) {
2259				ret = PTR_ERR(log_di);
2260				goto out;
2261			}
2262			cur += this_len;
2263			di = (struct btrfs_dir_item *)((char *)di + this_len);
2264		}
2265	}
2266	ret = btrfs_next_leaf(root, path);
2267	if (ret > 0)
2268		ret = 0;
2269	else if (ret == 0)
2270		goto process_leaf;
2271out:
2272	btrfs_free_path(log_path);
2273	btrfs_release_path(path);
2274	return ret;
2275}
2276
2277
2278/*
2279 * deletion replay happens before we copy any new directory items
2280 * out of the log or out of backreferences from inodes.  It
2281 * scans the log to find ranges of keys that log is authoritative for,
2282 * and then scans the directory to find items in those ranges that are
2283 * not present in the log.
2284 *
2285 * Anything we don't find in the log is unlinked and removed from the
2286 * directory.
2287 */
2288static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
2289				       struct btrfs_root *root,
2290				       struct btrfs_root *log,
2291				       struct btrfs_path *path,
2292				       u64 dirid, int del_all)
2293{
2294	u64 range_start;
2295	u64 range_end;
2296	int key_type = BTRFS_DIR_LOG_ITEM_KEY;
2297	int ret = 0;
2298	struct btrfs_key dir_key;
2299	struct btrfs_key found_key;
2300	struct btrfs_path *log_path;
2301	struct inode *dir;
2302
2303	dir_key.objectid = dirid;
2304	dir_key.type = BTRFS_DIR_ITEM_KEY;
2305	log_path = btrfs_alloc_path();
2306	if (!log_path)
2307		return -ENOMEM;
2308
2309	dir = read_one_inode(root, dirid);
2310	/* it isn't an error if the inode isn't there, that can happen
2311	 * because we replay the deletes before we copy in the inode item
2312	 * from the log
2313	 */
2314	if (!dir) {
2315		btrfs_free_path(log_path);
2316		return 0;
2317	}
2318again:
2319	range_start = 0;
2320	range_end = 0;
2321	while (1) {
2322		if (del_all)
2323			range_end = (u64)-1;
2324		else {
2325			ret = find_dir_range(log, path, dirid, key_type,
2326					     &range_start, &range_end);
2327			if (ret != 0)
2328				break;
2329		}
2330
2331		dir_key.offset = range_start;
2332		while (1) {
2333			int nritems;
2334			ret = btrfs_search_slot(NULL, root, &dir_key, path,
2335						0, 0);
2336			if (ret < 0)
2337				goto out;
2338
2339			nritems = btrfs_header_nritems(path->nodes[0]);
2340			if (path->slots[0] >= nritems) {
2341				ret = btrfs_next_leaf(root, path);
2342				if (ret == 1)
2343					break;
2344				else if (ret < 0)
2345					goto out;
2346			}
2347			btrfs_item_key_to_cpu(path->nodes[0], &found_key,
2348					      path->slots[0]);
2349			if (found_key.objectid != dirid ||
2350			    found_key.type != dir_key.type)
2351				goto next_type;
2352
2353			if (found_key.offset > range_end)
2354				break;
2355
2356			ret = check_item_in_log(trans, root, log, path,
2357						log_path, dir,
2358						&found_key);
2359			if (ret)
2360				goto out;
2361			if (found_key.offset == (u64)-1)
2362				break;
2363			dir_key.offset = found_key.offset + 1;
2364		}
2365		btrfs_release_path(path);
2366		if (range_end == (u64)-1)
2367			break;
2368		range_start = range_end + 1;
2369	}
2370
2371next_type:
2372	ret = 0;
2373	if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
2374		key_type = BTRFS_DIR_LOG_INDEX_KEY;
2375		dir_key.type = BTRFS_DIR_INDEX_KEY;
2376		btrfs_release_path(path);
2377		goto again;
2378	}
2379out:
2380	btrfs_release_path(path);
2381	btrfs_free_path(log_path);
2382	iput(dir);
2383	return ret;
2384}
2385
2386/*
2387 * the process_func used to replay items from the log tree.  This
2388 * gets called in two different stages.  The first stage just looks
2389 * for inodes and makes sure they are all copied into the subvolume.
2390 *
2391 * The second stage copies all the other item types from the log into
2392 * the subvolume.  The two stage approach is slower, but gets rid of
2393 * lots of complexity around inodes referencing other inodes that exist
2394 * only in the log (references come from either directory items or inode
2395 * back refs).
2396 */
2397static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
2398			     struct walk_control *wc, u64 gen, int level)
2399{
2400	int nritems;
2401	struct btrfs_path *path;
2402	struct btrfs_root *root = wc->replay_dest;
2403	struct btrfs_key key;
2404	int i;
2405	int ret;
2406
2407	ret = btrfs_read_buffer(eb, gen, level, NULL);
2408	if (ret)
2409		return ret;
2410
2411	level = btrfs_header_level(eb);
2412
2413	if (level != 0)
2414		return 0;
2415
2416	path = btrfs_alloc_path();
2417	if (!path)
2418		return -ENOMEM;
2419
2420	nritems = btrfs_header_nritems(eb);
2421	for (i = 0; i < nritems; i++) {
2422		btrfs_item_key_to_cpu(eb, &key, i);
2423
2424		/* inode keys are done during the first stage */
2425		if (key.type == BTRFS_INODE_ITEM_KEY &&
2426		    wc->stage == LOG_WALK_REPLAY_INODES) {
2427			struct btrfs_inode_item *inode_item;
2428			u32 mode;
2429
2430			inode_item = btrfs_item_ptr(eb, i,
2431					    struct btrfs_inode_item);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2432			ret = replay_xattr_deletes(wc->trans, root, log,
2433						   path, key.objectid);
2434			if (ret)
2435				break;
2436			mode = btrfs_inode_mode(eb, inode_item);
2437			if (S_ISDIR(mode)) {
2438				ret = replay_dir_deletes(wc->trans,
2439					 root, log, path, key.objectid, 0);
2440				if (ret)
2441					break;
2442			}
2443			ret = overwrite_item(wc->trans, root, path,
2444					     eb, i, &key);
2445			if (ret)
2446				break;
2447
2448			/*
2449			 * Before replaying extents, truncate the inode to its
2450			 * size. We need to do it now and not after log replay
2451			 * because before an fsync we can have prealloc extents
2452			 * added beyond the inode's i_size. If we did it after,
2453			 * through orphan cleanup for example, we would drop
2454			 * those prealloc extents just after replaying them.
2455			 */
2456			if (S_ISREG(mode)) {
 
2457				struct inode *inode;
2458				u64 from;
2459
2460				inode = read_one_inode(root, key.objectid);
2461				if (!inode) {
2462					ret = -EIO;
2463					break;
2464				}
2465				from = ALIGN(i_size_read(inode),
2466					     root->fs_info->sectorsize);
2467				ret = btrfs_drop_extents(wc->trans, root, inode,
2468							 from, (u64)-1, 1);
2469				/*
2470				 * If the nlink count is zero here, the iput
2471				 * will free the inode.  We bump it to make
2472				 * sure it doesn't get freed until the link
2473				 * count fixup is done.
2474				 */
2475				if (!ret) {
2476					if (inode->i_nlink == 0)
2477						inc_nlink(inode);
2478					/* Update link count and nbytes. */
2479					ret = btrfs_update_inode(wc->trans,
2480								 root, inode);
2481				}
2482				iput(inode);
2483				if (ret)
2484					break;
2485			}
2486
2487			ret = link_to_fixup_dir(wc->trans, root,
2488						path, key.objectid);
2489			if (ret)
2490				break;
2491		}
2492
 
 
 
2493		if (key.type == BTRFS_DIR_INDEX_KEY &&
2494		    wc->stage == LOG_WALK_REPLAY_DIR_INDEX) {
2495			ret = replay_one_dir_item(wc->trans, root, path,
2496						  eb, i, &key);
2497			if (ret)
2498				break;
2499		}
2500
2501		if (wc->stage < LOG_WALK_REPLAY_ALL)
2502			continue;
2503
2504		/* these keys are simply copied */
2505		if (key.type == BTRFS_XATTR_ITEM_KEY) {
2506			ret = overwrite_item(wc->trans, root, path,
2507					     eb, i, &key);
2508			if (ret)
2509				break;
2510		} else if (key.type == BTRFS_INODE_REF_KEY ||
2511			   key.type == BTRFS_INODE_EXTREF_KEY) {
2512			ret = add_inode_ref(wc->trans, root, log, path,
2513					    eb, i, &key);
2514			if (ret && ret != -ENOENT)
2515				break;
2516			ret = 0;
2517		} else if (key.type == BTRFS_EXTENT_DATA_KEY) {
2518			ret = replay_one_extent(wc->trans, root, path,
2519						eb, i, &key);
2520			if (ret)
2521				break;
2522		} else if (key.type == BTRFS_DIR_ITEM_KEY) {
2523			ret = replay_one_dir_item(wc->trans, root, path,
2524						  eb, i, &key);
2525			if (ret)
2526				break;
2527		}
2528	}
2529	btrfs_free_path(path);
2530	return ret;
2531}
2532
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2533static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
2534				   struct btrfs_root *root,
2535				   struct btrfs_path *path, int *level,
2536				   struct walk_control *wc)
2537{
2538	struct btrfs_fs_info *fs_info = root->fs_info;
2539	u64 root_owner;
2540	u64 bytenr;
2541	u64 ptr_gen;
2542	struct extent_buffer *next;
2543	struct extent_buffer *cur;
2544	struct extent_buffer *parent;
2545	u32 blocksize;
2546	int ret = 0;
2547
2548	WARN_ON(*level < 0);
2549	WARN_ON(*level >= BTRFS_MAX_LEVEL);
2550
2551	while (*level > 0) {
2552		struct btrfs_key first_key;
2553
2554		WARN_ON(*level < 0);
2555		WARN_ON(*level >= BTRFS_MAX_LEVEL);
2556		cur = path->nodes[*level];
2557
2558		WARN_ON(btrfs_header_level(cur) != *level);
2559
2560		if (path->slots[*level] >=
2561		    btrfs_header_nritems(cur))
2562			break;
2563
2564		bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
2565		ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
2566		btrfs_node_key_to_cpu(cur, &first_key, path->slots[*level]);
2567		blocksize = fs_info->nodesize;
2568
2569		parent = path->nodes[*level];
2570		root_owner = btrfs_header_owner(parent);
2571
2572		next = btrfs_find_create_tree_block(fs_info, bytenr);
2573		if (IS_ERR(next))
2574			return PTR_ERR(next);
2575
2576		if (*level == 1) {
2577			ret = wc->process_func(root, next, wc, ptr_gen,
2578					       *level - 1);
2579			if (ret) {
2580				free_extent_buffer(next);
2581				return ret;
2582			}
2583
2584			path->slots[*level]++;
2585			if (wc->free) {
2586				ret = btrfs_read_buffer(next, ptr_gen,
2587							*level - 1, &first_key);
2588				if (ret) {
2589					free_extent_buffer(next);
2590					return ret;
2591				}
2592
2593				if (trans) {
2594					btrfs_tree_lock(next);
2595					btrfs_set_lock_blocking(next);
2596					clean_tree_block(fs_info, next);
2597					btrfs_wait_tree_block_writeback(next);
2598					btrfs_tree_unlock(next);
 
 
 
 
 
 
 
 
2599				} else {
2600					if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
2601						clear_extent_buffer_dirty(next);
2602				}
2603
2604				WARN_ON(root_owner !=
2605					BTRFS_TREE_LOG_OBJECTID);
2606				ret = btrfs_free_and_pin_reserved_extent(
2607							fs_info, bytenr,
2608							blocksize);
2609				if (ret) {
2610					free_extent_buffer(next);
2611					return ret;
2612				}
2613			}
2614			free_extent_buffer(next);
2615			continue;
2616		}
2617		ret = btrfs_read_buffer(next, ptr_gen, *level - 1, &first_key);
2618		if (ret) {
2619			free_extent_buffer(next);
2620			return ret;
2621		}
2622
2623		WARN_ON(*level <= 0);
2624		if (path->nodes[*level-1])
2625			free_extent_buffer(path->nodes[*level-1]);
2626		path->nodes[*level-1] = next;
2627		*level = btrfs_header_level(next);
2628		path->slots[*level] = 0;
2629		cond_resched();
2630	}
2631	WARN_ON(*level < 0);
2632	WARN_ON(*level >= BTRFS_MAX_LEVEL);
2633
2634	path->slots[*level] = btrfs_header_nritems(path->nodes[*level]);
2635
2636	cond_resched();
2637	return 0;
2638}
2639
2640static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
2641				 struct btrfs_root *root,
2642				 struct btrfs_path *path, int *level,
2643				 struct walk_control *wc)
2644{
2645	struct btrfs_fs_info *fs_info = root->fs_info;
2646	u64 root_owner;
2647	int i;
2648	int slot;
2649	int ret;
2650
2651	for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
2652		slot = path->slots[i];
2653		if (slot + 1 < btrfs_header_nritems(path->nodes[i])) {
2654			path->slots[i]++;
2655			*level = i;
2656			WARN_ON(*level == 0);
2657			return 0;
2658		} else {
2659			struct extent_buffer *parent;
2660			if (path->nodes[*level] == root->node)
2661				parent = path->nodes[*level];
2662			else
2663				parent = path->nodes[*level + 1];
2664
2665			root_owner = btrfs_header_owner(parent);
2666			ret = wc->process_func(root, path->nodes[*level], wc,
2667				 btrfs_header_generation(path->nodes[*level]),
2668				 *level);
2669			if (ret)
2670				return ret;
2671
2672			if (wc->free) {
2673				struct extent_buffer *next;
2674
2675				next = path->nodes[*level];
2676
2677				if (trans) {
2678					btrfs_tree_lock(next);
2679					btrfs_set_lock_blocking(next);
2680					clean_tree_block(fs_info, next);
2681					btrfs_wait_tree_block_writeback(next);
2682					btrfs_tree_unlock(next);
 
 
 
 
 
2683				} else {
2684					if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
2685						clear_extent_buffer_dirty(next);
2686				}
2687
2688				WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
2689				ret = btrfs_free_and_pin_reserved_extent(
2690						fs_info,
2691						path->nodes[*level]->start,
2692						path->nodes[*level]->len);
2693				if (ret)
2694					return ret;
2695			}
2696			free_extent_buffer(path->nodes[*level]);
2697			path->nodes[*level] = NULL;
2698			*level = i + 1;
2699		}
2700	}
2701	return 1;
2702}
2703
2704/*
2705 * drop the reference count on the tree rooted at 'snap'.  This traverses
2706 * the tree freeing any blocks that have a ref count of zero after being
2707 * decremented.
2708 */
2709static int walk_log_tree(struct btrfs_trans_handle *trans,
2710			 struct btrfs_root *log, struct walk_control *wc)
2711{
2712	struct btrfs_fs_info *fs_info = log->fs_info;
2713	int ret = 0;
2714	int wret;
2715	int level;
2716	struct btrfs_path *path;
2717	int orig_level;
2718
2719	path = btrfs_alloc_path();
2720	if (!path)
2721		return -ENOMEM;
2722
2723	level = btrfs_header_level(log->node);
2724	orig_level = level;
2725	path->nodes[level] = log->node;
2726	extent_buffer_get(log->node);
2727	path->slots[level] = 0;
2728
2729	while (1) {
2730		wret = walk_down_log_tree(trans, log, path, &level, wc);
2731		if (wret > 0)
2732			break;
2733		if (wret < 0) {
2734			ret = wret;
2735			goto out;
2736		}
2737
2738		wret = walk_up_log_tree(trans, log, path, &level, wc);
2739		if (wret > 0)
2740			break;
2741		if (wret < 0) {
2742			ret = wret;
2743			goto out;
2744		}
2745	}
2746
2747	/* was the root node processed? if not, catch it here */
2748	if (path->nodes[orig_level]) {
2749		ret = wc->process_func(log, path->nodes[orig_level], wc,
2750			 btrfs_header_generation(path->nodes[orig_level]),
2751			 orig_level);
2752		if (ret)
2753			goto out;
2754		if (wc->free) {
2755			struct extent_buffer *next;
2756
2757			next = path->nodes[orig_level];
2758
2759			if (trans) {
2760				btrfs_tree_lock(next);
2761				btrfs_set_lock_blocking(next);
2762				clean_tree_block(fs_info, next);
2763				btrfs_wait_tree_block_writeback(next);
2764				btrfs_tree_unlock(next);
 
 
 
 
2765			} else {
2766				if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &next->bflags))
2767					clear_extent_buffer_dirty(next);
 
2768			}
2769
2770			WARN_ON(log->root_key.objectid !=
2771				BTRFS_TREE_LOG_OBJECTID);
2772			ret = btrfs_free_and_pin_reserved_extent(fs_info,
2773							next->start, next->len);
2774			if (ret)
2775				goto out;
2776		}
2777	}
2778
2779out:
2780	btrfs_free_path(path);
2781	return ret;
2782}
2783
2784/*
2785 * helper function to update the item for a given subvolumes log root
2786 * in the tree of log roots
2787 */
2788static int update_log_root(struct btrfs_trans_handle *trans,
2789			   struct btrfs_root *log)
 
