transaction.c 30.1 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>
C
Chris Mason 已提交
20
#include <linux/sched.h>
21
#include <linux/writeback.h>
22
#include <linux/pagemap.h>
23
#include <linux/blkdev.h>
C
Chris Mason 已提交
24 25 26
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
27
#include "locking.h"
Y
Yan Zheng 已提交
28
#include "ref-cache.h"
29
#include "tree-log.h"
C
Chris Mason 已提交
30

C
Chris Mason 已提交
31
static int total_trans = 0;
C
Chris Mason 已提交
32 33 34
extern struct kmem_cache *btrfs_trans_handle_cachep;
extern struct kmem_cache *btrfs_transaction_cachep;

35 36
#define BTRFS_ROOT_TRANS_TAG 0

37
static noinline void put_transaction(struct btrfs_transaction *transaction)
C
Chris Mason 已提交
38
{
C
Chris Mason 已提交
39
	WARN_ON(transaction->use_count == 0);
C
Chris Mason 已提交
40
	transaction->use_count--;
C
Chris Mason 已提交
41 42 43
	if (transaction->use_count == 0) {
		WARN_ON(total_trans == 0);
		total_trans--;
C
Chris Mason 已提交
44
		list_del_init(&transaction->list);
C
Chris Mason 已提交
45 46
		memset(transaction, 0, sizeof(*transaction));
		kmem_cache_free(btrfs_transaction_cachep, transaction);
C
Chris Mason 已提交
47
	}
C
Chris Mason 已提交
48 49
}

C
Chris Mason 已提交
50 51 52
/*
 * either allocate a new transaction or hop into the existing one
 */
53
static noinline int join_transaction(struct btrfs_root *root)
C
Chris Mason 已提交
54 55 56 57
{
	struct btrfs_transaction *cur_trans;
	cur_trans = root->fs_info->running_transaction;
	if (!cur_trans) {
C
Chris Mason 已提交
58 59
		cur_trans = kmem_cache_alloc(btrfs_transaction_cachep,
					     GFP_NOFS);
C
Chris Mason 已提交
60
		total_trans++;
C
Chris Mason 已提交
61
		BUG_ON(!cur_trans);
62
		root->fs_info->generation++;
63
		root->fs_info->last_alloc = 0;
64
		root->fs_info->last_data_alloc = 0;
65 66
		cur_trans->num_writers = 1;
		cur_trans->num_joined = 0;
67
		cur_trans->transid = root->fs_info->generation;
C
Chris Mason 已提交
68 69 70
		init_waitqueue_head(&cur_trans->writer_wait);
		init_waitqueue_head(&cur_trans->commit_wait);
		cur_trans->in_commit = 0;
71
		cur_trans->blocked = 0;
72
		cur_trans->use_count = 1;
C
Chris Mason 已提交
73
		cur_trans->commit_done = 0;
C
Chris Mason 已提交
74
		cur_trans->start_time = get_seconds();
75
		INIT_LIST_HEAD(&cur_trans->pending_snapshots);
C
Chris Mason 已提交
76
		list_add_tail(&cur_trans->list, &root->fs_info->trans_list);
77
		extent_io_tree_init(&cur_trans->dirty_pages,
78 79
				     root->fs_info->btree_inode->i_mapping,
				     GFP_NOFS);
80 81 82
		spin_lock(&root->fs_info->new_trans_lock);
		root->fs_info->running_transaction = cur_trans;
		spin_unlock(&root->fs_info->new_trans_lock);
83 84 85
	} else {
		cur_trans->num_writers++;
		cur_trans->num_joined++;
C
Chris Mason 已提交
86
	}
87

C
Chris Mason 已提交
88 89 90
	return 0;
}

C
Chris Mason 已提交
91 92 93 94 95 96
/*
 * 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
 */
97
noinline int btrfs_record_root_in_trans(struct btrfs_root *root)
98
{
99
	struct btrfs_dirty_root *dirty;
100 101 102 103 104 105 106
	u64 running_trans_id = root->fs_info->running_transaction->transid;
	if (root->ref_cows && root->last_trans < running_trans_id) {
		WARN_ON(root == root->fs_info->extent_root);
		if (root->root_item.refs != 0) {
			radix_tree_tag_set(&root->fs_info->fs_roots_radix,
				   (unsigned long)root->root_key.objectid,
				   BTRFS_ROOT_TRANS_TAG);
Y
Yan Zheng 已提交
107 108 109 110 111 112 113 114

			dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
			BUG_ON(!dirty);
			dirty->root = kmalloc(sizeof(*dirty->root), GFP_NOFS);
			BUG_ON(!dirty->root);
			dirty->latest_root = root;
			INIT_LIST_HEAD(&dirty->list);

115
			root->commit_root = btrfs_root_node(root);
Y
Yan Zheng 已提交
116 117 118

			memcpy(dirty->root, root, sizeof(*root));
			spin_lock_init(&dirty->root->node_lock);
119
			spin_lock_init(&dirty->root->list_lock);
Y
Yan Zheng 已提交
120
			mutex_init(&dirty->root->objectid_mutex);
121
			mutex_init(&dirty->root->log_mutex);
122
			INIT_LIST_HEAD(&dirty->root->dead_list);
Y
Yan Zheng 已提交
123 124
			dirty->root->node = root->commit_root;
			dirty->root->commit_root = NULL;
125 126 127 128 129 130

			spin_lock(&root->list_lock);
			list_add(&dirty->root->dead_list, &root->dead_list);
			spin_unlock(&root->list_lock);

			root->dirty_root = dirty;
131 132 133 134 135 136 137 138
		} else {
			WARN_ON(1);
		}
		root->last_trans = running_trans_id;
	}
	return 0;
}

