2009-10-06 22:31:10 +04:00
|
|
|
#ifndef _FS_CEPH_OSDMAP_H
|
|
|
|
#define _FS_CEPH_OSDMAP_H
|
|
|
|
|
|
|
|
#include <linux/rbtree.h>
|
2012-10-02 21:01:25 +04:00
|
|
|
#include <linux/ceph/types.h>
|
2013-04-02 03:58:26 +04:00
|
|
|
#include <linux/ceph/decode.h>
|
2012-10-02 21:01:25 +04:00
|
|
|
#include <linux/ceph/ceph_fs.h>
|
2010-04-07 02:14:15 +04:00
|
|
|
#include <linux/crush/crush.h>
|
2009-10-06 22:31:10 +04:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The osd map describes the current membership of the osd cluster and
|
|
|
|
* specifies the mapping of objects to placement groups and placement
|
|
|
|
* groups to (sets of) osds. That is, it completely specifies the
|
|
|
|
* (desired) distribution of all data objects in the system at some
|
|
|
|
* point in time.
|
|
|
|
*
|
|
|
|
* Each map version is identified by an epoch, which increases monotonically.
|
|
|
|
*
|
|
|
|
* The map can be updated either via an incremental map (diff) describing
|
|
|
|
* the change between two successive epochs, or as a fully encoded map.
|
|
|
|
*/
|
2013-02-23 22:38:16 +04:00
|
|
|
struct ceph_pg {
|
|
|
|
uint64_t pool;
|
|
|
|
uint32_t seed;
|
|
|
|
};
|
|
|
|
|
2013-02-26 22:39:09 +04:00
|
|
|
#define CEPH_POOL_FLAG_HASHPSPOOL 1
|
|
|
|
|
2009-10-06 22:31:10 +04:00
|
|
|
struct ceph_pg_pool_info {
|
2010-02-17 02:55:03 +03:00
|
|
|
struct rb_node node;
|
2013-02-23 22:41:09 +04:00
|
|
|
s64 id;
|
|
|
|
u8 type;
|
|
|
|
u8 size;
|
|
|
|
u8 crush_ruleset;
|
|
|
|
u8 object_hash;
|
|
|
|
u32 pg_num, pgp_num;
|
|
|
|
int pg_num_mask, pgp_num_mask;
|
2014-01-27 19:40:19 +04:00
|
|
|
s64 read_tier;
|
|
|
|
s64 write_tier; /* wins for read+write ops */
|
2013-02-23 22:41:09 +04:00
|
|
|
u64 flags;
|
2010-04-10 02:46:42 +04:00
|
|
|
char *name;
|
2009-10-06 22:31:10 +04:00
|
|
|
};
|
|
|
|
|
2013-02-23 22:41:09 +04:00
|
|
|
struct ceph_object_locator {
|
2014-01-27 19:40:18 +04:00
|
|
|
s64 pool;
|
2013-02-23 22:41:09 +04:00
|
|
|
};
|
|
|
|
|
2014-01-27 19:40:18 +04:00
|
|
|
/*
|
|
|
|
* Maximum supported by kernel client object name length
|
|
|
|
*
|
|
|
|
* (probably outdated: must be >= RBD_MAX_MD_NAME_LEN -- currently 100)
|
|
|
|
*/
|
|
|
|
#define CEPH_MAX_OID_NAME_LEN 100
|
|
|
|
|
|
|
|
struct ceph_object_id {
|
|
|
|
char name[CEPH_MAX_OID_NAME_LEN];
|
|
|
|
int name_len;
|
|
|
|
};
|
|
|
|
|
2009-10-06 22:31:10 +04:00
|
|
|
struct ceph_pg_mapping {
|
|
|
|
struct rb_node node;
|
2013-02-23 22:38:16 +04:00
|
|
|
struct ceph_pg pgid;
|
2014-03-21 21:05:29 +04:00
|
|
|
|
|
|
|
union {
|
|
|
|
struct {
|
|
|
|
int len;
|
|
|
|
int osds[];
|
|
|
|
} pg_temp;
|
|
|
|
};
|
2009-10-06 22:31:10 +04:00
|
|
|
};
|
|
|
|
|
|
|
|
struct ceph_osdmap {
|
|
|
|
struct ceph_fsid fsid;
|
|
|
|
u32 epoch;
|
|
|
|
u32 mkfs_epoch;
|
|
|
|
struct ceph_timespec created, modified;
|
|
|
|
|
|
|
|
u32 flags; /* CEPH_OSDMAP_* */
|
|
|
|
|
|
|
|
u32 max_osd; /* size of osd_state, _offload, _addr arrays */
|
|
|
|
u8 *osd_state; /* CEPH_OSD_* */
|
|
|
|
u32 *osd_weight; /* 0 = failed, 0x10000 = 100% normal */
|
|
|
|
struct ceph_entity_addr *osd_addr;
|
|
|
|
|
|
|
|
struct rb_root pg_temp;
|
2010-02-17 02:55:03 +03:00
|
|
|
struct rb_root pg_pools;
|
|
|
|
u32 pool_max;
|
2009-10-06 22:31:10 +04:00
|
|
|
|
|
|
|
/* the CRUSH map specifies the mapping of placement groups to
|
|
|
|
* the list of osds that store+replicate them. */
|
|
|
|
struct crush_map *crush;
|
2014-01-31 19:54:26 +04:00
|
|
|
|
|
|
|
struct mutex crush_scratch_mutex;
|
|
|
|
int crush_scratch_ary[CEPH_PG_MAX_SIZE * 3];
|
2009-10-06 22:31:10 +04:00
|
|
|
};
|
|
|
|
|
2014-01-27 19:40:18 +04:00
|
|
|
static inline void ceph_oid_set_name(struct ceph_object_id *oid,
|
|
|
|
const char *name)
|
|
|
|
{
|
|
|
|
