qgroup.c 65.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
/*
 * Copyright (C) 2011 STRATO.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include <linux/sched.h>
#include <linux/pagemap.h>
#include <linux/writeback.h>
#include <linux/blkdev.h>
#include <linux/rbtree.h>
#include <linux/slab.h>
#include <linux/workqueue.h>
26
#include <linux/btrfs.h>
27 28 29 30 31 32 33

#include "ctree.h"
#include "transaction.h"
#include "disk-io.h"
#include "locking.h"
#include "ulist.h"
#include "backref.h"
J
Jan Schmidt 已提交
34
#include "extent_io.h"
J
Josef Bacik 已提交
35
#include "qgroup.h"
36

37

38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
/* TODO XXX FIXME
 *  - subvol delete -> delete when ref goes to 0? delete limits also?
 *  - reorganize keys
 *  - compressed
 *  - sync
 *  - copy also limits on subvol creation
 *  - limit
 *  - caches fuer ulists
 *  - performance benchmarks
 *  - check all ioctl parameters
 */

/*
 * one struct for each qgroup, organized in fs_info->qgroup_tree.
 */
struct btrfs_qgroup {
	u64 qgroupid;

	/*
	 * state
	 */
	u64 rfer;	/* referenced */
	u64 rfer_cmpr;	/* referenced compressed */
	u64 excl;	/* exclusive */
	u64 excl_cmpr;	/* exclusive compressed */

	/*
	 * limits
	 */
	u64 lim_flags;	/* which limits are set */
	u64 max_rfer;
	u64 max_excl;
	u64 rsv_rfer;
	u64 rsv_excl;

	/*
	 * reservation tracking
	 */
	u64 reserved;

	/*
	 * lists
	 */
	struct list_head groups;  /* groups this group is member of */
	struct list_head members; /* groups that are members of this group */
	struct list_head dirty;   /* dirty groups */
	struct rb_node node;	  /* tree of qgroups */

	/*
	 * temp variables for accounting operations
88
	 * Refer to qgroup_shared_accouting() for details.
89
	 */
J
Josef Bacik 已提交
90 91
	u64 old_refcnt;
	u64 new_refcnt;
92 93
};

94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
static void btrfs_qgroup_update_old_refcnt(struct btrfs_qgroup *qg, u64 seq,
					   int mod)
{
	if (qg->old_refcnt < seq)
		qg->old_refcnt = seq;
	qg->old_refcnt += mod;
}

static void btrfs_qgroup_update_new_refcnt(struct btrfs_qgroup *qg, u64 seq,
					   int mod)
{
	if (qg->new_refcnt < seq)
		qg->new_refcnt = seq;
	qg->new_refcnt += mod;
}

static inline u64 btrfs_qgroup_get_old_refcnt(struct btrfs_qgroup *qg, u64 seq)
{
	if (qg->old_refcnt < seq)
		return 0;
	return qg->old_refcnt - seq;
}

static inline u64 btrfs_qgroup_get_new_refcnt(struct btrfs_qgroup *qg, u64 seq)
{
	if (qg->new_refcnt < seq)
		return 0;
	return qg->new_refcnt - seq;
}

124 125 126 127 128 129 130 131 132 133
/*
 * glue structure to represent the relations between qgroups.
 */
struct btrfs_qgroup_list {
	struct list_head next_group;
	struct list_head next_member;
	struct btrfs_qgroup *group;
	struct btrfs_qgroup *member;
};

J
Josef Bacik 已提交
134 135 136
#define ptr_to_u64(x) ((u64)(uintptr_t)x)
#define u64_to_ptr(x) ((struct btrfs_qgroup *)(uintptr_t)x)

137 138 139 140
static int
qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
		   int init_flags);
static void qgroup_rescan_zero_tracking(struct btrfs_fs_info *fs_info);
J
Jan Schmidt 已提交
141

142
/* must be called with qgroup_ioctl_lock held */
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
static struct btrfs_qgroup *find_qgroup_rb(struct btrfs_fs_info *fs_info,
					   u64 qgroupid)
{
	struct rb_node *n = fs_info->qgroup_tree.rb_node;
	struct btrfs_qgroup *qgroup;

	while (n) {
		qgroup = rb_entry(n, struct btrfs_qgroup, node);
		if (qgroup->qgroupid < qgroupid)
			n = n->rb_left;
		else if (qgroup->qgroupid > qgroupid)
			n = n->rb_right;
		else
			return qgroup;
	}
	return NULL;
}

/* must be called with qgroup_lock held */
static struct btrfs_qgroup *add_qgroup_rb(struct btrfs_fs_info *fs_info,
					  u64 qgroupid)
{
	struct rb_node **p = &fs_info->qgroup_tree.rb_node;
	struct rb_node *parent = NULL;
	struct btrfs_qgroup *qgroup;

	while (*p) {
		parent = *p;
		qgroup = rb_entry(parent, struct btrfs_qgroup, node);

		if (qgroup->qgroupid < qgroupid)
			p = &(*p)->rb_left;
		else if (qgroup->qgroupid > qgroupid)
			p = &(*p)->rb_right;
		else
			return qgroup;
	}

	qgroup = kzalloc(sizeof(*qgroup), GFP_ATOMIC);
	if (!qgroup)
		return ERR_PTR(-ENOMEM);

	qgroup->qgroupid = qgroupid;
	INIT_LIST_HEAD(&qgroup->groups);
	INIT_LIST_HEAD(&qgroup->members);
	INIT_LIST_HEAD(&qgroup->dirty);

	rb_link_node(&qgroup->node, parent, p);
	rb_insert_color(&qgroup->node, &fs_info->qgroup_tree);

	return qgroup;
}

196
static void __del_qgroup_rb(struct btrfs_qgroup *qgroup)
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216
{
	struct btrfs_qgroup_list *list;

	list_del(&qgroup->dirty);
	while (!list_empty(&qgroup->groups)) {
		list = list_first_entry(&qgroup->groups,
					struct btrfs_qgroup_list, next_group);
		list_del(&list->next_group);
		list_del(&list->next_member);
		kfree(list);
	}

	while (!list_empty(&qgroup->members)) {
		list = list_first_entry(&qgroup->members,
					struct btrfs_qgroup_list, next_member);
		list_del(&list->next_group);
		list_del(&list->next_member);
		kfree(list);
	}
	kfree(qgroup);
217
}
218

219 220 221 222 223 224 225 226 227 228
/* must be called with qgroup_lock held */
static int del_qgroup_rb(struct btrfs_fs_info *fs_info, u64 qgroupid)
{
	struct btrfs_qgroup *qgroup = find_qgroup_rb(fs_info, qgroupid);

	if (!qgroup)
		return -ENOENT;

	rb_erase(&qgroup->node, &fs_info->qgroup_tree);
	__del_qgroup_rb(qgroup);
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280
	return 0;
}

/* must be called with qgroup_lock held */
static int add_relation_rb(struct btrfs_fs_info *fs_info,
			   u64 memberid, u64 parentid)
{
	struct btrfs_qgroup *member;
	struct btrfs_qgroup *parent;
	struct btrfs_qgroup_list *list;

	member = find_qgroup_rb(fs_info, memberid);
	parent = find_qgroup_rb(fs_info, parentid);
	if (!member || !parent)
		return -ENOENT;

	list = kzalloc(sizeof(*list), GFP_ATOMIC);
	if (!list)
		return -ENOMEM;

	list->group = parent;
	list->member = member;
	list_add_tail(&list->next_group, &member->groups);
	list_add_tail(&list->next_member, &parent->members);

	return 0;
}

/* must be called with qgroup_lock held */
static int del_relation_rb(struct btrfs_fs_info *fs_info,
			   u64 memberid, u64 parentid)
{
	struct btrfs_qgroup *member;
	struct btrfs_qgroup *parent;
	struct btrfs_qgroup_list *list;

	member = find_qgroup_rb(fs_info, memberid);
	parent = find_qgroup_rb(fs_info, parentid);
	if (!member || !parent)
		return -ENOENT;

	list_for_each_entry(list, &member->groups, next_group) {
		if (list->group == parent) {
			list_del(&list->next_group);
			list_del(&list->next_member);
			kfree(list);
			return 0;
		}
	}
	return -ENOENT;
}

281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
int btrfs_verify_qgroup_counts(struct btrfs_fs_info *fs_info, u64 qgroupid,
			       u64 rfer, u64 excl)
{
	struct btrfs_qgroup *qgroup;

	qgroup = find_qgroup_rb(fs_info, qgroupid);
	if (!qgroup)
		return -EINVAL;
	if (qgroup->rfer != rfer || qgroup->excl != excl)
		return -EINVAL;
	return 0;
}
#endif

296 297 298 299 300 301 302 303 304 305 306 307 308 309
/*
 * The full config is read in one go, only called from open_ctree()
 * It doesn't use any locking, as at this point we're still single-threaded
 */
int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info)
{
	struct btrfs_key key;
	struct btrfs_key found_key;
	struct btrfs_root *quota_root = fs_info->quota_root;
	struct btrfs_path *path = NULL;
	struct extent_buffer *l;
	int slot;
	int ret = 0;
	u64 flags = 0;
310
	u64 rescan_progress = 0;
311 312 313 314

	if (!fs_info->quota_enabled)
		return 0;

315 316 317 318 319 320
	fs_info->qgroup_ulist = ulist_alloc(GFP_NOFS);
	if (!fs_info->qgroup_ulist) {
		ret = -ENOMEM;
		goto out;
	}

321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	/* default this to quota off, in case no status key is found */
	fs_info->qgroup_flags = 0;

	/*
	 * pass 1: read status, all qgroup infos and limits
	 */
	key.objectid = 0;
	key.type = 0;
	key.offset = 0;
	ret = btrfs_search_slot_for_read(quota_root, &key, path, 1, 1);
	if (ret)
		goto out;

	while (1) {
		struct btrfs_qgroup *qgroup;

