transaction.c 47.2 KB
Newer Older
C
Chris Mason 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

C
Chris Mason 已提交
19
#include <linux/fs.h>
20
#include <linux/slab.h>
C
Chris Mason 已提交
21
#include <linux/sched.h>
22
#include <linux/writeback.h>
23
#include <linux/pagemap.h>
24
#include <linux/blkdev.h>
25
#include <linux/uuid.h>
C
Chris Mason 已提交
26 27 28
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
29
#include "locking.h"
30
#include "tree-log.h"
31
#include "inode-map.h"
32
#include "volumes.h"
33
#include "dev-replace.h"
C
Chris Mason 已提交
34

35 36
#define BTRFS_ROOT_TRANS_TAG 0

37
void put_transaction(struct btrfs_transaction *transaction)
C
Chris Mason 已提交
38
{
39 40
	WARN_ON(atomic_read(&transaction->use_count) == 0);
	if (atomic_dec_and_test(&transaction->use_count)) {
J
Josef Bacik 已提交
41
		BUG_ON(!list_empty(&transaction->list));
42
		WARN_ON(transaction->delayed_refs.root.rb_node);
C
Chris Mason 已提交
43 44
		memset(transaction, 0, sizeof(*transaction));
		kmem_cache_free(btrfs_transaction_cachep, transaction);
C
Chris Mason 已提交
45
	}
C
Chris Mason 已提交
46 47
}

J
Josef Bacik 已提交
48 49 50 51 52 53
static noinline void switch_commit_root(struct btrfs_root *root)
{
	free_extent_buffer(root->commit_root);
	root->commit_root = btrfs_root_node(root);
}

C
Chris Mason 已提交
54 55 56
/*
 * either allocate a new transaction or hop into the existing one
 */
57
static noinline int join_transaction(struct btrfs_root *root, int type)
C
Chris Mason 已提交
58 59
{
	struct btrfs_transaction *cur_trans;
60
	struct btrfs_fs_info *fs_info = root->fs_info;
J
Josef Bacik 已提交
61

62
	spin_lock(&fs_info->trans_lock);
63
loop:
64
	/* The file system has been taken offline. No new transactions. */
65 66
	if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
		spin_unlock(&fs_info->trans_lock);
67 68 69
		return -EROFS;
	}

70
	if (fs_info->trans_no_join) {
71 72 73 74 75 76 77
		/* 
		 * If we are JOIN_NOLOCK we're already committing a current
		 * transaction, we just need a handle to deal with something
		 * when committing the transaction, such as inode cache and
		 * space cache. It is a special case.
		 */
		if (type != TRANS_JOIN_NOLOCK) {
78
			spin_unlock(&fs_info->trans_lock);
J
Josef Bacik 已提交
79 80 81 82
			return -EBUSY;
		}
	}

83
	cur_trans = fs_info->running_transaction;
J
Josef Bacik 已提交
84
	if (cur_trans) {
85
		if (cur_trans->aborted) {
86
			spin_unlock(&fs_info->trans_lock);
87
			return cur_trans->aborted;
88
		}
J
Josef Bacik 已提交
89
		atomic_inc(&cur_trans->use_count);
90
		atomic_inc(&cur_trans->num_writers);
91
		cur_trans->num_joined++;
92
		spin_unlock(&fs_info->trans_lock);
J
Josef Bacik 已提交
93
		return 0;
C
Chris Mason 已提交
94
	}
95
	spin_unlock(&fs_info->trans_lock);
J
Josef Bacik 已提交
96

97 98 99 100 101 102 103
	/*
	 * If we are ATTACH, we just want to catch the current transaction,
	 * and commit it. If there is no transaction, just return ENOENT.
	 */
	if (type == TRANS_ATTACH)
		return -ENOENT;

J
Josef Bacik 已提交
104 105 106
	cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS);
	if (!cur_trans)
		return -ENOMEM;
107

108 109
	spin_lock(&fs_info->trans_lock);
	if (fs_info->running_transaction) {
110 111 112 113
		/*
		 * someone started a transaction after we unlocked.  Make sure
		 * to redo the trans_no_join checks above
		 */
J
Josef Bacik 已提交
114
		kmem_cache_free(btrfs_transaction_cachep, cur_trans);
115
		cur_trans = fs_info->running_transaction;
116
		goto loop;
117 118
	} else if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
		spin_unlock(&fs_info->trans_lock);
119 120
		kmem_cache_free(btrfs_transaction_cachep, cur_trans);
		return -EROFS;
C
Chris Mason 已提交
121
	}
122

J
Josef Bacik 已提交
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142
	atomic_set(&cur_trans->num_writers, 1);
	cur_trans->num_joined = 0;
	init_waitqueue_head(&cur_trans->writer_wait);
	init_waitqueue_head(&cur_trans->commit_wait);
	cur_trans->in_commit = 0;
	cur_trans->blocked = 0;
	/*
	 * One for this trans handle, one so it will live on until we
	 * commit the transaction.
	 */
	atomic_set(&cur_trans->use_count, 2);
	cur_trans->commit_done = 0;
	cur_trans->start_time = get_seconds();

	cur_trans->delayed_refs.root = RB_ROOT;
	cur_trans->delayed_refs.num_entries = 0;
	cur_trans->delayed_refs.num_heads_ready = 0;
	cur_trans->delayed_refs.num_heads = 0;
	cur_trans->delayed_refs.flushing = 0;
	cur_trans->delayed_refs.run_delayed_start = 0;
143 144 145 146 147 148

	/*
	 * although the tree mod log is per file system and not per transaction,
	 * the log must never go across transaction boundaries.
	 */
	smp_mb();
J
Julia Lawall 已提交
149 150
	if (!list_empty(&fs_info->tree_mod_seq_list))
		WARN(1, KERN_ERR "btrfs: tree_mod_seq_list not empty when "
151
			"creating a fresh transaction\n");
J
Julia Lawall 已提交
152 153
	if (!RB_EMPTY_ROOT(&fs_info->tree_mod_log))
		WARN(1, KERN_ERR "btrfs: tree_mod_log rb tree not empty when "
154 155 156
			"creating a fresh transaction\n");
	atomic_set(&fs_info->tree_mod_seq, 0);

J
Josef Bacik 已提交
157 158 159 160
	spin_lock_init(&cur_trans->commit_lock);
	spin_lock_init(&cur_trans->delayed_refs.lock);

	INIT_LIST_HEAD(&cur_trans->pending_snapshots);
161
	list_add_tail(&cur_trans->list, &fs_info->trans_list);
J
Josef Bacik 已提交
162
	extent_io_tree_init(&cur_trans->dirty_pages,
163 164 165 166
			     fs_info->btree_inode->i_mapping);
	fs_info->generation++;
	cur_trans->transid = fs_info->generation;
	fs_info->running_transaction = cur_trans;
167
	cur_trans->aborted = 0;
168
	spin_unlock(&fs_info->trans_lock);
169

C
Chris Mason 已提交
170 171 172
	return 0;
}

C
Chris Mason 已提交
173
/*
C
Chris Mason 已提交
174 175 176 177
 * this does all the record keeping required to make sure that a reference
 * counted root is properly recorded in a given transaction.  This is required
 * to make sure the old root from before we joined the transaction is deleted
 * when the transaction commits
C
Chris Mason 已提交
178
 */
C
Chris Mason 已提交
179
static int record_root_in_trans(struct btrfs_trans_handle *trans,
J
Josef Bacik 已提交
180
			       struct btrfs_root *root)
181
{
182
	if (root->ref_cows && root->last_trans < trans->transid) {
183
		WARN_ON(root == root->fs_info->extent_root);
184 185
		WARN_ON(root->commit_root != root->node);

C
Chris Mason 已提交
186 187 188 189 190 191 192 193 194 195 196 197
		/*
		 * see below for in_trans_setup usage rules
		 * we have the reloc mutex held now, so there
		 * is only one writer in this function
		 */
		root->in_trans_setup = 1;

		/* make sure readers find in_trans_setup before
		 * they find our root->last_trans update
		 */
		smp_wmb();

J
Josef Bacik 已提交
198 199 200 201 202
		spin_lock(&root->fs_info->fs_roots_radix_lock);
		if (root->last_trans == trans->transid) {
			spin_unlock(&root->fs_info->fs_roots_radix_lock);
			return 0;
		}
203 204 205
		radix_tree_tag_set(&root->fs_info->fs_roots_radix,
			   (unsigned long)root->root_key.objectid,
			   BTRFS_ROOT_TRANS_TAG);
J
Josef Bacik 已提交
206
		spin_unlock(&root->fs_info->fs_roots_radix_lock);
C
Chris Mason 已提交
207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
		root->last_trans = trans->transid;

