[mirror_zfs.git] / include / sys / zap_impl.h

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */

/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
 * Copyright (c) 2013, 2016 by Delphix. All rights reserved.
 * Copyright 2017 Nexenta Systems, Inc.
 */

#ifndef	_SYS_ZAP_IMPL_H
#define	_SYS_ZAP_IMPL_H

#include <sys/zap.h>
#include <sys/zfs_context.h>
#include <sys/avl.h>

#ifdef	__cplusplus
extern "C" {
#endif

extern int fzap_default_block_shift;

#define	ZAP_MAGIC 0x2F52AB2ABULL

#define	FZAP_BLOCK_SHIFT(zap)	((zap)->zap_f.zap_block_shift)

#define	MZAP_ENT_LEN		64
#define	MZAP_NAME_LEN		(MZAP_ENT_LEN - 8 - 4 - 2)
#define	MZAP_MAX_BLKSZ		SPA_OLD_MAXBLOCKSIZE

#define	ZAP_NEED_CD		(-1U)

typedef struct mzap_ent_phys {
	uint64_t mze_value;
	uint32_t mze_cd;
	uint16_t mze_pad;	/* in case we want to chain them someday */
	char mze_name[MZAP_NAME_LEN];
} mzap_ent_phys_t;

typedef struct mzap_phys {
	uint64_t mz_block_type;	/* ZBT_MICRO */
	uint64_t mz_salt;
	uint64_t mz_normflags;
	uint64_t mz_pad[5];
	mzap_ent_phys_t mz_chunk[1];
	/* actually variable size depending on block size */
} mzap_phys_t;

typedef struct mzap_ent {
	avl_node_t mze_node;
	int mze_chunkid;
	uint64_t mze_hash;
	uint32_t mze_cd; /* copy from mze_phys->mze_cd */
} mzap_ent_t;

#define	MZE_PHYS(zap, mze) \
	(&zap_m_phys(zap)->mz_chunk[(mze)->mze_chunkid])

/*
 * The (fat) zap is stored in one object. It is an array of
 * 1<<FZAP_BLOCK_SHIFT byte blocks. The layout looks like one of:
 *
 * ptrtbl fits in first block:
 * 	[zap_phys_t zap_ptrtbl_shift < 6] [zap_leaf_t] ...
 *
 * ptrtbl too big for first block:
 * 	[zap_phys_t zap_ptrtbl_shift >= 6] [zap_leaf_t] [ptrtbl] ...
 *
 */

struct dmu_buf;
struct zap_leaf;

#define	ZBT_LEAF		((1ULL << 63) + 0)
#define	ZBT_HEADER		((1ULL << 63) + 1)
#define	ZBT_MICRO		((1ULL << 63) + 3)
/* any other values are ptrtbl blocks */

/*
 * the embedded pointer table takes up half a block:
 * block size / entry size (2^3) / 2
 */
#define	ZAP_EMBEDDED_PTRTBL_SHIFT(zap) (FZAP_BLOCK_SHIFT(zap) - 3 - 1)

/*
 * The embedded pointer table starts half-way through the block.  Since
 * the pointer table itself is half the block, it starts at (64-bit)
 * word number (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap)).
 */
#define	ZAP_EMBEDDED_PTRTBL_ENT(zap, idx) \
	((uint64_t *)zap_f_phys(zap)) \
	[(idx) + (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap))]

/*
 * TAKE NOTE:
 * If zap_phys_t is modified, zap_byteswap() must be modified.
 */
typedef struct zap_phys {
	uint64_t zap_block_type;	/* ZBT_HEADER */
	uint64_t zap_magic;		/* ZAP_MAGIC */

	struct zap_table_phys {
		uint64_t zt_blk;	/* starting block number */
		uint64_t zt_numblks;	/* number of blocks */
		uint64_t zt_shift;	/* bits to index it */
		uint64_t zt_nextblk;	/* next (larger) copy start block */
		uint64_t zt_blks_copied; /* number source blocks copied */
	} zap_ptrtbl;