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

		if (found_key.type == BTRFS_QGROUP_STATUS_KEY) {
			struct btrfs_qgroup_status_item *ptr;

			ptr = btrfs_item_ptr(l, slot,
					     struct btrfs_qgroup_status_item);

			if (btrfs_qgroup_status_version(l, ptr) !=
			    BTRFS_QGROUP_STATUS_VERSION) {
355 356
				btrfs_err(fs_info,
				 "old qgroup version, quota disabled");
357 358 359 360 361
				goto out;
			}
			if (btrfs_qgroup_status_generation(l, ptr) !=
			    fs_info->generation) {
				flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
362 363 364
				btrfs_err(fs_info,
					"qgroup generation mismatch, "
					"marked as inconsistent");
365 366 367
			}
			fs_info->qgroup_flags = btrfs_qgroup_status_flags(l,
									  ptr);
368
			rescan_progress = btrfs_qgroup_status_rescan(l, ptr);
369 370 371 372 373 374 375 376 377 378
			goto next1;
		}

		if (found_key.type != BTRFS_QGROUP_INFO_KEY &&
		    found_key.type != BTRFS_QGROUP_LIMIT_KEY)
			goto next1;

		qgroup = find_qgroup_rb(fs_info, found_key.offset);
		if ((qgroup && found_key.type == BTRFS_QGROUP_INFO_KEY) ||
		    (!qgroup && found_key.type == BTRFS_QGROUP_LIMIT_KEY)) {
379
			btrfs_err(fs_info, "inconsistent qgroup config");
380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
			flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
		}
		if (!qgroup) {
			qgroup = add_qgroup_rb(fs_info, found_key.offset);
			if (IS_ERR(qgroup)) {
				ret = PTR_ERR(qgroup);
				goto out;
			}
		}
		switch (found_key.type) {
		case BTRFS_QGROUP_INFO_KEY: {
			struct btrfs_qgroup_info_item *ptr;

			ptr = btrfs_item_ptr(l, slot,
					     struct btrfs_qgroup_info_item);
			qgroup->rfer = btrfs_qgroup_info_rfer(l, ptr);
			qgroup->rfer_cmpr = btrfs_qgroup_info_rfer_cmpr(l, ptr);
			qgroup->excl = btrfs_qgroup_info_excl(l, ptr);
			qgroup->excl_cmpr = btrfs_qgroup_info_excl_cmpr(l, ptr);
			/* generation currently unused */
			break;
		}
		case BTRFS_QGROUP_LIMIT_KEY: {
			struct btrfs_qgroup_limit_item *ptr;

			ptr = btrfs_item_ptr(l, slot,
					     struct btrfs_qgroup_limit_item);
			qgroup->lim_flags = btrfs_qgroup_limit_flags(l, ptr);
			qgroup->max_rfer = btrfs_qgroup_limit_max_rfer(l, ptr);
			qgroup->max_excl = btrfs_qgroup_limit_max_excl(l, ptr);
			qgroup->rsv_rfer = btrfs_qgroup_limit_rsv_rfer(l, ptr);
			qgroup->rsv_excl = btrfs_qgroup_limit_rsv_excl(l, ptr);
			break;
		}
		}
next1:
		ret = btrfs_next_item(quota_root, path);
		if (ret < 0)
			goto out;
		if (ret)
			break;
	}
	btrfs_release_path(path);

	/*
	 * pass 2: read all qgroup relations
	 */
	key.objectid = 0;
	key.type = BTRFS_QGROUP_RELATION_KEY;
	key.offset = 0;
	ret = btrfs_search_slot_for_read(quota_root, &key, path, 1, 0);
	if (ret)
		goto out;
	while (1) {
		slot = path->slots[0];
		l = path->nodes[0];
		btrfs_item_key_to_cpu(l, &found_key, slot);

		if (found_key.type != BTRFS_QGROUP_RELATION_KEY)
			goto next2;

		if (found_key.objectid > found_key.offset) {
			/* parent <- member, not needed to build config */
			/* FIXME should we omit the key completely? */
			goto next2;
		}

		ret = add_relation_rb(fs_info, found_key.objectid,
				      found_key.offset);
449
		if (ret == -ENOENT) {
450 451
			btrfs_warn(fs_info,
				"orphan qgroup relation 0x%llx->0x%llx",
452
				found_key.objectid, found_key.offset);
453 454
			ret = 0;	/* ignore the error */
		}
455 456 457 458 459 460 461 462 463 464 465 466 467 468
		if (ret)
			goto out;
next2:
		ret = btrfs_next_item(quota_root, path);
		if (ret < 0)
			goto out;
		if (ret)
			break;
	}
out:
	fs_info->qgroup_flags |= flags;
	if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON)) {
		fs_info->quota_enabled = 0;
		fs_info->pending_quota_state = 0;
469 470 471
	} else if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN &&
		   ret >= 0) {
		ret = qgroup_rescan_init(fs_info, rescan_progress, 0);
472 473 474
	}
	btrfs_free_path(path);

475
	if (ret < 0) {
476
		ulist_free(fs_info->qgroup_ulist);
477
		fs_info->qgroup_ulist = NULL;
478
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
479
	}
480

481 482 483 484
	return ret < 0 ? ret : 0;
}

/*
485 486 487 488
 * This is called from close_ctree() or open_ctree() or btrfs_quota_disable(),
 * first two are in single-threaded paths.And for the third one, we have set
 * quota_root to be null with qgroup_lock held before, so it is safe to clean
 * up the in-memory structures without qgroup_lock held.
489 490 491 492 493 494 495 496 497
 */
void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info)
{
	struct rb_node *n;
	struct btrfs_qgroup *qgroup;

	while ((n = rb_first(&fs_info->qgroup_tree))) {
		qgroup = rb_entry(n, struct btrfs_qgroup, node);
		rb_erase(n, &fs_info->qgroup_tree);
498
		__del_qgroup_rb(qgroup);
499
	}
500 501 502 503 504
	/*
	 * we call btrfs_free_qgroup_config() when umounting
	 * filesystem and disabling quota, so we set qgroup_ulit
	 * to be null here to avoid double free.
	 */
505
	ulist_free(fs_info->qgroup_ulist);
506
	fs_info->qgroup_ulist = NULL;
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
}

static int add_qgroup_relation_item(struct btrfs_trans_handle *trans,
				    struct btrfs_root *quota_root,
				    u64 src, u64 dst)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;

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

	key.objectid = src;
	key.type = BTRFS_QGROUP_RELATION_KEY;
	key.offset = dst;

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

	btrfs_mark_buffer_dirty(path->nodes[0]);

	btrfs_free_path(path);
	return ret;
}

static int del_qgroup_relation_item(struct btrfs_trans_handle *trans,
				    struct btrfs_root *quota_root,
				    u64 src, u64 dst)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;

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

	key.objectid = src;
	key.type = BTRFS_QGROUP_RELATION_KEY;
	key.offset = dst;

	ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	ret = btrfs_del_item(trans, quota_root, path);
out:
	btrfs_free_path(path);
	return ret;
}

static int add_qgroup_item(struct btrfs_trans_handle *trans,
			   struct btrfs_root *quota_root, u64 qgroupid)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_qgroup_info_item *qgroup_info;
	struct btrfs_qgroup_limit_item *qgroup_limit;
	struct extent_buffer *leaf;
	struct btrfs_key key;

574
	if (btrfs_test_is_dummy_root(quota_root))
575
		return 0;
576

577 578 579 580 581 582 583 584
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = 0;
	key.type = BTRFS_QGROUP_INFO_KEY;
	key.offset = qgroupid;

585 586 587 588 589 590
	/*
	 * Avoid a transaction abort by catching -EEXIST here. In that
	 * case, we proceed by re-initializing the existing structure
	 * on disk.
	 */

591 592
	ret = btrfs_insert_empty_item(trans, quota_root, path, &key,
				      sizeof(*qgroup_info));
593
	if (ret && ret != -EEXIST)
594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
		goto out;

	leaf = path->nodes[0];
	qgroup_info = btrfs_item_ptr(leaf, path->slots[0],
				 struct btrfs_qgroup_info_item);
	btrfs_set_qgroup_info_generation(leaf, qgroup_info, trans->transid);
	btrfs_set_qgroup_info_rfer(leaf, qgroup_info, 0);
	btrfs_set_qgroup_info_rfer_cmpr(leaf, qgroup_info, 0);
	btrfs_set_qgroup_info_excl(leaf, qgroup_info, 0);
	btrfs_set_qgroup_info_excl_cmpr(leaf, qgroup_info, 0);

	btrfs_mark_buffer_dirty(leaf);

	btrfs_release_path(path);

	key.type = BTRFS_QGROUP_LIMIT_KEY;
	ret = btrfs_insert_empty_item(trans, quota_root, path, &key,
				      sizeof(*qgroup_limit));
612
	if (ret && ret != -EEXIST)
613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678
		goto out;

	leaf = path->nodes[0];
	qgroup_limit = btrfs_item_ptr(leaf, path->slots[0],
				  struct btrfs_qgroup_limit_item);
	btrfs_set_qgroup_limit_flags(leaf, qgroup_limit, 0);
	btrfs_set_qgroup_limit_max_rfer(leaf, qgroup_limit, 0);
	btrfs_set_qgroup_limit_max_excl(leaf, qgroup_limit, 0);
	btrfs_set_qgroup_limit_rsv_rfer(leaf, qgroup_limit, 0);
	btrfs_set_qgroup_limit_rsv_excl(leaf, qgroup_limit, 0);

	btrfs_mark_buffer_dirty(leaf);

	ret = 0;
out:
	btrfs_free_path(path);
	return ret;
}

static int del_qgroup_item(struct btrfs_trans_handle *trans,
			   struct btrfs_root *quota_root, u64 qgroupid)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;

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

	key.objectid = 0;
	key.type = BTRFS_QGROUP_INFO_KEY;
	key.offset = qgroupid;
	ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	ret = btrfs_del_item(trans, quota_root, path);
	if (ret)
		goto out;

	btrfs_release_path(path);

	key.type = BTRFS_QGROUP_LIMIT_KEY;
	ret = btrfs_search_slot(trans, quota_root, &key, path, -1, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	ret = btrfs_del_item(trans, quota_root, path);

out:
	btrfs_free_path(path);
	return ret;
}

static int update_qgroup_limit_item(struct btrfs_trans_handle *trans,
679 680
				    struct btrfs_root *root,
				    struct btrfs_qgroup *qgroup)
681 682 683 684 685 686 687 688 689 690
{
	struct btrfs_path *path;
	struct btrfs_key key;
	struct extent_buffer *l;
	struct btrfs_qgroup_limit_item *qgroup_limit;
	int ret;
	int slot;

	key.objectid = 0;
	key.type = BTRFS_QGROUP_LIMIT_KEY;
691
	key.offset = qgroup->qgroupid;
692 693

	path = btrfs_alloc_path();
694 695 696
	if (!path)
		return -ENOMEM;

697 698 699 700 701 702 703 704 705
	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
	if (ret > 0)
		ret = -ENOENT;

	if (ret)
		goto out;

	l = path->nodes[0];
	slot = path->slots[0];
706
	qgroup_limit = btrfs_item_ptr(l, slot, struct btrfs_qgroup_limit_item);
707 708 709 710 711
	btrfs_set_qgroup_limit_flags(l, qgroup_limit, qgroup->lim_flags);
	btrfs_set_qgroup_limit_max_rfer(l, qgroup_limit, qgroup->max_rfer);
	btrfs_set_qgroup_limit_max_excl(l, qgroup_limit, qgroup->max_excl);
	btrfs_set_qgroup_limit_rsv_rfer(l, qgroup_limit, qgroup->rsv_rfer);
	btrfs_set_qgroup_limit_rsv_excl(l, qgroup_limit, qgroup->rsv_excl);
712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730

	btrfs_mark_buffer_dirty(l);

out:
	btrfs_free_path(path);
	return ret;
}

static int update_qgroup_info_item(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct btrfs_qgroup *qgroup)
{
	struct btrfs_path *path;
	struct btrfs_key key;
	struct extent_buffer *l;
	struct btrfs_qgroup_info_item *qgroup_info;
	int ret;
	int slot;

731
	if (btrfs_test_is_dummy_root(root))
732
		return 0;
733

734 735 736 737 738
	key.objectid = 0;
	key.type = BTRFS_QGROUP_INFO_KEY;
	key.offset = qgroup->qgroupid;

	path = btrfs_alloc_path();
739 740 741
	if (!path)
		return -ENOMEM;

