[linux.git] / fs / btrfs / qgroup.h

/* SPDX-License-Identifier: GPL-2.0 */
/*
 * Copyright (C) 2014 Facebook.  All rights reserved.
 */

#ifndef BTRFS_QGROUP_H
#define BTRFS_QGROUP_H

#include <linux/spinlock.h>
#include <linux/rbtree.h>
#include <linux/kobject.h>
#include "ulist.h"
#include "delayed-ref.h"
#include "misc.h"

/*
 * Btrfs qgroup overview
 *
 * Btrfs qgroup splits into 3 main part:
 * 1) Reserve
 *    Reserve metadata/data space for incoming operations
 *    Affect how qgroup limit works
 *
 * 2) Trace
 *    Tell btrfs qgroup to trace dirty extents.
 *
 *    Dirty extents including:
 *    - Newly allocated extents
 *    - Extents going to be deleted (in this trans)
 *    - Extents whose owner is going to be modified
 *
 *    This is the main part affects whether qgroup numbers will stay
 *    consistent.
 *    Btrfs qgroup can trace clean extents and won't cause any problem,
 *    but it will consume extra CPU time, it should be avoided if possible.
 *
 * 3) Account
 *    Btrfs qgroup will updates its numbers, based on dirty extents traced
 *    in previous step.
 *
 *    Normally at qgroup rescan and transaction commit time.
 */

/*
 * Special performance optimization for balance.
 *
 * For balance, we need to swap subtree of subvolume and reloc trees.
 * In theory, we need to trace all subtree blocks of both subvolume and reloc
 * trees, since their owner has changed during such swap.
 *
 * However since balance has ensured that both subtrees are containing the
 * same contents and have the same tree structures, such swap won't cause
 * qgroup number change.
 *
 * But there is a race window between subtree swap and transaction commit,
 * during that window, if we increase/decrease tree level or merge/split tree
 * blocks, we still need to trace the original subtrees.
 *
 * So for balance, we use a delayed subtree tracing, whose workflow is:
 *
 * 1) Record the subtree root block get swapped.
 *
 *    During subtree swap:
 *    O = Old tree blocks
 *    N = New tree blocks
 *          reloc tree                     subvolume tree X
 *             Root                               Root
 *            /    \                             /    \
 *          NA     OB                          OA      OB
 *        /  |     |  \                      /  |      |  \
 *      NC  ND     OE  OF                   OC  OD     OE  OF
 *
 *   In this case, NA and OA are going to be swapped, record (NA, OA) into
 *   subvolume tree X.
 *
 * 2) After subtree swap.
 *          reloc tree                     subvolume tree X
 *             Root                               Root
 *            /    \                             /    \
 *          OA     OB                          NA      OB
 *        /  |     |  \                      /  |      |  \
 *      OC  OD     OE  OF                   NC  ND     OE  OF
 *
 * 3a) COW happens for OB
 *     If we are going to COW tree block OB, we check OB's bytenr against
 *     tree X's swapped_blocks structure.
 *     If it doesn't fit any, nothing will happen.
 *
 * 3b) COW happens for NA
 *     Check NA's bytenr against tree X's swapped_blocks, and get a hit.
 *     Then we do subtree scan on both subtrees OA and NA.
 *     Resulting 6 tree blocks to be scanned (OA, OC, OD, NA, NC, ND).
 *
 *     Then no matter what we do to subvolume tree X, qgroup numbers will
 *     still be correct.
 *     Then NA's record gets removed from X's swapped_blocks.
 *
 * 4)  Transaction commit
 *     Any record in X's swapped_blocks gets removed, since there is no
 *     modification to the swapped subtrees, no need to trigger heavy qgroup
 *     subtree rescan for them.
 */

/*
 * These flags share the flags field of the btrfs_qgroup_status_item with the
 * persisted flags defined in btrfs_tree.h.
 *
 * To minimize the chance of collision with new persisted status flags, these
 * count backwards from the MSB.
 */
#define BTRFS_QGROUP_RUNTIME_FLAG_CANCEL_RESCAN		(1ULL << 63)
#define BTRFS_QGROUP_RUNTIME_FLAG_NO_ACCOUNTING		(1ULL << 62)

/*
 * Record a dirty extent, and info qgroup to update quota on it
 * TODO: Use kmem cache to alloc it.
 */
struct btrfs_qgroup_extent_record {
	struct rb_node node;
	u64 bytenr;
	u64 num_bytes;

