[deliverable/linux.git] / include / linux / crush / crush.h

#ifndef CEPH_CRUSH_CRUSH_H
#define CEPH_CRUSH_CRUSH_H

#include <linux/types.h>

/*
 * CRUSH is a pseudo-random data distribution algorithm that
 * efficiently distributes input values (typically, data objects)
 * across a heterogeneous, structured storage cluster.
 *
 * The algorithm was originally described in detail in this paper
 * (although the algorithm has evolved somewhat since then):
 *
 *     http://www.ssrc.ucsc.edu/Papers/weil-sc06.pdf
 *
 * LGPL2
 */


#define CRUSH_MAGIC 0x00010000ul   /* for detecting algorithm revisions */


#define CRUSH_MAX_DEPTH 10  /* max crush hierarchy depth */
#define CRUSH_MAX_SET   10  /* max size of a mapping result */


/*
 * CRUSH uses user-defined "rules" to describe how inputs should be
 * mapped to devices.  A rule consists of sequence of steps to perform
 * to generate the set of output devices.
 */
struct crush_rule_step {
	__u32 op;
	__s32 arg1;
	__s32 arg2;
};

/* step op codes */
enum {
	CRUSH_RULE_NOOP = 0,
	CRUSH_RULE_TAKE = 1,          /* arg1 = value to start with */
	CRUSH_RULE_CHOOSE_FIRSTN = 2, /* arg1 = num items to pick */
				      /* arg2 = type */
	CRUSH_RULE_CHOOSE_INDEP = 3,  /* same */
	CRUSH_RULE_EMIT = 4,          /* no args */
	CRUSH_RULE_CHOOSE_LEAF_FIRSTN = 6,
	CRUSH_RULE_CHOOSE_LEAF_INDEP = 7,
};

/*
 * for specifying choose num (arg1) relative to the max parameter
 * passed to do_rule
 */
#define CRUSH_CHOOSE_N            0
#define CRUSH_CHOOSE_N_MINUS(x)   (-(x))

/*
 * The rule mask is used to describe what the rule is intended for.
 * Given a ruleset and size of output set, we search through the
 * rule list for a matching rule_mask.
 */
struct crush_rule_mask {
	__u8 ruleset;
	__u8 type;
	__u8 min_size;
	__u8 max_size;
};

struct crush_rule {
	__u32 len;
	struct crush_rule_mask mask;
	struct crush_rule_step steps[0];
};

#define crush_rule_size(len) (sizeof(struct crush_rule) + \
			      (len)*sizeof(struct crush_rule_step))


/*
 * A bucket is a named container of other items (either devices or
 * other buckets).  Items within a bucket are chosen using one of a
 * few different algorithms.  The table summarizes how the speed of
 * each option measures up against mapping stability when items are
 * added or removed.
 *
 *  Bucket Alg     Speed       Additions    Removals
 *  ------------------------------------------------
 *  uniform         O(1)       poor         poor
 *  list            O(n)       optimal      poor
 *  tree            O(log n)   good         good
 *  straw           O(n)       optimal      optimal
 */
enum {
	CRUSH_BUCKET_UNIFORM = 1,
	CRUSH_BUCKET_LIST = 2,
	CRUSH_BUCKET_TREE = 3,
	CRUSH_BUCKET_STRAW = 4
};
extern const char *crush_bucket_alg_name(int alg);

struct crush_bucket {
	__s32 id;        /* this'll be negative */
	__u16 type;      /* non-zero; type=0 is reserved for devices */
	__u8 alg;        /* one of CRUSH_BUCKET_* */
	__u8 hash;       /* which hash function to use, CRUSH_HASH_* */
	__u32 weight;    /* 16-bit fixed point */
	__u32 size;      /* num items */
	__s32 *items;

	/*
	 * cached random permutation: used for uniform bucket and for
	 * the linear search fallback for the other bucket types.
	 */
	__u32 perm_x;  /* @x for which *perm is defined */
	__u32 perm_n;  /* num elements of *perm that are permuted/defined */
	__u32 *perm;
};

struct crush_bucket_uniform {
	struct crush_bucket h;
	__u32 item_weight;  /* 16-bit fixed point; all items equally weighted */
};

struct crush_bucket_list {
	struct crush_bucket h;
	__u32 *item_weights;  /* 16-bit fixed point */
	__u32 *sum_weights;   /* 16-bit fixed point.  element i is sum
				 of weights 0..i, inclusive */
};

struct crush_bucket_tree {
	struct crush_bucket h;  /* note: h.size is _tree_ size, not number of
				   actual items */
	__u8 num_nodes;
	__u32 *node_weights;
};

struct crush_bucket_straw {
	struct crush_bucket h;
	__u32 *item_weights;   /* 16-bit fixed point */
	__u32 *straws;         /* 16-bit fixed point */
};


