]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /* |
2 | * Copyright (C) 2003 Sistina Software Limited. | |
3 | * | |
4 | * This file is released under the GPL. | |
5 | */ | |
6 | ||
7 | #include "dm.h" | |
8 | #include "dm-bio-list.h" | |
06386bbf | 9 | #include "dm-bio-record.h" |
1da177e4 LT |
10 | #include "dm-io.h" |
11 | #include "dm-log.h" | |
12 | #include "kcopyd.h" | |
13 | ||
14 | #include <linux/ctype.h> | |
15 | #include <linux/init.h> | |
16 | #include <linux/mempool.h> | |
17 | #include <linux/module.h> | |
18 | #include <linux/pagemap.h> | |
19 | #include <linux/slab.h> | |
20 | #include <linux/time.h> | |
21 | #include <linux/vmalloc.h> | |
22 | #include <linux/workqueue.h> | |
6f3c3f0a | 23 | #include <linux/log2.h> |
72f4b314 | 24 | #include <linux/hardirq.h> |
1da177e4 | 25 | |
72d94861 | 26 | #define DM_MSG_PREFIX "raid1" |
88be163a | 27 | #define DM_IO_PAGES 64 |
72d94861 | 28 | |
a8e6afa2 | 29 | #define DM_RAID1_HANDLE_ERRORS 0x01 |
f44db678 | 30 | #define errors_handled(p) ((p)->features & DM_RAID1_HANDLE_ERRORS) |
a8e6afa2 | 31 | |
33184048 | 32 | static DECLARE_WAIT_QUEUE_HEAD(_kmirrord_recovery_stopped); |
1da177e4 | 33 | |
1da177e4 LT |
34 | /*----------------------------------------------------------------- |
35 | * Region hash | |
36 | * | |
37 | * The mirror splits itself up into discrete regions. Each | |
38 | * region can be in one of three states: clean, dirty, | |
39 | * nosync. There is no need to put clean regions in the hash. | |
40 | * | |
41 | * In addition to being present in the hash table a region _may_ | |
42 | * be present on one of three lists. | |
43 | * | |
44 | * clean_regions: Regions on this list have no io pending to | |
45 | * them, they are in sync, we are no longer interested in them, | |
46 | * they are dull. rh_update_states() will remove them from the | |
47 | * hash table. | |
48 | * | |
49 | * quiesced_regions: These regions have been spun down, ready | |
50 | * for recovery. rh_recovery_start() will remove regions from | |
51 | * this list and hand them to kmirrord, which will schedule the | |
52 | * recovery io with kcopyd. | |
53 | * | |
54 | * recovered_regions: Regions that kcopyd has successfully | |
55 | * recovered. rh_update_states() will now schedule any delayed | |
56 | * io, up the recovery_count, and remove the region from the | |
57 | * hash. | |
58 | * | |
59 | * There are 2 locks: | |
60 | * A rw spin lock 'hash_lock' protects just the hash table, | |
61 | * this is never held in write mode from interrupt context, | |
62 | * which I believe means that we only have to disable irqs when | |
63 | * doing a write lock. | |
64 | * | |
65 | * An ordinary spin lock 'region_lock' that protects the three | |
66 | * lists in the region_hash, with the 'state', 'list' and | |
67 | * 'bhs_delayed' fields of the regions. This is used from irq | |
68 | * context, so all other uses will have to suspend local irqs. | |
69 | *---------------------------------------------------------------*/ | |
70 | struct mirror_set; | |
71 | struct region_hash { | |
72 | struct mirror_set *ms; | |
73 | uint32_t region_size; | |
74 | unsigned region_shift; | |
75 | ||
76 | /* holds persistent region state */ | |
77 | struct dirty_log *log; | |
78 | ||
79 | /* hash table */ | |
80 | rwlock_t hash_lock; | |
81 | mempool_t *region_pool; | |
82 | unsigned int mask; | |
83 | unsigned int nr_buckets; | |
84 | struct list_head *buckets; | |
85 | ||
86 | spinlock_t region_lock; | |
33184048 | 87 | atomic_t recovery_in_flight; |
1da177e4 LT |
88 | struct semaphore recovery_count; |
89 | struct list_head clean_regions; | |
90 | struct list_head quiesced_regions; | |
91 | struct list_head recovered_regions; | |
f44db678 | 92 | struct list_head failed_recovered_regions; |
1da177e4 LT |
93 | }; |
94 | ||
95 | enum { | |
96 | RH_CLEAN, | |
97 | RH_DIRTY, | |
98 | RH_NOSYNC, | |
99 | RH_RECOVERING | |
100 | }; | |
101 | ||
102 | struct region { | |
103 | struct region_hash *rh; /* FIXME: can we get rid of this ? */ | |
104 | region_t key; | |
105 | int state; | |
106 | ||
107 | struct list_head hash_list; | |
108 | struct list_head list; | |
109 | ||
110 | atomic_t pending; | |
111 | struct bio_list delayed_bios; | |
112 | }; | |
113 | ||
e4c8b3ba NB |
114 | |
115 | /*----------------------------------------------------------------- | |
116 | * Mirror set structures. | |
117 | *---------------------------------------------------------------*/ | |
72f4b314 JB |
118 | enum dm_raid1_error { |
119 | DM_RAID1_WRITE_ERROR, | |
120 | DM_RAID1_SYNC_ERROR, | |
121 | DM_RAID1_READ_ERROR | |
122 | }; | |
123 | ||
e4c8b3ba | 124 | struct mirror { |
aa5617c5 | 125 | struct mirror_set *ms; |
e4c8b3ba | 126 | atomic_t error_count; |
39ed7adb | 127 | unsigned long error_type; |
e4c8b3ba NB |
128 | struct dm_dev *dev; |
129 | sector_t offset; | |
130 | }; | |
131 | ||
132 | struct mirror_set { | |
133 | struct dm_target *ti; | |
134 | struct list_head list; | |
135 | struct region_hash rh; | |
136 | struct kcopyd_client *kcopyd_client; | |
a8e6afa2 | 137 | uint64_t features; |
e4c8b3ba | 138 | |
72f4b314 | 139 | spinlock_t lock; /* protects the lists */ |
e4c8b3ba NB |
140 | struct bio_list reads; |
141 | struct bio_list writes; | |
72f4b314 | 142 | struct bio_list failures; |
e4c8b3ba | 143 | |
88be163a | 144 | struct dm_io_client *io_client; |
06386bbf | 145 | mempool_t *read_record_pool; |
88be163a | 146 | |
e4c8b3ba NB |
147 | /* recovery */ |
148 | region_t nr_regions; | |
149 | int in_sync; | |
fc1ff958 | 150 | int log_failure; |
b80aa7a0 | 151 | atomic_t suspend; |
e4c8b3ba | 152 | |
72f4b314 | 153 | atomic_t default_mirror; /* Default mirror */ |
e4c8b3ba | 154 | |
6ad36fe2 HS |
155 | struct workqueue_struct *kmirrord_wq; |
156 | struct work_struct kmirrord_work; | |
72f4b314 | 157 | struct work_struct trigger_event; |
6ad36fe2 | 158 | |
e4c8b3ba NB |
159 | unsigned int nr_mirrors; |
160 | struct mirror mirror[0]; | |
161 | }; | |
162 | ||
1da177e4 LT |
163 | /* |
164 | * Conversion fns | |
165 | */ | |
166 | static inline region_t bio_to_region(struct region_hash *rh, struct bio *bio) | |
167 | { | |
e4c8b3ba | 168 | return (bio->bi_sector - rh->ms->ti->begin) >> rh->region_shift; |
1da177e4 LT |
169 | } |
170 | ||
171 | static inline sector_t region_to_sector(struct region_hash *rh, region_t region) | |
172 | { | |
173 | return region << rh->region_shift; | |
174 | } | |
175 | ||
6ad36fe2 HS |
176 | static void wake(struct mirror_set *ms) |
177 | { | |
178 | queue_work(ms->kmirrord_wq, &ms->kmirrord_work); | |
179 | } | |
180 | ||
1da177e4 LT |
181 | /* FIXME move this */ |
182 | static void queue_bio(struct mirror_set *ms, struct bio *bio, int rw); | |
183 | ||
1da177e4 LT |
184 | #define MIN_REGIONS 64 |
185 | #define MAX_RECOVERY 1 | |
186 | static int rh_init(struct region_hash *rh, struct mirror_set *ms, | |
187 | struct dirty_log *log, uint32_t region_size, | |
188 | region_t nr_regions) | |
189 | { | |
190 | unsigned int nr_buckets, max_buckets; | |
191 | size_t i; | |
192 | ||
193 | /* | |
194 | * Calculate a suitable number of buckets for our hash | |
195 | * table. | |
196 | */ | |
197 | max_buckets = nr_regions >> 6; | |
198 | for (nr_buckets = 128u; nr_buckets < max_buckets; nr_buckets <<= 1) | |
199 | ; | |
200 | nr_buckets >>= 1; | |
201 | ||
202 | rh->ms = ms; | |
203 | rh->log = log; | |
204 | rh->region_size = region_size; | |
205 | rh->region_shift = ffs(region_size) - 1; | |
206 | rwlock_init(&rh->hash_lock); | |
207 | rh->mask = nr_buckets - 1; | |
208 | rh->nr_buckets = nr_buckets; | |
209 | ||
210 | rh->buckets = vmalloc(nr_buckets * sizeof(*rh->buckets)); | |
211 | if (!rh->buckets) { | |
212 | DMERR("unable to allocate region hash memory"); | |
213 | return -ENOMEM; | |
214 | } | |
215 | ||
216 | for (i = 0; i < nr_buckets; i++) | |
217 | INIT_LIST_HEAD(rh->buckets + i); | |
218 | ||
219 | spin_lock_init(&rh->region_lock); | |
220 | sema_init(&rh->recovery_count, 0); | |
33184048 | 221 | atomic_set(&rh->recovery_in_flight, 0); |
1da177e4 LT |
222 | INIT_LIST_HEAD(&rh->clean_regions); |
223 | INIT_LIST_HEAD(&rh->quiesced_regions); | |
224 | INIT_LIST_HEAD(&rh->recovered_regions); | |
f44db678 | 225 | INIT_LIST_HEAD(&rh->failed_recovered_regions); |
1da177e4 | 226 | |