C
Chris Mason 已提交
139 140 141 142
/* 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
143
static void wait_current_trans(struct btrfs_root *root)
C
Chris Mason 已提交
144
{
145
	struct btrfs_transaction *cur_trans;
C
Chris Mason 已提交
146

147
	cur_trans = root->fs_info->running_transaction;
C
Chris Mason 已提交
148
	if (cur_trans && cur_trans->blocked) {
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167
		DEFINE_WAIT(wait);
		cur_trans->use_count++;
		while(1) {
			prepare_to_wait(&root->fs_info->transaction_wait, &wait,
					TASK_UNINTERRUPTIBLE);
			if (cur_trans->blocked) {
				mutex_unlock(&root->fs_info->trans_mutex);
				schedule();
				mutex_lock(&root->fs_info->trans_mutex);
				finish_wait(&root->fs_info->transaction_wait,
					    &wait);
			} else {
				finish_wait(&root->fs_info->transaction_wait,
					    &wait);
				break;
			}
		}
		put_transaction(cur_trans);
	}
C
Chris Mason 已提交
168 169
}

170
static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root,
171
					     int num_blocks, int wait)
C
Chris Mason 已提交
172 173 174 175 176 177
{
	struct btrfs_trans_handle *h =
		kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS);
	int ret;

	mutex_lock(&root->fs_info->trans_mutex);
C
Chris Mason 已提交
178 179
	if (!root->fs_info->log_root_recovering &&
	    ((wait == 1 && !root->fs_info->open_ioctl_trans) || wait == 2))
C
Chris Mason 已提交
180
		wait_current_trans(root);
C
Chris Mason 已提交
181 182
	ret = join_transaction(root);
	BUG_ON(ret);
183

184
	btrfs_record_root_in_trans(root);
185
	h->transid = root->fs_info->running_transaction->transid;
C
Chris Mason 已提交
186 187 188
	h->transaction = root->fs_info->running_transaction;
	h->blocks_reserved = num_blocks;
	h->blocks_used = 0;
189
	h->block_group = NULL;
190 191
	h->alloc_exclude_nr = 0;
	h->alloc_exclude_start = 0;
C
Chris Mason 已提交
192 193 194 195 196
	root->fs_info->running_transaction->use_count++;
	mutex_unlock(&root->fs_info->trans_mutex);
	return h;
}

197 198 199
struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root,
						   int num_blocks)
{
200
	return start_transaction(root, num_blocks, 1);
201 202 203 204
}
struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root,
						   int num_blocks)
{
205
	return start_transaction(root, num_blocks, 0);
206 207
}

208 209 210 211 212 213
struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r,
							 int num_blocks)
{
	return start_transaction(r, num_blocks, 2);
}

C
Chris Mason 已提交
214
/* wait for a transaction commit to be fully complete */
215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
static noinline int wait_for_commit(struct btrfs_root *root,
				    struct btrfs_transaction *commit)
{
	DEFINE_WAIT(wait);
	mutex_lock(&root->fs_info->trans_mutex);
	while(!commit->commit_done) {
		prepare_to_wait(&commit->commit_wait, &wait,
				TASK_UNINTERRUPTIBLE);
		if (commit->commit_done)
			break;
		mutex_unlock(&root->fs_info->trans_mutex);
		schedule();
		mutex_lock(&root->fs_info->trans_mutex);
	}
	mutex_unlock(&root->fs_info->trans_mutex);
	finish_wait(&commit->commit_wait, &wait);
	return 0;
}

C
Chris Mason 已提交
234 235 236 237
/*
 * rate limit against the drop_snapshot code.  This helps to slow down new operations
 * if the drop_snapshot code isn't able to keep up.
 */
C
Chris Mason 已提交
238
static void throttle_on_drops(struct btrfs_root *root)
239 240
{
	struct btrfs_fs_info *info = root->fs_info;
C
Chris Mason 已提交
241
	int harder_count = 0;
242

C
Chris Mason 已提交
243
harder:
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
	if (atomic_read(&info->throttles)) {
		DEFINE_WAIT(wait);
		int thr;
		thr = atomic_read(&info->throttle_gen);

		do {
			prepare_to_wait(&info->transaction_throttle,
					&wait, TASK_UNINTERRUPTIBLE);
			if (!atomic_read(&info->throttles)) {
				finish_wait(&info->transaction_throttle, &wait);
				break;
			}
			schedule();
			finish_wait(&info->transaction_throttle, &wait);
		} while (thr == atomic_read(&info->throttle_gen));
C
Chris Mason 已提交
259 260 261 262 263 264 265 266 267 268 269 270 271
		harder_count++;

		if (root->fs_info->total_ref_cache_size > 1 * 1024 * 1024 &&
		    harder_count < 2)
			goto harder;

		if (root->fs_info->total_ref_cache_size > 5 * 1024 * 1024 &&
		    harder_count < 10)
			goto harder;

		if (root->fs_info->total_ref_cache_size > 10 * 1024 * 1024 &&
		    harder_count < 20)
			goto harder;
272 273 274
	}
}

