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 {
46 struct ceph_pg_mapping {
54 struct ceph_fsid fsid;
57 struct ceph_timespec created, modified;
59 u32 flags; /* CEPH_OSDMAP_* */
61 u32 max_osd; /* size of osd_state, _offload, _addr arrays */
62 u8 *osd_state; /* CEPH_OSD_* */
63 u32 *osd_weight; /* 0 = failed, 0x10000 = 100% normal */
64 struct ceph_entity_addr *osd_addr;
66 struct rb_root pg_temp;
67 struct rb_root pg_pools;
70 /* the CRUSH map specifies the mapping of placement groups to
71 * the list of osds that store+replicate them. */
72 struct crush_map *crush;
75 static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd)
77 return (osd < map->max_osd) && (map->osd_state[osd] & CEPH_OSD_UP);
80 static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag)
82 return map && (map->flags & flag);
85 extern char *ceph_osdmap_state_str(char *str, int len, int state);
87 static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map,
90 if (osd >= map->max_osd)
92 return &map->osd_addr[osd];
95 static inline int ceph_decode_pgid(void **p, void *end, struct ceph_pg *pgid)
99 if (!ceph_has_room(p, end, 1 + 8 + 4 + 4)) {
100 pr_warning("incomplete pg encoding");
104 version = ceph_decode_8(p);
106 pr_warning("do not understand pg encoding %d > 1",
111 pgid->pool = ceph_decode_64(p);
112 pgid->seed = ceph_decode_32(p);
113 *p += 4; /* skip deprecated preferred value */
118 extern struct ceph_osdmap *osdmap_decode(void **p, void *end);
119 extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
120 struct ceph_osdmap *map,
121 struct ceph_messenger *msgr);
122 extern void ceph_osdmap_destroy(struct ceph_osdmap *map);
124 /* calculate mapping of a file extent to an object */
125 extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
127 u64 *bno, u64 *oxoff, u64 *oxlen);
129 /* calculate mapping of object to a placement group */
130 extern int ceph_calc_ceph_pg(struct ceph_pg *pg, const char *oid,
131 struct ceph_osdmap *osdmap, uint64_t pool);
132 extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
135 extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
136 struct ceph_pg pgid);
138 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
139 extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);