0eaae62a MD |
227 | rh->region_pool = mempool_create_kmalloc_pool(MIN_REGIONS, |
228 | sizeof(struct region)); | |
1da177e4 LT |
229 | if (!rh->region_pool) { |
230 | vfree(rh->buckets); | |
231 | rh->buckets = NULL; | |
232 | return -ENOMEM; | |
233 | } | |
234 | ||
235 | return 0; | |
236 | } | |
237 | ||
238 | static void rh_exit(struct region_hash *rh) | |
239 | { | |
240 | unsigned int h; | |
241 | struct region *reg, *nreg; | |
242 | ||
243 | BUG_ON(!list_empty(&rh->quiesced_regions)); | |
244 | for (h = 0; h < rh->nr_buckets; h++) { | |
245 | list_for_each_entry_safe(reg, nreg, rh->buckets + h, hash_list) { | |
246 | BUG_ON(atomic_read(®->pending)); | |
247 | mempool_free(reg, rh->region_pool); | |
248 | } | |
249 | } | |
250 | ||
251 | if (rh->log) | |
252 | dm_destroy_dirty_log(rh->log); | |
253 | if (rh->region_pool) | |
254 | mempool_destroy(rh->region_pool); | |
255 | vfree(rh->buckets); | |
256 | } | |
257 | ||
258 | #define RH_HASH_MULT 2654435387U | |
259 | ||
260 | static inline unsigned int rh_hash(struct region_hash *rh, region_t region) | |
261 | { | |
262 | return (unsigned int) ((region * RH_HASH_MULT) >> 12) & rh->mask; | |
263 | } | |
264 | ||
265 | static struct region *__rh_lookup(struct region_hash *rh, region_t region) | |
266 | { | |
267 | struct region *reg; | |
268 | ||
269 | list_for_each_entry (reg, rh->buckets + rh_hash(rh, region), hash_list) | |
270 | if (reg->key == region) | |
271 | return reg; | |
272 | ||
273 | return NULL; | |
274 | } | |
275 | ||
276 | static void __rh_insert(struct region_hash *rh, struct region *reg) | |
277 | { | |
278 | unsigned int h = rh_hash(rh, reg->key); | |
279 | list_add(®->hash_list, rh->buckets + h); | |
280 | } | |
281 | ||
282 | static struct region *__rh_alloc(struct region_hash *rh, region_t region) | |
283 | { | |
284 | struct region *reg, *nreg; | |
285 | ||
286 | read_unlock(&rh->hash_lock); | |
c06aad85 DK |
287 | nreg = mempool_alloc(rh->region_pool, GFP_ATOMIC); |
288 | if (unlikely(!nreg)) | |
289 | nreg = kmalloc(sizeof(struct region), GFP_NOIO); | |
1da177e4 LT |
290 | nreg->state = rh->log->type->in_sync(rh->log, region, 1) ? |
291 | RH_CLEAN : RH_NOSYNC; | |
292 | nreg->rh = rh; | |
293 | nreg->key = region; | |
294 | ||
295 | INIT_LIST_HEAD(&nreg->list); | |
296 | ||
297 | atomic_set(&nreg->pending, 0); | |
298 | bio_list_init(&nreg->delayed_bios); | |
299 | write_lock_irq(&rh->hash_lock); | |
300 | ||
301 | reg = __rh_lookup(rh, region); | |
302 | if (reg) | |
303 | /* we lost the race */ | |
304 | mempool_free(nreg, rh->region_pool); | |
305 | ||
306 | else { | |
307 | __rh_insert(rh, nreg); | |
308 | if (nreg->state == RH_CLEAN) { | |
309 | spin_lock(&rh->region_lock); | |
310 | list_add(&nreg->list, &rh->clean_regions); | |
311 | spin_unlock(&rh->region_lock); | |
312 | } | |
313 | reg = nreg; | |
314 | } | |
315 | write_unlock_irq(&rh->hash_lock); | |
316 | read_lock(&rh->hash_lock); | |
317 | ||
318 | return reg; | |
319 | } | |
320 | ||
321 | static inline struct region *__rh_find(struct region_hash *rh, region_t region) | |
322 | { | |
323 | struct region *reg; | |
324 | ||
325 | reg = __rh_lookup(rh, region); | |
326 | if (!reg) | |
327 | reg = __rh_alloc(rh, region); | |
328 | ||
329 | return reg; | |
330 | } | |
331 | ||
332 | static int rh_state(struct region_hash *rh, region_t region, int may_block) | |
333 | { | |
334 | int r; | |
335 | struct region *reg; | |
336 | ||
337 | read_lock(&rh->hash_lock); | |
338 | reg = __rh_lookup(rh, region); | |
339 | read_unlock(&rh->hash_lock); | |
340 | ||
341 | if (reg) | |
342 | return reg->state; | |
343 | ||
344 | /* | |
345 | * The region wasn't in the hash, so we fall back to the | |
346 | * dirty log. | |
347 | */ | |
348 | r = rh->log->type->in_sync(rh->log, region, may_block); | |
349 | ||
350 | /* | |
351 | * Any error from the dirty log (eg. -EWOULDBLOCK) gets | |
352 | * taken as a RH_NOSYNC | |
353 | */ | |
354 | return r == 1 ? RH_CLEAN : RH_NOSYNC; | |
355 | } | |
356 | ||
357 | static inline int rh_in_sync(struct region_hash *rh, | |
358 | region_t region, int may_block) | |
359 | { | |
360 | int state = rh_state(rh, region, may_block); | |
361 | return state == RH_CLEAN || state == RH_DIRTY; | |
362 | } | |
363 | ||
364 | static void dispatch_bios(struct mirror_set *ms, struct bio_list *bio_list) | |
365 | { | |
366 | struct bio *bio; | |
367 | ||
368 | while ((bio = bio_list_pop(bio_list))) { | |
369 | queue_bio(ms, bio, WRITE); | |
370 | } | |
371 | } | |
372 | ||
f3ee6b2f JB |
373 | static void complete_resync_work(struct region *reg, int success) |
374 | { | |
375 | struct region_hash *rh = reg->rh; | |
376 | ||
377 | rh->log->type->set_region_sync(rh->log, reg->key, success); | |
b80aa7a0 JB |
378 | |
379 | /* | |
380 | * Dispatch the bios before we call 'wake_up_all'. | |
381 | * This is important because if we are suspending, | |
382 | * we want to know that recovery is complete and | |
383 | * the work queue is flushed. If we wake_up_all | |
384 | * before we dispatch_bios (queue bios and call wake()), | |
385 | * then we risk suspending before the work queue | |
386 | * has been properly flushed. | |
387 | */ | |
f3ee6b2f JB |
388 | dispatch_bios(rh->ms, ®->delayed_bios); |
389 | if (atomic_dec_and_test(&rh->recovery_in_flight)) | |
390 | wake_up_all(&_kmirrord_recovery_stopped); | |
391 | up(&rh->recovery_count); | |
392 | } | |
393 | ||
1da177e4 LT |
394 | static void rh_update_states(struct region_hash *rh) |
395 | { | |
396 | struct region *reg, *next; | |
397 | ||
398 | LIST_HEAD(clean); | |
399 | LIST_HEAD(recovered); | |
f44db678 | 400 | LIST_HEAD(failed_recovered); |
1da177e4 LT |
401 | |
402 | /* | |
403 | * Quickly grab the lists. | |
404 | */ | |
405 | write_lock_irq(&rh->hash_lock); | |
406 | spin_lock(&rh->region_lock); | |
407 | if (!list_empty(&rh->clean_regions)) { | |
c12bfc92 | 408 | list_splice_init(&rh->clean_regions, &clean); |
1da177e4 | 409 | |
943317ef | 410 | list_for_each_entry(reg, &clean, list) |
1da177e4 | 411 | list_del(®->hash_list); |
1da177e4 LT |
412 | } |
413 | ||
414 | if (!list_empty(&rh->recovered_regions)) { | |
c12bfc92 | 415 | list_splice_init(&rh->recovered_regions, &recovered); |
1da177e4 LT |
416 | |
417 | list_for_each_entry (reg, &recovered, list) | |
418 | list_del(®->hash_list); | |
419 | } | |
f44db678 JB |
420 | |
421 | if (!list_empty(&rh->failed_recovered_regions)) { | |
c12bfc92 RD |
422 | list_splice_init(&rh->failed_recovered_regions, |
423 | &failed_recovered); | |
f44db678 JB |
424 | |
425 | list_for_each_entry(reg, &failed_recovered, list) | |
426 | list_del(®->hash_list); | |
427 | } | |
428 | ||
1da177e4 LT |
429 | spin_unlock(&rh->region_lock); |
430 | write_unlock_irq(&rh->hash_lock); | |
431 | ||
432 | /* | |
433 | * All the regions on the recovered and clean lists have | |
434 | * now been pulled out of the system, so no need to do | |
435 | * any more locking. | |
436 | */ | |
437 | list_for_each_entry_safe (reg, next, &recovered, list) { | |
438 | rh->log->type->clear_region(rh->log, reg->key); | |
f3ee6b2f | 439 | complete_resync_work(reg, 1); |
1da177e4 LT |
440 | mempool_free(reg, rh->region_pool); |
441 | } | |
442 | ||
f44db678 JB |
443 | list_for_each_entry_safe(reg, next, &failed_recovered, list) { |
444 | complete_resync_work(reg, errors_handled(rh->ms) ? 0 : 1); | |
445 | mempool_free(reg, rh->region_pool); | |
446 | } | |
447 | ||
943317ef JB |
448 | list_for_each_entry_safe(reg, next, &clean, list) { |
449 | rh->log->type->clear_region(rh->log, reg->key); | |
1da177e4 | 450 | mempool_free(reg, rh->region_pool); |
943317ef JB |
451 | } |
452 | ||
453 | rh->log->type->flush(rh->log); | |
1da177e4 LT |
454 | } |
455 | ||
456 | static void rh_inc(struct region_hash *rh, region_t region) | |
457 | { | |
458 | struct region *reg; | |
459 | ||
460 | read_lock(&rh->hash_lock); | |
461 | reg = __rh_find(rh, region); | |
844e8d90 | 462 | |
7692c5dd | 463 | spin_lock_irq(&rh->region_lock); |
844e8d90 JN |
464 | atomic_inc(®->pending); |
465 | ||
1da177e4 | 466 | if (reg->state == RH_CLEAN) { |
1da177e4 LT |
467 | reg->state = RH_DIRTY; |
468 | list_del_init(®->list); /* take off the clean list */ | |
7692c5dd JB |
469 | spin_unlock_irq(&rh->region_lock); |
470 | ||
471 | rh->log->type->mark_region(rh->log, reg->key); | |