742 743 744 745 746 747 748 749 750
	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
	if (ret > 0)
		ret = -ENOENT;

	if (ret)
		goto out;

	l = path->nodes[0];
	slot = path->slots[0];
751
	qgroup_info = btrfs_item_ptr(l, slot, struct btrfs_qgroup_info_item);
752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
	btrfs_set_qgroup_info_generation(l, qgroup_info, trans->transid);
	btrfs_set_qgroup_info_rfer(l, qgroup_info, qgroup->rfer);
	btrfs_set_qgroup_info_rfer_cmpr(l, qgroup_info, qgroup->rfer_cmpr);
	btrfs_set_qgroup_info_excl(l, qgroup_info, qgroup->excl);
	btrfs_set_qgroup_info_excl_cmpr(l, qgroup_info, qgroup->excl_cmpr);

	btrfs_mark_buffer_dirty(l);

out:
	btrfs_free_path(path);
	return ret;
}

static int update_qgroup_status_item(struct btrfs_trans_handle *trans,
				     struct btrfs_fs_info *fs_info,
				    struct btrfs_root *root)
{
	struct btrfs_path *path;
	struct btrfs_key key;
	struct extent_buffer *l;
	struct btrfs_qgroup_status_item *ptr;
	int ret;
	int slot;

	key.objectid = 0;
	key.type = BTRFS_QGROUP_STATUS_KEY;
	key.offset = 0;

	path = btrfs_alloc_path();
781 782 783
	if (!path)
		return -ENOMEM;

784 785 786 787 788 789 790 791 792 793 794 795
	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
	if (ret > 0)
		ret = -ENOENT;

	if (ret)
		goto out;

	l = path->nodes[0];
	slot = path->slots[0];
	ptr = btrfs_item_ptr(l, slot, struct btrfs_qgroup_status_item);
	btrfs_set_qgroup_status_flags(l, ptr, fs_info->qgroup_flags);
	btrfs_set_qgroup_status_generation(l, ptr, trans->transid);
J
Jan Schmidt 已提交
796 797
	btrfs_set_qgroup_status_rescan(l, ptr,
				fs_info->qgroup_rescan_progress.objectid);
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813

	btrfs_mark_buffer_dirty(l);

out:
	btrfs_free_path(path);
	return ret;
}

/*
 * called with qgroup_lock held
 */
static int btrfs_clean_quota_tree(struct btrfs_trans_handle *trans,
				  struct btrfs_root *root)
{
	struct btrfs_path *path;
	struct btrfs_key key;
814
	struct extent_buffer *leaf = NULL;
815
	int ret;
816
	int nr = 0;
817 818 819 820 821

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

822 823 824 825 826
	path->leave_spinning = 1;

	key.objectid = 0;
	key.offset = 0;
	key.type = 0;
827

828
	while (1) {
829
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
830 831 832 833 834
		if (ret < 0)
			goto out;
		leaf = path->nodes[0];
		nr = btrfs_header_nritems(leaf);
		if (!nr)
835
			break;
836 837 838 839 840 841 842
		/*
		 * delete the leaf one by one
		 * since the whole tree is going
		 * to be deleted.
		 */
		path->slots[0] = 0;
		ret = btrfs_del_items(trans, root, path, 0, nr);
843 844
		if (ret)
			goto out;
845

846 847 848 849 850 851 852 853 854 855 856 857 858
		btrfs_release_path(path);
	}
	ret = 0;
out:
	root->fs_info->pending_quota_state = 0;
	btrfs_free_path(path);
	return ret;
}

int btrfs_quota_enable(struct btrfs_trans_handle *trans,
		       struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *quota_root;
859
	struct btrfs_root *tree_root = fs_info->tree_root;
860 861 862 863
	struct btrfs_path *path = NULL;
	struct btrfs_qgroup_status_item *ptr;
	struct extent_buffer *leaf;
	struct btrfs_key key;
864 865
	struct btrfs_key found_key;
	struct btrfs_qgroup *qgroup = NULL;
866
	int ret = 0;
867
	int slot;
868

869
	mutex_lock(&fs_info->qgroup_ioctl_lock);
870 871 872 873 874
	if (fs_info->quota_root) {
		fs_info->pending_quota_state = 1;
		goto out;
	}

875 876 877 878 879 880
	fs_info->qgroup_ulist = ulist_alloc(GFP_NOFS);
	if (!fs_info->qgroup_ulist) {
		ret = -ENOMEM;
		goto out;
	}

881 882 883 884 885 886 887 888 889 890 891
	/*
	 * initially create the quota tree
	 */
	quota_root = btrfs_create_tree(trans, fs_info,
				       BTRFS_QUOTA_TREE_OBJECTID);
	if (IS_ERR(quota_root)) {
		ret =  PTR_ERR(quota_root);
		goto out;
	}

	path = btrfs_alloc_path();
892 893 894 895
	if (!path) {
		ret = -ENOMEM;
		goto out_free_root;
	}
896 897 898 899 900 901 902 903

	key.objectid = 0;
	key.type = BTRFS_QGROUP_STATUS_KEY;
	key.offset = 0;

	ret = btrfs_insert_empty_item(trans, quota_root, path, &key,
				      sizeof(*ptr));
	if (ret)
904
		goto out_free_path;
905 906 907 908 909 910 911 912 913

	leaf = path->nodes[0];
	ptr = btrfs_item_ptr(leaf, path->slots[0],
				 struct btrfs_qgroup_status_item);
	btrfs_set_qgroup_status_generation(leaf, ptr, trans->transid);
	btrfs_set_qgroup_status_version(leaf, ptr, BTRFS_QGROUP_STATUS_VERSION);
	fs_info->qgroup_flags = BTRFS_QGROUP_STATUS_FLAG_ON |
				BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
	btrfs_set_qgroup_status_flags(leaf, ptr, fs_info->qgroup_flags);
J
Jan Schmidt 已提交
914
	btrfs_set_qgroup_status_rescan(leaf, ptr, 0);
915 916 917

	btrfs_mark_buffer_dirty(leaf);

918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964
	key.objectid = 0;
	key.type = BTRFS_ROOT_REF_KEY;
	key.offset = 0;

	btrfs_release_path(path);
	ret = btrfs_search_slot_for_read(tree_root, &key, path, 1, 0);
	if (ret > 0)
		goto out_add_root;
	if (ret < 0)
		goto out_free_path;


	while (1) {
		slot = path->slots[0];
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, slot);

		if (found_key.type == BTRFS_ROOT_REF_KEY) {
			ret = add_qgroup_item(trans, quota_root,
					      found_key.offset);
			if (ret)
				goto out_free_path;

			qgroup = add_qgroup_rb(fs_info, found_key.offset);
			if (IS_ERR(qgroup)) {
				ret = PTR_ERR(qgroup);
				goto out_free_path;
			}
		}
		ret = btrfs_next_item(tree_root, path);
		if (ret < 0)
			goto out_free_path;
		if (ret)
			break;
	}

out_add_root:
	btrfs_release_path(path);
	ret = add_qgroup_item(trans, quota_root, BTRFS_FS_TREE_OBJECTID);
	if (ret)
		goto out_free_path;

	qgroup = add_qgroup_rb(fs_info, BTRFS_FS_TREE_OBJECTID);
	if (IS_ERR(qgroup)) {
		ret = PTR_ERR(qgroup);
		goto out_free_path;
	}
965
	spin_lock(&fs_info->qgroup_lock);
966 967 968
	fs_info->quota_root = quota_root;
	fs_info->pending_quota_state = 1;
	spin_unlock(&fs_info->qgroup_lock);
969
out_free_path:
970
	btrfs_free_path(path);
971 972 973 974 975 976 977
out_free_root:
	if (ret) {
		free_extent_buffer(quota_root->node);
		free_extent_buffer(quota_root->commit_root);
		kfree(quota_root);
	}
out:
978
	if (ret) {
979
		ulist_free(fs_info->qgroup_ulist);
980 981
		fs_info->qgroup_ulist = NULL;
	}
982
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
983 984 985 986 987 988 989 990 991 992
	return ret;
}

int btrfs_quota_disable(struct btrfs_trans_handle *trans,
			struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *tree_root = fs_info->tree_root;
	struct btrfs_root *quota_root;
	int ret = 0;

993
	mutex_lock(&fs_info->qgroup_ioctl_lock);
994
	if (!fs_info->quota_root)
995
		goto out;
996
	spin_lock(&fs_info->qgroup_lock);
997 998 999 1000
	fs_info->quota_enabled = 0;
	fs_info->pending_quota_state = 0;
	quota_root = fs_info->quota_root;
	fs_info->quota_root = NULL;
1001
	fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_ON;
1002 1003
	spin_unlock(&fs_info->qgroup_lock);

1004 1005
	btrfs_free_qgroup_config(fs_info);

1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016
	ret = btrfs_clean_quota_tree(trans, quota_root);
	if (ret)
		goto out;

	ret = btrfs_del_root(trans, tree_root, &quota_root->root_key);
	if (ret)
		goto out;

	list_del(&quota_root->dirty_list);

	btrfs_tree_lock(quota_root->node);
1017
	clean_tree_block(trans, tree_root->fs_info, quota_root->node);
1018 1019 1020 1021 1022 1023 1024
	btrfs_tree_unlock(quota_root->node);
	btrfs_free_tree_block(trans, quota_root, quota_root->node, 0, 1);

	free_extent_buffer(quota_root->node);
	free_extent_buffer(quota_root->commit_root);
	kfree(quota_root);
out:
1025
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1026 1027 1028
	return ret;
}

J
Jan Schmidt 已提交
1029 1030
static void qgroup_dirty(struct btrfs_fs_info *fs_info,
			 struct btrfs_qgroup *qgroup)
1031
{
J
Jan Schmidt 已提交
1032 1033
	if (list_empty(&qgroup->dirty))
		list_add(&qgroup->dirty, &fs_info->dirty_qgroups);
1034 1035
}

1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
/*
 * The easy accounting, if we are adding/removing the only ref for an extent
 * then this qgroup and all of the parent qgroups get their refrence and
 * exclusive counts adjusted.
 *
 * Caller should hold fs_info->qgroup_lock.
 */
static int __qgroup_excl_accounting(struct btrfs_fs_info *fs_info,
				    struct ulist *tmp, u64 ref_root,
				    u64 num_bytes, int sign)
{
	struct btrfs_qgroup *qgroup;
	struct btrfs_qgroup_list *glist;
	struct ulist_node *unode;
	struct ulist_iterator uiter;
	int ret = 0;

	qgroup = find_qgroup_rb(fs_info, ref_root);
	if (!qgroup)
		goto out;

	qgroup->rfer += sign * num_bytes;
	qgroup->rfer_cmpr += sign * num_bytes;

	WARN_ON(sign < 0 && qgroup->excl < num_bytes);
	qgroup->excl += sign * num_bytes;
	qgroup->excl_cmpr += sign * num_bytes;
	if (sign > 0)
		qgroup->reserved -= num_bytes;

	qgroup_dirty(fs_info, qgroup);

	/* Get all of the parent groups that contain this qgroup */
	list_for_each_entry(glist, &qgroup->groups, next_group) {
		ret = ulist_add(tmp, glist->group->qgroupid,
				ptr_to_u64(glist->group), GFP_ATOMIC);
		if (ret < 0)
			goto out;
	}

	/* Iterate all of the parents and adjust their reference counts */
	ULIST_ITER_INIT(&uiter);
	while ((unode = ulist_next(tmp, &uiter))) {
		qgroup = u64_to_ptr(unode->aux);
		qgroup->rfer += sign * num_bytes;
		qgroup->rfer_cmpr += sign * num_bytes;
		WARN_ON(sign < 0 && qgroup->excl < num_bytes);
		qgroup->excl += sign * num_bytes;
		if (sign > 0)
			qgroup->reserved -= num_bytes;
		qgroup->excl_cmpr += sign * num_bytes;
		qgroup_dirty(fs_info, qgroup);