		/* this is pretty tricky.  We don't want to
		 * take the relocation lock in btrfs_record_root_in_trans
		 * unless we're really doing the first setup for this root in
		 * this transaction.
		 *
		 * Normally we'd use root->last_trans as a flag to decide
		 * if we want to take the expensive mutex.
		 *
		 * But, we have to set root->last_trans before we
		 * init the relocation root, otherwise, we trip over warnings
		 * in ctree.c.  The solution used here is to flag ourselves
		 * with root->in_trans_setup.  When this is 1, we're still
		 * fixing up the reloc trees and everyone must wait.
		 *
		 * When this is zero, they can trust root->last_trans and fly
		 * through btrfs_record_root_in_trans without having to take the
		 * lock.  smp_wmb() makes sure that all the writes above are
		 * done before we pop in the zero below
		 */
228
		btrfs_init_reloc_root(trans, root);
C
Chris Mason 已提交
229 230
		smp_wmb();
		root->in_trans_setup = 0;
231 232 233
	}
	return 0;
}
234

C
Chris Mason 已提交
235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257

int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root)
{
	if (!root->ref_cows)
		return 0;

	/*
	 * see record_root_in_trans for comments about in_trans_setup usage
	 * and barriers
	 */
	smp_rmb();
	if (root->last_trans == trans->transid &&
	    !root->in_trans_setup)
		return 0;

	mutex_lock(&root->fs_info->reloc_mutex);
	record_root_in_trans(trans, root);
	mutex_unlock(&root->fs_info->reloc_mutex);

	return 0;
}

C
Chris Mason 已提交
258 259 260 261
/* wait for commit against the current transaction to become unblocked
 * when this is done, it is safe to start a new transaction, but the current
 * transaction might not be fully on disk.
 */
C
Chris Mason 已提交
262
static void wait_current_trans(struct btrfs_root *root)
C
Chris Mason 已提交
263
{
264
	struct btrfs_transaction *cur_trans;
C
Chris Mason 已提交
265

J
Josef Bacik 已提交
266
	spin_lock(&root->fs_info->trans_lock);
267
	cur_trans = root->fs_info->running_transaction;
C
Chris Mason 已提交
268
	if (cur_trans && cur_trans->blocked) {
269
		atomic_inc(&cur_trans->use_count);
J
Josef Bacik 已提交
270
		spin_unlock(&root->fs_info->trans_lock);
L
Li Zefan 已提交
271 272 273

		wait_event(root->fs_info->transaction_wait,
			   !cur_trans->blocked);
274
		put_transaction(cur_trans);
J
Josef Bacik 已提交
275 276
	} else {
		spin_unlock(&root->fs_info->trans_lock);
277
	}
C
Chris Mason 已提交
278 279
}

280 281
static int may_wait_transaction(struct btrfs_root *root, int type)
{
J
Josef Bacik 已提交
282 283 284 285 286 287 288 289
	if (root->fs_info->log_root_recovering)
		return 0;

	if (type == TRANS_USERSPACE)
		return 1;

	if (type == TRANS_START &&
	    !atomic_read(&root->fs_info->open_ioctl_trans))
290
		return 1;
J
Josef Bacik 已提交
291

292 293 294
	return 0;
}

M
Miao Xie 已提交
295 296 297
static struct btrfs_trans_handle *
start_transaction(struct btrfs_root *root, u64 num_items, int type,
		  enum btrfs_reserve_flush_enum flush)
C
Chris Mason 已提交
298
{
299 300
	struct btrfs_trans_handle *h;
	struct btrfs_transaction *cur_trans;
301
	u64 num_bytes = 0;
C
Chris Mason 已提交
302
	int ret;
303
	u64 qgroup_reserved = 0;
L
liubo 已提交
304 305 306

	if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR)
		return ERR_PTR(-EROFS);
307 308 309 310 311

	if (current->journal_info) {
		WARN_ON(type != TRANS_JOIN && type != TRANS_JOIN_NOLOCK);
		h = current->journal_info;
		h->use_count++;
312
		WARN_ON(h->use_count > 2);
313 314 315 316
		h->orig_rsv = h->block_rsv;
		h->block_rsv = NULL;
		goto got_it;
	}
317 318 319 320 321 322

	/*
	 * Do the reservation before we join the transaction so we can do all
	 * the appropriate flushing if need be.
	 */
	if (num_items > 0 && root != root->fs_info->chunk_root) {
323 324 325 326 327 328 329 330
		if (root->fs_info->quota_enabled &&
		    is_fstree(root->root_key.objectid)) {
			qgroup_reserved = num_items * root->leafsize;
			ret = btrfs_qgroup_reserve(root, qgroup_reserved);
			if (ret)
				return ERR_PTR(ret);
		}

331
		num_bytes = btrfs_calc_trans_metadata_size(root, num_items);
M
Miao Xie 已提交
332 333 334
		ret = btrfs_block_rsv_add(root,
					  &root->fs_info->trans_block_rsv,
					  num_bytes, flush);
335 336 337
		if (ret)
			return ERR_PTR(ret);
	}
338 339 340 341
again:
	h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
	if (!h)
		return ERR_PTR(-ENOMEM);
C
Chris Mason 已提交
342

343 344 345 346 347 348
	/*
	 * If we are JOIN_NOLOCK we're already committing a transaction and
	 * waiting on this guy, so we don't need to do the sb_start_intwrite
	 * because we're already holding a ref.  We need this because we could
	 * have raced in and did an fsync() on a file which can kick a commit
	 * and then we deadlock with somebody doing a freeze.
349 350 351
	 *
	 * If we are ATTACH, it means we just want to catch the current
	 * transaction and commit it, so we needn't do sb_start_intwrite(). 
352
	 */
353
	if (type < TRANS_JOIN_NOLOCK)
354
		sb_start_intwrite(root->fs_info->sb);
355

356
	if (may_wait_transaction(root, type))
C
Chris Mason 已提交
357
		wait_current_trans(root);
358

J
Josef Bacik 已提交
359
	do {
360
		ret = join_transaction(root, type);
J
Josef Bacik 已提交
361 362 363 364
		if (ret == -EBUSY)
			wait_current_trans(root);
	} while (ret == -EBUSY);

T
Tsutomu Itoh 已提交
365
	if (ret < 0) {
366 367 368 369 370
		/* We must get the transaction if we are JOIN_NOLOCK. */
		BUG_ON(type == TRANS_JOIN_NOLOCK);

		if (type < TRANS_JOIN_NOLOCK)
			sb_end_intwrite(root->fs_info->sb);
371
		kmem_cache_free(btrfs_trans_handle_cachep, h);
T
Tsutomu Itoh 已提交
372 373
		return ERR_PTR(ret);
	}
374

375 376 377 378
	cur_trans = root->fs_info->running_transaction;

	h->transid = cur_trans->transid;
	h->transaction = cur_trans;
C
Chris Mason 已提交
379
	h->blocks_used = 0;
380
	h->bytes_reserved = 0;
381
	h->root = root;
382
	h->delayed_ref_updates = 0;
383
	h->use_count = 1;
384
	h->adding_csums = 0;
385
	h->block_rsv = NULL;
386
	h->orig_rsv = NULL;
387
	h->aborted = 0;
388
	h->qgroup_reserved = qgroup_reserved;
389
	h->delayed_ref_elem.seq = 0;
390
	h->type = type;
391
	INIT_LIST_HEAD(&h->qgroup_ref_list);
392
	INIT_LIST_HEAD(&h->new_bgs);
393

394 395 396 397 398 399
	smp_mb();
	if (cur_trans->blocked && may_wait_transaction(root, type)) {
		btrfs_commit_transaction(h, root);
		goto again;
	}

400
	if (num_bytes) {
J
Josef Bacik 已提交
401
		trace_btrfs_space_reservation(root->fs_info, "transaction",
402
					      h->transid, num_bytes, 1);
403 404
		h->block_rsv = &root->fs_info->trans_block_rsv;
		h->bytes_reserved = num_bytes;
405
	}
J
Josef Bacik 已提交
406

407
got_it:
J
Josef Bacik 已提交
408
	btrfs_record_root_in_trans(h, root);
409 410 411

	if (!current->journal_info && type != TRANS_USERSPACE)
		current->journal_info = h;
C
Chris Mason 已提交
412 413 414
	return h;
}

415
struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
416
						   int num_items)
417
{
M
Miao Xie 已提交
418 419
	return start_transaction(root, num_items, TRANS_START,
				 BTRFS_RESERVE_FLUSH_ALL);
420
}
421

M
Miao Xie 已提交
422
struct btrfs_trans_handle *btrfs_start_transaction_lflush(
423 424
					struct btrfs_root *root, int num_items)
{
M
Miao Xie 已提交
425 426
	return start_transaction(root, num_items, TRANS_START,
				 BTRFS_RESERVE_FLUSH_LIMIT);
427 428
}

