Blame view
drivers/md/raid5.c
159 KB
1da177e4c
|
1 2 3 4 |
/* * raid5.c : Multiple Devices driver for Linux * Copyright (C) 1996, 1997 Ingo Molnar, Miguel de Icaza, Gadi Oxman * Copyright (C) 1999, 2000 Ingo Molnar |
16a53ecc3
|
5 |
* Copyright (C) 2002, 2003 H. Peter Anvin |
1da177e4c
|
6 |
* |
16a53ecc3
|
7 8 9 |
* RAID-4/5/6 management functions. * Thanks to Penguin Computing for making the RAID-6 development possible * by donating a test server! |
1da177e4c
|
10 11 12 13 14 15 16 17 18 19 |
* * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2, or (at your option) * any later version. * * You should have received a copy of the GNU General Public License * (for example /usr/src/linux/COPYING); if not, write to the Free * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ |
ae3c20ccf
|
20 21 22 23 24 25 26 27 28 |
/* * BITMAP UNPLUGGING: * * The sequencing for updating the bitmap reliably is a little * subtle (and I got it wrong the first time) so it deserves some * explanation. * * We group bitmap updates into batches. Each batch has a number. * We may write out several batches at once, but that isn't very important. |
7c13edc87
|
29 30 |
* conf->seq_write is the number of the last batch successfully written. * conf->seq_flush is the number of the last batch that was closed to |
ae3c20ccf
|
31 32 33 |
* new additions. * When we discover that we will need to write to any block in a stripe * (in add_stripe_bio) we update the in-memory bitmap and record in sh->bm_seq |
7c13edc87
|
34 |
* the number of the batch it will be in. This is seq_flush+1. |
ae3c20ccf
|
35 36 37 38 39 40 41 42 43 |
* When we are ready to do a write, if that batch hasn't been written yet, * we plug the array and queue the stripe for later. * When an unplug happens, we increment bm_flush, thus closing the current * batch. * When we notice that bm_flush > bm_write, we write out all pending updates * to the bitmap, and advance bm_write to where bm_flush was. * This may occasionally write a bit out twice, but is sure never to * miss any bits. */ |
1da177e4c
|
44 |
|
bff61975b
|
45 |
#include <linux/blkdev.h> |
f67055780
|
46 |
#include <linux/kthread.h> |
f701d589a
|
47 |
#include <linux/raid/pq.h> |
91c009248
|
48 |
#include <linux/async_tx.h> |
056075c76
|
49 |
#include <linux/module.h> |
07a3b417d
|
50 |
#include <linux/async.h> |
bff61975b
|
51 |
#include <linux/seq_file.h> |
36d1c6476
|
52 |
#include <linux/cpu.h> |
5a0e3ad6a
|
53 |
#include <linux/slab.h> |
8bda470e8
|
54 |
#include <linux/ratelimit.h> |
43b2e5d86
|
55 |
#include "md.h" |
bff61975b
|
56 |
#include "raid5.h" |
54071b380
|
57 |
#include "raid0.h" |
ef740c372
|
58 |
#include "bitmap.h" |
72626685d
|
59 |
|
1da177e4c
|
60 61 62 63 64 65 66 67 68 |
/* * Stripe cache */ #define NR_STRIPES 256 #define STRIPE_SIZE PAGE_SIZE #define STRIPE_SHIFT (PAGE_SHIFT - 9) #define STRIPE_SECTORS (STRIPE_SIZE>>9) #define IO_THRESHOLD 1 |
8b3e6cdc5
|
69 |
#define BYPASS_THRESHOLD 1 |
fccddba06
|
70 |
#define NR_HASH (PAGE_SIZE / sizeof(struct hlist_head)) |
1da177e4c
|
71 |
#define HASH_MASK (NR_HASH - 1) |
d1688a6d5
|
72 |
static inline struct hlist_head *stripe_hash(struct r5conf *conf, sector_t sect) |
db298e194
|
73 74 75 76 |
{ int hash = (sect >> STRIPE_SHIFT) & HASH_MASK; return &conf->stripe_hashtbl[hash]; } |
1da177e4c
|
77 78 79 80 81 82 83 |
/* bio's attached to a stripe+device for I/O are linked together in bi_sector * order without overlap. There may be several bio's per stripe+device, and * a bio could span several devices. * When walking this list for a particular stripe+device, we must never proceed * beyond a bio that extends past this device, as the next bio might no longer * be valid. |
db298e194
|
84 |
* This function is used to determine the 'next' bio in the list, given the sector |
1da177e4c
|
85 86 |
* of the current stripe+device */ |
db298e194
|
87 88 89 90 91 92 93 94 |
static inline struct bio *r5_next_bio(struct bio *bio, sector_t sector) { int sectors = bio->bi_size >> 9; if (bio->bi_sector + sectors < sector + STRIPE_SECTORS) return bio->bi_next; else return NULL; } |
1da177e4c
|
95 |
|
960e739d9
|
96 |
/* |
5b99c2ffa
|
97 98 |
* We maintain a biased count of active stripes in the bottom 16 bits of * bi_phys_segments, and a count of processed stripes in the upper 16 bits |
960e739d9
|
99 100 101 |
*/ static inline int raid5_bi_phys_segments(struct bio *bio) { |
5b99c2ffa
|
102 |
return bio->bi_phys_segments & 0xffff; |
960e739d9
|
103 104 105 106 |
} static inline int raid5_bi_hw_segments(struct bio *bio) { |
5b99c2ffa
|
107 |
return (bio->bi_phys_segments >> 16) & 0xffff; |
960e739d9
|
108 109 110 111 112 113 114 115 116 117 118 119 120 |
} static inline int raid5_dec_bi_phys_segments(struct bio *bio) { --bio->bi_phys_segments; return raid5_bi_phys_segments(bio); } static inline int raid5_dec_bi_hw_segments(struct bio *bio) { unsigned short val = raid5_bi_hw_segments(bio); --val; |
5b99c2ffa
|
121 |
bio->bi_phys_segments = (val << 16) | raid5_bi_phys_segments(bio); |
960e739d9
|
122 123 124 125 126 |
return val; } static inline void raid5_set_bi_hw_segments(struct bio *bio, unsigned int cnt) { |
9b2dc8b66
|
127 |
bio->bi_phys_segments = raid5_bi_phys_segments(bio) | (cnt << 16); |
960e739d9
|
128 |
} |
d0dabf7e5
|
129 130 131 |
/* Find first data disk in a raid6 stripe */ static inline int raid6_d0(struct stripe_head *sh) { |
67cc2b816
|
132 133 134 135 |
if (sh->ddf_layout) /* ddf always start from first device */ return 0; /* md starts just after Q block */ |
d0dabf7e5
|
136 137 138 139 140 |
if (sh->qd_idx == sh->disks - 1) return 0; else return sh->qd_idx + 1; } |
16a53ecc3
|
141 142 143 144 145 |
static inline int raid6_next_disk(int disk, int raid_disks) { disk++; return (disk < raid_disks) ? disk : 0; } |
a44568564
|
146 |
|
d0dabf7e5
|
147 148 149 150 151 |
/* When walking through the disks in a raid5, starting at raid6_d0, * We need to map each disk to a 'slot', where the data disks are slot * 0 .. raid_disks-3, the parity disk is raid_disks-2 and the Q disk * is raid_disks-1. This help does that mapping. */ |
67cc2b816
|
152 153 |
static int raid6_idx_to_slot(int idx, struct stripe_head *sh, int *count, int syndrome_disks) |
d0dabf7e5
|
154 |
{ |
6629542e7
|
155 |
int slot = *count; |
67cc2b816
|
156 |
|
e4424fee1
|
157 |
if (sh->ddf_layout) |
6629542e7
|
158 |
(*count)++; |
d0dabf7e5
|
159 |
if (idx == sh->pd_idx) |
67cc2b816
|
160 |
return syndrome_disks; |
d0dabf7e5
|
161 |
if (idx == sh->qd_idx) |
67cc2b816
|
162 |
return syndrome_disks + 1; |
e4424fee1
|
163 |
if (!sh->ddf_layout) |
6629542e7
|
164 |
(*count)++; |
d0dabf7e5
|
165 166 |
return slot; } |
a44568564
|
167 168 169 170 |
static void return_io(struct bio *return_bi) { struct bio *bi = return_bi; while (bi) { |
a44568564
|
171 172 173 174 |
return_bi = bi->bi_next; bi->bi_next = NULL; bi->bi_size = 0; |
0e13fe23a
|
175 |
bio_endio(bi, 0); |
a44568564
|
176 177 178 |
bi = return_bi; } } |
d1688a6d5
|
179 |
static void print_raid5_conf (struct r5conf *conf); |
1da177e4c
|
180 |
|
600aa1099
|
181 182 183 184 185 186 |
static int stripe_operations_active(struct stripe_head *sh) { return sh->check_state || sh->reconstruct_state || test_bit(STRIPE_BIOFILL_RUN, &sh->state) || test_bit(STRIPE_COMPUTE_RUN, &sh->state); } |
d1688a6d5
|
187 |
static void __release_stripe(struct r5conf *conf, struct stripe_head *sh) |
1da177e4c
|
188 189 |
{ if (atomic_dec_and_test(&sh->count)) { |
78bafebd4
|
190 191 |
BUG_ON(!list_empty(&sh->lru)); BUG_ON(atomic_read(&conf->active_stripes)==0); |
1da177e4c
|
192 |
if (test_bit(STRIPE_HANDLE, &sh->state)) { |
482c08349
|
193 |
if (test_bit(STRIPE_DELAYED, &sh->state)) |
1da177e4c
|
194 |
list_add_tail(&sh->lru, &conf->delayed_list); |
482c08349
|
195 196 |
else if (test_bit(STRIPE_BIT_DELAY, &sh->state) && sh->bm_seq - conf->seq_write > 0) |
72626685d
|
197 |
list_add_tail(&sh->lru, &conf->bitmap_list); |
482c08349
|
198 |
else { |
72626685d
|
199 |
clear_bit(STRIPE_BIT_DELAY, &sh->state); |
1da177e4c
|
200 |
list_add_tail(&sh->lru, &conf->handle_list); |
72626685d
|
201 |
} |
1da177e4c
|
202 203 |
md_wakeup_thread(conf->mddev->thread); } else { |
600aa1099
|
204 |
BUG_ON(stripe_operations_active(sh)); |
1da177e4c
|
205 206 207 208 209 |
if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { atomic_dec(&conf->preread_active_stripes); if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD) md_wakeup_thread(conf->mddev->thread); } |
1da177e4c
|
210 |
atomic_dec(&conf->active_stripes); |
ccfcc3c10
|
211 212 |
if (!test_bit(STRIPE_EXPANDING, &sh->state)) { list_add_tail(&sh->lru, &conf->inactive_list); |
1da177e4c
|
213 |
wake_up(&conf->wait_for_stripe); |
46031f9a3
|
214 215 |
if (conf->retry_read_aligned) md_wakeup_thread(conf->mddev->thread); |
ccfcc3c10
|
216 |
} |
1da177e4c
|
217 218 219 |
} } } |
d0dabf7e5
|
220 |
|
1da177e4c
|
221 222 |
static void release_stripe(struct stripe_head *sh) { |
d1688a6d5
|
223 |
struct r5conf *conf = sh->raid_conf; |
1da177e4c
|
224 |
unsigned long flags; |
16a53ecc3
|
225 |
|
1da177e4c
|
226 227 228 229 |
spin_lock_irqsave(&conf->device_lock, flags); __release_stripe(conf, sh); spin_unlock_irqrestore(&conf->device_lock, flags); } |
fccddba06
|
230 |
static inline void remove_hash(struct stripe_head *sh) |
1da177e4c
|
231 |
{ |
45b4233ca
|
232 233 234 |
pr_debug("remove_hash(), stripe %llu ", (unsigned long long)sh->sector); |
1da177e4c
|
235 |
|
fccddba06
|
236 |
hlist_del_init(&sh->hash); |
1da177e4c
|
237 |
} |
d1688a6d5
|
238 |
static inline void insert_hash(struct r5conf *conf, struct stripe_head *sh) |
1da177e4c
|
239 |
{ |
fccddba06
|
240 |
struct hlist_head *hp = stripe_hash(conf, sh->sector); |
1da177e4c
|
241 |
|
45b4233ca
|
242 243 244 |
pr_debug("insert_hash(), stripe %llu ", (unsigned long long)sh->sector); |
1da177e4c
|
245 |
|
fccddba06
|
246 |
hlist_add_head(&sh->hash, hp); |
1da177e4c
|
247 248 249 250 |
} /* find an idle stripe, make sure it is unhashed, and return it. */ |
d1688a6d5
|
251 |
static struct stripe_head *get_free_stripe(struct r5conf *conf) |
1da177e4c
|
252 253 254 |
{ struct stripe_head *sh = NULL; struct list_head *first; |
1da177e4c
|
255 256 257 258 259 260 261 262 263 264 |
if (list_empty(&conf->inactive_list)) goto out; first = conf->inactive_list.next; sh = list_entry(first, struct stripe_head, lru); list_del_init(first); remove_hash(sh); atomic_inc(&conf->active_stripes); out: return sh; } |
e4e11e385
|
265 |
static void shrink_buffers(struct stripe_head *sh) |
1da177e4c
|
266 267 268 |
{ struct page *p; int i; |
e4e11e385
|
269 |
int num = sh->raid_conf->pool_size; |
1da177e4c
|
270 |
|
e4e11e385
|
271 |
for (i = 0; i < num ; i++) { |
1da177e4c
|
272 273 274 275 |
p = sh->dev[i].page; if (!p) continue; sh->dev[i].page = NULL; |
2d1f3b5d1
|
276 |
put_page(p); |
1da177e4c
|
277 278 |
} } |
e4e11e385
|
279 |
static int grow_buffers(struct stripe_head *sh) |
1da177e4c
|
280 281 |
{ int i; |
e4e11e385
|
282 |
int num = sh->raid_conf->pool_size; |
1da177e4c
|
283 |
|
e4e11e385
|
284 |
for (i = 0; i < num; i++) { |
1da177e4c
|
285 286 287 288 289 290 291 292 293 |
struct page *page; if (!(page = alloc_page(GFP_KERNEL))) { return 1; } sh->dev[i].page = page; } return 0; } |
784052ecc
|
294 |
static void raid5_build_block(struct stripe_head *sh, int i, int previous); |
d1688a6d5
|
295 |
static void stripe_set_idx(sector_t stripe, struct r5conf *conf, int previous, |
911d4ee85
|
296 |
struct stripe_head *sh); |
1da177e4c
|
297 |
|
b5663ba40
|
298 |
static void init_stripe(struct stripe_head *sh, sector_t sector, int previous) |
1da177e4c
|
299 |
{ |
d1688a6d5
|
300 |
struct r5conf *conf = sh->raid_conf; |
7ecaa1e6a
|
301 |
int i; |
1da177e4c
|
302 |
|
78bafebd4
|
303 304 |
BUG_ON(atomic_read(&sh->count) != 0); BUG_ON(test_bit(STRIPE_HANDLE, &sh->state)); |
600aa1099
|
305 |
BUG_ON(stripe_operations_active(sh)); |
d84e0f10d
|
306 |
|
45b4233ca
|
307 308 |
pr_debug("init_stripe called, stripe %llu ", |
1da177e4c
|
309 310 311 |
(unsigned long long)sh->sector); remove_hash(sh); |
16a53ecc3
|
312 |
|
86b42c713
|
313 |
sh->generation = conf->generation - previous; |
b5663ba40
|
314 |
sh->disks = previous ? conf->previous_raid_disks : conf->raid_disks; |
1da177e4c
|
315 |
sh->sector = sector; |
911d4ee85
|
316 |
stripe_set_idx(sector, conf, previous, sh); |
1da177e4c
|
317 |
sh->state = 0; |
7ecaa1e6a
|
318 319 |
for (i = sh->disks; i--; ) { |
1da177e4c
|
320 |
struct r5dev *dev = &sh->dev[i]; |
d84e0f10d
|
321 |
if (dev->toread || dev->read || dev->towrite || dev->written || |
1da177e4c
|
322 |
test_bit(R5_LOCKED, &dev->flags)) { |
d84e0f10d
|
323 324 |
printk(KERN_ERR "sector=%llx i=%d %p %p %p %p %d ", |
1da177e4c
|
325 |
(unsigned long long)sh->sector, i, dev->toread, |
d84e0f10d
|
326 |
dev->read, dev->towrite, dev->written, |
1da177e4c
|
327 |
test_bit(R5_LOCKED, &dev->flags)); |
8cfa7b0f6
|
328 |
WARN_ON(1); |
1da177e4c
|
329 330 |
} dev->flags = 0; |
784052ecc
|
331 |
raid5_build_block(sh, i, previous); |
1da177e4c
|
332 333 334 |
} insert_hash(conf, sh); } |
d1688a6d5
|
335 |
static struct stripe_head *__find_stripe(struct r5conf *conf, sector_t sector, |
86b42c713
|
336 |
short generation) |
1da177e4c
|
337 338 |
{ struct stripe_head *sh; |
fccddba06
|
339 |
struct hlist_node *hn; |
1da177e4c
|
340 |
|
45b4233ca
|
341 342 |
pr_debug("__find_stripe, sector %llu ", (unsigned long long)sector); |
fccddba06
|
343 |
hlist_for_each_entry(sh, hn, stripe_hash(conf, sector), hash) |
86b42c713
|
344 |
if (sh->sector == sector && sh->generation == generation) |
1da177e4c
|
345 |
return sh; |
45b4233ca
|
346 347 |
pr_debug("__stripe %llu not in cache ", (unsigned long long)sector); |
1da177e4c
|
348 349 |
return NULL; } |
674806d62
|
350 351 352 353 354 355 356 357 358 359 360 361 362 |
/* * Need to check if array has failed when deciding whether to: * - start an array * - remove non-faulty devices * - add a spare * - allow a reshape * This determination is simple when no reshape is happening. * However if there is a reshape, we need to carefully check * both the before and after sections. * This is because some failed devices may only affect one * of the two sections, and some non-in_sync devices may * be insync in the section most affected by failed devices. */ |
