1 // -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
2 // vim: ts=8 sw=2 smarttab
4 * Ceph - scalable distributed file system
6 * Copyright (C) 2011 New Dream Network
8 * This is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License version 2.1, as published by the Free Software
11 * Foundation. See file COPYING.
22 #include <netinet/in.h>
23 #if defined(__linux__)
24 #include <linux/types.h>
25 #elif defined(__FreeBSD__)
26 #include <sys/types.h>
31 #include "../rados/librados.h"
34 #define LIBRBD_VER_MAJOR 1
35 #define LIBRBD_VER_MINOR 12
36 #define LIBRBD_VER_EXTRA 0
38 #define LIBRBD_VERSION(maj, min, extra) ((maj << 16) + (min << 8) + extra)
40 #define LIBRBD_VERSION_CODE LIBRBD_VERSION(LIBRBD_VER_MAJOR, LIBRBD_VER_MINOR, LIBRBD_VER_EXTRA)
42 #define LIBRBD_SUPPORTS_AIO_FLUSH 1
43 #define LIBRBD_SUPPORTS_AIO_OPEN 1
44 #define LIBRBD_SUPPORTS_LOCKING 1
45 #define LIBRBD_SUPPORTS_INVALIDATE 1
46 #define LIBRBD_SUPPORTS_IOVEC 1
47 #define LIBRBD_SUPPORTS_WATCH 0
48 #define LIBRBD_SUPPORTS_WRITESAME 1
51 #define CEPH_RBD_API __attribute__ ((visibility ("default")))
56 #define RBD_FLAG_OBJECT_MAP_INVALID (1<<0)
57 #define RBD_FLAG_FAST_DIFF_INVALID (1<<1)
59 typedef void *rbd_snap_t;
60 typedef void *rbd_image_t;
61 typedef void *rbd_image_options_t;
63 typedef void *rbd_completion_t;
64 typedef void (*rbd_callback_t)(rbd_completion_t cb, void *arg);
66 typedef int (*librbd_progress_fn_t)(uint64_t offset, uint64_t total, void *ptr);
68 typedef void (*rbd_update_callback_t)(void *arg);
76 #define RBD_MAX_IMAGE_NAME_SIZE 96
77 #define RBD_MAX_BLOCK_NAME_SIZE 24
79 #define RBD_SNAP_REMOVE_UNPROTECT 1 << 0
80 #define RBD_SNAP_REMOVE_FLATTEN 1 << 1
81 #define RBD_SNAP_REMOVE_FORCE (RBD_SNAP_REMOVE_UNPROTECT | RBD_SNAP_REMOVE_FLATTEN)
84 * These types used to in set_image_notification to indicate the type of event
89 EVENT_TYPE_EVENTFD = 2
97 char block_name_prefix[RBD_MAX_BLOCK_NAME_SIZE]; /* deprecated */
98 int64_t parent_pool; /* deprecated */
99 char parent_name[RBD_MAX_IMAGE_NAME_SIZE]; /* deprecated */
103 RBD_MIRROR_MODE_DISABLED, /* mirroring is disabled */
104 RBD_MIRROR_MODE_IMAGE, /* mirroring enabled on a per-image basis */
105 RBD_MIRROR_MODE_POOL /* mirroring enabled on all journaled images */
115 RBD_MIRROR_IMAGE_DISABLING = 0,
116 RBD_MIRROR_IMAGE_ENABLED = 1,
117 RBD_MIRROR_IMAGE_DISABLED = 2
118 } rbd_mirror_image_state_t;
122 rbd_mirror_image_state_t state;
124 } rbd_mirror_image_info_t;
127 MIRROR_IMAGE_STATUS_STATE_UNKNOWN = 0,
128 MIRROR_IMAGE_STATUS_STATE_ERROR = 1,
129 MIRROR_IMAGE_STATUS_STATE_SYNCING = 2,
130 MIRROR_IMAGE_STATUS_STATE_STARTING_REPLAY = 3,
131 MIRROR_IMAGE_STATUS_STATE_REPLAYING = 4,
132 MIRROR_IMAGE_STATUS_STATE_STOPPING_REPLAY = 5,
133 MIRROR_IMAGE_STATUS_STATE_STOPPED = 6,
134 } rbd_mirror_image_status_state_t;
138 rbd_mirror_image_info_t info;
139 rbd_mirror_image_status_state_t state;
143 } rbd_mirror_image_status_t;
146 RBD_LOCK_MODE_EXCLUSIVE = 0,
147 RBD_LOCK_MODE_SHARED = 1,
150 CEPH_RBD_API void rbd_version(int *major, int *minor, int *extra);
154 RBD_IMAGE_OPTION_FORMAT = 0,
155 RBD_IMAGE_OPTION_FEATURES = 1,
156 RBD_IMAGE_OPTION_ORDER = 2,
157 RBD_IMAGE_OPTION_STRIPE_UNIT = 3,
158 RBD_IMAGE_OPTION_STRIPE_COUNT = 4,
159 RBD_IMAGE_OPTION_JOURNAL_ORDER = 5,