	/*
	 * For qgroup reserved data space freeing.
	 *
	 * @data_rsv_refroot and @data_rsv will be recorded after
	 * BTRFS_ADD_DELAYED_EXTENT is called.
	 * And will be used to free reserved qgroup space at
	 * transaction commit time.
	 */
	u32 data_rsv;		/* reserved data space needs to be freed */
	u64 data_rsv_refroot;	/* which root the reserved data belongs to */
	struct ulist *old_roots;
};

struct btrfs_qgroup_swapped_block {
	struct rb_node node;

	int level;
	bool trace_leaf;

	/* bytenr/generation of the tree block in subvolume tree after swap */
	u64 subvol_bytenr;
	u64 subvol_generation;

	/* bytenr/generation of the tree block in reloc tree after swap */
	u64 reloc_bytenr;
	u64 reloc_generation;

	u64 last_snapshot;
	struct btrfs_key first_key;
};

/*
 * Qgroup reservation types:
 *
 * DATA:
 *	space reserved for data
 *
 * META_PERTRANS:
 * 	Space reserved for metadata (per-transaction)
 * 	Due to the fact that qgroup data is only updated at transaction commit
 * 	time, reserved space for metadata must be kept until transaction
 * 	commits.
 * 	Any metadata reserved that are used in btrfs_start_transaction() should
 * 	be of this type.
 *
 * META_PREALLOC:
 *	There are cases where metadata space is reserved before starting
 *	transaction, and then btrfs_join_transaction() to get a trans handle.
 *	Any metadata reserved for such usage should be of this type.
 *	And after join_transaction() part (or all) of such reservation should
 *	be converted into META_PERTRANS.
 */
enum btrfs_qgroup_rsv_type {
	BTRFS_QGROUP_RSV_DATA,
	BTRFS_QGROUP_RSV_META_PERTRANS,
	BTRFS_QGROUP_RSV_META_PREALLOC,
	BTRFS_QGROUP_RSV_LAST,
};

/*
 * Represents how many bytes we have reserved for this qgroup.
 *
 * Each type should have different reservation behavior.
 * E.g, data follows its io_tree flag modification, while
 * *currently* meta is just reserve-and-clear during transaction.
 *
 * TODO: Add new type for reservation which can survive transaction commit.
 * Current metadata reservation behavior is not suitable for such case.
 */
struct btrfs_qgroup_rsv {
	u64 values[BTRFS_QGROUP_RSV_LAST];
};

/*
 * 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
	 */
	struct btrfs_qgroup_rsv rsv;

	/*
	 * 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 */

	/*
	 * For qgroup iteration usage.
	 *
	 * The iteration list should always be empty until qgroup_iterator_add()
	 * is called.  And should be reset to empty after the iteration is
	 * finished.
	 */
	struct list_head iterator;

	/*
	 * For nested iterator usage.
	 *
	 * Here we support at most one level of nested iterator calls like:
	 *
	 *	LIST_HEAD(all_qgroups);
	 *	{
	 *		LIST_HEAD(local_qgroups);
	 *		qgroup_iterator_add(local_qgroups, qg);
	 *		qgroup_iterator_nested_add(all_qgroups, qg);
	 *		do_some_work(local_qgroups);
	 *		qgroup_iterator_clean(local_qgroups);
	 *	}
	 *	do_some_work(all_qgroups);
	 *	qgroup_iterator_nested_clean(all_qgroups);
	 */
	struct list_head nested_iterator;
	struct rb_node node;	  /* tree of qgroups */

	/*
	 * temp variables for accounting operations
	 * Refer to qgroup_shared_accounting() for details.
	 */
	u64 old_refcnt;
	u64 new_refcnt;

	/*
	 * Sysfs kobjectid
	 */
	struct kobject kobj;
};

struct btrfs_squota_delta {
	/* The fstree root this delta counts against. */
	u64 root;
	/* The number of bytes in the extent being counted. */
	u64 num_bytes;
	/* The generation the extent was created in. */
	u64 generation;
	/* Whether we are using or freeing the extent. */
	bool is_inc;
	/* Whether the extent is data or metadata. */
	bool is_data;
};

static inline u64 btrfs_qgroup_subvolid(u64 qgroupid)
{
	return (qgroupid & ((1ULL << BTRFS_QGROUP_LEVEL_SHIFT) - 1));
}