d1688a6d5
|
363 |
static int has_failed(struct r5conf *conf) |
674806d62
|
364 365 366 367 368 369 370 371 372 |
{ int degraded; int i; if (conf->mddev->reshape_position == MaxSector) return conf->mddev->degraded > conf->max_degraded; rcu_read_lock(); degraded = 0; for (i = 0; i < conf->previous_raid_disks; i++) { |
3cb030020
|
373 |
struct md_rdev *rdev = rcu_dereference(conf->disks[i].rdev); |
674806d62
|
374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 |
if (!rdev || test_bit(Faulty, &rdev->flags)) degraded++; else if (test_bit(In_sync, &rdev->flags)) ; else /* not in-sync or faulty. * If the reshape increases the number of devices, * this is being recovered by the reshape, so * this 'previous' section is not in_sync. * If the number of devices is being reduced however, * the device can only be part of the array if * we are reverting a reshape, so this section will * be in-sync. */ if (conf->raid_disks >= conf->previous_raid_disks) degraded++; } rcu_read_unlock(); if (degraded > conf->max_degraded) return 1; rcu_read_lock(); degraded = 0; for (i = 0; i < conf->raid_disks; i++) { |
3cb030020
|
397 |
struct md_rdev *rdev = rcu_dereference(conf->disks[i].rdev); |
674806d62
|
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 |
if (!rdev || test_bit(Faulty, &rdev->flags)) degraded++; else if (test_bit(In_sync, &rdev->flags)) ; else /* not in-sync or faulty. * If reshape increases the number of devices, this * section has already been recovered, else it * almost certainly hasn't. */ if (conf->raid_disks <= conf->previous_raid_disks) degraded++; } rcu_read_unlock(); if (degraded > conf->max_degraded) return 1; return 0; } |
b5663ba40
|
416 |
static struct stripe_head * |
d1688a6d5
|
417 |
get_active_stripe(struct r5conf *conf, sector_t sector, |
a8c906ca3
|
418 |
int previous, int noblock, int noquiesce) |
1da177e4c
|
419 420 |
{ struct stripe_head *sh; |
45b4233ca
|
421 422 |
pr_debug("get_stripe, sector %llu ", (unsigned long long)sector); |
1da177e4c
|
423 424 425 426 |
spin_lock_irq(&conf->device_lock); do { |
72626685d
|
427 |
wait_event_lock_irq(conf->wait_for_stripe, |
a8c906ca3
|
428 |
conf->quiesce == 0 || noquiesce, |
72626685d
|
429 |
conf->device_lock, /* nothing */); |
86b42c713
|
430 |
sh = __find_stripe(conf, sector, conf->generation - previous); |
1da177e4c
|
431 432 433 434 435 436 437 438 439 |
if (!sh) { if (!conf->inactive_blocked) sh = get_free_stripe(conf); if (noblock && sh == NULL) break; if (!sh) { conf->inactive_blocked = 1; wait_event_lock_irq(conf->wait_for_stripe, !list_empty(&conf->inactive_list) && |
5036805be
|
440 441 |
(atomic_read(&conf->active_stripes) < (conf->max_nr_stripes *3/4) |
1da177e4c
|
442 443 |
|| !conf->inactive_blocked), conf->device_lock, |
7c13edc87
|
444 |
); |
1da177e4c
|
445 446 |
conf->inactive_blocked = 0; } else |
b5663ba40
|
447 |
init_stripe(sh, sector, previous); |
1da177e4c
|
448 449 |
} else { if (atomic_read(&sh->count)) { |
ab69ae12c
|
450 451 |
BUG_ON(!list_empty(&sh->lru) && !test_bit(STRIPE_EXPANDING, &sh->state)); |
1da177e4c
|
452 453 454 |
} else { if (!test_bit(STRIPE_HANDLE, &sh->state)) atomic_inc(&conf->active_stripes); |
ff4e8d9a9
|
455 456 |
if (list_empty(&sh->lru) && !test_bit(STRIPE_EXPANDING, &sh->state)) |
16a53ecc3
|
457 458 |
BUG(); list_del_init(&sh->lru); |
1da177e4c
|
459 460 461 462 463 464 465 466 467 468 |
} } } while (sh == NULL); if (sh) atomic_inc(&sh->count); spin_unlock_irq(&conf->device_lock); return sh; } |
6712ecf8f
|
469 470 471 472 |
static void raid5_end_read_request(struct bio *bi, int error); static void raid5_end_write_request(struct bio *bi, int error); |
91c009248
|
473 |
|
c4e5ac0a2
|
474 |
static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s) |
91c009248
|
475 |
{ |
d1688a6d5
|
476 |
struct r5conf *conf = sh->raid_conf; |
91c009248
|
477 478 479 480 481 482 483 |
int i, disks = sh->disks; might_sleep(); for (i = disks; i--; ) { int rw; struct bio *bi; |
3cb030020
|
484 |
struct md_rdev *rdev; |
e9c7469bb
|
485 486 487 488 489 490 |
if (test_and_clear_bit(R5_Wantwrite, &sh->dev[i].flags)) { if (test_and_clear_bit(R5_WantFUA, &sh->dev[i].flags)) rw = WRITE_FUA; else rw = WRITE; } else if (test_and_clear_bit(R5_Wantread, &sh->dev[i].flags)) |
91c009248
|
491 492 493 494 495 496 497 |
rw = READ; else continue; bi = &sh->dev[i].req; bi->bi_rw = rw; |
b062962ed
|
498 |
if (rw & WRITE) |
91c009248
|
499 500 501 502 503 504 505 506 507 508 509 |
bi->bi_end_io = raid5_end_write_request; else bi->bi_end_io = raid5_end_read_request; rcu_read_lock(); rdev = rcu_dereference(conf->disks[i].rdev); if (rdev && test_bit(Faulty, &rdev->flags)) rdev = NULL; if (rdev) atomic_inc(&rdev->nr_pending); rcu_read_unlock(); |
73e92e51b
|
510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 |
/* We have already checked bad blocks for reads. Now * need to check for writes. */ while ((rw & WRITE) && rdev && test_bit(WriteErrorSeen, &rdev->flags)) { sector_t first_bad; int bad_sectors; int bad = is_badblock(rdev, sh->sector, STRIPE_SECTORS, &first_bad, &bad_sectors); if (!bad) break; if (bad < 0) { set_bit(BlockedBadBlocks, &rdev->flags); if (!conf->mddev->external && conf->mddev->flags) { /* It is very unlikely, but we might * still need to write out the * bad block log - better give it * a chance*/ md_check_recovery(conf->mddev); } md_wait_for_blocked_rdev(rdev, conf->mddev); } else { /* Acknowledged bad block - skip the write */ rdev_dec_pending(rdev, conf->mddev); rdev = NULL; } } |
91c009248
|
539 |
if (rdev) { |
c4e5ac0a2
|
540 |
if (s->syncing || s->expanding || s->expanded) |
91c009248
|
541 |
md_sync_acct(rdev->bdev, STRIPE_SECTORS); |
2b7497f0e
|
542 |
set_bit(STRIPE_IO_STARTED, &sh->state); |
91c009248
|
543 544 545 |
bi->bi_bdev = rdev->bdev; pr_debug("%s: for %llu schedule op %ld on disc %d ", |
e46b272b6
|
546 |
__func__, (unsigned long long)sh->sector, |
91c009248
|
547 548 549 550 551 552 553 554 555 556 557 558 |
bi->bi_rw, i); atomic_inc(&sh->count); bi->bi_sector = sh->sector + rdev->data_offset; bi->bi_flags = 1 << BIO_UPTODATE; bi->bi_vcnt = 1; bi->bi_max_vecs = 1; bi->bi_idx = 0; bi->bi_io_vec = &sh->dev[i].vec; bi->bi_io_vec[0].bv_len = STRIPE_SIZE; bi->bi_io_vec[0].bv_offset = 0; bi->bi_size = STRIPE_SIZE; bi->bi_next = NULL; |
91c009248
|
559 560 |
generic_make_request(bi); } else { |
b062962ed
|
561 |
if (rw & WRITE) |
91c009248
|
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 |
set_bit(STRIPE_DEGRADED, &sh->state); pr_debug("skip op %ld on disc %d for sector %llu ", bi->bi_rw, i, (unsigned long long)sh->sector); clear_bit(R5_LOCKED, &sh->dev[i].flags); set_bit(STRIPE_HANDLE, &sh->state); } } } static struct dma_async_tx_descriptor * async_copy_data(int frombio, struct bio *bio, struct page *page, sector_t sector, struct dma_async_tx_descriptor *tx) { struct bio_vec *bvl; struct page *bio_page; int i; int page_offset; |
a08abd8ca
|
580 |
struct async_submit_ctl submit; |
0403e3827
|
581 |
enum async_tx_flags flags = 0; |
91c009248
|
582 583 584 585 586 |
if (bio->bi_sector >= sector) page_offset = (signed)(bio->bi_sector - sector) * 512; else page_offset = (signed)(sector - bio->bi_sector) * -512; |
a08abd8ca
|
587 |
|
0403e3827
|
588 589 590 |
if (frombio) flags |= ASYNC_TX_FENCE; init_async_submit(&submit, flags, tx, NULL, NULL, NULL); |
91c009248
|
591 |
bio_for_each_segment(bvl, bio, i) { |
fcde90759
|
592 |
int len = bvl->bv_len; |
91c009248
|
593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 |
int clen; int b_offset = 0; if (page_offset < 0) { b_offset = -page_offset; page_offset += b_offset; len -= b_offset; } if (len > 0 && page_offset + len > STRIPE_SIZE) clen = STRIPE_SIZE - page_offset; else clen = len; if (clen > 0) { |
fcde90759
|
608 609 |
b_offset += bvl->bv_offset; bio_page = bvl->bv_page; |
91c009248
|
610 611 |
if (frombio) tx = async_memcpy(page, bio_page, page_offset, |
a08abd8ca
|
612 |
b_offset, clen, &submit); |
91c009248
|
613 614 |
else tx = async_memcpy(bio_page, page, b_offset, |
a08abd8ca
|
615 |
page_offset, clen, &submit); |
91c009248
|
616 |
} |
a08abd8ca
|
617 618 |
/* chain the operations */ submit.depend_tx = tx; |
91c009248
|
619 620 621 622 623 624 625 626 627 628 629 630 |
if (clen < len) /* hit end of page */ break; page_offset += len; } return tx; } static void ops_complete_biofill(void *stripe_head_ref) { struct stripe_head *sh = stripe_head_ref; struct bio *return_bi = NULL; |
d1688a6d5
|
631 |
struct r5conf *conf = sh->raid_conf; |
e4d84909d
|
632 |
int i; |
91c009248
|
633 |
|
e46b272b6
|
634 635 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
636 637 638 |
(unsigned long long)sh->sector); /* clear completed biofills */ |
83de75cc9
|
639 |
spin_lock_irq(&conf->device_lock); |
91c009248
|
640 641 |
for (i = sh->disks; i--; ) { struct r5dev *dev = &sh->dev[i]; |
91c009248
|
642 643 |
/* acknowledge completion of a biofill operation */ |
e4d84909d
|
644 645 |
/* and check if we need to reply to a read request, * new R5_Wantfill requests are held off until |
83de75cc9
|
646 |
* !STRIPE_BIOFILL_RUN |
e4d84909d
|
647 648 |
*/ if (test_and_clear_bit(R5_Wantfill, &dev->flags)) { |
91c009248
|
649 |
struct bio *rbi, *rbi2; |
91c009248
|
650 |
|
91c009248
|
651 652 653 654 655 656 |
BUG_ON(!dev->read); rbi = dev->read; dev->read = NULL; while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) { rbi2 = r5_next_bio(rbi, dev->sector); |
960e739d9
|
657 |
if (!raid5_dec_bi_phys_segments(rbi)) { |
91c009248
|
658 659 660 |
rbi->bi_next = return_bi; return_bi = rbi; } |
91c009248
|
661 662 663 664 |
rbi = rbi2; } } } |
83de75cc9
|
665 666 |
spin_unlock_irq(&conf->device_lock); clear_bit(STRIPE_BIOFILL_RUN, &sh->state); |
91c009248
|
667 668 |
return_io(return_bi); |
e4d84909d
|
669 |
set_bit(STRIPE_HANDLE, &sh->state); |
91c009248
|
670 671 672 673 674 675 |
release_stripe(sh); } static void ops_run_biofill(struct stripe_head *sh) { struct dma_async_tx_descriptor *tx = NULL; |
d1688a6d5
|
676 |
struct r5conf *conf = sh->raid_conf; |
a08abd8ca
|
677 |
struct async_submit_ctl submit; |
91c009248
|
678 |
int i; |
e46b272b6
|
679 680 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 |
(unsigned long long)sh->sector); for (i = sh->disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (test_bit(R5_Wantfill, &dev->flags)) { struct bio *rbi; spin_lock_irq(&conf->device_lock); dev->read = rbi = dev->toread; dev->toread = NULL; spin_unlock_irq(&conf->device_lock); while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) { tx = async_copy_data(0, rbi, dev->page, dev->sector, tx); rbi = r5_next_bio(rbi, dev->sector); } } } atomic_inc(&sh->count); |
a08abd8ca
|
701 702 |
init_async_submit(&submit, ASYNC_TX_ACK, tx, ops_complete_biofill, sh, NULL); async_trigger_callback(&submit); |
91c009248
|
703 |
} |
4e7d2c0ae
|
704 |
static void mark_target_uptodate(struct stripe_head *sh, int target) |
91c009248
|
705 |
{ |
4e7d2c0ae
|
706 |
struct r5dev *tgt; |
91c009248
|
707 |
|
4e7d2c0ae
|
708 709 |
if (target < 0) return; |
91c009248
|
710 |
|
4e7d2c0ae
|
711 |
tgt = &sh->dev[target]; |
91c009248
|
712 713 714 |
set_bit(R5_UPTODATE, &tgt->flags); BUG_ON(!test_bit(R5_Wantcompute, &tgt->flags)); clear_bit(R5_Wantcompute, &tgt->flags); |
4e7d2c0ae
|
715 |
} |
ac6b53b6e
|
716 |
static void ops_complete_compute(void *stripe_head_ref) |
91c009248
|
717 718 |
{ struct stripe_head *sh = stripe_head_ref; |
91c009248
|
719 |
|
e46b272b6
|
720 721 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
722 |
(unsigned long long)sh->sector); |
ac6b53b6e
|
723 |
/* mark the computed target(s) as uptodate */ |
4e7d2c0ae
|
724 |
mark_target_uptodate(sh, sh->ops.target); |
ac6b53b6e
|
725 |
mark_target_uptodate(sh, sh->ops.target2); |
4e7d2c0ae
|
726 |
|
ecc65c9b3
|
727 728 729 |
clear_bit(STRIPE_COMPUTE_RUN, &sh->state); if (sh->check_state == check_state_compute_run) sh->check_state = check_state_compute_result; |
91c009248
|
730 731 732 |
set_bit(STRIPE_HANDLE, &sh->state); release_stripe(sh); } |
d6f38f31f
|
733 734 735 736 737 738 739 740 741 |
/* return a pointer to the address conversion region of the scribble buffer */ static addr_conv_t *to_addr_conv(struct stripe_head *sh, struct raid5_percpu *percpu) { return percpu->scribble + sizeof(struct page *) * (sh->disks + 2); } static struct dma_async_tx_descriptor * ops_run_compute5(struct stripe_head *sh, struct raid5_percpu *percpu) |
91c009248
|
742 |
{ |
91c009248
|
743 |
int disks = sh->disks; |
d6f38f31f
|
744 |
struct page **xor_srcs = percpu->scribble; |
91c009248
|
745 746 747 748 749 |
int target = sh->ops.target; struct r5dev *tgt = &sh->dev[target]; struct page *xor_dest = tgt->page; int count = 0; struct dma_async_tx_descriptor *tx; |
a08abd8ca
|
750 |
struct async_submit_ctl submit; |
91c009248
|
751 752 753 754 |
int i; pr_debug("%s: stripe %llu block: %d ", |
e46b272b6
|
755 |
__func__, (unsigned long long)sh->sector, target); |
91c009248
|
756 757 758 759 760 761 762 |
BUG_ON(!test_bit(R5_Wantcompute, &tgt->flags)); for (i = disks; i--; ) if (i != target) xor_srcs[count++] = sh->dev[i].page; atomic_inc(&sh->count); |
0403e3827
|
763 |
init_async_submit(&submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_ZERO_DST, NULL, |
ac6b53b6e
|
764 |
ops_complete_compute, sh, to_addr_conv(sh, percpu)); |
91c009248
|
765 |
if (unlikely(count == 1)) |
a08abd8ca
|
766 |
tx = async_memcpy(xor_dest, xor_srcs[0], 0, 0, STRIPE_SIZE, &submit); |
91c009248
|
767 |
else |
a08abd8ca
|
768 |
tx = async_xor(xor_dest, xor_srcs, 0, count, STRIPE_SIZE, &submit); |
91c009248
|
769 |
|
91c009248
|
770 771 |
return tx; } |
ac6b53b6e
|
772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 |