429
struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root)
430
{
431
	return start_transaction(root, 0, TRANS_JOIN, 0);
432 433
}

434
struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root)
435
{
436
	return start_transaction(root, 0, TRANS_JOIN_NOLOCK, 0);
437 438
}

439
struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root)
440
{
441
	return start_transaction(root, 0, TRANS_USERSPACE, 0);
442 443
}

444
struct btrfs_trans_handle *btrfs_attach_transaction(struct btrfs_root *root)
445
{
446
	return start_transaction(root, 0, TRANS_ATTACH, 0);
447 448
}

C
Chris Mason 已提交
449
/* wait for a transaction commit to be fully complete */
450
static noinline void wait_for_commit(struct btrfs_root *root,
451 452
				    struct btrfs_transaction *commit)
{
L
Li Zefan 已提交
453
	wait_event(commit->commit_wait, commit->commit_done);
454 455
}

456 457 458
int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid)
{
	struct btrfs_transaction *cur_trans = NULL, *t;
459
	int ret = 0;
460 461 462

	if (transid) {
		if (transid <= root->fs_info->last_trans_committed)
J
Josef Bacik 已提交
463
			goto out;
464

465
		ret = -EINVAL;
466
		/* find specified transaction */
J
Josef Bacik 已提交
467
		spin_lock(&root->fs_info->trans_lock);
468 469 470
		list_for_each_entry(t, &root->fs_info->trans_list, list) {
			if (t->transid == transid) {
				cur_trans = t;
J
Josef Bacik 已提交
471
				atomic_inc(&cur_trans->use_count);
472
				ret = 0;
473 474
				break;
			}
475 476
			if (t->transid > transid) {
				ret = 0;
477
				break;
478
			}
479
		}
J
Josef Bacik 已提交
480
		spin_unlock(&root->fs_info->trans_lock);
481
		/* The specified transaction doesn't exist */
482
		if (!cur_trans)
483
			goto out;
484 485
	} else {
		/* find newest transaction that is committing | committed */
J
Josef Bacik 已提交
486
		spin_lock(&root->fs_info->trans_lock);
487 488 489 490
		list_for_each_entry_reverse(t, &root->fs_info->trans_list,
					    list) {
			if (t->in_commit) {
				if (t->commit_done)
491
					break;
492
				cur_trans = t;
J
Josef Bacik 已提交
493
				atomic_inc(&cur_trans->use_count);
494 495 496
				break;
			}
		}
J
Josef Bacik 已提交
497
		spin_unlock(&root->fs_info->trans_lock);
498
		if (!cur_trans)
J
Josef Bacik 已提交
499
			goto out;  /* nothing committing|committed */
500 501 502 503
	}

	wait_for_commit(root, cur_trans);
	put_transaction(cur_trans);
J
Josef Bacik 已提交
504
out:
505 506 507
	return ret;
}

C
Chris Mason 已提交
508 509
void btrfs_throttle(struct btrfs_root *root)
{
J
Josef Bacik 已提交
510
	if (!atomic_read(&root->fs_info->open_ioctl_trans))
511
		wait_current_trans(root);
C
Chris Mason 已提交
512 513
}

514 515 516 517
static int should_end_transaction(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root)
{
	int ret;
518 519

	ret = btrfs_block_rsv_check(root, &root->fs_info->global_block_rsv, 5);
520 521 522 523 524 525 526 527
	return ret ? 1 : 0;
}

int btrfs_should_end_transaction(struct btrfs_trans_handle *trans,
				 struct btrfs_root *root)
{
	struct btrfs_transaction *cur_trans = trans->transaction;
	int updates;
528
	int err;
529

J
Josef Bacik 已提交
530
	smp_mb();
531 532 533 534 535
	if (cur_trans->blocked || cur_trans->delayed_refs.flushing)
		return 1;

	updates = trans->delayed_ref_updates;
	trans->delayed_ref_updates = 0;
536 537 538 539 540
	if (updates) {
		err = btrfs_run_delayed_refs(trans, root, updates);
		if (err) /* Error code will also eval true */
			return err;
	}
541 542 543 544

	return should_end_transaction(trans, root);
}

545
static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
546
			  struct btrfs_root *root, int throttle)
C
Chris Mason 已提交
547
{
548
	struct btrfs_transaction *cur_trans = trans->transaction;
549
	struct btrfs_fs_info *info = root->fs_info;
550
	int count = 0;
551
	int lock = (trans->type != TRANS_JOIN_NOLOCK);
552
	int err = 0;
553

554 555 556 557 558
	if (--trans->use_count) {
		trans->block_rsv = trans->orig_rsv;
		return 0;
	}

559 560 561 562 563
	/*
	 * do the qgroup accounting as early as possible
	 */
	err = btrfs_delayed_refs_qgroup_accounting(trans, info);

564
	btrfs_trans_release_metadata(trans, root);
565
	trans->block_rsv = NULL;
566 567 568 569 570
	/*
	 * the same root has to be passed to start_transaction and
	 * end_transaction. Subvolume quota depends on this.
	 */
	WARN_ON(trans->root != root);
571 572 573 574 575 576

	if (trans->qgroup_reserved) {
		btrfs_qgroup_free(root, trans->qgroup_reserved);
		trans->qgroup_reserved = 0;
	}

577 578 579
	if (!list_empty(&trans->new_bgs))
		btrfs_create_pending_block_groups(trans, root);

580
	while (count < 2) {
581 582 583 584 585 586 587 588 589 590
		unsigned long cur = trans->delayed_ref_updates;
		trans->delayed_ref_updates = 0;
		if (cur &&
		    trans->transaction->delayed_refs.num_heads_ready > 64) {
			trans->delayed_ref_updates = 0;
			btrfs_run_delayed_refs(trans, root, cur);
		} else {
			break;
		}
		count++;
591
	}
592 593
	btrfs_trans_release_metadata(trans, root);
	trans->block_rsv = NULL;
594

595 596 597
	if (!list_empty(&trans->new_bgs))
		btrfs_create_pending_block_groups(trans, root);

J
Josef Bacik 已提交
598 599
	if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) &&
	    should_end_transaction(trans, root)) {
600
		trans->transaction->blocked = 1;
J
Josef Bacik 已提交
601 602
		smp_wmb();
	}
603

604
	if (lock && cur_trans->blocked && !cur_trans->in_commit) {
605 606 607 608 609 610 611
		if (throttle) {
			/*
			 * We may race with somebody else here so end up having
			 * to call end_transaction on ourselves again, so inc
			 * our use_count.
			 */
			trans->use_count++;
612
			return btrfs_commit_transaction(trans, root);
613
		} else {
614
			wake_up_process(info->transaction_kthread);
615
		}
616 617
	}

618
	if (trans->type < TRANS_JOIN_NOLOCK)
619
		sb_end_intwrite(root->fs_info->sb);
620

621
	WARN_ON(cur_trans != info->running_transaction);
622 623
	WARN_ON(atomic_read(&cur_trans->num_writers) < 1);
	atomic_dec(&cur_trans->num_writers);
624

625
	smp_mb();
C
Chris Mason 已提交
626 627 628
	if (waitqueue_active(&cur_trans->writer_wait))
		wake_up(&cur_trans->writer_wait);
	put_transaction(cur_trans);
J
Josef Bacik 已提交
629 630 631

	if (current->journal_info == trans)
		current->journal_info = NULL;
632

Y
Yan, Zheng 已提交
633 634 635
	if (throttle)
		btrfs_run_delayed_iputs(root);

636 637
	if (trans->aborted ||
	    root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) {
638
		err = -EIO;
639
	}
640
	assert_qgroups_uptodate(trans);
641

642 643 644
	memset(trans, 0, sizeof(*trans));
	kmem_cache_free(btrfs_trans_handle_cachep, trans);
	return err;
C
Chris Mason 已提交
645 646
}

647 648 649
int btrfs_end_transaction(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root)
{
650 651
	int ret;

652
	ret = __btrfs_end_transaction(trans, root, 0);
653 654 655
	if (ret)
		return ret;
	return 0;
656 657 658 659 660
}

int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root)
{
661 662
	int ret;

663
	ret = __btrfs_end_transaction(trans, root, 1);
664 665 666 667 668 669 670 671
	if (ret)
		return ret;
	return 0;
}

int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans,
				struct btrfs_root *root)
{
672
	return __btrfs_end_transaction(trans, root, 1);
673 674
}