/*
 * For qgroup event trace points only
 */
enum {
	ENUM_BIT(QGROUP_RESERVE),
	ENUM_BIT(QGROUP_RELEASE),
	ENUM_BIT(QGROUP_FREE),
};

enum btrfs_qgroup_mode {
	BTRFS_QGROUP_MODE_DISABLED,
	BTRFS_QGROUP_MODE_FULL,
	BTRFS_QGROUP_MODE_SIMPLE
};

enum btrfs_qgroup_mode btrfs_qgroup_mode(struct btrfs_fs_info *fs_info);
bool btrfs_qgroup_enabled(struct btrfs_fs_info *fs_info);
bool btrfs_qgroup_full_accounting(struct btrfs_fs_info *fs_info);
int btrfs_quota_enable(struct btrfs_fs_info *fs_info,
		       struct btrfs_ioctl_quota_ctl_args *quota_ctl_args);
int btrfs_quota_disable(struct btrfs_fs_info *fs_info);
int btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info);
void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info);
int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info,
				     bool interruptible);
int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans, u64 src, u64 dst);
int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans, u64 src,
			      u64 dst);
int btrfs_create_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid);
int btrfs_remove_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid);
int btrfs_limit_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid,
		       struct btrfs_qgroup_limit *limit);
int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info);
void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info);
struct btrfs_delayed_extent_op;

int btrfs_qgroup_trace_extent_nolock(
		struct btrfs_fs_info *fs_info,
		struct btrfs_delayed_ref_root *delayed_refs,
		struct btrfs_qgroup_extent_record *record);
int btrfs_qgroup_trace_extent_post(struct btrfs_trans_handle *trans,
				   struct btrfs_qgroup_extent_record *qrecord);
int btrfs_qgroup_trace_extent(struct btrfs_trans_handle *trans, u64 bytenr,
			      u64 num_bytes);
int btrfs_qgroup_trace_leaf_items(struct btrfs_trans_handle *trans,
				  struct extent_buffer *eb);
int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
			       struct extent_buffer *root_eb,
			       u64 root_gen, int root_level);
int btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, u64 bytenr,
				u64 num_bytes, struct ulist *old_roots,
				struct ulist *new_roots);
int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans);
int btrfs_run_qgroups(struct btrfs_trans_handle *trans);
int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
			 u64 objectid, u64 inode_rootid,
			 struct btrfs_qgroup_inherit *inherit);
void btrfs_qgroup_free_refroot(struct btrfs_fs_info *fs_info,
			       u64 ref_root, u64 num_bytes,
			       enum btrfs_qgroup_rsv_type type);

#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
int btrfs_verify_qgroup_counts(struct btrfs_fs_info *fs_info, u64 qgroupid,
			       u64 rfer, u64 excl);
#endif

/* New io_tree based accurate qgroup reserve API */
int btrfs_qgroup_reserve_data(struct btrfs_inode *inode,
			struct extent_changeset **reserved, u64 start, u64 len);
int btrfs_qgroup_release_data(struct btrfs_inode *inode, u64 start, u64 len, u64 *released);
int btrfs_qgroup_free_data(struct btrfs_inode *inode,
			   struct extent_changeset *reserved, u64 start,
			   u64 len, u64 *freed);
int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
			      enum btrfs_qgroup_rsv_type type, bool enforce);
int __btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
				enum btrfs_qgroup_rsv_type type, bool enforce,
				bool noflush);
/* Reserve metadata space for pertrans and prealloc type */
static inline int btrfs_qgroup_reserve_meta_pertrans(struct btrfs_root *root,
				int num_bytes, bool enforce)
{
	return __btrfs_qgroup_reserve_meta(root, num_bytes,
					   BTRFS_QGROUP_RSV_META_PERTRANS,
					   enforce, false);
}
static inline int btrfs_qgroup_reserve_meta_prealloc(struct btrfs_root *root,
						     int num_bytes, bool enforce,
						     bool noflush)
{
	return __btrfs_qgroup_reserve_meta(root, num_bytes,
					   BTRFS_QGROUP_RSV_META_PREALLOC,
					   enforce, noflush);
}

void __btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes,
			     enum btrfs_qgroup_rsv_type type);

/* Free per-transaction meta reservation for error handling */
static inline void btrfs_qgroup_free_meta_pertrans(struct btrfs_root *root,
						   int num_bytes)
{
	__btrfs_qgroup_free_meta(root, num_bytes,
			BTRFS_QGROUP_RSV_META_PERTRANS);
}

/* Pre-allocated meta reservation can be freed at need */
static inline void btrfs_qgroup_free_meta_prealloc(struct btrfs_root *root,
						   int num_bytes)
{
	__btrfs_qgroup_free_meta(root, num_bytes,
			BTRFS_QGROUP_RSV_META_PREALLOC);
}

void btrfs_qgroup_free_meta_all_pertrans(struct btrfs_root *root);
void btrfs_qgroup_convert_reserved_meta(struct btrfs_root *root, int num_bytes);
void btrfs_qgroup_check_reserved_leak(struct btrfs_inode *inode);