/*
 * CRUSH map includes all buckets, rules, etc.
 */
struct crush_map {
	struct crush_bucket **buckets;
	struct crush_rule **rules;

	__s32 max_buckets;
	__u32 max_rules;
	__s32 max_devices;

	/* choose local retries before re-descent */
	__u32 choose_local_tries;
	/* choose local attempts using a fallback permutation before
	 * re-descent */
	__u32 choose_local_fallback_tries;
	/* choose attempts before giving up */ 
	__u32 choose_total_tries;
	/* attempt chooseleaf inner descent once; on failure retry outer descent */
	__u32 chooseleaf_descend_once;
};


/* crush.c */
extern int crush_get_bucket_item_weight(const struct crush_bucket *b, int pos);
extern void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b);
extern void crush_destroy_bucket_list(struct crush_bucket_list *b);
extern void crush_destroy_bucket_tree(struct crush_bucket_tree *b);
extern void crush_destroy_bucket_straw(struct crush_bucket_straw *b);
extern void crush_destroy_bucket(struct crush_bucket *b);
extern void crush_destroy(struct crush_map *map);

static inline int crush_calc_tree_node(int i)
{
	return ((i+1) << 1)-1;
}

#endif
Commit	Line	Data
5cd068c2 SW	1	#ifndef CEPH_CRUSH_CRUSH_H
5cd068c2 SW	2	#define CEPH_CRUSH_CRUSH_H
5ecc0a0f SW	3
	4	#include <linux/types.h>
	5
	6	/*
	7	* CRUSH is a pseudo-random data distribution algorithm that
	8	* efficiently distributes input values (typically, data objects)
	9	* across a heterogeneous, structured storage cluster.
	10	*
	11	* The algorithm was originally described in detail in this paper
	12	* (although the algorithm has evolved somewhat since then):
	13	*
	14	* http://www.ssrc.ucsc.edu/Papers/weil-sc06.pdf
	15	*
	16	* LGPL2
	17	*/
	18
	19
	20	#define CRUSH_MAGIC 0x00010000ul /* for detecting algorithm revisions */
	21
	22
	23	#define CRUSH_MAX_DEPTH 10 /* max crush hierarchy depth */
	24	#define CRUSH_MAX_SET 10 /* max size of a mapping result */
	25
	26
	27	/*
	28	* CRUSH uses user-defined "rules" to describe how inputs should be
	29	* mapped to devices. A rule consists of sequence of steps to perform
	30	* to generate the set of output devices.
	31	*/
	32	struct crush_rule_step {
	33	__u32 op;
	34	__s32 arg1;
	35	__s32 arg2;
	36	};
	37
	38	/* step op codes */
	39	enum {
	40	CRUSH_RULE_NOOP = 0,
	41	CRUSH_RULE_TAKE = 1, /* arg1 = value to start with */
	42	CRUSH_RULE_CHOOSE_FIRSTN = 2, /* arg1 = num items to pick */
	43	/* arg2 = type */
	44	CRUSH_RULE_CHOOSE_INDEP = 3, /* same */
	45	CRUSH_RULE_EMIT = 4, /* no args */
	46	CRUSH_RULE_CHOOSE_LEAF_FIRSTN = 6,
	47	CRUSH_RULE_CHOOSE_LEAF_INDEP = 7,
	48	};
	49
	50	/*
	51	* for specifying choose num (arg1) relative to the max parameter
	52	* passed to do_rule
	53	*/
	54	#define CRUSH_CHOOSE_N 0
	55	#define CRUSH_CHOOSE_N_MINUS(x) (-(x))
	56
	57	/*
	58	* The rule mask is used to describe what the rule is intended for.
	59	* Given a ruleset and size of output set, we search through the
	60	* rule list for a matching rule_mask.
	61	*/
	62	struct crush_rule_mask {
	63	__u8 ruleset;
	64	__u8 type;
	65	__u8 min_size;
	66	__u8 max_size;
67	};
68
69	struct crush_rule {
70	__u32 len;
71	struct crush_rule_mask mask;
72	struct crush_rule_step steps[0];
73	};
74
75	#define crush_rule_size(len) (sizeof(struct crush_rule) + \
76	(len)*sizeof(struct crush_rule_step))
77
78
79
80	/*
81	* A bucket is a named container of other items (either devices or
82	* other buckets). Items within a bucket are chosen using one of a