2790{
2791	struct btrfs_fs_info *fs_info = log->fs_info;
2792	int ret;
2793
2794	if (log->log_transid == 1) {
2795		/* insert root item on the first sync */
2796		ret = btrfs_insert_root(trans, fs_info->log_root_tree,
2797				&log->root_key, &log->root_item);
2798	} else {
2799		ret = btrfs_update_root(trans, fs_info->log_root_tree,
2800				&log->root_key, &log->root_item);
2801	}
2802	return ret;
2803}
2804
2805static void wait_log_commit(struct btrfs_root *root, int transid)
2806{
2807	DEFINE_WAIT(wait);
2808	int index = transid % 2;
2809
2810	/*
2811	 * we only allow two pending log transactions at a time,
2812	 * so we know that if ours is more than 2 older than the
2813	 * current transaction, we're done
2814	 */
2815	for (;;) {
2816		prepare_to_wait(&root->log_commit_wait[index],
2817				&wait, TASK_UNINTERRUPTIBLE);
2818
2819		if (!(root->log_transid_committed < transid &&
2820		      atomic_read(&root->log_commit[index])))
2821			break;
2822
2823		mutex_unlock(&root->log_mutex);
2824		schedule();
2825		mutex_lock(&root->log_mutex);
2826	}
2827	finish_wait(&root->log_commit_wait[index], &wait);
2828}
2829
2830static void wait_for_writer(struct btrfs_root *root)
2831{
2832	DEFINE_WAIT(wait);
2833
2834	for (;;) {
2835		prepare_to_wait(&root->log_writer_wait, &wait,
2836				TASK_UNINTERRUPTIBLE);
2837		if (!atomic_read(&root->log_writers))
2838			break;
2839
2840		mutex_unlock(&root->log_mutex);
2841		schedule();
2842		mutex_lock(&root->log_mutex);
2843	}
2844	finish_wait(&root->log_writer_wait, &wait);
2845}
2846
2847static inline void btrfs_remove_log_ctx(struct btrfs_root *root,
2848					struct btrfs_log_ctx *ctx)
2849{
2850	if (!ctx)
2851		return;
2852
2853	mutex_lock(&root->log_mutex);
2854	list_del_init(&ctx->list);
2855	mutex_unlock(&root->log_mutex);
2856}
2857
2858/* 
2859 * Invoked in log mutex context, or be sure there is no other task which
2860 * can access the list.
2861 */
2862static inline void btrfs_remove_all_log_ctxs(struct btrfs_root *root,
2863					     int index, int error)
2864{
2865	struct btrfs_log_ctx *ctx;
2866	struct btrfs_log_ctx *safe;
2867
2868	list_for_each_entry_safe(ctx, safe, &root->log_ctxs[index], list) {
2869		list_del_init(&ctx->list);
2870		ctx->log_ret = error;
2871	}
2872
2873	INIT_LIST_HEAD(&root->log_ctxs[index]);
2874}
2875
2876/*
2877 * btrfs_sync_log does sends a given tree log down to the disk and
2878 * updates the super blocks to record it.  When this call is done,
2879 * you know that any inodes previously logged are safely on disk only
2880 * if it returns 0.
2881 *
2882 * Any other return value means you need to call btrfs_commit_transaction.
2883 * Some of the edge cases for fsyncing directories that have had unlinks
2884 * or renames done in the past mean that sometimes the only safe
2885 * fsync is to commit the whole FS.  When btrfs_sync_log returns -EAGAIN,
2886 * that has happened.
2887 */
2888int btrfs_sync_log(struct btrfs_trans_handle *trans,
2889		   struct btrfs_root *root, struct btrfs_log_ctx *ctx)
2890{
2891	int index1;
2892	int index2;
2893	int mark;
2894	int ret;
2895	struct btrfs_fs_info *fs_info = root->fs_info;
2896	struct btrfs_root *log = root->log_root;
2897	struct btrfs_root *log_root_tree = fs_info->log_root_tree;
 
2898	int log_transid = 0;
2899	struct btrfs_log_ctx root_log_ctx;
2900	struct blk_plug plug;
 
 
2901
2902	mutex_lock(&root->log_mutex);
2903	log_transid = ctx->log_transid;
2904	if (root->log_transid_committed >= log_transid) {
2905		mutex_unlock(&root->log_mutex);
2906		return ctx->log_ret;
2907	}
2908
2909	index1 = log_transid % 2;
2910	if (atomic_read(&root->log_commit[index1])) {
2911		wait_log_commit(root, log_transid);
2912		mutex_unlock(&root->log_mutex);
2913		return ctx->log_ret;
2914	}
2915	ASSERT(log_transid == root->log_transid);
2916	atomic_set(&root->log_commit[index1], 1);
2917
2918	/* wait for previous tree log sync to complete */
2919	if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
2920		wait_log_commit(root, log_transid - 1);
2921
2922	while (1) {
2923		int batch = atomic_read(&root->log_batch);
2924		/* when we're on an ssd, just kick the log commit out */
2925		if (!btrfs_test_opt(fs_info, SSD) &&
2926		    test_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state)) {
2927			mutex_unlock(&root->log_mutex);
2928			schedule_timeout_uninterruptible(1);
2929			mutex_lock(&root->log_mutex);
2930		}
2931		wait_for_writer(root);
2932		if (batch == atomic_read(&root->log_batch))
2933			break;
2934	}
2935
2936	/* bail out if we need to do a full commit */
2937	if (btrfs_need_log_full_commit(fs_info, trans)) {
2938		ret = -EAGAIN;
2939		btrfs_free_logged_extents(log, log_transid);
2940		mutex_unlock(&root->log_mutex);
2941		goto out;
2942	}
2943
2944	if (log_transid % 2 == 0)
2945		mark = EXTENT_DIRTY;
2946	else
2947		mark = EXTENT_NEW;
2948
2949	/* we start IO on  all the marked extents here, but we don't actually
2950	 * wait for them until later.
2951	 */
2952	blk_start_plug(&plug);
2953	ret = btrfs_write_marked_extents(fs_info, &log->dirty_log_pages, mark);
 
 
 
 
 
 
 
 
 
 
 
2954	if (ret) {
2955		blk_finish_plug(&plug);
2956		btrfs_abort_transaction(trans, ret);
2957		btrfs_free_logged_extents(log, log_transid);
2958		btrfs_set_log_full_commit(fs_info, trans);
2959		mutex_unlock(&root->log_mutex);
2960		goto out;
2961	}
2962
 
 
 
 
 
 
 
 
 
 
 
 
 
2963	btrfs_set_root_node(&log->root_item, log->node);
 
2964
2965	root->log_transid++;
2966	log->log_transid = root->log_transid;
2967	root->log_start_pid = 0;
2968	/*
2969	 * IO has been started, blocks of the log tree have WRITTEN flag set
2970	 * in their headers. new modifications of the log will be written to
2971	 * new positions. so it's safe to allow log writers to go in.
2972	 */
2973	mutex_unlock(&root->log_mutex);
2974
 
 
 
 
 
 
 
 
 
 
 
 
2975	btrfs_init_log_ctx(&root_log_ctx, NULL);
2976
2977	mutex_lock(&log_root_tree->log_mutex);
2978	atomic_inc(&log_root_tree->log_batch);
2979	atomic_inc(&log_root_tree->log_writers);
2980
2981	index2 = log_root_tree->log_transid % 2;
2982	list_add_tail(&root_log_ctx.list, &log_root_tree->log_ctxs[index2]);
2983	root_log_ctx.log_transid = log_root_tree->log_transid;
2984
2985	mutex_unlock(&log_root_tree->log_mutex);
2986
2987	ret = update_log_root(trans, log);
2988
2989	mutex_lock(&log_root_tree->log_mutex);
2990	if (atomic_dec_and_test(&log_root_tree->log_writers)) {
2991		/*
2992		 * Implicit memory barrier after atomic_dec_and_test
2993		 */
2994		if (waitqueue_active(&log_root_tree->log_writer_wait))
2995			wake_up(&log_root_tree->log_writer_wait);
2996	}
2997
2998	if (ret) {
2999		if (!list_empty(&root_log_ctx.list))
3000			list_del_init(&root_log_ctx.list);
3001
3002		blk_finish_plug(&plug);
3003		btrfs_set_log_full_commit(fs_info, trans);
3004
3005		if (ret != -ENOSPC) {
3006			btrfs_abort_transaction(trans, ret);
3007			mutex_unlock(&log_root_tree->log_mutex);
3008			goto out;
3009		}
3010		btrfs_wait_tree_log_extents(log, mark);
3011		btrfs_free_logged_extents(log, log_transid);
3012		mutex_unlock(&log_root_tree->log_mutex);
3013		ret = -EAGAIN;
3014		goto out;
3015	}
3016
3017	if (log_root_tree->log_transid_committed >= root_log_ctx.log_transid) {
3018		blk_finish_plug(&plug);
3019		list_del_init(&root_log_ctx.list);
3020		mutex_unlock(&log_root_tree->log_mutex);
3021		ret = root_log_ctx.log_ret;
3022		goto out;
3023	}
3024
3025	index2 = root_log_ctx.log_transid % 2;
3026	if (atomic_read(&log_root_tree->log_commit[index2])) {
3027		blk_finish_plug(&plug);
3028		ret = btrfs_wait_tree_log_extents(log, mark);
3029		btrfs_wait_logged_extents(trans, log, log_transid);
3030		wait_log_commit(log_root_tree,
3031				root_log_ctx.log_transid);
3032		mutex_unlock(&log_root_tree->log_mutex);
3033		if (!ret)
3034			ret = root_log_ctx.log_ret;
3035		goto out;
3036	}
3037	ASSERT(root_log_ctx.log_transid == log_root_tree->log_transid);
3038	atomic_set(&log_root_tree->log_commit[index2], 1);
3039
3040	if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2])) {
3041		wait_log_commit(log_root_tree,
3042				root_log_ctx.log_transid - 1);
3043	}
3044
3045	wait_for_writer(log_root_tree);
3046
3047	/*
3048	 * now that we've moved on to the tree of log tree roots,
3049	 * check the full commit flag again
3050	 */
3051	if (btrfs_need_log_full_commit(fs_info, trans)) {
3052		blk_finish_plug(&plug);
3053		btrfs_wait_tree_log_extents(log, mark);
3054		btrfs_free_logged_extents(log, log_transid);
3055		mutex_unlock(&log_root_tree->log_mutex);
3056		ret = -EAGAIN;
3057		goto out_wake_log_root;
3058	}
3059
3060	ret = btrfs_write_marked_extents(fs_info,
3061					 &log_root_tree->dirty_log_pages,
3062					 EXTENT_DIRTY | EXTENT_NEW);
3063	blk_finish_plug(&plug);
3064	if (ret) {
3065		btrfs_set_log_full_commit(fs_info, trans);
 
 
 
 
 
 
 
 
 
 
3066		btrfs_abort_transaction(trans, ret);
3067		btrfs_free_logged_extents(log, log_transid);
3068		mutex_unlock(&log_root_tree->log_mutex);
3069		goto out_wake_log_root;
3070	}
3071	ret = btrfs_wait_tree_log_extents(log, mark);
3072	if (!ret)
3073		ret = btrfs_wait_tree_log_extents(log_root_tree,
3074						  EXTENT_NEW | EXTENT_DIRTY);
3075	if (ret) {
3076		btrfs_set_log_full_commit(fs_info, trans);
3077		btrfs_free_logged_extents(log, log_transid);
3078		mutex_unlock(&log_root_tree->log_mutex);
3079		goto out_wake_log_root;
3080	}
3081	btrfs_wait_logged_extents(trans, log, log_transid);
3082
3083	btrfs_set_super_log_root(fs_info->super_for_commit,
3084				 log_root_tree->node->start);
3085	btrfs_set_super_log_root_level(fs_info->super_for_commit,
3086				       btrfs_header_level(log_root_tree->node));
3087
 
 
3088	log_root_tree->log_transid++;
3089	mutex_unlock(&log_root_tree->log_mutex);
3090
3091	/*
3092	 * nobody else is going to jump in and write the the ctree
3093	 * super here because the log_commit atomic below is protecting
3094	 * us.  We must be called with a transaction handle pinning
3095	 * the running transaction open, so a full commit can't hop
3096	 * in and cause problems either.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3097	 */
 
 
 
 
 
 
 
 
 
 
3098	ret = write_all_supers(fs_info, 1);
 
3099	if (ret) {
3100		btrfs_set_log_full_commit(fs_info, trans);
3101		btrfs_abort_transaction(trans, ret);
3102		goto out_wake_log_root;
3103	}
3104
3105	mutex_lock(&root->log_mutex);
3106	if (root->last_log_commit < log_transid)
3107		root->last_log_commit = log_transid;
3108	mutex_unlock(&root->log_mutex);
3109
3110out_wake_log_root:
3111	mutex_lock(&log_root_tree->log_mutex);
3112	btrfs_remove_all_log_ctxs(log_root_tree, index2, ret);
3113
3114	log_root_tree->log_transid_committed++;
3115	atomic_set(&log_root_tree->log_commit[index2], 0);
3116	mutex_unlock(&log_root_tree->log_mutex);
3117
3118	/*
3119	 * The barrier before waitqueue_active is implied by mutex_unlock
 
 
3120	 */
3121	if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
3122		wake_up(&log_root_tree->log_commit_wait[index2]);
3123out:
3124	mutex_lock(&root->log_mutex);
3125	btrfs_remove_all_log_ctxs(root, index1, ret);
3126	root->log_transid_committed++;
3127	atomic_set(&root->log_commit[index1], 0);
3128	mutex_unlock(&root->log_mutex);
3129
3130	/*
3131	 * The barrier before waitqueue_active is implied by mutex_unlock
 
 
3132	 */
3133	if (waitqueue_active(&root->log_commit_wait[index1]))
3134		wake_up(&root->log_commit_wait[index1]);
3135	return ret;
3136}
3137
3138static void free_log_tree(struct btrfs_trans_handle *trans,
3139			  struct btrfs_root *log)
3140{
3141	int ret;
3142	u64 start;
3143	u64 end;
3144	struct walk_control wc = {
3145		.free = 1,
3146		.process_func = process_one_buffer
3147	};
3148
3149	ret = walk_log_tree(trans, log, &wc);
3150	/* I don't think this can happen but just in case */
3151	if (ret)
3152		btrfs_abort_transaction(trans, ret);
3153
3154	while (1) {
3155		ret = find_first_extent_bit(&log->dirty_log_pages,
3156				0, &start, &end,
3157				EXTENT_DIRTY | EXTENT_NEW | EXTENT_NEED_WAIT,
3158				NULL);
3159		if (ret)
3160			break;
3161
3162		clear_extent_bits(&log->dirty_log_pages, start, end,
3163				  EXTENT_DIRTY | EXTENT_NEW | EXTENT_NEED_WAIT);
3164	}
3165
3166	/*
3167	 * We may have short-circuited the log tree with the full commit logic
3168	 * and left ordered extents on our list, so clear these out to keep us
3169	 * from leaking inodes and memory.
3170	 */
3171	btrfs_free_logged_extents(log, 0);
3172	btrfs_free_logged_extents(log, 1);
3173
3174	free_extent_buffer(log->node);
3175	kfree(log);
 
3176}
3177
3178/*
3179 * free all the extents used by the tree log.  This should be called
3180 * at commit time of the full transaction
3181 */
3182int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
3183{
3184	if (root->log_root) {
3185		free_log_tree(trans, root->log_root);
3186		root->log_root = NULL;
 
3187	}
3188	return 0;
3189}
3190
3191int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
3192			     struct btrfs_fs_info *fs_info)
3193{
3194	if (fs_info->log_root_tree) {
3195		free_log_tree(trans, fs_info->log_root_tree);
3196		fs_info->log_root_tree = NULL;
 
3197	}
3198	return 0;
3199}
3200
3201/*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3202 * If both a file and directory are logged, and unlinks or renames are
3203 * mixed in, we have a few interesting corners:
3204 *
3205 * create file X in dir Y
3206 * link file X to X.link in dir Y
3207 * fsync file X
3208 * unlink file X but leave X.link
3209 * fsync dir Y
3210 *
3211 * After a crash we would expect only X.link to exist.  But file X
3212 * didn't get fsync'd again so the log has back refs for X and X.link.
3213 *
3214 * We solve this by removing directory entries and inode backrefs from the
3215 * log when a file that was logged in the current transaction is
3216 * unlinked.  Any later fsync will include the updated log entries, and
3217 * we'll be able to reconstruct the proper directory items from backrefs.
3218 *
3219 * This optimizations allows us to avoid relogging the entire inode
3220 * or the entire directory.
3221 */
3222int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
3223				 struct btrfs_root *root,
3224				 const char *name, int name_len,
3225				 struct btrfs_inode *dir, u64 index)
3226{
3227	struct btrfs_root *log;
3228	struct btrfs_dir_item *di;
3229	struct btrfs_path *path;
3230	int ret;
3231	int err = 0;
3232	int bytes_del = 0;
3233	u64 dir_ino = btrfs_ino(dir);
3234
3235	if (dir->logged_trans < trans->transid)
3236		return 0;
3237
3238	ret = join_running_log_trans(root);
3239	if (ret)
3240		return 0;
3241
3242	mutex_lock(&dir->log_mutex);
3243
3244	log = root->log_root;
3245	path = btrfs_alloc_path();
3246	if (!path) {
3247		err = -ENOMEM;
3248		goto out_unlock;
3249	}
3250
3251	di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
3252				   name, name_len, -1);
3253	if (IS_ERR(di)) {
3254		err = PTR_ERR(di);
3255		goto fail;
3256	}
3257	if (di) {
3258		ret = btrfs_delete_one_dir_name(trans, log, path, di);
3259		bytes_del += name_len;
3260		if (ret) {
3261			err = ret;
3262			goto fail;
3263		}
3264	}
3265	btrfs_release_path(path);
3266	di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
3267					 index, name, name_len, -1);
3268	if (IS_ERR(di)) {
3269		err = PTR_ERR(di);
3270		goto fail;
3271	}
3272	if (di) {
3273		ret = btrfs_delete_one_dir_name(trans, log, path, di);
3274		bytes_del += name_len;
3275		if (ret) {
3276			err = ret;
3277			goto fail;
3278		}
3279	}
3280
3281	/* update the directory size in the log to reflect the names
3282	 * we have removed
 