/* btrfs_qgroup_swapped_blocks related functions */
void btrfs_qgroup_init_swapped_blocks(
	struct btrfs_qgroup_swapped_blocks *swapped_blocks);

void btrfs_qgroup_clean_swapped_blocks(struct btrfs_root *root);
int btrfs_qgroup_add_swapped_blocks(struct btrfs_trans_handle *trans,
		struct btrfs_root *subvol_root,
		struct btrfs_block_group *bg,
		struct extent_buffer *subvol_parent, int subvol_slot,
		struct extent_buffer *reloc_parent, int reloc_slot,
		u64 last_snapshot);
int btrfs_qgroup_trace_subtree_after_cow(struct btrfs_trans_handle *trans,
		struct btrfs_root *root, struct extent_buffer *eb);
void btrfs_qgroup_destroy_extent_records(struct btrfs_transaction *trans);
bool btrfs_check_quota_leak(struct btrfs_fs_info *fs_info);
void btrfs_free_squota_rsv(struct btrfs_fs_info *fs_info, u64 root, u64 rsv_bytes);
int btrfs_record_squota_delta(struct btrfs_fs_info *fs_info,
			      struct btrfs_squota_delta *delta);

#endif
Commit	Line	Data
9888c340	1	/* SPDX-License-Identifier: GPL-2.0 */
fcebe456 JB	2	/*
fcebe456 JB	3	* Copyright (C) 2014 Facebook. All rights reserved.
fcebe456 JB	4	*/
fcebe456 JB	5
9888c340 DS	6	#ifndef BTRFS_QGROUP_H
9888c340 DS	7	#define BTRFS_QGROUP_H
fcebe456	8
370a11b8 QW	9	#include <linux/spinlock.h>
370a11b8 QW	10	#include <linux/rbtree.h>
49e5fb46	11	#include <linux/kobject.h>
3368d001 QW	12	#include "ulist.h"
3368d001 QW	13	#include "delayed-ref.h"
e0a8b9a7	14	#include "misc.h"
3368d001	15
1d2beaa9 QW	16	/*
	17	* Btrfs qgroup overview
	18	*
	19	* Btrfs qgroup splits into 3 main part:
	20	* 1) Reserve
	21	* Reserve metadata/data space for incoming operations
	22	* Affect how qgroup limit works
	23	*
	24	* 2) Trace
	25	* Tell btrfs qgroup to trace dirty extents.
	26	*
	27	* Dirty extents including:
	28	* - Newly allocated extents
	29	* - Extents going to be deleted (in this trans)
	30	* - Extents whose owner is going to be modified
	31	*
	32	* This is the main part affects whether qgroup numbers will stay
	33	* consistent.
	34	* Btrfs qgroup can trace clean extents and won't cause any problem,
	35	* but it will consume extra CPU time, it should be avoided if possible.
	36	*
	37	* 3) Account
	38	* Btrfs qgroup will updates its numbers, based on dirty extents traced
	39	* in previous step.
	40	*
	41	* Normally at qgroup rescan and transaction commit time.
	42	*/
	43
370a11b8 QW	44	/*
	45	* Special performance optimization for balance.
	46	*
	47	* For balance, we need to swap subtree of subvolume and reloc trees.
	48	* In theory, we need to trace all subtree blocks of both subvolume and reloc
	49	* trees, since their owner has changed during such swap.
	50	*
	51	* However since balance has ensured that both subtrees are containing the
	52	* same contents and have the same tree structures, such swap won't cause
	53	* qgroup number change.
	54	*
	55	* But there is a race window between subtree swap and transaction commit,
	56	* during that window, if we increase/decrease tree level or merge/split tree
	57	* blocks, we still need to trace the original subtrees.
	58	*
	59	* So for balance, we use a delayed subtree tracing, whose workflow is:
	60	*
	61	* 1) Record the subtree root block get swapped.
	62	*
	63	* During subtree swap:
	64	* O = Old tree blocks
	65	* N = New tree blocks
	66	* reloc tree subvolume tree X
	67	* Root Root
	68	* / \ / \
	69	* NA OB OA OB
	70	* / \| \| \ / \| \| \
	71	* NC ND OE OF OC OD OE OF
	72	*
	73	* In this case, NA and OA are going to be swapped, record (NA, OA) into
