1 #ifndef _FS_CEPH_OSDMAP_H
2 #define _FS_CEPH_OSDMAP_H
4 #include <linux/rbtree.h>
5 #include <linux/ceph/types.h>
6 #include <linux/ceph/decode.h>
7 #include <linux/ceph/ceph_fs.h>
8 #include <linux/crush/crush.h>
11 * The osd map describes the current membership of the osd cluster and
12 * specifies the mapping of objects to placement groups and placement
13 * groups to (sets of) osds. That is, it completely specifies the
14 * (desired) distribution of all data objects in the system at some
17 * Each map version is identified by an epoch, which increases monotonically.
19 * The map can be updated either via an incremental map (diff) describing
20 * the change between two successive epochs, or as a fully encoded map.
27 #define CEPH_POOL_FLAG_HASHPSPOOL 1
29 struct ceph_pg_pool_info {
37 int pg_num_mask, pgp_num_mask;
42 struct ceph_object_locator {
47 * Maximum supported by kernel client object name length
49 * (probably outdated: must be >= RBD_MAX_MD_NAME_LEN -- currently 100)
51 #define CEPH_MAX_OID_NAME_LEN 100
53 struct ceph_object_id {
54 char name[CEPH_MAX_OID_NAME_LEN];
58 struct ceph_pg_mapping {
66 struct ceph_fsid fsid;
69 struct ceph_timespec created, modified;
71 u32 flags; /* CEPH_OSDMAP_* */
73 u32 max_osd; /* size of osd_state, _offload, _addr arrays */
74 u8 *osd_state; /* CEPH_OSD_* */
75 u32 *osd_weight; /* 0 = failed, 0x10000 = 100% normal */
76 struct ceph_entity_addr *osd_addr;
78 struct rb_root pg_temp;
79 struct rb_root pg_pools;
82 /* the CRUSH map specifies the mapping of placement groups to
83 * the list of osds that store+replicate them. */
84 struct crush_map *crush;
87 static inline void ceph_oid_set_name(struct ceph_object_id *oid,
93 if (len > sizeof(oid->name)) {
94 WARN(1, "ceph_oid_set_name '%s' len %d vs %zu, truncating\n",
95 name, len, sizeof(oid->name));
96 len = sizeof(oid->name);
99 memcpy(oid->name, name, len);
103 static inline void ceph_oid_copy(struct ceph_object_id *dest,
104 struct ceph_object_id *src)
106 BUG_ON(src->name_len > sizeof(dest->name));
107 memcpy(dest->name, src->name, src->name_len);
108 dest->name_len = src->name_len;
111 static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd)
113 return (osd < map->max_osd) && (map->osd_state[osd] & CEPH_OSD_UP);
116 static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag)
118 return map && (map->flags & flag);
121 extern char *ceph_osdmap_state_str(char *str, int len, int state);
123 static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map,
126 if (osd >= map->max_osd)
128 return &map->osd_addr[osd];
131 static inline int ceph_decode_pgid(void **p, void *end, struct ceph_pg *pgid)
135 if (!ceph_has_room(p, end, 1 + 8 + 4 + 4)) {
136 pr_warning("incomplete pg encoding");
140 version = ceph_decode_8(p);
142 pr_warning("do not understand pg encoding %d > 1",
147 pgid->pool = ceph_decode_64(p);
148 pgid->seed = ceph_decode_32(p);
149 *p += 4; /* skip deprecated preferred value */
154 extern struct ceph_osdmap *osdmap_decode(void **p, void *end);
155 extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
156 struct ceph_osdmap *map,
157 struct ceph_messenger *msgr);
158 extern void ceph_osdmap_destroy(struct ceph_osdmap *map);
160 /* calculate mapping of a file extent to an object */
161 extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
163 u64 *bno, u64 *oxoff, u64 *oxlen);
165 /* calculate mapping of object to a placement group */
166 extern int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap,
167 struct ceph_object_locator *oloc,
168 struct ceph_object_id *oid,
169 struct ceph_pg *pg_out);
171 extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
174 extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
175 struct ceph_pg pgid);
177 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
178 extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);