 
3283	 */
3284	if (bytes_del) {
3285		struct btrfs_key key;
3286
3287		key.objectid = dir_ino;
3288		key.offset = 0;
3289		key.type = BTRFS_INODE_ITEM_KEY;
3290		btrfs_release_path(path);
3291
3292		ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
3293		if (ret < 0) {
3294			err = ret;
3295			goto fail;
3296		}
3297		if (ret == 0) {
3298			struct btrfs_inode_item *item;
3299			u64 i_size;
3300
3301			item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3302					      struct btrfs_inode_item);
3303			i_size = btrfs_inode_size(path->nodes[0], item);
3304			if (i_size > bytes_del)
3305				i_size -= bytes_del;
3306			else
3307				i_size = 0;
3308			btrfs_set_inode_size(path->nodes[0], item, i_size);
3309			btrfs_mark_buffer_dirty(path->nodes[0]);
3310		} else
3311			ret = 0;
3312		btrfs_release_path(path);
3313	}
3314fail:
3315	btrfs_free_path(path);
3316out_unlock:
3317	mutex_unlock(&dir->log_mutex);
3318	if (ret == -ENOSPC) {
3319		btrfs_set_log_full_commit(root->fs_info, trans);
3320		ret = 0;
3321	} else if (ret < 0)
3322		btrfs_abort_transaction(trans, ret);
 
 
3323
3324	btrfs_end_log_trans(root);
3325
3326	return err;
3327}
3328
3329/* see comments for btrfs_del_dir_entries_in_log */
3330int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
3331			       struct btrfs_root *root,
3332			       const char *name, int name_len,
3333			       struct btrfs_inode *inode, u64 dirid)
3334{
3335	struct btrfs_fs_info *fs_info = root->fs_info;
3336	struct btrfs_root *log;
3337	u64 index;
3338	int ret;
3339
3340	if (inode->logged_trans < trans->transid)
3341		return 0;
3342
3343	ret = join_running_log_trans(root);
3344	if (ret)
3345		return 0;
3346	log = root->log_root;
3347	mutex_lock(&inode->log_mutex);
3348
3349	ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
3350				  dirid, &index);
3351	mutex_unlock(&inode->log_mutex);
3352	if (ret == -ENOSPC) {
3353		btrfs_set_log_full_commit(fs_info, trans);
3354		ret = 0;
3355	} else if (ret < 0 && ret != -ENOENT)
3356		btrfs_abort_transaction(trans, ret);
3357	btrfs_end_log_trans(root);
3358
3359	return ret;
3360}
3361
3362/*
3363 * creates a range item in the log for 'dirid'.  first_offset and
3364 * last_offset tell us which parts of the key space the log should
3365 * be considered authoritative for.
3366 */
3367static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
3368				       struct btrfs_root *log,
3369				       struct btrfs_path *path,
3370				       int key_type, u64 dirid,
3371				       u64 first_offset, u64 last_offset)
3372{
3373	int ret;
3374	struct btrfs_key key;
3375	struct btrfs_dir_log_item *item;
3376
3377	key.objectid = dirid;
3378	key.offset = first_offset;
3379	if (key_type == BTRFS_DIR_ITEM_KEY)
3380		key.type = BTRFS_DIR_LOG_ITEM_KEY;
3381	else
3382		key.type = BTRFS_DIR_LOG_INDEX_KEY;
3383	ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
3384	if (ret)
3385		return ret;
3386
3387	item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3388			      struct btrfs_dir_log_item);
3389	btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
3390	btrfs_mark_buffer_dirty(path->nodes[0]);
3391	btrfs_release_path(path);
3392	return 0;
3393}
3394
3395/*
3396 * log all the items included in the current transaction for a given
3397 * directory.  This also creates the range items in the log tree required
3398 * to replay anything deleted before the fsync
3399 */
3400static noinline int log_dir_items(struct btrfs_trans_handle *trans,
3401			  struct btrfs_root *root, struct btrfs_inode *inode,
3402			  struct btrfs_path *path,
3403			  struct btrfs_path *dst_path, int key_type,
3404			  struct btrfs_log_ctx *ctx,
3405			  u64 min_offset, u64 *last_offset_ret)
3406{
3407	struct btrfs_key min_key;
3408	struct btrfs_root *log = root->log_root;
3409	struct extent_buffer *src;
3410	int err = 0;
3411	int ret;
3412	int i;
3413	int nritems;
3414	u64 first_offset = min_offset;
3415	u64 last_offset = (u64)-1;
3416	u64 ino = btrfs_ino(inode);
3417
3418	log = root->log_root;
3419
3420	min_key.objectid = ino;
3421	min_key.type = key_type;
3422	min_key.offset = min_offset;
3423
3424	ret = btrfs_search_forward(root, &min_key, path, trans->transid);
3425
3426	/*
3427	 * we didn't find anything from this transaction, see if there
3428	 * is anything at all
3429	 */
3430	if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
3431		min_key.objectid = ino;
3432		min_key.type = key_type;
3433		min_key.offset = (u64)-1;
3434		btrfs_release_path(path);
3435		ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
3436		if (ret < 0) {
3437			btrfs_release_path(path);
3438			return ret;
3439		}
3440		ret = btrfs_previous_item(root, path, ino, key_type);
3441
3442		/* if ret == 0 there are items for this type,
3443		 * create a range to tell us the last key of this type.
3444		 * otherwise, there are no items in this directory after
3445		 * *min_offset, and we create a range to indicate that.
3446		 */
3447		if (ret == 0) {
3448			struct btrfs_key tmp;
3449			btrfs_item_key_to_cpu(path->nodes[0], &tmp,
3450					      path->slots[0]);
3451			if (key_type == tmp.type)
3452				first_offset = max(min_offset, tmp.offset) + 1;
3453		}
3454		goto done;
3455	}
3456
3457	/* go backward to find any previous key */
3458	ret = btrfs_previous_item(root, path, ino, key_type);
3459	if (ret == 0) {
3460		struct btrfs_key tmp;
3461		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
3462		if (key_type == tmp.type) {
3463			first_offset = tmp.offset;
3464			ret = overwrite_item(trans, log, dst_path,
3465					     path->nodes[0], path->slots[0],
3466					     &tmp);
3467			if (ret) {
3468				err = ret;
3469				goto done;
3470			}
3471		}
3472	}
3473	btrfs_release_path(path);
3474
3475	/* find the first key from this transaction again */
 
 
 
 
 
 
 
 
3476	ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
3477	if (WARN_ON(ret != 0))
3478		goto done;
3479
3480	/*
3481	 * we have a block from this transaction, log every item in it
3482	 * from our directory
3483	 */
3484	while (1) {
3485		struct btrfs_key tmp;
3486		src = path->nodes[0];
3487		nritems = btrfs_header_nritems(src);
3488		for (i = path->slots[0]; i < nritems; i++) {
3489			struct btrfs_dir_item *di;
3490
3491			btrfs_item_key_to_cpu(src, &min_key, i);
3492
3493			if (min_key.objectid != ino || min_key.type != key_type)
3494				goto done;
 
 
 
 
 
 
 
3495			ret = overwrite_item(trans, log, dst_path, src, i,
3496					     &min_key);
3497			if (ret) {
3498				err = ret;
3499				goto done;
3500			}
3501
3502			/*
3503			 * We must make sure that when we log a directory entry,
3504			 * the corresponding inode, after log replay, has a
3505			 * matching link count. For example:
3506			 *
3507			 * touch foo
3508			 * mkdir mydir
3509			 * sync
3510			 * ln foo mydir/bar
3511			 * xfs_io -c "fsync" mydir
3512			 * <crash>
3513			 * <mount fs and log replay>
3514			 *
3515			 * Would result in a fsync log that when replayed, our
3516			 * file inode would have a link count of 1, but we get
3517			 * two directory entries pointing to the same inode.
3518			 * After removing one of the names, it would not be
3519			 * possible to remove the other name, which resulted
3520			 * always in stale file handle errors, and would not
3521			 * be possible to rmdir the parent directory, since
3522			 * its i_size could never decrement to the value
3523			 * BTRFS_EMPTY_DIR_SIZE, resulting in -ENOTEMPTY errors.
3524			 */
3525			di = btrfs_item_ptr(src, i, struct btrfs_dir_item);
3526			btrfs_dir_item_key_to_cpu(src, di, &tmp);
3527			if (ctx &&
3528			    (btrfs_dir_transid(src, di) == trans->transid ||
3529			     btrfs_dir_type(src, di) == BTRFS_FT_DIR) &&
3530			    tmp.type != BTRFS_ROOT_ITEM_KEY)
3531				ctx->log_new_dentries = true;
3532		}
3533		path->slots[0] = nritems;
3534
3535		/*
3536		 * look ahead to the next item and see if it is also
3537		 * from this directory and from this transaction
3538		 */
3539		ret = btrfs_next_leaf(root, path);
3540		if (ret) {
3541			if (ret == 1)
3542				last_offset = (u64)-1;
3543			else
3544				err = ret;
3545			goto done;
3546		}
3547		btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
3548		if (tmp.objectid != ino || tmp.type != key_type) {
3549			last_offset = (u64)-1;
3550			goto done;
3551		}
3552		if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
3553			ret = overwrite_item(trans, log, dst_path,
3554					     path->nodes[0], path->slots[0],
3555					     &tmp);
3556			if (ret)
3557				err = ret;
3558			else
3559				last_offset = tmp.offset;
3560			goto done;
3561		}
3562	}
3563done:
3564	btrfs_release_path(path);
3565	btrfs_release_path(dst_path);
3566
3567	if (err == 0) {
3568		*last_offset_ret = last_offset;
3569		/*
3570		 * insert the log range keys to indicate where the log
3571		 * is valid
3572		 */
3573		ret = insert_dir_log_key(trans, log, path, key_type,
3574					 ino, first_offset, last_offset);
3575		if (ret)
3576			err = ret;
3577	}
3578	return err;
3579}
3580
3581/*
3582 * logging directories is very similar to logging inodes, We find all the items
3583 * from the current transaction and write them to the log.
3584 *
3585 * The recovery code scans the directory in the subvolume, and if it finds a
3586 * key in the range logged that is not present in the log tree, then it means
3587 * that dir entry was unlinked during the transaction.
3588 *
3589 * In order for that scan to work, we must include one key smaller than
3590 * the smallest logged by this transaction and one key larger than the largest
3591 * key logged by this transaction.
3592 */
3593static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
3594			  struct btrfs_root *root, struct btrfs_inode *inode,
3595			  struct btrfs_path *path,
3596			  struct btrfs_path *dst_path,
3597			  struct btrfs_log_ctx *ctx)
3598{
3599	u64 min_key;
3600	u64 max_key;
3601	int ret;
3602	int key_type = BTRFS_DIR_ITEM_KEY;
3603
3604again:
3605	min_key = 0;
3606	max_key = 0;
3607	while (1) {
3608		ret = log_dir_items(trans, root, inode, path, dst_path, key_type,
3609				ctx, min_key, &max_key);
3610		if (ret)
3611			return ret;
3612		if (max_key == (u64)-1)
3613			break;
3614		min_key = max_key + 1;
3615	}
3616
3617	if (key_type == BTRFS_DIR_ITEM_KEY) {
3618		key_type = BTRFS_DIR_INDEX_KEY;
3619		goto again;
3620	}
3621	return 0;
3622}
3623
3624/*
3625 * a helper function to drop items from the log before we relog an
3626 * inode.  max_key_type indicates the highest item type to remove.
3627 * This cannot be run for file data extents because it does not
3628 * free the extents they point to.
3629 */
3630static int drop_objectid_items(struct btrfs_trans_handle *trans,
3631				  struct btrfs_root *log,
3632				  struct btrfs_path *path,
3633				  u64 objectid, int max_key_type)
3634{
3635	int ret;
3636	struct btrfs_key key;
3637	struct btrfs_key found_key;
3638	int start_slot;
3639
3640	key.objectid = objectid;
3641	key.type = max_key_type;
3642	key.offset = (u64)-1;
3643
3644	while (1) {
3645		ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
3646		BUG_ON(ret == 0); /* Logic error */
3647		if (ret < 0)
3648			break;
3649
3650		if (path->slots[0] == 0)
3651			break;
3652
3653		path->slots[0]--;
3654		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
3655				      path->slots[0]);
3656
3657		if (found_key.objectid != objectid)
3658			break;
3659
3660		found_key.offset = 0;
3661		found_key.type = 0;
3662		ret = btrfs_bin_search(path->nodes[0], &found_key, 0,
3663				       &start_slot);
 
3664
3665		ret = btrfs_del_items(trans, log, path, start_slot,
3666				      path->slots[0] - start_slot + 1);
3667		/*
3668		 * If start slot isn't 0 then we don't need to re-search, we've
3669		 * found the last guy with the objectid in this tree.
3670		 */
3671		if (ret || start_slot != 0)
3672			break;
3673		btrfs_release_path(path);
3674	}
3675	btrfs_release_path(path);
3676	if (ret > 0)
3677		ret = 0;
3678	return ret;
3679}
3680
3681static void fill_inode_item(struct btrfs_trans_handle *trans,
3682			    struct extent_buffer *leaf,
3683			    struct btrfs_inode_item *item,
3684			    struct inode *inode, int log_inode_only,
3685			    u64 logged_isize)
3686{
3687	struct btrfs_map_token token;
3688
3689	btrfs_init_map_token(&token);
3690
3691	if (log_inode_only) {
3692		/* set the generation to zero so the recover code
3693		 * can tell the difference between an logging
3694		 * just to say 'this inode exists' and a logging
3695		 * to say 'update this inode with these values'
3696		 */
3697		btrfs_set_token_inode_generation(leaf, item, 0, &token);
3698		btrfs_set_token_inode_size(leaf, item, logged_isize, &token);
3699	} else {
3700		btrfs_set_token_inode_generation(leaf, item,
3701						 BTRFS_I(inode)->generation,
3702						 &token);
3703		btrfs_set_token_inode_size(leaf, item, inode->i_size, &token);
3704	}
3705
3706	btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token);
3707	btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token);
3708	btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token);
3709	btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token);
3710
3711	btrfs_set_token_timespec_sec(leaf, &item->atime,
3712				     inode->i_atime.tv_sec, &token);
3713	btrfs_set_token_timespec_nsec(leaf, &item->atime,
3714				      inode->i_atime.tv_nsec, &token);
3715
3716	btrfs_set_token_timespec_sec(leaf, &item->mtime,
3717				     inode->i_mtime.tv_sec, &token);
3718	btrfs_set_token_timespec_nsec(leaf, &item->mtime,
3719				      inode->i_mtime.tv_nsec, &token);
3720
3721	btrfs_set_token_timespec_sec(leaf, &item->ctime,
3722				     inode->i_ctime.tv_sec, &token);
3723	btrfs_set_token_timespec_nsec(leaf, &item->ctime,
3724				      inode->i_ctime.tv_nsec, &token);
3725
3726	btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode),
3727				     &token);
3728
3729	btrfs_set_token_inode_sequence(leaf, item,
3730				       inode_peek_iversion(inode), &token);
3731	btrfs_set_token_inode_transid(leaf, item, trans->transid, &token);
3732	btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token);
3733	btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token);
3734	btrfs_set_token_inode_block_group(leaf, item, 0, &token);
 
 
 