C
Chris Mason 已提交
275 276 277
void btrfs_throttle(struct btrfs_root *root)
{
	mutex_lock(&root->fs_info->trans_mutex);
278 279
	if (!root->fs_info->open_ioctl_trans)
		wait_current_trans(root);
C
Chris Mason 已提交
280 281 282 283 284
	mutex_unlock(&root->fs_info->trans_mutex);

	throttle_on_drops(root);
}

285 286
static int __btrfs_end_transaction(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root, int throttle)
C
Chris Mason 已提交
287 288
{
	struct btrfs_transaction *cur_trans;
289
	struct btrfs_fs_info *info = root->fs_info;
290

291 292
	mutex_lock(&info->trans_mutex);
	cur_trans = info->running_transaction;
C
Chris Mason 已提交
293
	WARN_ON(cur_trans != trans->transaction);
294
	WARN_ON(cur_trans->num_writers < 1);
C
Chris Mason 已提交
295
	cur_trans->num_writers--;
296

C
Chris Mason 已提交
297 298 299
	if (waitqueue_active(&cur_trans->writer_wait))
		wake_up(&cur_trans->writer_wait);
	put_transaction(cur_trans);
300
	mutex_unlock(&info->trans_mutex);
C
Chris Mason 已提交
301
	memset(trans, 0, sizeof(*trans));
C
Chris Mason 已提交
302
	kmem_cache_free(btrfs_trans_handle_cachep, trans);
303 304

	if (throttle)
C
Chris Mason 已提交
305
		throttle_on_drops(root);
306

C
Chris Mason 已提交
307 308 309
	return 0;
}

310 311 312 313 314 315 316 317 318 319 320 321
int btrfs_end_transaction(struct btrfs_trans_handle *trans,
			  struct btrfs_root *root)
{
	return __btrfs_end_transaction(trans, root, 0);
}

int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root)
{
	return __btrfs_end_transaction(trans, root, 1);
}

C
Chris Mason 已提交
322 323 324 325 326
/*
 * 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
 */
327 328
int btrfs_write_and_wait_marked_extents(struct btrfs_root *root,
					struct extent_io_tree *dirty_pages)
C
Chris Mason 已提交
329
{
330
	int ret;
331
	int err = 0;
332 333 334
	int werr = 0;
	struct page *page;
	struct inode *btree_inode = root->fs_info->btree_inode;
335
	struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree;
336
	u64 start = 0;
337 338
	u64 end;
	unsigned long index;
339 340

	while(1) {
341
		ret = find_first_extent_bit(dirty_pages, start, &start, &end,
342 343
					    EXTENT_DIRTY);
		if (ret)
344
			break;
345
		while(start <= end) {
346 347
			cond_resched();

348
			index = start >> PAGE_CACHE_SHIFT;
349
			start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
C
Chris Mason 已提交
350
			page = find_get_page(btree_inode->i_mapping, index);
351 352
			if (!page)
				continue;
C
Chris Mason 已提交
353 354 355 356 357 358 359 360

			btree_lock_page_hook(page);
			if (!page->mapping) {
				unlock_page(page);
				page_cache_release(page);
				continue;
			}

361 362 363 364 365 366 367 368 369
			if (PageWriteback(page)) {
				if (PageDirty(page))
					wait_on_page_writeback(page);
				else {
					unlock_page(page);
					page_cache_release(page);
					continue;
				}
			}
370 371 372 373 374 375
			err = write_one_page(page, 0);
			if (err)
				werr = err;
			page_cache_release(page);
		}
	}
376 377 378 379 380
	/*
	 * we unplug once and then use the wait_on_extent_bit for
	 * everything else
	 */
	blk_run_address_space(btree_inode->i_mapping);
381 382 383 384 385 386 387 388 389 390 391 392 393 394
	while(1) {
		ret = find_first_extent_bit(dirty_pages, 0, &start, &end,
					    EXTENT_DIRTY);
		if (ret)
			break;

		clear_extent_dirty(dirty_pages, start, end, GFP_NOFS);
		while(start <= end) {
			index = start >> PAGE_CACHE_SHIFT;
			start = (u64)(index + 1) << PAGE_CACHE_SHIFT;
			page = find_get_page(btree_inode->i_mapping, index);
			if (!page)
				continue;
			if (PageDirty(page)) {
C
Chris Mason 已提交
395 396
				btree_lock_page_hook(page);
				wait_on_page_writeback(page);
397 398 399 400
				err = write_one_page(page, 0);
				if (err)
					werr = err;
			}
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
			if (PageWriteback(page)) {
				/*
				 * we don't wait on the page writeback bit
				 * because that triggers a lot of unplugs.
				 * The extent bits are much nicer to
				 * the disks, but come with a slightly
				 * higher latency because we aren't forcing
				 * unplugs.
				 */
				wait_on_extent_writeback(io_tree,
					 page_offset(page),
					 page_offset(page) +
					 PAGE_CACHE_SIZE - 1);
			}
			if (PageWriteback(page)) {
				/*
				 * the state bits get cleared before the
				 * page bits, lets add some extra
				 * paranoia here
				 */
				wait_on_page_writeback(page);
			}
423 424 425 426
			page_cache_release(page);
			cond_resched();
		}
	}
427 428 429
	if (err)
		werr = err;
	return werr;
C
Chris Mason 已提交
430 431
}