	uint64_t zap_freeblk;		/* the next free block */
	uint64_t zap_num_leafs;		/* number of leafs */
	uint64_t zap_num_entries;	/* number of entries */
	uint64_t zap_salt;		/* salt to stir into hash function */
	uint64_t zap_normflags;		/* flags for u8_textprep_str() */
	uint64_t zap_flags;		/* zap_flags_t */
	/*
	 * This structure is followed by padding, and then the embedded
	 * pointer table.  The embedded pointer table takes up second
	 * half of the block.  It is accessed using the
	 * ZAP_EMBEDDED_PTRTBL_ENT() macro.
	 */
} zap_phys_t;

typedef struct zap_table_phys zap_table_phys_t;

typedef struct zap {
	dmu_buf_user_t zap_dbu;
	objset_t *zap_objset;
	uint64_t zap_object;
	struct dmu_buf *zap_dbuf;
	krwlock_t zap_rwlock;
	boolean_t zap_ismicro;
	int zap_normflags;
	uint64_t zap_salt;
	union {
		struct {
			/*
			 * zap_num_entries_mtx protects
			 * zap_num_entries
			 */
			kmutex_t zap_num_entries_mtx;
			int zap_block_shift;
		} zap_fat;
		struct {
			int16_t zap_num_entries;
			int16_t zap_num_chunks;
			int16_t zap_alloc_next;
			avl_tree_t zap_avl;
		} zap_micro;
	} zap_u;
} zap_t;

static inline zap_phys_t *
zap_f_phys(zap_t *zap)
{
	return (zap->zap_dbuf->db_data);
}

static inline mzap_phys_t *
zap_m_phys(zap_t *zap)
{
	return (zap->zap_dbuf->db_data);
}

typedef struct zap_name {
	zap_t *zn_zap;
	int zn_key_intlen;
	const void *zn_key_orig;
	int zn_key_orig_numints;
	const void *zn_key_norm;
	int zn_key_norm_numints;
	uint64_t zn_hash;
	matchtype_t zn_matchtype;
	int zn_normflags;
	char zn_normbuf[ZAP_MAXNAMELEN];
} zap_name_t;

#define	zap_f	zap_u.zap_fat
#define	zap_m	zap_u.zap_micro

boolean_t zap_match(zap_name_t *zn, const char *matchname);
int zap_lockdir(objset_t *os, uint64_t obj, dmu_tx_t *tx,
    krw_t lti, boolean_t fatreader, boolean_t adding, void *tag, zap_t **zapp);
void zap_unlockdir(zap_t *zap, void *tag);
void zap_evict_sync(void *dbu);
zap_name_t *zap_name_alloc(zap_t *zap, const char *key, matchtype_t mt);
void zap_name_free(zap_name_t *zn);
int zap_hashbits(zap_t *zap);
uint32_t zap_maxcd(zap_t *zap);
uint64_t zap_getflags(zap_t *zap);

#define	ZAP_HASH_IDX(hash, n) (((n) == 0) ? 0 : ((hash) >> (64 - (n))))

void fzap_byteswap(void *buf, size_t size);
int fzap_count(zap_t *zap, uint64_t *count);
int fzap_lookup(zap_name_t *zn,
    uint64_t integer_size, uint64_t num_integers, void *buf,
    char *realname, int rn_len, boolean_t *normalization_conflictp);
void fzap_prefetch(zap_name_t *zn);
int fzap_add(zap_name_t *zn, uint64_t integer_size, uint64_t num_integers,
    const void *val, void *tag, dmu_tx_t *tx);
int fzap_update(zap_name_t *zn,
    int integer_size, uint64_t num_integers, const void *val,
    void *tag, dmu_tx_t *tx);
int fzap_length(zap_name_t *zn,
    uint64_t *integer_size, uint64_t *num_integers);
int fzap_remove(zap_name_t *zn, dmu_tx_t *tx);
int fzap_cursor_retrieve(zap_t *zap, zap_cursor_t *zc, zap_attribute_t *za);
void fzap_get_stats(zap_t *zap, zap_stats_t *zs);
void zap_put_leaf(struct zap_leaf *l);

int fzap_add_cd(zap_name_t *zn,
    uint64_t integer_size, uint64_t num_integers,
    const void *val, uint32_t cd, void *tag, dmu_tx_t *tx);
void fzap_upgrade(zap_t *zap, dmu_tx_t *tx, zap_flags_t flags);

