2 * Copyright (C) 2007 Oracle. All rights reserved.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
19 #ifndef __BTRFS_VOLUMES_
20 #define __BTRFS_VOLUMES_
22 #include <linux/bio.h>
23 #include <linux/sort.h>
24 #include <linux/btrfs.h>
25 #include "async-thread.h"
27 #define BTRFS_STRIPE_LEN (64 * 1024)
30 struct btrfs_pending_bios {
36 struct list_head dev_list;
37 struct list_head dev_alloc_list;
38 struct btrfs_fs_devices *fs_devices;
40 struct btrfs_root *dev_root;
42 struct rcu_string *name;
46 spinlock_t io_lock ____cacheline_aligned;
48 /* regular prio bios */
49 struct btrfs_pending_bios pending_bios;
51 struct btrfs_pending_bios pending_sync_bios;
53 struct block_device *bdev;
55 /* the mode sent to blkdev_get */
62 int is_tgtdev_for_dev_replace;
64 /* the internal btrfs device id */
67 /* size of the device */
70 /* size of the disk */
76 /* optimal io alignment for this device */
79 /* optimal io width for this device */
81 /* type and info about this device */
84 /* minimal io size for this device */
87 /* physical drive uuid (or lvm uuid) */
88 u8 uuid[BTRFS_UUID_SIZE];
90 /* for sending down flush barriers */
92 struct bio *flush_bio;
93 struct completion flush_wait;
95 /* per-device scrub information */
96 struct scrub_ctx *scrub_device;
98 struct btrfs_work work;
100 struct work_struct rcu_work;
102 /* readahead state */
103 spinlock_t reada_lock;
104 atomic_t reada_in_flight;
106 struct reada_zone *reada_curr_zone;
107 struct radix_tree_root reada_zones;
108 struct radix_tree_root reada_extents;
110 /* disk I/O failure stats. For detailed description refer to
111 * enum btrfs_dev_stat_values in ioctl.h */
114 /* Counter to record the change of device stats */
115 atomic_t dev_stats_ccnt;
116 atomic_t dev_stat_values[BTRFS_DEV_STAT_VALUES_MAX];
119 struct btrfs_fs_devices {
120 u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
122 /* the device with this id has the most recent copy of the super */
132 struct block_device *latest_bdev;
134 /* all of the devices in the FS, protected by a mutex
135 * so we can safely walk it to write out the supers without
136 * worrying about add/remove by the multi-device code.
137 * Scrubbing super can kick off supers writing by holding
140 struct mutex device_list_mutex;
141 struct list_head devices;
143 /* devices not currently being allocated */
144 struct list_head alloc_list;
145 struct list_head list;
147 struct btrfs_fs_devices *seed;
152 /* set when we find or add a device that doesn't have the
158 #define BTRFS_BIO_INLINE_CSUM_SIZE 64
161 * we need the mirror number and stripe index to be passed around
162 * the call chain while we are processing end_io (especially errors).
163 * Really, what we need is a btrfs_bio structure that has this info
164 * and is properly sized with its stripe array, but we're not there
165 * quite yet. We have our own btrfs bioset, and all of the bios
166 * we allocate are actually btrfs_io_bios. We'll cram as much of
167 * struct btrfs_bio as we can into this over time.