432 433 434 435 436 437 438 439 440 441 442 443
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,
					   &trans->transaction->dirty_pages);
}

C
Chris Mason 已提交
444 445 446 447 448 449 450 451 452 453
/*
 * 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.
 */
454 455
static int update_cowonly_root(struct btrfs_trans_handle *trans,
			       struct btrfs_root *root)
C
Chris Mason 已提交
456 457
{
	int ret;
458 459
	u64 old_root_bytenr;
	struct btrfs_root *tree_root = root->fs_info->tree_root;
C
Chris Mason 已提交
460

461
	btrfs_extent_post_op(trans, root);
462
	btrfs_write_dirty_block_groups(trans, root);
463 464
	btrfs_extent_post_op(trans, root);

C
Chris Mason 已提交
465
	while(1) {
466 467
		old_root_bytenr = btrfs_root_bytenr(&root->root_item);
		if (old_root_bytenr == root->node->start)
C
Chris Mason 已提交
468
			break;
469 470 471 472
		btrfs_set_root_bytenr(&root->root_item,
				       root->node->start);
		btrfs_set_root_level(&root->root_item,
				     btrfs_header_level(root->node));
473
		btrfs_set_root_generation(&root->root_item, trans->transid);
474 475 476

		btrfs_extent_post_op(trans, root);

C
Chris Mason 已提交
477
		ret = btrfs_update_root(trans, tree_root,
478 479
					&root->root_key,
					&root->root_item);
C
Chris Mason 已提交
480
		BUG_ON(ret);
481
		btrfs_write_dirty_block_groups(trans, root);
482
		btrfs_extent_post_op(trans, root);
483 484 485 486
	}
	return 0;
}

C
Chris Mason 已提交
487 488 489
/*
 * update all the cowonly tree roots on disk
 */
490 491 492 493 494
int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct list_head *next;
495 496
	struct extent_buffer *eb;

497 498
	btrfs_extent_post_op(trans, fs_info->tree_root);

499 500 501 502
	eb = btrfs_lock_root_node(fs_info->tree_root);
	btrfs_cow_block(trans, fs_info->tree_root, eb, NULL, 0, &eb, 0);
	btrfs_tree_unlock(eb);
	free_extent_buffer(eb);
503

504 505
	btrfs_extent_post_op(trans, fs_info->tree_root);

506 507 508 509
	while(!list_empty(&fs_info->dirty_cowonly_roots)) {
		next = fs_info->dirty_cowonly_roots.next;
		list_del_init(next);
		root = list_entry(next, struct btrfs_root, dirty_list);
510

511
		update_cowonly_root(trans, root);
C
Chris Mason 已提交
512 513 514 515
	}
	return 0;
}

C
Chris Mason 已提交
516 517 518 519 520
/*
 * 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
 */
Y
Yan Zheng 已提交
521
int btrfs_add_dead_root(struct btrfs_root *root, struct btrfs_root *latest)
522
{
523
	struct btrfs_dirty_root *dirty;
524 525 526 527 528

	dirty = kmalloc(sizeof(*dirty), GFP_NOFS);
	if (!dirty)
		return -ENOMEM;
	dirty->root = root;
529
	dirty->latest_root = latest;
Y
Yan Zheng 已提交
530 531 532 533

	mutex_lock(&root->fs_info->trans_mutex);
	list_add(&dirty->list, &latest->fs_info->dead_roots);
	mutex_unlock(&root->fs_info->trans_mutex);
534 535 536
	return 0;
}

C
Chris Mason 已提交
537 538 539 540 541 542
/*
 * at transaction commit time we need to schedule the old roots for
 * deletion via btrfs_drop_snapshot.  This runs through all the
 * reference counted roots that were modified in the current
 * transaction and puts them into the drop list
 */
543 544 545
static noinline int add_dirty_roots(struct btrfs_trans_handle *trans,
				    struct radix_tree_root *radix,
				    struct list_head *list)
546
{
547
	struct btrfs_dirty_root *dirty;
548 549 550 551
	struct btrfs_root *gang[8];
	struct btrfs_root *root;
	int i;
	int ret;
552
	int err = 0;
553
	u32 refs;
554

555 556 557 558 559 560 561 562
	while(1) {
		ret = radix_tree_gang_lookup_tag(radix, (void **)gang, 0,
						 ARRAY_SIZE(gang),
						 BTRFS_ROOT_TRANS_TAG);
		if (ret == 0)
			break;
		for (i = 0; i < ret; i++) {
			root = gang[i];
C
Chris Mason 已提交
563 564 565
			radix_tree_tag_clear(radix,
				     (unsigned long)root->root_key.objectid,
				     BTRFS_ROOT_TRANS_TAG);
Y
Yan Zheng 已提交
566 567

			BUG_ON(!root->ref_tree);
C
Chris Mason 已提交
568
			dirty = root->dirty_root;
Y
Yan Zheng 已提交
569

570
			btrfs_free_log(trans, root);
Y
Yan Zheng 已提交
571
			btrfs_free_reloc_root(trans, root);
572

573
			if (root->commit_root == root->node) {
574 575
				WARN_ON(root->node->start !=
					btrfs_root_bytenr(&root->root_item));
Y
Yan Zheng 已提交
576