#ifdef	__cplusplus
}
#endif

#endif /* _SYS_ZAP_IMPL_H */
Commit	Line	Data
34dc7c2f BB	1	/*
	2	* CDDL HEADER START
	3	*
	4	* The contents of this file are subject to the terms of the
	5	* Common Development and Distribution License (the "License").
	6	* You may not use this file except in compliance with the License.
	7	*
	8	* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
	9	* or http://www.opensolaris.org/os/licensing.
	10	* See the License for the specific language governing permissions
	11	* and limitations under the License.
	12	*
	13	* When distributing Covered Code, include this CDDL HEADER in each
	14	* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
	15	* If applicable, add the following below this CDDL HEADER, with the
	16	* fields enclosed by brackets "[]" replaced with your own identifying
	17	* information: Portions Copyright [yyyy] [name of copyright owner]
	18	*
	19	* CDDL HEADER END
	20	*/
9b7b9cd3	21
34dc7c2f	22	/*
428870ff	23	* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
0c66c32d	24	* Copyright (c) 2014 Spectra Logic Corporation, All rights reserved.
8bea9815	25	* Copyright (c) 2013, 2016 by Delphix. All rights reserved.
9b7b9cd3	26	* Copyright 2017 Nexenta Systems, Inc.
34dc7c2f BB	27	*/
	28
	29	#ifndef _SYS_ZAP_IMPL_H
	30	#define _SYS_ZAP_IMPL_H
	31
34dc7c2f BB	32	#include <sys/zap.h>
	33	#include <sys/zfs_context.h>
	34	#include <sys/avl.h>
	35
	36	#ifdef __cplusplus
	37	extern "C" {
	38	#endif
	39
	40	extern int fzap_default_block_shift;
	41
	42	#define ZAP_MAGIC 0x2F52AB2ABULL
	43
	44	#define FZAP_BLOCK_SHIFT(zap) ((zap)->zap_f.zap_block_shift)
	45
34dc7c2f BB	46	#define MZAP_ENT_LEN 64
34dc7c2f BB	47	#define MZAP_NAME_LEN (MZAP_ENT_LEN - 8 - 4 - 2)
f1512ee6	48	#define MZAP_MAX_BLKSZ SPA_OLD_MAXBLOCKSIZE
34dc7c2f	49
428870ff BB	50	#define ZAP_NEED_CD (-1U)
428870ff BB	51
34dc7c2f BB	52	typedef struct mzap_ent_phys {
	53	uint64_t mze_value;
	54	uint32_t mze_cd;
	55	uint16_t mze_pad; /* in case we want to chain them someday */
	56	char mze_name[MZAP_NAME_LEN];
	57	} mzap_ent_phys_t;
	58
	59	typedef struct mzap_phys {
	60	uint64_t mz_block_type; /* ZBT_MICRO */
	61	uint64_t mz_salt;
	62	uint64_t mz_normflags;
	63	uint64_t mz_pad[5];
	64	mzap_ent_phys_t mz_chunk[1];
	65	/* actually variable size depending on block size */
	66	} mzap_phys_t;
	67
	68	typedef struct mzap_ent {
	69	avl_node_t mze_node;
	70	int mze_chunkid;
	71	uint64_t mze_hash;
428870ff	72	uint32_t mze_cd; /* copy from mze_phys->mze_cd */
34dc7c2f BB	73	} mzap_ent_t;
34dc7c2f BB	74
428870ff	75	#define MZE_PHYS(zap, mze) \
d683ddbb	76	(&zap_m_phys(zap)->mz_chunk[(mze)->mze_chunkid])
34dc7c2f BB	77
	78	/*
	79	* The (fat) zap is stored in one object. It is an array of
	80	* 1<<FZAP_BLOCK_SHIFT byte blocks. The layout looks like one of:
	81	*