	74	* subvolume tree X.
	75	*
	76	* 2) After subtree swap.
	77	* reloc tree subvolume tree X
	78	* Root Root
	79	* / \ / \
	80	* OA OB NA OB
	81	* / \| \| \ / \| \| \
	82	* OC OD OE OF NC ND OE OF
	83	*
	84	* 3a) COW happens for OB
	85	* If we are going to COW tree block OB, we check OB's bytenr against
	86	* tree X's swapped_blocks structure.
	87	* If it doesn't fit any, nothing will happen.
	88	*
	89	* 3b) COW happens for NA
	90	* Check NA's bytenr against tree X's swapped_blocks, and get a hit.
	91	* Then we do subtree scan on both subtrees OA and NA.
	92	* Resulting 6 tree blocks to be scanned (OA, OC, OD, NA, NC, ND).
	93	*
	94	* Then no matter what we do to subvolume tree X, qgroup numbers will
	95	* still be correct.
	96	* Then NA's record gets removed from X's swapped_blocks.
	97	*
	98	* 4) Transaction commit
	99	* Any record in X's swapped_blocks gets removed, since there is no
	100	* modification to the swapped subtrees, no need to trigger heavy qgroup
	101	* subtree rescan for them.
	102	*/
	103
182940f4 BB	104	/*
	105	* These flags share the flags field of the btrfs_qgroup_status_item with the
	106	* persisted flags defined in btrfs_tree.h.
	107	*
	108	* To minimize the chance of collision with new persisted status flags, these
	109	* count backwards from the MSB.
	110	*/
	111	#define BTRFS_QGROUP_RUNTIME_FLAG_CANCEL_RESCAN (1ULL << 63)
	112	#define BTRFS_QGROUP_RUNTIME_FLAG_NO_ACCOUNTING (1ULL << 62)
e562a8bd	113
3368d001 QW	114	/*
	115	* Record a dirty extent, and info qgroup to update quota on it
	116	* TODO: Use kmem cache to alloc it.
	117	*/
	118	struct btrfs_qgroup_extent_record {
	119	struct rb_node node;
	120	u64 bytenr;
	121	u64 num_bytes;
1418bae1 QW	122
	123	/*
	124	* For qgroup reserved data space freeing.
	125	*
	126	* @data_rsv_refroot and @data_rsv will be recorded after
	127	* BTRFS_ADD_DELAYED_EXTENT is called.
	128	* And will be used to free reserved qgroup space at
	129	* transaction commit time.
	130	*/
	131	u32 data_rsv; /* reserved data space needs to be freed */
	132	u64 data_rsv_refroot; /* which root the reserved data belongs to */
3368d001 QW	133	struct ulist *old_roots;
	134	};
	135
370a11b8 QW	136	struct btrfs_qgroup_swapped_block {
	137	struct rb_node node;
	138
	139	int level;
	140	bool trace_leaf;
	141
	142	/* bytenr/generation of the tree block in subvolume tree after swap */
	143	u64 subvol_bytenr;
	144	u64 subvol_generation;
	145
	146	/* bytenr/generation of the tree block in reloc tree after swap */
	147	u64 reloc_bytenr;
	148	u64 reloc_generation;
	149
	150	u64 last_snapshot;
	151	struct btrfs_key first_key;
	152	};
	153
733e03a0 QW	154	/*
	155	* Qgroup reservation types:
	156	*
	157	* DATA:
	158	* space reserved for data
	159	*
	160	* META_PERTRANS:
	161	* Space reserved for metadata (per-transaction)
	162	* Due to the fact that qgroup data is only updated at transaction commit
	163	* time, reserved space for metadata must be kept until transaction
	164	* commits.
	165	* Any metadata reserved that are used in btrfs_start_transaction() should
	166	* be of this type.
	167	*
	168	* META_PREALLOC:
	169	* There are cases where metadata space is reserved before starting
	170	* transaction, and then btrfs_join_transaction() to get a trans handle.
	171	* Any metadata reserved for such usage should be of this type.
	172	* And after join_transaction() part (or all) of such reservation should
	173	* be converted into META_PERTRANS.