169 typedef void (btrfs_io_bio_end_io_t) (struct btrfs_io_bio *bio, int err);
170 struct btrfs_io_bio {
171 unsigned long mirror_num;
172 unsigned long stripe_index;
174 u8 csum_inline[BTRFS_BIO_INLINE_CSUM_SIZE];
176 btrfs_io_bio_end_io_t *end_io;
180 static inline struct btrfs_io_bio *btrfs_io_bio(struct bio *bio)
182 return container_of(bio, struct btrfs_io_bio, bio);
185 struct btrfs_bio_stripe {
186 struct btrfs_device *dev;
188 u64 length; /* only used for discard mappings */
192 typedef void (btrfs_bio_end_io_t) (struct btrfs_bio *bio, int err);
194 #define BTRFS_BIO_ORIG_BIO_SUBMITTED 0x1
197 atomic_t stripes_pending;
198 struct btrfs_fs_info *fs_info;
199 bio_end_io_t *end_io;
200 struct bio *orig_bio;
207 struct btrfs_bio_stripe stripes[];
210 struct btrfs_device_info {
211 struct btrfs_device *dev;
217 struct btrfs_raid_attr {
218 int sub_stripes; /* sub_stripes info for map */
219 int dev_stripes; /* stripes per dev */
220 int devs_max; /* max devs to use */
221 int devs_min; /* min devs needed */
222 int devs_increment; /* ndevs has to be a multiple of this */
223 int ncopies; /* how many copies to data has */
234 struct btrfs_bio_stripe stripes[];
237 #define map_lookup_size(n) (sizeof(struct map_lookup) + \
238 (sizeof(struct btrfs_bio_stripe) * (n)))
241 * Restriper's general type filter
243 #define BTRFS_BALANCE_DATA (1ULL << 0)
244 #define BTRFS_BALANCE_SYSTEM (1ULL << 1)
245 #define BTRFS_BALANCE_METADATA (1ULL << 2)
247 #define BTRFS_BALANCE_TYPE_MASK (BTRFS_BALANCE_DATA | \
248 BTRFS_BALANCE_SYSTEM | \
249 BTRFS_BALANCE_METADATA)
251 #define BTRFS_BALANCE_FORCE (1ULL << 3)
252 #define BTRFS_BALANCE_RESUME (1ULL << 4)
257 #define BTRFS_BALANCE_ARGS_PROFILES (1ULL << 0)
258 #define BTRFS_BALANCE_ARGS_USAGE (1ULL << 1)
259 #define BTRFS_BALANCE_ARGS_DEVID (1ULL << 2)
260 #define BTRFS_BALANCE_ARGS_DRANGE (1ULL << 3)
261 #define BTRFS_BALANCE_ARGS_VRANGE (1ULL << 4)
262 #define BTRFS_BALANCE_ARGS_LIMIT (1ULL << 5)
265 * Profile changing flags. When SOFT is set we won't relocate chunk if
266 * it already has the target profile (even though it may be
269 #define BTRFS_BALANCE_ARGS_CONVERT (1ULL << 8)
270 #define BTRFS_BALANCE_ARGS_SOFT (1ULL << 9)
272 struct btrfs_balance_args;
273 struct btrfs_balance_progress;
274 struct btrfs_balance_control {
275 struct btrfs_fs_info *fs_info;
277 struct btrfs_balance_args data;
278 struct btrfs_balance_args meta;
279 struct btrfs_balance_args sys;
283 struct btrfs_balance_progress stat;
286 int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
287 u64 end, u64 *length);
289 #define btrfs_bio_size(n) (sizeof(struct btrfs_bio) + \
290 (sizeof(struct btrfs_bio_stripe) * (n)))
292 int btrfs_map_block(struct btrfs_fs_info *fs_info, int rw,
293 u64 logical, u64 *length,
294 struct btrfs_bio **bbio_ret, int mirror_num);
295 int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
296 u64 chunk_start, u64 physical, u64 devid,
297 u64 **logical, int *naddrs, int *stripe_len);
298 int btrfs_read_sys_array(struct btrfs_root *root);
299 int btrfs_read_chunk_tree(struct btrfs_root *root);
300 int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
301 struct btrfs_root *extent_root, u64 type);
302 void btrfs_mapping_init(struct btrfs_mapping_tree *tree);
303 void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree);
304 int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
305 int mirror_num, int async_submit);
306 int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
307 fmode_t flags, void *holder);
308 int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
309 struct btrfs_fs_devices **fs_devices_ret);
310 int btrfs_close_devices(struct btrfs_fs_devices *fs_devices);