472 | } else | |
473 | spin_unlock_irq(&rh->region_lock); | |
474 | ||
1da177e4 | 475 | |
1da177e4 LT |
476 | read_unlock(&rh->hash_lock); |
477 | } | |
478 | ||
479 | static void rh_inc_pending(struct region_hash *rh, struct bio_list *bios) | |
480 | { | |
481 | struct bio *bio; | |
482 | ||
483 | for (bio = bios->head; bio; bio = bio->bi_next) | |
484 | rh_inc(rh, bio_to_region(rh, bio)); | |
485 | } | |
486 | ||
487 | static void rh_dec(struct region_hash *rh, region_t region) | |
488 | { | |
489 | unsigned long flags; | |
490 | struct region *reg; | |
491 | int should_wake = 0; | |
492 | ||
493 | read_lock(&rh->hash_lock); | |
494 | reg = __rh_lookup(rh, region); | |
495 | read_unlock(&rh->hash_lock); | |
496 | ||
7692c5dd | 497 | spin_lock_irqsave(&rh->region_lock, flags); |
1da177e4 | 498 | if (atomic_dec_and_test(®->pending)) { |
930d332a JN |
499 | /* |
500 | * There is no pending I/O for this region. | |
501 | * We can move the region to corresponding list for next action. | |
502 | * At this point, the region is not yet connected to any list. | |
503 | * | |
504 | * If the state is RH_NOSYNC, the region should be kept off | |
505 | * from clean list. | |
506 | * The hash entry for RH_NOSYNC will remain in memory | |
507 | * until the region is recovered or the map is reloaded. | |
508 | */ | |
509 | ||
510 | /* do nothing for RH_NOSYNC */ | |
1da177e4 LT |
511 | if (reg->state == RH_RECOVERING) { |
512 | list_add_tail(®->list, &rh->quiesced_regions); | |
930d332a | 513 | } else if (reg->state == RH_DIRTY) { |
1da177e4 LT |
514 | reg->state = RH_CLEAN; |
515 | list_add(®->list, &rh->clean_regions); | |
516 | } | |
1da177e4 LT |
517 | should_wake = 1; |
518 | } | |
7692c5dd | 519 | spin_unlock_irqrestore(&rh->region_lock, flags); |
1da177e4 LT |
520 | |
521 | if (should_wake) | |
6ad36fe2 | 522 | wake(rh->ms); |
1da177e4 LT |
523 | } |
524 | ||
525 | /* | |
526 | * Starts quiescing a region in preparation for recovery. | |
527 | */ | |
528 | static int __rh_recovery_prepare(struct region_hash *rh) | |
529 | { | |
530 | int r; | |
531 | struct region *reg; | |
532 | region_t region; | |
533 | ||
534 | /* | |
535 | * Ask the dirty log what's next. | |
536 | */ | |
537 | r = rh->log->type->get_resync_work(rh->log, ®ion); | |
538 | if (r <= 0) | |
539 | return r; | |
540 | ||
541 | /* | |
542 | * Get this region, and start it quiescing by setting the | |
543 | * recovering flag. | |
544 | */ | |
545 | read_lock(&rh->hash_lock); | |
546 | reg = __rh_find(rh, region); | |
547 | read_unlock(&rh->hash_lock); | |
548 | ||
549 | spin_lock_irq(&rh->region_lock); | |
550 | reg->state = RH_RECOVERING; | |
551 | ||
552 | /* Already quiesced ? */ | |
553 | if (atomic_read(®->pending)) | |
554 | list_del_init(®->list); | |
179e0917 AM |
555 | else |
556 | list_move(®->list, &rh->quiesced_regions); | |
1da177e4 | 557 | |
1da177e4 LT |
558 | spin_unlock_irq(&rh->region_lock); |
559 | ||
560 | return 1; | |
561 | } | |
562 | ||
563 | static void rh_recovery_prepare(struct region_hash *rh) | |
564 | { | |
33184048 JB |
565 | /* Extra reference to avoid race with rh_stop_recovery */ |
566 | atomic_inc(&rh->recovery_in_flight); | |
567 | ||
568 | while (!down_trylock(&rh->recovery_count)) { | |
569 | atomic_inc(&rh->recovery_in_flight); | |
1da177e4 | 570 | if (__rh_recovery_prepare(rh) <= 0) { |
33184048 | 571 | atomic_dec(&rh->recovery_in_flight); |
1da177e4 LT |
572 | up(&rh->recovery_count); |
573 | break; | |
574 | } | |
33184048 JB |
575 | } |
576 | ||
577 | /* Drop the extra reference */ | |
578 | if (atomic_dec_and_test(&rh->recovery_in_flight)) | |
579 | wake_up_all(&_kmirrord_recovery_stopped); | |
1da177e4 LT |
580 | } |
581 | ||
582 | /* | |
583 | * Returns any quiesced regions. | |
584 | */ | |
585 | static struct region *rh_recovery_start(struct region_hash *rh) | |
586 | { | |
587 | struct region *reg = NULL; | |
588 | ||
589 | spin_lock_irq(&rh->region_lock); | |
590 | if (!list_empty(&rh->quiesced_regions)) { | |
591 | reg = list_entry(rh->quiesced_regions.next, | |
592 | struct region, list); | |
593 | list_del_init(®->list); /* remove from the quiesced list */ | |
594 | } | |
595 | spin_unlock_irq(&rh->region_lock); | |
596 | ||
597 | return reg; | |
598 | } | |
599 | ||
1da177e4 LT |
600 | static void rh_recovery_end(struct region *reg, int success) |
601 | { | |
602 | struct region_hash *rh = reg->rh; | |
603 | ||
604 | spin_lock_irq(&rh->region_lock); | |
f44db678 JB |
605 | if (success) |
606 | list_add(®->list, ®->rh->recovered_regions); | |
607 | else { | |
608 | reg->state = RH_NOSYNC; | |
609 | list_add(®->list, ®->rh->failed_recovered_regions); | |
610 | } | |
1da177e4 LT |
611 | spin_unlock_irq(&rh->region_lock); |
612 | ||
6ad36fe2 | 613 | wake(rh->ms); |
1da177e4 LT |
614 | } |
615 | ||
fc1ff958 | 616 | static int rh_flush(struct region_hash *rh) |
1da177e4 | 617 | { |
fc1ff958 | 618 | return rh->log->type->flush(rh->log); |
1da177e4 LT |
619 | } |
620 | ||
621 | static void rh_delay(struct region_hash *rh, struct bio *bio) | |
622 | { | |
623 | struct region *reg; | |
624 | ||
625 | read_lock(&rh->hash_lock); | |
626 | reg = __rh_find(rh, bio_to_region(rh, bio)); | |
627 | bio_list_add(®->delayed_bios, bio); | |
628 | read_unlock(&rh->hash_lock); | |
629 | } | |
630 | ||
631 | static void rh_stop_recovery(struct region_hash *rh) | |
632 | { | |
633 | int i; | |
634 | ||
635 | /* wait for any recovering regions */ | |
636 | for (i = 0; i < MAX_RECOVERY; i++) | |
637 | down(&rh->recovery_count); | |
638 | } | |
639 | ||
640 | static void rh_start_recovery(struct region_hash *rh) | |
641 | { | |
642 | int i; | |
643 | ||
644 | for (i = 0; i < MAX_RECOVERY; i++) | |
645 | up(&rh->recovery_count); | |
646 | ||
6ad36fe2 | 647 | wake(rh->ms); |
1da177e4 LT |
648 | } |
649 | ||
06386bbf JB |
650 | #define MIN_READ_RECORDS 20 |
651 | struct dm_raid1_read_record { | |
652 | struct mirror *m; | |
653 | struct dm_bio_details details; | |
654 | }; | |
655 | ||
1da177e4 LT |
656 | /* |
657 | * Every mirror should look like this one. | |
658 | */ | |
659 | #define DEFAULT_MIRROR 0 | |
660 | ||
661 | /* | |
06386bbf JB |
662 | * This is yucky. We squirrel the mirror struct away inside |
663 | * bi_next for read/write buffers. This is safe since the bh | |
1da177e4 LT |
664 | * doesn't get submitted to the lower levels of block layer. |
665 | */ | |
06386bbf | 666 | static struct mirror *bio_get_m(struct bio *bio) |
1da177e4 | 667 | { |
06386bbf | 668 | return (struct mirror *) bio->bi_next; |
1da177e4 LT |
669 | } |
670 | ||
06386bbf | 671 | static void bio_set_m(struct bio *bio, struct mirror *m) |
1da177e4 | 672 | { |
06386bbf | 673 | bio->bi_next = (struct bio *) m; |
1da177e4 LT |
674 | } |
675 | ||
72f4b314 JB |
676 | static struct mirror *get_default_mirror(struct mirror_set *ms) |
677 | { | |
678 | return &ms->mirror[atomic_read(&ms->default_mirror)]; | |
679 | } | |
680 | ||
681 | static void set_default_mirror(struct mirror *m) | |
682 | { | |
683 | struct mirror_set *ms = m->ms; | |
684 | struct mirror *m0 = &(ms->mirror[0]); | |
685 | ||
686 | atomic_set(&ms->default_mirror, m - m0); | |
687 | } | |
688 | ||
689 | /* fail_mirror | |
690 | * @m: mirror device to fail | |
691 | * @error_type: one of the enum's, DM_RAID1_*_ERROR | |
692 | * | |
693 | * If errors are being handled, record the type of | |
694 | * error encountered for this device. If this type | |
695 | * of error has already been recorded, we can return; | |
696 | * otherwise, we must signal userspace by triggering | |
697 | * an event. Additionally, if the device is the | |
698 | * primary device, we must choose a new primary, but | |
699 | * only if the mirror is in-sync. | |
700 | * | |
701 | * This function must not block. | |
702 | */ | |
703 | static void fail_mirror(struct mirror *m, enum dm_raid1_error error_type) | |
704 | { | |
705 | struct mirror_set *ms = m->ms; | |
706 | struct mirror *new; | |
707 | ||
708 | if (!errors_handled(ms)) | |
709 | return; | |
710 | ||
711 | /* | |
712 | * error_count is used for nothing more than a | |
713 | * simple way to tell if a device has encountered | |
714 | * errors. | |
715 | */ | |
716 | atomic_inc(&m->error_count); | |
717 | ||
718 | if (test_and_set_bit(error_type, &m->error_type)) | |