577
				free_extent_buffer(root->commit_root);
578
				root->commit_root = NULL;
579
				root->dirty_root = NULL;
580 581 582 583 584

				spin_lock(&root->list_lock);
				list_del_init(&dirty->root->dead_list);
				spin_unlock(&root->list_lock);

Y
Yan Zheng 已提交
585 586
				kfree(dirty->root);
				kfree(dirty);
587 588 589 590 591 592 593 594 595

				/* make sure to update the root on disk
				 * so we get any updates to the block used
				 * counts
				 */
				err = btrfs_update_root(trans,
						root->fs_info->tree_root,
						&root->root_key,
						&root->root_item);
596 597
				continue;
			}
598 599 600 601

			memset(&root->root_item.drop_progress, 0,
			       sizeof(struct btrfs_disk_key));
			root->root_item.drop_level = 0;
602
			root->commit_root = NULL;
603
			root->dirty_root = NULL;
604
			root->root_key.offset = root->fs_info->generation;
605 606 607 608
			btrfs_set_root_bytenr(&root->root_item,
					      root->node->start);
			btrfs_set_root_level(&root->root_item,
					     btrfs_header_level(root->node));
609 610 611
			btrfs_set_root_generation(&root->root_item,
						  root->root_key.offset);

612 613 614
			err = btrfs_insert_root(trans, root->fs_info->tree_root,
						&root->root_key,
						&root->root_item);
615 616
			if (err)
				break;
617 618 619

			refs = btrfs_root_refs(&dirty->root->root_item);
			btrfs_set_root_refs(&dirty->root->root_item, refs - 1);
620
			err = btrfs_update_root(trans, root->fs_info->tree_root,
621 622
						&dirty->root->root_key,
						&dirty->root->root_item);
623 624

			BUG_ON(err);
625
			if (refs == 1) {
626
				list_add(&dirty->list, list);
627 628
			} else {
				WARN_ON(1);
Y
Yan Zheng 已提交
629
				free_extent_buffer(dirty->root->node);
630
				kfree(dirty->root);
631
				kfree(dirty);
632
			}
633 634
		}
	}
635
	return err;
636 637
}

C
Chris Mason 已提交
638 639 640 641
/*
 * defrag a given btree.  If cacheonly == 1, this won't read from the disk,
 * otherwise every leaf in the btree is read and defragged.
 */
642 643 644 645 646
int btrfs_defrag_root(struct btrfs_root *root, int cacheonly)
{
	struct btrfs_fs_info *info = root->fs_info;
	int ret;
	struct btrfs_trans_handle *trans;
647
	unsigned long nr;
648

649
	smp_mb();
650 651 652
	if (root->defrag_running)
		return 0;
	trans = btrfs_start_transaction(root, 1);
653
	while (1) {
654 655
		root->defrag_running = 1;
		ret = btrfs_defrag_leaves(trans, root, cacheonly);
656
		nr = trans->blocks_used;
657
		btrfs_end_transaction(trans, root);
658
		btrfs_btree_balance_dirty(info->tree_root, nr);
659 660 661
		cond_resched();

		trans = btrfs_start_transaction(root, 1);
662
		if (root->fs_info->closing || ret != -EAGAIN)
663 664 665
			break;
	}
	root->defrag_running = 0;
666
	smp_mb();
667 668 669 670
	btrfs_end_transaction(trans, root);
	return 0;
}

C
Chris Mason 已提交
671 672 673 674
/*
 * Given a list of roots that need to be deleted, call btrfs_drop_snapshot on
 * all of them
 */
675 676
static noinline int drop_dirty_roots(struct btrfs_root *tree_root,
				     struct list_head *list)
677
{
678
	struct btrfs_dirty_root *dirty;
679
	struct btrfs_trans_handle *trans;
680
	unsigned long nr;
681 682
	u64 num_bytes;
	u64 bytes_used;
683
	u64 max_useless;
684
	int ret = 0;
685 686
	int err;

687
	while(!list_empty(list)) {
688 689
		struct btrfs_root *root;

690
		dirty = list_entry(list->prev, struct btrfs_dirty_root, list);
691
		list_del_init(&dirty->list);
692

693
		num_bytes = btrfs_root_used(&dirty->root->root_item);
694
		root = dirty->latest_root;
695
		atomic_inc(&root->fs_info->throttles);
696

697 698
		while(1) {
			trans = btrfs_start_transaction(tree_root, 1);
699
			mutex_lock(&root->fs_info->drop_mutex);
700 701 702 703
			ret = btrfs_drop_snapshot(trans, dirty->root);
			if (ret != -EAGAIN) {
				break;
			}
704
			mutex_unlock(&root->fs_info->drop_mutex);
705

706 707 708 709 710 711
			err = btrfs_update_root(trans,
					tree_root,
					&dirty->root->root_key,
					&dirty->root->root_item);
			if (err)
				ret = err;
712
			nr = trans->blocks_used;
C
Chris Mason 已提交
713
			ret = btrfs_end_transaction(trans, tree_root);
714
			BUG_ON(ret);
715

716
			btrfs_btree_balance_dirty(tree_root, nr);
717
			cond_resched();
718
		}
719
		BUG_ON(ret);
720
		atomic_dec(&root->fs_info->throttles);
C
Chris Mason 已提交
721
		wake_up(&root->fs_info->transaction_throttle);
722