83	* few different algorithms. The table summarizes how the speed of
84	* each option measures up against mapping stability when items are
85	* added or removed.
86	*
87	* Bucket Alg Speed Additions Removals
88	* ------------------------------------------------
89	* uniform O(1) poor poor
90	* list O(n) optimal poor
91	* tree O(log n) good good
92	* straw O(n) optimal optimal
93	*/
94	enum {
95	CRUSH_BUCKET_UNIFORM = 1,
96	CRUSH_BUCKET_LIST = 2,
97	CRUSH_BUCKET_TREE = 3,
98	CRUSH_BUCKET_STRAW = 4
99	};
c6cf7263	100	extern const char *crush_bucket_alg_name(int alg);
5ecc0a0f SW	101
	102	struct crush_bucket {
	103	__s32 id; /* this'll be negative */
	104	__u16 type; /* non-zero; type=0 is reserved for devices */
fb690390 SW	105	__u8 alg; /* one of CRUSH_BUCKET_* */
fb690390 SW	106	__u8 hash; /* which hash function to use, CRUSH_HASH_* */
5ecc0a0f SW	107	__u32 weight; /* 16-bit fixed point */
	108	__u32 size; /* num items */
	109	__s32 *items;
	110
	111	/*
	112	* cached random permutation: used for uniform bucket and for
	113	* the linear search fallback for the other bucket types.
	114	*/
	115	__u32 perm_x; /* @x for which perm is defined /
	116	__u32 perm_n; /* num elements of perm that are permuted/defined /
	117	__u32 *perm;
	118	};
	119
	120	struct crush_bucket_uniform {
	121	struct crush_bucket h;
	122	__u32 item_weight; /* 16-bit fixed point; all items equally weighted */
	123	};
	124
	125	struct crush_bucket_list {
	126	struct crush_bucket h;
	127	__u32 item_weights; / 16-bit fixed point */
	128	__u32 sum_weights; / 16-bit fixed point. element i is sum
	129	of weights 0..i, inclusive */
	130	};
	131
	132	struct crush_bucket_tree {
	133	struct crush_bucket h; /* note: h.size is _tree_ size, not number of
	134	actual items */
	135	__u8 num_nodes;
	136	__u32 *node_weights;
	137	};
	138
	139	struct crush_bucket_straw {
	140	struct crush_bucket h;
	141	__u32 item_weights; / 16-bit fixed point */
	142	__u32 straws; / 16-bit fixed point */
	143	};
	144
	145
	146
	147	/*
	148	* CRUSH map includes all buckets, rules, etc.
	149	*/
	150	struct crush_map {
	151	struct crush_bucket **buckets;
	152	struct crush_rule **rules;
	153
5ecc0a0f SW	154	__s32 max_buckets;
	155	__u32 max_rules;
	156	__s32 max_devices;
546f04ef SW	157
	158	/* choose local retries before re-descent */
	159	__u32 choose_local_tries;
	160	/* choose local attempts using a fallback permutation before
	161	* re-descent */
	162	__u32 choose_local_fallback_tries;
	163	/* choose attempts before giving up */
	164	__u32 choose_total_tries;
1604f488 JS	165	/* attempt chooseleaf inner descent once; on failure retry outer descent */
1604f488 JS	166	__u32 chooseleaf_descend_once;
5ecc0a0f SW	167	};
	168
	169
	170	/* crush.c */
8b12d47b	171	extern int crush_get_bucket_item_weight(const struct crush_bucket *b, int pos);
5ecc0a0f SW	172	extern void crush_destroy_bucket_uniform(struct crush_bucket_uniform *b);
	173	extern void crush_destroy_bucket_list(struct crush_bucket_list *b);
	174	extern void crush_destroy_bucket_tree(struct crush_bucket_tree *b);
	175	extern void crush_destroy_bucket_straw(struct crush_bucket_straw *b);
	176	extern void crush_destroy_bucket(struct crush_bucket *b);
	177	extern void crush_destroy(struct crush_map *map);
	178
f671d4cd SW	179	static inline int crush_calc_tree_node(int i)
	180	{
	181	return ((i+1) << 1)-1;
	182	}
	183
5ecc0a0f	184	#endif