		/* Add any parents of the parents */
		list_for_each_entry(glist, &qgroup->groups, next_group) {
			ret = ulist_add(tmp, glist->group->qgroupid,
					ptr_to_u64(glist->group), GFP_ATOMIC);
			if (ret < 0)
				goto out;
		}
	}
	ret = 0;
out:
	return ret;
}


/*
 * Quick path for updating qgroup with only excl refs.
 *
 * In that case, just update all parent will be enough.
 * Or we needs to do a full rescan.
 * Caller should also hold fs_info->qgroup_lock.
 *
 * Return 0 for quick update, return >0 for need to full rescan
 * and mark INCONSISTENT flag.
 * Return < 0 for other error.
 */
static int quick_update_accounting(struct btrfs_fs_info *fs_info,
				   struct ulist *tmp, u64 src, u64 dst,
				   int sign)
{
	struct btrfs_qgroup *qgroup;
	int ret = 1;
	int err = 0;

	qgroup = find_qgroup_rb(fs_info, src);
	if (!qgroup)
		goto out;
	if (qgroup->excl == qgroup->rfer) {
		ret = 0;
		err = __qgroup_excl_accounting(fs_info, tmp, dst,
					       qgroup->excl, sign);
		if (err < 0) {
			ret = err;
			goto out;
		}
	}
out:
	if (ret)
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
	return ret;
}

1140 1141 1142 1143
int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans,
			      struct btrfs_fs_info *fs_info, u64 src, u64 dst)
{
	struct btrfs_root *quota_root;
1144 1145
	struct btrfs_qgroup *parent;
	struct btrfs_qgroup *member;
1146
	struct btrfs_qgroup_list *list;
1147
	struct ulist *tmp;
1148 1149
	int ret = 0;

1150 1151 1152 1153
	/* Check the level of src and dst first */
	if (btrfs_qgroup_level(src) >= btrfs_qgroup_level(dst))
		return -EINVAL;

1154 1155 1156 1157
	tmp = ulist_alloc(GFP_NOFS);
	if (!tmp)
		return -ENOMEM;

1158
	mutex_lock(&fs_info->qgroup_ioctl_lock);
1159
	quota_root = fs_info->quota_root;
1160 1161 1162 1163
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1164 1165 1166 1167 1168 1169
	member = find_qgroup_rb(fs_info, src);
	parent = find_qgroup_rb(fs_info, dst);
	if (!member || !parent) {
		ret = -EINVAL;
		goto out;
	}
1170

1171 1172 1173 1174 1175 1176 1177 1178
	/* check if such qgroup relation exist firstly */
	list_for_each_entry(list, &member->groups, next_group) {
		if (list->group == parent) {
			ret = -EEXIST;
			goto out;
		}
	}

1179 1180
	ret = add_qgroup_relation_item(trans, quota_root, src, dst);
	if (ret)
1181
		goto out;
1182 1183 1184 1185

	ret = add_qgroup_relation_item(trans, quota_root, dst, src);
	if (ret) {
		del_qgroup_relation_item(trans, quota_root, src, dst);
1186
		goto out;
1187 1188 1189 1190
	}

	spin_lock(&fs_info->qgroup_lock);
	ret = add_relation_rb(quota_root->fs_info, src, dst);
1191 1192 1193 1194 1195
	if (ret < 0) {
		spin_unlock(&fs_info->qgroup_lock);
		goto out;
	}
	ret = quick_update_accounting(fs_info, tmp, src, dst, 1);
1196
	spin_unlock(&fs_info->qgroup_lock);
1197 1198
out:
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1199
	ulist_free(tmp);
1200 1201 1202
	return ret;
}

1203
int __del_qgroup_relation(struct btrfs_trans_handle *trans,
1204 1205 1206
			      struct btrfs_fs_info *fs_info, u64 src, u64 dst)
{
	struct btrfs_root *quota_root;
1207 1208 1209
	struct btrfs_qgroup *parent;
	struct btrfs_qgroup *member;
	struct btrfs_qgroup_list *list;
1210
	struct ulist *tmp;
1211 1212 1213
	int ret = 0;
	int err;

1214 1215 1216 1217
	tmp = ulist_alloc(GFP_NOFS);
	if (!tmp)
		return -ENOMEM;

1218
	quota_root = fs_info->quota_root;
1219 1220 1221 1222
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1223

1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
	member = find_qgroup_rb(fs_info, src);
	parent = find_qgroup_rb(fs_info, dst);
	if (!member || !parent) {
		ret = -EINVAL;
		goto out;
	}

	/* check if such qgroup relation exist firstly */
	list_for_each_entry(list, &member->groups, next_group) {
		if (list->group == parent)
			goto exist;
	}
	ret = -ENOENT;
	goto out;
exist:
1239 1240 1241 1242 1243 1244 1245
	ret = del_qgroup_relation_item(trans, quota_root, src, dst);
	err = del_qgroup_relation_item(trans, quota_root, dst, src);
	if (err && !ret)
		ret = err;

	spin_lock(&fs_info->qgroup_lock);
	del_relation_rb(fs_info, src, dst);
1246
	ret = quick_update_accounting(fs_info, tmp, src, dst, -1);
1247
	spin_unlock(&fs_info->qgroup_lock);
1248
out:
1249
	ulist_free(tmp);
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
	return ret;
}

int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans,
			      struct btrfs_fs_info *fs_info, u64 src, u64 dst)
{
	int ret = 0;

	mutex_lock(&fs_info->qgroup_ioctl_lock);
	ret = __del_qgroup_relation(trans, fs_info, src, dst);
1260
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1261

1262 1263 1264 1265
	return ret;
}

int btrfs_create_qgroup(struct btrfs_trans_handle *trans,
1266
			struct btrfs_fs_info *fs_info, u64 qgroupid)
1267 1268 1269 1270 1271
{
	struct btrfs_root *quota_root;
	struct btrfs_qgroup *qgroup;
	int ret = 0;

1272
	mutex_lock(&fs_info->qgroup_ioctl_lock);
1273
	quota_root = fs_info->quota_root;
1274 1275 1276 1277
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1278 1279 1280 1281 1282
	qgroup = find_qgroup_rb(fs_info, qgroupid);
	if (qgroup) {
		ret = -EEXIST;
		goto out;
	}
1283 1284

	ret = add_qgroup_item(trans, quota_root, qgroupid);
1285 1286
	if (ret)
		goto out;
1287 1288 1289 1290 1291 1292 1293

	spin_lock(&fs_info->qgroup_lock);
	qgroup = add_qgroup_rb(fs_info, qgroupid);
	spin_unlock(&fs_info->qgroup_lock);

	if (IS_ERR(qgroup))
		ret = PTR_ERR(qgroup);
1294 1295
out:
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1296 1297 1298 1299 1300 1301 1302
	return ret;
}

int btrfs_remove_qgroup(struct btrfs_trans_handle *trans,
			struct btrfs_fs_info *fs_info, u64 qgroupid)
{
	struct btrfs_root *quota_root;
1303
	struct btrfs_qgroup *qgroup;
1304
	struct btrfs_qgroup_list *list;
1305 1306
	int ret = 0;

1307
	mutex_lock(&fs_info->qgroup_ioctl_lock);
1308
	quota_root = fs_info->quota_root;
1309 1310 1311 1312
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1313

1314
	qgroup = find_qgroup_rb(fs_info, qgroupid);
1315 1316 1317 1318
	if (!qgroup) {
		ret = -ENOENT;
		goto out;
	} else {
1319 1320
		/* check if there are no children of this qgroup */
		if (!list_empty(&qgroup->members)) {
1321 1322
			ret = -EBUSY;
			goto out;
1323 1324
		}
	}
1325 1326
	ret = del_qgroup_item(trans, quota_root, qgroupid);

1327 1328 1329 1330 1331 1332 1333 1334 1335 1336
	while (!list_empty(&qgroup->groups)) {
		list = list_first_entry(&qgroup->groups,
					struct btrfs_qgroup_list, next_group);
		ret = __del_qgroup_relation(trans, fs_info,
					   qgroupid,
					   list->group->qgroupid);
		if (ret)
			goto out;
	}

1337 1338 1339
	spin_lock(&fs_info->qgroup_lock);
	del_qgroup_rb(quota_root->fs_info, qgroupid);
	spin_unlock(&fs_info->qgroup_lock);
1340 1341
out:
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1342 1343 1344 1345 1346 1347 1348
	return ret;
}

int btrfs_limit_qgroup(struct btrfs_trans_handle *trans,
		       struct btrfs_fs_info *fs_info, u64 qgroupid,
		       struct btrfs_qgroup_limit *limit)
{
1349
	struct btrfs_root *quota_root;
1350 1351
	struct btrfs_qgroup *qgroup;
	int ret = 0;
1352 1353 1354 1355 1356
	/* Sometimes we would want to clear the limit on this qgroup.
	 * To meet this requirement, we treat the -1 as a special value
	 * which tell kernel to clear the limit on this qgroup.
	 */
	const u64 CLEAR_VALUE = -1;
1357

1358 1359 1360 1361 1362 1363
	mutex_lock(&fs_info->qgroup_ioctl_lock);
	quota_root = fs_info->quota_root;
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1364

1365 1366 1367 1368 1369
	qgroup = find_qgroup_rb(fs_info, qgroupid);
	if (!qgroup) {
		ret = -ENOENT;
		goto out;
	}
1370

1371
	spin_lock(&fs_info->qgroup_lock);
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_RFER) {
		if (limit->max_rfer == CLEAR_VALUE) {
			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_MAX_RFER;
			limit->flags &= ~BTRFS_QGROUP_LIMIT_MAX_RFER;
			qgroup->max_rfer = 0;
		} else {
			qgroup->max_rfer = limit->max_rfer;
		}
	}
	if (limit->flags & BTRFS_QGROUP_LIMIT_MAX_EXCL) {
		if (limit->max_excl == CLEAR_VALUE) {
			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_MAX_EXCL;
			limit->flags &= ~BTRFS_QGROUP_LIMIT_MAX_EXCL;
			qgroup->max_excl = 0;
		} else {
			qgroup->max_excl = limit->max_excl;
		}
	}
	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_RFER) {
		if (limit->rsv_rfer == CLEAR_VALUE) {
			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_RSV_RFER;
			limit->flags &= ~BTRFS_QGROUP_LIMIT_RSV_RFER;
			qgroup->rsv_rfer = 0;
		} else {
			qgroup->rsv_rfer = limit->rsv_rfer;
		}
	}
	if (limit->flags & BTRFS_QGROUP_LIMIT_RSV_EXCL) {
		if (limit->rsv_excl == CLEAR_VALUE) {
			qgroup->lim_flags &= ~BTRFS_QGROUP_LIMIT_RSV_EXCL;
			limit->flags &= ~BTRFS_QGROUP_LIMIT_RSV_EXCL;
			qgroup->rsv_excl = 0;
		} else {
			qgroup->rsv_excl = limit->rsv_excl;
		}
	}
1408 1409
	qgroup->lim_flags |= limit->flags;