719 | return; | |
720 | ||
721 | if (m != get_default_mirror(ms)) | |
722 | goto out; | |
723 | ||
724 | if (!ms->in_sync) { | |
725 | /* | |
726 | * Better to issue requests to same failing device | |
727 | * than to risk returning corrupt data. | |
728 | */ | |
729 | DMERR("Primary mirror (%s) failed while out-of-sync: " | |
730 | "Reads may fail.", m->dev->name); | |
731 | goto out; | |
732 | } | |
733 | ||
734 | for (new = ms->mirror; new < ms->mirror + ms->nr_mirrors; new++) | |
735 | if (!atomic_read(&new->error_count)) { | |
736 | set_default_mirror(new); | |
737 | break; | |
738 | } | |
739 | ||
740 | if (unlikely(new == ms->mirror + ms->nr_mirrors)) | |
741 | DMWARN("All sides of mirror have failed."); | |
742 | ||
743 | out: | |
744 | schedule_work(&ms->trigger_event); | |
745 | } | |
746 | ||
1da177e4 LT |
747 | /*----------------------------------------------------------------- |
748 | * Recovery. | |
749 | * | |
750 | * When a mirror is first activated we may find that some regions | |
751 | * are in the no-sync state. We have to recover these by | |
752 | * recopying from the default mirror to all the others. | |
753 | *---------------------------------------------------------------*/ | |
4cdc1d1f | 754 | static void recovery_complete(int read_err, unsigned long write_err, |
1da177e4 LT |
755 | void *context) |
756 | { | |
8f0205b7 JB |
757 | struct region *reg = (struct region *)context; |
758 | struct mirror_set *ms = reg->rh->ms; | |
759 | int m, bit = 0; | |
1da177e4 | 760 | |
8f0205b7 | 761 | if (read_err) { |
f44db678 JB |
762 | /* Read error means the failure of default mirror. */ |
763 | DMERR_LIMIT("Unable to read primary mirror during recovery"); | |
8f0205b7 JB |
764 | fail_mirror(get_default_mirror(ms), DM_RAID1_SYNC_ERROR); |
765 | } | |
f44db678 | 766 | |
8f0205b7 | 767 | if (write_err) { |
4cdc1d1f | 768 | DMERR_LIMIT("Write error during recovery (error = 0x%lx)", |
f44db678 | 769 | write_err); |
8f0205b7 JB |
770 | /* |
771 | * Bits correspond to devices (excluding default mirror). | |
772 | * The default mirror cannot change during recovery. | |
773 | */ | |
774 | for (m = 0; m < ms->nr_mirrors; m++) { | |
775 | if (&ms->mirror[m] == get_default_mirror(ms)) | |
776 | continue; | |
777 | if (test_bit(bit, &write_err)) | |
778 | fail_mirror(ms->mirror + m, | |
779 | DM_RAID1_SYNC_ERROR); | |
780 | bit++; | |
781 | } | |
782 | } | |
f44db678 | 783 | |
ce503f59 | 784 | rh_recovery_end(reg, !(read_err || write_err)); |
1da177e4 LT |
785 | } |
786 | ||
787 | static int recover(struct mirror_set *ms, struct region *reg) | |
788 | { | |
789 | int r; | |
790 | unsigned int i; | |
22a1ceb1 | 791 | struct dm_io_region from, to[KCOPYD_MAX_REGIONS], *dest; |
1da177e4 LT |
792 | struct mirror *m; |
793 | unsigned long flags = 0; | |
794 | ||
795 | /* fill in the source */ | |
72f4b314 | 796 | m = get_default_mirror(ms); |
1da177e4 LT |
797 | from.bdev = m->dev->bdev; |
798 | from.sector = m->offset + region_to_sector(reg->rh, reg->key); | |
799 | if (reg->key == (ms->nr_regions - 1)) { | |
800 | /* | |
801 | * The final region may be smaller than | |
802 | * region_size. | |
803 | */ | |
804 | from.count = ms->ti->len & (reg->rh->region_size - 1); | |
805 | if (!from.count) | |
806 | from.count = reg->rh->region_size; | |
807 | } else | |
808 | from.count = reg->rh->region_size; | |
809 | ||
810 | /* fill in the destinations */ | |
811 | for (i = 0, dest = to; i < ms->nr_mirrors; i++) { | |
72f4b314 | 812 | if (&ms->mirror[i] == get_default_mirror(ms)) |
1da177e4 LT |
813 | continue; |
814 | ||
815 | m = ms->mirror + i; | |
816 | dest->bdev = m->dev->bdev; | |
817 | dest->sector = m->offset + region_to_sector(reg->rh, reg->key); | |
818 | dest->count = from.count; | |
819 | dest++; | |
820 | } | |
821 | ||
822 | /* hand to kcopyd */ | |
823 | set_bit(KCOPYD_IGNORE_ERROR, &flags); | |
824 | r = kcopyd_copy(ms->kcopyd_client, &from, ms->nr_mirrors - 1, to, flags, | |
825 | recovery_complete, reg); | |
826 | ||
827 | return r; | |
828 | } | |
829 | ||
830 | static void do_recovery(struct mirror_set *ms) | |
831 | { | |
832 | int r; | |
833 | struct region *reg; | |
834 | struct dirty_log *log = ms->rh.log; | |
835 | ||
836 | /* | |
837 | * Start quiescing some regions. | |
838 | */ | |
839 | rh_recovery_prepare(&ms->rh); | |
840 | ||
841 | /* | |
842 | * Copy any already quiesced regions. | |
843 | */ | |
844 | while ((reg = rh_recovery_start(&ms->rh))) { | |
845 | r = recover(ms, reg); | |
846 | if (r) | |
847 | rh_recovery_end(reg, 0); | |
848 | } | |
849 | ||
850 | /* | |
851 | * Update the in sync flag. | |
852 | */ | |
853 | if (!ms->in_sync && | |
854 | (log->type->get_sync_count(log) == ms->nr_regions)) { | |
855 | /* the sync is complete */ | |
856 | dm_table_event(ms->ti->table); | |
857 | ms->in_sync = 1; | |
858 | } | |
859 | } | |
860 | ||
861 | /*----------------------------------------------------------------- | |
862 | * Reads | |
863 | *---------------------------------------------------------------*/ | |
864 | static struct mirror *choose_mirror(struct mirror_set *ms, sector_t sector) | |
865 | { | |
06386bbf JB |
866 | struct mirror *m = get_default_mirror(ms); |
867 | ||
868 | do { | |
869 | if (likely(!atomic_read(&m->error_count))) | |
870 | return m; | |
871 | ||
872 | if (m-- == ms->mirror) | |
873 | m += ms->nr_mirrors; | |
874 | } while (m != get_default_mirror(ms)); | |
875 | ||
876 | return NULL; | |
877 | } | |
878 | ||
879 | static int default_ok(struct mirror *m) | |
880 | { | |
881 | struct mirror *default_mirror = get_default_mirror(m->ms); | |
882 | ||
883 | return !atomic_read(&default_mirror->error_count); | |
884 | } | |
885 | ||
886 | static int mirror_available(struct mirror_set *ms, struct bio *bio) | |
887 | { | |
888 | region_t region = bio_to_region(&ms->rh, bio); | |
889 | ||
890 | if (ms->rh.log->type->in_sync(ms->rh.log, region, 0)) | |
891 | return choose_mirror(ms, bio->bi_sector) ? 1 : 0; | |
892 | ||
893 | return 0; | |
1da177e4 LT |
894 | } |
895 | ||
896 | /* | |
897 | * remap a buffer to a particular mirror. | |
898 | */ | |
06386bbf JB |
899 | static sector_t map_sector(struct mirror *m, struct bio *bio) |
900 | { | |
901 | return m->offset + (bio->bi_sector - m->ms->ti->begin); | |
902 | } | |
903 | ||
904 | static void map_bio(struct mirror *m, struct bio *bio) | |
1da177e4 LT |
905 | { |
906 | bio->bi_bdev = m->dev->bdev; | |
06386bbf JB |
907 | bio->bi_sector = map_sector(m, bio); |
908 | } | |
909 | ||
22a1ceb1 | 910 | static void map_region(struct dm_io_region *io, struct mirror *m, |
06386bbf JB |
911 | struct bio *bio) |
912 | { | |
913 | io->bdev = m->dev->bdev; | |
914 | io->sector = map_sector(m, bio); | |
915 | io->count = bio->bi_size >> 9; | |
916 | } | |
917 | ||
918 | /*----------------------------------------------------------------- | |
919 | * Reads | |
920 | *---------------------------------------------------------------*/ | |
921 | static void read_callback(unsigned long error, void *context) | |
922 | { | |
923 | struct bio *bio = context; | |
924 | struct mirror *m; | |
925 | ||
926 | m = bio_get_m(bio); | |
927 | bio_set_m(bio, NULL); | |
928 | ||
929 | if (likely(!error)) { | |
930 | bio_endio(bio, 0); | |
931 | return; | |
932 | } | |
933 | ||
934 | fail_mirror(m, DM_RAID1_READ_ERROR); | |
935 | ||
936 | if (likely(default_ok(m)) || mirror_available(m->ms, bio)) { | |
937 | DMWARN_LIMIT("Read failure on mirror device %s. " | |
938 | "Trying alternative device.", | |
939 | m->dev->name); | |
940 | queue_bio(m->ms, bio, bio_rw(bio)); | |
941 | return; | |
942 | } | |
943 | ||
944 | DMERR_LIMIT("Read failure on mirror device %s. Failing I/O.", | |
945 | m->dev->name); | |
946 | bio_endio(bio, -EIO); | |
947 | } | |
948 | ||
949 | /* Asynchronous read. */ | |
950 | static void read_async_bio(struct mirror *m, struct bio *bio) | |
951 | { | |
22a1ceb1 | 952 | struct dm_io_region io; |
06386bbf JB |
953 | struct dm_io_request io_req = { |
954 | .bi_rw = READ, | |
955 | .mem.type = DM_IO_BVEC, | |
956 | .mem.ptr.bvec = bio->bi_io_vec + bio->bi_idx, | |
957 | .notify.fn = read_callback, | |
958 | .notify.context = bio, | |
959 | .client = m->ms->io_client, | |
960 | }; | |
961 | ||
962 | map_region(&io, m, bio); | |