160 RBD_IMAGE_OPTION_JOURNAL_SPLAY_WIDTH = 6,
161 RBD_IMAGE_OPTION_JOURNAL_POOL = 7,
162 RBD_IMAGE_OPTION_FEATURES_SET = 8,
163 RBD_IMAGE_OPTION_FEATURES_CLEAR = 9,
164 RBD_IMAGE_OPTION_DATA_POOL = 10
168 RBD_TRASH_IMAGE_SOURCE_USER = 0,
169 RBD_TRASH_IMAGE_SOURCE_MIRRORING = 1
170 } rbd_trash_image_source_t;
175 rbd_trash_image_source_t source;
176 time_t deletion_time;
177 time_t deferment_end_time;
178 } rbd_trash_image_info_t;
180 CEPH_RBD_API void rbd_image_options_create(rbd_image_options_t* opts);
181 CEPH_RBD_API void rbd_image_options_destroy(rbd_image_options_t opts);
182 CEPH_RBD_API int rbd_image_options_set_string(rbd_image_options_t opts,
183 int optname, const char* optval);
184 CEPH_RBD_API int rbd_image_options_set_uint64(rbd_image_options_t opts,
185 int optname, uint64_t optval);
186 CEPH_RBD_API int rbd_image_options_get_string(rbd_image_options_t opts,
187 int optname, char* optval,
189 CEPH_RBD_API int rbd_image_options_get_uint64(rbd_image_options_t opts,
190 int optname, uint64_t* optval);
191 CEPH_RBD_API int rbd_image_options_is_set(rbd_image_options_t opts,
192 int optname, bool* is_set);
193 CEPH_RBD_API int rbd_image_options_unset(rbd_image_options_t opts, int optname);
194 CEPH_RBD_API void rbd_image_options_clear(rbd_image_options_t opts);
195 CEPH_RBD_API int rbd_image_options_is_empty(rbd_image_options_t opts);
198 CEPH_RBD_API int rbd_list(rados_ioctx_t io, char *names, size_t *size);
199 CEPH_RBD_API int rbd_create(rados_ioctx_t io, const char *name, uint64_t size,
201 CEPH_RBD_API int rbd_create2(rados_ioctx_t io, const char *name, uint64_t size,
202 uint64_t features, int *order);
204 * create new rbd image
206 * The stripe_unit must be a factor of the object size (1 << order).
207 * The stripe_count can be one (no intra-object striping) or greater
208 * than one. The RBD_FEATURE_STRIPINGV2 must be specified if the
209 * stripe_unit != the object size and the stripe_count is != 1.
212 * @param name image name
213 * @param size image size in bytes
214 * @param features initial feature bits
215 * @param order object/block size, as a power of two (object size == 1 << order)
216 * @param stripe_unit stripe unit size, in bytes.
217 * @param stripe_count number of objects to stripe over before looping
218 * @return 0 on success, or negative error code
220 CEPH_RBD_API int rbd_create3(rados_ioctx_t io, const char *name, uint64_t size,
221 uint64_t features, int *order,
222 uint64_t stripe_unit, uint64_t stripe_count);
223 CEPH_RBD_API int rbd_create4(rados_ioctx_t io, const char *name, uint64_t size,
224 rbd_image_options_t opts);
225 CEPH_RBD_API int rbd_clone(rados_ioctx_t p_ioctx, const char *p_name,
226 const char *p_snapname, rados_ioctx_t c_ioctx,
227 const char *c_name, uint64_t features, int *c_order);
228 CEPH_RBD_API int rbd_clone2(rados_ioctx_t p_ioctx, const char *p_name,
229 const char *p_snapname, rados_ioctx_t c_ioctx,
230 const char *c_name, uint64_t features, int *c_order,
231 uint64_t stripe_unit, int stripe_count);
232 CEPH_RBD_API int rbd_clone3(rados_ioctx_t p_ioctx, const char *p_name,
233 const char *p_snapname, rados_ioctx_t c_ioctx,
234 const char *c_name, rbd_image_options_t c_opts);
235 CEPH_RBD_API int rbd_remove(rados_ioctx_t io, const char *name);