723 724 725
		num_bytes -= btrfs_root_used(&dirty->root->root_item);
		bytes_used = btrfs_root_used(&root->root_item);
		if (num_bytes) {
726
			btrfs_record_root_in_trans(root);
727
			btrfs_set_root_used(&root->root_item,
728
					    bytes_used - num_bytes);
729
		}
730

731
		ret = btrfs_del_root(trans, tree_root, &dirty->root->root_key);
732 733
		if (ret) {
			BUG();
734
			break;
735
		}
736 737
		mutex_unlock(&root->fs_info->drop_mutex);

738 739 740 741 742 743 744 745 746 747 748 749
		spin_lock(&root->list_lock);
		list_del_init(&dirty->root->dead_list);
		if (!list_empty(&root->dead_list)) {
			struct btrfs_root *oldest;
			oldest = list_entry(root->dead_list.prev,
					    struct btrfs_root, dead_list);
			max_useless = oldest->root_key.offset - 1;
		} else {
			max_useless = root->root_key.offset - 1;
		}
		spin_unlock(&root->list_lock);

750
		nr = trans->blocks_used;
751 752
		ret = btrfs_end_transaction(trans, tree_root);
		BUG_ON(ret);
753

Z
Zheng Yan 已提交
754
		ret = btrfs_remove_leaf_refs(root, max_useless, 0);
755 756
		BUG_ON(ret);

757
		free_extent_buffer(dirty->root->node);
758
		kfree(dirty->root);
759
		kfree(dirty);
760 761

		btrfs_btree_balance_dirty(tree_root, nr);
762
		cond_resched();
763
	}
764
	return ret;
765 766
}

C
Chris Mason 已提交
767 768 769 770
/*
 * new snapshots need to be created at a very specific time in the
 * transaction commit.  This does the actual creation
 */
771
static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
772 773 774 775
				   struct btrfs_fs_info *fs_info,
				   struct btrfs_pending_snapshot *pending)
{
	struct btrfs_key key;
776
	struct btrfs_root_item *new_root_item;
777 778 779
	struct btrfs_root *tree_root = fs_info->tree_root;
	struct btrfs_root *root = pending->root;
	struct extent_buffer *tmp;
780
	struct extent_buffer *old;
781 782 783
	int ret;
	u64 objectid;

784 785 786 787 788
	new_root_item = kmalloc(sizeof(*new_root_item), GFP_NOFS);
	if (!new_root_item) {
		ret = -ENOMEM;
		goto fail;
	}
789 790 791 792
	ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid);
	if (ret)
		goto fail;

Y
Yan Zheng 已提交
793 794
	btrfs_record_root_in_trans(root);
	btrfs_set_root_last_snapshot(&root->root_item, trans->transid);
795
	memcpy(new_root_item, &root->root_item, sizeof(*new_root_item));
796 797

	key.objectid = objectid;
798
	key.offset = trans->transid;
799 800
	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);

801
	old = btrfs_lock_root_node(root);
802
	btrfs_cow_block(trans, root, old, NULL, 0, &old, 0);
803

804 805 806
	btrfs_copy_root(trans, root, old, &tmp, objectid);
	btrfs_tree_unlock(old);
	free_extent_buffer(old);
807

808 809
	btrfs_set_root_bytenr(new_root_item, tmp->start);
	btrfs_set_root_level(new_root_item, btrfs_header_level(tmp));
810
	btrfs_set_root_generation(new_root_item, trans->transid);
811
	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
812
				new_root_item);
813
	btrfs_tree_unlock(tmp);
814 815 816 817
	free_extent_buffer(tmp);
	if (ret)
		goto fail;

818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
	key.offset = (u64)-1;
	memcpy(&pending->root_key, &key, sizeof(key));
fail:
	kfree(new_root_item);
	return ret;
}

static noinline int finish_pending_snapshot(struct btrfs_fs_info *fs_info,
				   struct btrfs_pending_snapshot *pending)
{
	int ret;
	int namelen;
	u64 index = 0;
	struct btrfs_trans_handle *trans;
	struct inode *parent_inode;
	struct inode *inode;

835 836
	parent_inode = pending->dentry->d_parent->d_inode;
	trans = btrfs_start_transaction(BTRFS_I(parent_inode)->root, 1);
837

838 839 840
	/*
	 * insert the directory item
	 */
841
	namelen = strlen(pending->name);
842 843 844 845 846 847
	ret = btrfs_set_inode_index(parent_inode, &index);
	ret = btrfs_insert_dir_item(trans,
			    BTRFS_I(parent_inode)->root,
			    pending->name, namelen,
			    parent_inode->i_ino,
			    &pending->root_key, BTRFS_FT_DIR, index);
848 849 850

	if (ret)
		goto fail;
851
#if 0
852 853
	ret = btrfs_insert_inode_ref(trans, root->fs_info->tree_root,
			     pending->name, strlen(pending->name), objectid,
854
			     root->fs_info->sb->s_root->d_inode->i_ino, 0);
855 856 857
#endif
	inode = btrfs_lookup_dentry(parent_inode, pending->dentry);
	d_instantiate(pending->dentry, inode);
858
fail:
859
	btrfs_end_transaction(trans, fs_info->fs_root);
860 861 862
	return ret;
}