/* set_syndrome_sources - populate source buffers for gen_syndrome * @srcs - (struct page *) array of size sh->disks * @sh - stripe_head to parse * * Populates srcs in proper layout order for the stripe and returns the * 'count' of sources to be used in a call to async_gen_syndrome. The P * destination buffer is recorded in srcs[count] and the Q destination * is recorded in srcs[count+1]]. */ static int set_syndrome_sources(struct page **srcs, struct stripe_head *sh) { int disks = sh->disks; int syndrome_disks = sh->ddf_layout ? disks : (disks - 2); int d0_idx = raid6_d0(sh); int count; int i; for (i = 0; i < disks; i++) |
5dd33c9a4
|
790 |
srcs[i] = NULL; |
ac6b53b6e
|
791 792 793 794 795 796 797 798 799 |
count = 0; i = d0_idx; do { int slot = raid6_idx_to_slot(i, sh, &count, syndrome_disks); srcs[slot] = sh->dev[i].page; i = raid6_next_disk(i, disks); } while (i != d0_idx); |
ac6b53b6e
|
800 |
|
e4424fee1
|
801 |
return syndrome_disks; |
ac6b53b6e
|
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 |
} static struct dma_async_tx_descriptor * ops_run_compute6_1(struct stripe_head *sh, struct raid5_percpu *percpu) { int disks = sh->disks; struct page **blocks = percpu->scribble; int target; int qd_idx = sh->qd_idx; struct dma_async_tx_descriptor *tx; struct async_submit_ctl submit; struct r5dev *tgt; struct page *dest; int i; int count; if (sh->ops.target < 0) target = sh->ops.target2; else if (sh->ops.target2 < 0) target = sh->ops.target; |
91c009248
|
822 |
else |
ac6b53b6e
|
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 |
/* we should only have one valid target */ BUG(); BUG_ON(target < 0); pr_debug("%s: stripe %llu block: %d ", __func__, (unsigned long long)sh->sector, target); tgt = &sh->dev[target]; BUG_ON(!test_bit(R5_Wantcompute, &tgt->flags)); dest = tgt->page; atomic_inc(&sh->count); if (target == qd_idx) { count = set_syndrome_sources(blocks, sh); blocks[count] = NULL; /* regenerating p is not necessary */ BUG_ON(blocks[count+1] != dest); /* q should already be set */ |
0403e3827
|
840 841 |
init_async_submit(&submit, ASYNC_TX_FENCE, NULL, ops_complete_compute, sh, |
ac6b53b6e
|
842 843 844 845 846 847 848 849 850 851 |
to_addr_conv(sh, percpu)); tx = async_gen_syndrome(blocks, 0, count+2, STRIPE_SIZE, &submit); } else { /* Compute any data- or p-drive using XOR */ count = 0; for (i = disks; i-- ; ) { if (i == target || i == qd_idx) continue; blocks[count++] = sh->dev[i].page; } |
0403e3827
|
852 853 |
init_async_submit(&submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_ZERO_DST, NULL, ops_complete_compute, sh, |
ac6b53b6e
|
854 855 856 |
to_addr_conv(sh, percpu)); tx = async_xor(dest, blocks, 0, count, STRIPE_SIZE, &submit); } |
91c009248
|
857 |
|
91c009248
|
858 859 |
return tx; } |
ac6b53b6e
|
860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 |
static struct dma_async_tx_descriptor * ops_run_compute6_2(struct stripe_head *sh, struct raid5_percpu *percpu) { int i, count, disks = sh->disks; int syndrome_disks = sh->ddf_layout ? disks : disks-2; int d0_idx = raid6_d0(sh); int faila = -1, failb = -1; int target = sh->ops.target; int target2 = sh->ops.target2; struct r5dev *tgt = &sh->dev[target]; struct r5dev *tgt2 = &sh->dev[target2]; struct dma_async_tx_descriptor *tx; struct page **blocks = percpu->scribble; struct async_submit_ctl submit; pr_debug("%s: stripe %llu block1: %d block2: %d ", __func__, (unsigned long long)sh->sector, target, target2); BUG_ON(target < 0 || target2 < 0); BUG_ON(!test_bit(R5_Wantcompute, &tgt->flags)); BUG_ON(!test_bit(R5_Wantcompute, &tgt2->flags)); |
6c910a78e
|
881 |
/* we need to open-code set_syndrome_sources to handle the |
ac6b53b6e
|
882 883 884 |
* slot number conversion for 'faila' and 'failb' */ for (i = 0; i < disks ; i++) |
5dd33c9a4
|
885 |
blocks[i] = NULL; |
ac6b53b6e
|
886 887 888 889 890 891 892 893 894 895 896 897 898 |
count = 0; i = d0_idx; do { int slot = raid6_idx_to_slot(i, sh, &count, syndrome_disks); blocks[slot] = sh->dev[i].page; if (i == target) faila = slot; if (i == target2) failb = slot; i = raid6_next_disk(i, disks); } while (i != d0_idx); |
ac6b53b6e
|
899 900 901 902 903 904 905 906 907 908 909 910 911 912 |
BUG_ON(faila == failb); if (failb < faila) swap(faila, failb); pr_debug("%s: stripe: %llu faila: %d failb: %d ", __func__, (unsigned long long)sh->sector, faila, failb); atomic_inc(&sh->count); if (failb == syndrome_disks+1) { /* Q disk is one of the missing disks */ if (faila == syndrome_disks) { /* Missing P+Q, just recompute */ |
0403e3827
|
913 914 915 |
init_async_submit(&submit, ASYNC_TX_FENCE, NULL, ops_complete_compute, sh, to_addr_conv(sh, percpu)); |
e4424fee1
|
916 |
return async_gen_syndrome(blocks, 0, syndrome_disks+2, |
ac6b53b6e
|
917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 |
STRIPE_SIZE, &submit); } else { struct page *dest; int data_target; int qd_idx = sh->qd_idx; /* Missing D+Q: recompute D from P, then recompute Q */ if (target == qd_idx) data_target = target2; else data_target = target; count = 0; for (i = disks; i-- ; ) { if (i == data_target || i == qd_idx) continue; blocks[count++] = sh->dev[i].page; } dest = sh->dev[data_target].page; |
0403e3827
|
936 937 938 939 |
init_async_submit(&submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_ZERO_DST, NULL, NULL, NULL, to_addr_conv(sh, percpu)); |
ac6b53b6e
|
940 941 942 943 |
tx = async_xor(dest, blocks, 0, count, STRIPE_SIZE, &submit); count = set_syndrome_sources(blocks, sh); |
0403e3827
|
944 945 946 |
init_async_submit(&submit, ASYNC_TX_FENCE, tx, ops_complete_compute, sh, to_addr_conv(sh, percpu)); |
ac6b53b6e
|
947 948 949 |
return async_gen_syndrome(blocks, 0, count+2, STRIPE_SIZE, &submit); } |
ac6b53b6e
|
950 |
} else { |
6c910a78e
|
951 952 953 954 955 956 957 958 959 960 961 962 963 964 |
init_async_submit(&submit, ASYNC_TX_FENCE, NULL, ops_complete_compute, sh, to_addr_conv(sh, percpu)); if (failb == syndrome_disks) { /* We're missing D+P. */ return async_raid6_datap_recov(syndrome_disks+2, STRIPE_SIZE, faila, blocks, &submit); } else { /* We're missing D+D. */ return async_raid6_2data_recov(syndrome_disks+2, STRIPE_SIZE, faila, failb, blocks, &submit); } |
ac6b53b6e
|
965 966 |
} } |
91c009248
|
967 968 969 |
static void ops_complete_prexor(void *stripe_head_ref) { struct stripe_head *sh = stripe_head_ref; |
e46b272b6
|
970 971 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
972 |
(unsigned long long)sh->sector); |
91c009248
|
973 974 975 |
} static struct dma_async_tx_descriptor * |
d6f38f31f
|
976 977 |
ops_run_prexor(struct stripe_head *sh, struct raid5_percpu *percpu, struct dma_async_tx_descriptor *tx) |
91c009248
|
978 |
{ |
91c009248
|
979 |
int disks = sh->disks; |
d6f38f31f
|
980 |
struct page **xor_srcs = percpu->scribble; |
91c009248
|
981 |
int count = 0, pd_idx = sh->pd_idx, i; |
a08abd8ca
|
982 |
struct async_submit_ctl submit; |
91c009248
|
983 984 985 |
/* existing parity data subtracted */ struct page *xor_dest = xor_srcs[count++] = sh->dev[pd_idx].page; |
e46b272b6
|
986 987 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
988 989 990 991 992 |
(unsigned long long)sh->sector); for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; /* Only process blocks that are known to be uptodate */ |
d8ee0728b
|
993 |
if (test_bit(R5_Wantdrain, &dev->flags)) |
91c009248
|
994 995 |
xor_srcs[count++] = dev->page; } |
0403e3827
|
996 |
init_async_submit(&submit, ASYNC_TX_FENCE|ASYNC_TX_XOR_DROP_DST, tx, |
d6f38f31f
|
997 |
ops_complete_prexor, sh, to_addr_conv(sh, percpu)); |
a08abd8ca
|
998 |
tx = async_xor(xor_dest, xor_srcs, 0, count, STRIPE_SIZE, &submit); |
91c009248
|
999 1000 1001 1002 1003 |
return tx; } static struct dma_async_tx_descriptor * |
d8ee0728b
|
1004 |
ops_run_biodrain(struct stripe_head *sh, struct dma_async_tx_descriptor *tx) |
91c009248
|
1005 1006 |
{ int disks = sh->disks; |
d8ee0728b
|
1007 |
int i; |
91c009248
|
1008 |
|
e46b272b6
|
1009 1010 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
1011 1012 1013 1014 1015 |
(unsigned long long)sh->sector); for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; struct bio *chosen; |
91c009248
|
1016 |
|
d8ee0728b
|
1017 |
if (test_and_clear_bit(R5_Wantdrain, &dev->flags)) { |
91c009248
|
1018 |
struct bio *wbi; |
cbe47ec55
|
1019 |
spin_lock_irq(&sh->raid_conf->device_lock); |
91c009248
|
1020 1021 1022 1023 |
chosen = dev->towrite; dev->towrite = NULL; BUG_ON(dev->written); wbi = dev->written = chosen; |
cbe47ec55
|
1024 |
spin_unlock_irq(&sh->raid_conf->device_lock); |
91c009248
|
1025 1026 1027 |
while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) { |
e9c7469bb
|
1028 1029 |
if (wbi->bi_rw & REQ_FUA) set_bit(R5_WantFUA, &dev->flags); |
91c009248
|
1030 1031 1032 1033 1034 1035 1036 1037 1038 |
tx = async_copy_data(1, wbi, dev->page, dev->sector, tx); wbi = r5_next_bio(wbi, dev->sector); } } } return tx; } |
ac6b53b6e
|
1039 |
static void ops_complete_reconstruct(void *stripe_head_ref) |
91c009248
|
1040 1041 |
{ struct stripe_head *sh = stripe_head_ref; |
ac6b53b6e
|
1042 1043 1044 1045 |
int disks = sh->disks; int pd_idx = sh->pd_idx; int qd_idx = sh->qd_idx; int i; |
e9c7469bb
|
1046 |
bool fua = false; |
91c009248
|
1047 |
|
e46b272b6
|
1048 1049 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
1050 |
(unsigned long long)sh->sector); |
e9c7469bb
|
1051 1052 |
for (i = disks; i--; ) fua |= test_bit(R5_WantFUA, &sh->dev[i].flags); |
91c009248
|
1053 1054 |
for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; |
ac6b53b6e
|
1055 |
|
e9c7469bb
|
1056 |
if (dev->written || i == pd_idx || i == qd_idx) { |
91c009248
|
1057 |
set_bit(R5_UPTODATE, &dev->flags); |
e9c7469bb
|
1058 1059 1060 |
if (fua) set_bit(R5_WantFUA, &dev->flags); } |
91c009248
|
1061 |
} |
d8ee0728b
|
1062 1063 1064 1065 1066 1067 1068 1069 |
if (sh->reconstruct_state == reconstruct_state_drain_run) sh->reconstruct_state = reconstruct_state_drain_result; else if (sh->reconstruct_state == reconstruct_state_prexor_drain_run) sh->reconstruct_state = reconstruct_state_prexor_drain_result; else { BUG_ON(sh->reconstruct_state != reconstruct_state_run); sh->reconstruct_state = reconstruct_state_result; } |
91c009248
|
1070 1071 1072 1073 1074 1075 |
set_bit(STRIPE_HANDLE, &sh->state); release_stripe(sh); } static void |
ac6b53b6e
|
1076 1077 |
ops_run_reconstruct5(struct stripe_head *sh, struct raid5_percpu *percpu, struct dma_async_tx_descriptor *tx) |
91c009248
|
1078 |
{ |
91c009248
|
1079 |
int disks = sh->disks; |
d6f38f31f
|
1080 |
struct page **xor_srcs = percpu->scribble; |
a08abd8ca
|
1081 |
struct async_submit_ctl submit; |
91c009248
|
1082 1083 |
int count = 0, pd_idx = sh->pd_idx, i; struct page *xor_dest; |
d8ee0728b
|
1084 |
int prexor = 0; |
91c009248
|
1085 |
unsigned long flags; |
91c009248
|
1086 |
|
e46b272b6
|
1087 1088 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
1089 1090 1091 1092 1093 |
(unsigned long long)sh->sector); /* check if prexor is active which means only process blocks * that are part of a read-modify-write (written) */ |
d8ee0728b
|
1094 1095 |
if (sh->reconstruct_state == reconstruct_state_prexor_drain_run) { prexor = 1; |
91c009248
|
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 |
xor_dest = xor_srcs[count++] = sh->dev[pd_idx].page; for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (dev->written) xor_srcs[count++] = dev->page; } } else { xor_dest = sh->dev[pd_idx].page; for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (i != pd_idx) xor_srcs[count++] = dev->page; } } |
91c009248
|
1110 1111 1112 1113 1114 |
/* 1/ if we prexor'd then the dest is reused as a source * 2/ if we did not prexor then we are redoing the parity * set ASYNC_TX_XOR_DROP_DST and ASYNC_TX_XOR_ZERO_DST * for the synchronous xor case */ |
88ba2aa58
|
1115 |
flags = ASYNC_TX_ACK | |
91c009248
|
1116 1117 1118 |
(prexor ? ASYNC_TX_XOR_DROP_DST : ASYNC_TX_XOR_ZERO_DST); atomic_inc(&sh->count); |
ac6b53b6e
|
1119 |
init_async_submit(&submit, flags, tx, ops_complete_reconstruct, sh, |
d6f38f31f
|
1120 |
to_addr_conv(sh, percpu)); |
a08abd8ca
|
1121 1122 1123 1124 |
if (unlikely(count == 1)) tx = async_memcpy(xor_dest, xor_srcs[0], 0, 0, STRIPE_SIZE, &submit); else tx = async_xor(xor_dest, xor_srcs, 0, count, STRIPE_SIZE, &submit); |
91c009248
|
1125 |
} |
ac6b53b6e
|
1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 |
static void ops_run_reconstruct6(struct stripe_head *sh, struct raid5_percpu *percpu, struct dma_async_tx_descriptor *tx) { struct async_submit_ctl submit; struct page **blocks = percpu->scribble; int count; pr_debug("%s: stripe %llu ", __func__, (unsigned long long)sh->sector); count = set_syndrome_sources(blocks, sh); atomic_inc(&sh->count); init_async_submit(&submit, ASYNC_TX_ACK, tx, ops_complete_reconstruct, sh, to_addr_conv(sh, percpu)); async_gen_syndrome(blocks, 0, count+2, STRIPE_SIZE, &submit); |
91c009248
|
1144 1145 1146 1147 1148 |
} static void ops_complete_check(void *stripe_head_ref) { struct stripe_head *sh = stripe_head_ref; |
91c009248
|
1149 |
|
e46b272b6
|
1150 1151 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
1152 |
(unsigned long long)sh->sector); |
ecc65c9b3
|
1153 |
sh->check_state = check_state_check_result; |
91c009248
|
1154 1155 1156 |
set_bit(STRIPE_HANDLE, &sh->state); release_stripe(sh); } |
ac6b53b6e
|
1157 |
static void ops_run_check_p(struct stripe_head *sh, struct raid5_percpu *percpu) |
91c009248
|
1158 |
{ |
91c009248
|
1159 |
int disks = sh->disks; |
ac6b53b6e
|
1160 1161 1162 |
int pd_idx = sh->pd_idx; int qd_idx = sh->qd_idx; struct page *xor_dest; |
d6f38f31f
|
1163 |
struct page **xor_srcs = percpu->scribble; |
91c009248
|
1164 |
struct dma_async_tx_descriptor *tx; |
a08abd8ca
|
1165 |
struct async_submit_ctl submit; |
ac6b53b6e
|
1166 1167 |
int count; int i; |
91c009248
|
1168 |
|
e46b272b6
|
1169 1170 |
pr_debug("%s: stripe %llu ", __func__, |
91c009248
|
1171 |
(unsigned long long)sh->sector); |
ac6b53b6e
|
1172 1173 1174 |
count = 0; xor_dest = sh->dev[pd_idx].page; xor_srcs[count++] = xor_dest; |
91c009248
|
1175 |
for (i = disks; i--; ) { |
ac6b53b6e
|
1176 1177 1178 |
if (i == pd_idx || i == qd_idx) continue; xor_srcs[count++] = sh->dev[i].page; |
91c009248
|
1179 |