int len;
|
|
|
|
|
|
|
|
len = strlen(name);
|
|
|
|
if (len > sizeof(oid->name)) {
|
|
|
|
WARN(1, "ceph_oid_set_name '%s' len %d vs %zu, truncating\n",
|
|
|
|
name, len, sizeof(oid->name));
|
|
|
|
len = sizeof(oid->name);
|
|
|
|
}
|
|
|
|
|
|
|
|
memcpy(oid->name, name, len);
|
|
|
|
oid->name_len = len;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void ceph_oid_copy(struct ceph_object_id *dest,
|
|
|
|
struct ceph_object_id *src)
|
|
|
|
{
|
|
|
|
BUG_ON(src->name_len > sizeof(dest->name));
|
|
|
|
memcpy(dest->name, src->name, src->name_len);
|
|
|
|
dest->name_len = src->name_len;
|
|
|
|
}
|
|
|
|
|
2009-10-06 22:31:10 +04:00
|
|
|
static inline int ceph_osd_is_up(struct ceph_osdmap *map, int osd)
|
|
|
|
{
|
|
|
|
return (osd < map->max_osd) && (map->osd_state[osd] & CEPH_OSD_UP);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline bool ceph_osdmap_flag(struct ceph_osdmap *map, int flag)
|
|
|
|
{
|
|
|
|
return map && (map->flags & flag);
|
|
|
|
}
|
|
|
|
|
|
|
|
extern char *ceph_osdmap_state_str(char *str, int len, int state);
|
|
|
|
|
|
|
|
static inline struct ceph_entity_addr *ceph_osd_addr(struct ceph_osdmap *map,
|
|
|
|
int osd)
|
|
|
|
{
|
|
|
|
if (osd >= map->max_osd)
|
|
|
|
return NULL;
|
|
|
|
return &map->osd_addr[osd];
|
|
|
|
}
|
|
|
|
|
2013-04-02 03:58:26 +04:00
|
|
|
static inline int ceph_decode_pgid(void **p, void *end, struct ceph_pg *pgid)
|
|
|
|
{
|
|
|
|
__u8 version;
|
|
|
|
|
|
|
|
if (!ceph_has_room(p, end, 1 + 8 + 4 + 4)) {
|
|
|
|
pr_warning("incomplete pg encoding");
|
|
|
|
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
version = ceph_decode_8(p);
|
|
|
|
if (version > 1) {
|
|
|
|
pr_warning("do not understand pg encoding %d > 1",
|
|
|
|
(int)version);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
pgid->pool = ceph_decode_64(p);
|
|
|
|
pgid->seed = ceph_decode_32(p);
|
|
|
|
*p += 4; /* skip deprecated preferred value */
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-03-13 18:36:13 +04:00
|
|
|
extern struct ceph_osdmap *ceph_osdmap_decode(void **p, void *end);
|
2009-10-06 22:31:10 +04:00
|
|
|
extern struct ceph_osdmap *osdmap_apply_incremental(void **p, void *end,
|
|
|
|
struct ceph_osdmap *map,
|
|
|
|
struct ceph_messenger *msgr);
|
|
|
|
extern void ceph_osdmap_destroy(struct ceph_osdmap *map);
|
|
|
|
|
|
|
|
/* calculate mapping of a file extent to an object */
|
2012-09-25 07:59:48 +04:00
|
|
|
extern int ceph_calc_file_object_mapping(struct ceph_file_layout *layout,
|
2012-11-14 19:38:19 +04:00
|
|
|
u64 off, u64 len,
|
2012-09-25 07:59:48 +04:00
|
|
|
u64 *bno, u64 *oxoff, u64 *oxlen);
|
2009-10-06 22:31:10 +04:00
|
|
|
|
|
|
|
/* calculate mapping of object to a placement group */
|
2014-01-27 19:40:19 +04:00
|
|
|
extern int ceph_oloc_oid_to_pg(struct ceph_osdmap *osdmap,
|
|
|
|
struct ceph_object_locator *oloc,
|
|
|
|
struct ceph_object_id *oid,
|
|
|
|
struct ceph_pg *pg_out);
|
|
|
|
|
2013-01-08 21:15:10 +04:00
|
|
|
extern int ceph_calc_pg_acting(struct ceph_osdmap *osdmap,
|
2013-02-23 22:38:16 +04:00
|
|
|
struct ceph_pg pgid,
|
2010-05-10 21:24:48 +04:00
|
|
|
int *acting);
|
2009-11-04 22:39:12 +03:00
|
|
|
extern int ceph_calc_pg_primary(struct ceph_osdmap *osdmap,
|
2013-02-23 22:38:16 +04:00
|
|
|
struct ceph_pg pgid);
|
2009-10-06 22:31:10 +04:00
|
|
|
|
2014-01-27 19:40:19 +04:00
|
|
|
extern struct ceph_pg_pool_info *ceph_pg_pool_by_id(struct ceph_osdmap *map,
|
|
|
|
u64 id);
|
|
|
|
|
2012-10-31 04:40:33 +04:00
|
|
|
extern const char *ceph_pg_pool_name_by_id(struct ceph_osdmap *map, u64 id);
|
2010-05-17 23:31:35 +04:00
|
|
|
extern int ceph_pg_poolid_by_name(struct ceph_osdmap *map, const char *name);
|
|
|
|
|
2009-10-06 22:31:10 +04:00
|
|
|
#endif
|