	82	* ptrtbl fits in first block:
	83	* [zap_phys_t zap_ptrtbl_shift < 6] [zap_leaf_t] ...
	84	*
	85	* ptrtbl too big for first block:
	86	* [zap_phys_t zap_ptrtbl_shift >= 6] [zap_leaf_t] [ptrtbl] ...
	87	*
	88	*/
	89
	90	struct dmu_buf;
	91	struct zap_leaf;
	92
	93	#define ZBT_LEAF ((1ULL << 63) + 0)
	94	#define ZBT_HEADER ((1ULL << 63) + 1)
	95	#define ZBT_MICRO ((1ULL << 63) + 3)
	96	/* any other values are ptrtbl blocks */
	97
	98	/*
	99	* the embedded pointer table takes up half a block:
	100	* block size / entry size (2^3) / 2
	101	*/
	102	#define ZAP_EMBEDDED_PTRTBL_SHIFT(zap) (FZAP_BLOCK_SHIFT(zap) - 3 - 1)
	103
	104	/*
	105	* The embedded pointer table starts half-way through the block. Since
	106	* the pointer table itself is half the block, it starts at (64-bit)
	107	* word number (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap)).
	108	*/
	109	#define ZAP_EMBEDDED_PTRTBL_ENT(zap, idx) \
d683ddbb	110	((uint64_t *)zap_f_phys(zap)) \
34dc7c2f BB	111	[(idx) + (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap))]
	112
	113	/*
	114	* TAKE NOTE:
	115	* If zap_phys_t is modified, zap_byteswap() must be modified.
	116	*/
	117	typedef struct zap_phys {
	118	uint64_t zap_block_type; /* ZBT_HEADER */
	119	uint64_t zap_magic; /* ZAP_MAGIC */
	120
	121	struct zap_table_phys {
	122	uint64_t zt_blk; /* starting block number */
	123	uint64_t zt_numblks; /* number of blocks */
	124	uint64_t zt_shift; /* bits to index it */
	125	uint64_t zt_nextblk; /* next (larger) copy start block */
	126	uint64_t zt_blks_copied; /* number source blocks copied */
	127	} zap_ptrtbl;
	128
	129	uint64_t zap_freeblk; /* the next free block */
	130	uint64_t zap_num_leafs; /* number of leafs */
	131	uint64_t zap_num_entries; /* number of entries */
	132	uint64_t zap_salt; /* salt to stir into hash function */
	133	uint64_t zap_normflags; /* flags for u8_textprep_str() */
428870ff	134	uint64_t zap_flags; /* zap_flags_t */
34dc7c2f BB	135	/*
	136	* This structure is followed by padding, and then the embedded
	137	* pointer table. The embedded pointer table takes up second
	138	* half of the block. It is accessed using the
	139	* ZAP_EMBEDDED_PTRTBL_ENT() macro.
	140	*/
	141	} zap_phys_t;
	142
	143	typedef struct zap_table_phys zap_table_phys_t;
	144
	145	typedef struct zap {
0c66c32d	146	dmu_buf_user_t zap_dbu;
34dc7c2f BB	147	objset_t *zap_objset;
	148	uint64_t zap_object;
	149	struct dmu_buf *zap_dbuf;
	150	krwlock_t zap_rwlock;
	151	boolean_t zap_ismicro;
	152	int zap_normflags;
	153	uint64_t zap_salt;
	154	union {
	155	struct {
34dc7c2f BB	156	/*
	157	* zap_num_entries_mtx protects
	158	* zap_num_entries
	159	*/
	160	kmutex_t zap_num_entries_mtx;
	161	int zap_block_shift;
	162	} zap_fat;
	163	struct {
34dc7c2f BB	164	int16_t zap_num_entries;