} |
d6f38f31f
|
1180 1181 |
init_async_submit(&submit, 0, NULL, NULL, NULL, to_addr_conv(sh, percpu)); |
099f53cb5
|
1182 |
tx = async_xor_val(xor_dest, xor_srcs, 0, count, STRIPE_SIZE, |
a08abd8ca
|
1183 |
&sh->ops.zero_sum_result, &submit); |
91c009248
|
1184 |
|
91c009248
|
1185 |
atomic_inc(&sh->count); |
a08abd8ca
|
1186 1187 |
init_async_submit(&submit, ASYNC_TX_ACK, tx, ops_complete_check, sh, NULL); tx = async_trigger_callback(&submit); |
91c009248
|
1188 |
} |
ac6b53b6e
|
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 |
static void ops_run_check_pq(struct stripe_head *sh, struct raid5_percpu *percpu, int checkp) { struct page **srcs = percpu->scribble; struct async_submit_ctl submit; int count; pr_debug("%s: stripe %llu checkp: %d ", __func__, (unsigned long long)sh->sector, checkp); count = set_syndrome_sources(srcs, sh); if (!checkp) srcs[count] = NULL; |
91c009248
|
1202 |
|
91c009248
|
1203 |
atomic_inc(&sh->count); |
ac6b53b6e
|
1204 1205 1206 1207 |
init_async_submit(&submit, ASYNC_TX_ACK, NULL, ops_complete_check, sh, to_addr_conv(sh, percpu)); async_syndrome_val(srcs, 0, count+2, STRIPE_SIZE, &sh->ops.zero_sum_result, percpu->spare_page, &submit); |
91c009248
|
1208 |
} |
417b8d4ac
|
1209 |
static void __raid_run_ops(struct stripe_head *sh, unsigned long ops_request) |
91c009248
|
1210 1211 1212 |
{ int overlap_clear = 0, i, disks = sh->disks; struct dma_async_tx_descriptor *tx = NULL; |
d1688a6d5
|
1213 |
struct r5conf *conf = sh->raid_conf; |
ac6b53b6e
|
1214 |
int level = conf->level; |
d6f38f31f
|
1215 1216 |
struct raid5_percpu *percpu; unsigned long cpu; |
91c009248
|
1217 |
|
d6f38f31f
|
1218 1219 |
cpu = get_cpu(); percpu = per_cpu_ptr(conf->percpu, cpu); |
83de75cc9
|
1220 |
if (test_bit(STRIPE_OP_BIOFILL, &ops_request)) { |
91c009248
|
1221 1222 1223 |
ops_run_biofill(sh); overlap_clear++; } |
7b3a871ed
|
1224 |
if (test_bit(STRIPE_OP_COMPUTE_BLK, &ops_request)) { |
ac6b53b6e
|
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 |
if (level < 6) tx = ops_run_compute5(sh, percpu); else { if (sh->ops.target2 < 0 || sh->ops.target < 0) tx = ops_run_compute6_1(sh, percpu); else tx = ops_run_compute6_2(sh, percpu); } /* terminate the chain if reconstruct is not set to be run */ if (tx && !test_bit(STRIPE_OP_RECONSTRUCT, &ops_request)) |
7b3a871ed
|
1235 1236 |
async_tx_ack(tx); } |
91c009248
|
1237 |
|
600aa1099
|
1238 |
if (test_bit(STRIPE_OP_PREXOR, &ops_request)) |
d6f38f31f
|
1239 |
tx = ops_run_prexor(sh, percpu, tx); |
91c009248
|
1240 |
|
600aa1099
|
1241 |
if (test_bit(STRIPE_OP_BIODRAIN, &ops_request)) { |
d8ee0728b
|
1242 |
tx = ops_run_biodrain(sh, tx); |
91c009248
|
1243 1244 |
overlap_clear++; } |
ac6b53b6e
|
1245 1246 1247 1248 1249 1250 |
if (test_bit(STRIPE_OP_RECONSTRUCT, &ops_request)) { if (level < 6) ops_run_reconstruct5(sh, percpu, tx); else ops_run_reconstruct6(sh, percpu, tx); } |
91c009248
|
1251 |
|
ac6b53b6e
|
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 |
if (test_bit(STRIPE_OP_CHECK, &ops_request)) { if (sh->check_state == check_state_run) ops_run_check_p(sh, percpu); else if (sh->check_state == check_state_run_q) ops_run_check_pq(sh, percpu, 0); else if (sh->check_state == check_state_run_pq) ops_run_check_pq(sh, percpu, 1); else BUG(); } |
91c009248
|
1262 |
|
91c009248
|
1263 1264 1265 1266 1267 1268 |
if (overlap_clear) for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (test_and_clear_bit(R5_Overlap, &dev->flags)) wake_up(&sh->raid_conf->wait_for_overlap); } |
d6f38f31f
|
1269 |
put_cpu(); |
91c009248
|
1270 |
} |
417b8d4ac
|
1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 |
#ifdef CONFIG_MULTICORE_RAID456 static void async_run_ops(void *param, async_cookie_t cookie) { struct stripe_head *sh = param; unsigned long ops_request = sh->ops.request; clear_bit_unlock(STRIPE_OPS_REQ_PENDING, &sh->state); wake_up(&sh->ops.wait_for_ops); __raid_run_ops(sh, ops_request); release_stripe(sh); } static void raid_run_ops(struct stripe_head *sh, unsigned long ops_request) { /* since handle_stripe can be called outside of raid5d context * we need to ensure sh->ops.request is de-staged before another * request arrives */ wait_event(sh->ops.wait_for_ops, !test_and_set_bit_lock(STRIPE_OPS_REQ_PENDING, &sh->state)); sh->ops.request = ops_request; atomic_inc(&sh->count); async_schedule(async_run_ops, sh); } #else #define raid_run_ops __raid_run_ops #endif |
d1688a6d5
|
1300 |
static int grow_one_stripe(struct r5conf *conf) |
1da177e4c
|
1301 1302 |
{ struct stripe_head *sh; |
6ce328462
|
1303 |
sh = kmem_cache_zalloc(conf->slab_cache, GFP_KERNEL); |
3f294f4fb
|
1304 1305 |
if (!sh) return 0; |
6ce328462
|
1306 |
|
3f294f4fb
|
1307 |
sh->raid_conf = conf; |
417b8d4ac
|
1308 1309 1310 |
#ifdef CONFIG_MULTICORE_RAID456 init_waitqueue_head(&sh->ops.wait_for_ops); #endif |
3f294f4fb
|
1311 |
|
e4e11e385
|
1312 1313 |
if (grow_buffers(sh)) { shrink_buffers(sh); |
3f294f4fb
|
1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 |
kmem_cache_free(conf->slab_cache, sh); return 0; } /* we just created an active stripe so... */ atomic_set(&sh->count, 1); atomic_inc(&conf->active_stripes); INIT_LIST_HEAD(&sh->lru); release_stripe(sh); return 1; } |
d1688a6d5
|
1324 |
static int grow_stripes(struct r5conf *conf, int num) |
3f294f4fb
|
1325 |
{ |
e18b890bb
|
1326 |
struct kmem_cache *sc; |
5e5e3e78e
|
1327 |
int devs = max(conf->raid_disks, conf->previous_raid_disks); |
1da177e4c
|
1328 |
|
f4be6b43f
|
1329 1330 1331 1332 1333 1334 1335 |
if (conf->mddev->gendisk) sprintf(conf->cache_name[0], "raid%d-%s", conf->level, mdname(conf->mddev)); else sprintf(conf->cache_name[0], "raid%d-%p", conf->level, conf->mddev); sprintf(conf->cache_name[1], "%s-alt", conf->cache_name[0]); |
ad01c9e37
|
1336 1337 |
conf->active_name = 0; sc = kmem_cache_create(conf->cache_name[conf->active_name], |
1da177e4c
|
1338 |
sizeof(struct stripe_head)+(devs-1)*sizeof(struct r5dev), |
20c2df83d
|
1339 |
0, 0, NULL); |
1da177e4c
|
1340 1341 1342 |
if (!sc) return 1; conf->slab_cache = sc; |
ad01c9e37
|
1343 |
conf->pool_size = devs; |
16a53ecc3
|
1344 |
while (num--) |
3f294f4fb
|
1345 |
if (!grow_one_stripe(conf)) |
1da177e4c
|
1346 |
return 1; |
1da177e4c
|
1347 1348 |
return 0; } |
292695531
|
1349 |
|
d6f38f31f
|
1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 |
/** * scribble_len - return the required size of the scribble region * @num - total number of disks in the array * * The size must be enough to contain: * 1/ a struct page pointer for each device in the array +2 * 2/ room to convert each entry in (1) to its corresponding dma * (dma_map_page()) or page (page_address()) address. * * Note: the +2 is for the destination buffers of the ddf/raid6 case where we * calculate over all devices (not just the data blocks), using zeros in place * of the P and Q blocks. */ static size_t scribble_len(int num) { size_t len; len = sizeof(struct page *) * (num+2) + sizeof(addr_conv_t) * (num+2); return len; } |
d1688a6d5
|
1371 |
static int resize_stripes(struct r5conf *conf, int newsize) |
ad01c9e37
|
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 |
{ /* Make all the stripes able to hold 'newsize' devices. * New slots in each stripe get 'page' set to a new page. * * This happens in stages: * 1/ create a new kmem_cache and allocate the required number of * stripe_heads. * 2/ gather all the old stripe_heads and tranfer the pages across * to the new stripe_heads. This will have the side effect of * freezing the array as once all stripe_heads have been collected, * no IO will be possible. Old stripe heads are freed once their * pages have been transferred over, and the old kmem_cache is * freed when all stripes are done. * 3/ reallocate conf->disks to be suitable bigger. If this fails, * we simple return a failre status - no need to clean anything up. * 4/ allocate new pages for the new slots in the new stripe_heads. * If this fails, we don't bother trying the shrink the * stripe_heads down again, we just leave them as they are. * As each stripe_head is processed the new one is released into * active service. * * Once step2 is started, we cannot afford to wait for a write, * so we use GFP_NOIO allocations. */ struct stripe_head *osh, *nsh; LIST_HEAD(newstripes); struct disk_info *ndisks; |
d6f38f31f
|
1399 |
unsigned long cpu; |
b5470dc5f
|
1400 |
int err; |
e18b890bb
|
1401 |
struct kmem_cache *sc; |
ad01c9e37
|
1402 1403 1404 1405 |
int i; if (newsize <= conf->pool_size) return 0; /* never bother to shrink */ |
b5470dc5f
|
1406 1407 1408 |
err = md_allow_write(conf->mddev); if (err) return err; |
2a2275d63
|
1409 |
|
ad01c9e37
|
1410 1411 1412 |
/* Step 1 */ sc = kmem_cache_create(conf->cache_name[1-conf->active_name], sizeof(struct stripe_head)+(newsize-1)*sizeof(struct r5dev), |
20c2df83d
|
1413 |
0, 0, NULL); |
ad01c9e37
|
1414 1415 1416 1417 |
if (!sc) return -ENOMEM; for (i = conf->max_nr_stripes; i; i--) { |
6ce328462
|
1418 |
nsh = kmem_cache_zalloc(sc, GFP_KERNEL); |
ad01c9e37
|
1419 1420 |
if (!nsh) break; |
ad01c9e37
|
1421 |
nsh->raid_conf = conf; |
417b8d4ac
|
1422 1423 1424 |
#ifdef CONFIG_MULTICORE_RAID456 init_waitqueue_head(&nsh->ops.wait_for_ops); #endif |
ad01c9e37
|
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 |
list_add(&nsh->lru, &newstripes); } if (i) { /* didn't get enough, give up */ while (!list_empty(&newstripes)) { nsh = list_entry(newstripes.next, struct stripe_head, lru); list_del(&nsh->lru); kmem_cache_free(sc, nsh); } kmem_cache_destroy(sc); return -ENOMEM; } /* Step 2 - Must use GFP_NOIO now. * OK, we have enough stripes, start collecting inactive * stripes and copying them over */ list_for_each_entry(nsh, &newstripes, lru) { spin_lock_irq(&conf->device_lock); wait_event_lock_irq(conf->wait_for_stripe, !list_empty(&conf->inactive_list), conf->device_lock, |
482c08349
|
1447 |
); |
ad01c9e37
|
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 |
osh = get_free_stripe(conf); spin_unlock_irq(&conf->device_lock); atomic_set(&nsh->count, 1); for(i=0; i<conf->pool_size; i++) nsh->dev[i].page = osh->dev[i].page; for( ; i<newsize; i++) nsh->dev[i].page = NULL; kmem_cache_free(conf->slab_cache, osh); } kmem_cache_destroy(conf->slab_cache); /* Step 3. * At this point, we are holding all the stripes so the array * is completely stalled, so now is a good time to resize |
d6f38f31f
|
1462 |
* conf->disks and the scribble region |
ad01c9e37
|
1463 1464 1465 1466 1467 1468 1469 1470 1471 |
*/ ndisks = kzalloc(newsize * sizeof(struct disk_info), GFP_NOIO); if (ndisks) { for (i=0; i<conf->raid_disks; i++) ndisks[i] = conf->disks[i]; kfree(conf->disks); conf->disks = ndisks; } else err = -ENOMEM; |
d6f38f31f
|
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 |
get_online_cpus(); conf->scribble_len = scribble_len(newsize); for_each_present_cpu(cpu) { struct raid5_percpu *percpu; void *scribble; percpu = per_cpu_ptr(conf->percpu, cpu); scribble = kmalloc(conf->scribble_len, GFP_NOIO); if (scribble) { kfree(percpu->scribble); percpu->scribble = scribble; } else { err = -ENOMEM; break; } } put_online_cpus(); |
ad01c9e37
|
1490 1491 1492 1493 |
/* Step 4, return new stripes to service */ while(!list_empty(&newstripes)) { nsh = list_entry(newstripes.next, struct stripe_head, lru); list_del_init(&nsh->lru); |
d6f38f31f
|
1494 |
|
ad01c9e37
|
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 |
for (i=conf->raid_disks; i < newsize; i++) if (nsh->dev[i].page == NULL) { struct page *p = alloc_page(GFP_NOIO); nsh->dev[i].page = p; if (!p) err = -ENOMEM; } release_stripe(nsh); } /* critical section pass, GFP_NOIO no longer needed */ conf->slab_cache = sc; conf->active_name = 1-conf->active_name; conf->pool_size = newsize; return err; } |
1da177e4c
|
1511 |
|
d1688a6d5
|
1512 |
static int drop_one_stripe(struct r5conf *conf) |
1da177e4c
|
1513 1514 |
{ struct stripe_head *sh; |
3f294f4fb
|
1515 1516 1517 1518 1519 |
spin_lock_irq(&conf->device_lock); sh = get_free_stripe(conf); spin_unlock_irq(&conf->device_lock); if (!sh) return 0; |
78bafebd4
|
1520 |
BUG_ON(atomic_read(&sh->count)); |
e4e11e385
|
1521 |
shrink_buffers(sh); |
3f294f4fb
|
1522 1523 1524 1525 |
kmem_cache_free(conf->slab_cache, sh); atomic_dec(&conf->active_stripes); return 1; } |
d1688a6d5
|
1526 |
static void shrink_stripes(struct r5conf *conf) |
3f294f4fb
|
1527 1528 1529 |
{ while (drop_one_stripe(conf)) ; |
29fc7e3e7
|
1530 1531 |
if (conf->slab_cache) kmem_cache_destroy(conf->slab_cache); |
1da177e4c
|
1532 1533 |
conf->slab_cache = NULL; } |
6712ecf8f
|
1534 |
static void raid5_end_read_request(struct bio * bi, int error) |
1da177e4c
|
1535 |
{ |
99c0fb5f9
|
1536 |
struct stripe_head *sh = bi->bi_private; |
d1688a6d5
|
1537 |
struct r5conf *conf = sh->raid_conf; |
7ecaa1e6a
|
1538 |
int disks = sh->disks, i; |
1da177e4c
|
1539 |
int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags); |
d69504325
|
1540 |
char b[BDEVNAME_SIZE]; |
3cb030020
|
1541 |
struct md_rdev *rdev; |
1da177e4c
|
1542 |
|
1da177e4c
|
1543 1544 1545 1546 |
for (i=0 ; i<disks; i++) if (bi == &sh->dev[i].req) break; |
45b4233ca
|
1547 1548 1549 |
pr_debug("end_read_request %llu/%d, count: %d, uptodate %d. ", (unsigned long long)sh->sector, i, atomic_read(&sh->count), |
1da177e4c
|
1550 1551 1552 |
uptodate); if (i == disks) { BUG(); |
6712ecf8f
|
1553 |
return; |
1da177e4c
|
1554 1555 1556 |
} if (uptodate) { |
1da177e4c
|
1557 |
set_bit(R5_UPTODATE, &sh->dev[i].flags); |
4e5314b56
|
1558 |
if (test_bit(R5_ReadError, &sh->dev[i].flags)) { |
d69504325
|
1559 |
rdev = conf->disks[i].rdev; |
8bda470e8
|
1560 1561 1562 1563 1564 1565 1566 1567 1568 |
printk_ratelimited( KERN_INFO "md/raid:%s: read error corrected" " (%lu sectors at %llu on %s) ", mdname(conf->mddev), STRIPE_SECTORS, (unsigned long long)(sh->sector + rdev->data_offset), bdevname(rdev->bdev, b)); |
ddd5115fe
|
1569 |
atomic_add(STRIPE_SECTORS, &rdev->corrected_errors); |
4e5314b56
|
1570 1571 1572 |
clear_bit(R5_ReadError, &sh->dev[i].flags); clear_bit(R5_ReWrite, &sh->dev[i].flags); } |
ba22dcbf1
|
1573 1574 |