963 | bio_set_m(bio, m); | |
964 | (void) dm_io(&io_req, 1, &io, NULL); | |
1da177e4 LT |
965 | } |
966 | ||
967 | static void do_reads(struct mirror_set *ms, struct bio_list *reads) | |
968 | { | |
969 | region_t region; | |
970 | struct bio *bio; | |
971 | struct mirror *m; | |
972 | ||
973 | while ((bio = bio_list_pop(reads))) { | |
974 | region = bio_to_region(&ms->rh, bio); | |
06386bbf | 975 | m = get_default_mirror(ms); |
1da177e4 LT |
976 | |
977 | /* | |
978 | * We can only read balance if the region is in sync. | |
979 | */ | |
06386bbf | 980 | if (likely(rh_in_sync(&ms->rh, region, 1))) |
1da177e4 | 981 | m = choose_mirror(ms, bio->bi_sector); |
06386bbf JB |
982 | else if (m && atomic_read(&m->error_count)) |
983 | m = NULL; | |
1da177e4 | 984 | |
06386bbf JB |
985 | if (likely(m)) |
986 | read_async_bio(m, bio); | |
987 | else | |
988 | bio_endio(bio, -EIO); | |
1da177e4 LT |
989 | } |
990 | } | |
991 | ||
992 | /*----------------------------------------------------------------- | |
993 | * Writes. | |
994 | * | |
995 | * We do different things with the write io depending on the | |
996 | * state of the region that it's in: | |
997 | * | |
998 | * SYNC: increment pending, use kcopyd to write to *all* mirrors | |
999 | * RECOVERING: delay the io until recovery completes | |
1000 | * NOSYNC: increment pending, just write to the default mirror | |
1001 | *---------------------------------------------------------------*/ | |
72f4b314 JB |
1002 | |
1003 | /* __bio_mark_nosync | |
1004 | * @ms | |
1005 | * @bio | |
1006 | * @done | |
1007 | * @error | |
1008 | * | |
1009 | * The bio was written on some mirror(s) but failed on other mirror(s). | |
1010 | * We can successfully endio the bio but should avoid the region being | |
1011 | * marked clean by setting the state RH_NOSYNC. | |
1012 | * | |
1013 | * This function is _not_ safe in interrupt context! | |
1014 | */ | |
1015 | static void __bio_mark_nosync(struct mirror_set *ms, | |
1016 | struct bio *bio, unsigned done, int error) | |
1017 | { | |
1018 | unsigned long flags; | |
1019 | struct region_hash *rh = &ms->rh; | |
1020 | struct dirty_log *log = ms->rh.log; | |
1021 | struct region *reg; | |
1022 | region_t region = bio_to_region(rh, bio); | |
1023 | int recovering = 0; | |
1024 | ||
1025 | /* We must inform the log that the sync count has changed. */ | |
1026 | log->type->set_region_sync(log, region, 0); | |
1027 | ms->in_sync = 0; | |
1028 | ||
1029 | read_lock(&rh->hash_lock); | |
1030 | reg = __rh_find(rh, region); | |
1031 | read_unlock(&rh->hash_lock); | |
1032 | ||
1033 | /* region hash entry should exist because write was in-flight */ | |
1034 | BUG_ON(!reg); | |
1035 | BUG_ON(!list_empty(®->list)); | |
1036 | ||
1037 | spin_lock_irqsave(&rh->region_lock, flags); | |
1038 | /* | |
1039 | * Possible cases: | |
1040 | * 1) RH_DIRTY | |
1041 | * 2) RH_NOSYNC: was dirty, other preceeding writes failed | |
1042 | * 3) RH_RECOVERING: flushing pending writes | |
1043 | * Either case, the region should have not been connected to list. | |
1044 | */ | |
1045 | recovering = (reg->state == RH_RECOVERING); | |
1046 | reg->state = RH_NOSYNC; | |
1047 | BUG_ON(!list_empty(®->list)); | |
1048 | spin_unlock_irqrestore(&rh->region_lock, flags); | |
1049 | ||
1050 | bio_endio(bio, error); | |
1051 | if (recovering) | |
1052 | complete_resync_work(reg, 0); | |
1053 | } | |
1054 | ||
1da177e4 LT |
1055 | static void write_callback(unsigned long error, void *context) |
1056 | { | |
72f4b314 | 1057 | unsigned i, ret = 0; |
1da177e4 LT |
1058 | struct bio *bio = (struct bio *) context; |
1059 | struct mirror_set *ms; | |
72f4b314 JB |
1060 | int uptodate = 0; |
1061 | int should_wake = 0; | |
1062 | unsigned long flags; | |
1da177e4 | 1063 | |
06386bbf JB |
1064 | ms = bio_get_m(bio)->ms; |
1065 | bio_set_m(bio, NULL); | |
1da177e4 LT |
1066 | |
1067 | /* | |
1068 | * NOTE: We don't decrement the pending count here, | |
1069 | * instead it is done by the targets endio function. | |
1070 | * This way we handle both writes to SYNC and NOSYNC | |
1071 | * regions with the same code. | |
1072 | */ | |
72f4b314 JB |
1073 | if (likely(!error)) |
1074 | goto out; | |
1da177e4 | 1075 | |
72f4b314 JB |
1076 | for (i = 0; i < ms->nr_mirrors; i++) |
1077 | if (test_bit(i, &error)) | |
1078 | fail_mirror(ms->mirror + i, DM_RAID1_WRITE_ERROR); | |
1079 | else | |
1080 | uptodate = 1; | |
1081 | ||
1082 | if (unlikely(!uptodate)) { | |
1083 | DMERR("All replicated volumes dead, failing I/O"); | |
1084 | /* None of the writes succeeded, fail the I/O. */ | |
1085 | ret = -EIO; | |
1086 | } else if (errors_handled(ms)) { | |
1da177e4 | 1087 | /* |
72f4b314 JB |
1088 | * Need to raise event. Since raising |
1089 | * events can block, we need to do it in | |
1090 | * the main thread. | |
1da177e4 | 1091 | */ |
72f4b314 JB |
1092 | spin_lock_irqsave(&ms->lock, flags); |
1093 | if (!ms->failures.head) | |
1094 | should_wake = 1; | |
1095 | bio_list_add(&ms->failures, bio); | |
1096 | spin_unlock_irqrestore(&ms->lock, flags); | |
1097 | if (should_wake) | |
1098 | wake(ms); | |
1099 | return; | |
1da177e4 | 1100 | } |
72f4b314 JB |
1101 | out: |
1102 | bio_endio(bio, ret); | |
1da177e4 LT |
1103 | } |
1104 | ||
1105 | static void do_write(struct mirror_set *ms, struct bio *bio) | |
1106 | { | |
1107 | unsigned int i; | |
22a1ceb1 | 1108 | struct dm_io_region io[ms->nr_mirrors], *dest = io; |
1da177e4 | 1109 | struct mirror *m; |
88be163a MB |
1110 | struct dm_io_request io_req = { |
1111 | .bi_rw = WRITE, | |
1112 | .mem.type = DM_IO_BVEC, | |
1113 | .mem.ptr.bvec = bio->bi_io_vec + bio->bi_idx, | |
1114 | .notify.fn = write_callback, | |
1115 | .notify.context = bio, | |
1116 | .client = ms->io_client, | |
1117 | }; | |
1da177e4 | 1118 | |
06386bbf JB |
1119 | for (i = 0, m = ms->mirror; i < ms->nr_mirrors; i++, m++) |
1120 | map_region(dest++, m, bio); | |
1da177e4 | 1121 | |
06386bbf JB |
1122 | /* |
1123 | * Use default mirror because we only need it to retrieve the reference | |
1124 | * to the mirror set in write_callback(). | |
1125 | */ | |
1126 | bio_set_m(bio, get_default_mirror(ms)); | |
88be163a MB |
1127 | |
1128 | (void) dm_io(&io_req, ms->nr_mirrors, io, NULL); | |
1da177e4 LT |
1129 | } |
1130 | ||
1131 | static void do_writes(struct mirror_set *ms, struct bio_list *writes) | |
1132 | { | |
1133 | int state; | |
1134 | struct bio *bio; | |
1135 | struct bio_list sync, nosync, recover, *this_list = NULL; | |
1136 | ||
1137 | if (!writes->head) | |
1138 | return; | |
1139 | ||
1140 | /* | |
1141 | * Classify each write. | |
1142 | */ | |
1143 | bio_list_init(&sync); | |
1144 | bio_list_init(&nosync); | |
1145 | bio_list_init(&recover); | |
1146 | ||
1147 | while ((bio = bio_list_pop(writes))) { | |
1148 | state = rh_state(&ms->rh, bio_to_region(&ms->rh, bio), 1); | |
1149 | switch (state) { | |
1150 | case RH_CLEAN: | |
1151 | case RH_DIRTY: | |
1152 | this_list = &sync; | |
1153 | break; | |
1154 | ||
1155 | case RH_NOSYNC: | |
1156 | this_list = &nosync; | |
1157 | break; | |
1158 | ||
1159 | case RH_RECOVERING: | |
1160 | this_list = &recover; | |
1161 | break; | |
1162 | } | |
1163 | ||
1164 | bio_list_add(this_list, bio); | |
1165 | } | |
1166 | ||
1167 | /* | |
1168 | * Increment the pending counts for any regions that will | |
1169 | * be written to (writes to recover regions are going to | |
1170 | * be delayed). | |
1171 | */ | |
1172 | rh_inc_pending(&ms->rh, &sync); | |
1173 | rh_inc_pending(&ms->rh, &nosync); | |
fc1ff958 | 1174 | ms->log_failure = rh_flush(&ms->rh) ? 1 : 0; |
1da177e4 LT |
1175 | |
1176 | /* | |
1177 | * Dispatch io. | |
1178 | */ | |
b80aa7a0 JB |
1179 | if (unlikely(ms->log_failure)) { |
1180 | spin_lock_irq(&ms->lock); | |
1181 | bio_list_merge(&ms->failures, &sync); | |
1182 | spin_unlock_irq(&ms->lock); | |
1183 | } else | |
fc1ff958 | 1184 | while ((bio = bio_list_pop(&sync))) |
b80aa7a0 | 1185 | do_write(ms, bio); |
1da177e4 LT |
1186 | |
1187 | while ((bio = bio_list_pop(&recover))) | |
1188 | rh_delay(&ms->rh, bio); | |
1189 | ||
1190 | while ((bio = bio_list_pop(&nosync))) { | |
06386bbf | 1191 | map_bio(get_default_mirror(ms), bio); |
1da177e4 LT |
1192 | generic_make_request(bio); |
1193 | } | |
1194 | } | |
1195 | ||