236 CEPH_RBD_API int rbd_remove_with_progress(rados_ioctx_t io, const char *name,
237 librbd_progress_fn_t cb,
239 CEPH_RBD_API int rbd_rename(rados_ioctx_t src_io_ctx, const char *srcname,
240 const char *destname);
242 CEPH_RBD_API int rbd_trash_move(rados_ioctx_t io, const char *name,
244 CEPH_RBD_API int rbd_trash_get(rados_ioctx_t io, const char *id,
245 rbd_trash_image_info_t *info);
246 CEPH_RBD_API void rbd_trash_get_cleanup(rbd_trash_image_info_t *info);
247 CEPH_RBD_API int rbd_trash_list(rados_ioctx_t io,
248 rbd_trash_image_info_t *trash_entries,
249 size_t *num_entries);
250 CEPH_RBD_API void rbd_trash_list_cleanup(rbd_trash_image_info_t *trash_entries,
252 CEPH_RBD_API int rbd_trash_remove(rados_ioctx_t io, const char *id, bool force);
253 CEPH_RBD_API int rbd_trash_remove_with_progress(rados_ioctx_t io, const char *id,
254 bool force, librbd_progress_fn_t cb,
256 CEPH_RBD_API int rbd_trash_restore(rados_ioctx_t io, const char *id,
260 CEPH_RBD_API int rbd_mirror_mode_get(rados_ioctx_t io_ctx,
261 rbd_mirror_mode_t *mirror_mode);
262 CEPH_RBD_API int rbd_mirror_mode_set(rados_ioctx_t io_ctx,
263 rbd_mirror_mode_t mirror_mode);
264 CEPH_RBD_API int rbd_mirror_peer_add(rados_ioctx_t io_ctx,
265 char *uuid, size_t uuid_max_length,
266 const char *cluster_name,
267 const char *client_name);
268 CEPH_RBD_API int rbd_mirror_peer_remove(rados_ioctx_t io_ctx,
270 CEPH_RBD_API int rbd_mirror_peer_list(rados_ioctx_t io_ctx,
271 rbd_mirror_peer_t *peers, int *max_peers);
272 CEPH_RBD_API void rbd_mirror_peer_list_cleanup(rbd_mirror_peer_t *peers,
274 CEPH_RBD_API int rbd_mirror_peer_set_client(rados_ioctx_t io_ctx,
276 const char *client_name);
277 CEPH_RBD_API int rbd_mirror_peer_set_cluster(rados_ioctx_t io_ctx,
279 const char *cluster_name);
280 CEPH_RBD_API int rbd_mirror_image_status_list(rados_ioctx_t io_ctx,
281 const char *start_id, size_t max,
283 rbd_mirror_image_status_t *images,
285 CEPH_RBD_API void rbd_mirror_image_status_list_cleanup(char **image_ids,
286 rbd_mirror_image_status_t *images, size_t len);
287 CEPH_RBD_API int rbd_mirror_image_status_summary(rados_ioctx_t io_ctx,
288 rbd_mirror_image_status_state_t *states, int *counts, size_t *maxlen);
290 CEPH_RBD_API int rbd_open(rados_ioctx_t io, const char *name,
291 rbd_image_t *image, const char *snap_name);
292 CEPH_RBD_API int rbd_open_by_id(rados_ioctx_t io, const char *id,
293 rbd_image_t *image, const char *snap_name);
295 CEPH_RBD_API int rbd_aio_open(rados_ioctx_t io, const char *name,
296 rbd_image_t *image, const char *snap_name,
298 CEPH_RBD_API int rbd_aio_open_by_id(rados_ioctx_t io, const char *id,
299 rbd_image_t *image, const char *snap_name,
303 * Open an image in read-only mode.
305 * This is intended for use by clients that cannot write to a block
306 * device due to cephx restrictions. There will be no watch
307 * established on the header object, since a watch is a write. This
308 * means the metadata reported about this image (parents, snapshots,
309 * size, etc.) may become stale. This should not be used for
310 * long-running operations, unless you can be sure that one of these
311 * properties changing is safe.
313 * Attempting to write to a read-only image will return -EROFS.