1410
	spin_unlock(&fs_info->qgroup_lock);
1411 1412 1413 1414 1415 1416 1417 1418

	ret = update_qgroup_limit_item(trans, quota_root, qgroup);
	if (ret) {
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
		btrfs_info(fs_info, "unable to update quota limit for %llu",
		       qgroupid);
	}

1419 1420
out:
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
1421 1422
	return ret;
}
1423

1424 1425 1426 1427 1428 1429
int btrfs_qgroup_prepare_account_extents(struct btrfs_trans_handle *trans,
					 struct btrfs_fs_info *fs_info)
{
	struct btrfs_qgroup_extent_record *record;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct rb_node *node;
1430
	u64 qgroup_to_skip;
1431 1432 1433
	int ret = 0;

	delayed_refs = &trans->transaction->delayed_refs;
1434
	qgroup_to_skip = delayed_refs->qgroup_to_skip;
1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447

	/*
	 * No need to do lock, since this function will only be called in
	 * btrfs_commmit_transaction().
	 */
	node = rb_first(&delayed_refs->dirty_extent_root);
	while (node) {
		record = rb_entry(node, struct btrfs_qgroup_extent_record,
				  node);
		ret = btrfs_find_all_roots(NULL, fs_info, record->bytenr, 0,
					   &record->old_roots);
		if (ret < 0)
			break;
1448 1449
		if (qgroup_to_skip)
			ulist_del(record->old_roots, qgroup_to_skip, 0);
1450 1451 1452 1453 1454
		node = rb_next(node);
	}
	return ret;
}

1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
struct btrfs_qgroup_extent_record
*btrfs_qgroup_insert_dirty_extent(struct btrfs_delayed_ref_root *delayed_refs,
				  struct btrfs_qgroup_extent_record *record)
{
	struct rb_node **p = &delayed_refs->dirty_extent_root.rb_node;
	struct rb_node *parent_node = NULL;
	struct btrfs_qgroup_extent_record *entry;
	u64 bytenr = record->bytenr;

	while (*p) {
		parent_node = *p;
		entry = rb_entry(parent_node, struct btrfs_qgroup_extent_record,
				 node);
		if (bytenr < entry->bytenr)
			p = &(*p)->rb_left;
		else if (bytenr > entry->bytenr)
			p = &(*p)->rb_right;
		else
			return entry;
	}

	rb_link_node(&record->node, parent_node, p);
	rb_insert_color(&record->node, &delayed_refs->dirty_extent_root);
	return NULL;
}

1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538
#define UPDATE_NEW	0
#define UPDATE_OLD	1
/*
 * Walk all of the roots that points to the bytenr and adjust their refcnts.
 */
static int qgroup_update_refcnt(struct btrfs_fs_info *fs_info,
				struct ulist *roots, struct ulist *tmp,
				struct ulist *qgroups, u64 seq, int update_old)
{
	struct ulist_node *unode;
	struct ulist_iterator uiter;
	struct ulist_node *tmp_unode;
	struct ulist_iterator tmp_uiter;
	struct btrfs_qgroup *qg;
	int ret = 0;

	if (!roots)
		return 0;
	ULIST_ITER_INIT(&uiter);
	while ((unode = ulist_next(roots, &uiter))) {
		qg = find_qgroup_rb(fs_info, unode->val);
		if (!qg)
			continue;

		ulist_reinit(tmp);
		ret = ulist_add(qgroups, qg->qgroupid, ptr_to_u64(qg),
				GFP_ATOMIC);
		if (ret < 0)
			return ret;
		ret = ulist_add(tmp, qg->qgroupid, ptr_to_u64(qg), GFP_ATOMIC);
		if (ret < 0)
			return ret;
		ULIST_ITER_INIT(&tmp_uiter);
		while ((tmp_unode = ulist_next(tmp, &tmp_uiter))) {
			struct btrfs_qgroup_list *glist;

			qg = u64_to_ptr(tmp_unode->aux);
			if (update_old)
				btrfs_qgroup_update_old_refcnt(qg, seq, 1);
			else
				btrfs_qgroup_update_new_refcnt(qg, seq, 1);
			list_for_each_entry(glist, &qg->groups, next_group) {
				ret = ulist_add(qgroups, glist->group->qgroupid,
						ptr_to_u64(glist->group),
						GFP_ATOMIC);
				if (ret < 0)
					return ret;
				ret = ulist_add(tmp, glist->group->qgroupid,
						ptr_to_u64(glist->group),
						GFP_ATOMIC);
				if (ret < 0)
					return ret;
			}
		}
	}
	return 0;
}

1539 1540 1541
/*
 * Update qgroup rfer/excl counters.
 * Rfer update is easy, codes can explain themselves.
1542
 *
1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
 * Excl update is tricky, the update is split into 2 part.
 * Part 1: Possible exclusive <-> sharing detect:
 *	|	A	|	!A	|
 *  -------------------------------------
 *  B	|	*	|	-	|
 *  -------------------------------------
 *  !B	|	+	|	**	|
 *  -------------------------------------
 *
 * Conditions:
 * A:	cur_old_roots < nr_old_roots	(not exclusive before)
 * !A:	cur_old_roots == nr_old_roots	(possible exclusive before)
 * B:	cur_new_roots < nr_new_roots	(not exclusive now)
 * !B:	cur_new_roots == nr_new_roots	(possible exclsuive now)
 *
 * Results:
 * +: Possible sharing -> exclusive	-: Possible exclusive -> sharing
 * *: Definitely not changed.		**: Possible unchanged.
 *
 * For !A and !B condition, the exception is cur_old/new_roots == 0 case.
 *
 * To make the logic clear, we first use condition A and B to split
 * combination into 4 results.
 *
 * Then, for result "+" and "-", check old/new_roots == 0 case, as in them
 * only on variant maybe 0.
 *
 * Lastly, check result **, since there are 2 variants maybe 0, split them
 * again(2x2).
 * But this time we don't need to consider other things, the codes and logic
 * is easy to understand now.
 */
static int qgroup_update_counters(struct btrfs_fs_info *fs_info,
				  struct ulist *qgroups,
				  u64 nr_old_roots,
				  u64 nr_new_roots,
				  u64 num_bytes, u64 seq)
{
	struct ulist_node *unode;
	struct ulist_iterator uiter;
	struct btrfs_qgroup *qg;
	u64 cur_new_count, cur_old_count;

	ULIST_ITER_INIT(&uiter);
	while ((unode = ulist_next(qgroups, &uiter))) {
		bool dirty = false;

		qg = u64_to_ptr(unode->aux);
		cur_old_count = btrfs_qgroup_get_old_refcnt(qg, seq);
		cur_new_count = btrfs_qgroup_get_new_refcnt(qg, seq);

		/* Rfer update part */
		if (cur_old_count == 0 && cur_new_count > 0) {
			qg->rfer += num_bytes;
			qg->rfer_cmpr += num_bytes;
			dirty = true;
		}
		if (cur_old_count > 0 && cur_new_count == 0) {
			qg->rfer -= num_bytes;
			qg->rfer_cmpr -= num_bytes;
			dirty = true;
		}

		/* Excl update part */
		/* Exclusive/none -> shared case */
		if (cur_old_count == nr_old_roots &&
		    cur_new_count < nr_new_roots) {
			/* Exclusive -> shared */
			if (cur_old_count != 0) {
				qg->excl -= num_bytes;
				qg->excl_cmpr -= num_bytes;
				dirty = true;
			}
		}

		/* Shared -> exclusive/none case */
		if (cur_old_count < nr_old_roots &&
		    cur_new_count == nr_new_roots) {
			/* Shared->exclusive */
			if (cur_new_count != 0) {
				qg->excl += num_bytes;
				qg->excl_cmpr += num_bytes;
				dirty = true;
			}
		}

		/* Exclusive/none -> exclusive/none case */
		if (cur_old_count == nr_old_roots &&
		    cur_new_count == nr_new_roots) {
			if (cur_old_count == 0) {
				/* None -> exclusive/none */

				if (cur_new_count != 0) {
					/* None -> exclusive */
					qg->excl += num_bytes;
					qg->excl_cmpr += num_bytes;
					dirty = true;
				}
				/* None -> none, nothing changed */
			} else {
				/* Exclusive -> exclusive/none */

				if (cur_new_count == 0) {
					/* Exclusive -> none */
					qg->excl -= num_bytes;
					qg->excl_cmpr -= num_bytes;
					dirty = true;
				}
				/* Exclusive -> exclusive, nothing changed */
			}
		}
1654

1655 1656 1657 1658 1659 1660
		if (dirty)
			qgroup_dirty(fs_info, qg);
	}
	return 0;
}

1661
int
1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans,
			    struct btrfs_fs_info *fs_info,
			    u64 bytenr, u64 num_bytes,
			    struct ulist *old_roots, struct ulist *new_roots)
{
	struct ulist *qgroups = NULL;
	struct ulist *tmp = NULL;
	u64 seq;
	u64 nr_new_roots = 0;
	u64 nr_old_roots = 0;
	int ret = 0;

	if (new_roots)
		nr_new_roots = new_roots->nnodes;
	if (old_roots)
		nr_old_roots = old_roots->nnodes;

	if (!fs_info->quota_enabled)
		goto out_free;
	BUG_ON(!fs_info->quota_root);

	qgroups = ulist_alloc(GFP_NOFS);
	if (!qgroups) {
		ret = -ENOMEM;
		goto out_free;
	}
	tmp = ulist_alloc(GFP_NOFS);
	if (!tmp) {
		ret = -ENOMEM;
		goto out_free;
	}

	mutex_lock(&fs_info->qgroup_rescan_lock);
	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
		if (fs_info->qgroup_rescan_progress.objectid <= bytenr) {
			mutex_unlock(&fs_info->qgroup_rescan_lock);
			ret = 0;
			goto out_free;
		}
	}
	mutex_unlock(&fs_info->qgroup_rescan_lock);

	spin_lock(&fs_info->qgroup_lock);
	seq = fs_info->qgroup_seq;

	/* Update old refcnts using old_roots */
	ret = qgroup_update_refcnt(fs_info, old_roots, tmp, qgroups, seq,
				   UPDATE_OLD);
	if (ret < 0)
		goto out;

	/* Update new refcnts using new_roots */
	ret = qgroup_update_refcnt(fs_info, new_roots, tmp, qgroups, seq,
				   UPDATE_NEW);
	if (ret < 0)
		goto out;

	qgroup_update_counters(fs_info, qgroups, nr_old_roots, nr_new_roots,
			       num_bytes, seq);

	/*
	 * Bump qgroup_seq to avoid seq overlap
	 */
	fs_info->qgroup_seq += max(nr_old_roots, nr_new_roots) + 1;
out:
	spin_unlock(&fs_info->qgroup_lock);
out_free:
	ulist_free(tmp);
	ulist_free(qgroups);
	ulist_free(old_roots);
	ulist_free(new_roots);
	return ret;
}

int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans,
				 struct btrfs_fs_info *fs_info)
{
	struct btrfs_qgroup_extent_record *record;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct ulist *new_roots = NULL;
	struct rb_node *node;
1743
	u64 qgroup_to_skip;
1744 1745 1746
	int ret = 0;

	delayed_refs = &trans->transaction->delayed_refs;
1747
	qgroup_to_skip = delayed_refs->qgroup_to_skip;
1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
	while ((node = rb_first(&delayed_refs->dirty_extent_root))) {
		record = rb_entry(node, struct btrfs_qgroup_extent_record,
				  node);

		if (!ret) {
			/*
			 * Use (u64)-1 as time_seq to do special search, which
			 * doesn't lock tree or delayed_refs and search current
			 * root. It's safe inside commit_transaction().
			 */
			ret = btrfs_find_all_roots(trans, fs_info,
					record->bytenr, (u64)-1, &new_roots);
			if (ret < 0)
				goto cleanup;
1762 1763
			if (qgroup_to_skip)
				ulist_del(new_roots, qgroup_to_skip, 0);
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780
			ret = btrfs_qgroup_account_extent(trans, fs_info,
					record->bytenr, record->num_bytes,
					record->old_roots, new_roots);
			record->old_roots = NULL;
			new_roots = NULL;
		}
cleanup:
		ulist_free(record->old_roots);
		ulist_free(new_roots);
		new_roots = NULL;
		rb_erase(node, &delayed_refs->dirty_extent_root);
		kfree(record);