	174	*/
d4e5c920	175	enum btrfs_qgroup_rsv_type {
bbe339cc	176	BTRFS_QGROUP_RSV_DATA,
733e03a0 QW	177	BTRFS_QGROUP_RSV_META_PERTRANS,
733e03a0 QW	178	BTRFS_QGROUP_RSV_META_PREALLOC,
d4e5c920 QW	179	BTRFS_QGROUP_RSV_LAST,
	180	};
	181
	182	/*
	183	* Represents how many bytes we have reserved for this qgroup.
	184	*
	185	* Each type should have different reservation behavior.
	186	* E.g, data follows its io_tree flag modification, while
52042d8e	187	* currently meta is just reserve-and-clear during transaction.
d4e5c920 QW	188	*
d4e5c920 QW	189	* TODO: Add new type for reservation which can survive transaction commit.
52042d8e	190	* Current metadata reservation behavior is not suitable for such case.
d4e5c920 QW	191	*/
	192	struct btrfs_qgroup_rsv {
	193	u64 values[BTRFS_QGROUP_RSV_LAST];
	194	};
	195
3159fe7b QW	196	/*
	197	* one struct for each qgroup, organized in fs_info->qgroup_tree.
	198	*/
	199	struct btrfs_qgroup {
	200	u64 qgroupid;
	201
	202	/*
	203	* state
	204	*/
	205	u64 rfer; /* referenced */
	206	u64 rfer_cmpr; /* referenced compressed */
	207	u64 excl; /* exclusive */
	208	u64 excl_cmpr; /* exclusive compressed */
	209
	210	/*
	211	* limits
	212	*/
	213	u64 lim_flags; /* which limits are set */
	214	u64 max_rfer;
	215	u64 max_excl;
	216	u64 rsv_rfer;
	217	u64 rsv_excl;
	218
	219	/*
	220	* reservation tracking
	221	*/
d4e5c920	222	struct btrfs_qgroup_rsv rsv;
3159fe7b QW	223
	224	/*
	225	* lists
	226	*/
	227	struct list_head groups; /* groups this group is member of */
	228	struct list_head members; /* groups that are members of this group */
	229	struct list_head dirty; /* dirty groups */
686c4a5a QW	230
	231	/*
	232	* For qgroup iteration usage.
	233	*
	234	* The iteration list should always be empty until qgroup_iterator_add()
	235	* is called. And should be reset to empty after the iteration is
	236	* finished.
	237	*/
	238	struct list_head iterator;
dce28769 QW	239
	240	/*
	241	* For nested iterator usage.
	242	*
	243	* Here we support at most one level of nested iterator calls like:
	244	*
	245	* LIST_HEAD(all_qgroups);
	246	* {
	247	* LIST_HEAD(local_qgroups);
	248	* qgroup_iterator_add(local_qgroups, qg);
	249	* qgroup_iterator_nested_add(all_qgroups, qg);
	250	* do_some_work(local_qgroups);
	251	* qgroup_iterator_clean(local_qgroups);
	252	* }
	253	* do_some_work(all_qgroups);
	254	* qgroup_iterator_nested_clean(all_qgroups);
	255	*/
	256	struct list_head nested_iterator;
3159fe7b QW	257	struct rb_node node; /* tree of qgroups */
	258
	259	/*
	260	* temp variables for accounting operations
	261	* Refer to qgroup_shared_accounting() for details.
	262	*/
	263	u64 old_refcnt;
	264	u64 new_refcnt;
49e5fb46 QW	265
	266	/*
	267	* Sysfs kobjectid
	268	*/
	269	struct kobject kobj;
3159fe7b QW	270	};
3159fe7b QW	271
1e0e9d57 BB	272	struct btrfs_squota_delta {
	273	/* The fstree root this delta counts against. */
	274	u64 root;
	275	/* The number of bytes in the extent being counted. */
	276	u64 num_bytes;
bd7c1ea3 BB	277	/* The generation the extent was created in. */
bd7c1ea3 BB	278	u64 generation;
1e0e9d57 BB	279	/* Whether we are using or freeing the extent. */
	280	bool is_inc;
	281	/* Whether the extent is data or metadata. */
	282	bool is_data;
	283	};
	284
49e5fb46 QW	285	static inline u64 btrfs_qgroup_subvolid(u64 qgroupid)
	286	{
	287	return (qgroupid & ((1ULL << BTRFS_QGROUP_LEVEL_SHIFT) - 1));
	288	}
	289
81fb6f77 QW	290	/*
	291	* For qgroup event trace points only