315 * @param io ioctx to determine the pool the image is in
316 * @param name image name
317 * @param image where to store newly opened image handle
318 * @param snap_name name of snapshot to open at, or NULL for no snapshot
319 * @returns 0 on success, negative error code on failure
321 CEPH_RBD_API int rbd_open_read_only(rados_ioctx_t io, const char *name,
322 rbd_image_t *image, const char *snap_name);
323 CEPH_RBD_API int rbd_open_by_id_read_only(rados_ioctx_t io, const char *id,
324 rbd_image_t *image, const char *snap_name);
325 CEPH_RBD_API int rbd_aio_open_read_only(rados_ioctx_t io, const char *name,
326 rbd_image_t *image, const char *snap_name,
328 CEPH_RBD_API int rbd_aio_open_by_id_read_only(rados_ioctx_t io, const char *id,
329 rbd_image_t *image, const char *snap_name,
331 CEPH_RBD_API int rbd_close(rbd_image_t image);
332 CEPH_RBD_API int rbd_aio_close(rbd_image_t image, rbd_completion_t c);
333 CEPH_RBD_API int rbd_resize(rbd_image_t image, uint64_t size);
334 CEPH_RBD_API int rbd_resize2(rbd_image_t image, uint64_t size, bool allow_shrink,
335 librbd_progress_fn_t cb, void *cbdata);
336 CEPH_RBD_API int rbd_resize_with_progress(rbd_image_t image, uint64_t size,
337 librbd_progress_fn_t cb, void *cbdata);
338 CEPH_RBD_API int rbd_stat(rbd_image_t image, rbd_image_info_t *info,
340 CEPH_RBD_API int rbd_get_old_format(rbd_image_t image, uint8_t *old);
341 CEPH_RBD_API int rbd_get_size(rbd_image_t image, uint64_t *size);
342 CEPH_RBD_API int rbd_get_features(rbd_image_t image, uint64_t *features);
343 CEPH_RBD_API int rbd_update_features(rbd_image_t image, uint64_t features,
345 CEPH_RBD_API int rbd_get_stripe_unit(rbd_image_t image, uint64_t *stripe_unit);
346 CEPH_RBD_API int rbd_get_stripe_count(rbd_image_t image,
347 uint64_t *stripe_count);
349 CEPH_RBD_API int rbd_get_create_timestamp(rbd_image_t image,
350 struct timespec *timestamp);
352 CEPH_RBD_API int rbd_get_overlap(rbd_image_t image, uint64_t *overlap);
353 CEPH_RBD_API int rbd_get_id(rbd_image_t image, char *id, size_t id_len);
354 CEPH_RBD_API int rbd_get_block_name_prefix(rbd_image_t image,
355 char *prefix, size_t prefix_len);
356 CEPH_RBD_API int64_t rbd_get_data_pool_id(rbd_image_t image);
357 CEPH_RBD_API int rbd_get_parent_info(rbd_image_t image,
358 char *parent_poolname, size_t ppoolnamelen,
359 char *parent_name, size_t pnamelen,
360 char *parent_snapname,
361 size_t psnapnamelen);
362 CEPH_RBD_API int rbd_get_parent_info2(rbd_image_t image,
363 char *parent_poolname,
365 char *parent_name, size_t pnamelen,
366 char *parent_id, size_t pidlen,
367 char *parent_snapname,
368 size_t psnapnamelen);
369 CEPH_RBD_API int rbd_get_flags(rbd_image_t image, uint64_t *flags);
370 CEPH_RBD_API int rbd_set_image_notification(rbd_image_t image, int fd, int type);
372 /* exclusive lock feature */
373 CEPH_RBD_API int rbd_is_exclusive_lock_owner(rbd_image_t image, int *is_owner);
374 CEPH_RBD_API int rbd_lock_acquire(rbd_image_t image, rbd_lock_mode_t lock_mode);
375 CEPH_RBD_API int rbd_lock_release(rbd_image_t image);
376 CEPH_RBD_API int rbd_lock_get_owners(rbd_image_t image,
377 rbd_lock_mode_t *lock_mode,
379 size_t *max_lock_owners);
380 CEPH_RBD_API void rbd_lock_get_owners_cleanup(char **lock_owners,
381 size_t lock_owner_count);
382 CEPH_RBD_API int rbd_lock_break(rbd_image_t image, rbd_lock_mode_t lock_mode,
383 const char *lock_owner);
385 /* object map feature */
386 CEPH_RBD_API int rbd_rebuild_object_map(rbd_image_t image,
387 librbd_progress_fn_t cb, void *cbdata);
389 CEPH_RBD_API int rbd_copy(rbd_image_t image, rados_ioctx_t dest_io_ctx,
390 const char *destname);
391 CEPH_RBD_API int rbd_copy2(rbd_image_t src, rbd_image_t dest);
392 CEPH_RBD_API int rbd_copy3(rbd_image_t src, rados_ioctx_t dest_io_ctx,
393 const char *destname, rbd_image_options_t dest_opts);
394 CEPH_RBD_API int rbd_copy4(rbd_image_t src, rados_ioctx_t dest_io_ctx,
395 const char *destname, rbd_image_options_t dest_opts,
397 CEPH_RBD_API int rbd_copy_with_progress(rbd_image_t image, rados_ioctx_t dest_p,
398 const char *destname,
399 librbd_progress_fn_t cb, void *cbdata);
400 CEPH_RBD_API int rbd_copy_with_progress2(rbd_image_t src, rbd_image_t dest,
401 librbd_progress_fn_t cb, void *cbdata);
402 CEPH_RBD_API int rbd_copy_with_progress3(rbd_image_t image,
403 rados_ioctx_t dest_p,
404 const char *destname,
405 rbd_image_options_t dest_opts,