	}
	return ret;
}

1781 1782 1783 1784 1785 1786 1787 1788
/*
 * called from commit_transaction. Writes all changed qgroups to disk.
 */
int btrfs_run_qgroups(struct btrfs_trans_handle *trans,
		      struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *quota_root = fs_info->quota_root;
	int ret = 0;
1789
	int start_rescan_worker = 0;
1790 1791 1792 1793

	if (!quota_root)
		goto out;

1794 1795 1796
	if (!fs_info->quota_enabled && fs_info->pending_quota_state)
		start_rescan_worker = 1;

1797 1798 1799 1800 1801 1802 1803 1804 1805 1806
	fs_info->quota_enabled = fs_info->pending_quota_state;

	spin_lock(&fs_info->qgroup_lock);
	while (!list_empty(&fs_info->dirty_qgroups)) {
		struct btrfs_qgroup *qgroup;
		qgroup = list_first_entry(&fs_info->dirty_qgroups,
					  struct btrfs_qgroup, dirty);
		list_del_init(&qgroup->dirty);
		spin_unlock(&fs_info->qgroup_lock);
		ret = update_qgroup_info_item(trans, quota_root, qgroup);
1807 1808 1809 1810
		if (ret)
			fs_info->qgroup_flags |=
					BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
		ret = update_qgroup_limit_item(trans, quota_root, qgroup);
1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
		if (ret)
			fs_info->qgroup_flags |=
					BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
		spin_lock(&fs_info->qgroup_lock);
	}
	if (fs_info->quota_enabled)
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_ON;
	else
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_ON;
	spin_unlock(&fs_info->qgroup_lock);

	ret = update_qgroup_status_item(trans, fs_info, quota_root);
	if (ret)
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;

1826
	if (!ret && start_rescan_worker) {
1827 1828 1829
		ret = qgroup_rescan_init(fs_info, 0, 1);
		if (!ret) {
			qgroup_rescan_zero_tracking(fs_info);
1830 1831
			btrfs_queue_work(fs_info->qgroup_rescan_workers,
					 &fs_info->qgroup_rescan_work);
1832
		}
1833 1834 1835
		ret = 0;
	}

1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855
out:

	return ret;
}

/*
 * copy the acounting information between qgroups. This is necessary when a
 * snapshot or a subvolume is created
 */
int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans,
			 struct btrfs_fs_info *fs_info, u64 srcid, u64 objectid,
			 struct btrfs_qgroup_inherit *inherit)
{
	int ret = 0;
	int i;
	u64 *i_qgroups;
	struct btrfs_root *quota_root = fs_info->quota_root;
	struct btrfs_qgroup *srcgroup;
	struct btrfs_qgroup *dstgroup;
	u32 level_size = 0;
1856
	u64 nums;
1857

1858
	mutex_lock(&fs_info->qgroup_ioctl_lock);
1859
	if (!fs_info->quota_enabled)
1860
		goto out;
1861

1862 1863 1864 1865
	if (!quota_root) {
		ret = -EINVAL;
		goto out;
	}
1866

1867 1868 1869 1870 1871 1872 1873 1874 1875 1876
	if (inherit) {
		i_qgroups = (u64 *)(inherit + 1);
		nums = inherit->num_qgroups + 2 * inherit->num_ref_copies +
		       2 * inherit->num_excl_copies;
		for (i = 0; i < nums; ++i) {
			srcgroup = find_qgroup_rb(fs_info, *i_qgroups);
			if (!srcgroup) {
				ret = -EINVAL;
				goto out;
			}
1877 1878 1879 1880 1881

			if ((srcgroup->qgroupid >> 48) <= (objectid >> 48)) {
				ret = -EINVAL;
				goto out;
			}
1882 1883 1884 1885
			++i_qgroups;
		}
	}

1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906
	/*
	 * create a tracking group for the subvol itself
	 */
	ret = add_qgroup_item(trans, quota_root, objectid);
	if (ret)
		goto out;

	if (srcid) {
		struct btrfs_root *srcroot;
		struct btrfs_key srckey;

		srckey.objectid = srcid;
		srckey.type = BTRFS_ROOT_ITEM_KEY;
		srckey.offset = (u64)-1;
		srcroot = btrfs_read_fs_root_no_name(fs_info, &srckey);
		if (IS_ERR(srcroot)) {
			ret = PTR_ERR(srcroot);
			goto out;
		}

		rcu_read_lock();
1907
		level_size = srcroot->nodesize;
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932
		rcu_read_unlock();
	}

	/*
	 * add qgroup to all inherited groups
	 */
	if (inherit) {
		i_qgroups = (u64 *)(inherit + 1);
		for (i = 0; i < inherit->num_qgroups; ++i) {
			ret = add_qgroup_relation_item(trans, quota_root,
						       objectid, *i_qgroups);
			if (ret)
				goto out;
			ret = add_qgroup_relation_item(trans, quota_root,
						       *i_qgroups, objectid);
			if (ret)
				goto out;
			++i_qgroups;
		}
	}


	spin_lock(&fs_info->qgroup_lock);

	dstgroup = add_qgroup_rb(fs_info, objectid);
1933 1934
	if (IS_ERR(dstgroup)) {
		ret = PTR_ERR(dstgroup);
1935
		goto unlock;
1936
	}
1937

1938 1939 1940 1941 1942 1943
	if (inherit && inherit->flags & BTRFS_QGROUP_INHERIT_SET_LIMITS) {
		dstgroup->lim_flags = inherit->lim.flags;
		dstgroup->max_rfer = inherit->lim.max_rfer;
		dstgroup->max_excl = inherit->lim.max_excl;
		dstgroup->rsv_rfer = inherit->lim.rsv_rfer;
		dstgroup->rsv_excl = inherit->lim.rsv_excl;
1944 1945 1946 1947 1948 1949 1950 1951

		ret = update_qgroup_limit_item(trans, quota_root, dstgroup);
		if (ret) {
			fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
			btrfs_info(fs_info, "unable to update quota limit for %llu",
			       dstgroup->qgroupid);
			goto unlock;
		}
1952 1953
	}

1954 1955
	if (srcid) {
		srcgroup = find_qgroup_rb(fs_info, srcid);
1956
		if (!srcgroup)
1957
			goto unlock;
J
Josef Bacik 已提交
1958 1959 1960 1961 1962 1963 1964 1965 1966 1967

		/*
		 * We call inherit after we clone the root in order to make sure
		 * our counts don't go crazy, so at this point the only
		 * difference between the two roots should be the root node.
		 */
		dstgroup->rfer = srcgroup->rfer;
		dstgroup->rfer_cmpr = srcgroup->rfer_cmpr;
		dstgroup->excl = level_size;
		dstgroup->excl_cmpr = level_size;
1968 1969
		srcgroup->excl = level_size;
		srcgroup->excl_cmpr = level_size;
1970 1971 1972 1973 1974 1975 1976 1977

		/* inherit the limit info */
		dstgroup->lim_flags = srcgroup->lim_flags;
		dstgroup->max_rfer = srcgroup->max_rfer;
		dstgroup->max_excl = srcgroup->max_excl;
		dstgroup->rsv_rfer = srcgroup->rsv_rfer;
		dstgroup->rsv_excl = srcgroup->rsv_excl;

1978 1979 1980 1981
		qgroup_dirty(fs_info, dstgroup);
		qgroup_dirty(fs_info, srcgroup);
	}

1982
	if (!inherit)
1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
		goto unlock;

	i_qgroups = (u64 *)(inherit + 1);
	for (i = 0; i < inherit->num_qgroups; ++i) {
		ret = add_relation_rb(quota_root->fs_info, objectid,
				      *i_qgroups);
		if (ret)
			goto unlock;
		++i_qgroups;
	}

	for (i = 0; i <  inherit->num_ref_copies; ++i) {
		struct btrfs_qgroup *src;
		struct btrfs_qgroup *dst;

		src = find_qgroup_rb(fs_info, i_qgroups[0]);
		dst = find_qgroup_rb(fs_info, i_qgroups[1]);

		if (!src || !dst) {
			ret = -EINVAL;
			goto unlock;
		}

		dst->rfer = src->rfer - level_size;
		dst->rfer_cmpr = src->rfer_cmpr - level_size;
		i_qgroups += 2;
	}
	for (i = 0; i <  inherit->num_excl_copies; ++i) {
		struct btrfs_qgroup *src;
		struct btrfs_qgroup *dst;

		src = find_qgroup_rb(fs_info, i_qgroups[0]);
		dst = find_qgroup_rb(fs_info, i_qgroups[1]);

		if (!src || !dst) {
			ret = -EINVAL;
			goto unlock;
		}

		dst->excl = src->excl + level_size;
		dst->excl_cmpr = src->excl_cmpr + level_size;
		i_qgroups += 2;
	}

unlock:
	spin_unlock(&fs_info->qgroup_lock);
out:
2030
	mutex_unlock(&fs_info->qgroup_ioctl_lock);
2031 2032 2033
	return ret;
}

2034
static int qgroup_reserve(struct btrfs_root *root, u64 num_bytes)
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062
{
	struct btrfs_root *quota_root;
	struct btrfs_qgroup *qgroup;
	struct btrfs_fs_info *fs_info = root->fs_info;
	u64 ref_root = root->root_key.objectid;
	int ret = 0;
	struct ulist_node *unode;
	struct ulist_iterator uiter;

	if (!is_fstree(ref_root))
		return 0;

	if (num_bytes == 0)
		return 0;

	spin_lock(&fs_info->qgroup_lock);
	quota_root = fs_info->quota_root;
	if (!quota_root)
		goto out;

	qgroup = find_qgroup_rb(fs_info, ref_root);
	if (!qgroup)
		goto out;

	/*
	 * in a first step, we check all affected qgroups if any limits would
	 * be exceeded
	 */
2063 2064
	ulist_reinit(fs_info->qgroup_ulist);
	ret = ulist_add(fs_info->qgroup_ulist, qgroup->qgroupid,
2065 2066 2067
			(uintptr_t)qgroup, GFP_ATOMIC);
	if (ret < 0)
		goto out;
2068
	ULIST_ITER_INIT(&uiter);
2069
	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
2070 2071 2072
		struct btrfs_qgroup *qg;
		struct btrfs_qgroup_list *glist;