311 void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info,
312 struct btrfs_fs_devices *fs_devices, int step);
313 int btrfs_find_device_missing_or_by_path(struct btrfs_root *root,
315 struct btrfs_device **device);
316 struct btrfs_device *btrfs_alloc_device(struct btrfs_fs_info *fs_info,
319 int btrfs_rm_device(struct btrfs_root *root, char *device_path);
320 void btrfs_cleanup_fs_uuids(void);
321 int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len);
322 int btrfs_grow_device(struct btrfs_trans_handle *trans,
323 struct btrfs_device *device, u64 new_size);
324 struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid,
326 int btrfs_shrink_device(struct btrfs_device *device, u64 new_size);
327 int btrfs_init_new_device(struct btrfs_root *root, char *path);
328 int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
329 struct btrfs_device **device_out);
330 int btrfs_balance(struct btrfs_balance_control *bctl,
331 struct btrfs_ioctl_balance_args *bargs);
332 int btrfs_resume_balance_async(struct btrfs_fs_info *fs_info);
333 int btrfs_recover_balance(struct btrfs_fs_info *fs_info);
334 int btrfs_pause_balance(struct btrfs_fs_info *fs_info);
335 int btrfs_cancel_balance(struct btrfs_fs_info *fs_info);
336 int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info);
337 int btrfs_check_uuid_tree(struct btrfs_fs_info *fs_info);
338 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset);
339 int find_free_dev_extent(struct btrfs_trans_handle *trans,
340 struct btrfs_device *device, u64 num_bytes,
341 u64 *start, u64 *max_avail);
342 void btrfs_dev_stat_inc_and_print(struct btrfs_device *dev, int index);
343 int btrfs_get_dev_stats(struct btrfs_root *root,
344 struct btrfs_ioctl_get_dev_stats *stats);
345 void btrfs_init_devices_late(struct btrfs_fs_info *fs_info);
346 int btrfs_init_dev_stats(struct btrfs_fs_info *fs_info);
347 int btrfs_run_dev_stats(struct btrfs_trans_handle *trans,
348 struct btrfs_fs_info *fs_info);
349 void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
350 struct btrfs_device *srcdev);
351 void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
352 struct btrfs_device *tgtdev);
353 void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info,
354 struct btrfs_device *tgtdev);
355 int btrfs_scratch_superblock(struct btrfs_device *device);
356 int btrfs_is_parity_mirror(struct btrfs_mapping_tree *map_tree,
357 u64 logical, u64 len, int mirror_num);
358 unsigned long btrfs_full_stripe_len(struct btrfs_root *root,
359 struct btrfs_mapping_tree *map_tree,
361 int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
362 struct btrfs_root *extent_root,
363 u64 chunk_offset, u64 chunk_size);
365 static inline int btrfs_dev_stats_dirty(struct btrfs_device *dev)
367 return atomic_read(&dev->dev_stats_ccnt);
370 static inline void btrfs_dev_stat_inc(struct btrfs_device *dev,
373 atomic_inc(dev->dev_stat_values + index);
374 smp_mb__before_atomic();
375 atomic_inc(&dev->dev_stats_ccnt);
378 static inline int btrfs_dev_stat_read(struct btrfs_device *dev,
381 return atomic_read(dev->dev_stat_values + index);
384 static inline int btrfs_dev_stat_read_and_reset(struct btrfs_device *dev,
389 ret = atomic_xchg(dev->dev_stat_values + index, 0);
390 smp_mb__before_atomic();
391 atomic_inc(&dev->dev_stats_ccnt);
395 static inline void btrfs_dev_stat_set(struct btrfs_device *dev,
396 int index, unsigned long val)
398 atomic_set(dev->dev_stat_values + index, val);
399 smp_mb__before_atomic();
400 atomic_inc(&dev->dev_stats_ccnt);
403 static inline void btrfs_dev_stat_reset(struct btrfs_device *dev,
406 btrfs_dev_stat_set(dev, index, 0);