if (atomic_read(&conf->disks[i].rdev->read_errors)) atomic_set(&conf->disks[i].rdev->read_errors, 0); |
1da177e4c
|
1575 |
} else { |
d69504325
|
1576 |
const char *bdn = bdevname(conf->disks[i].rdev->bdev, b); |
ba22dcbf1
|
1577 |
int retry = 0; |
d69504325
|
1578 |
rdev = conf->disks[i].rdev; |
1da177e4c
|
1579 |
clear_bit(R5_UPTODATE, &sh->dev[i].flags); |
d69504325
|
1580 |
atomic_inc(&rdev->read_errors); |
7b0bb5368
|
1581 |
if (conf->mddev->degraded >= conf->max_degraded) |
8bda470e8
|
1582 1583 1584 1585 1586 1587 1588 1589 1590 |
printk_ratelimited( KERN_WARNING "md/raid:%s: read error not correctable " "(sector %llu on %s). ", mdname(conf->mddev), (unsigned long long)(sh->sector + rdev->data_offset), bdn); |
ba22dcbf1
|
1591 |
else if (test_bit(R5_ReWrite, &sh->dev[i].flags)) |
4e5314b56
|
1592 |
/* Oh, no!!! */ |
8bda470e8
|
1593 1594 1595 1596 1597 1598 1599 1600 1601 |
printk_ratelimited( KERN_WARNING "md/raid:%s: read error NOT corrected!! " "(sector %llu on %s). ", mdname(conf->mddev), (unsigned long long)(sh->sector + rdev->data_offset), bdn); |
d69504325
|
1602 |
else if (atomic_read(&rdev->read_errors) |
ba22dcbf1
|
1603 |
> conf->max_nr_stripes) |
14f8d26b8
|
1604 |
printk(KERN_WARNING |
0c55e0225
|
1605 1606 |
"md/raid:%s: Too many read errors, failing device %s. ", |
d69504325
|
1607 |
mdname(conf->mddev), bdn); |
ba22dcbf1
|
1608 1609 1610 1611 1612 |
else retry = 1; if (retry) set_bit(R5_ReadError, &sh->dev[i].flags); else { |
4e5314b56
|
1613 1614 |
clear_bit(R5_ReadError, &sh->dev[i].flags); clear_bit(R5_ReWrite, &sh->dev[i].flags); |
d69504325
|
1615 |
md_error(conf->mddev, rdev); |
ba22dcbf1
|
1616 |
} |
1da177e4c
|
1617 1618 |
} rdev_dec_pending(conf->disks[i].rdev, conf->mddev); |
1da177e4c
|
1619 1620 1621 |
clear_bit(R5_LOCKED, &sh->dev[i].flags); set_bit(STRIPE_HANDLE, &sh->state); release_stripe(sh); |
1da177e4c
|
1622 |
} |
d710e1381
|
1623 |
static void raid5_end_write_request(struct bio *bi, int error) |
1da177e4c
|
1624 |
{ |
99c0fb5f9
|
1625 |
struct stripe_head *sh = bi->bi_private; |
d1688a6d5
|
1626 |
struct r5conf *conf = sh->raid_conf; |
7ecaa1e6a
|
1627 |
int disks = sh->disks, i; |
1da177e4c
|
1628 |
int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags); |
b84db560e
|
1629 1630 |
sector_t first_bad; int bad_sectors; |
1da177e4c
|
1631 |
|
1da177e4c
|
1632 1633 1634 |
for (i=0 ; i<disks; i++) if (bi == &sh->dev[i].req) break; |
45b4233ca
|
1635 1636 |
pr_debug("end_write_request %llu/%d, count %d, uptodate: %d. ", |
1da177e4c
|
1637 1638 1639 1640 |
(unsigned long long)sh->sector, i, atomic_read(&sh->count), uptodate); if (i == disks) { BUG(); |
6712ecf8f
|
1641 |
return; |
1da177e4c
|
1642 |
} |
bc2607f39
|
1643 1644 1645 |
if (!uptodate) { set_bit(WriteErrorSeen, &conf->disks[i].rdev->flags); set_bit(R5_WriteError, &sh->dev[i].flags); |
b84db560e
|
1646 1647 1648 |
} else if (is_badblock(conf->disks[i].rdev, sh->sector, STRIPE_SECTORS, &first_bad, &bad_sectors)) set_bit(R5_MadeGood, &sh->dev[i].flags); |
1da177e4c
|
1649 1650 1651 1652 1653 |
rdev_dec_pending(conf->disks[i].rdev, conf->mddev); clear_bit(R5_LOCKED, &sh->dev[i].flags); set_bit(STRIPE_HANDLE, &sh->state); |
c04be0aa8
|
1654 |
release_stripe(sh); |
1da177e4c
|
1655 |
} |
784052ecc
|
1656 |
static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous); |
1da177e4c
|
1657 |
|
784052ecc
|
1658 |
static void raid5_build_block(struct stripe_head *sh, int i, int previous) |
1da177e4c
|
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 |
{ struct r5dev *dev = &sh->dev[i]; bio_init(&dev->req); dev->req.bi_io_vec = &dev->vec; dev->req.bi_vcnt++; dev->req.bi_max_vecs++; dev->vec.bv_page = dev->page; dev->vec.bv_len = STRIPE_SIZE; dev->vec.bv_offset = 0; dev->req.bi_sector = sh->sector; dev->req.bi_private = sh; dev->flags = 0; |
784052ecc
|
1674 |
dev->sector = compute_blocknr(sh, i, previous); |
1da177e4c
|
1675 |
} |
fd01b88c7
|
1676 |
static void error(struct mddev *mddev, struct md_rdev *rdev) |
1da177e4c
|
1677 1678 |
{ char b[BDEVNAME_SIZE]; |
d1688a6d5
|
1679 |
struct r5conf *conf = mddev->private; |
0c55e0225
|
1680 1681 |
pr_debug("raid456: error called "); |
1da177e4c
|
1682 |
|
6f8d0c77c
|
1683 1684 1685 1686 1687 1688 1689 1690 1691 |
if (test_and_clear_bit(In_sync, &rdev->flags)) { unsigned long flags; spin_lock_irqsave(&conf->device_lock, flags); mddev->degraded++; spin_unlock_irqrestore(&conf->device_lock, flags); /* * if recovery was running, make sure it aborts. */ set_bit(MD_RECOVERY_INTR, &mddev->recovery); |
1da177e4c
|
1692 |
} |
de393cdea
|
1693 |
set_bit(Blocked, &rdev->flags); |
6f8d0c77c
|
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 |
set_bit(Faulty, &rdev->flags); set_bit(MD_CHANGE_DEVS, &mddev->flags); printk(KERN_ALERT "md/raid:%s: Disk failure on %s, disabling device. " "md/raid:%s: Operation continuing on %d devices. ", mdname(mddev), bdevname(rdev->bdev, b), mdname(mddev), conf->raid_disks - mddev->degraded); |
16a53ecc3
|
1705 |
} |
1da177e4c
|
1706 1707 1708 1709 1710 |
/* * Input: a 'big' sector number, * Output: index of the data and parity disk, and the sector # in them. */ |
d1688a6d5
|
1711 |
static sector_t raid5_compute_sector(struct r5conf *conf, sector_t r_sector, |
911d4ee85
|
1712 1713 |
int previous, int *dd_idx, struct stripe_head *sh) |
1da177e4c
|
1714 |
{ |
6e3b96ed6
|
1715 |
sector_t stripe, stripe2; |
35f2a5911
|
1716 |
sector_t chunk_number; |
1da177e4c
|
1717 |
unsigned int chunk_offset; |
911d4ee85
|
1718 |
int pd_idx, qd_idx; |
67cc2b816
|
1719 |
int ddf_layout = 0; |
1da177e4c
|
1720 |
sector_t new_sector; |
e183eaedd
|
1721 1722 |
int algorithm = previous ? conf->prev_algo : conf->algorithm; |
09c9e5fa1
|
1723 1724 |
int sectors_per_chunk = previous ? conf->prev_chunk_sectors : conf->chunk_sectors; |
112bf8970
|
1725 1726 1727 |
int raid_disks = previous ? conf->previous_raid_disks : conf->raid_disks; int data_disks = raid_disks - conf->max_degraded; |
1da177e4c
|
1728 1729 1730 1731 1732 1733 1734 1735 |
/* First compute the information on this sector */ /* * Compute the chunk number and the sector offset inside the chunk */ chunk_offset = sector_div(r_sector, sectors_per_chunk); chunk_number = r_sector; |
1da177e4c
|
1736 1737 1738 1739 |
/* * Compute the stripe number */ |
35f2a5911
|
1740 1741 |
stripe = chunk_number; *dd_idx = sector_div(stripe, data_disks); |
6e3b96ed6
|
1742 |
stripe2 = stripe; |
1da177e4c
|
1743 1744 1745 |
/* * Select the parity disk based on the user selected algorithm. */ |
84789554e
|
1746 |
pd_idx = qd_idx = -1; |
16a53ecc3
|
1747 1748 |
switch(conf->level) { case 4: |
911d4ee85
|
1749 |
pd_idx = data_disks; |
16a53ecc3
|
1750 1751 |
break; case 5: |
e183eaedd
|
1752 |
switch (algorithm) { |
1da177e4c
|
1753 |
case ALGORITHM_LEFT_ASYMMETRIC: |
6e3b96ed6
|
1754 |
pd_idx = data_disks - sector_div(stripe2, raid_disks); |
911d4ee85
|
1755 |
if (*dd_idx >= pd_idx) |
1da177e4c
|
1756 1757 1758 |
(*dd_idx)++; break; case ALGORITHM_RIGHT_ASYMMETRIC: |
6e3b96ed6
|
1759 |
pd_idx = sector_div(stripe2, raid_disks); |
911d4ee85
|
1760 |
if (*dd_idx >= pd_idx) |
1da177e4c
|
1761 1762 1763 |
(*dd_idx)++; break; case ALGORITHM_LEFT_SYMMETRIC: |
6e3b96ed6
|
1764 |
pd_idx = data_disks - sector_div(stripe2, raid_disks); |
911d4ee85
|
1765 |
*dd_idx = (pd_idx + 1 + *dd_idx) % raid_disks; |
1da177e4c
|
1766 1767 |
break; case ALGORITHM_RIGHT_SYMMETRIC: |
6e3b96ed6
|
1768 |
pd_idx = sector_div(stripe2, raid_disks); |
911d4ee85
|
1769 |
*dd_idx = (pd_idx + 1 + *dd_idx) % raid_disks; |
1da177e4c
|
1770 |
break; |
99c0fb5f9
|
1771 1772 1773 1774 1775 1776 1777 |
case ALGORITHM_PARITY_0: pd_idx = 0; (*dd_idx)++; break; case ALGORITHM_PARITY_N: pd_idx = data_disks; break; |
1da177e4c
|
1778 |
default: |
99c0fb5f9
|
1779 |
BUG(); |
16a53ecc3
|
1780 1781 1782 |
} break; case 6: |
e183eaedd
|
1783 |
switch (algorithm) { |
16a53ecc3
|
1784 |
case ALGORITHM_LEFT_ASYMMETRIC: |
6e3b96ed6
|
1785 |
pd_idx = raid_disks - 1 - sector_div(stripe2, raid_disks); |
911d4ee85
|
1786 1787 |
qd_idx = pd_idx + 1; if (pd_idx == raid_disks-1) { |
99c0fb5f9
|
1788 |
(*dd_idx)++; /* Q D D D P */ |
911d4ee85
|
1789 1790 |
qd_idx = 0; } else if (*dd_idx >= pd_idx) |
16a53ecc3
|
1791 1792 1793 |
(*dd_idx) += 2; /* D D P Q D */ break; case ALGORITHM_RIGHT_ASYMMETRIC: |
6e3b96ed6
|
1794 |
pd_idx = sector_div(stripe2, raid_disks); |
911d4ee85
|
1795 1796 |
qd_idx = pd_idx + 1; if (pd_idx == raid_disks-1) { |
99c0fb5f9
|
1797 |
(*dd_idx)++; /* Q D D D P */ |
911d4ee85
|
1798 1799 |
qd_idx = 0; } else if (*dd_idx >= pd_idx) |
16a53ecc3
|
1800 1801 1802 |
(*dd_idx) += 2; /* D D P Q D */ break; case ALGORITHM_LEFT_SYMMETRIC: |
6e3b96ed6
|
1803 |
pd_idx = raid_disks - 1 - sector_div(stripe2, raid_disks); |
911d4ee85
|
1804 1805 |
qd_idx = (pd_idx + 1) % raid_disks; *dd_idx = (pd_idx + 2 + *dd_idx) % raid_disks; |
16a53ecc3
|
1806 1807 |
break; case ALGORITHM_RIGHT_SYMMETRIC: |
6e3b96ed6
|
1808 |
pd_idx = sector_div(stripe2, raid_disks); |
911d4ee85
|
1809 1810 |
qd_idx = (pd_idx + 1) % raid_disks; *dd_idx = (pd_idx + 2 + *dd_idx) % raid_disks; |
16a53ecc3
|
1811 |
break; |
99c0fb5f9
|
1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 |
case ALGORITHM_PARITY_0: pd_idx = 0; qd_idx = 1; (*dd_idx) += 2; break; case ALGORITHM_PARITY_N: pd_idx = data_disks; qd_idx = data_disks + 1; break; case ALGORITHM_ROTATING_ZERO_RESTART: /* Exactly the same as RIGHT_ASYMMETRIC, but or * of blocks for computing Q is different. */ |
6e3b96ed6
|
1827 |
pd_idx = sector_div(stripe2, raid_disks); |
99c0fb5f9
|
1828 1829 1830 1831 1832 1833 |
qd_idx = pd_idx + 1; if (pd_idx == raid_disks-1) { (*dd_idx)++; /* Q D D D P */ qd_idx = 0; } else if (*dd_idx >= pd_idx) (*dd_idx) += 2; /* D D P Q D */ |
67cc2b816
|
1834 |
ddf_layout = 1; |
99c0fb5f9
|
1835 1836 1837 1838 1839 1840 1841 |
break; case ALGORITHM_ROTATING_N_RESTART: /* Same a left_asymmetric, by first stripe is * D D D P Q rather than * Q D D D P */ |
6e3b96ed6
|
1842 1843 |
stripe2 += 1; pd_idx = raid_disks - 1 - sector_div(stripe2, raid_disks); |
99c0fb5f9
|
1844 1845 1846 1847 1848 1849 |
qd_idx = pd_idx + 1; if (pd_idx == raid_disks-1) { (*dd_idx)++; /* Q D D D P */ qd_idx = 0; } else if (*dd_idx >= pd_idx) (*dd_idx) += 2; /* D D P Q D */ |
67cc2b816
|
1850 |
ddf_layout = 1; |
99c0fb5f9
|
1851 1852 1853 1854 |
break; case ALGORITHM_ROTATING_N_CONTINUE: /* Same as left_symmetric but Q is before P */ |
6e3b96ed6
|
1855 |
pd_idx = raid_disks - 1 - sector_div(stripe2, raid_disks); |
99c0fb5f9
|
1856 1857 |
qd_idx = (pd_idx + raid_disks - 1) % raid_disks; *dd_idx = (pd_idx + 1 + *dd_idx) % raid_disks; |
67cc2b816
|
1858 |
ddf_layout = 1; |
99c0fb5f9
|
1859 1860 1861 1862 |
break; case ALGORITHM_LEFT_ASYMMETRIC_6: /* RAID5 left_asymmetric, with Q on last device */ |
6e3b96ed6
|
1863 |
pd_idx = data_disks - sector_div(stripe2, raid_disks-1); |
99c0fb5f9
|
1864 1865 1866 1867 1868 1869 |
if (*dd_idx >= pd_idx) (*dd_idx)++; qd_idx = raid_disks - 1; break; case ALGORITHM_RIGHT_ASYMMETRIC_6: |
6e3b96ed6
|
1870 |
pd_idx = sector_div(stripe2, raid_disks-1); |
99c0fb5f9
|
1871 1872 1873 1874 1875 1876 |
if (*dd_idx >= pd_idx) (*dd_idx)++; qd_idx = raid_disks - 1; break; case ALGORITHM_LEFT_SYMMETRIC_6: |
6e3b96ed6
|
1877 |
pd_idx = data_disks - sector_div(stripe2, raid_disks-1); |
99c0fb5f9
|
1878 1879 1880 1881 1882 |
*dd_idx = (pd_idx + 1 + *dd_idx) % (raid_disks-1); qd_idx = raid_disks - 1; break; case ALGORITHM_RIGHT_SYMMETRIC_6: |
6e3b96ed6
|
1883 |
pd_idx = sector_div(stripe2, raid_disks-1); |
99c0fb5f9
|
1884 1885 1886 1887 1888 1889 1890 1891 1892 |
*dd_idx = (pd_idx + 1 + *dd_idx) % (raid_disks-1); qd_idx = raid_disks - 1; break; case ALGORITHM_PARITY_0_6: pd_idx = 0; (*dd_idx)++; qd_idx = raid_disks - 1; break; |
16a53ecc3
|
1893 |
default: |
99c0fb5f9
|
1894 |
BUG(); |
16a53ecc3
|
1895 1896 |
} break; |
1da177e4c
|
1897 |
} |
911d4ee85
|
1898 1899 1900 |
if (sh) { sh->pd_idx = pd_idx; sh->qd_idx = qd_idx; |
67cc2b816
|
1901 |
sh->ddf_layout = ddf_layout; |
911d4ee85
|
1902 |
} |
1da177e4c
|
1903 1904 1905 1906 1907 1908 |
/* * Finally, compute the new sector number */ new_sector = (sector_t)stripe * sectors_per_chunk + chunk_offset; return new_sector; } |
784052ecc
|
1909 |
static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous) |
1da177e4c
|
1910 |
{ |
d1688a6d5
|
1911 |
struct r5conf *conf = sh->raid_conf; |
b875e531f
|
1912 1913 |
int raid_disks = sh->disks; int data_disks = raid_disks - conf->max_degraded; |
1da177e4c
|
1914 |
sector_t new_sector = sh->sector, check; |
09c9e5fa1
|
1915 1916 |
int sectors_per_chunk = previous ? conf->prev_chunk_sectors : conf->chunk_sectors; |
e183eaedd
|
1917 1918 |
int algorithm = previous ? conf->prev_algo : conf->algorithm; |
1da177e4c
|
1919 1920 |
sector_t stripe; int chunk_offset; |
35f2a5911
|
1921 1922 |
sector_t chunk_number; int dummy1, dd_idx = i; |
1da177e4c
|
1923 |
sector_t r_sector; |
911d4ee85
|
1924 |
struct stripe_head sh2; |
1da177e4c
|
1925 |
|
16a53ecc3
|
1926 |
|
1da177e4c
|
1927 1928 |
chunk_offset = sector_div(new_sector, sectors_per_chunk); stripe = new_sector; |
1da177e4c
|
1929 |
|
16a53ecc3
|
1930 1931 1932 1933 1934 |
if (i == sh->pd_idx) return 0; switch(conf->level) { case 4: break; case 5: |
e183eaedd
|
1935 |
switch (algorithm) { |
1da177e4c
|
1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 |
case ALGORITHM_LEFT_ASYMMETRIC: case ALGORITHM_RIGHT_ASYMMETRIC: if (i > sh->pd_idx) i--; break; case ALGORITHM_LEFT_SYMMETRIC: case ALGORITHM_RIGHT_SYMMETRIC: if (i < sh->pd_idx) i += raid_disks; i -= (sh->pd_idx + 1); break; |
99c0fb5f9
|
1947 1948 1949 1950 1951 |
case ALGORITHM_PARITY_0: i -= 1; break; case ALGORITHM_PARITY_N: break; |
1da177e4c
|
1952 |
default: |
99c0fb5f9
|
1953 |
BUG(); |
16a53ecc3
|
1954 1955 1956 |
} break; case 6: |
d0dabf7e5
|
1957 |
if (i == sh->qd_idx) |
16a53ecc3
|
1958 |
return 0; /* It is the Q disk */ |
e183eaedd
|
1959 |
switch (algorithm) { |
16a53ecc3
|
1960 1961 |
case ALGORITHM_LEFT_ASYMMETRIC: case ALGORITHM_RIGHT_ASYMMETRIC: |
99c0fb5f9
|
1962 1963 1964 1965 |