 
3735}
3736
3737static int log_inode_item(struct btrfs_trans_handle *trans,
3738			  struct btrfs_root *log, struct btrfs_path *path,
3739			  struct btrfs_inode *inode)
3740{
3741	struct btrfs_inode_item *inode_item;
3742	int ret;
3743
3744	ret = btrfs_insert_empty_item(trans, log, path,
3745				      &inode->location, sizeof(*inode_item));
3746	if (ret && ret != -EEXIST)
3747		return ret;
3748	inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3749				    struct btrfs_inode_item);
3750	fill_inode_item(trans, path->nodes[0], inode_item, &inode->vfs_inode,
3751			0, 0);
3752	btrfs_release_path(path);
3753	return 0;
3754}
3755
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3756static noinline int copy_items(struct btrfs_trans_handle *trans,
3757			       struct btrfs_inode *inode,
3758			       struct btrfs_path *dst_path,
3759			       struct btrfs_path *src_path, u64 *last_extent,
3760			       int start_slot, int nr, int inode_only,
3761			       u64 logged_isize)
3762{
3763	struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb);
3764	unsigned long src_offset;
3765	unsigned long dst_offset;
3766	struct btrfs_root *log = inode->root->log_root;
3767	struct btrfs_file_extent_item *extent;
3768	struct btrfs_inode_item *inode_item;
3769	struct extent_buffer *src = src_path->nodes[0];
3770	struct btrfs_key first_key, last_key, key;
3771	int ret;
3772	struct btrfs_key *ins_keys;
3773	u32 *ins_sizes;
3774	char *ins_data;
3775	int i;
3776	struct list_head ordered_sums;
3777	int skip_csum = inode->flags & BTRFS_INODE_NODATASUM;
3778	bool has_extents = false;
3779	bool need_find_last_extent = true;
3780	bool done = false;
3781
3782	INIT_LIST_HEAD(&ordered_sums);
3783
3784	ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
3785			   nr * sizeof(u32), GFP_NOFS);
3786	if (!ins_data)
3787		return -ENOMEM;
3788
3789	first_key.objectid = (u64)-1;
3790
3791	ins_sizes = (u32 *)ins_data;
3792	ins_keys = (struct btrfs_key *)(ins_data + nr * sizeof(u32));
3793
3794	for (i = 0; i < nr; i++) {
3795		ins_sizes[i] = btrfs_item_size_nr(src, i + start_slot);
3796		btrfs_item_key_to_cpu(src, ins_keys + i, i + start_slot);
3797	}
3798	ret = btrfs_insert_empty_items(trans, log, dst_path,
3799				       ins_keys, ins_sizes, nr);
3800	if (ret) {
3801		kfree(ins_data);
3802		return ret;
3803	}
3804
3805	for (i = 0; i < nr; i++, dst_path->slots[0]++) {
3806		dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
3807						   dst_path->slots[0]);
3808
3809		src_offset = btrfs_item_ptr_offset(src, start_slot + i);
3810
3811		if (i == nr - 1)
3812			last_key = ins_keys[i];
3813
3814		if (ins_keys[i].type == BTRFS_INODE_ITEM_KEY) {
3815			inode_item = btrfs_item_ptr(dst_path->nodes[0],
3816						    dst_path->slots[0],
3817						    struct btrfs_inode_item);
3818			fill_inode_item(trans, dst_path->nodes[0], inode_item,
3819					&inode->vfs_inode,
3820					inode_only == LOG_INODE_EXISTS,
3821					logged_isize);
3822		} else {
3823			copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
3824					   src_offset, ins_sizes[i]);
3825		}
3826
3827		/*
3828		 * We set need_find_last_extent here in case we know we were
3829		 * processing other items and then walk into the first extent in
3830		 * the inode.  If we don't hit an extent then nothing changes,
3831		 * we'll do the last search the next time around.
3832		 */
3833		if (ins_keys[i].type == BTRFS_EXTENT_DATA_KEY) {
3834			has_extents = true;
3835			if (first_key.objectid == (u64)-1)
3836				first_key = ins_keys[i];
3837		} else {
3838			need_find_last_extent = false;
3839		}
3840
3841		/* take a reference on file data extents so that truncates
3842		 * or deletes of this inode don't have to relog the inode
3843		 * again
3844		 */
3845		if (ins_keys[i].type == BTRFS_EXTENT_DATA_KEY &&
3846		    !skip_csum) {
3847			int found_type;
3848			extent = btrfs_item_ptr(src, start_slot + i,
3849						struct btrfs_file_extent_item);
3850
3851			if (btrfs_file_extent_generation(src, extent) < trans->transid)
3852				continue;
3853
3854			found_type = btrfs_file_extent_type(src, extent);
3855			if (found_type == BTRFS_FILE_EXTENT_REG) {
3856				u64 ds, dl, cs, cl;
3857				ds = btrfs_file_extent_disk_bytenr(src,
3858								extent);
3859				/* ds == 0 is a hole */
3860				if (ds == 0)
3861					continue;
3862
3863				dl = btrfs_file_extent_disk_num_bytes(src,
3864								extent);
3865				cs = btrfs_file_extent_offset(src, extent);
3866				cl = btrfs_file_extent_num_bytes(src,
3867								extent);
3868				if (btrfs_file_extent_compression(src,
3869								  extent)) {
3870					cs = 0;
3871					cl = dl;
3872				}
3873
3874				ret = btrfs_lookup_csums_range(
3875						fs_info->csum_root,
3876						ds + cs, ds + cs + cl - 1,
3877						&ordered_sums, 0);
3878				if (ret) {
3879					btrfs_release_path(dst_path);
3880					kfree(ins_data);
3881					return ret;
3882				}
3883			}
3884		}
3885	}
3886
3887	btrfs_mark_buffer_dirty(dst_path->nodes[0]);
3888	btrfs_release_path(dst_path);
3889	kfree(ins_data);
3890
3891	/*
3892	 * we have to do this after the loop above to avoid changing the
3893	 * log tree while trying to change the log tree.
3894	 */
3895	ret = 0;
3896	while (!list_empty(&ordered_sums)) {
3897		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
3898						   struct btrfs_ordered_sum,
3899						   list);
3900		if (!ret)
3901			ret = btrfs_csum_file_blocks(trans, log, sums);
3902		list_del(&sums->list);
3903		kfree(sums);
3904	}
3905
3906	if (!has_extents)
3907		return ret;
3908
3909	if (need_find_last_extent && *last_extent == first_key.offset) {
3910		/*
3911		 * We don't have any leafs between our current one and the one
3912		 * we processed before that can have file extent items for our
3913		 * inode (and have a generation number smaller than our current
3914		 * transaction id).
3915		 */
3916		need_find_last_extent = false;
3917	}
3918
3919	/*
3920	 * Because we use btrfs_search_forward we could skip leaves that were
3921	 * not modified and then assume *last_extent is valid when it really
3922	 * isn't.  So back up to the previous leaf and read the end of the last
3923	 * extent before we go and fill in holes.
3924	 */
3925	if (need_find_last_extent) {
3926		u64 len;
3927
3928		ret = btrfs_prev_leaf(inode->root, src_path);
3929		if (ret < 0)
3930			return ret;
3931		if (ret)
3932			goto fill_holes;
3933		if (src_path->slots[0])
3934			src_path->slots[0]--;
3935		src = src_path->nodes[0];
3936		btrfs_item_key_to_cpu(src, &key, src_path->slots[0]);
3937		if (key.objectid != btrfs_ino(inode) ||
3938		    key.type != BTRFS_EXTENT_DATA_KEY)
3939			goto fill_holes;
3940		extent = btrfs_item_ptr(src, src_path->slots[0],
3941					struct btrfs_file_extent_item);
3942		if (btrfs_file_extent_type(src, extent) ==
3943		    BTRFS_FILE_EXTENT_INLINE) {
3944			len = btrfs_file_extent_inline_len(src,
3945							   src_path->slots[0],
3946							   extent);
3947			*last_extent = ALIGN(key.offset + len,
3948					     fs_info->sectorsize);
3949		} else {
3950			len = btrfs_file_extent_num_bytes(src, extent);
3951			*last_extent = key.offset + len;
3952		}
3953	}
3954fill_holes:
3955	/* So we did prev_leaf, now we need to move to the next leaf, but a few
3956	 * things could have happened
3957	 *
3958	 * 1) A merge could have happened, so we could currently be on a leaf
3959	 * that holds what we were copying in the first place.
3960	 * 2) A split could have happened, and now not all of the items we want
3961	 * are on the same leaf.
3962	 *
3963	 * So we need to adjust how we search for holes, we need to drop the
3964	 * path and re-search for the first extent key we found, and then walk
3965	 * forward until we hit the last one we copied.
3966	 */
3967	if (need_find_last_extent) {
3968		/* btrfs_prev_leaf could return 1 without releasing the path */
3969		btrfs_release_path(src_path);
3970		ret = btrfs_search_slot(NULL, inode->root, &first_key,
3971				src_path, 0, 0);
3972		if (ret < 0)
3973			return ret;
3974		ASSERT(ret == 0);
3975		src = src_path->nodes[0];
3976		i = src_path->slots[0];
3977	} else {
3978		i = start_slot;
3979	}
3980
3981	/*
3982	 * Ok so here we need to go through and fill in any holes we may have
3983	 * to make sure that holes are punched for those areas in case they had
3984	 * extents previously.
3985	 */
3986	while (!done) {
3987		u64 offset, len;
3988		u64 extent_end;
3989
3990		if (i >= btrfs_header_nritems(src_path->nodes[0])) {
3991			ret = btrfs_next_leaf(inode->root, src_path);
3992			if (ret < 0)
3993				return ret;
3994			ASSERT(ret == 0);
3995			src = src_path->nodes[0];
3996			i = 0;
3997			need_find_last_extent = true;
3998		}
3999
4000		btrfs_item_key_to_cpu(src, &key, i);
4001		if (!btrfs_comp_cpu_keys(&key, &last_key))
4002			done = true;
4003		if (key.objectid != btrfs_ino(inode) ||
4004		    key.type != BTRFS_EXTENT_DATA_KEY) {
4005			i++;
4006			continue;
4007		}
4008		extent = btrfs_item_ptr(src, i, struct btrfs_file_extent_item);
4009		if (btrfs_file_extent_type(src, extent) ==
4010		    BTRFS_FILE_EXTENT_INLINE) {
4011			len = btrfs_file_extent_inline_len(src, i, extent);
4012			extent_end = ALIGN(key.offset + len,
4013					   fs_info->sectorsize);
4014		} else {
4015			len = btrfs_file_extent_num_bytes(src, extent);
4016			extent_end = key.offset + len;
4017		}
4018		i++;
4019
4020		if (*last_extent == key.offset) {
4021			*last_extent = extent_end;
4022			continue;
4023		}
4024		offset = *last_extent;
4025		len = key.offset - *last_extent;
4026		ret = btrfs_insert_file_extent(trans, log, btrfs_ino(inode),
4027				offset, 0, 0, len, 0, len, 0, 0, 0);
4028		if (ret)
4029			break;
4030		*last_extent = extent_end;
4031	}
4032
4033	/*
4034	 * Check if there is a hole between the last extent found in our leaf
4035	 * and the first extent in the next leaf. If there is one, we need to
4036	 * log an explicit hole so that at replay time we can punch the hole.
4037	 */
4038	if (ret == 0 &&
4039	    key.objectid == btrfs_ino(inode) &&
4040	    key.type == BTRFS_EXTENT_DATA_KEY &&
4041	    i == btrfs_header_nritems(src_path->nodes[0])) {
4042		ret = btrfs_next_leaf(inode->root, src_path);
4043		need_find_last_extent = true;
4044		if (ret > 0) {
4045			ret = 0;
4046		} else if (ret == 0) {
4047			btrfs_item_key_to_cpu(src_path->nodes[0], &key,
4048					      src_path->slots[0]);
4049			if (key.objectid == btrfs_ino(inode) &&
4050			    key.type == BTRFS_EXTENT_DATA_KEY &&
4051			    *last_extent < key.offset) {
4052				const u64 len = key.offset - *last_extent;
4053
4054				ret = btrfs_insert_file_extent(trans, log,
4055							       btrfs_ino(inode),
4056							       *last_extent, 0,
4057							       0, len, 0, len,
4058							       0, 0, 0);
4059			}
4060		}
4061	}
4062	/*
4063	 * Need to let the callers know we dropped the path so they should
4064	 * re-search.
4065	 */
4066	if (!ret && need_find_last_extent)
4067		ret = 1;
4068	return ret;
4069}
4070
4071static int extent_cmp(void *priv, struct list_head *a, struct list_head *b)
 
4072{
4073	struct extent_map *em1, *em2;
4074
4075	em1 = list_entry(a, struct extent_map, list);
4076	em2 = list_entry(b, struct extent_map, list);
4077
4078	if (em1->start < em2->start)
4079		return -1;
4080	else if (em1->start > em2->start)
4081		return 1;
4082	return 0;
4083}
4084
4085static int wait_ordered_extents(struct btrfs_trans_handle *trans,
4086				struct inode *inode,
4087				struct btrfs_root *root,
4088				const struct extent_map *em,
4089				const struct list_head *logged_list,
4090				bool *ordered_io_error)
4091{
4092	struct btrfs_fs_info *fs_info = root->fs_info;
4093	struct btrfs_ordered_extent *ordered;
4094	struct btrfs_root *log = root->log_root;
4095	u64 mod_start = em->mod_start;
4096	u64 mod_len = em->mod_len;
4097	const bool skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
4098	u64 csum_offset;
4099	u64 csum_len;
 
 
4100	LIST_HEAD(ordered_sums);
4101	int ret = 0;
4102
4103	*ordered_io_error = false;
4104
4105	if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) ||
4106	    em->block_start == EXTENT_MAP_HOLE)
4107		return 0;
4108
4109	/*
4110	 * Wait far any ordered extent that covers our extent map. If it
4111	 * finishes without an error, first check and see if our csums are on
4112	 * our outstanding ordered extents.
4113	 */
4114	list_for_each_entry(ordered, logged_list, log_list) {
4115		struct btrfs_ordered_sum *sum;
4116
4117		if (!mod_len)
4118			break;
4119
4120		if (ordered->file_offset + ordered->len <= mod_start ||
4121		    mod_start + mod_len <= ordered->file_offset)
4122			continue;
4123
4124		if (!test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) &&
4125		    !test_bit(BTRFS_ORDERED_IOERR, &ordered->flags) &&
4126		    !test_bit(BTRFS_ORDERED_DIRECT, &ordered->flags)) {
4127			const u64 start = ordered->file_offset;
4128			const u64 end = ordered->file_offset + ordered->len - 1;
4129
4130			WARN_ON(ordered->inode != inode);
4131			filemap_fdatawrite_range(inode->i_mapping, start, end);
4132		}
4133
4134		wait_event(ordered->wait,
4135			   (test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) ||
4136			    test_bit(BTRFS_ORDERED_IOERR, &ordered->flags)));
4137
4138		if (test_bit(BTRFS_ORDERED_IOERR, &ordered->flags)) {
4139			/*
4140			 * Clear the AS_EIO/AS_ENOSPC flags from the inode's
4141			 * i_mapping flags, so that the next fsync won't get
4142			 * an outdated io error too.
4143			 */
4144			filemap_check_errors(inode->i_mapping);
4145			*ordered_io_error = true;
4146			break;
4147		}
4148		/*
4149		 * We are going to copy all the csums on this ordered extent, so
4150		 * go ahead and adjust mod_start and mod_len in case this
4151		 * ordered extent has already been logged.
4152		 */
4153		if (ordered->file_offset > mod_start) {
4154			if (ordered->file_offset + ordered->len >=
4155			    mod_start + mod_len)
4156				mod_len = ordered->file_offset - mod_start;
4157			/*
4158			 * If we have this case
4159			 *
4160			 * |--------- logged extent ---------|
4161			 *       |----- ordered extent ----|
4162			 *
4163			 * Just don't mess with mod_start and mod_len, we'll
4164			 * just end up logging more csums than we need and it
4165			 * will be ok.
4166			 */
4167		} else {
4168			if (ordered->file_offset + ordered->len <
4169			    mod_start + mod_len) {
4170				mod_len = (mod_start + mod_len) -
4171					(ordered->file_offset + ordered->len);
4172				mod_start = ordered->file_offset +
4173					ordered->len;
4174			} else {
4175				mod_len = 0;
4176			}
4177		}
4178
4179		if (skip_csum)
4180			continue;
4181
4182		/*
4183		 * To keep us from looping for the above case of an ordered
4184		 * extent that falls inside of the logged extent.
4185		 */
4186		if (test_and_set_bit(BTRFS_ORDERED_LOGGED_CSUM,
4187				     &ordered->flags))
4188			continue;
4189
4190		list_for_each_entry(sum, &ordered->list, list) {
4191			ret = btrfs_csum_file_blocks(trans, log, sum);
4192			if (ret)
4193				break;
4194		}
4195	}
4196
4197	if (*ordered_io_error || !mod_len || ret || skip_csum)
4198		return ret;
 
4199
 
4200	if (em->compress_type) {
4201		csum_offset = 0;
4202		csum_len = max(em->block_len, em->orig_block_len);
4203	} else {
4204		csum_offset = mod_start - em->start;
4205		csum_len = mod_len;
4206	}
4207
4208	/* block start is already adjusted for the file extent offset. */
4209	ret = btrfs_lookup_csums_range(fs_info->csum_root,
4210				       em->block_start + csum_offset,
4211				       em->block_start + csum_offset +
4212				       csum_len - 1, &ordered_sums, 0);
4213	if (ret)
4214		return ret;
4215
4216	while (!list_empty(&ordered_sums)) {
4217		struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
4218						   struct btrfs_ordered_sum,
4219						   list);
4220		if (!ret)
4221			ret = btrfs_csum_file_blocks(trans, log, sums);
4222		list_del(&sums->list);
4223		kfree(sums);
4224	}
4225
4226	return ret;
4227}
4228
4229static int log_one_extent(struct btrfs_trans_handle *trans,
4230			  struct btrfs_inode *inode, struct btrfs_root *root,
4231			  const struct extent_map *em,
4232			  struct btrfs_path *path,
4233			  const struct list_head *logged_list,
4234			  struct btrfs_log_ctx *ctx)
4235{
 