	292	*/
e0a8b9a7 DS	293	enum {
	294	ENUM_BIT(QGROUP_RESERVE),
	295	ENUM_BIT(QGROUP_RELEASE),
	296	ENUM_BIT(QGROUP_FREE),
	297	};
81fb6f77	298
6b0cd63b BB	299	enum btrfs_qgroup_mode {
	300	BTRFS_QGROUP_MODE_DISABLED,
	301	BTRFS_QGROUP_MODE_FULL,
182940f4	302	BTRFS_QGROUP_MODE_SIMPLE
6b0cd63b BB	303	};
	304
	305	enum btrfs_qgroup_mode btrfs_qgroup_mode(struct btrfs_fs_info *fs_info);
182940f4 BB	306	bool btrfs_qgroup_enabled(struct btrfs_fs_info *fs_info);
	307	bool btrfs_qgroup_full_accounting(struct btrfs_fs_info *fs_info);
	308	int btrfs_quota_enable(struct btrfs_fs_info *fs_info,
	309	struct btrfs_ioctl_quota_ctl_args *quota_ctl_args);
340f1aa2	310	int btrfs_quota_disable(struct btrfs_fs_info *fs_info);
fcebe456 JB	311	int btrfs_qgroup_rescan(struct btrfs_fs_info *fs_info);
fcebe456 JB	312	void btrfs_qgroup_rescan_resume(struct btrfs_fs_info *fs_info);
d06f23d6 JM	313	int btrfs_qgroup_wait_for_completion(struct btrfs_fs_info *fs_info,
d06f23d6 JM	314	bool interruptible);
5343cd93	315	int btrfs_add_qgroup_relation(struct btrfs_trans_handle *trans, u64 src, u64 dst);
39616c27 LF	316	int btrfs_del_qgroup_relation(struct btrfs_trans_handle *trans, u64 src,
39616c27 LF	317	u64 dst);
49a05ecd	318	int btrfs_create_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid);
3efbee1d	319	int btrfs_remove_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid);
f0042d5e	320	int btrfs_limit_qgroup(struct btrfs_trans_handle *trans, u64 qgroupid,
fcebe456 JB	321	struct btrfs_qgroup_limit *limit);
	322	int btrfs_read_qgroup_config(struct btrfs_fs_info *fs_info);
	323	void btrfs_free_qgroup_config(struct btrfs_fs_info *fs_info);
	324	struct btrfs_delayed_extent_op;
d1b8b94a	325
50b3e040	326	int btrfs_qgroup_trace_extent_nolock(
cb93b52c QW	327	struct btrfs_fs_info *fs_info,
	328	struct btrfs_delayed_ref_root *delayed_refs,
	329	struct btrfs_qgroup_extent_record *record);
8949b9a1	330	int btrfs_qgroup_trace_extent_post(struct btrfs_trans_handle *trans,
fb235dc0	331	struct btrfs_qgroup_extent_record *qrecord);
a95f3aaf	332	int btrfs_qgroup_trace_extent(struct btrfs_trans_handle *trans, u64 bytenr,
e2896e79	333	u64 num_bytes);
33d1f05c	334	int btrfs_qgroup_trace_leaf_items(struct btrfs_trans_handle *trans,
33d1f05c	335	struct extent_buffer *eb);
33d1f05c	336	int btrfs_qgroup_trace_subtree(struct btrfs_trans_handle *trans,
33d1f05c QW	337	struct extent_buffer *root_eb,
33d1f05c QW	338	u64 root_gen, int root_level);
8696d760 LF	339	int btrfs_qgroup_account_extent(struct btrfs_trans_handle *trans, u64 bytenr,
	340	u64 num_bytes, struct ulist *old_roots,
	341	struct ulist *new_roots);
460fb20a	342	int btrfs_qgroup_account_extents(struct btrfs_trans_handle *trans);
280f8bd2	343	int btrfs_run_qgroups(struct btrfs_trans_handle *trans);
a9377422	344	int btrfs_qgroup_inherit(struct btrfs_trans_handle *trans, u64 srcid,
5343cd93 BB	345	u64 objectid, u64 inode_rootid,
5343cd93 BB	346	struct btrfs_qgroup_inherit *inherit);
297d750b	347	void btrfs_qgroup_free_refroot(struct btrfs_fs_info *fs_info,
d4e5c920 QW	348	u64 ref_root, u64 num_bytes,
d4e5c920 QW	349	enum btrfs_qgroup_rsv_type type);
fcebe456 JB	350
	351	#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
	352	int btrfs_verify_qgroup_counts(struct btrfs_fs_info *fs_info, u64 qgroupid,
	353	u64 rfer, u64 excl);
	354	#endif
	355
52472553	356	/* New io_tree based accurate qgroup reserve API */
7661a3e0	357	int btrfs_qgroup_reserve_data(struct btrfs_inode *inode,