case ALGORITHM_ROTATING_ZERO_RESTART: case ALGORITHM_ROTATING_N_RESTART: if (sh->pd_idx == raid_disks-1) i--; /* Q D D D P */ |
16a53ecc3
|
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 |
else if (i > sh->pd_idx) i -= 2; /* D D P Q D */ break; case ALGORITHM_LEFT_SYMMETRIC: case ALGORITHM_RIGHT_SYMMETRIC: if (sh->pd_idx == raid_disks-1) i--; /* Q D D D P */ else { /* D D P Q D */ if (i < sh->pd_idx) i += raid_disks; i -= (sh->pd_idx + 2); } break; |
99c0fb5f9
|
1980 1981 1982 1983 1984 1985 |
case ALGORITHM_PARITY_0: i -= 2; break; case ALGORITHM_PARITY_N: break; case ALGORITHM_ROTATING_N_CONTINUE: |
e4424fee1
|
1986 |
/* Like left_symmetric, but P is before Q */ |
99c0fb5f9
|
1987 1988 |
if (sh->pd_idx == 0) i--; /* P D D D Q */ |
e4424fee1
|
1989 1990 1991 1992 1993 1994 |
else { /* D D Q P D */ if (i < sh->pd_idx) i += raid_disks; i -= (sh->pd_idx + 1); } |
99c0fb5f9
|
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 |
break; case ALGORITHM_LEFT_ASYMMETRIC_6: case ALGORITHM_RIGHT_ASYMMETRIC_6: if (i > sh->pd_idx) i--; break; case ALGORITHM_LEFT_SYMMETRIC_6: case ALGORITHM_RIGHT_SYMMETRIC_6: if (i < sh->pd_idx) i += data_disks + 1; i -= (sh->pd_idx + 1); break; case ALGORITHM_PARITY_0_6: i -= 1; break; |
16a53ecc3
|
2010 |
default: |
99c0fb5f9
|
2011 |
BUG(); |
16a53ecc3
|
2012 2013 |
} break; |
1da177e4c
|
2014 2015 2016 |
} chunk_number = stripe * data_disks + i; |
35f2a5911
|
2017 |
r_sector = chunk_number * sectors_per_chunk + chunk_offset; |
1da177e4c
|
2018 |
|
112bf8970
|
2019 |
check = raid5_compute_sector(conf, r_sector, |
784052ecc
|
2020 |
previous, &dummy1, &sh2); |
911d4ee85
|
2021 2022 |
if (check != sh->sector || dummy1 != dd_idx || sh2.pd_idx != sh->pd_idx || sh2.qd_idx != sh->qd_idx) { |
0c55e0225
|
2023 2024 2025 |
printk(KERN_ERR "md/raid:%s: compute_blocknr: map not correct ", mdname(conf->mddev)); |
1da177e4c
|
2026 2027 2028 2029 |
return 0; } return r_sector; } |
600aa1099
|
2030 |
static void |
c0f7bddbe
|
2031 |
schedule_reconstruction(struct stripe_head *sh, struct stripe_head_state *s, |
600aa1099
|
2032 |
int rcw, int expand) |
e33129d84
|
2033 2034 |
{ int i, pd_idx = sh->pd_idx, disks = sh->disks; |
d1688a6d5
|
2035 |
struct r5conf *conf = sh->raid_conf; |
c0f7bddbe
|
2036 |
int level = conf->level; |
e33129d84
|
2037 2038 2039 2040 2041 2042 2043 |
if (rcw) { /* if we are not expanding this is a proper write request, and * there will be bios with new data to be drained into the * stripe cache */ if (!expand) { |
600aa1099
|
2044 2045 2046 2047 |
sh->reconstruct_state = reconstruct_state_drain_run; set_bit(STRIPE_OP_BIODRAIN, &s->ops_request); } else sh->reconstruct_state = reconstruct_state_run; |
16a53ecc3
|
2048 |
|
ac6b53b6e
|
2049 |
set_bit(STRIPE_OP_RECONSTRUCT, &s->ops_request); |
e33129d84
|
2050 2051 2052 2053 2054 2055 |
for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (dev->towrite) { set_bit(R5_LOCKED, &dev->flags); |
d8ee0728b
|
2056 |
set_bit(R5_Wantdrain, &dev->flags); |
e33129d84
|
2057 2058 |
if (!expand) clear_bit(R5_UPTODATE, &dev->flags); |
600aa1099
|
2059 |
s->locked++; |
e33129d84
|
2060 2061 |
} } |
c0f7bddbe
|
2062 |
if (s->locked + conf->max_degraded == disks) |
8b3e6cdc5
|
2063 |
if (!test_and_set_bit(STRIPE_FULL_WRITE, &sh->state)) |
c0f7bddbe
|
2064 |
atomic_inc(&conf->pending_full_writes); |
e33129d84
|
2065 |
} else { |
c0f7bddbe
|
2066 |
BUG_ON(level == 6); |
e33129d84
|
2067 2068 |
BUG_ON(!(test_bit(R5_UPTODATE, &sh->dev[pd_idx].flags) || test_bit(R5_Wantcompute, &sh->dev[pd_idx].flags))); |
d8ee0728b
|
2069 |
sh->reconstruct_state = reconstruct_state_prexor_drain_run; |
600aa1099
|
2070 2071 |
set_bit(STRIPE_OP_PREXOR, &s->ops_request); set_bit(STRIPE_OP_BIODRAIN, &s->ops_request); |
ac6b53b6e
|
2072 |
set_bit(STRIPE_OP_RECONSTRUCT, &s->ops_request); |
e33129d84
|
2073 2074 2075 2076 2077 |
for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (i == pd_idx) continue; |
e33129d84
|
2078 2079 |
if (dev->towrite && (test_bit(R5_UPTODATE, &dev->flags) || |
d8ee0728b
|
2080 2081 |
test_bit(R5_Wantcompute, &dev->flags))) { set_bit(R5_Wantdrain, &dev->flags); |
e33129d84
|
2082 2083 |
set_bit(R5_LOCKED, &dev->flags); clear_bit(R5_UPTODATE, &dev->flags); |
600aa1099
|
2084 |
s->locked++; |
e33129d84
|
2085 2086 2087 |
} } } |
c0f7bddbe
|
2088 |
/* keep the parity disk(s) locked while asynchronous operations |
e33129d84
|
2089 2090 2091 2092 |
* are in flight */ set_bit(R5_LOCKED, &sh->dev[pd_idx].flags); clear_bit(R5_UPTODATE, &sh->dev[pd_idx].flags); |
600aa1099
|
2093 |
s->locked++; |
e33129d84
|
2094 |
|
c0f7bddbe
|
2095 2096 2097 2098 2099 2100 2101 2102 |
if (level == 6) { int qd_idx = sh->qd_idx; struct r5dev *dev = &sh->dev[qd_idx]; set_bit(R5_LOCKED, &dev->flags); clear_bit(R5_UPTODATE, &dev->flags); s->locked++; } |
600aa1099
|
2103 2104 |
pr_debug("%s: stripe %llu locked: %d ops_request: %lx ", |
e46b272b6
|
2105 |
__func__, (unsigned long long)sh->sector, |
600aa1099
|
2106 |
s->locked, s->ops_request); |
e33129d84
|
2107 |
} |
16a53ecc3
|
2108 |
|
1da177e4c
|
2109 2110 |
/* * Each stripe/dev can have one or more bion attached. |
16a53ecc3
|
2111 |
* toread/towrite point to the first in a chain. |
1da177e4c
|
2112 2113 2114 2115 2116 |
* The bi_next chain must be in order. */ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, int forwrite) { struct bio **bip; |
d1688a6d5
|
2117 |
struct r5conf *conf = sh->raid_conf; |
72626685d
|
2118 |
int firstwrite=0; |
1da177e4c
|
2119 |
|
cbe47ec55
|
2120 2121 |
pr_debug("adding bi b#%llu to stripe s#%llu ", |
1da177e4c
|
2122 2123 |
(unsigned long long)bi->bi_sector, (unsigned long long)sh->sector); |
1da177e4c
|
2124 |
spin_lock_irq(&conf->device_lock); |
72626685d
|
2125 |
if (forwrite) { |
1da177e4c
|
2126 |
bip = &sh->dev[dd_idx].towrite; |
72626685d
|
2127 2128 2129 |
if (*bip == NULL && sh->dev[dd_idx].written == NULL) firstwrite = 1; } else |
1da177e4c
|
2130 2131 2132 2133 2134 2135 2136 2137 |
bip = &sh->dev[dd_idx].toread; while (*bip && (*bip)->bi_sector < bi->bi_sector) { if ((*bip)->bi_sector + ((*bip)->bi_size >> 9) > bi->bi_sector) goto overlap; bip = & (*bip)->bi_next; } if (*bip && (*bip)->bi_sector < bi->bi_sector + ((bi->bi_size)>>9)) goto overlap; |
78bafebd4
|
2138 |
BUG_ON(*bip && bi->bi_next && (*bip) != bi->bi_next); |
1da177e4c
|
2139 2140 2141 |
if (*bip) bi->bi_next = *bip; *bip = bi; |
960e739d9
|
2142 |
bi->bi_phys_segments++; |
72626685d
|
2143 |
|
1da177e4c
|
2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 |
if (forwrite) { /* check if page is covered */ sector_t sector = sh->dev[dd_idx].sector; for (bi=sh->dev[dd_idx].towrite; sector < sh->dev[dd_idx].sector + STRIPE_SECTORS && bi && bi->bi_sector <= sector; bi = r5_next_bio(bi, sh->dev[dd_idx].sector)) { if (bi->bi_sector + (bi->bi_size>>9) >= sector) sector = bi->bi_sector + (bi->bi_size>>9); } if (sector >= sh->dev[dd_idx].sector + STRIPE_SECTORS) set_bit(R5_OVERWRITE, &sh->dev[dd_idx].flags); } |
cbe47ec55
|
2157 |
spin_unlock_irq(&conf->device_lock); |
cbe47ec55
|
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 |
pr_debug("added bi b#%llu to stripe s#%llu, disk %d. ", (unsigned long long)(*bip)->bi_sector, (unsigned long long)sh->sector, dd_idx); if (conf->mddev->bitmap && firstwrite) { bitmap_startwrite(conf->mddev->bitmap, sh->sector, STRIPE_SECTORS, 0); sh->bm_seq = conf->seq_flush+1; set_bit(STRIPE_BIT_DELAY, &sh->state); } |
1da177e4c
|
2170 2171 2172 2173 2174 |
return 1; overlap: set_bit(R5_Overlap, &sh->dev[dd_idx].flags); spin_unlock_irq(&conf->device_lock); |
1da177e4c
|
2175 2176 |
return 0; } |
d1688a6d5
|
2177 |
static void end_reshape(struct r5conf *conf); |
292695531
|
2178 |
|
d1688a6d5
|
2179 |
static void stripe_set_idx(sector_t stripe, struct r5conf *conf, int previous, |
911d4ee85
|
2180 |
struct stripe_head *sh) |
ccfcc3c10
|
2181 |
{ |
784052ecc
|
2182 |
int sectors_per_chunk = |
09c9e5fa1
|
2183 |
previous ? conf->prev_chunk_sectors : conf->chunk_sectors; |
911d4ee85
|
2184 |
int dd_idx; |
2d2063cea
|
2185 |
int chunk_offset = sector_div(stripe, sectors_per_chunk); |
112bf8970
|
2186 |
int disks = previous ? conf->previous_raid_disks : conf->raid_disks; |
2d2063cea
|
2187 |
|
112bf8970
|
2188 2189 |
raid5_compute_sector(conf, stripe * (disks - conf->max_degraded) |
b875e531f
|
2190 |
*sectors_per_chunk + chunk_offset, |
112bf8970
|
2191 |
previous, |
911d4ee85
|
2192 |
&dd_idx, sh); |
ccfcc3c10
|
2193 |
} |
a44568564
|
2194 |
static void |
d1688a6d5
|
2195 |
handle_failed_stripe(struct r5conf *conf, struct stripe_head *sh, |
a44568564
|
2196 2197 2198 2199 2200 2201 2202 2203 2204 |
struct stripe_head_state *s, int disks, struct bio **return_bi) { int i; for (i = disks; i--; ) { struct bio *bi; int bitmap_end = 0; if (test_bit(R5_ReadError, &sh->dev[i].flags)) { |
3cb030020
|
2205 |
struct md_rdev *rdev; |
a44568564
|
2206 2207 2208 |
rcu_read_lock(); rdev = rcu_dereference(conf->disks[i].rdev); if (rdev && test_bit(In_sync, &rdev->flags)) |
7f0da59bd
|
2209 2210 2211 |
atomic_inc(&rdev->nr_pending); else rdev = NULL; |
a44568564
|
2212 |
rcu_read_unlock(); |
7f0da59bd
|
2213 2214 2215 2216 2217 2218 2219 2220 |
if (rdev) { if (!rdev_set_badblocks( rdev, sh->sector, STRIPE_SECTORS, 0)) md_error(conf->mddev, rdev); rdev_dec_pending(rdev, conf->mddev); } |
a44568564
|
2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 |
} spin_lock_irq(&conf->device_lock); /* fail all writes first */ bi = sh->dev[i].towrite; sh->dev[i].towrite = NULL; if (bi) { s->to_write--; bitmap_end = 1; } if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags)) wake_up(&conf->wait_for_overlap); while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); |
960e739d9
|
2238 |
if (!raid5_dec_bi_phys_segments(bi)) { |
a44568564
|
2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 |
md_write_end(conf->mddev); bi->bi_next = *return_bi; *return_bi = bi; } bi = nextbi; } /* and fail all 'written' */ bi = sh->dev[i].written; sh->dev[i].written = NULL; if (bi) bitmap_end = 1; while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); |
960e739d9
|
2253 |
if (!raid5_dec_bi_phys_segments(bi)) { |
a44568564
|
2254 2255 2256 2257 2258 2259 |
md_write_end(conf->mddev); bi->bi_next = *return_bi; *return_bi = bi; } bi = bi2; } |
b5e98d65d
|
2260 2261 2262 2263 2264 2265 |
/* fail any reads if this device is non-operational and * the data has not reached the cache yet. */ if (!test_bit(R5_Wantfill, &sh->dev[i].flags) && (!test_bit(R5_Insync, &sh->dev[i].flags) || test_bit(R5_ReadError, &sh->dev[i].flags))) { |
a44568564
|
2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 |
bi = sh->dev[i].toread; sh->dev[i].toread = NULL; if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags)) wake_up(&conf->wait_for_overlap); if (bi) s->to_read--; while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS) { struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector); clear_bit(BIO_UPTODATE, &bi->bi_flags); |
960e739d9
|
2276 |
if (!raid5_dec_bi_phys_segments(bi)) { |
a44568564
|
2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 |
bi->bi_next = *return_bi; *return_bi = bi; } bi = nextbi; } } spin_unlock_irq(&conf->device_lock); if (bitmap_end) bitmap_endwrite(conf->mddev->bitmap, sh->sector, STRIPE_SECTORS, 0, 0); |
8cfa7b0f6
|
2287 2288 2289 2290 |
/* If we were in the middle of a write the parity block might * still be locked - so just clear all R5_LOCKED flags */ clear_bit(R5_LOCKED, &sh->dev[i].flags); |
a44568564
|
2291 |
} |
8b3e6cdc5
|
2292 2293 2294 |
if (test_and_clear_bit(STRIPE_FULL_WRITE, &sh->state)) if (atomic_dec_and_test(&conf->pending_full_writes)) md_wakeup_thread(conf->mddev->thread); |
a44568564
|
2295 |
} |
7f0da59bd
|
2296 |
static void |
d1688a6d5
|
2297 |
handle_failed_sync(struct r5conf *conf, struct stripe_head *sh, |
7f0da59bd
|
2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 |
struct stripe_head_state *s) { int abort = 0; int i; md_done_sync(conf->mddev, STRIPE_SECTORS, 0); clear_bit(STRIPE_SYNCING, &sh->state); s->syncing = 0; /* There is nothing more to do for sync/check/repair. * For recover we need to record a bad block on all * non-sync devices, or abort the recovery */ if (!test_bit(MD_RECOVERY_RECOVER, &conf->mddev->recovery)) return; /* During recovery devices cannot be removed, so locking and * refcounting of rdevs is not needed */ for (i = 0; i < conf->raid_disks; i++) { |
3cb030020
|
2316 |
struct md_rdev *rdev = conf->disks[i].rdev; |
7f0da59bd
|
2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 |
if (!rdev || test_bit(Faulty, &rdev->flags) || test_bit(In_sync, &rdev->flags)) continue; if (!rdev_set_badblocks(rdev, sh->sector, STRIPE_SECTORS, 0)) abort = 1; } if (abort) { conf->recovery_disabled = conf->mddev->recovery_disabled; set_bit(MD_RECOVERY_INTR, &conf->mddev->recovery); } } |
93b3dbce6
|
2330 |
/* fetch_block - checks the given member device to see if its data needs |
1fe797e67
|
2331 2332 2333 |
* to be read or computed to satisfy a request. * * Returns 1 when no more member devices need to be checked, otherwise returns |
93b3dbce6
|
2334 |
* 0 to tell the loop in handle_stripe_fill to continue |
f38e12199
|
2335 |
*/ |
93b3dbce6
|
2336 2337 |
static int fetch_block(struct stripe_head *sh, struct stripe_head_state *s, int disk_idx, int disks) |
a44568564
|
2338 |
{ |
5599becca
|
2339 |
struct r5dev *dev = &sh->dev[disk_idx]; |
f2b3b44de
|
2340 2341 |
struct r5dev *fdev[2] = { &sh->dev[s->failed_num[0]], &sh->dev[s->failed_num[1]] }; |
5599becca
|
2342 |
|
93b3dbce6
|
2343 |
/* is the data in this block needed, and can we get it? */ |
5599becca
|
2344 2345 2346 2347 2348 |
if (!test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) && (dev->toread || (dev->towrite && !test_bit(R5_OVERWRITE, &dev->flags)) || s->syncing || s->expanding || |
5d35e09ca
|
2349 2350 |
(s->failed >= 1 && fdev[0]->toread) || (s->failed >= 2 && fdev[1]->toread) || |
93b3dbce6
|
2351 2352 2353 |
(sh->raid_conf->level <= 5 && s->failed && fdev[0]->towrite && !test_bit(R5_OVERWRITE, &fdev[0]->flags)) || (sh->raid_conf->level == 6 && s->failed && s->to_write))) { |