C
Chris Mason 已提交
675 676 677
/*
 * when btree blocks are allocated, they have some corresponding bits set for
 * them in one of two extent_io trees.  This is used to make sure all of
678
 * those extents are sent to disk but does not wait on them
C
Chris Mason 已提交
679
 */
680
int btrfs_write_marked_extents(struct btrfs_root *root,
681
			       struct extent_io_tree *dirty_pages, int mark)
C
Chris Mason 已提交
682
{
683
	int err = 0;
684
	int werr = 0;
J
Josef Bacik 已提交
685
	struct address_space *mapping = root->fs_info->btree_inode->i_mapping;
686
	struct extent_state *cached_state = NULL;
687
	u64 start = 0;
688
	u64 end;
689

J
Josef Bacik 已提交
690
	while (!find_first_extent_bit(dirty_pages, start, &start, &end,
691 692 693 694
				      mark, &cached_state)) {
		convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT,
				   mark, &cached_state, GFP_NOFS);
		cached_state = NULL;
J
Josef Bacik 已提交
695 696 697 698 699
		err = filemap_fdatawrite_range(mapping, start, end);
		if (err)
			werr = err;
		cond_resched();
		start = end + 1;
700
	}
701 702 703 704 705 706 707 708 709 710 711 712
	if (err)
		werr = err;
	return werr;
}

/*
 * when btree blocks are allocated, they have some corresponding bits set for
 * them in one of two extent_io trees.  This is used to make sure all of
 * those extents are on disk for transaction or log commit.  We wait
 * on all the pages and clear them from the dirty pages state tree
 */
int btrfs_wait_marked_extents(struct btrfs_root *root,
713
			      struct extent_io_tree *dirty_pages, int mark)
714 715 716
{
	int err = 0;
	int werr = 0;
J
Josef Bacik 已提交
717
	struct address_space *mapping = root->fs_info->btree_inode->i_mapping;
718
	struct extent_state *cached_state = NULL;
719 720
	u64 start = 0;
	u64 end;
721

J
Josef Bacik 已提交
722
	while (!find_first_extent_bit(dirty_pages, start, &start, &end,
723 724 725
				      EXTENT_NEED_WAIT, &cached_state)) {
		clear_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT,
				 0, 0, &cached_state, GFP_NOFS);
J
Josef Bacik 已提交
726 727 728 729 730
		err = filemap_fdatawait_range(mapping, start, end);
		if (err)
			werr = err;
		cond_resched();
		start = end + 1;
731
	}
732 733 734
	if (err)
		werr = err;
	return werr;
C
Chris Mason 已提交
735 736
}

737 738 739 740 741 742
/*
 * when btree blocks are allocated, they have some corresponding bits set for
 * them in one of two extent_io trees.  This is used to make sure all of
 * those extents are on disk for transaction or log commit
 */
int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
743
				struct extent_io_tree *dirty_pages, int mark)
744 745 746 747
{
	int ret;
	int ret2;

748 749
	ret = btrfs_write_marked_extents(root, dirty_pages, mark);
	ret2 = btrfs_wait_marked_extents(root, dirty_pages, mark);
750 751 752 753 754 755

	if (ret)
		return ret;
	if (ret2)
		return ret2;
	return 0;
756 757
}

758 759 760 761 762 763 764 765 766
int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root)
{
	if (!trans || !trans->transaction) {
		struct inode *btree_inode;
		btree_inode = root->fs_info->btree_inode;
		return filemap_write_and_wait(btree_inode->i_mapping);
	}
	return btrfs_write_and_wait_marked_extents(root,
767 768
					   &trans->transaction->dirty_pages,
					   EXTENT_DIRTY);
769 770
}

C
Chris Mason 已提交
771 772 773 774 775 776 777 778 779 780
/*
 * this is used to update the root pointer in the tree of tree roots.
 *
 * But, in the case of the extent allocation tree, updating the root
 * pointer may allocate blocks which may change the root of the extent
 * allocation tree.
 *
 * So, this loops and repeats and makes sure the cowonly root didn't
 * change while the root pointer was being updated in the metadata.
 */
781 782
static int update_cowonly_root(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root)
C
Chris Mason 已提交
783 784
{
	int ret;
785
	u64 old_root_bytenr;
786
	u64 old_root_used;
787
	struct btrfs_root *tree_root = root->fs_info->tree_root;
C
Chris Mason 已提交
788

789
	old_root_used = btrfs_root_used(&root->root_item);
790
	btrfs_write_dirty_block_groups(trans, root);
791

C
Chris Mason 已提交
792
	while (1) {
793
		old_root_bytenr = btrfs_root_bytenr(&root->root_item);
794 795
		if (old_root_bytenr == root->node->start &&
		    old_root_used == btrfs_root_used(&root->root_item))
C
Chris Mason 已提交
796
			break;
797

798
		btrfs_set_root_node(&root->root_item, root->node);
C
Chris Mason 已提交
799
		ret = btrfs_update_root(trans, tree_root,
800 801
					&root->root_key,
					&root->root_item);
802 803
		if (ret)
			return ret;
804

805
		old_root_used = btrfs_root_used(&root->root_item);
806
		ret = btrfs_write_dirty_block_groups(trans, root);
807 808
		if (ret)
			return ret;
809
	}
810 811 812 813

	if (root != root->fs_info->extent_root)
		switch_commit_root(root);

814 815 816
	return 0;
}

C
Chris Mason 已提交
817 818
/*
 * update all the cowonly tree roots on disk
819 820 821 822
 *
 * The error handling in this function may not be obvious. Any of the
 * failures will cause the file system to go offline. We still need
 * to clean up the delayed refs.
C
Chris Mason 已提交
823
 */
824 825
static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans,
					 struct btrfs_root *root)
826 827 828
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct list_head *next;
829
	struct extent_buffer *eb;
830
	int ret;
831

832
	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
833 834
	if (ret)
		return ret;
835

836
	eb = btrfs_lock_root_node(fs_info->tree_root);
837 838
	ret = btrfs_cow_block(trans, fs_info->tree_root, eb, NULL,
			      0, &eb);
839 840
	btrfs_tree_unlock(eb);
	free_extent_buffer(eb);
841

842 843 844
	if (ret)
		return ret;

845
	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
846 847
	if (ret)
		return ret;
848

849
	ret = btrfs_run_dev_stats(trans, root->fs_info);
850 851 852
	WARN_ON(ret);
	ret = btrfs_run_dev_replace(trans, root->fs_info);
	WARN_ON(ret);
853

854 855 856 857 858 859 860
	ret = btrfs_run_qgroups(trans, root->fs_info);
	BUG_ON(ret);

	/* run_qgroups might have added some more refs */
	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
	BUG_ON(ret);

C
Chris Mason 已提交
861
	while (!list_empty(&fs_info->dirty_cowonly_roots)) {
862 863 864
		next = fs_info->dirty_cowonly_roots.next;
		list_del_init(next);
		root = list_entry(next, struct btrfs_root, dirty_list);
865

866 867 868
		ret = update_cowonly_root(trans, root);
		if (ret)
			return ret;
C
Chris Mason 已提交
869
	}
870 871 872 873 874

	down_write(&fs_info->extent_commit_sem);
	switch_commit_root(fs_info->extent_root);
	up_write(&fs_info->extent_commit_sem);

875 876
	btrfs_after_dev_replace_commit(fs_info);

C
Chris Mason 已提交
877 878 879
	return 0;
}

C
Chris Mason 已提交
880 881 882 883 884
/*
 * dead roots are old snapshots that need to be deleted.  This allocates
 * a dirty root struct and adds it into the list of dead roots that need to
 * be deleted
 */
885
int btrfs_add_dead_root(struct btrfs_root *root)
886
{
J
Josef Bacik 已提交
887
	spin_lock(&root->fs_info->trans_lock);
888
	list_add(&root->root_list, &root->fs_info->dead_roots);
J
Josef Bacik 已提交
889
	spin_unlock(&root->fs_info->trans_lock);
890 891 892
	return 0;
}

C
Chris Mason 已提交
893
/*
894
 * update all the cowonly tree roots on disk
C
Chris Mason 已提交
895
 */
896 897
static noinline int commit_fs_roots(struct btrfs_trans_handle *trans,
				    struct btrfs_root *root)
898 899
{
	struct btrfs_root *gang[8];
900
	struct btrfs_fs_info *fs_info = root->fs_info;
901 902
	int i;
	int ret;
903 904
	int err = 0;

J
Josef Bacik 已提交
905
	spin_lock(&fs_info->fs_roots_radix_lock);
C
Chris Mason 已提交
906
	while (1) {
907 908
		ret = radix_tree_gang_lookup_tag(&fs_info->fs_roots_radix,
						 (void **)gang, 0,
909 910 911 912 913 914
						 ARRAY_SIZE(gang),
						 BTRFS_ROOT_TRANS_TAG);
		if (ret == 0)
			break;
		for (i = 0; i < ret; i++) {
			root = gang[i];
915 916 917
			radix_tree_tag_clear(&fs_info->fs_roots_radix,
					(unsigned long)root->root_key.objectid,
					BTRFS_ROOT_TRANS_TAG);
J
Josef Bacik 已提交
918
			spin_unlock(&fs_info->fs_roots_radix_lock);
Y
Yan Zheng 已提交
919