J
Josef Bacik 已提交
2073
		qg = u64_to_ptr(unode->aux);
2074 2075

		if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_RFER) &&
2076
		    qg->reserved + (s64)qg->rfer + num_bytes >
2077
		    qg->max_rfer) {
2078
			ret = -EDQUOT;
2079 2080
			goto out;
		}
2081 2082

		if ((qg->lim_flags & BTRFS_QGROUP_LIMIT_MAX_EXCL) &&
2083
		    qg->reserved + (s64)qg->excl + num_bytes >
2084
		    qg->max_excl) {
2085
			ret = -EDQUOT;
2086 2087
			goto out;
		}
2088 2089

		list_for_each_entry(glist, &qg->groups, next_group) {
2090 2091
			ret = ulist_add(fs_info->qgroup_ulist,
					glist->group->qgroupid,
2092 2093 2094
					(uintptr_t)glist->group, GFP_ATOMIC);
			if (ret < 0)
				goto out;
2095 2096
		}
	}
2097
	ret = 0;
2098 2099 2100 2101
	/*
	 * no limits exceeded, now record the reservation into all qgroups
	 */
	ULIST_ITER_INIT(&uiter);
2102
	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
2103 2104
		struct btrfs_qgroup *qg;

J
Josef Bacik 已提交
2105
		qg = u64_to_ptr(unode->aux);
2106

2107
		qg->reserved += num_bytes;
2108 2109 2110 2111 2112 2113 2114
	}

out:
	spin_unlock(&fs_info->qgroup_lock);
	return ret;
}

2115 2116
void btrfs_qgroup_free_refroot(struct btrfs_fs_info *fs_info,
			       u64 ref_root, u64 num_bytes)
2117 2118 2119 2120 2121
{
	struct btrfs_root *quota_root;
	struct btrfs_qgroup *qgroup;
	struct ulist_node *unode;
	struct ulist_iterator uiter;
2122
	int ret = 0;
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139

	if (!is_fstree(ref_root))
		return;

	if (num_bytes == 0)
		return;

	spin_lock(&fs_info->qgroup_lock);

	quota_root = fs_info->quota_root;
	if (!quota_root)
		goto out;

	qgroup = find_qgroup_rb(fs_info, ref_root);
	if (!qgroup)
		goto out;

2140 2141
	ulist_reinit(fs_info->qgroup_ulist);
	ret = ulist_add(fs_info->qgroup_ulist, qgroup->qgroupid,
2142 2143 2144
			(uintptr_t)qgroup, GFP_ATOMIC);
	if (ret < 0)
		goto out;
2145
	ULIST_ITER_INIT(&uiter);
2146
	while ((unode = ulist_next(fs_info->qgroup_ulist, &uiter))) {
2147 2148 2149
		struct btrfs_qgroup *qg;
		struct btrfs_qgroup_list *glist;

J
Josef Bacik 已提交
2150
		qg = u64_to_ptr(unode->aux);
2151

2152
		qg->reserved -= num_bytes;
2153 2154

		list_for_each_entry(glist, &qg->groups, next_group) {
2155 2156
			ret = ulist_add(fs_info->qgroup_ulist,
					glist->group->qgroupid,
2157 2158 2159
					(uintptr_t)glist->group, GFP_ATOMIC);
			if (ret < 0)
				goto out;
2160 2161 2162 2163 2164 2165 2166
		}
	}

out:
	spin_unlock(&fs_info->qgroup_lock);
}

2167 2168 2169 2170 2171
static inline void qgroup_free(struct btrfs_root *root, u64 num_bytes)
{
	return btrfs_qgroup_free_refroot(root->fs_info, root->objectid,
					 num_bytes);
}
2172 2173 2174 2175
void assert_qgroups_uptodate(struct btrfs_trans_handle *trans)
{
	if (list_empty(&trans->qgroup_ref_list) && !trans->delayed_ref_elem.seq)
		return;
2176 2177 2178
	btrfs_err(trans->root->fs_info,
		"qgroups not uptodate in trans handle %p:  list is%s empty, "
		"seq is %#x.%x",
2179
		trans, list_empty(&trans->qgroup_ref_list) ? "" : " not",
2180 2181
		(u32)(trans->delayed_ref_elem.seq >> 32),
		(u32)trans->delayed_ref_elem.seq);
2182 2183
	BUG();
}
J
Jan Schmidt 已提交
2184 2185 2186

/*
 * returns < 0 on error, 0 when more leafs are to be scanned.
2187
 * returns 1 when done.
J
Jan Schmidt 已提交
2188 2189
 */
static int
2190
qgroup_rescan_leaf(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
2191
		   struct btrfs_trans_handle *trans)
J
Jan Schmidt 已提交
2192 2193
{
	struct btrfs_key found;
2194
	struct extent_buffer *scratch_leaf = NULL;
J
Jan Schmidt 已提交
2195
	struct ulist *roots = NULL;
2196
	struct seq_list tree_mod_seq_elem = SEQ_LIST_INIT(tree_mod_seq_elem);
J
Josef Bacik 已提交
2197
	u64 num_bytes;
J
Jan Schmidt 已提交
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207
	int slot;
	int ret;

	path->leave_spinning = 1;
	mutex_lock(&fs_info->qgroup_rescan_lock);
	ret = btrfs_search_slot_for_read(fs_info->extent_root,
					 &fs_info->qgroup_rescan_progress,
					 path, 1, 0);

	pr_debug("current progress key (%llu %u %llu), search_slot ret %d\n",
2208
		 fs_info->qgroup_rescan_progress.objectid,
J
Jan Schmidt 已提交
2209
		 fs_info->qgroup_rescan_progress.type,
2210
		 fs_info->qgroup_rescan_progress.offset, ret);
J
Jan Schmidt 已提交
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231

	if (ret) {
		/*
		 * The rescan is about to end, we will not be scanning any
		 * further blocks. We cannot unset the RESCAN flag here, because
		 * we want to commit the transaction if everything went well.
		 * To make the live accounting work in this phase, we set our
		 * scan progress pointer such that every real extent objectid
		 * will be smaller.
		 */
		fs_info->qgroup_rescan_progress.objectid = (u64)-1;
		btrfs_release_path(path);
		mutex_unlock(&fs_info->qgroup_rescan_lock);
		return ret;
	}

	btrfs_item_key_to_cpu(path->nodes[0], &found,
			      btrfs_header_nritems(path->nodes[0]) - 1);
	fs_info->qgroup_rescan_progress.objectid = found.objectid + 1;

	btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
2232 2233 2234 2235 2236 2237 2238 2239 2240
	scratch_leaf = btrfs_clone_extent_buffer(path->nodes[0]);
	if (!scratch_leaf) {
		ret = -ENOMEM;
		mutex_unlock(&fs_info->qgroup_rescan_lock);
		goto out;
	}
	extent_buffer_get(scratch_leaf);
	btrfs_tree_read_lock(scratch_leaf);
	btrfs_set_lock_blocking_rw(scratch_leaf, BTRFS_READ_LOCK);
J
Jan Schmidt 已提交
2241 2242 2243 2244 2245 2246
	slot = path->slots[0];
	btrfs_release_path(path);
	mutex_unlock(&fs_info->qgroup_rescan_lock);

	for (; slot < btrfs_header_nritems(scratch_leaf); ++slot) {
		btrfs_item_key_to_cpu(scratch_leaf, &found, slot);
2247 2248
		if (found.type != BTRFS_EXTENT_ITEM_KEY &&
		    found.type != BTRFS_METADATA_ITEM_KEY)
J
Jan Schmidt 已提交
2249
			continue;
2250
		if (found.type == BTRFS_METADATA_ITEM_KEY)
2251
			num_bytes = fs_info->extent_root->nodesize;
2252 2253 2254
		else
			num_bytes = found.offset;

J
Josef Bacik 已提交
2255 2256
		ret = btrfs_find_all_roots(NULL, fs_info, found.objectid, 0,
					   &roots);
J
Jan Schmidt 已提交
2257 2258
		if (ret < 0)
			goto out;
2259 2260 2261 2262
		/* For rescan, just pass old_roots as NULL */
		ret = btrfs_qgroup_account_extent(trans, fs_info,
				found.objectid, num_bytes, NULL, roots);
		if (ret < 0)
J
Josef Bacik 已提交
2263
			goto out;
J
Jan Schmidt 已提交
2264 2265
	}
out:
2266 2267 2268 2269
	if (scratch_leaf) {
		btrfs_tree_read_unlock_blocking(scratch_leaf);
		free_extent_buffer(scratch_leaf);
	}
J
Jan Schmidt 已提交
2270 2271 2272 2273 2274
	btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);

	return ret;
}

2275
static void btrfs_qgroup_rescan_worker(struct btrfs_work *work)
J
Jan Schmidt 已提交
2276
{
2277 2278
	struct btrfs_fs_info *fs_info = container_of(work, struct btrfs_fs_info,
						     qgroup_rescan_work);
J
Jan Schmidt 已提交
2279 2280 2281
	struct btrfs_path *path;
	struct btrfs_trans_handle *trans = NULL;
	int err = -ENOMEM;
2282
	int ret = 0;
J
Jan Schmidt 已提交
2283 2284 2285 2286 2287 2288

	path = btrfs_alloc_path();
	if (!path)
		goto out;

	err = 0;
2289
	while (!err && !btrfs_fs_closing(fs_info)) {
J
Jan Schmidt 已提交
2290 2291 2292 2293 2294 2295 2296 2297
		trans = btrfs_start_transaction(fs_info->fs_root, 0);
		if (IS_ERR(trans)) {
			err = PTR_ERR(trans);
			break;
		}
		if (!fs_info->quota_enabled) {
			err = -EINTR;
		} else {
2298
			err = qgroup_rescan_leaf(fs_info, path, trans);
J
Jan Schmidt 已提交
2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309
		}
		if (err > 0)
			btrfs_commit_transaction(trans, fs_info->fs_root);
		else
			btrfs_end_transaction(trans, fs_info->fs_root);
	}

out:
	btrfs_free_path(path);

	mutex_lock(&fs_info->qgroup_rescan_lock);
2310 2311
	if (!btrfs_fs_closing(fs_info))
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
J
Jan Schmidt 已提交
2312

2313
	if (err > 0 &&
J
Jan Schmidt 已提交
2314 2315 2316 2317 2318 2319 2320
	    fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT) {
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
	} else if (err < 0) {
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_INCONSISTENT;
	}
	mutex_unlock(&fs_info->qgroup_rescan_lock);

2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335
	/*
	 * only update status, since the previous part has alreay updated the
	 * qgroup info.
	 */
	trans = btrfs_start_transaction(fs_info->quota_root, 1);
	if (IS_ERR(trans)) {
		err = PTR_ERR(trans);
		btrfs_err(fs_info,
			  "fail to start transaction for status update: %d\n",
			  err);
		goto done;
	}
	ret = update_qgroup_status_item(trans, fs_info, fs_info->quota_root);
	if (ret < 0) {
		err = ret;
2336
		btrfs_err(fs_info, "fail to update qgroup status: %d\n", err);
2337 2338 2339
	}
	btrfs_end_transaction(trans, fs_info->quota_root);

2340 2341 2342
	if (btrfs_fs_closing(fs_info)) {
		btrfs_info(fs_info, "qgroup scan paused");
	} else if (err >= 0) {
2343
		btrfs_info(fs_info, "qgroup scan completed%s",
2344
			err > 0 ? " (inconsistency flag cleared)" : "");
J
Jan Schmidt 已提交
2345
	} else {
2346
		btrfs_err(fs_info, "qgroup scan failed with %d", err);
J
Jan Schmidt 已提交
2347
	}
2348