406 librbd_progress_fn_t cb, void *cbdata);
407 CEPH_RBD_API int rbd_copy_with_progress4(rbd_image_t image,
408 rados_ioctx_t dest_p,
409 const char *destname,
410 rbd_image_options_t dest_opts,
411 librbd_progress_fn_t cb, void *cbdata,
415 CEPH_RBD_API int rbd_snap_list(rbd_image_t image, rbd_snap_info_t *snaps,
417 CEPH_RBD_API void rbd_snap_list_end(rbd_snap_info_t *snaps);
418 CEPH_RBD_API int rbd_snap_create(rbd_image_t image, const char *snapname);
419 CEPH_RBD_API int rbd_snap_remove(rbd_image_t image, const char *snapname);
420 CEPH_RBD_API int rbd_snap_remove2(rbd_image_t image, const char *snap_name, uint32_t flags,
421 librbd_progress_fn_t cb, void *cbdata);
422 CEPH_RBD_API int rbd_snap_rollback(rbd_image_t image, const char *snapname);
423 CEPH_RBD_API int rbd_snap_rollback_with_progress(rbd_image_t image,
424 const char *snapname,
425 librbd_progress_fn_t cb,
427 CEPH_RBD_API int rbd_snap_rename(rbd_image_t image, const char *snapname,
428 const char* dstsnapsname);
430 * Prevent a snapshot from being deleted until it is unprotected.
432 * @param snap_name which snapshot to protect
433 * @returns 0 on success, negative error code on failure
434 * @returns -EBUSY if snap is already protected
436 CEPH_RBD_API int rbd_snap_protect(rbd_image_t image, const char *snap_name);
438 * Allow a snaphshot to be deleted.
440 * @param snap_name which snapshot to unprotect
441 * @returns 0 on success, negative error code on failure
442 * @returns -EINVAL if snap is not protected
444 CEPH_RBD_API int rbd_snap_unprotect(rbd_image_t image, const char *snap_name);
446 * Determine whether a snapshot is protected.
448 * @param snap_name which snapshot query
449 * @param is_protected where to store the result (0 or 1)
450 * @returns 0 on success, negative error code on failure
452 CEPH_RBD_API int rbd_snap_is_protected(rbd_image_t image, const char *snap_name,
455 * Get the current snapshot limit for an image. If no limit is set,
456 * UINT64_MAX is returned.
458 * @param limit pointer where the limit will be stored on success
459 * @returns 0 on success, negative error code on failure
461 CEPH_RBD_API int rbd_snap_get_limit(rbd_image_t image, uint64_t *limit);
464 * Set a limit for the number of snapshots that may be taken of an image.
466 * @param limit the maximum number of snapshots allowed in the future.
467 * @returns 0 on success, negative error code on failure
469 CEPH_RBD_API int rbd_snap_set_limit(rbd_image_t image, uint64_t limit);
472 * Get the timestamp of a snapshot for an image.
474 * @param snap_id the snap id of a snapshot of input image.
475 * @param timestamp the timestamp of input snapshot.
476 * @returns 0 on success, negative error code on failure
478 CEPH_RBD_API int rbd_snap_get_timestamp(rbd_image_t image, uint64_t snap_id, struct timespec *timestamp);
480 CEPH_RBD_API int rbd_snap_set(rbd_image_t image, const char *snapname);
482 CEPH_RBD_API int rbd_flatten(rbd_image_t image);
484 CEPH_RBD_API int rbd_flatten_with_progress(rbd_image_t image,
485 librbd_progress_fn_t cb,
489 * List all images that are cloned from the image at the
490 * snapshot that is set via rbd_snap_set().
492 * This iterates over all pools, so it should be run by a user with
493 * read access to all of them. pools_len and images_len are filled in
494 * with the number of bytes put into the pools and images buffers.
496 * If the provided buffers are too short, the required lengths are
497 * still filled in, but the data is not and -ERANGE is returned.
498 * Otherwise, the buffers are filled with the pool and image names
499 * of the children, with a '\0' after each.
501 * @param image which image (and implicitly snapshot) to list clones of
502 * @param pools buffer in which to store pool names
503 * @param pools_len number of bytes in pools buffer
504 * @param images buffer in which to store image names
505 * @param images_len number of bytes in images buffer
506 * @returns number of children on success, negative error code on failure
507 * @returns -ERANGE if either buffer is too short
509 CEPH_RBD_API ssize_t rbd_list_children(rbd_image_t image, char *pools,
510 size_t *pools_len, char *images,
514 * @defgroup librbd_h_locking Advisory Locking
516 * An rbd image may be locking exclusively, or shared, to facilitate
517 * e.g. live migration where the image may be open in two places at once.