920
			btrfs_free_log(trans, root);
921
			btrfs_update_reloc_root(trans, root);
922
			btrfs_orphan_commit_root(trans, root);
923

924 925
			btrfs_save_ino_cache(root, trans);

926 927 928 929
			/* see comments in should_cow_block() */
			root->force_cow = 0;
			smp_wmb();

930
			if (root->commit_root != root->node) {
931
				mutex_lock(&root->fs_commit_mutex);
J
Josef Bacik 已提交
932
				switch_commit_root(root);
933 934 935
				btrfs_unpin_free_ino(root);
				mutex_unlock(&root->fs_commit_mutex);

936 937 938
				btrfs_set_root_node(&root->root_item,
						    root->node);
			}
939 940

			err = btrfs_update_root(trans, fs_info->tree_root,
941 942
						&root->root_key,
						&root->root_item);
J
Josef Bacik 已提交
943
			spin_lock(&fs_info->fs_roots_radix_lock);
944 945
			if (err)
				break;
946 947
		}
	}
J
Josef Bacik 已提交
948
	spin_unlock(&fs_info->fs_roots_radix_lock);
949
	return err;
950 951
}

C
Chris Mason 已提交
952 953 954 955
/*
 * defrag a given btree.  If cacheonly == 1, this won't read from the disk,
 * otherwise every leaf in the btree is read and defragged.
 */
956 957 958 959
int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
{
	struct btrfs_fs_info *info = root->fs_info;
	struct btrfs_trans_handle *trans;
960
	int ret;
961

962
	if (xchg(&root->defrag_running, 1))
963
		return 0;
964

965
	while (1) {
966 967 968 969
		trans = btrfs_start_transaction(root, 0);
		if (IS_ERR(trans))
			return PTR_ERR(trans);

970
		ret = btrfs_defrag_leaves(trans, root, cacheonly);
971

972
		btrfs_end_transaction(trans, root);
973
		btrfs_btree_balance_dirty(info->tree_root);
974 975
		cond_resched();

976
		if (btrfs_fs_closing(root->fs_info) || ret != -EAGAIN)
977 978 979
			break;
	}
	root->defrag_running = 0;
980
	return ret;
981 982
}

C
Chris Mason 已提交
983 984 985 986
/*
 * new snapshots need to be created at a very specific time in the
 * transaction commit.  This does the actual creation
 */
987
static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
988 989 990 991
				   struct btrfs_fs_info *fs_info,
				   struct btrfs_pending_snapshot *pending)
{
	struct btrfs_key key;
992
	struct btrfs_root_item *new_root_item;
993 994
	struct btrfs_root *tree_root = fs_info->tree_root;
	struct btrfs_root *root = pending->root;
995
	struct btrfs_root *parent_root;
L
Liu Bo 已提交
996
	struct btrfs_block_rsv *rsv;
997
	struct inode *parent_inode;
998 999
	struct btrfs_path *path;
	struct btrfs_dir_item *dir_item;
1000
	struct dentry *parent;
1001
	struct dentry *dentry;
1002
	struct extent_buffer *tmp;
1003
	struct extent_buffer *old;
1004
	struct timespec cur_time = CURRENT_TIME;
1005
	int ret;
1006
	u64 to_reserve = 0;
1007
	u64 index = 0;
1008
	u64 objectid;
L
Li Zefan 已提交
1009
	u64 root_flags;
1010
	uuid_le new_uuid;
1011

1012 1013 1014 1015 1016 1017
	path = btrfs_alloc_path();
	if (!path) {
		ret = pending->error = -ENOMEM;
		goto path_alloc_fail;
	}

1018 1019
	new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS);
	if (!new_root_item) {
1020
		ret = pending->error = -ENOMEM;
1021
		goto root_item_alloc_fail;
1022
	}
1023

1024
	ret = btrfs_find_free_objectid(tree_root, &objectid);
1025 1026
	if (ret) {
		pending->error = ret;
1027
		goto no_free_objectid;
1028
	}
1029

1030
	btrfs_reloc_pre_snapshot(trans, pending, &to_reserve);
1031 1032

	if (to_reserve > 0) {
M
Miao Xie 已提交
1033 1034 1035
		ret = btrfs_block_rsv_add(root, &pending->block_rsv,
					  to_reserve,
					  BTRFS_RESERVE_NO_FLUSH);
1036 1037
		if (ret) {
			pending->error = ret;
1038
			goto no_free_objectid;
1039 1040 1041
		}
	}

A
Arne Jansen 已提交
1042 1043 1044 1045
	ret = btrfs_qgroup_inherit(trans, fs_info, root->root_key.objectid,
				   objectid, pending->inherit);
	if (ret) {
		pending->error = ret;
1046
		goto no_free_objectid;
A
Arne Jansen 已提交
1047 1048
	}

1049
	key.objectid = objectid;
1050 1051
	key.offset = (u64)-1;
	key.type = BTRFS_ROOT_ITEM_KEY;
1052

1053
	rsv = trans->block_rsv;
1054
	trans->block_rsv = &pending->block_rsv;
1055

1056
	dentry = pending->dentry;
1057 1058
	parent = dget_parent(dentry);
	parent_inode = parent->d_inode;
1059
	parent_root = BTRFS_I(parent_inode)->root;
C
Chris Mason 已提交
1060
	record_root_in_trans(trans, parent_root);
1061

1062 1063 1064
	/*
	 * insert the directory item
	 */
1065
	ret = btrfs_set_inode_index(parent_inode, &index);
1066
	BUG_ON(ret); /* -ENOMEM */
1067 1068 1069 1070 1071 1072 1073

	/* check if there is a file/dir which has the same name. */
	dir_item = btrfs_lookup_dir_item(NULL, parent_root, path,
					 btrfs_ino(parent_inode),
					 dentry->d_name.name,
					 dentry->d_name.len, 0);
	if (dir_item != NULL && !IS_ERR(dir_item)) {
1074 1075
		pending->error = -EEXIST;
		goto fail;
1076 1077
	} else if (IS_ERR(dir_item)) {
		ret = PTR_ERR(dir_item);
1078 1079
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
1080
	}
1081
	btrfs_release_path(path);
1082

1083 1084 1085 1086 1087 1088 1089
	/*
	 * pull in the delayed directory update
	 * and the delayed inode item
	 * otherwise we corrupt the FS during
	 * snapshot
	 */
	ret = btrfs_run_delayed_items(trans, root);
1090 1091 1092 1093
	if (ret) {	/* Transaction aborted */
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1094

C
Chris Mason 已提交
1095
	record_root_in_trans(trans, root);
1096 1097
	btrfs_set_root_last_snapshot(&root->root_item, trans->transid);
	memcpy(new_root_item, &root->root_item, sizeof(*new_root_item));
1098
	btrfs_check_and_init_root_item(new_root_item);
1099

L
Li Zefan 已提交
1100 1101 1102 1103 1104 1105 1106
	root_flags = btrfs_root_flags(new_root_item);
	if (pending->readonly)
		root_flags |= BTRFS_ROOT_SUBVOL_RDONLY;
	else
		root_flags &= ~BTRFS_ROOT_SUBVOL_RDONLY;
	btrfs_set_root_flags(new_root_item, root_flags);

1107 1108 1109 1110 1111 1112 1113
	btrfs_set_root_generation_v2(new_root_item,
			trans->transid);
	uuid_le_gen(&new_uuid);
	memcpy(new_root_item->uuid, new_uuid.b, BTRFS_UUID_SIZE);
	memcpy(new_root_item->parent_uuid, root->root_item.uuid,
			BTRFS_UUID_SIZE);
	new_root_item->otime.sec = cpu_to_le64(cur_time.tv_sec);
1114
	new_root_item->otime.nsec = cpu_to_le32(cur_time.tv_nsec);
1115 1116 1117 1118 1119 1120
	btrfs_set_root_otransid(new_root_item, trans->transid);
	memset(&new_root_item->stime, 0, sizeof(new_root_item->stime));
	memset(&new_root_item->rtime, 0, sizeof(new_root_item->rtime));
	btrfs_set_root_stransid(new_root_item, 0);
	btrfs_set_root_rtransid(new_root_item, 0);

1121
	old = btrfs_lock_root_node(root);
1122
	ret = btrfs_cow_block(trans, root, old, NULL, 0, &old);
1123 1124 1125
	if (ret) {
		btrfs_tree_unlock(old);
		free_extent_buffer(old);
1126 1127
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
1128
	}
1129