4236	struct btrfs_root *log = root->log_root;
4237	struct btrfs_file_extent_item *fi;
4238	struct extent_buffer *leaf;
4239	struct btrfs_map_token token;
4240	struct btrfs_key key;
4241	u64 extent_offset = em->start - em->orig_start;
4242	u64 block_len;
4243	int ret;
4244	int extent_inserted = 0;
4245	bool ordered_io_err = false;
4246
4247	ret = wait_ordered_extents(trans, &inode->vfs_inode, root, em,
4248			logged_list, &ordered_io_err);
4249	if (ret)
4250		return ret;
4251
4252	if (ordered_io_err) {
4253		ctx->io_err = -EIO;
4254		return ctx->io_err;
4255	}
4256
4257	btrfs_init_map_token(&token);
4258
4259	ret = __btrfs_drop_extents(trans, log, &inode->vfs_inode, path, em->start,
4260				   em->start + em->len, NULL, 0, 1,
4261				   sizeof(*fi), &extent_inserted);
4262	if (ret)
4263		return ret;
4264
4265	if (!extent_inserted) {
4266		key.objectid = btrfs_ino(inode);
4267		key.type = BTRFS_EXTENT_DATA_KEY;
4268		key.offset = em->start;
4269
4270		ret = btrfs_insert_empty_item(trans, log, path, &key,
4271					      sizeof(*fi));
4272		if (ret)
4273			return ret;
4274	}
4275	leaf = path->nodes[0];
 
4276	fi = btrfs_item_ptr(leaf, path->slots[0],
4277			    struct btrfs_file_extent_item);
4278
4279	btrfs_set_token_file_extent_generation(leaf, fi, trans->transid,
4280					       &token);
4281	if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
4282		btrfs_set_token_file_extent_type(leaf, fi,
4283						 BTRFS_FILE_EXTENT_PREALLOC,
4284						 &token);
4285	else
4286		btrfs_set_token_file_extent_type(leaf, fi,
4287						 BTRFS_FILE_EXTENT_REG,
4288						 &token);
4289
4290	block_len = max(em->block_len, em->orig_block_len);
4291	if (em->compress_type != BTRFS_COMPRESS_NONE) {
4292		btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
4293							em->block_start,
4294							&token);
4295		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
4296							   &token);
4297	} else if (em->block_start < EXTENT_MAP_LAST_BYTE) {
4298		btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
4299							em->block_start -
4300							extent_offset, &token);
4301		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
4302							   &token);
4303	} else {
4304		btrfs_set_token_file_extent_disk_bytenr(leaf, fi, 0, &token);
4305		btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, 0,
4306							   &token);
4307	}
4308
4309	btrfs_set_token_file_extent_offset(leaf, fi, extent_offset, &token);
4310	btrfs_set_token_file_extent_num_bytes(leaf, fi, em->len, &token);
4311	btrfs_set_token_file_extent_ram_bytes(leaf, fi, em->ram_bytes, &token);
4312	btrfs_set_token_file_extent_compression(leaf, fi, em->compress_type,
4313						&token);
4314	btrfs_set_token_file_extent_encryption(leaf, fi, 0, &token);
4315	btrfs_set_token_file_extent_other_encoding(leaf, fi, 0, &token);
4316	btrfs_mark_buffer_dirty(leaf);
4317
4318	btrfs_release_path(path);
4319
4320	return ret;
4321}
4322
4323/*
4324 * Log all prealloc extents beyond the inode's i_size to make sure we do not
4325 * lose them after doing a fast fsync and replaying the log. We scan the
4326 * subvolume's root instead of iterating the inode's extent map tree because
4327 * otherwise we can log incorrect extent items based on extent map conversion.
4328 * That can happen due to the fact that extent maps are merged when they
4329 * are not in the extent map tree's list of modified extents.
4330 */
4331static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
4332				      struct btrfs_inode *inode,
4333				      struct btrfs_path *path)
4334{
4335	struct btrfs_root *root = inode->root;
4336	struct btrfs_key key;
4337	const u64 i_size = i_size_read(&inode->vfs_inode);
4338	const u64 ino = btrfs_ino(inode);
4339	struct btrfs_path *dst_path = NULL;
4340	u64 last_extent = (u64)-1;
 
 
 
4341	int ins_nr = 0;
4342	int start_slot;
4343	int ret;
4344
4345	if (!(inode->flags & BTRFS_INODE_PREALLOC))
4346		return 0;
4347
4348	key.objectid = ino;
4349	key.type = BTRFS_EXTENT_DATA_KEY;
4350	key.offset = i_size;
4351	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4352	if (ret < 0)
4353		goto out;
4354
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4355	while (true) {
4356		struct extent_buffer *leaf = path->nodes[0];
4357		int slot = path->slots[0];
4358
4359		if (slot >= btrfs_header_nritems(leaf)) {
4360			if (ins_nr > 0) {
4361				ret = copy_items(trans, inode, dst_path, path,
4362						 &last_extent, start_slot,
4363						 ins_nr, 1, 0);
4364				if (ret < 0)
4365					goto out;
4366				ins_nr = 0;
4367			}
4368			ret = btrfs_next_leaf(root, path);
4369			if (ret < 0)
4370				goto out;
4371			if (ret > 0) {
4372				ret = 0;
4373				break;
4374			}
4375			continue;
4376		}
4377
4378		btrfs_item_key_to_cpu(leaf, &key, slot);
4379		if (key.objectid > ino)
4380			break;
4381		if (WARN_ON_ONCE(key.objectid < ino) ||
4382		    key.type < BTRFS_EXTENT_DATA_KEY ||
4383		    key.offset < i_size) {
4384			path->slots[0]++;
4385			continue;
4386		}
4387		if (last_extent == (u64)-1) {
4388			last_extent = key.offset;
4389			/*
4390			 * Avoid logging extent items logged in past fsync calls
4391			 * and leading to duplicate keys in the log tree.
4392			 */
4393			do {
4394				ret = btrfs_truncate_inode_items(trans,
4395							 root->log_root,
4396							 &inode->vfs_inode,
4397							 i_size,
4398							 BTRFS_EXTENT_DATA_KEY);
4399			} while (ret == -EAGAIN);
4400			if (ret)
4401				goto out;
 
4402		}
4403		if (ins_nr == 0)
4404			start_slot = slot;
4405		ins_nr++;
4406		path->slots[0]++;
4407		if (!dst_path) {
4408			dst_path = btrfs_alloc_path();
4409			if (!dst_path) {
4410				ret = -ENOMEM;
4411				goto out;
4412			}
4413		}
4414	}
4415	if (ins_nr > 0) {
4416		ret = copy_items(trans, inode, dst_path, path, &last_extent,
4417				 start_slot, ins_nr, 1, 0);
4418		if (ret > 0)
4419			ret = 0;
4420	}
4421out:
4422	btrfs_release_path(path);
4423	btrfs_free_path(dst_path);
4424	return ret;
4425}
4426
4427static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
4428				     struct btrfs_root *root,
4429				     struct btrfs_inode *inode,
4430				     struct btrfs_path *path,
4431				     struct list_head *logged_list,
4432				     struct btrfs_log_ctx *ctx,
4433				     const u64 start,
4434				     const u64 end)
4435{
 
 
4436	struct extent_map *em, *n;
4437	struct list_head extents;
4438	struct extent_map_tree *tree = &inode->extent_tree;
4439	u64 logged_start, logged_end;
4440	u64 test_gen;
4441	int ret = 0;
4442	int num = 0;
4443
4444	INIT_LIST_HEAD(&extents);
4445
4446	down_write(&inode->dio_sem);
4447	write_lock(&tree->lock);
4448	test_gen = root->fs_info->last_trans_committed;
4449	logged_start = start;
4450	logged_end = end;
4451
4452	list_for_each_entry_safe(em, n, &tree->modified_extents, list) {
4453		list_del_init(&em->list);
4454		/*
4455		 * Just an arbitrary number, this can be really CPU intensive
4456		 * once we start getting a lot of extents, and really once we
4457		 * have a bunch of extents we just want to commit since it will
4458		 * be faster.
4459		 */
4460		if (++num > 32768) {
4461			list_del_init(&tree->modified_extents);
4462			ret = -EFBIG;
4463			goto process;
4464		}
4465
4466		if (em->generation <= test_gen)
4467			continue;
4468
4469		/* We log prealloc extents beyond eof later. */
4470		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) &&
4471		    em->start >= i_size_read(&inode->vfs_inode))
4472			continue;
4473
4474		if (em->start < logged_start)
4475			logged_start = em->start;
4476		if ((em->start + em->len - 1) > logged_end)
4477			logged_end = em->start + em->len - 1;
4478
4479		/* Need a ref to keep it from getting evicted from cache */
4480		refcount_inc(&em->refs);
4481		set_bit(EXTENT_FLAG_LOGGING, &em->flags);
4482		list_add_tail(&em->list, &extents);
4483		num++;
4484	}
4485
4486	list_sort(NULL, &extents, extent_cmp);
4487	btrfs_get_logged_extents(inode, logged_list, logged_start, logged_end);
4488	/*
4489	 * Some ordered extents started by fsync might have completed
4490	 * before we could collect them into the list logged_list, which
4491	 * means they're gone, not in our logged_list nor in the inode's
4492	 * ordered tree. We want the application/user space to know an
4493	 * error happened while attempting to persist file data so that
4494	 * it can take proper action. If such error happened, we leave
4495	 * without writing to the log tree and the fsync must report the
4496	 * file data write error and not commit the current transaction.
4497	 */
4498	ret = filemap_check_errors(inode->vfs_inode.i_mapping);
4499	if (ret)
4500		ctx->io_err = ret;
4501process:
4502	while (!list_empty(&extents)) {
4503		em = list_entry(extents.next, struct extent_map, list);
4504
4505		list_del_init(&em->list);
4506
4507		/*
4508		 * If we had an error we just need to delete everybody from our
4509		 * private list.
4510		 */
4511		if (ret) {
4512			clear_em_logging(tree, em);
4513			free_extent_map(em);
4514			continue;
4515		}
4516
4517		write_unlock(&tree->lock);
4518
4519		ret = log_one_extent(trans, inode, root, em, path, logged_list,
4520				     ctx);
4521		write_lock(&tree->lock);
4522		clear_em_logging(tree, em);
4523		free_extent_map(em);
4524	}
4525	WARN_ON(!list_empty(&extents));
4526	write_unlock(&tree->lock);
4527	up_write(&inode->dio_sem);
4528
4529	btrfs_release_path(path);
4530	if (!ret)
4531		ret = btrfs_log_prealloc_extents(trans, inode, path);
 
 
4532
4533	return ret;
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4534}
4535
4536static int logged_inode_size(struct btrfs_root *log, struct btrfs_inode *inode,
4537			     struct btrfs_path *path, u64 *size_ret)
4538{
4539	struct btrfs_key key;
4540	int ret;
4541
4542	key.objectid = btrfs_ino(inode);
4543	key.type = BTRFS_INODE_ITEM_KEY;
4544	key.offset = 0;
4545
4546	ret = btrfs_search_slot(NULL, log, &key, path, 0, 0);
4547	if (ret < 0) {
4548		return ret;
4549	} else if (ret > 0) {
4550		*size_ret = 0;
4551	} else {
4552		struct btrfs_inode_item *item;
4553
4554		item = btrfs_item_ptr(path->nodes[0], path->slots[0],
4555				      struct btrfs_inode_item);
4556		*size_ret = btrfs_inode_size(path->nodes[0], item);
 
 
 
 
 
 
 
 
 
 
 
 
 
4557	}
4558
4559	btrfs_release_path(path);
4560	return 0;
4561}
4562
4563/*
4564 * At the moment we always log all xattrs. This is to figure out at log replay
4565 * time which xattrs must have their deletion replayed. If a xattr is missing
4566 * in the log tree and exists in the fs/subvol tree, we delete it. This is
4567 * because if a xattr is deleted, the inode is fsynced and a power failure
4568 * happens, causing the log to be replayed the next time the fs is mounted,
4569 * we want the xattr to not exist anymore (same behaviour as other filesystems
4570 * with a journal, ext3/4, xfs, f2fs, etc).
4571 */
4572static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
4573				struct btrfs_root *root,
4574				struct btrfs_inode *inode,
4575				struct btrfs_path *path,
4576				struct btrfs_path *dst_path)
4577{
4578	int ret;
4579	struct btrfs_key key;
4580	const u64 ino = btrfs_ino(inode);
4581	int ins_nr = 0;
4582	int start_slot = 0;
 
 
 
 
4583
4584	key.objectid = ino;
4585	key.type = BTRFS_XATTR_ITEM_KEY;
4586	key.offset = 0;
4587
4588	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4589	if (ret < 0)
4590		return ret;
4591
4592	while (true) {
4593		int slot = path->slots[0];
4594		struct extent_buffer *leaf = path->nodes[0];
4595		int nritems = btrfs_header_nritems(leaf);
4596
4597		if (slot >= nritems) {
4598			if (ins_nr > 0) {
4599				u64 last_extent = 0;
4600
4601				ret = copy_items(trans, inode, dst_path, path,
4602						 &last_extent, start_slot,
4603						 ins_nr, 1, 0);
4604				/* can't be 1, extent items aren't processed */
4605				ASSERT(ret <= 0);
4606				if (ret < 0)
4607					return ret;
4608				ins_nr = 0;
4609			}
4610			ret = btrfs_next_leaf(root, path);
4611			if (ret < 0)
4612				return ret;
4613			else if (ret > 0)
4614				break;
4615			continue;
4616		}
4617
4618		btrfs_item_key_to_cpu(leaf, &key, slot);
4619		if (key.objectid != ino || key.type != BTRFS_XATTR_ITEM_KEY)
4620			break;
4621
4622		if (ins_nr == 0)
4623			start_slot = slot;
4624		ins_nr++;
4625		path->slots[0]++;
 
4626		cond_resched();
4627	}
4628	if (ins_nr > 0) {
4629		u64 last_extent = 0;
4630
4631		ret = copy_items(trans, inode, dst_path, path,
4632				 &last_extent, start_slot,
4633				 ins_nr, 1, 0);
4634		/* can't be 1, extent items aren't processed */
4635		ASSERT(ret <= 0);
4636		if (ret < 0)
4637			return ret;
4638	}
4639
 
 
 
4640	return 0;
4641}
4642
4643/*
4644 * If the no holes feature is enabled we need to make sure any hole between the
4645 * last extent and the i_size of our inode is explicitly marked in the log. This
4646 * is to make sure that doing something like:
4647 *
4648 *      1) create file with 128Kb of data
4649 *      2) truncate file to 64Kb
4650 *      3) truncate file to 256Kb
4651 *      4) fsync file
4652 *      5) <crash/power failure>
4653 *      6) mount fs and trigger log replay
4654 *
4655 * Will give us a file with a size of 256Kb, the first 64Kb of data match what
4656 * the file had in its first 64Kb of data at step 1 and the last 192Kb of the
4657 * file correspond to a hole. The presence of explicit holes in a log tree is
4658 * what guarantees that log replay will remove/adjust file extent items in the
4659 * fs/subvol tree.
4660 *
4661 * Here we do not need to care about holes between extents, that is already done
4662 * by copy_items(). We also only need to do this in the full sync path, where we
4663 * lookup for extents from the fs/subvol tree only. In the fast path case, we
4664 * lookup the list of modified extent maps and if any represents a hole, we
4665 * insert a corresponding extent representing a hole in the log tree.
4666 */
4667static int btrfs_log_trailing_hole(struct btrfs_trans_handle *trans,
4668				   struct btrfs_root *root,
4669				   struct btrfs_inode *inode,
4670				   struct btrfs_path *path)
4671{
4672	struct btrfs_fs_info *fs_info = root->fs_info;
4673	int ret;
4674	struct btrfs_key key;
4675	u64 hole_start;
4676	u64 hole_size;
4677	struct extent_buffer *leaf;
4678	struct btrfs_root *log = root->log_root;
4679	const u64 ino = btrfs_ino(inode);
4680	const u64 i_size = i_size_read(&inode->vfs_inode);
 
 
4681
4682	if (!btrfs_fs_incompat(fs_info, NO_HOLES))
4683		return 0;
4684
4685	key.objectid = ino;
4686	key.type = BTRFS_EXTENT_DATA_KEY;
4687	key.offset = (u64)-1;
4688
4689	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4690	ASSERT(ret != 0);
4691	if (ret < 0)
4692		return ret;
4693
4694	ASSERT(path->slots[0] > 0);
4695	path->slots[0]--;
4696	leaf = path->nodes[0];
4697	btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
4698
4699	if (key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY) {
4700		/* inode does not have any extents */
4701		hole_start = 0;
4702		hole_size = i_size;
4703	} else {
4704		struct btrfs_file_extent_item *extent;
4705		u64 len;
 
 
 
4706
4707		/*
4708		 * If there's an extent beyond i_size, an explicit hole was
4709		 * already inserted by copy_items().
4710		 */
4711		if (key.offset >= i_size)
4712			return 0;
4713
4714		extent = btrfs_item_ptr(leaf, path->slots[0],
4715					struct btrfs_file_extent_item);
 
4716
4717		if (btrfs_file_extent_type(leaf, extent) ==
4718		    BTRFS_FILE_EXTENT_INLINE) {
4719			len = btrfs_file_extent_inline_len(leaf,
4720							   path->slots[0],
4721							   extent);
4722			ASSERT(len == i_size ||
4723			       (len == fs_info->sectorsize &&
4724				btrfs_file_extent_compression(leaf, extent) !=
4725				BTRFS_COMPRESS_NONE));
4726			return 0;
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4727		}
4728
4729		len = btrfs_file_extent_num_bytes(leaf, extent);
4730		/* Last extent goes beyond i_size, no need to log a hole. */
4731		if (key.offset + len > i_size)
4732			return 0;
4733		hole_start = key.offset + len;
4734		hole_size = i_size - hole_start;
4735	}
4736	btrfs_release_path(path);
4737
4738	/* Last extent ends at i_size. */
4739	if (hole_size == 0)
4740		return 0;
4741
4742	hole_size = ALIGN(hole_size, fs_info->sectorsize);
4743	ret = btrfs_insert_file_extent(trans, log, ino, hole_start, 0, 0,
4744				       hole_size, 0, hole_size, 0, 0, 0);
4745	return ret;
 