518 * These locks are intended to guard against more than one client
519 * writing to an image without coordination. They don't need to
520 * be used for snapshots, since snapshots are read-only.
522 * Currently locks only guard against locks being acquired.
523 * They do not prevent anything else.
525 * A locker is identified by the internal rados client id of the
526 * holder and a user-defined cookie. This (client id, cookie) pair
527 * must be unique for each locker.
529 * A shared lock also has a user-defined tag associated with it. Each
530 * additional shared lock must specify the same tag or lock
531 * acquisition will fail. This can be used by e.g. groups of hosts
532 * using a clustered filesystem on top of an rbd image to make sure
533 * they're accessing the correct image.
538 * List clients that have locked the image and information about the lock.
540 * The number of bytes required in each buffer is put in the
541 * corresponding size out parameter. If any of the provided buffers
542 * are too short, -ERANGE is returned after these sizes are filled in.
544 * @param exclusive where to store whether the lock is exclusive (1) or shared (0)
545 * @param tag where to store the tag associated with the image
546 * @param tag_len number of bytes in tag buffer
547 * @param clients buffer in which locker clients are stored, separated by '\0'
548 * @param clients_len number of bytes in the clients buffer
549 * @param cookies buffer in which locker cookies are stored, separated by '\0'
550 * @param cookies_len number of bytes in the cookies buffer
551 * @param addrs buffer in which locker addresses are stored, separated by '\0'
552 * @param addrs_len number of bytes in the clients buffer
553 * @returns number of lockers on success, negative error code on failure
554 * @returns -ERANGE if any of the buffers are too short
556 CEPH_RBD_API ssize_t rbd_list_lockers(rbd_image_t image, int *exclusive,
557 char *tag, size_t *tag_len,
558 char *clients, size_t *clients_len,
559 char *cookies, size_t *cookies_len,
560 char *addrs, size_t *addrs_len);
563 * Take an exclusive lock on the image.
565 * @param image the image to lock
566 * @param cookie user-defined identifier for this instance of the lock
567 * @returns 0 on success, negative error code on failure
568 * @returns -EBUSY if the lock is already held by another (client, cookie) pair
569 * @returns -EEXIST if the lock is already held by the same (client, cookie) pair
571 CEPH_RBD_API int rbd_lock_exclusive(rbd_image_t image, const char *cookie);
574 * Take a shared lock on the image.
576 * Other clients may also take a shared lock, as lock as they use the
579 * @param image the image to lock
580 * @param cookie user-defined identifier for this instance of the lock
581 * @param tag user-defined identifier for this shared use of the lock
582 * @returns 0 on success, negative error code on failure
583 * @returns -EBUSY if the lock is already held by another (client, cookie) pair
584 * @returns -EEXIST if the lock is already held by the same (client, cookie) pair
586 CEPH_RBD_API int rbd_lock_shared(rbd_image_t image, const char *cookie,
590 * Release a shared or exclusive lock on the image.
592 * @param image the image to unlock
593 * @param cookie user-defined identifier for the instance of the lock
594 * @returns 0 on success, negative error code on failure
595 * @returns -ENOENT if the lock is not held by the specified (client, cookie) pair
597 CEPH_RBD_API int rbd_unlock(rbd_image_t image, const char *cookie);
600 * Release a shared or exclusive lock that was taken by the specified client.
602 * @param image the image to unlock
603 * @param client the entity holding the lock (as given by rbd_list_lockers())
604 * @param cookie user-defined identifier for the instance of the lock to break
605 * @returns 0 on success, negative error code on failure
606 * @returns -ENOENT if the lock is not held by the specified (client, cookie) pair
608 CEPH_RBD_API int rbd_break_lock(rbd_image_t image, const char *client,
614 CEPH_RBD_API ssize_t rbd_read(rbd_image_t image, uint64_t ofs, size_t len,
617 * @param op_flags: see librados.h constants beginning with LIBRADOS_OP_FLAG
619 CEPH_RBD_API ssize_t rbd_read2(rbd_image_t image, uint64_t ofs, size_t len,
620 char *buf, int op_flags);
621 /* DEPRECATED; use rbd_read_iterate2 */
622 CEPH_RBD_API int64_t rbd_read_iterate(rbd_image_t image, uint64_t ofs, size_t len,
623 int (*cb)(uint64_t, size_t, const char *, void *),
627 * iterate read over an image
629 * Reads each region of the image and calls the callback. If the
630 * buffer pointer passed to the callback is NULL, the given extent is
631 * defined to be zeros (a hole). Normally the granularity for the
632 * callback is the image stripe size.