1130 1131
	btrfs_set_lock_blocking(old);

1132
	ret = btrfs_copy_root(trans, root, old, &tmp, objectid);
1133
	/* clean up in any case */
1134 1135
	btrfs_tree_unlock(old);
	free_extent_buffer(old);
1136 1137 1138 1139
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1140

1141 1142 1143 1144
	/* see comments in should_cow_block() */
	root->force_cow = 1;
	smp_wmb();

1145
	btrfs_set_root_node(new_root_item, tmp);
1146 1147 1148
	/* record when the snapshot was created in key.offset */
	key.offset = trans->transid;
	ret = btrfs_insert_root(trans, tree_root, &key, new_root_item);
1149 1150
	btrfs_tree_unlock(tmp);
	free_extent_buffer(tmp);
1151 1152 1153 1154
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1155

1156 1157 1158 1159
	/*
	 * insert root back/forward references
	 */
	ret = btrfs_add_root_ref(trans, tree_root, objectid,
1160
				 parent_root->root_key.objectid,
L
Li Zefan 已提交
1161
				 btrfs_ino(parent_inode), index,
1162
				 dentry->d_name.name, dentry->d_name.len);
1163 1164 1165 1166
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1167

1168 1169
	key.offset = (u64)-1;
	pending->snap = btrfs_read_fs_root_no_name(root->fs_info, &key);
1170 1171
	if (IS_ERR(pending->snap)) {
		ret = PTR_ERR(pending->snap);
1172 1173
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
1174
	}
1175

1176
	ret = btrfs_reloc_post_snapshot(trans, pending);
1177 1178 1179 1180
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1181 1182

	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
1183 1184 1185 1186
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1187 1188 1189 1190 1191 1192

	ret = btrfs_insert_dir_item(trans, parent_root,
				    dentry->d_name.name, dentry->d_name.len,
				    parent_inode, &key,
				    BTRFS_FT_DIR, index);
	/* We have check then name at the beginning, so it is impossible. */
C
Chris Mason 已提交
1193
	BUG_ON(ret == -EEXIST || ret == -EOVERFLOW);
1194 1195 1196 1197
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto fail;
	}
1198 1199 1200 1201

	btrfs_i_size_write(parent_inode, parent_inode->i_size +
					 dentry->d_name.len * 2);
	parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
1202
	ret = btrfs_update_inode_fallback(trans, parent_root, parent_inode);
1203
	if (ret)
1204
		btrfs_abort_transaction(trans, root, ret);
1205
fail:
1206
	dput(parent);
L
Liu Bo 已提交
1207
	trans->block_rsv = rsv;
1208 1209 1210
no_free_objectid:
	kfree(new_root_item);
root_item_alloc_fail:
1211 1212
	btrfs_free_path(path);
path_alloc_fail:
1213
	btrfs_block_rsv_release(root, &pending->block_rsv, (u64)-1);
1214
	return ret;
1215 1216
}

C
Chris Mason 已提交
1217 1218 1219
/*
 * create all the snapshots we've scheduled for creation
 */
1220 1221
static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
					     struct btrfs_fs_info *fs_info)
1222 1223 1224 1225
{
	struct btrfs_pending_snapshot *pending;
	struct list_head *head = &trans->transaction->pending_snapshots;

1226 1227
	list_for_each_entry(pending, head, list)
		create_pending_snapshot(trans, fs_info, pending);
1228 1229 1230
	return 0;
}

1231 1232 1233 1234 1235
static void update_super_roots(struct btrfs_root *root)
{
	struct btrfs_root_item *root_item;
	struct btrfs_super_block *super;

1236
	super = root->fs_info->super_copy;
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246

	root_item = &root->fs_info->chunk_root->root_item;
	super->chunk_root = root_item->bytenr;
	super->chunk_root_generation = root_item->generation;
	super->chunk_root_level = root_item->level;

	root_item = &root->fs_info->tree_root->root_item;
	super->root = root_item->bytenr;
	super->generation = root_item->generation;
	super->root_level = root_item->level;
1247
	if (btrfs_test_opt(root, SPACE_CACHE))
1248
		super->cache_generation = root_item->generation;
1249 1250
}

1251 1252 1253
int btrfs_transaction_in_commit(struct btrfs_fs_info *info)
{
	int ret = 0;
J
Josef Bacik 已提交
1254
	spin_lock(&info->trans_lock);
1255 1256
	if (info->running_transaction)
		ret = info->running_transaction->in_commit;
J
Josef Bacik 已提交
1257
	spin_unlock(&info->trans_lock);
1258 1259 1260
	return ret;
}

1261 1262 1263
int btrfs_transaction_blocked(struct btrfs_fs_info *info)
{
	int ret = 0;
J
Josef Bacik 已提交
1264
	spin_lock(&info->trans_lock);
1265 1266
	if (info->running_transaction)
		ret = info->running_transaction->blocked;
J
Josef Bacik 已提交
1267
	spin_unlock(&info->trans_lock);
1268 1269 1270
	return ret;
}

S
Sage Weil 已提交
1271 1272 1273 1274 1275 1276 1277
/*
 * wait for the current transaction commit to start and block subsequent
 * transaction joins
 */
static void wait_current_trans_commit_start(struct btrfs_root *root,
					    struct btrfs_transaction *trans)
{
L
Li Zefan 已提交
1278
	wait_event(root->fs_info->transaction_blocked_wait, trans->in_commit);
S
Sage Weil 已提交
1279 1280 1281 1282 1283 1284 1285 1286 1287
}

/*
 * wait for the current transaction to start and then become unblocked.
 * caller holds ref.
 */
static void wait_current_trans_commit_start_and_unblock(struct btrfs_root *root,
					 struct btrfs_transaction *trans)
{
L
Li Zefan 已提交
1288 1289
	wait_event(root->fs_info->transaction_wait,
		   trans->commit_done || (trans->in_commit && !trans->blocked));
S
Sage Weil 已提交
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
}

/*
 * commit transactions asynchronously. once btrfs_commit_transaction_async
 * returns, any subsequent transaction will not be allowed to join.
 */
struct btrfs_async_commit {
	struct btrfs_trans_handle *newtrans;
	struct btrfs_root *root;
	struct delayed_work work;
};

static void do_async_commit(struct work_struct *work)
{
	struct btrfs_async_commit *ac =
		container_of(work, struct btrfs_async_commit, work.work);

1307 1308 1309 1310
	/*
	 * We've got freeze protection passed with the transaction.
	 * Tell lockdep about it.
	 */
1311 1312 1313 1314
	if (ac->newtrans->type < TRANS_JOIN_NOLOCK)
		rwsem_acquire_read(
		     &ac->root->fs_info->sb->s_writers.lock_map[SB_FREEZE_FS-1],
		     0, 1, _THIS_IP_);
1315

1316 1317
	current->journal_info = ac->newtrans;

S
Sage Weil 已提交
1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329
	btrfs_commit_transaction(ac->newtrans, ac->root);
	kfree(ac);
}

int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   int wait_for_unblock)
{
	struct btrfs_async_commit *ac;
	struct btrfs_transaction *cur_trans;

	ac = kmalloc(sizeof(*ac), GFP_NOFS);
T
Tsutomu Itoh 已提交
1330 1331
	if (!ac)
		return -ENOMEM;
S
Sage Weil 已提交
1332 1333 1334

	INIT_DELAYED_WORK(&ac->work, do_async_commit);
	ac->root = root;
1335
	ac->newtrans = btrfs_join_transaction(root);
1336 1337 1338 1339 1340
	if (IS_ERR(ac->newtrans)) {
		int err = PTR_ERR(ac->newtrans);
		kfree(ac);
		return err;
	}
S
Sage Weil 已提交
1341 1342 1343

	/* take transaction reference */
	cur_trans = trans->transaction;
1344
	atomic_inc(&cur_trans->use_count);
S
Sage Weil 已提交
1345 1346

	btrfs_end_transaction(trans, root);
1347 1348 1349 1350 1351

	/*
	 * Tell lockdep we've released the freeze rwsem, since the
	 * async commit thread will be the one to unlock it.
	 */
1352 1353 1354 1355
	if (trans->type < TRANS_JOIN_NOLOCK)
		rwsem_release(
			&root->fs_info->sb->s_writers.lock_map[SB_FREEZE_FS-1],
			1, _THIS_IP_);
1356

S
Sage Weil 已提交
1357 1358 1359 1360 1361 1362 1363 1364
	schedule_delayed_work(&ac->work, 0);

	/* wait for transaction to start and unblock */
	if (wait_for_unblock)
		wait_current_trans_commit_start_and_unblock(root, cur_trans);
	else
		wait_current_trans_commit_start(root, cur_trans);