 
 
 
 
 
 
4746}
4747
4748/*
4749 * When we are logging a new inode X, check if it doesn't have a reference that
4750 * matches the reference from some other inode Y created in a past transaction
4751 * and that was renamed in the current transaction. If we don't do this, then at
4752 * log replay time we can lose inode Y (and all its files if it's a directory):
4753 *
4754 * mkdir /mnt/x
4755 * echo "hello world" > /mnt/x/foobar
4756 * sync
4757 * mv /mnt/x /mnt/y
4758 * mkdir /mnt/x                 # or touch /mnt/x
4759 * xfs_io -c fsync /mnt/x
4760 * <power fail>
4761 * mount fs, trigger log replay
4762 *
4763 * After the log replay procedure, we would lose the first directory and all its
4764 * files (file foobar).
4765 * For the case where inode Y is not a directory we simply end up losing it:
4766 *
4767 * echo "123" > /mnt/foo
4768 * sync
4769 * mv /mnt/foo /mnt/bar
4770 * echo "abc" > /mnt/foo
4771 * xfs_io -c fsync /mnt/foo
4772 * <power fail>
4773 *
4774 * We also need this for cases where a snapshot entry is replaced by some other
4775 * entry (file or directory) otherwise we end up with an unreplayable log due to
4776 * attempts to delete the snapshot entry (entry of type BTRFS_ROOT_ITEM_KEY) as
4777 * if it were a regular entry:
4778 *
4779 * mkdir /mnt/x
4780 * btrfs subvolume snapshot /mnt /mnt/x/snap
4781 * btrfs subvolume delete /mnt/x/snap
4782 * rmdir /mnt/x
4783 * mkdir /mnt/x
4784 * fsync /mnt/x or fsync some new file inside it
4785 * <power fail>
4786 *
4787 * The snapshot delete, rmdir of x, mkdir of a new x and the fsync all happen in
4788 * the same transaction.
4789 */
4790static int btrfs_check_ref_name_override(struct extent_buffer *eb,
4791					 const int slot,
4792					 const struct btrfs_key *key,
4793					 struct btrfs_inode *inode,
4794					 u64 *other_ino)
4795{
4796	int ret;
4797	struct btrfs_path *search_path;
4798	char *name = NULL;
4799	u32 name_len = 0;
4800	u32 item_size = btrfs_item_size_nr(eb, slot);
4801	u32 cur_offset = 0;
4802	unsigned long ptr = btrfs_item_ptr_offset(eb, slot);
4803
4804	search_path = btrfs_alloc_path();
4805	if (!search_path)
4806		return -ENOMEM;
4807	search_path->search_commit_root = 1;
4808	search_path->skip_locking = 1;
4809
4810	while (cur_offset < item_size) {
4811		u64 parent;
4812		u32 this_name_len;
4813		u32 this_len;
4814		unsigned long name_ptr;
4815		struct btrfs_dir_item *di;
4816
4817		if (key->type == BTRFS_INODE_REF_KEY) {
4818			struct btrfs_inode_ref *iref;
4819
4820			iref = (struct btrfs_inode_ref *)(ptr + cur_offset);
4821			parent = key->offset;
4822			this_name_len = btrfs_inode_ref_name_len(eb, iref);
4823			name_ptr = (unsigned long)(iref + 1);
4824			this_len = sizeof(*iref) + this_name_len;
4825		} else {
4826			struct btrfs_inode_extref *extref;
4827
4828			extref = (struct btrfs_inode_extref *)(ptr +
4829							       cur_offset);
4830			parent = btrfs_inode_extref_parent(eb, extref);
4831			this_name_len = btrfs_inode_extref_name_len(eb, extref);
4832			name_ptr = (unsigned long)&extref->name;
4833			this_len = sizeof(*extref) + this_name_len;
4834		}
4835
4836		if (this_name_len > name_len) {
4837			char *new_name;
4838
4839			new_name = krealloc(name, this_name_len, GFP_NOFS);
4840			if (!new_name) {
4841				ret = -ENOMEM;
4842				goto out;
4843			}
4844			name_len = this_name_len;
4845			name = new_name;
4846		}
4847
4848		read_extent_buffer(eb, name, name_ptr, this_name_len);
4849		di = btrfs_lookup_dir_item(NULL, inode->root, search_path,
4850				parent, name, this_name_len, 0);
4851		if (di && !IS_ERR(di)) {
4852			struct btrfs_key di_key;
4853
4854			btrfs_dir_item_key_to_cpu(search_path->nodes[0],
4855						  di, &di_key);
4856			if (di_key.type == BTRFS_INODE_ITEM_KEY) {
4857				ret = 1;
4858				*other_ino = di_key.objectid;
 
 
 
 
 
4859			} else {
4860				ret = -EAGAIN;
4861			}
4862			goto out;
4863		} else if (IS_ERR(di)) {
4864			ret = PTR_ERR(di);
4865			goto out;
4866		}
4867		btrfs_release_path(search_path);
4868
4869		cur_offset += this_len;
4870	}
4871	ret = 0;
4872out:
4873	btrfs_free_path(search_path);
4874	kfree(name);
4875	return ret;
4876}
4877
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4878/* log a single inode in the tree log.
4879 * At least one parent directory for this inode must exist in the tree
4880 * or be logged already.
4881 *
4882 * Any items from this inode changed by the current transaction are copied
4883 * to the log tree.  An extra reference is taken on any extents in this
4884 * file, allowing us to avoid a whole pile of corner cases around logging
4885 * blocks that have been removed from the tree.
4886 *
4887 * See LOG_INODE_ALL and related defines for a description of what inode_only
4888 * does.
4889 *
4890 * This handles both files and directories.
4891 */
4892static int btrfs_log_inode(struct btrfs_trans_handle *trans,
4893			   struct btrfs_root *root, struct btrfs_inode *inode,
4894			   int inode_only,
4895			   const loff_t start,
4896			   const loff_t end,
4897			   struct btrfs_log_ctx *ctx)
4898{
4899	struct btrfs_fs_info *fs_info = root->fs_info;
4900	struct btrfs_path *path;
4901	struct btrfs_path *dst_path;
4902	struct btrfs_key min_key;
4903	struct btrfs_key max_key;
4904	struct btrfs_root *log = root->log_root;
4905	LIST_HEAD(logged_list);
4906	u64 last_extent = 0;
4907	int err = 0;
4908	int ret;
4909	int nritems;
4910	int ins_start_slot = 0;
4911	int ins_nr;
4912	bool fast_search = false;
4913	u64 ino = btrfs_ino(inode);
4914	struct extent_map_tree *em_tree = &inode->extent_tree;
4915	u64 logged_isize = 0;
4916	bool need_log_inode_item = true;
4917	bool xattrs_logged = false;
 
4918
4919	path = btrfs_alloc_path();
4920	if (!path)
4921		return -ENOMEM;
4922	dst_path = btrfs_alloc_path();
4923	if (!dst_path) {
4924		btrfs_free_path(path);
4925		return -ENOMEM;
4926	}
4927
4928	min_key.objectid = ino;
4929	min_key.type = BTRFS_INODE_ITEM_KEY;
4930	min_key.offset = 0;
4931
4932	max_key.objectid = ino;
4933
4934
4935	/* today the code can only do partial logging of directories */
4936	if (S_ISDIR(inode->vfs_inode.i_mode) ||
4937	    (!test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
4938		       &inode->runtime_flags) &&
4939	     inode_only >= LOG_INODE_EXISTS))
4940		max_key.type = BTRFS_XATTR_ITEM_KEY;
4941	else
4942		max_key.type = (u8)-1;
4943	max_key.offset = (u64)-1;
4944
4945	/*
4946	 * Only run delayed items if we are a dir or a new file.
4947	 * Otherwise commit the delayed inode only, which is needed in
4948	 * order for the log replay code to mark inodes for link count
4949	 * fixup (create temporary BTRFS_TREE_LOG_FIXUP_OBJECTID items).
 
 
 
 
 
4950	 */
4951	if (S_ISDIR(inode->vfs_inode.i_mode) ||
4952	    inode->generation > fs_info->last_trans_committed)
4953		ret = btrfs_commit_inode_delayed_items(trans, inode);
4954	else
4955		ret = btrfs_commit_inode_delayed_inode(inode);
4956
4957	if (ret) {
4958		btrfs_free_path(path);
4959		btrfs_free_path(dst_path);
4960		return ret;
4961	}
4962
4963	if (inode_only == LOG_OTHER_INODE) {
4964		inode_only = LOG_INODE_EXISTS;
 
 
 
 
4965		mutex_lock_nested(&inode->log_mutex, SINGLE_DEPTH_NESTING);
4966	} else {
4967		mutex_lock(&inode->log_mutex);
4968	}
4969
4970	/*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4971	 * a brute force approach to making sure we get the most uptodate
4972	 * copies of everything.
4973	 */
4974	if (S_ISDIR(inode->vfs_inode.i_mode)) {
4975		int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;
4976
 
4977		if (inode_only == LOG_INODE_EXISTS)
4978			max_key_type = BTRFS_XATTR_ITEM_KEY;
4979		ret = drop_objectid_items(trans, log, path, ino, max_key_type);
4980	} else {
4981		if (inode_only == LOG_INODE_EXISTS) {
4982			/*
4983			 * Make sure the new inode item we write to the log has
4984			 * the same isize as the current one (if it exists).
4985			 * This is necessary to prevent data loss after log
4986			 * replay, and also to prevent doing a wrong expanding
4987			 * truncate - for e.g. create file, write 4K into offset
4988			 * 0, fsync, write 4K into offset 4096, add hard link,
4989			 * fsync some other file (to sync log), power fail - if
4990			 * we use the inode's current i_size, after log replay
4991			 * we get a 8Kb file, with the last 4Kb extent as a hole
4992			 * (zeroes), as if an expanding truncate happened,
4993			 * instead of getting a file of 4Kb only.
4994			 */
4995			err = logged_inode_size(log, inode, path, &logged_isize);
4996			if (err)
4997				goto out_unlock;
4998		}
4999		if (test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
5000			     &inode->runtime_flags)) {
5001			if (inode_only == LOG_INODE_EXISTS) {
5002				max_key.type = BTRFS_XATTR_ITEM_KEY;
5003				ret = drop_objectid_items(trans, log, path, ino,
5004							  max_key.type);
5005			} else {
5006				clear_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
5007					  &inode->runtime_flags);
5008				clear_bit(BTRFS_INODE_COPY_EVERYTHING,
5009					  &inode->runtime_flags);
5010				while(1) {
5011					ret = btrfs_truncate_inode_items(trans,
5012						log, &inode->vfs_inode, 0, 0);
5013					if (ret != -EAGAIN)
5014						break;
5015				}
5016			}
5017		} else if (test_and_clear_bit(BTRFS_INODE_COPY_EVERYTHING,
5018					      &inode->runtime_flags) ||
5019			   inode_only == LOG_INODE_EXISTS) {
5020			if (inode_only == LOG_INODE_ALL)
5021				fast_search = true;
5022			max_key.type = BTRFS_XATTR_ITEM_KEY;
5023			ret = drop_objectid_items(trans, log, path, ino,
5024						  max_key.type);
5025		} else {
5026			if (inode_only == LOG_INODE_ALL)
5027				fast_search = true;
5028			goto log_extents;
5029		}
5030
5031	}
5032	if (ret) {
5033		err = ret;
5034		goto out_unlock;
5035	}
5036
5037	while (1) {
5038		ins_nr = 0;
5039		ret = btrfs_search_forward(root, &min_key,
5040					   path, trans->transid);
5041		if (ret < 0) {
5042			err = ret;
5043			goto out_unlock;
5044		}
5045		if (ret != 0)
5046			break;
5047again:
5048		/* note, ins_nr might be > 0 here, cleanup outside the loop */
5049		if (min_key.objectid != ino)
5050			break;
5051		if (min_key.type > max_key.type)
5052			break;
5053
5054		if (min_key.type == BTRFS_INODE_ITEM_KEY)
5055			need_log_inode_item = false;
5056
5057		if ((min_key.type == BTRFS_INODE_REF_KEY ||
5058		     min_key.type == BTRFS_INODE_EXTREF_KEY) &&
5059		    inode->generation == trans->transid) {
5060			u64 other_ino = 0;
5061
5062			ret = btrfs_check_ref_name_override(path->nodes[0],
5063					path->slots[0], &min_key, inode,
5064					&other_ino);
5065			if (ret < 0) {
5066				err = ret;
5067				goto out_unlock;
5068			} else if (ret > 0 && ctx &&
5069				   other_ino != btrfs_ino(BTRFS_I(ctx->inode))) {
5070				struct btrfs_key inode_key;
5071				struct inode *other_inode;
5072
5073				if (ins_nr > 0) {
5074					ins_nr++;
5075				} else {
5076					ins_nr = 1;
5077					ins_start_slot = path->slots[0];
5078				}
5079				ret = copy_items(trans, inode, dst_path, path,
5080						 &last_extent, ins_start_slot,
5081						 ins_nr, inode_only,
5082						 logged_isize);
5083				if (ret < 0) {
5084					err = ret;
5085					goto out_unlock;
5086				}
5087				ins_nr = 0;
5088				btrfs_release_path(path);
5089				inode_key.objectid = other_ino;
5090				inode_key.type = BTRFS_INODE_ITEM_KEY;
5091				inode_key.offset = 0;
5092				other_inode = btrfs_iget(fs_info->sb,
5093							 &inode_key, root,
5094							 NULL);
5095				/*
5096				 * If the other inode that had a conflicting dir
5097				 * entry was deleted in the current transaction,
5098				 * we don't need to do more work nor fallback to
5099				 * a transaction commit.
5100				 */
5101				if (IS_ERR(other_inode) &&
5102				    PTR_ERR(other_inode) == -ENOENT) {
5103					goto next_key;
5104				} else if (IS_ERR(other_inode)) {
5105					err = PTR_ERR(other_inode);
5106					goto out_unlock;
5107				}
5108				/*
5109				 * We are safe logging the other inode without
5110				 * acquiring its i_mutex as long as we log with
5111				 * the LOG_INODE_EXISTS mode. We're safe against
5112				 * concurrent renames of the other inode as well
5113				 * because during a rename we pin the log and
5114				 * update the log with the new name before we
5115				 * unpin it.
5116				 */
5117				err = btrfs_log_inode(trans, root,
5118						BTRFS_I(other_inode),
5119						LOG_OTHER_INODE, 0, LLONG_MAX,
5120						ctx);
5121				iput(other_inode);
5122				if (err)
5123					goto out_unlock;
5124				else
5125					goto next_key;
5126			}
5127		}
5128
5129		/* Skip xattrs, we log them later with btrfs_log_all_xattrs() */
5130		if (min_key.type == BTRFS_XATTR_ITEM_KEY) {
5131			if (ins_nr == 0)
5132				goto next_slot;
5133			ret = copy_items(trans, inode, dst_path, path,
5134					 &last_extent, ins_start_slot,
5135					 ins_nr, inode_only, logged_isize);
5136			if (ret < 0) {
5137				err = ret;
5138				goto out_unlock;
5139			}
5140			ins_nr = 0;
5141			if (ret) {
5142				btrfs_release_path(path);
5143				continue;
5144			}
5145			goto next_slot;
5146		}
5147
5148		if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
5149			ins_nr++;
5150			goto next_slot;
5151		} else if (!ins_nr) {
5152			ins_start_slot = path->slots[0];
5153			ins_nr = 1;
5154			goto next_slot;
5155		}
5156
5157		ret = copy_items(trans, inode, dst_path, path, &last_extent,
5158				 ins_start_slot, ins_nr, inode_only,
5159				 logged_isize);
5160		if (ret < 0) {
5161			err = ret;
5162			goto out_unlock;
5163		}
5164		if (ret) {
5165			ins_nr = 0;
5166			btrfs_release_path(path);
5167			continue;
5168		}
5169		ins_nr = 1;
5170		ins_start_slot = path->slots[0];
5171next_slot:
5172
5173		nritems = btrfs_header_nritems(path->nodes[0]);
5174		path->slots[0]++;
5175		if (path->slots[0] < nritems) {
5176			btrfs_item_key_to_cpu(path->nodes[0], &min_key,
5177					      path->slots[0]);
5178			goto again;
5179		}
5180		if (ins_nr) {
5181			ret = copy_items(trans, inode, dst_path, path,
5182					 &last_extent, ins_start_slot,
5183					 ins_nr, inode_only, logged_isize);
5184			if (ret < 0) {
5185				err = ret;
5186				goto out_unlock;
5187			}
5188			ret = 0;
5189			ins_nr = 0;
5190		}
5191		btrfs_release_path(path);
5192next_key:
5193		if (min_key.offset < (u64)-1) {
5194			min_key.offset++;
5195		} else if (min_key.type < max_key.type) {
5196			min_key.type++;
5197			min_key.offset = 0;
5198		} else {
5199			break;
5200		}
5201	}
5202	if (ins_nr) {
5203		ret = copy_items(trans, inode, dst_path, path, &last_extent,
5204				 ins_start_slot, ins_nr, inode_only,
5205				 logged_isize);
5206		if (ret < 0) {
5207			err = ret;
5208			goto out_unlock;
5209		}
5210		ret = 0;
5211		ins_nr = 0;
5212	}
5213
5214	btrfs_release_path(path);
5215	btrfs_release_path(dst_path);
5216	err = btrfs_log_all_xattrs(trans, root, inode, path, dst_path);
5217	if (err)
5218		goto out_unlock;
5219	xattrs_logged = true;
5220	if (max_key.type >= BTRFS_EXTENT_DATA_KEY && !fast_search) {
5221		btrfs_release_path(path);
5222		btrfs_release_path(dst_path);
5223		err = btrfs_log_trailing_hole(trans, root, inode, path);
5224		if (err)
5225			goto out_unlock;
5226	}
5227log_extents:
5228	btrfs_release_path(path);
5229	btrfs_release_path(dst_path);
5230	if (need_log_inode_item) {
5231		err = log_inode_item(trans, log, dst_path, inode);
5232		if (!err && !xattrs_logged) {
 
