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
;
39 s64 write_tier
; /* wins for read+write ops */
44 static inline bool ceph_can_shift_osds(struct ceph_pg_pool_info
*pool
)
47 case CEPH_POOL_TYPE_REP
:
49 case CEPH_POOL_TYPE_EC
:
56 struct ceph_object_locator
{
61 * Maximum supported by kernel client object name length
63 * (probably outdated: must be >= RBD_MAX_MD_NAME_LEN -- currently 100)
65 #define CEPH_MAX_OID_NAME_LEN 100
67 struct ceph_object_id
{
68 char name
[CEPH_MAX_OID_NAME_LEN
];
72 struct ceph_pg_mapping
{
88 struct ceph_fsid fsid
;
91 struct ceph_timespec created
, modified
;
93 u32 flags
; /* CEPH_OSDMAP_* */
95 u32 max_osd
; /* size of osd_state, _offload, _addr arrays */
96 u8
*osd_state
; /* CEPH_OSD_* */
97 u32
*osd_weight
; /* 0 = failed, 0x10000 = 100% normal */
98 struct ceph_entity_addr
*osd_addr
;
100 struct rb_root pg_temp
;
101 struct rb_root primary_temp
;
103 u32
*osd_primary_affinity
;
105 struct rb_root pg_pools
;
108 /* the CRUSH map specifies the mapping of placement groups to
109 * the list of osds that store+replicate them. */
110 struct crush_map
*crush
;
112 struct mutex crush_scratch_mutex
;
113 int crush_scratch_ary
[CEPH_PG_MAX_SIZE
* 3];
116 static inline void ceph_oid_set_name(struct ceph_object_id
*oid
,
122 if (len
> sizeof(oid
->name
)) {
123 WARN(1, "ceph_oid_set_name '%s' len %d vs %zu, truncating\n",
124 name
, len
, sizeof(oid
->name
));
125 len
= sizeof(oid
->name
);
128 memcpy(oid
->name
, name
, len
);
132 static inline void ceph_oid_copy(struct ceph_object_id
*dest
,
133 struct ceph_object_id
*src
)
135 BUG_ON(src
->name_len
> sizeof(dest
->name
));
136 memcpy(dest
->name
, src
->name
, src
->name_len
);
137 dest
->name_len
= src
->name_len
;
140 static inline int ceph_osd_exists(struct ceph_osdmap
*map
, int osd
)
142 return osd
>= 0 && osd
< map
->max_osd
&&
143 (map
->osd_state
[osd
] & CEPH_OSD_EXISTS
);
146 static inline int ceph_osd_is_up(struct ceph_osdmap
*map
, int osd
)
148 return ceph_osd_exists(map
, osd
) &&
149 (map
->osd_state
[osd
] & CEPH_OSD_UP
);
152 static inline int ceph_osd_is_down(struct ceph_osdmap
*map
, int osd
)
154 return !ceph_osd_is_up(map
, osd
);
157 static inline bool ceph_osdmap_flag(struct ceph_osdmap
*map
, int flag
)
159 return map
&& (map
->flags
& flag
);
162 extern char *ceph_osdmap_state_str(char *str
, int len
, int state
);
163 extern u32
ceph_get_primary_affinity(struct ceph_osdmap
*map
, int osd
);
165 static inline struct ceph_entity_addr
*ceph_osd_addr(struct ceph_osdmap
*map
,
168 if (osd
>= map
->max_osd
)
170 return &map
->osd_addr
[osd
];
173 static inline int ceph_decode_pgid(void **p
, void *end
, struct ceph_pg
*pgid
)
177 if (!ceph_has_room(p
, end
, 1 + 8 + 4 + 4)) {
178 pr_warning("incomplete pg encoding");
182 version
= ceph_decode_8(p
);
184 pr_warning("do not understand pg encoding %d > 1",
189 pgid
->pool
= ceph_decode_64(p
);
190 pgid
->seed
= ceph_decode_32(p
);
191 *p
+= 4; /* skip deprecated preferred value */
196 extern struct ceph_osdmap
*ceph_osdmap_decode(void **p
, void *end
);
197 extern struct ceph_osdmap
*osdmap_apply_incremental(void **p
, void *end
,
198 struct ceph_osdmap
*map
,
199 struct ceph_messenger
*msgr
);
200 extern void ceph_osdmap_destroy(struct ceph_osdmap
*map
);
202 /* calculate mapping of a file extent to an object */
203 extern int ceph_calc_file_object_mapping(struct ceph_file_layout
*layout
,
205 u64
*bno
, u64
*oxoff
, u64
*oxlen
);
207 /* calculate mapping of object to a placement group */
208 extern int ceph_oloc_oid_to_pg(struct ceph_osdmap
*osdmap
,
209 struct ceph_object_locator
*oloc
,
210 struct ceph_object_id
*oid
,
211 struct ceph_pg
*pg_out
);
213 extern int ceph_calc_pg_acting(struct ceph_osdmap
*osdmap
,
215 int *osds
, int *primary
);
216 extern int ceph_calc_pg_primary(struct ceph_osdmap
*osdmap
,
217 struct ceph_pg pgid
);
219 extern struct ceph_pg_pool_info
*ceph_pg_pool_by_id(struct ceph_osdmap
*map
,
222 extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap
*map
, u64 id
);
223 extern int ceph_pg_poolid_by_name(struct ceph_osdmap
*map
, const char *name
);