C
Chris Mason 已提交
863 864 865
/*
 * create all the snapshots we've scheduled for creation
 */
866 867
static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans,
					     struct btrfs_fs_info *fs_info)
868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883
{
	struct btrfs_pending_snapshot *pending;
	struct list_head *head = &trans->transaction->pending_snapshots;
	struct list_head *cur;
	int ret;

	list_for_each(cur, head) {
		pending = list_entry(cur, struct btrfs_pending_snapshot, list);
		ret = create_pending_snapshot(trans, fs_info, pending);
		BUG_ON(ret);
	}
	return 0;
}

static noinline int finish_pending_snapshots(struct btrfs_trans_handle *trans,
					     struct btrfs_fs_info *fs_info)
884 885 886 887 888 889 890 891
{
	struct btrfs_pending_snapshot *pending;
	struct list_head *head = &trans->transaction->pending_snapshots;
	int ret;

	while(!list_empty(head)) {
		pending = list_entry(head->next,
				     struct btrfs_pending_snapshot, list);
892
		ret = finish_pending_snapshot(fs_info, pending);
893 894 895 896 897
		BUG_ON(ret);
		list_del(&pending->list);
		kfree(pending->name);
		kfree(pending);
	}
C
Chris Mason 已提交
898 899 900
	return 0;
}

C
Chris Mason 已提交
901 902 903
int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
			     struct btrfs_root *root)
{
904 905
	unsigned long joined = 0;
	unsigned long timeout = 1;
C
Chris Mason 已提交
906
	struct btrfs_transaction *cur_trans;
C
Chris Mason 已提交
907
	struct btrfs_transaction *prev_trans = NULL;
908
	struct btrfs_root *chunk_root = root->fs_info->chunk_root;
909
	struct list_head dirty_fs_roots;
910
	struct extent_io_tree *pinned_copy;
C
Chris Mason 已提交
911
	DEFINE_WAIT(wait);
912
	int ret;
C
Chris Mason 已提交
913

914
	INIT_LIST_HEAD(&dirty_fs_roots);
C
Chris Mason 已提交
915 916 917 918
	mutex_lock(&root->fs_info->trans_mutex);
	if (trans->transaction->in_commit) {
		cur_trans = trans->transaction;
		trans->transaction->use_count++;
C
Chris Mason 已提交
919
		mutex_unlock(&root->fs_info->trans_mutex);
C
Chris Mason 已提交
920
		btrfs_end_transaction(trans, root);
C
Chris Mason 已提交
921

C
Chris Mason 已提交
922 923
		ret = wait_for_commit(root, cur_trans);
		BUG_ON(ret);
924 925

		mutex_lock(&root->fs_info->trans_mutex);
C
Chris Mason 已提交
926
		put_transaction(cur_trans);
927 928
		mutex_unlock(&root->fs_info->trans_mutex);

C
Chris Mason 已提交
929 930
		return 0;
	}
931 932 933 934 935

	pinned_copy = kmalloc(sizeof(*pinned_copy), GFP_NOFS);
	if (!pinned_copy)
		return -ENOMEM;

936
	extent_io_tree_init(pinned_copy,
937 938
			     root->fs_info->btree_inode->i_mapping, GFP_NOFS);

C
Chris Mason 已提交
939
	trans->transaction->in_commit = 1;
940
	trans->transaction->blocked = 1;
C
Chris Mason 已提交
941 942 943 944 945 946 947 948 949 950 951
	cur_trans = trans->transaction;
	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) {
			prev_trans->use_count++;
			mutex_unlock(&root->fs_info->trans_mutex);

			wait_for_commit(root, prev_trans);

			mutex_lock(&root->fs_info->trans_mutex);
952
			put_transaction(prev_trans);
C
Chris Mason 已提交
953 954
		}
	}
955 956

	do {
957
		int snap_pending = 0;
958
		joined = cur_trans->num_joined;
959 960 961
		if (!list_empty(&trans->transaction->pending_snapshots))
			snap_pending = 1;

C
Chris Mason 已提交
962
		WARN_ON(cur_trans != trans->transaction);
963
		prepare_to_wait(&cur_trans->writer_wait, &wait,
C
Chris Mason 已提交
964
				TASK_UNINTERRUPTIBLE);
965 966 967 968 969 970

		if (cur_trans->num_writers > 1)
			timeout = MAX_SCHEDULE_TIMEOUT;
		else
			timeout = 1;

C
Chris Mason 已提交
971
		mutex_unlock(&root->fs_info->trans_mutex);
972

973 974 975 976 977
		if (snap_pending) {
			ret = btrfs_wait_ordered_extents(root, 1);
			BUG_ON(ret);
		}

978 979
		schedule_timeout(timeout);

C
Chris Mason 已提交
980
		mutex_lock(&root->fs_info->trans_mutex);
981 982 983 984
		finish_wait(&cur_trans->writer_wait, &wait);
	} while (cur_trans->num_writers > 1 ||
		 (cur_trans->num_joined != joined));

985 986 987
	ret = create_pending_snapshots(trans, root->fs_info);
	BUG_ON(ret);

C
Chris Mason 已提交
988
	WARN_ON(cur_trans != trans->transaction);
C
Chris Mason 已提交
989

990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003
	/* 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);
Z
Zheng Yan 已提交
1004 1005 1006 1007 1008
	/*
	 * keep tree reloc code from adding new reloc trees
	 */
	mutex_lock(&root->fs_info->tree_reloc_mutex);