72f4b314 JB |
1196 | static void do_failures(struct mirror_set *ms, struct bio_list *failures) |
1197 | { | |
1198 | struct bio *bio; | |
1199 | ||
1200 | if (!failures->head) | |
1201 | return; | |
1202 | ||
b80aa7a0 JB |
1203 | if (!ms->log_failure) { |
1204 | while ((bio = bio_list_pop(failures))) | |
1205 | __bio_mark_nosync(ms, bio, bio->bi_size, 0); | |
1206 | return; | |
1207 | } | |
1208 | ||
1209 | /* | |
1210 | * If the log has failed, unattempted writes are being | |
1211 | * put on the failures list. We can't issue those writes | |
1212 | * until a log has been marked, so we must store them. | |
1213 | * | |
1214 | * If a 'noflush' suspend is in progress, we can requeue | |
1215 | * the I/O's to the core. This give userspace a chance | |
1216 | * to reconfigure the mirror, at which point the core | |
1217 | * will reissue the writes. If the 'noflush' flag is | |
1218 | * not set, we have no choice but to return errors. | |
1219 | * | |
1220 | * Some writes on the failures list may have been | |
1221 | * submitted before the log failure and represent a | |
1222 | * failure to write to one of the devices. It is ok | |
1223 | * for us to treat them the same and requeue them | |
1224 | * as well. | |
1225 | */ | |
1226 | if (dm_noflush_suspending(ms->ti)) { | |
1227 | while ((bio = bio_list_pop(failures))) | |
1228 | bio_endio(bio, DM_ENDIO_REQUEUE); | |
1229 | return; | |
1230 | } | |
1231 | ||
1232 | if (atomic_read(&ms->suspend)) { | |
1233 | while ((bio = bio_list_pop(failures))) | |
1234 | bio_endio(bio, -EIO); | |
1235 | return; | |
1236 | } | |
1237 | ||
1238 | spin_lock_irq(&ms->lock); | |
1239 | bio_list_merge(&ms->failures, failures); | |
1240 | spin_unlock_irq(&ms->lock); | |
1241 | ||
1242 | wake(ms); | |
72f4b314 JB |
1243 | } |
1244 | ||
1245 | static void trigger_event(struct work_struct *work) | |
1246 | { | |
1247 | struct mirror_set *ms = | |
1248 | container_of(work, struct mirror_set, trigger_event); | |
1249 | ||
1250 | dm_table_event(ms->ti->table); | |
1251 | } | |
1252 | ||
1da177e4 LT |
1253 | /*----------------------------------------------------------------- |
1254 | * kmirrord | |
1255 | *---------------------------------------------------------------*/ | |
72f4b314 | 1256 | static int _do_mirror(struct work_struct *work) |
1da177e4 | 1257 | { |
6ad36fe2 HS |
1258 | struct mirror_set *ms =container_of(work, struct mirror_set, |
1259 | kmirrord_work); | |
72f4b314 JB |
1260 | struct bio_list reads, writes, failures; |
1261 | unsigned long flags; | |
1da177e4 | 1262 | |
72f4b314 | 1263 | spin_lock_irqsave(&ms->lock, flags); |
1da177e4 LT |
1264 | reads = ms->reads; |
1265 | writes = ms->writes; | |
72f4b314 | 1266 | failures = ms->failures; |
1da177e4 LT |
1267 | bio_list_init(&ms->reads); |
1268 | bio_list_init(&ms->writes); | |
72f4b314 JB |
1269 | bio_list_init(&ms->failures); |
1270 | spin_unlock_irqrestore(&ms->lock, flags); | |
1da177e4 LT |
1271 | |
1272 | rh_update_states(&ms->rh); | |
1273 | do_recovery(ms); | |
1274 | do_reads(ms, &reads); | |
1275 | do_writes(ms, &writes); | |
72f4b314 JB |
1276 | do_failures(ms, &failures); |
1277 | ||
1278 | return (ms->failures.head) ? 1 : 0; | |
1279 | } | |
1280 | ||
1281 | static void do_mirror(struct work_struct *work) | |
1282 | { | |
1283 | /* | |
1284 | * If _do_mirror returns 1, we give it | |
1285 | * another shot. This helps for cases like | |
1286 | * 'suspend' where we call flush_workqueue | |
1287 | * and expect all work to be finished. If | |
1288 | * a failure happens during a suspend, we | |
1289 | * couldn't issue a 'wake' because it would | |
1290 | * not be honored. Therefore, we return '1' | |
1291 | * from _do_mirror, and retry here. | |
1292 | */ | |
1293 | while (_do_mirror(work)) | |
1294 | schedule(); | |
1da177e4 LT |
1295 | } |
1296 | ||
72f4b314 | 1297 | |
1da177e4 LT |
1298 | /*----------------------------------------------------------------- |
1299 | * Target functions | |
1300 | *---------------------------------------------------------------*/ | |
1301 | static struct mirror_set *alloc_context(unsigned int nr_mirrors, | |
1302 | uint32_t region_size, | |
1303 | struct dm_target *ti, | |
1304 | struct dirty_log *dl) | |
1305 | { | |
1306 | size_t len; | |
1307 | struct mirror_set *ms = NULL; | |
1308 | ||
1309 | if (array_too_big(sizeof(*ms), sizeof(ms->mirror[0]), nr_mirrors)) | |
1310 | return NULL; | |
1311 | ||
1312 | len = sizeof(*ms) + (sizeof(ms->mirror[0]) * nr_mirrors); | |
1313 | ||
dd00cc48 | 1314 | ms = kzalloc(len, GFP_KERNEL); |
1da177e4 | 1315 | if (!ms) { |
72d94861 | 1316 | ti->error = "Cannot allocate mirror context"; |
1da177e4 LT |
1317 | return NULL; |
1318 | } | |
1319 | ||
1da177e4 LT |
1320 | spin_lock_init(&ms->lock); |
1321 | ||
1322 | ms->ti = ti; | |
1323 | ms->nr_mirrors = nr_mirrors; | |
1324 | ms->nr_regions = dm_sector_div_up(ti->len, region_size); | |
1325 | ms->in_sync = 0; | |
b80aa7a0 JB |
1326 | ms->log_failure = 0; |
1327 | atomic_set(&ms->suspend, 0); | |
72f4b314 | 1328 | atomic_set(&ms->default_mirror, DEFAULT_MIRROR); |
1da177e4 | 1329 | |
06386bbf JB |
1330 | len = sizeof(struct dm_raid1_read_record); |
1331 | ms->read_record_pool = mempool_create_kmalloc_pool(MIN_READ_RECORDS, | |
1332 | len); | |
1333 | if (!ms->read_record_pool) { | |
1334 | ti->error = "Error creating mirror read_record_pool"; | |
1335 | kfree(ms); | |
1336 | return NULL; | |
1337 | } | |
1338 | ||
88be163a MB |
1339 | ms->io_client = dm_io_client_create(DM_IO_PAGES); |
1340 | if (IS_ERR(ms->io_client)) { | |
1341 | ti->error = "Error creating dm_io client"; | |
06386bbf | 1342 | mempool_destroy(ms->read_record_pool); |
88be163a MB |
1343 | kfree(ms); |
1344 | return NULL; | |
1345 | } | |
1346 | ||
1da177e4 | 1347 | if (rh_init(&ms->rh, ms, dl, region_size, ms->nr_regions)) { |
72d94861 | 1348 | ti->error = "Error creating dirty region hash"; |
a72cf737 | 1349 | dm_io_client_destroy(ms->io_client); |
06386bbf | 1350 | mempool_destroy(ms->read_record_pool); |
1da177e4 LT |
1351 | kfree(ms); |
1352 | return NULL; | |
1353 | } | |
1354 | ||
1355 | return ms; | |
1356 | } | |
1357 | ||
1358 | static void free_context(struct mirror_set *ms, struct dm_target *ti, | |
1359 | unsigned int m) | |
1360 | { | |
1361 | while (m--) | |
1362 | dm_put_device(ti, ms->mirror[m].dev); | |
1363 | ||
88be163a | 1364 | dm_io_client_destroy(ms->io_client); |
1da177e4 | 1365 | rh_exit(&ms->rh); |
06386bbf | 1366 | mempool_destroy(ms->read_record_pool); |
1da177e4 LT |
1367 | kfree(ms); |
1368 | } | |
1369 | ||
1370 | static inline int _check_region_size(struct dm_target *ti, uint32_t size) | |
1371 | { | |
6f3c3f0a | 1372 | return !(size % (PAGE_SIZE >> 9) || !is_power_of_2(size) || |
1da177e4 LT |
1373 | size > ti->len); |
1374 | } | |
1375 | ||
1376 | static int get_mirror(struct mirror_set *ms, struct dm_target *ti, | |
1377 | unsigned int mirror, char **argv) | |
1378 | { | |
4ee218cd | 1379 | unsigned long long offset; |
1da177e4 | 1380 | |
4ee218cd | 1381 | if (sscanf(argv[1], "%llu", &offset) != 1) { |
72d94861 | 1382 | ti->error = "Invalid offset"; |
1da177e4 LT |
1383 | return -EINVAL; |
1384 | } | |
1385 | ||
1386 | if (dm_get_device(ti, argv[0], offset, ti->len, | |
1387 | dm_table_get_mode(ti->table), | |
1388 | &ms->mirror[mirror].dev)) { | |
72d94861 | 1389 | ti->error = "Device lookup failure"; |
1da177e4 LT |
1390 | return -ENXIO; |
1391 | } | |
1392 | ||
aa5617c5 | 1393 | ms->mirror[mirror].ms = ms; |
72f4b314 JB |
1394 | atomic_set(&(ms->mirror[mirror].error_count), 0); |
1395 | ms->mirror[mirror].error_type = 0; | |
1da177e4 LT |
1396 | ms->mirror[mirror].offset = offset; |
1397 | ||
1398 | return 0; | |
1399 | } | |
1400 | ||
1da177e4 LT |
1401 | /* |
1402 | * Create dirty log: log_type #log_params <log_params> | |
1403 | */ | |
1404 | static struct dirty_log *create_dirty_log(struct dm_target *ti, | |
1405 | unsigned int argc, char **argv, | |
1406 | unsigned int *args_used) | |
1407 | { | |
1408 | unsigned int param_count; | |
1409 | struct dirty_log *dl; | |
1410 | ||
1411 | if (argc < 2) { | |
72d94861 | 1412 | ti->error = "Insufficient mirror log arguments"; |
1da177e4 LT |
1413 | return NULL; |
1414 | } | |
1415 | ||
1416 | if (sscanf(argv[1], "%u", ¶m_count) != 1) { | |