	165	int16_t zap_num_chunks;
	166	int16_t zap_alloc_next;
	167	avl_tree_t zap_avl;
	168	} zap_micro;
	169	} zap_u;
	170	} zap_t;
	171
d683ddbb JG	172	static inline zap_phys_t *
	173	zap_f_phys(zap_t *zap)
	174	{
	175	return (zap->zap_dbuf->db_data);
	176	}
	177
	178	static inline mzap_phys_t *
	179	zap_m_phys(zap_t *zap)
	180	{
	181	return (zap->zap_dbuf->db_data);
	182	}
	183
34dc7c2f BB	184	typedef struct zap_name {
34dc7c2f BB	185	zap_t *zn_zap;
428870ff BB	186	int zn_key_intlen;
	187	const void *zn_key_orig;
	188	int zn_key_orig_numints;
	189	const void *zn_key_norm;
	190	int zn_key_norm_numints;
34dc7c2f BB	191	uint64_t zn_hash;
34dc7c2f BB	192	matchtype_t zn_matchtype;
9b7b9cd3	193	int zn_normflags;
34dc7c2f BB	194	char zn_normbuf[ZAP_MAXNAMELEN];
	195	} zap_name_t;
	196
	197	#define zap_f zap_u.zap_fat
	198	#define zap_m zap_u.zap_micro
	199
	200	boolean_t zap_match(zap_name_t zn, const char matchname);
	201	int zap_lockdir(objset_t os, uint64_t obj, dmu_tx_t tx,
8bea9815 MA	202	krw_t lti, boolean_t fatreader, boolean_t adding, void tag, zap_t *zapp);
8bea9815 MA	203	void zap_unlockdir(zap_t zap, void tag);
39efbde7	204	void zap_evict_sync(void *dbu);
428870ff	205	zap_name_t zap_name_alloc(zap_t zap, const char *key, matchtype_t mt);
34dc7c2f	206	void zap_name_free(zap_name_t *zn);
428870ff BB	207	int zap_hashbits(zap_t *zap);
	208	uint32_t zap_maxcd(zap_t *zap);
	209	uint64_t zap_getflags(zap_t *zap);
34dc7c2f BB	210
	211	#define ZAP_HASH_IDX(hash, n) (((n) == 0) ? 0 : ((hash) >> (64 - (n))))
	212
	213	void fzap_byteswap(void *buf, size_t size);
	214	int fzap_count(zap_t zap, uint64_t count);
	215	int fzap_lookup(zap_name_t *zn,
	216	uint64_t integer_size, uint64_t num_integers, void *buf,
	217	char realname, int rn_len, boolean_t normalization_conflictp);
428870ff	218	void fzap_prefetch(zap_name_t *zn);
34dc7c2f	219	int fzap_add(zap_name_t *zn, uint64_t integer_size, uint64_t num_integers,
8bea9815	220	const void val, void tag, dmu_tx_t *tx);
34dc7c2f	221	int fzap_update(zap_name_t *zn,
8bea9815 MA	222	int integer_size, uint64_t num_integers, const void *val,
8bea9815 MA	223	void tag, dmu_tx_t tx);
34dc7c2f BB	224	int fzap_length(zap_name_t *zn,
	225	uint64_t integer_size, uint64_t num_integers);
	226	int fzap_remove(zap_name_t zn, dmu_tx_t tx);
	227	int fzap_cursor_retrieve(zap_t zap, zap_cursor_t zc, zap_attribute_t *za);
	228	void fzap_get_stats(zap_t zap, zap_stats_t zs);
	229	void zap_put_leaf(struct zap_leaf *l);
	230
	231	int fzap_add_cd(zap_name_t *zn,
	232	uint64_t integer_size, uint64_t num_integers,
8bea9815	233	const void val, uint32_t cd, void tag, dmu_tx_t *tx);
428870ff	234	void fzap_upgrade(zap_t zap, dmu_tx_t tx, zap_flags_t flags);
34dc7c2f BB	235
	236	#ifdef __cplusplus
	237	}
	238	#endif
	239
	240	#endif /* _SYS_ZAP_IMPL_H */