 
 
 
 
 
 
 
 
 
5233			err = btrfs_log_all_xattrs(trans, root, inode, path,
5234						   dst_path);
 
 
5235			btrfs_release_path(path);
5236		}
5237		if (err)
5238			goto out_unlock;
5239	}
5240	if (fast_search) {
5241		ret = btrfs_log_changed_extents(trans, root, inode, dst_path,
5242						&logged_list, ctx, start, end);
5243		if (ret) {
5244			err = ret;
5245			goto out_unlock;
5246		}
5247	} else if (inode_only == LOG_INODE_ALL) {
5248		struct extent_map *em, *n;
5249
5250		write_lock(&em_tree->lock);
5251		/*
5252		 * We can't just remove every em if we're called for a ranged
5253		 * fsync - that is, one that doesn't cover the whole possible
5254		 * file range (0 to LLONG_MAX). This is because we can have
5255		 * em's that fall outside the range we're logging and therefore
5256		 * their ordered operations haven't completed yet
5257		 * (btrfs_finish_ordered_io() not invoked yet). This means we
5258		 * didn't get their respective file extent item in the fs/subvol
5259		 * tree yet, and need to let the next fast fsync (one which
5260		 * consults the list of modified extent maps) find the em so
5261		 * that it logs a matching file extent item and waits for the
5262		 * respective ordered operation to complete (if it's still
5263		 * running).
5264		 *
5265		 * Removing every em outside the range we're logging would make
5266		 * the next fast fsync not log their matching file extent items,
5267		 * therefore making us lose data after a log replay.
5268		 */
5269		list_for_each_entry_safe(em, n, &em_tree->modified_extents,
5270					 list) {
5271			const u64 mod_end = em->mod_start + em->mod_len - 1;
5272
5273			if (em->mod_start >= start && mod_end <= end)
5274				list_del_init(&em->list);
5275		}
5276		write_unlock(&em_tree->lock);
5277	}
5278
5279	if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->vfs_inode.i_mode)) {
5280		ret = log_directory_changes(trans, root, inode, path, dst_path,
5281					ctx);
5282		if (ret) {
5283			err = ret;
5284			goto out_unlock;
5285		}
5286	}
5287
5288	spin_lock(&inode->lock);
5289	inode->logged_trans = trans->transid;
5290	inode->last_log_commit = inode->last_sub_trans;
5291	spin_unlock(&inode->lock);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5292out_unlock:
5293	if (unlikely(err))
5294		btrfs_put_logged_extents(&logged_list);
5295	else
5296		btrfs_submit_logged_extents(&logged_list, log);
5297	mutex_unlock(&inode->log_mutex);
5298
5299	btrfs_free_path(path);
5300	btrfs_free_path(dst_path);
5301	return err;
5302}
5303
5304/*
5305 * Check if we must fallback to a transaction commit when logging an inode.
5306 * This must be called after logging the inode and is used only in the context
5307 * when fsyncing an inode requires the need to log some other inode - in which
5308 * case we can't lock the i_mutex of each other inode we need to log as that
5309 * can lead to deadlocks with concurrent fsync against other inodes (as we can
5310 * log inodes up or down in the hierarchy) or rename operations for example. So
5311 * we take the log_mutex of the inode after we have logged it and then check for
5312 * its last_unlink_trans value - this is safe because any task setting
5313 * last_unlink_trans must take the log_mutex and it must do this before it does
5314 * the actual unlink operation, so if we do this check before a concurrent task
5315 * sets last_unlink_trans it means we've logged a consistent version/state of
5316 * all the inode items, otherwise we are not sure and must do a transaction
5317 * commit (the concurrent task might have only updated last_unlink_trans before
5318 * we logged the inode or it might have also done the unlink).
5319 */
5320static bool btrfs_must_commit_transaction(struct btrfs_trans_handle *trans,
5321					  struct btrfs_inode *inode)
5322{
5323	struct btrfs_fs_info *fs_info = inode->root->fs_info;
5324	bool ret = false;
5325
5326	mutex_lock(&inode->log_mutex);
5327	if (inode->last_unlink_trans > fs_info->last_trans_committed) {
5328		/*
5329		 * Make sure any commits to the log are forced to be full
5330		 * commits.
5331		 */
5332		btrfs_set_log_full_commit(fs_info, trans);
5333		ret = true;
5334	}
5335	mutex_unlock(&inode->log_mutex);
5336
5337	return ret;
5338}
5339
5340/*
5341 * follow the dentry parent pointers up the chain and see if any
5342 * of the directories in it require a full commit before they can
5343 * be logged.  Returns zero if nothing special needs to be done or 1 if
5344 * a full commit is required.
5345 */
5346static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
5347					       struct btrfs_inode *inode,
5348					       struct dentry *parent,
5349					       struct super_block *sb,
5350					       u64 last_committed)
5351{
5352	int ret = 0;
5353	struct dentry *old_parent = NULL;
5354	struct btrfs_inode *orig_inode = inode;
5355
5356	/*
5357	 * for regular files, if its inode is already on disk, we don't
5358	 * have to worry about the parents at all.  This is because
5359	 * we can use the last_unlink_trans field to record renames
5360	 * and other fun in this file.
5361	 */
5362	if (S_ISREG(inode->vfs_inode.i_mode) &&
5363	    inode->generation <= last_committed &&
5364	    inode->last_unlink_trans <= last_committed)
5365		goto out;
 
 
5366
5367	if (!S_ISDIR(inode->vfs_inode.i_mode)) {
5368		if (!parent || d_really_is_negative(parent) || sb != parent->d_sb)
5369			goto out;
5370		inode = BTRFS_I(d_inode(parent));
5371	}
5372
5373	while (1) {
5374		/*
5375		 * If we are logging a directory then we start with our inode,
5376		 * not our parent's inode, so we need to skip setting the
5377		 * logged_trans so that further down in the log code we don't
5378		 * think this inode has already been logged.
5379		 */
5380		if (inode != orig_inode)
5381			inode->logged_trans = trans->transid;
5382		smp_mb();
5383
5384		if (btrfs_must_commit_transaction(trans, inode)) {
5385			ret = 1;
5386			break;
5387		}
5388
5389		if (!parent || d_really_is_negative(parent) || sb != parent->d_sb)
5390			break;
5391
5392		if (IS_ROOT(parent)) {
5393			inode = BTRFS_I(d_inode(parent));
5394			if (btrfs_must_commit_transaction(trans, inode))
5395				ret = 1;
5396			break;
5397		}
5398
5399		parent = dget_parent(parent);
5400		dput(old_parent);
5401		old_parent = parent;
5402		inode = BTRFS_I(d_inode(parent));
5403
5404	}
5405	dput(old_parent);
5406out:
5407	return ret;
5408}
5409
5410struct btrfs_dir_list {
5411	u64 ino;
5412	struct list_head list;
5413};
5414
5415/*
5416 * Log the inodes of the new dentries of a directory. See log_dir_items() for
5417 * details about the why it is needed.
5418 * This is a recursive operation - if an existing dentry corresponds to a
5419 * directory, that directory's new entries are logged too (same behaviour as
5420 * ext3/4, xfs, f2fs, reiserfs, nilfs2). Note that when logging the inodes
5421 * the dentries point to we do not lock their i_mutex, otherwise lockdep
5422 * complains about the following circular lock dependency / possible deadlock:
5423 *
5424 *        CPU0                                        CPU1
5425 *        ----                                        ----
5426 * lock(&type->i_mutex_dir_key#3/2);
5427 *                                            lock(sb_internal#2);
5428 *                                            lock(&type->i_mutex_dir_key#3/2);
5429 * lock(&sb->s_type->i_mutex_key#14);
5430 *
5431 * Where sb_internal is the lock (a counter that works as a lock) acquired by
5432 * sb_start_intwrite() in btrfs_start_transaction().
5433 * Not locking i_mutex of the inodes is still safe because:
5434 *
5435 * 1) For regular files we log with a mode of LOG_INODE_EXISTS. It's possible
5436 *    that while logging the inode new references (names) are added or removed
5437 *    from the inode, leaving the logged inode item with a link count that does
5438 *    not match the number of logged inode reference items. This is fine because
5439 *    at log replay time we compute the real number of links and correct the
5440 *    link count in the inode item (see replay_one_buffer() and
5441 *    link_to_fixup_dir());
5442 *
5443 * 2) For directories we log with a mode of LOG_INODE_ALL. It's possible that
5444 *    while logging the inode's items new items with keys BTRFS_DIR_ITEM_KEY and
5445 *    BTRFS_DIR_INDEX_KEY are added to fs/subvol tree and the logged inode item
5446 *    has a size that doesn't match the sum of the lengths of all the logged
5447 *    names. This does not result in a problem because if a dir_item key is
5448 *    logged but its matching dir_index key is not logged, at log replay time we
5449 *    don't use it to replay the respective name (see replay_one_name()). On the
5450 *    other hand if only the dir_index key ends up being logged, the respective
5451 *    name is added to the fs/subvol tree with both the dir_item and dir_index
5452 *    keys created (see replay_one_name()).
5453 *    The directory's inode item with a wrong i_size is not a problem as well,
5454 *    since we don't use it at log replay time to set the i_size in the inode
5455 *    item of the fs/subvol tree (see overwrite_item()).
5456 */
5457static int log_new_dir_dentries(struct btrfs_trans_handle *trans,
5458				struct btrfs_root *root,
5459				struct btrfs_inode *start_inode,
5460				struct btrfs_log_ctx *ctx)
5461{
5462	struct btrfs_fs_info *fs_info = root->fs_info;
5463	struct btrfs_root *log = root->log_root;
5464	struct btrfs_path *path;
5465	LIST_HEAD(dir_list);
5466	struct btrfs_dir_list *dir_elem;
5467	int ret = 0;
5468
5469	path = btrfs_alloc_path();
5470	if (!path)
5471		return -ENOMEM;
5472
5473	dir_elem = kmalloc(sizeof(*dir_elem), GFP_NOFS);
5474	if (!dir_elem) {
5475		btrfs_free_path(path);
5476		return -ENOMEM;
5477	}
5478	dir_elem->ino = btrfs_ino(start_inode);
5479	list_add_tail(&dir_elem->list, &dir_list);
5480
5481	while (!list_empty(&dir_list)) {
5482		struct extent_buffer *leaf;
5483		struct btrfs_key min_key;
5484		int nritems;
5485		int i;
5486
5487		dir_elem = list_first_entry(&dir_list, struct btrfs_dir_list,
5488					    list);
5489		if (ret)
5490			goto next_dir_inode;
5491
5492		min_key.objectid = dir_elem->ino;
5493		min_key.type = BTRFS_DIR_ITEM_KEY;
5494		min_key.offset = 0;
5495again:
5496		btrfs_release_path(path);
5497		ret = btrfs_search_forward(log, &min_key, path, trans->transid);
5498		if (ret < 0) {
5499			goto next_dir_inode;
5500		} else if (ret > 0) {
5501			ret = 0;
5502			goto next_dir_inode;
5503		}
5504
5505process_leaf:
5506		leaf = path->nodes[0];
5507		nritems = btrfs_header_nritems(leaf);
5508		for (i = path->slots[0]; i < nritems; i++) {
5509			struct btrfs_dir_item *di;
5510			struct btrfs_key di_key;
5511			struct inode *di_inode;
5512			struct btrfs_dir_list *new_dir_elem;
5513			int log_mode = LOG_INODE_EXISTS;
5514			int type;
5515
5516			btrfs_item_key_to_cpu(leaf, &min_key, i);
5517			if (min_key.objectid != dir_elem->ino ||
5518			    min_key.type != BTRFS_DIR_ITEM_KEY)
5519				goto next_dir_inode;
5520
5521			di = btrfs_item_ptr(leaf, i, struct btrfs_dir_item);
5522			type = btrfs_dir_type(leaf, di);
5523			if (btrfs_dir_transid(leaf, di) < trans->transid &&
5524			    type != BTRFS_FT_DIR)
5525				continue;
5526			btrfs_dir_item_key_to_cpu(leaf, di, &di_key);
5527			if (di_key.type == BTRFS_ROOT_ITEM_KEY)
5528				continue;
5529
5530			btrfs_release_path(path);
5531			di_inode = btrfs_iget(fs_info->sb, &di_key, root, NULL);
5532			if (IS_ERR(di_inode)) {
5533				ret = PTR_ERR(di_inode);
5534				goto next_dir_inode;
5535			}
5536
5537			if (btrfs_inode_in_log(BTRFS_I(di_inode), trans->transid)) {
5538				iput(di_inode);
5539				break;
5540			}
5541
5542			ctx->log_new_dentries = false;
5543			if (type == BTRFS_FT_DIR || type == BTRFS_FT_SYMLINK)
5544				log_mode = LOG_INODE_ALL;
5545			ret = btrfs_log_inode(trans, root, BTRFS_I(di_inode),
5546					      log_mode, 0, LLONG_MAX, ctx);
5547			if (!ret &&
5548			    btrfs_must_commit_transaction(trans, BTRFS_I(di_inode)))
5549				ret = 1;
5550			iput(di_inode);
5551			if (ret)
5552				goto next_dir_inode;
5553			if (ctx->log_new_dentries) {
5554				new_dir_elem = kmalloc(sizeof(*new_dir_elem),
5555						       GFP_NOFS);
5556				if (!new_dir_elem) {
5557					ret = -ENOMEM;
5558					goto next_dir_inode;
5559				}
5560				new_dir_elem->ino = di_key.objectid;
5561				list_add_tail(&new_dir_elem->list, &dir_list);
5562			}
5563			break;
5564		}
5565		if (i == nritems) {
5566			ret = btrfs_next_leaf(log, path);
5567			if (ret < 0) {
5568				goto next_dir_inode;
5569			} else if (ret > 0) {
5570				ret = 0;
5571				goto next_dir_inode;
5572			}
5573			goto process_leaf;
5574		}
5575		if (min_key.offset < (u64)-1) {
5576			min_key.offset++;
5577			goto again;
5578		}
5579next_dir_inode:
5580		list_del(&dir_elem->list);
5581		kfree(dir_elem);
5582	}
5583
5584	btrfs_free_path(path);
5585	return ret;
5586}
5587
5588static int btrfs_log_all_parents(struct btrfs_trans_handle *trans,
5589				 struct btrfs_inode *inode,
5590				 struct btrfs_log_ctx *ctx)
5591{
5592	struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb);
5593	int ret;
5594	struct btrfs_path *path;
5595	struct btrfs_key key;
5596	struct btrfs_root *root = inode->root;
5597	const u64 ino = btrfs_ino(inode);
5598
5599	path = btrfs_alloc_path();
5600	if (!path)
5601		return -ENOMEM;
5602	path->skip_locking = 1;
5603	path->search_commit_root = 1;
5604
5605	key.objectid = ino;
5606	key.type = BTRFS_INODE_REF_KEY;
5607	key.offset = 0;
5608	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
5609	if (ret < 0)
5610		goto out;
5611
5612	while (true) {
5613		struct extent_buffer *leaf = path->nodes[0];
5614		int slot = path->slots[0];
5615		u32 cur_offset = 0;
5616		u32 item_size;
5617		unsigned long ptr;
5618
5619		if (slot >= btrfs_header_nritems(leaf)) {
5620			ret = btrfs_next_leaf(root, path);
5621			if (ret < 0)
5622				goto out;
5623			else if (ret > 0)
5624				break;
5625			continue;
5626		}
5627
5628		btrfs_item_key_to_cpu(leaf, &key, slot);
5629		/* BTRFS_INODE_EXTREF_KEY is BTRFS_INODE_REF_KEY + 1 */
5630		if (key.objectid != ino || key.type > BTRFS_INODE_EXTREF_KEY)
5631			break;
5632
5633		item_size = btrfs_item_size_nr(leaf, slot);
5634		ptr = btrfs_item_ptr_offset(leaf, slot);
5635		while (cur_offset < item_size) {
5636			struct btrfs_key inode_key;
5637			struct inode *dir_inode;
5638
5639			inode_key.type = BTRFS_INODE_ITEM_KEY;
5640			inode_key.offset = 0;
5641
5642			if (key.type == BTRFS_INODE_EXTREF_KEY) {
5643				struct btrfs_inode_extref *extref;
5644
5645				extref = (struct btrfs_inode_extref *)
5646					(ptr + cur_offset);
5647				inode_key.objectid = btrfs_inode_extref_parent(
5648					leaf, extref);
5649				cur_offset += sizeof(*extref);
5650				cur_offset += btrfs_inode_extref_name_len(leaf,
5651					extref);
5652			} else {
5653				inode_key.objectid = key.offset;
5654				cur_offset = item_size;
5655			}
5656
5657			dir_inode = btrfs_iget(fs_info->sb, &inode_key,
5658					       root, NULL);
5659			/* If parent inode was deleted, skip it. */
5660			if (IS_ERR(dir_inode))
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5661				continue;
 