2349
done:
2350
	complete_all(&fs_info->qgroup_rescan_completion);
J
Jan Schmidt 已提交
2351 2352
}

2353 2354 2355 2356 2357 2358 2359
/*
 * Checks that (a) no rescan is running and (b) quota is enabled. Allocates all
 * memory required for the rescan context.
 */
static int
qgroup_rescan_init(struct btrfs_fs_info *fs_info, u64 progress_objectid,
		   int init_flags)
J
Jan Schmidt 已提交
2360 2361 2362
{
	int ret = 0;

2363 2364 2365 2366 2367 2368
	if (!init_flags &&
	    (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) ||
	     !(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON))) {
		ret = -EINVAL;
		goto err;
	}
J
Jan Schmidt 已提交
2369 2370 2371

	mutex_lock(&fs_info->qgroup_rescan_lock);
	spin_lock(&fs_info->qgroup_lock);
2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384

	if (init_flags) {
		if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
			ret = -EINPROGRESS;
		else if (!(fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_ON))
			ret = -EINVAL;

		if (ret) {
			spin_unlock(&fs_info->qgroup_lock);
			mutex_unlock(&fs_info->qgroup_rescan_lock);
			goto err;
		}
		fs_info->qgroup_flags |= BTRFS_QGROUP_STATUS_FLAG_RESCAN;
J
Jan Schmidt 已提交
2385 2386 2387 2388
	}

	memset(&fs_info->qgroup_rescan_progress, 0,
		sizeof(fs_info->qgroup_rescan_progress));
2389 2390 2391 2392 2393
	fs_info->qgroup_rescan_progress.objectid = progress_objectid;

	spin_unlock(&fs_info->qgroup_lock);
	mutex_unlock(&fs_info->qgroup_rescan_lock);

2394
	init_completion(&fs_info->qgroup_rescan_completion);
J
Jan Schmidt 已提交
2395

2396 2397
	memset(&fs_info->qgroup_rescan_work, 0,
	       sizeof(fs_info->qgroup_rescan_work));
2398
	btrfs_init_work(&fs_info->qgroup_rescan_work,
2399
			btrfs_qgroup_rescan_helper,
2400
			btrfs_qgroup_rescan_worker, NULL, NULL);
2401 2402 2403

	if (ret) {
err:
2404
		btrfs_info(fs_info, "qgroup_rescan_init failed with %d", ret);
2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417
		return ret;
	}

	return 0;
}

static void
qgroup_rescan_zero_tracking(struct btrfs_fs_info *fs_info)
{
	struct rb_node *n;
	struct btrfs_qgroup *qgroup;

	spin_lock(&fs_info->qgroup_lock);
J
Jan Schmidt 已提交
2418 2419 2420 2421 2422 2423 2424 2425 2426
	/* clear all current qgroup tracking information */
	for (n = rb_first(&fs_info->qgroup_tree); n; n = rb_next(n)) {
		qgroup = rb_entry(n, struct btrfs_qgroup, node);
		qgroup->rfer = 0;
		qgroup->rfer_cmpr = 0;
		qgroup->excl = 0;
		qgroup->excl_cmpr = 0;
	}
	spin_unlock(&fs_info->qgroup_lock);
2427
}
J
Jan Schmidt 已提交
2428

2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462
int
btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info)
{
	int ret = 0;
	struct btrfs_trans_handle *trans;

	ret = qgroup_rescan_init(fs_info, 0, 1);
	if (ret)
		return ret;

	/*
	 * We have set the rescan_progress to 0, which means no more
	 * delayed refs will be accounted by btrfs_qgroup_account_ref.
	 * However, btrfs_qgroup_account_ref may be right after its call
	 * to btrfs_find_all_roots, in which case it would still do the
	 * accounting.
	 * To solve this, we're committing the transaction, which will
	 * ensure we run all delayed refs and only after that, we are
	 * going to clear all tracking information for a clean start.
	 */

	trans = btrfs_join_transaction(fs_info->fs_root);
	if (IS_ERR(trans)) {
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
		return PTR_ERR(trans);
	}
	ret = btrfs_commit_transaction(trans, fs_info->fs_root);
	if (ret) {
		fs_info->qgroup_flags &= ~BTRFS_QGROUP_STATUS_FLAG_RESCAN;
		return ret;
	}

	qgroup_rescan_zero_tracking(fs_info);

2463 2464
	btrfs_queue_work(fs_info->qgroup_rescan_workers,
			 &fs_info->qgroup_rescan_work);
J
Jan Schmidt 已提交
2465 2466 2467

	return 0;
}
2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485

int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info)
{
	int running;
	int ret = 0;

	mutex_lock(&fs_info->qgroup_rescan_lock);
	spin_lock(&fs_info->qgroup_lock);
	running = fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN;
	spin_unlock(&fs_info->qgroup_lock);
	mutex_unlock(&fs_info->qgroup_rescan_lock);

	if (running)
		ret = wait_for_completion_interruptible(
					&fs_info->qgroup_rescan_completion);

	return ret;
}
2486 2487 2488 2489 2490 2491 2492 2493 2494

/*
 * this is only called from open_ctree where we're still single threaded, thus
 * locking is omitted here.
 */
void
btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info)
{
	if (fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN)
2495 2496
		btrfs_queue_work(fs_info->qgroup_rescan_workers,
				 &fs_info->qgroup_rescan_work);
2497
}
2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526

/*
 * Reserve qgroup space for range [start, start + len).
 *
 * This function will either reserve space from related qgroups or doing
 * nothing if the range is already reserved.
 *
 * Return 0 for successful reserve
 * Return <0 for error (including -EQUOT)
 *
 * NOTE: this function may sleep for memory allocation.
 */
int btrfs_qgroup_reserve_data(struct inode *inode, u64 start, u64 len)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_changeset changeset;
	struct ulist_node *unode;
	struct ulist_iterator uiter;
	int ret;

	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid) ||
	    len == 0)
		return 0;

	changeset.bytes_changed = 0;
	changeset.range_changed = ulist_alloc(GFP_NOFS);
	ret = set_record_extent_bits(&BTRFS_I(inode)->io_tree, start,
			start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS,
			&changeset);
2527 2528 2529
	trace_btrfs_qgroup_reserve_data(inode, start, len,
					changeset.bytes_changed,
					QGROUP_RESERVE);
2530 2531
	if (ret < 0)
		goto cleanup;
2532
	ret = qgroup_reserve(root, changeset.bytes_changed);
2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548
	if (ret < 0)
		goto cleanup;

	ulist_free(changeset.range_changed);
	return ret;

cleanup:
	/* cleanup already reserved ranges */
	ULIST_ITER_INIT(&uiter);
	while ((unode = ulist_next(changeset.range_changed, &uiter)))
		clear_extent_bit(&BTRFS_I(inode)->io_tree, unode->val,
				 unode->aux, EXTENT_QGROUP_RESERVED, 0, 0, NULL,
				 GFP_NOFS);
	ulist_free(changeset.range_changed);
	return ret;
}
2549 2550 2551 2552 2553

static int __btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len,
				       int free)
{
	struct extent_changeset changeset;
2554
	int trace_op = QGROUP_RELEASE;
2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567
	int ret;

	changeset.bytes_changed = 0;
	changeset.range_changed = ulist_alloc(GFP_NOFS);
	if (!changeset.range_changed)
		return -ENOMEM;

	ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, start, 
			start + len -1, EXTENT_QGROUP_RESERVED, GFP_NOFS,
			&changeset);
	if (ret < 0)
		goto out;

2568
	if (free) {
2569
		qgroup_free(BTRFS_I(inode)->root, changeset.bytes_changed);
2570 2571 2572 2573
		trace_op = QGROUP_FREE;
	}
	trace_btrfs_qgroup_release_data(inode, start, len,
					changeset.bytes_changed, trace_op);
2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612
out:
	ulist_free(changeset.range_changed);
	return ret;
}

/*
 * Free a reserved space range from io_tree and related qgroups
 *
 * Should be called when a range of pages get invalidated before reaching disk.
 * Or for error cleanup case.
 *
 * For data written to disk, use btrfs_qgroup_release_data().
 *
 * NOTE: This function may sleep for memory allocation.
 */
int btrfs_qgroup_free_data(struct inode *inode, u64 start, u64 len)
{
	return __btrfs_qgroup_release_data(inode, start, len, 1);
}

/*
 * Release a reserved space range from io_tree only.
 *
 * Should be called when a range of pages get written to disk and corresponding
 * FILE_EXTENT is inserted into corresponding root.
 *
 * Since new qgroup accounting framework will only update qgroup numbers at
 * commit_transaction() time, its reserved space shouldn't be freed from
 * related qgroups.
 *
 * But we should release the range from io_tree, to allow further write to be
 * COWed.
 *
 * NOTE: This function may sleep for memory allocation.
 */
int btrfs_qgroup_release_data(struct inode *inode, u64 start, u64 len)
{
	return __btrfs_qgroup_release_data(inode, start, len, 0);
}
2613 2614 2615 2616 2617 2618 2619 2620 2621 2622

int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes)
{
	int ret;

	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid) ||
	    num_bytes == 0)
		return 0;

	BUG_ON(num_bytes != round_down(num_bytes, root->nodesize));
2623
	ret = qgroup_reserve(root, num_bytes);
2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639
	if (ret < 0)
		return ret;
	atomic_add(num_bytes, &root->qgroup_meta_rsv);
	return ret;
}

void btrfs_qgroup_free_meta_all(struct btrfs_root *root)
{
	int reserved;

	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid))
		return;

	reserved = atomic_xchg(&root->qgroup_meta_rsv, 0);
	if (reserved == 0)
		return;
2640
	qgroup_free(root, reserved);
2641 2642 2643 2644 2645 2646 2647 2648 2649 2650
}

void btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes)
{
	if (!root->fs_info->quota_enabled || !is_fstree(root->objectid))
		return;

	BUG_ON(num_bytes != round_down(num_bytes, root->nodesize));
	WARN_ON(atomic_read(&root->qgroup_meta_rsv) < num_bytes);
	atomic_sub(num_bytes, &root->qgroup_meta_rsv);
2651
	qgroup_free(root, num_bytes);
2652
}
2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684

/*
 * Check qgroup reserved space leaking, normally at destory inode
 * time
 */
void btrfs_qgroup_check_reserved_leak(struct inode *inode)
{
	struct extent_changeset changeset;
	struct ulist_node *unode;
	struct ulist_iterator iter;
	int ret;

	changeset.bytes_changed = 0;
	changeset.range_changed = ulist_alloc(GFP_NOFS);
	if (WARN_ON(!changeset.range_changed))
		return;

	ret = clear_record_extent_bits(&BTRFS_I(inode)->io_tree, 0, (u64)-1,
			EXTENT_QGROUP_RESERVED, GFP_NOFS, &changeset);

	WARN_ON(ret < 0);
	if (WARN_ON(changeset.bytes_changed)) {
		ULIST_ITER_INIT(&iter);
		while ((unode = ulist_next(changeset.range_changed, &iter))) {
			btrfs_warn(BTRFS_I(inode)->root->fs_info,
				"leaking qgroup reserved space, ino: %lu, start: %llu, end: %llu",
				inode->i_ino, unode->val, unode->aux);
		}
		qgroup_free(BTRFS_I(inode)->root, changeset.bytes_changed);
	}
	ulist_free(changeset.range_changed);
}