1009

1010 1011 1012 1013
	ret = add_dirty_roots(trans, &root->fs_info->fs_roots_radix,
			      &dirty_fs_roots);
	BUG_ON(ret);

1014 1015 1016 1017 1018
	/* add_dirty_roots gets rid of all the tree log roots, it is now
	 * safe to free the root of tree log roots
	 */
	btrfs_free_log_root_tree(trans, root->fs_info);

C
Chris Mason 已提交
1019 1020
	ret = btrfs_commit_tree_roots(trans, root);
	BUG_ON(ret);
1021

C
Chris Mason 已提交
1022
	cur_trans = root->fs_info->running_transaction;
1023
	spin_lock(&root->fs_info->new_trans_lock);
C
Chris Mason 已提交
1024
	root->fs_info->running_transaction = NULL;
1025
	spin_unlock(&root->fs_info->new_trans_lock);
1026 1027 1028
	btrfs_set_super_generation(&root->fs_info->super_copy,
				   cur_trans->transid);
	btrfs_set_super_root(&root->fs_info->super_copy,
1029 1030 1031
			     root->fs_info->tree_root->node->start);
	btrfs_set_super_root_level(&root->fs_info->super_copy,
			   btrfs_header_level(root->fs_info->tree_root->node));
1032

1033 1034 1035 1036
	btrfs_set_super_chunk_root(&root->fs_info->super_copy,
				   chunk_root->node->start);
	btrfs_set_super_chunk_root_level(&root->fs_info->super_copy,
					 btrfs_header_level(chunk_root->node));
1037 1038
	btrfs_set_super_chunk_root_generation(&root->fs_info->super_copy,
				btrfs_header_generation(chunk_root->node));
1039 1040 1041 1042 1043 1044

	if (!root->fs_info->log_root_recovering) {
		btrfs_set_super_log_root(&root->fs_info->super_copy, 0);
		btrfs_set_super_log_root_level(&root->fs_info->super_copy, 0);
	}

1045 1046
	memcpy(&root->fs_info->super_for_commit, &root->fs_info->super_copy,
	       sizeof(root->fs_info->super_copy));
C
Chris Mason 已提交
1047

1048
	btrfs_copy_pinned(root, pinned_copy);
C
Chris Mason 已提交
1049

1050
	trans->transaction->blocked = 0;
1051
	wake_up(&root->fs_info->transaction_throttle);
1052
	wake_up(&root->fs_info->transaction_wait);
1053

C
Chris Mason 已提交
1054
	mutex_unlock(&root->fs_info->trans_mutex);
C
Chris Mason 已提交
1055 1056 1057
	ret = btrfs_write_and_wait_transaction(trans, root);
	BUG_ON(ret);
	write_ctree_super(trans, root);
1058

1059 1060 1061 1062 1063 1064
	/*
	 * the super is written, we can safely allow the tree-loggers
	 * to go about their business
	 */
	mutex_unlock(&root->fs_info->tree_log_mutex);

1065 1066 1067
	btrfs_finish_extent_commit(trans, root, pinned_copy);
	kfree(pinned_copy);

Z
Zheng Yan 已提交
1068 1069 1070
	btrfs_drop_dead_reloc_roots(root);
	mutex_unlock(&root->fs_info->tree_reloc_mutex);

1071 1072 1073
	/* do the directory inserts of any pending snapshot creations */
	finish_pending_snapshots(trans, root->fs_info);

Z
Zheng Yan 已提交
1074 1075
	mutex_lock(&root->fs_info->trans_mutex);

C
Chris Mason 已提交
1076
	cur_trans->commit_done = 1;
1077
	root->fs_info->last_trans_committed = cur_trans->transid;
C
Chris Mason 已提交
1078
	wake_up(&cur_trans->commit_wait);
1079

C
Chris Mason 已提交
1080
	put_transaction(cur_trans);
C
Chris Mason 已提交
1081
	put_transaction(cur_trans);
1082

1083
	list_splice_init(&dirty_fs_roots, &root->fs_info->dead_roots);
1084 1085
	if (root->fs_info->closing)
		list_splice_init(&root->fs_info->dead_roots, &dirty_fs_roots);
1086

C
Chris Mason 已提交
1087
	mutex_unlock(&root->fs_info->trans_mutex);
1088

C
Chris Mason 已提交
1089
	kmem_cache_free(btrfs_trans_handle_cachep, trans);
C
Chris Mason 已提交
1090

1091 1092 1093
	if (root->fs_info->closing) {
		drop_dirty_roots(root->fs_info->tree_root, &dirty_fs_roots);
	}
C
Chris Mason 已提交
1094 1095 1096
	return ret;
}

C
Chris Mason 已提交
1097 1098 1099
/*
 * interface function to delete all the snapshots we have scheduled for deletion
 */
1100 1101 1102 1103
int btrfs_clean_old_snapshots(struct btrfs_root *root)
{
	struct list_head dirty_roots;
	INIT_LIST_HEAD(&dirty_roots);
1104
again:
1105 1106 1107 1108 1109 1110
	mutex_lock(&root->fs_info->trans_mutex);
	list_splice_init(&root->fs_info->dead_roots, &dirty_roots);
	mutex_unlock(&root->fs_info->trans_mutex);

	if (!list_empty(&dirty_roots)) {
		drop_dirty_roots(root, &dirty_roots);
1111
		goto again;
1112 1113 1114
	}
	return 0;
}