5662
5663			if (ctx)
5664				ctx->log_new_dentries = false;
5665			ret = btrfs_log_inode(trans, root, BTRFS_I(dir_inode),
5666					      LOG_INODE_ALL, 0, LLONG_MAX, ctx);
5667			if (!ret &&
5668			    btrfs_must_commit_transaction(trans, BTRFS_I(dir_inode)))
5669				ret = 1;
5670			if (!ret && ctx && ctx->log_new_dentries)
5671				ret = log_new_dir_dentries(trans, root,
5672						   BTRFS_I(dir_inode), ctx);
5673			iput(dir_inode);
5674			if (ret)
5675				goto out;
5676		}
5677		path->slots[0]++;
5678	}
5679	ret = 0;
5680out:
5681	btrfs_free_path(path);
5682	return ret;
5683}
5684
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5685/*
5686 * helper function around btrfs_log_inode to make sure newly created
5687 * parent directories also end up in the log.  A minimal inode and backref
5688 * only logging is done of any parent directories that are older than
5689 * the last committed transaction
5690 */
5691static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
5692				  struct btrfs_inode *inode,
5693				  struct dentry *parent,
5694				  const loff_t start,
5695				  const loff_t end,
5696				  int inode_only,
5697				  struct btrfs_log_ctx *ctx)
5698{
5699	struct btrfs_root *root = inode->root;
5700	struct btrfs_fs_info *fs_info = root->fs_info;
5701	struct super_block *sb;
5702	struct dentry *old_parent = NULL;
5703	int ret = 0;
5704	u64 last_committed = fs_info->last_trans_committed;
5705	bool log_dentries = false;
5706	struct btrfs_inode *orig_inode = inode;
5707
5708	sb = inode->vfs_inode.i_sb;
5709
5710	if (btrfs_test_opt(fs_info, NOTREELOG)) {
5711		ret = 1;
5712		goto end_no_trans;
5713	}
5714
5715	/*
5716	 * The prev transaction commit doesn't complete, we need do
5717	 * full commit by ourselves.
5718	 */
5719	if (fs_info->last_trans_log_full_commit >
5720	    fs_info->last_trans_committed) {
5721		ret = 1;
5722		goto end_no_trans;
5723	}
5724
5725	if (btrfs_root_refs(&root->root_item) == 0) {
5726		ret = 1;
5727		goto end_no_trans;
5728	}
5729
5730	ret = check_parent_dirs_for_sync(trans, inode, parent, sb,
5731			last_committed);
5732	if (ret)
5733		goto end_no_trans;
5734
5735	if (btrfs_inode_in_log(inode, trans->transid)) {
 
 
5736		ret = BTRFS_NO_LOG_SYNC;
5737		goto end_no_trans;
5738	}
5739
5740	ret = start_log_trans(trans, root, ctx);
5741	if (ret)
5742		goto end_no_trans;
5743
5744	ret = btrfs_log_inode(trans, root, inode, inode_only, start, end, ctx);
5745	if (ret)
5746		goto end_trans;
5747
5748	/*
5749	 * for regular files, if its inode is already on disk, we don't
5750	 * have to worry about the parents at all.  This is because
5751	 * we can use the last_unlink_trans field to record renames
5752	 * and other fun in this file.
5753	 */
5754	if (S_ISREG(inode->vfs_inode.i_mode) &&
5755	    inode->generation <= last_committed &&
5756	    inode->last_unlink_trans <= last_committed) {
5757		ret = 0;
5758		goto end_trans;
5759	}
5760
5761	if (S_ISDIR(inode->vfs_inode.i_mode) && ctx && ctx->log_new_dentries)
5762		log_dentries = true;
5763
5764	/*
5765	 * On unlink we must make sure all our current and old parent directory
5766	 * inodes are fully logged. This is to prevent leaving dangling
5767	 * directory index entries in directories that were our parents but are
5768	 * not anymore. Not doing this results in old parent directory being
5769	 * impossible to delete after log replay (rmdir will always fail with
5770	 * error -ENOTEMPTY).
5771	 *
5772	 * Example 1:
5773	 *
5774	 * mkdir testdir
5775	 * touch testdir/foo
5776	 * ln testdir/foo testdir/bar
5777	 * sync
5778	 * unlink testdir/bar
5779	 * xfs_io -c fsync testdir/foo
5780	 * <power failure>
5781	 * mount fs, triggers log replay
5782	 *
5783	 * If we don't log the parent directory (testdir), after log replay the
5784	 * directory still has an entry pointing to the file inode using the bar
5785	 * name, but a matching BTRFS_INODE_[REF|EXTREF]_KEY does not exist and
5786	 * the file inode has a link count of 1.
5787	 *
5788	 * Example 2:
5789	 *
5790	 * mkdir testdir
5791	 * touch foo
5792	 * ln foo testdir/foo2
5793	 * ln foo testdir/foo3
5794	 * sync
5795	 * unlink testdir/foo3
5796	 * xfs_io -c fsync foo
5797	 * <power failure>
5798	 * mount fs, triggers log replay
5799	 *
5800	 * Similar as the first example, after log replay the parent directory
5801	 * testdir still has an entry pointing to the inode file with name foo3
5802	 * but the file inode does not have a matching BTRFS_INODE_REF_KEY item
5803	 * and has a link count of 2.
5804	 */
5805	if (inode->last_unlink_trans > last_committed) {
5806		ret = btrfs_log_all_parents(trans, orig_inode, ctx);
5807		if (ret)
5808			goto end_trans;
5809	}
5810
5811	while (1) {
5812		if (!parent || d_really_is_negative(parent) || sb != parent->d_sb)
5813			break;
5814
5815		inode = BTRFS_I(d_inode(parent));
5816		if (root != inode->root)
5817			break;
5818
5819		if (inode->generation > last_committed) {
5820			ret = btrfs_log_inode(trans, root, inode,
5821					LOG_INODE_EXISTS, 0, LLONG_MAX, ctx);
5822			if (ret)
5823				goto end_trans;
5824		}
5825		if (IS_ROOT(parent))
5826			break;
5827
5828		parent = dget_parent(parent);
5829		dput(old_parent);
5830		old_parent = parent;
5831	}
5832	if (log_dentries)
5833		ret = log_new_dir_dentries(trans, root, orig_inode, ctx);
5834	else
5835		ret = 0;
5836end_trans:
5837	dput(old_parent);
5838	if (ret < 0) {
5839		btrfs_set_log_full_commit(fs_info, trans);
5840		ret = 1;
5841	}
5842
5843	if (ret)
5844		btrfs_remove_log_ctx(root, ctx);
5845	btrfs_end_log_trans(root);
5846end_no_trans:
5847	return ret;
5848}
5849
5850/*
5851 * it is not safe to log dentry if the chunk root has added new
5852 * chunks.  This returns 0 if the dentry was logged, and 1 otherwise.
5853 * If this returns 1, you must commit the transaction to safely get your
5854 * data on disk.
5855 */
5856int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
5857			  struct dentry *dentry,
5858			  const loff_t start,
5859			  const loff_t end,
5860			  struct btrfs_log_ctx *ctx)
5861{
5862	struct dentry *parent = dget_parent(dentry);
5863	int ret;
5864
5865	ret = btrfs_log_inode_parent(trans, BTRFS_I(d_inode(dentry)), parent,
5866				     start, end, LOG_INODE_ALL, ctx);
5867	dput(parent);
5868
5869	return ret;
5870}
5871
5872/*
5873 * should be called during mount to recover any replay any log trees
5874 * from the FS
5875 */
5876int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
5877{
5878	int ret;
5879	struct btrfs_path *path;
5880	struct btrfs_trans_handle *trans;
5881	struct btrfs_key key;
5882	struct btrfs_key found_key;
5883	struct btrfs_key tmp_key;
5884	struct btrfs_root *log;
5885	struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
5886	struct walk_control wc = {
5887		.process_func = process_one_buffer,
5888		.stage = 0,
5889	};
5890
5891	path = btrfs_alloc_path();
5892	if (!path)
5893		return -ENOMEM;
5894
5895	set_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags);
5896
5897	trans = btrfs_start_transaction(fs_info->tree_root, 0);
5898	if (IS_ERR(trans)) {
5899		ret = PTR_ERR(trans);
5900		goto error;
5901	}
5902
5903	wc.trans = trans;
5904	wc.pin = 1;
5905
5906	ret = walk_log_tree(trans, log_root_tree, &wc);
5907	if (ret) {
5908		btrfs_handle_fs_error(fs_info, ret,
5909			"Failed to pin buffers while recovering log root tree.");
5910		goto error;
5911	}
5912
5913again:
5914	key.objectid = BTRFS_TREE_LOG_OBJECTID;
5915	key.offset = (u64)-1;
5916	key.type = BTRFS_ROOT_ITEM_KEY;
5917
5918	while (1) {
5919		ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
5920
5921		if (ret < 0) {
5922			btrfs_handle_fs_error(fs_info, ret,
5923				    "Couldn't find tree log root.");
5924			goto error;
5925		}
5926		if (ret > 0) {
5927			if (path->slots[0] == 0)
5928				break;
5929			path->slots[0]--;
5930		}
5931		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
5932				      path->slots[0]);
5933		btrfs_release_path(path);
5934		if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
5935			break;
5936
5937		log = btrfs_read_fs_root(log_root_tree, &found_key);
5938		if (IS_ERR(log)) {
5939			ret = PTR_ERR(log);
5940			btrfs_handle_fs_error(fs_info, ret,
5941				    "Couldn't read tree log root.");
5942			goto error;
5943		}
5944
5945		tmp_key.objectid = found_key.offset;
5946		tmp_key.type = BTRFS_ROOT_ITEM_KEY;
5947		tmp_key.offset = (u64)-1;
5948
5949		wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
5950		if (IS_ERR(wc.replay_dest)) {
5951			ret = PTR_ERR(wc.replay_dest);
5952			free_extent_buffer(log->node);
5953			free_extent_buffer(log->commit_root);
5954			kfree(log);
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5955			btrfs_handle_fs_error(fs_info, ret,
5956				"Couldn't read target root for tree log recovery.");
5957			goto error;
5958		}
5959
5960		wc.replay_dest->log_root = log;
5961		btrfs_record_root_in_trans(trans, wc.replay_dest);
5962		ret = walk_log_tree(trans, log, &wc);
 
 
 
 
 
5963
5964		if (!ret && wc.stage == LOG_WALK_REPLAY_ALL) {
5965			ret = fixup_inode_link_counts(trans, wc.replay_dest,
5966						      path);
5967		}
5968
5969		if (!ret && wc.stage == LOG_WALK_REPLAY_ALL) {
5970			struct btrfs_root *root = wc.replay_dest;
5971
5972			btrfs_release_path(path);
5973
5974			/*
5975			 * We have just replayed everything, and the highest
5976			 * objectid of fs roots probably has changed in case
5977			 * some inode_item's got replayed.
5978			 *
5979			 * root->objectid_mutex is not acquired as log replay
5980			 * could only happen during mount.
5981			 */
5982			ret = btrfs_find_highest_objectid(root,
5983						  &root->highest_objectid);
5984		}
5985
5986		key.offset = found_key.offset - 1;
5987		wc.replay_dest->log_root = NULL;
5988		free_extent_buffer(log->node);
5989		free_extent_buffer(log->commit_root);
5990		kfree(log);
5991
5992		if (ret)
5993			goto error;
5994
5995		if (found_key.offset == 0)
5996			break;
 
5997	}
5998	btrfs_release_path(path);
5999
6000	/* step one is to pin it all, step two is to replay just inodes */
6001	if (wc.pin) {
6002		wc.pin = 0;
6003		wc.process_func = replay_one_buffer;
6004		wc.stage = LOG_WALK_REPLAY_INODES;
6005		goto again;
6006	}
6007	/* step three is to replay everything */
6008	if (wc.stage < LOG_WALK_REPLAY_ALL) {
6009		wc.stage++;
6010		goto again;
6011	}
6012
6013	btrfs_free_path(path);
6014
6015	/* step 4: commit the transaction, which also unpins the blocks */
6016	ret = btrfs_commit_transaction(trans);
6017	if (ret)
6018		return ret;
6019
6020	free_extent_buffer(log_root_tree->node);
6021	log_root_tree->log_root = NULL;
6022	clear_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags);
6023	kfree(log_root_tree);
6024
6025	return 0;
6026error:
6027	if (wc.trans)
6028		btrfs_end_transaction(wc.trans);
 
6029	btrfs_free_path(path);
6030	return ret;
6031}
6032
6033/*
6034 * there are some corner cases where we want to force a full
6035 * commit instead of allowing a directory to be logged.
6036 *
6037 * They revolve around files there were unlinked from the directory, and
6038 * this function updates the parent directory so that a full commit is
6039 * properly done if it is fsync'd later after the unlinks are done.
6040 *
6041 * Must be called before the unlink operations (updates to the subvolume tree,
6042 * inodes, etc) are done.
6043 */
6044void btrfs_record_unlink_dir(struct btrfs_trans_handle *trans,
6045			     struct btrfs_inode *dir, struct btrfs_inode *inode,
6046			     int for_rename)
6047{
6048	/*
6049	 * when we're logging a file, if it hasn't been renamed
6050	 * or unlinked, and its inode is fully committed on disk,
6051	 * we don't have to worry about walking up the directory chain
6052	 * to log its parents.
6053	 *
6054	 * So, we use the last_unlink_trans field to put this transid
6055	 * into the file.  When the file is logged we check it and
6056	 * don't log the parents if the file is fully on disk.
6057	 */
6058	mutex_lock(&inode->log_mutex);
6059	inode->last_unlink_trans = trans->transid;
6060	mutex_unlock(&inode->log_mutex);
6061
6062	/*
6063	 * if this directory was already logged any new
6064	 * names for this file/dir will get recorded
6065	 */
6066	smp_mb();
6067	if (dir->logged_trans == trans->transid)
6068		return;
6069
6070	/*
6071	 * if the inode we're about to unlink was logged,
6072	 * the log will be properly updated for any new names
6073	 */
6074	if (inode->logged_trans == trans->transid)
6075		return;
6076
6077	/*
6078	 * when renaming files across directories, if the directory
6079	 * there we're unlinking from gets fsync'd later on, there's
6080	 * no way to find the destination directory later and fsync it
6081	 * properly.  So, we have to be conservative and force commits
6082	 * so the new name gets discovered.
6083	 */
6084	if (for_rename)
6085		goto record;
6086
6087	/* we can safely do the unlink without any special recording */
6088	return;
6089
6090record:
6091	mutex_lock(&dir->log_mutex);
6092	dir->last_unlink_trans = trans->transid;
6093	mutex_unlock(&dir->log_mutex);
6094}
6095
6096/*
6097 * Make sure that if someone attempts to fsync the parent directory of a deleted
6098 * snapshot, it ends up triggering a transaction commit. This is to guarantee
6099 * that after replaying the log tree of the parent directory's root we will not
6100 * see the snapshot anymore and at log replay time we will not see any log tree
6101 * corresponding to the deleted snapshot's root, which could lead to replaying
6102 * it after replaying the log tree of the parent directory (which would replay
6103 * the snapshot delete operation).
6104 *
6105 * Must be called before the actual snapshot destroy operation (updates to the
6106 * parent root and tree of tree roots trees, etc) are done.
6107 */
6108void btrfs_record_snapshot_destroy(struct btrfs_trans_handle *trans,
6109				   struct btrfs_inode *dir)
6110{
6111	mutex_lock(&dir->log_mutex);
6112	dir->last_unlink_trans = trans->transid;
6113	mutex_unlock(&dir->log_mutex);
6114}
6115
6116/*
6117 * Call this after adding a new name for a file and it will properly
6118 * update the log to reflect the new name.
6119 *
6120 * It will return zero if all goes well, and it will return 1 if a
6121 * full transaction commit is required.
6122 */
6123int btrfs_log_new_name(struct btrfs_trans_handle *trans,
6124			struct btrfs_inode *inode, struct btrfs_inode *old_dir,
6125			struct dentry *parent)
6126{
6127	struct btrfs_fs_info *fs_info = btrfs_sb(inode->vfs_inode.i_sb);
6128
6129	/*
6130	 * this will force the logging code to walk the dentry chain
6131	 * up for the file
6132	 */
6133	if (!S_ISDIR(inode->vfs_inode.i_mode))
6134		inode->last_unlink_trans = trans->transid;
6135
6136	/*
6137	 * if this inode hasn't been logged and directory we're renaming it
6138	 * from hasn't been logged, we don't need to log it
6139	 */
6140	if (inode->logged_trans <= fs_info->last_trans_committed &&
6141	    (!old_dir || old_dir->logged_trans <= fs_info->last_trans_committed))
6142		return 0;
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6143
6144	return btrfs_log_inode_parent(trans, inode, parent, 0, LLONG_MAX,
6145				      LOG_INODE_EXISTS, NULL);
 
 
 
 
 
 
 
 
6146}
6147