5599becca
|
2354 2355 2356 2357 2358 2359 |
/* we would like to get this block, possibly by computing it, * otherwise read it if the backing disk is insync */ BUG_ON(test_bit(R5_Wantcompute, &dev->flags)); BUG_ON(test_bit(R5_Wantread, &dev->flags)); if ((s->uptodate == disks - 1) && |
f2b3b44de
|
2360 2361 |
(s->failed && (disk_idx == s->failed_num[0] || disk_idx == s->failed_num[1]))) { |
5599becca
|
2362 2363 |
/* have disk failed, and we're requested to fetch it; * do compute it |
a44568564
|
2364 |
*/ |
5599becca
|
2365 2366 2367 2368 2369 2370 2371 2372 2373 |
pr_debug("Computing stripe %llu block %d ", (unsigned long long)sh->sector, disk_idx); set_bit(STRIPE_COMPUTE_RUN, &sh->state); set_bit(STRIPE_OP_COMPUTE_BLK, &s->ops_request); set_bit(R5_Wantcompute, &dev->flags); sh->ops.target = disk_idx; sh->ops.target2 = -1; /* no 2nd target */ s->req_compute = 1; |
93b3dbce6
|
2374 2375 2376 2377 2378 2379 |
/* Careful: from this point on 'uptodate' is in the eye * of raid_run_ops which services 'compute' operations * before writes. R5_Wantcompute flags a block that will * be R5_UPTODATE by the time it is needed for a * subsequent operation. */ |
5599becca
|
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 |
s->uptodate++; return 1; } else if (s->uptodate == disks-2 && s->failed >= 2) { /* Computing 2-failure is *very* expensive; only * do it if failed >= 2 */ int other; for (other = disks; other--; ) { if (other == disk_idx) continue; if (!test_bit(R5_UPTODATE, &sh->dev[other].flags)) break; |
a44568564
|
2393 |
} |
5599becca
|
2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 |
BUG_ON(other < 0); pr_debug("Computing stripe %llu blocks %d,%d ", (unsigned long long)sh->sector, disk_idx, other); set_bit(STRIPE_COMPUTE_RUN, &sh->state); set_bit(STRIPE_OP_COMPUTE_BLK, &s->ops_request); set_bit(R5_Wantcompute, &sh->dev[disk_idx].flags); set_bit(R5_Wantcompute, &sh->dev[other].flags); sh->ops.target = disk_idx; sh->ops.target2 = other; s->uptodate += 2; s->req_compute = 1; return 1; } else if (test_bit(R5_Insync, &dev->flags)) { set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantread, &dev->flags); s->locked++; pr_debug("Reading block %d (sync=%d) ", disk_idx, s->syncing); |
a44568564
|
2415 2416 |
} } |
5599becca
|
2417 2418 2419 2420 2421 |
return 0; } /** |
93b3dbce6
|
2422 |
* handle_stripe_fill - read or compute data to satisfy pending requests. |
5599becca
|
2423 |
*/ |
93b3dbce6
|
2424 2425 2426 |
static void handle_stripe_fill(struct stripe_head *sh, struct stripe_head_state *s, int disks) |
5599becca
|
2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 |
{ int i; /* look for blocks to read/compute, skip this if a compute * is already in flight, or if the stripe contents are in the * midst of changing due to a write */ if (!test_bit(STRIPE_COMPUTE_RUN, &sh->state) && !sh->check_state && !sh->reconstruct_state) for (i = disks; i--; ) |
93b3dbce6
|
2437 |
if (fetch_block(sh, s, i, disks)) |
5599becca
|
2438 |
break; |
a44568564
|
2439 2440 |
set_bit(STRIPE_HANDLE, &sh->state); } |
1fe797e67
|
2441 |
/* handle_stripe_clean_event |
a44568564
|
2442 2443 2444 2445 |
* any written block on an uptodate or failed drive can be returned. * Note that if we 'wrote' to a failed drive, it will be UPTODATE, but * never LOCKED, so we don't need to test 'failed' directly. */ |
d1688a6d5
|
2446 |
static void handle_stripe_clean_event(struct r5conf *conf, |
a44568564
|
2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 |
struct stripe_head *sh, int disks, struct bio **return_bi) { int i; struct r5dev *dev; for (i = disks; i--; ) if (sh->dev[i].written) { dev = &sh->dev[i]; if (!test_bit(R5_LOCKED, &dev->flags) && test_bit(R5_UPTODATE, &dev->flags)) { /* We can return any write requests */ struct bio *wbi, *wbi2; int bitmap_end = 0; |
45b4233ca
|
2460 2461 |
pr_debug("Return write for disc %d ", i); |
a44568564
|
2462 2463 2464 2465 2466 2467 |
spin_lock_irq(&conf->device_lock); wbi = dev->written; dev->written = NULL; while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) { wbi2 = r5_next_bio(wbi, dev->sector); |
960e739d9
|
2468 |
if (!raid5_dec_bi_phys_segments(wbi)) { |
a44568564
|
2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 |
md_write_end(conf->mddev); wbi->bi_next = *return_bi; *return_bi = wbi; } wbi = wbi2; } if (dev->towrite == NULL) bitmap_end = 1; spin_unlock_irq(&conf->device_lock); if (bitmap_end) bitmap_endwrite(conf->mddev->bitmap, sh->sector, STRIPE_SECTORS, !test_bit(STRIPE_DEGRADED, &sh->state), 0); } } |
8b3e6cdc5
|
2486 2487 2488 2489 |
if (test_and_clear_bit(STRIPE_FULL_WRITE, &sh->state)) if (atomic_dec_and_test(&conf->pending_full_writes)) md_wakeup_thread(conf->mddev->thread); |
a44568564
|
2490 |
} |
d1688a6d5
|
2491 |
static void handle_stripe_dirtying(struct r5conf *conf, |
c8ac1803f
|
2492 2493 2494 |
struct stripe_head *sh, struct stripe_head_state *s, int disks) |
a44568564
|
2495 2496 |
{ int rmw = 0, rcw = 0, i; |
c8ac1803f
|
2497 2498 2499 2500 2501 2502 2503 |
if (conf->max_degraded == 2) { /* RAID6 requires 'rcw' in current implementation * Calculate the real rcw later - for now fake it * look like rcw is cheaper */ rcw = 1; rmw = 2; } else for (i = disks; i--; ) { |
a44568564
|
2504 2505 2506 2507 |
/* would I have to read this buffer for read_modify_write */ struct r5dev *dev = &sh->dev[i]; if ((dev->towrite || i == sh->pd_idx) && !test_bit(R5_LOCKED, &dev->flags) && |
f38e12199
|
2508 2509 |
!(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { |
a44568564
|
2510 2511 2512 2513 2514 2515 2516 2517 |
if (test_bit(R5_Insync, &dev->flags)) rmw++; else rmw += 2*disks; /* cannot read it */ } /* Would I have to read this buffer for reconstruct_write */ if (!test_bit(R5_OVERWRITE, &dev->flags) && i != sh->pd_idx && !test_bit(R5_LOCKED, &dev->flags) && |
f38e12199
|
2518 2519 2520 |
!(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags))) { if (test_bit(R5_Insync, &dev->flags)) rcw++; |
a44568564
|
2521 2522 2523 2524 |
else rcw += 2*disks; } } |
45b4233ca
|
2525 2526 |
pr_debug("for sector %llu, rmw=%d rcw=%d ", |
a44568564
|
2527 2528 2529 2530 2531 2532 2533 2534 |
(unsigned long long)sh->sector, rmw, rcw); set_bit(STRIPE_HANDLE, &sh->state); if (rmw < rcw && rmw > 0) /* prefer read-modify-write, but need to get some data */ for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if ((dev->towrite || i == sh->pd_idx) && !test_bit(R5_LOCKED, &dev->flags) && |
f38e12199
|
2535 2536 |
!(test_bit(R5_UPTODATE, &dev->flags) || test_bit(R5_Wantcompute, &dev->flags)) && |
a44568564
|
2537 2538 2539 |
test_bit(R5_Insync, &dev->flags)) { if ( test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { |
45b4233ca
|
2540 |
pr_debug("Read_old block " |
a44568564
|
2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 |
"%d for r-m-w ", i); set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantread, &dev->flags); s->locked++; } else { set_bit(STRIPE_DELAYED, &sh->state); set_bit(STRIPE_HANDLE, &sh->state); } } } |
c8ac1803f
|
2552 |
if (rcw <= rmw && rcw > 0) { |
a44568564
|
2553 |
/* want reconstruct write, but need to get some data */ |
c8ac1803f
|
2554 |
rcw = 0; |
a44568564
|
2555 2556 2557 |
for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (!test_bit(R5_OVERWRITE, &dev->flags) && |
c8ac1803f
|
2558 |
i != sh->pd_idx && i != sh->qd_idx && |
a44568564
|
2559 |
!test_bit(R5_LOCKED, &dev->flags) && |
f38e12199
|
2560 |
!(test_bit(R5_UPTODATE, &dev->flags) || |
c8ac1803f
|
2561 2562 2563 2564 |
test_bit(R5_Wantcompute, &dev->flags))) { rcw++; if (!test_bit(R5_Insync, &dev->flags)) continue; /* it's a failed drive */ |
a44568564
|
2565 2566 |
if ( test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) { |
45b4233ca
|
2567 |
pr_debug("Read_old block " |
a44568564
|
2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 |
"%d for Reconstruct ", i); set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantread, &dev->flags); s->locked++; } else { set_bit(STRIPE_DELAYED, &sh->state); set_bit(STRIPE_HANDLE, &sh->state); } } } |
c8ac1803f
|
2579 |
} |
a44568564
|
2580 2581 2582 |
/* now if nothing is locked, and if we have enough data, * we can start a write request */ |
f38e12199
|
2583 2584 |
/* since handle_stripe can be called at any time we need to handle the * case where a compute block operation has been submitted and then a |
ac6b53b6e
|
2585 2586 |
* subsequent call wants to start a write request. raid_run_ops only * handles the case where compute block and reconstruct are requested |
f38e12199
|
2587 2588 2589 |
* simultaneously. If this is not the case then new writes need to be * held off until the compute completes. */ |
976ea8d47
|
2590 2591 2592 |
if ((s->req_compute || !test_bit(STRIPE_COMPUTE_RUN, &sh->state)) && (s->locked == 0 && (rcw == 0 || rmw == 0) && !test_bit(STRIPE_BIT_DELAY, &sh->state))) |
c0f7bddbe
|
2593 |
schedule_reconstruction(sh, s, rcw == 0, 0); |
a44568564
|
2594 |
} |
d1688a6d5
|
2595 |
static void handle_parity_checks5(struct r5conf *conf, struct stripe_head *sh, |
a44568564
|
2596 2597 |
struct stripe_head_state *s, int disks) { |
ecc65c9b3
|
2598 |
struct r5dev *dev = NULL; |
bd2ab6703
|
2599 |
|
a44568564
|
2600 |
set_bit(STRIPE_HANDLE, &sh->state); |
e89f89629
|
2601 |
|
ecc65c9b3
|
2602 2603 2604 |
switch (sh->check_state) { case check_state_idle: /* start a new check operation if there are no failures */ |
bd2ab6703
|
2605 |
if (s->failed == 0) { |
bd2ab6703
|
2606 |
BUG_ON(s->uptodate != disks); |
ecc65c9b3
|
2607 2608 |
sh->check_state = check_state_run; set_bit(STRIPE_OP_CHECK, &s->ops_request); |
bd2ab6703
|
2609 |
clear_bit(R5_UPTODATE, &sh->dev[sh->pd_idx].flags); |
bd2ab6703
|
2610 |
s->uptodate--; |
ecc65c9b3
|
2611 |
break; |
bd2ab6703
|
2612 |
} |
f2b3b44de
|
2613 |
dev = &sh->dev[s->failed_num[0]]; |
ecc65c9b3
|
2614 2615 2616 2617 2618 2619 2620 2621 2622 |
/* fall through */ case check_state_compute_result: sh->check_state = check_state_idle; if (!dev) dev = &sh->dev[sh->pd_idx]; /* check that a write has not made the stripe insync */ if (test_bit(STRIPE_INSYNC, &sh->state)) break; |
c8894419a
|
2623 |
|
a44568564
|
2624 |
/* either failed parity check, or recovery is happening */ |
a44568564
|
2625 2626 2627 2628 |
BUG_ON(!test_bit(R5_UPTODATE, &dev->flags)); BUG_ON(s->uptodate != disks); set_bit(R5_LOCKED, &dev->flags); |
ecc65c9b3
|
2629 |
s->locked++; |
a44568564
|
2630 |
set_bit(R5_Wantwrite, &dev->flags); |
830ea0167
|
2631 |
|
a44568564
|
2632 |
clear_bit(STRIPE_DEGRADED, &sh->state); |
a44568564
|
2633 |
set_bit(STRIPE_INSYNC, &sh->state); |
ecc65c9b3
|
2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 |
break; case check_state_run: break; /* we will be called again upon completion */ case check_state_check_result: sh->check_state = check_state_idle; /* if a failure occurred during the check operation, leave * STRIPE_INSYNC not set and let the stripe be handled again */ if (s->failed) break; /* handle a successful check operation, if parity is correct * we are done. Otherwise update the mismatch count and repair * parity if !MD_RECOVERY_CHECK */ |
ad283ea4a
|
2650 |
if ((sh->ops.zero_sum_result & SUM_CHECK_P_RESULT) == 0) |
ecc65c9b3
|
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 |
/* parity is correct (on disc, * not in buffer any more) */ set_bit(STRIPE_INSYNC, &sh->state); else { conf->mddev->resync_mismatches += STRIPE_SECTORS; if (test_bit(MD_RECOVERY_CHECK, &conf->mddev->recovery)) /* don't try to repair!! */ set_bit(STRIPE_INSYNC, &sh->state); else { sh->check_state = check_state_compute_run; |
976ea8d47
|
2662 |
set_bit(STRIPE_COMPUTE_RUN, &sh->state); |
ecc65c9b3
|
2663 2664 2665 2666 |
set_bit(STRIPE_OP_COMPUTE_BLK, &s->ops_request); set_bit(R5_Wantcompute, &sh->dev[sh->pd_idx].flags); sh->ops.target = sh->pd_idx; |
ac6b53b6e
|
2667 |
sh->ops.target2 = -1; |
ecc65c9b3
|
2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 |
s->uptodate++; } } break; case check_state_compute_run: break; default: printk(KERN_ERR "%s: unknown check_state: %d sector: %llu ", __func__, sh->check_state, (unsigned long long) sh->sector); BUG(); |
a44568564
|
2680 2681 |
} } |
d1688a6d5
|
2682 |
static void handle_parity_checks6(struct r5conf *conf, struct stripe_head *sh, |
36d1c6476
|
2683 |
struct stripe_head_state *s, |
f2b3b44de
|
2684 |
int disks) |
a44568564
|
2685 |
{ |
a44568564
|
2686 |
int pd_idx = sh->pd_idx; |
34e04e87f
|
2687 |
int qd_idx = sh->qd_idx; |
d82dfee0a
|
2688 |
struct r5dev *dev; |
a44568564
|
2689 2690 2691 2692 |
set_bit(STRIPE_HANDLE, &sh->state); BUG_ON(s->failed > 2); |
d82dfee0a
|
2693 |
|
a44568564
|
2694 2695 2696 2697 2698 |
/* Want to check and possibly repair P and Q. * However there could be one 'failed' device, in which * case we can only check one of them, possibly using the * other to generate missing data */ |
d82dfee0a
|
2699 2700 2701 |
switch (sh->check_state) { case check_state_idle: /* start a new check operation if there are < 2 failures */ |
f2b3b44de
|
2702 |
if (s->failed == s->q_failed) { |
d82dfee0a
|
2703 |
/* The only possible failed device holds Q, so it |
a44568564
|
2704 2705 2706 |
* makes sense to check P (If anything else were failed, * we would have used P to recreate it). */ |
d82dfee0a
|
2707 |
sh->check_state = check_state_run; |
a44568564
|
2708 |
} |
f2b3b44de
|
2709 |
if (!s->q_failed && s->failed < 2) { |
d82dfee0a
|
2710 |
/* Q is not failed, and we didn't use it to generate |
a44568564
|
2711 2712 |
* anything, so it makes sense to check it */ |
d82dfee0a
|
2713 2714 2715 2716 |
if (sh->check_state == check_state_run) sh->check_state = check_state_run_pq; else sh->check_state = check_state_run_q; |
a44568564
|
2717 |
} |
a44568564
|
2718 |
|
d82dfee0a
|
2719 2720 |
/* discard potentially stale zero_sum_result */ sh->ops.zero_sum_result = 0; |
a44568564
|
2721 |
|
d82dfee0a
|
2722 2723 2724 2725 |
if (sh->check_state == check_state_run) { /* async_xor_zero_sum destroys the contents of P */ clear_bit(R5_UPTODATE, &sh->dev[pd_idx].flags); s->uptodate--; |