634 * @param image image to read
635 * @param ofs offset to start from
636 * @param len bytes of source image to cover
637 * @param cb callback for each region
638 * @returns 0 success, error otherwise
640 CEPH_RBD_API int rbd_read_iterate2(rbd_image_t image, uint64_t ofs, uint64_t len,
641 int (*cb)(uint64_t, size_t, const char *, void *),
644 * get difference between two versions of an image
646 * This will return the differences between two versions of an image
647 * via a callback, which gets the offset and length and a flag
648 * indicating whether the extent exists (1), or is known/defined to
649 * be zeros (a hole, 0). If the source snapshot name is NULL, we
650 * interpret that as the beginning of time and return all allocated
651 * regions of the image. The end version is whatever is currently
652 * selected for the image handle (either a snapshot or the writeable
655 * @param fromsnapname start snapshot name, or NULL
656 * @param ofs start offset
657 * @param len len in bytes of region to report on
658 * @param include_parent 1 if full history diff should include parent
659 * @param whole_object 1 if diff extents should cover whole object
660 * @param cb callback to call for each allocated region
661 * @param arg argument to pass to the callback
662 * @returns 0 on success, or negative error code on error
664 CEPH_RBD_API int rbd_diff_iterate(rbd_image_t image,
665 const char *fromsnapname,
666 uint64_t ofs, uint64_t len,
667 int (*cb)(uint64_t, size_t, int, void *),
669 CEPH_RBD_API int rbd_diff_iterate2(rbd_image_t image,
670 const char *fromsnapname,
671 uint64_t ofs, uint64_t len,
672 uint8_t include_parent, uint8_t whole_object,
673 int (*cb)(uint64_t, size_t, int, void *),
675 CEPH_RBD_API ssize_t rbd_write(rbd_image_t image, uint64_t ofs, size_t len,
678 * @param op_flags: see librados.h constants beginning with LIBRADOS_OP_FLAG
680 CEPH_RBD_API ssize_t rbd_write2(rbd_image_t image, uint64_t ofs, size_t len,
681 const char *buf, int op_flags);
682 CEPH_RBD_API int rbd_discard(rbd_image_t image, uint64_t ofs, uint64_t len);
683 CEPH_RBD_API ssize_t rbd_writesame(rbd_image_t image, uint64_t ofs, size_t len,
684 const char *buf, size_t data_len, int op_flags);
685 CEPH_RBD_API ssize_t rbd_compare_and_write(rbd_image_t image, uint64_t ofs,
686 size_t len, const char *cmp_buf,
687 const char *buf, uint64_t *mismatch_off,
690 CEPH_RBD_API int rbd_aio_write(rbd_image_t image, uint64_t off, size_t len,
691 const char *buf, rbd_completion_t c);
694 * @param op_flags: see librados.h constants beginning with LIBRADOS_OP_FLAG
696 CEPH_RBD_API int rbd_aio_write2(rbd_image_t image, uint64_t off, size_t len,
697 const char *buf, rbd_completion_t c,
699 CEPH_RBD_API int rbd_aio_writev(rbd_image_t image, const struct iovec *iov,
700 int iovcnt, uint64_t off, rbd_completion_t c);
701 CEPH_RBD_API int rbd_aio_read(rbd_image_t image, uint64_t off, size_t len,
702 char *buf, rbd_completion_t c);
704 * @param op_flags: see librados.h constants beginning with LIBRADOS_OP_FLAG
706 CEPH_RBD_API int rbd_aio_read2(rbd_image_t image, uint64_t off, size_t len,
707 char *buf, rbd_completion_t c, int op_flags);
708 CEPH_RBD_API int rbd_aio_readv(rbd_image_t image, const struct iovec *iov,
709 int iovcnt, uint64_t off, rbd_completion_t c);
710 CEPH_RBD_API int rbd_aio_discard(rbd_image_t image, uint64_t off, uint64_t len,
712 CEPH_RBD_API int rbd_aio_writesame(rbd_image_t image, uint64_t off, size_t len,
713 const char *buf, size_t data_len,
714 rbd_completion_t c, int op_flags);
715 CEPH_RBD_API ssize_t rbd_aio_compare_and_write(rbd_image_t image,
716 uint64_t off, size_t len,
717 const char *cmp_buf, const char *buf,
718 rbd_completion_t c, uint64_t *mismatch_off,
721 CEPH_RBD_API int rbd_aio_create_completion(void *cb_arg,
722 rbd_callback_t complete_cb,
723 rbd_completion_t *c);
724 CEPH_RBD_API int rbd_aio_is_complete(rbd_completion_t c);
725 CEPH_RBD_API int rbd_aio_wait_for_complete(rbd_completion_t c);
726 CEPH_RBD_API ssize_t rbd_aio_get_return_value(rbd_completion_t c);
727 CEPH_RBD_API void *rbd_aio_get_arg(rbd_completion_t c);
728 CEPH_RBD_API void rbd_aio_release(rbd_completion_t c);
729 CEPH_RBD_API int rbd_flush(rbd_image_t image);
731 * Start a flush if caching is enabled. Get a callback when
732 * the currently pending writes are on disk.