364ecf36	358	struct extent_changeset **reserved, u64 start, u64 len);
9e65bfca	359	int btrfs_qgroup_release_data(struct btrfs_inode inode, u64 start, u64 len, u64 released);
8b8a979f NB	360	int btrfs_qgroup_free_data(struct btrfs_inode *inode,
8b8a979f NB	361	struct extent_changeset *reserved, u64 start,
9e65bfca	362	u64 len, u64 *freed);
80e9baed NB	363	int btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
80e9baed NB	364	enum btrfs_qgroup_rsv_type type, bool enforce);
733e03a0	365	int __btrfs_qgroup_reserve_meta(struct btrfs_root *root, int num_bytes,
d4135134 FM	366	enum btrfs_qgroup_rsv_type type, bool enforce,
d4135134 FM	367	bool noflush);
733e03a0 QW	368	/* Reserve metadata space for pertrans and prealloc type */
	369	static inline int btrfs_qgroup_reserve_meta_pertrans(struct btrfs_root *root,
	370	int num_bytes, bool enforce)
	371	{
	372	return __btrfs_qgroup_reserve_meta(root, num_bytes,
d4135134 FM	373	BTRFS_QGROUP_RSV_META_PERTRANS,
d4135134 FM	374	enforce, false);
733e03a0 QW	375	}
733e03a0 QW	376	static inline int btrfs_qgroup_reserve_meta_prealloc(struct btrfs_root *root,
d4135134 FM	377	int num_bytes, bool enforce,
d4135134 FM	378	bool noflush)
733e03a0 QW	379	{
733e03a0 QW	380	return __btrfs_qgroup_reserve_meta(root, num_bytes,
d4135134 FM	381	BTRFS_QGROUP_RSV_META_PREALLOC,
d4135134 FM	382	enforce, noflush);
733e03a0 QW	383	}
	384
	385	void __btrfs_qgroup_free_meta(struct btrfs_root *root, int num_bytes,
	386	enum btrfs_qgroup_rsv_type type);
	387
	388	/* Free per-transaction meta reservation for error handling */
	389	static inline void btrfs_qgroup_free_meta_pertrans(struct btrfs_root *root,
	390	int num_bytes)
	391	{
	392	__btrfs_qgroup_free_meta(root, num_bytes,
	393	BTRFS_QGROUP_RSV_META_PERTRANS);
	394	}
	395
	396	/* Pre-allocated meta reservation can be freed at need */
	397	static inline void btrfs_qgroup_free_meta_prealloc(struct btrfs_root *root,
	398	int num_bytes)
	399	{
	400	__btrfs_qgroup_free_meta(root, num_bytes,
	401	BTRFS_QGROUP_RSV_META_PREALLOC);
	402	}
	403
733e03a0	404	void btrfs_qgroup_free_meta_all_pertrans(struct btrfs_root *root);
64cfaef6	405	void btrfs_qgroup_convert_reserved_meta(struct btrfs_root *root, int num_bytes);
cfdd4592	406	void btrfs_qgroup_check_reserved_leak(struct btrfs_inode *inode);
9888c340	407
370a11b8 QW	408	/* btrfs_qgroup_swapped_blocks related functions */
	409	void btrfs_qgroup_init_swapped_blocks(
	410	struct btrfs_qgroup_swapped_blocks *swapped_blocks);
	411
	412	void btrfs_qgroup_clean_swapped_blocks(struct btrfs_root *root);
	413	int btrfs_qgroup_add_swapped_blocks(struct btrfs_trans_handle *trans,
	414	struct btrfs_root *subvol_root,
32da5386	415	struct btrfs_block_group *bg,
370a11b8 QW	416	struct extent_buffer *subvol_parent, int subvol_slot,
	417	struct extent_buffer *reloc_parent, int reloc_slot,
	418	u64 last_snapshot);
f616f5cd QW	419	int btrfs_qgroup_trace_subtree_after_cow(struct btrfs_trans_handle *trans,
f616f5cd QW	420	struct btrfs_root root, struct extent_buffer eb);
81f7eb00	421	void btrfs_qgroup_destroy_extent_records(struct btrfs_transaction *trans);
5958253c	422	bool btrfs_check_quota_leak(struct btrfs_fs_info *fs_info);
e85a0ada	423	void btrfs_free_squota_rsv(struct btrfs_fs_info *fs_info, u64 root, u64 rsv_bytes);
1e0e9d57 BB	424	int btrfs_record_squota_delta(struct btrfs_fs_info *fs_info,
1e0e9d57 BB	425	struct btrfs_squota_delta *delta);
370a11b8	426
9888c340	427	#endif