72d94861 | 1417 | ti->error = "Invalid mirror log argument count"; |
1da177e4 LT |
1418 | return NULL; |
1419 | } | |
1420 | ||
1421 | *args_used = 2 + param_count; | |
1422 | ||
1423 | if (argc < *args_used) { | |
72d94861 | 1424 | ti->error = "Insufficient mirror log arguments"; |
1da177e4 LT |
1425 | return NULL; |
1426 | } | |
1427 | ||
1428 | dl = dm_create_dirty_log(argv[0], ti, param_count, argv + 2); | |
1429 | if (!dl) { | |
72d94861 | 1430 | ti->error = "Error creating mirror dirty log"; |
1da177e4 LT |
1431 | return NULL; |
1432 | } | |
1433 | ||
1434 | if (!_check_region_size(ti, dl->type->get_region_size(dl))) { | |
72d94861 | 1435 | ti->error = "Invalid region size"; |
1da177e4 LT |
1436 | dm_destroy_dirty_log(dl); |
1437 | return NULL; | |
1438 | } | |
1439 | ||
1440 | return dl; | |
1441 | } | |
1442 | ||
a8e6afa2 JB |
1443 | static int parse_features(struct mirror_set *ms, unsigned argc, char **argv, |
1444 | unsigned *args_used) | |
1445 | { | |
1446 | unsigned num_features; | |
1447 | struct dm_target *ti = ms->ti; | |
1448 | ||
1449 | *args_used = 0; | |
1450 | ||
1451 | if (!argc) | |
1452 | return 0; | |
1453 | ||
1454 | if (sscanf(argv[0], "%u", &num_features) != 1) { | |
1455 | ti->error = "Invalid number of features"; | |
1456 | return -EINVAL; | |
1457 | } | |
1458 | ||
1459 | argc--; | |
1460 | argv++; | |
1461 | (*args_used)++; | |
1462 | ||
1463 | if (num_features > argc) { | |
1464 | ti->error = "Not enough arguments to support feature count"; | |
1465 | return -EINVAL; | |
1466 | } | |
1467 | ||
1468 | if (!strcmp("handle_errors", argv[0])) | |
1469 | ms->features |= DM_RAID1_HANDLE_ERRORS; | |
1470 | else { | |
1471 | ti->error = "Unrecognised feature requested"; | |
1472 | return -EINVAL; | |
1473 | } | |
1474 | ||
1475 | (*args_used)++; | |
1476 | ||
1477 | return 0; | |
1478 | } | |
1479 | ||
1da177e4 LT |
1480 | /* |
1481 | * Construct a mirror mapping: | |
1482 | * | |
1483 | * log_type #log_params <log_params> | |
1484 | * #mirrors [mirror_path offset]{2,} | |
a8e6afa2 | 1485 | * [#features <features>] |
1da177e4 LT |
1486 | * |
1487 | * log_type is "core" or "disk" | |
1488 | * #log_params is between 1 and 3 | |
a8e6afa2 JB |
1489 | * |
1490 | * If present, features must be "handle_errors". | |
1da177e4 | 1491 | */ |
1da177e4 LT |
1492 | static int mirror_ctr(struct dm_target *ti, unsigned int argc, char **argv) |
1493 | { | |
1494 | int r; | |
1495 | unsigned int nr_mirrors, m, args_used; | |
1496 | struct mirror_set *ms; | |
1497 | struct dirty_log *dl; | |
1498 | ||
1499 | dl = create_dirty_log(ti, argc, argv, &args_used); | |
1500 | if (!dl) | |
1501 | return -EINVAL; | |
1502 | ||
1503 | argv += args_used; | |
1504 | argc -= args_used; | |
1505 | ||
1506 | if (!argc || sscanf(argv[0], "%u", &nr_mirrors) != 1 || | |
1507 | nr_mirrors < 2 || nr_mirrors > KCOPYD_MAX_REGIONS + 1) { | |
72d94861 | 1508 | ti->error = "Invalid number of mirrors"; |
1da177e4 LT |
1509 | dm_destroy_dirty_log(dl); |
1510 | return -EINVAL; | |
1511 | } | |
1512 | ||
1513 | argv++, argc--; | |
1514 | ||
a8e6afa2 JB |
1515 | if (argc < nr_mirrors * 2) { |
1516 | ti->error = "Too few mirror arguments"; | |
1da177e4 LT |
1517 | dm_destroy_dirty_log(dl); |
1518 | return -EINVAL; | |
1519 | } | |
1520 | ||
1521 | ms = alloc_context(nr_mirrors, dl->type->get_region_size(dl), ti, dl); | |
1522 | if (!ms) { | |
1523 | dm_destroy_dirty_log(dl); | |
1524 | return -ENOMEM; | |
1525 | } | |
1526 | ||
1527 | /* Get the mirror parameter sets */ | |
1528 | for (m = 0; m < nr_mirrors; m++) { | |
1529 | r = get_mirror(ms, ti, m, argv); | |
1530 | if (r) { | |
1531 | free_context(ms, ti, m); | |
1532 | return r; | |
1533 | } | |
1534 | argv += 2; | |
1535 | argc -= 2; | |
1536 | } | |
1537 | ||
1538 | ti->private = ms; | |
d88854f0 | 1539 | ti->split_io = ms->rh.region_size; |
1da177e4 | 1540 | |
6ad36fe2 HS |
1541 | ms->kmirrord_wq = create_singlethread_workqueue("kmirrord"); |
1542 | if (!ms->kmirrord_wq) { | |
1543 | DMERR("couldn't start kmirrord"); | |
a72cf737 DM |
1544 | r = -ENOMEM; |
1545 | goto err_free_context; | |
6ad36fe2 HS |
1546 | } |
1547 | INIT_WORK(&ms->kmirrord_work, do_mirror); | |
72f4b314 | 1548 | INIT_WORK(&ms->trigger_event, trigger_event); |
6ad36fe2 | 1549 | |
a8e6afa2 | 1550 | r = parse_features(ms, argc, argv, &args_used); |
a72cf737 DM |
1551 | if (r) |
1552 | goto err_destroy_wq; | |
a8e6afa2 JB |
1553 | |
1554 | argv += args_used; | |
1555 | argc -= args_used; | |
1556 | ||
f44db678 JB |
1557 | /* |
1558 | * Any read-balancing addition depends on the | |
1559 | * DM_RAID1_HANDLE_ERRORS flag being present. | |
1560 | * This is because the decision to balance depends | |
1561 | * on the sync state of a region. If the above | |
1562 | * flag is not present, we ignore errors; and | |
1563 | * the sync state may be inaccurate. | |
1564 | */ | |
1565 | ||
a8e6afa2 JB |
1566 | if (argc) { |
1567 | ti->error = "Too many mirror arguments"; | |
a72cf737 DM |
1568 | r = -EINVAL; |
1569 | goto err_destroy_wq; | |
a8e6afa2 JB |
1570 | } |
1571 | ||
1da177e4 | 1572 | r = kcopyd_client_create(DM_IO_PAGES, &ms->kcopyd_client); |
a72cf737 DM |
1573 | if (r) |
1574 | goto err_destroy_wq; | |
1da177e4 | 1575 | |
6ad36fe2 | 1576 | wake(ms); |
1da177e4 | 1577 | return 0; |
a72cf737 DM |
1578 | |
1579 | err_destroy_wq: | |
1580 | destroy_workqueue(ms->kmirrord_wq); | |
1581 | err_free_context: | |
1582 | free_context(ms, ti, ms->nr_mirrors); | |
1583 | return r; | |
1da177e4 LT |
1584 | } |
1585 | ||
1586 | static void mirror_dtr(struct dm_target *ti) | |
1587 | { | |
1588 | struct mirror_set *ms = (struct mirror_set *) ti->private; | |
1589 | ||
6ad36fe2 | 1590 | flush_workqueue(ms->kmirrord_wq); |
1da177e4 | 1591 | kcopyd_client_destroy(ms->kcopyd_client); |
6ad36fe2 | 1592 | destroy_workqueue(ms->kmirrord_wq); |
1da177e4 LT |
1593 | free_context(ms, ti, ms->nr_mirrors); |
1594 | } | |
1595 | ||
1596 | static void queue_bio(struct mirror_set *ms, struct bio *bio, int rw) | |
1597 | { | |
72f4b314 | 1598 | unsigned long flags; |
1da177e4 LT |
1599 | int should_wake = 0; |
1600 | struct bio_list *bl; | |
1601 | ||
1602 | bl = (rw == WRITE) ? &ms->writes : &ms->reads; | |
72f4b314 | 1603 | spin_lock_irqsave(&ms->lock, flags); |
1da177e4 LT |
1604 | should_wake = !(bl->head); |
1605 | bio_list_add(bl, bio); | |
72f4b314 | 1606 | spin_unlock_irqrestore(&ms->lock, flags); |
1da177e4 LT |
1607 | |
1608 | if (should_wake) | |
6ad36fe2 | 1609 | wake(ms); |
1da177e4 LT |
1610 | } |
1611 | ||
1612 | /* | |
1613 | * Mirror mapping function | |
1614 | */ | |
1615 | static int mirror_map(struct dm_target *ti, struct bio *bio, | |
1616 | union map_info *map_context) | |
1617 | { | |
1618 | int r, rw = bio_rw(bio); | |
1619 | struct mirror *m; | |
1620 | struct mirror_set *ms = ti->private; | |
06386bbf | 1621 | struct dm_raid1_read_record *read_record = NULL; |
1da177e4 LT |
1622 | |
1623 | if (rw == WRITE) { | |
06386bbf JB |
1624 | /* Save region for mirror_end_io() handler */ |
1625 | map_context->ll = bio_to_region(&ms->rh, bio); | |
1da177e4 | 1626 | queue_bio(ms, bio, rw); |
d2a7ad29 | 1627 | return DM_MAPIO_SUBMITTED; |
1da177e4 LT |
1628 | } |
1629 | ||
1630 | r = ms->rh.log->type->in_sync(ms->rh.log, | |
1631 | bio_to_region(&ms->rh, bio), 0); | |
1632 | if (r < 0 && r != -EWOULDBLOCK) | |
1633 | return r; | |
1634 | ||
1da177e4 | 1635 | /* |
06386bbf | 1636 | * If region is not in-sync queue the bio. |
1da177e4 | 1637 | */ |
06386bbf JB |
1638 | if (!r || (r == -EWOULDBLOCK)) { |
1639 | if (rw == READA) | |
1640 | return -EWOULDBLOCK; | |
1da177e4 | 1641 | |
1da177e4 | 1642 | queue_bio(ms, bio, rw); |
d2a7ad29 | 1643 | return DM_MAPIO_SUBMITTED; |
1da177e4 LT |
1644 | } |
1645 | ||
06386bbf JB |
1646 | /* |
1647 | * The region is in-sync and we can perform reads directly. | |
1648 | * Store enough information so we can retry if it fails. | |
1649 | */ | |
1da177e4 | 1650 | m = choose_mirror(ms, bio->bi_sector); |
06386bbf | 1651 | if (unlikely(!m)) |
1da177e4 LT |
1652 | return -EIO; |
1653 | ||
06386bbf JB |
1654 | read_record = mempool_alloc(ms->read_record_pool, GFP_NOIO); |
1655 | if (likely(read_record)) { | |
1656 | dm_bio_record(&read_record->details, bio); | |
1657 | map_context->ptr = read_record; | |
1658 | read_record->m = m; | |