734 * @param image the image to flush writes to
735 * @param c what to call when flushing is complete
736 * @returns 0 on success, negative error code on failure
738 CEPH_RBD_API int rbd_aio_flush(rbd_image_t image, rbd_completion_t c);
741 * Drop any cached data for an image
743 * @param image the image to invalidate cached data for
744 * @returns 0 on success, negative error code on failure
746 CEPH_RBD_API int rbd_invalidate_cache(rbd_image_t image);
748 CEPH_RBD_API int rbd_poll_io_events(rbd_image_t image, rbd_completion_t *comps, int numcomp);
750 CEPH_RBD_API int rbd_metadata_get(rbd_image_t image, const char *key, char *value, size_t *val_len);
751 CEPH_RBD_API int rbd_metadata_set(rbd_image_t image, const char *key, const char *value);
752 CEPH_RBD_API int rbd_metadata_remove(rbd_image_t image, const char *key);
754 * List all metadatas associated with this image.
756 * This iterates over all metadatas, key_len and val_len are filled in
757 * with the number of bytes put into the keys and values buffers.
759 * If the provided buffers are too short, the required lengths are
760 * still filled in, but the data is not and -ERANGE is returned.
761 * Otherwise, the buffers are filled with the keys and values
762 * of the image, with a '\0' after each.
764 * @param image which image (and implicitly snapshot) to list clones of
765 * @param start_after which name to begin listing after
766 * (use the empty string to start at the beginning)
767 * @param max the maximum number of names to lis(if 0 means no limit)
768 * @param keys buffer in which to store pool names
769 * @param keys_len number of bytes in pools buffer
770 * @param values buffer in which to store image names
771 * @param vals_len number of bytes in images buffer
772 * @returns number of children on success, negative error code on failure
773 * @returns -ERANGE if either buffer is too short
775 CEPH_RBD_API int rbd_metadata_list(rbd_image_t image, const char *start, uint64_t max,
776 char *keys, size_t *key_len, char *values, size_t *vals_len);
778 // RBD image mirroring support functions
779 CEPH_RBD_API int rbd_mirror_image_enable(rbd_image_t image);
780 CEPH_RBD_API int rbd_mirror_image_disable(rbd_image_t image, bool force);
781 CEPH_RBD_API int rbd_mirror_image_promote(rbd_image_t image, bool force);
782 CEPH_RBD_API int rbd_mirror_image_demote(rbd_image_t image);
783 CEPH_RBD_API int rbd_mirror_image_resync(rbd_image_t image);
784 CEPH_RBD_API int rbd_mirror_image_get_info(rbd_image_t image,
785 rbd_mirror_image_info_t *mirror_image_info,
787 CEPH_RBD_API int rbd_mirror_image_get_status(rbd_image_t image,
788 rbd_mirror_image_status_t *mirror_image_status,
790 CEPH_RBD_API int rbd_aio_mirror_image_promote(rbd_image_t image, bool force,
792 CEPH_RBD_API int rbd_aio_mirror_image_demote(rbd_image_t image,
794 CEPH_RBD_API int rbd_aio_mirror_image_get_info(rbd_image_t image,
795 rbd_mirror_image_info_t *mirror_image_info,
798 CEPH_RBD_API int rbd_aio_mirror_image_get_status(rbd_image_t image,
799 rbd_mirror_image_status_t *mirror_image_status,
804 * Register an image metadata change watcher.
806 * @param image the image to watch
807 * @param handle where to store the internal id assigned to this watch
808 * @param watch_cb what to do when a notify is received on this image
809 * @param arg opaque value to pass to the callback
810 * @returns 0 on success, negative error code on failure
812 CEPH_RBD_API int rbd_update_watch(rbd_image_t image, uint64_t *handle,
813 rbd_update_callback_t watch_cb, void *arg);
816 * Unregister an image watcher.
818 * @param image the image to unwatch
819 * @param handle which watch to unregister
820 * @returns 0 on success, negative error code on failure
822 CEPH_RBD_API int rbd_update_unwatch(rbd_image_t image, uint64_t handle);