1365 1366 1367 1368
	if (current->journal_info == trans)
		current->journal_info = NULL;

	put_transaction(cur_trans);
S
Sage Weil 已提交
1369 1370 1371
	return 0;
}

1372 1373

static void cleanup_transaction(struct btrfs_trans_handle *trans,
1374
				struct btrfs_root *root, int err)
1375 1376 1377 1378 1379
{
	struct btrfs_transaction *cur_trans = trans->transaction;

	WARN_ON(trans->use_count > 1);

1380 1381
	btrfs_abort_transaction(trans, root, err);

1382 1383
	spin_lock(&root->fs_info->trans_lock);
	list_del_init(&cur_trans->list);
1384 1385 1386 1387
	if (cur_trans == root->fs_info->running_transaction) {
		root->fs_info->running_transaction = NULL;
		root->fs_info->trans_no_join = 0;
	}
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
	spin_unlock(&root->fs_info->trans_lock);

	btrfs_cleanup_one_transaction(trans->transaction, root);

	put_transaction(cur_trans);
	put_transaction(cur_trans);

	trace_btrfs_transaction_commit(root);

	btrfs_scrub_continue(root);

	if (current->journal_info == trans)
		current->journal_info = NULL;

	kmem_cache_free(btrfs_trans_handle_cachep, trans);
}

1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446
static int btrfs_flush_all_pending_stuffs(struct btrfs_trans_handle *trans,
					  struct btrfs_root *root)
{
	int flush_on_commit = btrfs_test_opt(root, FLUSHONCOMMIT);
	int snap_pending = 0;
	int ret;

	if (!flush_on_commit) {
		spin_lock(&root->fs_info->trans_lock);
		if (!list_empty(&trans->transaction->pending_snapshots))
			snap_pending = 1;
		spin_unlock(&root->fs_info->trans_lock);
	}

	if (flush_on_commit || snap_pending) {
		btrfs_start_delalloc_inodes(root, 1);
		btrfs_wait_ordered_extents(root, 1);
	}

	ret = btrfs_run_delayed_items(trans, root);
	if (ret)
		return ret;

	/*
	 * running the delayed items may have added new refs. account
	 * them now so that they hinder processing of more delayed refs
	 * as little as possible.
	 */
	btrfs_delayed_refs_qgroup_accounting(trans, root->fs_info);

	/*
	 * rename don't use btrfs_join_transaction, so, once we
	 * set the transaction to blocked above, we aren't going
	 * to get any new ordered operations.  We can safely run
	 * it here and no for sure that nothing new will be added
	 * to the list
	 */
	btrfs_run_ordered_operations(root, 1);

	return 0;
}

S
Sage Weil 已提交
1447 1448 1449 1450 1451 1452 1453
/*
 * btrfs_transaction state sequence:
 *    in_commit = 0, blocked = 0  (initial)
 *    in_commit = 1, blocked = 1
 *    blocked = 0
 *    commit_done = 1
 */
C
Chris Mason 已提交
1454 1455 1456
int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
			     struct btrfs_root *root)
{
1457
	unsigned long joined = 0;
1458
	struct btrfs_transaction *cur_trans = trans->transaction;
C
Chris Mason 已提交
1459
	struct btrfs_transaction *prev_trans = NULL;
C
Chris Mason 已提交
1460
	DEFINE_WAIT(wait);
1461
	int ret;
1462 1463
	int should_grow = 0;
	unsigned long now = get_seconds();
C
Chris Mason 已提交
1464

1465 1466 1467 1468 1469
	ret = btrfs_run_ordered_operations(root, 0);
	if (ret) {
		btrfs_abort_transaction(trans, root, ret);
		goto cleanup_transaction;
	}
1470

1471 1472
	/* Stop the commit early if ->aborted is set */
	if (unlikely(ACCESS_ONCE(cur_trans->aborted))) {
1473
		ret = cur_trans->aborted;
1474
		goto cleanup_transaction;
1475
	}
1476

1477 1478 1479 1480
	/* make a pass through all the delayed refs we have so far
	 * any runnings procs may add more while we are here
	 */
	ret = btrfs_run_delayed_refs(trans, root, 0);
1481 1482
	if (ret)
		goto cleanup_transaction;
1483

1484 1485 1486
	btrfs_trans_release_metadata(trans, root);
	trans->block_rsv = NULL;

1487
	cur_trans = trans->transaction;
1488

1489 1490 1491 1492
	/*
	 * set the flushing flag so procs in this transaction have to
	 * start sending their work down.
	 */
1493
	cur_trans->delayed_refs.flushing = 1;
1494

1495 1496 1497
	if (!list_empty(&trans->new_bgs))
		btrfs_create_pending_block_groups(trans, root);

1498
	ret = btrfs_run_delayed_refs(trans, root, 0);
1499 1500
	if (ret)
		goto cleanup_transaction;
1501

J
Josef Bacik 已提交
1502
	spin_lock(&cur_trans->commit_lock);
1503
	if (cur_trans->in_commit) {
J
Josef Bacik 已提交
1504
		spin_unlock(&cur_trans->commit_lock);
1505
		atomic_inc(&cur_trans->use_count);
1506
		ret = btrfs_end_transaction(trans, root);
C
Chris Mason 已提交
1507

1508
		wait_for_commit(root, cur_trans);
1509

C
Chris Mason 已提交
1510
		put_transaction(cur_trans);
1511

1512
		return ret;
C
Chris Mason 已提交
1513
	}
1514

C
Chris Mason 已提交
1515
	trans->transaction->in_commit = 1;
1516
	trans->transaction->blocked = 1;
J
Josef Bacik 已提交
1517
	spin_unlock(&cur_trans->commit_lock);
S
Sage Weil 已提交
1518 1519
	wake_up(&root->fs_info->transaction_blocked_wait);

J
Josef Bacik 已提交
1520
	spin_lock(&root->fs_info->trans_lock);
C
Chris Mason 已提交
1521 1522 1523 1524
	if (cur_trans->list.prev != &root->fs_info->trans_list) {
		prev_trans = list_entry(cur_trans->list.prev,
					struct btrfs_transaction, list);
		if (!prev_trans->commit_done) {
1525
			atomic_inc(&prev_trans->use_count);
J
Josef Bacik 已提交
1526
			spin_unlock(&root->fs_info->trans_lock);
C
Chris Mason 已提交
1527 1528 1529

			wait_for_commit(root, prev_trans);

1530
			put_transaction(prev_trans);
J
Josef Bacik 已提交
1531 1532
		} else {
			spin_unlock(&root->fs_info->trans_lock);
C
Chris Mason 已提交
1533
		}
J
Josef Bacik 已提交
1534 1535
	} else {
		spin_unlock(&root->fs_info->trans_lock);
C
Chris Mason 已提交
1536
	}
1537

1538 1539
	if (!btrfs_test_opt(root, SSD) &&
	    (now < cur_trans->start_time || now - cur_trans->start_time < 1))
1540 1541
		should_grow = 1;

1542 1543
	do {
		joined = cur_trans->num_joined;
1544

C
Chris Mason 已提交
1545
		WARN_ON(cur_trans != trans->transaction);
1546

1547
		ret = btrfs_flush_all_pending_stuffs(trans, root);
1548 1549
		if (ret)
			goto cleanup_transaction;
1550

1551 1552 1553
		prepare_to_wait(&cur_trans->writer_wait, &wait,
				TASK_UNINTERRUPTIBLE);

1554
		if (atomic_read(&cur_trans->num_writers) > 1)
1555 1556 1557
			schedule_timeout(MAX_SCHEDULE_TIMEOUT);
		else if (should_grow)
			schedule_timeout(1);
1558 1559

		finish_wait(&cur_trans->writer_wait, &wait);
1560
	} while (atomic_read(&cur_trans->num_writers) > 1 ||
1561
		 (should_grow && cur_trans->num_joined != joined));
1562

1563 1564 1565 1566
	ret = btrfs_flush_all_pending_stuffs(trans, root);
	if (ret)
		goto cleanup_transaction;

1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577
	/*
	 * Ok now we need to make sure to block out any other joins while we
	 * commit the transaction.  We could have started a join before setting
	 * no_join so make sure to wait for num_writers to == 1 again.
	 */
	spin_lock(&root->fs_info->trans_lock);
	root->fs_info->trans_no_join = 1;
	spin_unlock(&root->fs_info->trans_lock);
	wait_event(cur_trans->writer_wait,
		   atomic_read(&cur_trans->num_writers) == 1);

C
Chris Mason 已提交
1578 1579 1580 1581 1582 1583 1584
	/*
	 * the reloc mutex makes sure that we stop
	 * the balancing code from coming in and moving
	 * extents around in the middle of the commit
	 */
	mutex_lock(&root->fs_info->reloc_mutex);