1659 | } | |
1660 | ||
1661 | map_bio(m, bio); | |
1662 | ||
d2a7ad29 | 1663 | return DM_MAPIO_REMAPPED; |
1da177e4 LT |
1664 | } |
1665 | ||
1666 | static int mirror_end_io(struct dm_target *ti, struct bio *bio, | |
1667 | int error, union map_info *map_context) | |
1668 | { | |
1669 | int rw = bio_rw(bio); | |
1670 | struct mirror_set *ms = (struct mirror_set *) ti->private; | |
06386bbf JB |
1671 | struct mirror *m = NULL; |
1672 | struct dm_bio_details *bd = NULL; | |
1673 | struct dm_raid1_read_record *read_record = map_context->ptr; | |
1da177e4 LT |
1674 | |
1675 | /* | |
1676 | * We need to dec pending if this was a write. | |
1677 | */ | |
06386bbf JB |
1678 | if (rw == WRITE) { |
1679 | rh_dec(&ms->rh, map_context->ll); | |
1680 | return error; | |
1681 | } | |
1da177e4 | 1682 | |
06386bbf JB |
1683 | if (error == -EOPNOTSUPP) |
1684 | goto out; | |
1685 | ||
1686 | if ((error == -EWOULDBLOCK) && bio_rw_ahead(bio)) | |
1687 | goto out; | |
1688 | ||
1689 | if (unlikely(error)) { | |
1690 | if (!read_record) { | |
1691 | /* | |
1692 | * There wasn't enough memory to record necessary | |
1693 | * information for a retry or there was no other | |
1694 | * mirror in-sync. | |
1695 | */ | |
e03f1a84 | 1696 | DMERR_LIMIT("Mirror read failed."); |
06386bbf JB |
1697 | return -EIO; |
1698 | } | |
e03f1a84 AB |
1699 | |
1700 | m = read_record->m; | |
1701 | ||
06386bbf JB |
1702 | DMERR("Mirror read failed from %s. Trying alternative device.", |
1703 | m->dev->name); | |
1704 | ||
06386bbf JB |
1705 | fail_mirror(m, DM_RAID1_READ_ERROR); |
1706 | ||
1707 | /* | |
1708 | * A failed read is requeued for another attempt using an intact | |
1709 | * mirror. | |
1710 | */ | |
1711 | if (default_ok(m) || mirror_available(ms, bio)) { | |
1712 | bd = &read_record->details; | |
1713 | ||
1714 | dm_bio_restore(bd, bio); | |
1715 | mempool_free(read_record, ms->read_record_pool); | |
1716 | map_context->ptr = NULL; | |
1717 | queue_bio(ms, bio, rw); | |
1718 | return 1; | |
1719 | } | |
1720 | DMERR("All replicated volumes dead, failing I/O"); | |
1721 | } | |
1722 | ||
1723 | out: | |
1724 | if (read_record) { | |
1725 | mempool_free(read_record, ms->read_record_pool); | |
1726 | map_context->ptr = NULL; | |
1727 | } | |
1728 | ||
1729 | return error; | |
1da177e4 LT |
1730 | } |
1731 | ||
b80aa7a0 | 1732 | static void mirror_presuspend(struct dm_target *ti) |
1da177e4 LT |
1733 | { |
1734 | struct mirror_set *ms = (struct mirror_set *) ti->private; | |
1735 | struct dirty_log *log = ms->rh.log; | |
1736 | ||
b80aa7a0 JB |
1737 | atomic_set(&ms->suspend, 1); |
1738 | ||
1739 | /* | |
1740 | * We must finish up all the work that we've | |
1741 | * generated (i.e. recovery work). | |
1742 | */ | |
1da177e4 | 1743 | rh_stop_recovery(&ms->rh); |
33184048 | 1744 | |
33184048 JB |
1745 | wait_event(_kmirrord_recovery_stopped, |
1746 | !atomic_read(&ms->rh.recovery_in_flight)); | |
1747 | ||
b80aa7a0 JB |
1748 | if (log->type->presuspend && log->type->presuspend(log)) |
1749 | /* FIXME: need better error handling */ | |
1750 | DMWARN("log presuspend failed"); | |
1751 | ||
1752 | /* | |
1753 | * Now that recovery is complete/stopped and the | |
1754 | * delayed bios are queued, we need to wait for | |
1755 | * the worker thread to complete. This way, | |
1756 | * we know that all of our I/O has been pushed. | |
1757 | */ | |
1758 | flush_workqueue(ms->kmirrord_wq); | |
1759 | } | |
1760 | ||
1761 | static void mirror_postsuspend(struct dm_target *ti) | |
1762 | { | |
1763 | struct mirror_set *ms = ti->private; | |
1764 | struct dirty_log *log = ms->rh.log; | |
1765 | ||
6b3df0d7 | 1766 | if (log->type->postsuspend && log->type->postsuspend(log)) |
1da177e4 | 1767 | /* FIXME: need better error handling */ |
b80aa7a0 | 1768 | DMWARN("log postsuspend failed"); |
1da177e4 LT |
1769 | } |
1770 | ||
1771 | static void mirror_resume(struct dm_target *ti) | |
1772 | { | |
b80aa7a0 | 1773 | struct mirror_set *ms = ti->private; |
1da177e4 | 1774 | struct dirty_log *log = ms->rh.log; |
b80aa7a0 JB |
1775 | |
1776 | atomic_set(&ms->suspend, 0); | |
1da177e4 LT |
1777 | if (log->type->resume && log->type->resume(log)) |
1778 | /* FIXME: need better error handling */ | |
1779 | DMWARN("log resume failed"); | |
1780 | rh_start_recovery(&ms->rh); | |
1781 | } | |
1782 | ||
af195ac8 JB |
1783 | /* |
1784 | * device_status_char | |
1785 | * @m: mirror device/leg we want the status of | |
1786 | * | |
1787 | * We return one character representing the most severe error | |
1788 | * we have encountered. | |
1789 | * A => Alive - No failures | |
1790 | * D => Dead - A write failure occurred leaving mirror out-of-sync | |
1791 | * S => Sync - A sychronization failure occurred, mirror out-of-sync | |
1792 | * R => Read - A read failure occurred, mirror data unaffected | |
1793 | * | |
1794 | * Returns: <char> | |
1795 | */ | |
1796 | static char device_status_char(struct mirror *m) | |
1797 | { | |
1798 | if (!atomic_read(&(m->error_count))) | |
1799 | return 'A'; | |
1800 | ||
1801 | return (test_bit(DM_RAID1_WRITE_ERROR, &(m->error_type))) ? 'D' : | |
1802 | (test_bit(DM_RAID1_SYNC_ERROR, &(m->error_type))) ? 'S' : | |
1803 | (test_bit(DM_RAID1_READ_ERROR, &(m->error_type))) ? 'R' : 'U'; | |
1804 | } | |
1805 | ||
1806 | ||
1da177e4 LT |
1807 | static int mirror_status(struct dm_target *ti, status_type_t type, |
1808 | char *result, unsigned int maxlen) | |
1809 | { | |
315dcc22 | 1810 | unsigned int m, sz = 0; |
1da177e4 | 1811 | struct mirror_set *ms = (struct mirror_set *) ti->private; |
af195ac8 JB |
1812 | struct dirty_log *log = ms->rh.log; |
1813 | char buffer[ms->nr_mirrors + 1]; | |
1da177e4 | 1814 | |
1da177e4 LT |
1815 | switch (type) { |
1816 | case STATUSTYPE_INFO: | |
1817 | DMEMIT("%d ", ms->nr_mirrors); | |
af195ac8 | 1818 | for (m = 0; m < ms->nr_mirrors; m++) { |
1da177e4 | 1819 | DMEMIT("%s ", ms->mirror[m].dev->name); |
af195ac8 JB |
1820 | buffer[m] = device_status_char(&(ms->mirror[m])); |
1821 | } | |
1822 | buffer[m] = '\0'; | |
1da177e4 | 1823 | |
af195ac8 JB |
1824 | DMEMIT("%llu/%llu 1 %s ", |
1825 | (unsigned long long)log->type->get_sync_count(ms->rh.log), | |
1826 | (unsigned long long)ms->nr_regions, buffer); | |
315dcc22 | 1827 | |
af195ac8 | 1828 | sz += log->type->status(ms->rh.log, type, result+sz, maxlen-sz); |
315dcc22 | 1829 | |
1da177e4 LT |
1830 | break; |
1831 | ||
1832 | case STATUSTYPE_TABLE: | |
af195ac8 | 1833 | sz = log->type->status(ms->rh.log, type, result, maxlen); |
315dcc22 | 1834 | |
e52b8f6d | 1835 | DMEMIT("%d", ms->nr_mirrors); |
1da177e4 | 1836 | for (m = 0; m < ms->nr_mirrors; m++) |
e52b8f6d | 1837 | DMEMIT(" %s %llu", ms->mirror[m].dev->name, |
b80aa7a0 | 1838 | (unsigned long long)ms->mirror[m].offset); |
a8e6afa2 JB |
1839 | |
1840 | if (ms->features & DM_RAID1_HANDLE_ERRORS) | |
1841 | DMEMIT(" 1 handle_errors"); | |
1da177e4 LT |
1842 | } |
1843 | ||
1844 | return 0; | |
1845 | } | |
1846 | ||
1847 | static struct target_type mirror_target = { | |
1848 | .name = "mirror", | |
af195ac8 | 1849 | .version = {1, 0, 20}, |
1da177e4 LT |
1850 | .module = THIS_MODULE, |
1851 | .ctr = mirror_ctr, | |
1852 | .dtr = mirror_dtr, | |
1853 | .map = mirror_map, | |
1854 | .end_io = mirror_end_io, | |
b80aa7a0 | 1855 | .presuspend = mirror_presuspend, |
1da177e4 LT |
1856 | .postsuspend = mirror_postsuspend, |
1857 | .resume = mirror_resume, | |
1858 | .status = mirror_status, | |
1859 | }; | |
1860 | ||
1861 | static int __init dm_mirror_init(void) | |
1862 | { | |
1863 | int r; | |
1864 | ||
1da177e4 | 1865 | r = dm_register_target(&mirror_target); |
769aef30 | 1866 | if (r < 0) |
0cd33124 | 1867 | DMERR("Failed to register mirror target"); |
1da177e4 LT |
1868 | |
1869 | return r; | |
1870 | } | |
1871 | ||
1872 | static void __exit dm_mirror_exit(void) | |
1873 | { | |
1874 | int r; | |
1875 | ||
1876 | r = dm_unregister_target(&mirror_target); | |
1877 | if (r < 0) | |
0cd33124 | 1878 | DMERR("unregister failed %d", r); |
1da177e4 LT |
1879 | } |
1880 | ||
1881 | /* Module hooks */ | |
1882 | module_init(dm_mirror_init); | |
1883 | module_exit(dm_mirror_exit); | |
1884 | ||
1885 | MODULE_DESCRIPTION(DM_NAME " mirror target"); | |
1886 | MODULE_AUTHOR("Joe Thornber"); | |
1887 | MODULE_LICENSE("GPL"); |