1585 1586 1587 1588 1589 1590
	/*
	 * We needn't worry about the delayed items because we will
	 * deal with them in create_pending_snapshot(), which is the
	 * core function of the snapshot creation.
	 */
	ret = create_pending_snapshots(trans, root->fs_info);
1591 1592 1593 1594
	if (ret) {
		mutex_unlock(&root->fs_info->reloc_mutex);
		goto cleanup_transaction;
	}
1595

1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606
	/*
	 * We insert the dir indexes of the snapshots and update the inode
	 * of the snapshots' parents after the snapshot creation, so there
	 * are some delayed items which are not dealt with. Now deal with
	 * them.
	 *
	 * We needn't worry that this operation will corrupt the snapshots,
	 * because all the tree which are snapshoted will be forced to COW
	 * the nodes and leaves.
	 */
	ret = btrfs_run_delayed_items(trans, root);
1607 1608 1609 1610
	if (ret) {
		mutex_unlock(&root->fs_info->reloc_mutex);
		goto cleanup_transaction;
	}
1611

1612
	ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1);
1613 1614 1615 1616
	if (ret) {
		mutex_unlock(&root->fs_info->reloc_mutex);
		goto cleanup_transaction;
	}
1617

1618 1619 1620 1621 1622 1623
	/*
	 * make sure none of the code above managed to slip in a
	 * delayed item
	 */
	btrfs_assert_delayed_root_empty(root);

C
Chris Mason 已提交
1624
	WARN_ON(cur_trans != trans->transaction);
C
Chris Mason 已提交
1625

A
Arne Jansen 已提交
1626
	btrfs_scrub_pause(root);
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
	/* btrfs_commit_tree_roots is responsible for getting the
	 * various roots consistent with each other.  Every pointer
	 * in the tree of tree roots has to point to the most up to date
	 * root for every subvolume and other tree.  So, we have to keep
	 * the tree logging code from jumping in and changing any
	 * of the trees.
	 *
	 * At this point in the commit, there can't be any tree-log
	 * writers, but a little lower down we drop the trans mutex
	 * and let new people in.  By holding the tree_log_mutex
	 * from now until after the super is written, we avoid races
	 * with the tree-log code.
	 */
	mutex_lock(&root->fs_info->tree_log_mutex);

1642
	ret = commit_fs_roots(trans, root);
1643 1644
	if (ret) {
		mutex_unlock(&root->fs_info->tree_log_mutex);
1645
		mutex_unlock(&root->fs_info->reloc_mutex);
1646 1647
		goto cleanup_transaction;
	}
1648

1649
	/* commit_fs_roots gets rid of all the tree log roots, it is now
1650 1651 1652 1653
	 * safe to free the root of tree log roots
	 */
	btrfs_free_log_root_tree(trans, root->fs_info);

1654
	ret = commit_cowonly_roots(trans, root);
1655 1656
	if (ret) {
		mutex_unlock(&root->fs_info->tree_log_mutex);
1657
		mutex_unlock(&root->fs_info->reloc_mutex);
1658 1659
		goto cleanup_transaction;
	}
1660

1661 1662
	btrfs_prepare_extent_commit(trans, root);

C
Chris Mason 已提交
1663
	cur_trans = root->fs_info->running_transaction;
1664 1665 1666

	btrfs_set_root_node(&root->fs_info->tree_root->root_item,
			    root->fs_info->tree_root->node);
J
Josef Bacik 已提交
1667
	switch_commit_root(root->fs_info->tree_root);
1668 1669 1670

	btrfs_set_root_node(&root->fs_info->chunk_root->root_item,
			    root->fs_info->chunk_root->node);
J
Josef Bacik 已提交
1671
	switch_commit_root(root->fs_info->chunk_root);
1672

1673
	assert_qgroups_uptodate(trans);
1674
	update_super_roots(root);
1675 1676

	if (!root->fs_info->log_root_recovering) {
1677 1678
		btrfs_set_super_log_root(root->fs_info->super_copy, 0);
		btrfs_set_super_log_root_level(root->fs_info->super_copy, 0);
1679 1680
	}

1681 1682
	memcpy(root->fs_info->super_for_commit, root->fs_info->super_copy,
	       sizeof(*root->fs_info->super_copy));
C
Chris Mason 已提交
1683

1684
	trans->transaction->blocked = 0;
J
Josef Bacik 已提交
1685 1686 1687 1688
	spin_lock(&root->fs_info->trans_lock);
	root->fs_info->running_transaction = NULL;
	root->fs_info->trans_no_join = 0;
	spin_unlock(&root->fs_info->trans_lock);
C
Chris Mason 已提交
1689
	mutex_unlock(&root->fs_info->reloc_mutex);
1690

1691
	wake_up(&root->fs_info->transaction_wait);
1692

C
Chris Mason 已提交
1693
	ret = btrfs_write_and_wait_transaction(trans, root);
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705
	if (ret) {
		btrfs_error(root->fs_info, ret,
			    "Error while writing out transaction.");
		mutex_unlock(&root->fs_info->tree_log_mutex);
		goto cleanup_transaction;
	}

	ret = write_ctree_super(trans, root, 0);
	if (ret) {
		mutex_unlock(&root->fs_info->tree_log_mutex);
		goto cleanup_transaction;
	}
1706

1707 1708 1709 1710 1711 1712
	/*
	 * the super is written, we can safely allow the tree-loggers
	 * to go about their business
	 */
	mutex_unlock(&root->fs_info->tree_log_mutex);

1713
	btrfs_finish_extent_commit(trans, root);
1714

C
Chris Mason 已提交
1715
	cur_trans->commit_done = 1;
1716

1717
	root->fs_info->last_trans_committed = cur_trans->transid;
J
Josef Bacik 已提交
1718

C
Chris Mason 已提交
1719
	wake_up(&cur_trans->commit_wait);
1720

J
Josef Bacik 已提交
1721
	spin_lock(&root->fs_info->trans_lock);
1722
	list_del_init(&cur_trans->list);
J
Josef Bacik 已提交
1723 1724
	spin_unlock(&root->fs_info->trans_lock);

C
Chris Mason 已提交
1725
	put_transaction(cur_trans);
C
Chris Mason 已提交
1726
	put_transaction(cur_trans);
1727

1728 1729
	if (trans->type < TRANS_JOIN_NOLOCK)
		sb_end_intwrite(root->fs_info->sb);
1730

1731 1732
	trace_btrfs_transaction_commit(root);

A
Arne Jansen 已提交
1733 1734
	btrfs_scrub_continue(root);

J
Josef Bacik 已提交
1735 1736 1737
	if (current->journal_info == trans)
		current->journal_info = NULL;

C
Chris Mason 已提交
1738
	kmem_cache_free(btrfs_trans_handle_cachep, trans);
Y
Yan, Zheng 已提交
1739 1740 1741 1742

	if (current != root->fs_info->transaction_kthread)
		btrfs_run_delayed_iputs(root);

C
Chris Mason 已提交
1743
	return ret;
1744 1745

cleanup_transaction:
1746 1747
	btrfs_trans_release_metadata(trans, root);
	trans->block_rsv = NULL;
1748 1749 1750 1751
	btrfs_printk(root->fs_info, "Skipping commit of aborted transaction.\n");
//	WARN_ON(1);
	if (current->journal_info == trans)
		current->journal_info = NULL;
1752
	cleanup_transaction(trans, root, ret);
1753 1754

	return ret;
C
Chris Mason 已提交
1755 1756
}

C
Chris Mason 已提交
1757 1758 1759
/*
 * interface function to delete all the snapshots we have scheduled for deletion
 */
1760 1761
int btrfs_clean_old_snapshots(struct btrfs_root *root)
{
1762 1763 1764
	LIST_HEAD(list);
	struct btrfs_fs_info *fs_info = root->fs_info;

J
Josef Bacik 已提交
1765
	spin_lock(&fs_info->trans_lock);
1766
	list_splice_init(&fs_info->dead_roots, &list);
J
Josef Bacik 已提交
1767
	spin_unlock(&fs_info->trans_lock);
1768

1769
	while (!list_empty(&list)) {
1770 1771
		int ret;

1772
		root = list_entry(list.next, struct btrfs_root, root_list);
1773 1774
		list_del(&root->root_list);

1775 1776
		btrfs_kill_all_delayed_nodes(root);

1777 1778
		if (btrfs_header_backref_rev(root->node) <
		    BTRFS_MIXED_BACKREF_REV)
1779
			ret = btrfs_drop_snapshot(root, NULL, 0, 0);
1780
		else
1781 1782
			ret =btrfs_drop_snapshot(root, NULL, 1, 0);
		BUG_ON(ret < 0);
1783 1784 1785
	}
	return 0;
}