a44568564
|
2726 |
} |
d82dfee0a
|
2727 2728 2729 2730 2731 2732 2733 |
if (sh->check_state >= check_state_run && sh->check_state <= check_state_run_pq) { /* async_syndrome_zero_sum preserves P and Q, so * no need to mark them !uptodate here */ set_bit(STRIPE_OP_CHECK, &s->ops_request); break; |
a44568564
|
2734 |
} |
d82dfee0a
|
2735 2736 2737 2738 2739 |
/* we have 2-disk failure */ BUG_ON(s->failed != 2); /* fall through */ case check_state_compute_result: sh->check_state = check_state_idle; |
a44568564
|
2740 |
|
d82dfee0a
|
2741 2742 2743 |
/* check that a write has not made the stripe insync */ if (test_bit(STRIPE_INSYNC, &sh->state)) break; |
a44568564
|
2744 2745 |
/* now write out any block on a failed drive, |
d82dfee0a
|
2746 |
* or P or Q if they were recomputed |
a44568564
|
2747 |
*/ |
d82dfee0a
|
2748 |
BUG_ON(s->uptodate < disks - 1); /* We don't need Q to recover */ |
a44568564
|
2749 |
if (s->failed == 2) { |
f2b3b44de
|
2750 |
dev = &sh->dev[s->failed_num[1]]; |
a44568564
|
2751 2752 2753 2754 2755 |
s->locked++; set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantwrite, &dev->flags); } if (s->failed >= 1) { |
f2b3b44de
|
2756 |
dev = &sh->dev[s->failed_num[0]]; |
a44568564
|
2757 2758 2759 2760 |
s->locked++; set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantwrite, &dev->flags); } |
d82dfee0a
|
2761 |
if (sh->ops.zero_sum_result & SUM_CHECK_P_RESULT) { |
a44568564
|
2762 2763 2764 2765 2766 |
dev = &sh->dev[pd_idx]; s->locked++; set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantwrite, &dev->flags); } |
d82dfee0a
|
2767 |
if (sh->ops.zero_sum_result & SUM_CHECK_Q_RESULT) { |
a44568564
|
2768 2769 2770 2771 2772 2773 2774 2775 |
dev = &sh->dev[qd_idx]; s->locked++; set_bit(R5_LOCKED, &dev->flags); set_bit(R5_Wantwrite, &dev->flags); } clear_bit(STRIPE_DEGRADED, &sh->state); set_bit(STRIPE_INSYNC, &sh->state); |
d82dfee0a
|
2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 |
break; case check_state_run: case check_state_run_q: case check_state_run_pq: break; /* we will be called again upon completion */ case check_state_check_result: sh->check_state = check_state_idle; /* handle a successful check operation, if parity is correct * we are done. Otherwise update the mismatch count and repair * parity if !MD_RECOVERY_CHECK */ if (sh->ops.zero_sum_result == 0) { /* both parities are correct */ if (!s->failed) set_bit(STRIPE_INSYNC, &sh->state); else { /* in contrast to the raid5 case we can validate * parity, but still have a failure to write * back */ sh->check_state = check_state_compute_result; /* Returning at this point means that we may go * off and bring p and/or q uptodate again so * we make sure to check zero_sum_result again * to verify if p or q need writeback */ } } else { conf->mddev->resync_mismatches += STRIPE_SECTORS; if (test_bit(MD_RECOVERY_CHECK, &conf->mddev->recovery)) /* don't try to repair!! */ set_bit(STRIPE_INSYNC, &sh->state); else { int *target = &sh->ops.target; sh->ops.target = -1; sh->ops.target2 = -1; sh->check_state = check_state_compute_run; set_bit(STRIPE_COMPUTE_RUN, &sh->state); set_bit(STRIPE_OP_COMPUTE_BLK, &s->ops_request); if (sh->ops.zero_sum_result & SUM_CHECK_P_RESULT) { set_bit(R5_Wantcompute, &sh->dev[pd_idx].flags); *target = pd_idx; target = &sh->ops.target2; s->uptodate++; } if (sh->ops.zero_sum_result & SUM_CHECK_Q_RESULT) { set_bit(R5_Wantcompute, &sh->dev[qd_idx].flags); *target = qd_idx; s->uptodate++; } } } break; case check_state_compute_run: break; default: printk(KERN_ERR "%s: unknown check_state: %d sector: %llu ", __func__, sh->check_state, (unsigned long long) sh->sector); BUG(); |
a44568564
|
2841 2842 |
} } |
d1688a6d5
|
2843 |
static void handle_stripe_expansion(struct r5conf *conf, struct stripe_head *sh) |
a44568564
|
2844 2845 2846 2847 2848 2849 |
{ int i; /* We have read all the blocks in this stripe and now we need to * copy some of them into a target stripe for expand. */ |
f0a50d375
|
2850 |
struct dma_async_tx_descriptor *tx = NULL; |
a44568564
|
2851 2852 |
clear_bit(STRIPE_EXPAND_SOURCE, &sh->state); for (i = 0; i < sh->disks; i++) |
34e04e87f
|
2853 |
if (i != sh->pd_idx && i != sh->qd_idx) { |
911d4ee85
|
2854 |
int dd_idx, j; |
a44568564
|
2855 |
struct stripe_head *sh2; |
a08abd8ca
|
2856 |
struct async_submit_ctl submit; |
a44568564
|
2857 |
|
784052ecc
|
2858 |
sector_t bn = compute_blocknr(sh, i, 1); |
911d4ee85
|
2859 2860 |
sector_t s = raid5_compute_sector(conf, bn, 0, &dd_idx, NULL); |
a8c906ca3
|
2861 |
sh2 = get_active_stripe(conf, s, 0, 1, 1); |
a44568564
|
2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 |
if (sh2 == NULL) /* so far only the early blocks of this stripe * have been requested. When later blocks * get requested, we will try again */ continue; if (!test_bit(STRIPE_EXPANDING, &sh2->state) || test_bit(R5_Expanded, &sh2->dev[dd_idx].flags)) { /* must have already done this block */ release_stripe(sh2); continue; } |
f0a50d375
|
2874 2875 |
/* place all the copies on one channel */ |
a08abd8ca
|
2876 |
init_async_submit(&submit, 0, tx, NULL, NULL, NULL); |
f0a50d375
|
2877 |
tx = async_memcpy(sh2->dev[dd_idx].page, |
88ba2aa58
|
2878 |
sh->dev[i].page, 0, 0, STRIPE_SIZE, |
a08abd8ca
|
2879 |
&submit); |
f0a50d375
|
2880 |
|
a44568564
|
2881 2882 2883 2884 |
set_bit(R5_Expanded, &sh2->dev[dd_idx].flags); set_bit(R5_UPTODATE, &sh2->dev[dd_idx].flags); for (j = 0; j < conf->raid_disks; j++) if (j != sh2->pd_idx && |
86c374ba9
|
2885 |
j != sh2->qd_idx && |
a44568564
|
2886 2887 2888 2889 2890 2891 2892 |
!test_bit(R5_Expanded, &sh2->dev[j].flags)) break; if (j == conf->raid_disks) { set_bit(STRIPE_EXPAND_READY, &sh2->state); set_bit(STRIPE_HANDLE, &sh2->state); } release_stripe(sh2); |
f0a50d375
|
2893 |
|
a44568564
|
2894 |
} |
a2e085518
|
2895 2896 2897 2898 2899 |
/* done submitting copies, wait for them to complete */ if (tx) { async_tx_ack(tx); dma_wait_for_async_tx(tx); } |
a44568564
|
2900 |
} |
1da177e4c
|
2901 |
|
6bfe0b499
|
2902 |
|
1da177e4c
|
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 |
/* * handle_stripe - do things to a stripe. * * We lock the stripe and then examine the state of various bits * to see what needs to be done. * Possible results: * return some read request which now have data * return some write requests which are safely on disc * schedule a read on some buffers * schedule a write of some buffers * return confirmation of parity correctness * |
1da177e4c
|
2915 2916 2917 2918 |
* buffers are taken off read_list or write_list, and bh_cache buffers * get BH_Lock set before the stripe lock is released. * */ |
a44568564
|
2919 |
|
acfe726bd
|
2920 |
static void analyse_stripe(struct stripe_head *sh, struct stripe_head_state *s) |
1da177e4c
|
2921 |
{ |
d1688a6d5
|
2922 |
struct r5conf *conf = sh->raid_conf; |
f416885ef
|
2923 |
int disks = sh->disks; |
474af965f
|
2924 2925 |
struct r5dev *dev; int i; |
1da177e4c
|
2926 |
|
acfe726bd
|
2927 2928 2929 2930 2931 2932 2933 |
memset(s, 0, sizeof(*s)); s->syncing = test_bit(STRIPE_SYNCING, &sh->state); s->expanding = test_bit(STRIPE_EXPAND_SOURCE, &sh->state); s->expanded = test_bit(STRIPE_EXPAND_READY, &sh->state); s->failed_num[0] = -1; s->failed_num[1] = -1; |
1da177e4c
|
2934 |
|
acfe726bd
|
2935 |
/* Now to look around and see what can be done */ |
1da177e4c
|
2936 |
rcu_read_lock(); |
c4c1663be
|
2937 |
spin_lock_irq(&conf->device_lock); |
16a53ecc3
|
2938 |
for (i=disks; i--; ) { |
3cb030020
|
2939 |
struct md_rdev *rdev; |
31c176ecd
|
2940 2941 2942 |
sector_t first_bad; int bad_sectors; int is_bad = 0; |
acfe726bd
|
2943 |
|
16a53ecc3
|
2944 |
dev = &sh->dev[i]; |
1da177e4c
|
2945 |
|
45b4233ca
|
2946 2947 |
pr_debug("check %d: state 0x%lx read %p write %p written %p ", |
16a53ecc3
|
2948 |
i, dev->flags, dev->toread, dev->towrite, dev->written); |
6c0069c0a
|
2949 2950 2951 2952 2953 2954 2955 2956 |
/* maybe we can reply to a read * * new wantfill requests are only permitted while * ops_complete_biofill is guaranteed to be inactive */ if (test_bit(R5_UPTODATE, &dev->flags) && dev->toread && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) set_bit(R5_Wantfill, &dev->flags); |
1da177e4c
|
2957 |
|
16a53ecc3
|
2958 |
/* now count some things */ |
cc94015a9
|
2959 2960 2961 2962 |
if (test_bit(R5_LOCKED, &dev->flags)) s->locked++; if (test_bit(R5_UPTODATE, &dev->flags)) s->uptodate++; |
2d6e4ecc8
|
2963 |
if (test_bit(R5_Wantcompute, &dev->flags)) { |
cc94015a9
|
2964 2965 |
s->compute++; BUG_ON(s->compute > 2); |
2d6e4ecc8
|
2966 |
} |
1da177e4c
|
2967 |
|
acfe726bd
|
2968 |
if (test_bit(R5_Wantfill, &dev->flags)) |
cc94015a9
|
2969 |
s->to_fill++; |
acfe726bd
|
2970 |
else if (dev->toread) |
cc94015a9
|
2971 |
s->to_read++; |
16a53ecc3
|
2972 |
if (dev->towrite) { |
cc94015a9
|
2973 |
s->to_write++; |
16a53ecc3
|
2974 |
if (!test_bit(R5_OVERWRITE, &dev->flags)) |
cc94015a9
|
2975 |
s->non_overwrite++; |
16a53ecc3
|
2976 |
} |
a44568564
|
2977 |
if (dev->written) |
cc94015a9
|
2978 |
s->written++; |
16a53ecc3
|
2979 |
rdev = rcu_dereference(conf->disks[i].rdev); |
9283d8c5a
|
2980 2981 |
if (rdev && test_bit(Faulty, &rdev->flags)) rdev = NULL; |
31c176ecd
|
2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 |
if (rdev) { is_bad = is_badblock(rdev, sh->sector, STRIPE_SECTORS, &first_bad, &bad_sectors); if (s->blocked_rdev == NULL && (test_bit(Blocked, &rdev->flags) || is_bad < 0)) { if (is_bad < 0) set_bit(BlockedBadBlocks, &rdev->flags); s->blocked_rdev = rdev; atomic_inc(&rdev->nr_pending); } |
6bfe0b499
|
2994 |
} |
415e72d03
|
2995 2996 2997 |
clear_bit(R5_Insync, &dev->flags); if (!rdev) /* Not in-sync */; |
31c176ecd
|
2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 |
else if (is_bad) { /* also not in-sync */ if (!test_bit(WriteErrorSeen, &rdev->flags)) { /* treat as in-sync, but with a read error * which we can now try to correct */ set_bit(R5_Insync, &dev->flags); set_bit(R5_ReadError, &dev->flags); } } else if (test_bit(In_sync, &rdev->flags)) |
415e72d03
|
3008 |
set_bit(R5_Insync, &dev->flags); |
30d7a4836
|
3009 |
else if (sh->sector + STRIPE_SECTORS <= rdev->recovery_offset) |
415e72d03
|
3010 |
/* in sync if before recovery_offset */ |
30d7a4836
|
3011 3012 3013 3014 3015 3016 3017 3018 |
set_bit(R5_Insync, &dev->flags); else if (test_bit(R5_UPTODATE, &dev->flags) && test_bit(R5_Expanded, &dev->flags)) /* If we've reshaped into here, we assume it is Insync. * We will shortly update recovery_offset to make * it official. */ set_bit(R5_Insync, &dev->flags); |
5d8c71f9e
|
3019 |
if (rdev && test_bit(R5_WriteError, &dev->flags)) { |
bc2607f39
|
3020 3021 3022 3023 3024 3025 3026 |
clear_bit(R5_Insync, &dev->flags); if (!test_bit(Faulty, &rdev->flags)) { s->handle_bad_blocks = 1; atomic_inc(&rdev->nr_pending); } else clear_bit(R5_WriteError, &dev->flags); } |
5d8c71f9e
|
3027 |
if (rdev && test_bit(R5_MadeGood, &dev->flags)) { |
b84db560e
|
3028 3029 3030 3031 3032 3033 |
if (!test_bit(Faulty, &rdev->flags)) { s->handle_bad_blocks = 1; atomic_inc(&rdev->nr_pending); } else clear_bit(R5_MadeGood, &dev->flags); } |
415e72d03
|
3034 |
if (!test_bit(R5_Insync, &dev->flags)) { |
16a53ecc3
|
3035 3036 3037 |
/* The ReadError flag will just be confusing now */ clear_bit(R5_ReadError, &dev->flags); clear_bit(R5_ReWrite, &dev->flags); |
1da177e4c
|
3038 |
} |
415e72d03
|
3039 3040 3041 |
if (test_bit(R5_ReadError, &dev->flags)) clear_bit(R5_Insync, &dev->flags); if (!test_bit(R5_Insync, &dev->flags)) { |
cc94015a9
|
3042 3043 3044 |
if (s->failed < 2) s->failed_num[s->failed] = i; s->failed++; |
415e72d03
|
3045 |
} |
1da177e4c
|
3046 |
} |
c4c1663be
|
3047 |
spin_unlock_irq(&conf->device_lock); |
1da177e4c
|
3048 |
rcu_read_unlock(); |
cc94015a9
|
3049 3050 3051 3052 3053 |
} static void handle_stripe(struct stripe_head *sh) { struct stripe_head_state s; |
d1688a6d5
|
3054 |
struct r5conf *conf = sh->raid_conf; |
3687c0618
|
3055 |
int i; |
84789554e
|
3056 3057 |
int prexor; int disks = sh->disks; |
474af965f
|
3058 |
struct r5dev *pdev, *qdev; |
cc94015a9
|
3059 3060 |
clear_bit(STRIPE_HANDLE, &sh->state); |
257a4b42a
|
3061 |
if (test_and_set_bit_lock(STRIPE_ACTIVE, &sh->state)) { |
cc94015a9
|
3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 |
/* already being handled, ensure it gets handled * again when current action finishes */ set_bit(STRIPE_HANDLE, &sh->state); return; } if (test_and_clear_bit(STRIPE_SYNC_REQUESTED, &sh->state)) { set_bit(STRIPE_SYNCING, &sh->state); clear_bit(STRIPE_INSYNC, &sh->state); } clear_bit(STRIPE_DELAYED, &sh->state); pr_debug("handling stripe %llu, state=%#lx cnt=%d, " "pd_idx=%d, qd_idx=%d , check:%d, reconstruct:%d ", (unsigned long long)sh->sector, sh->state, atomic_read(&sh->count), sh->pd_idx, sh->qd_idx, sh->check_state, sh->reconstruct_state); |
3687c0618
|
3081 |
|
acfe726bd
|
3082 |
analyse_stripe(sh, &s); |
c5a310006
|
3083 |
|
bc2607f39
|
3084 3085 3086 3087 |
if (s.handle_bad_blocks) { set_bit(STRIPE_HANDLE, &sh->state); goto finish; } |
474af965f
|
3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 |
if (unlikely(s.blocked_rdev)) { if (s.syncing || s.expanding || s.expanded || s.to_write || s.written) { set_bit(STRIPE_HANDLE, &sh->state); goto finish; } /* There is nothing for the blocked_rdev to block */ rdev_dec_pending(s.blocked_rdev, conf->mddev); s.blocked_rdev = NULL; } if (s.to_fill && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) { set_bit(STRIPE_OP_BIOFILL, &s.ops_request); set_bit(STRIPE_BIOFILL_RUN, &sh->state); } pr_debug("locked=%d uptodate=%d to_read=%d" " to_write=%d failed=%d failed_num=%d,%d ", s.locked, s.uptodate, s.to_read, s.to_write, s.failed, s.failed_num[0], s.failed_num[1]); /* check if the array has lost more than max_degraded devices and, * if so, some requests might need to be failed. */ |
9a3f530f3
|
3112 3113 3114 3115 3116 3117 3118 3119 |
if (s.failed > conf->max_degraded) { sh->check_state = 0; sh->reconstruct_state = 0; if (s.to_read+s.to_write+s.written) handle_failed_stripe(conf, sh, &s, disks, &s.return_bi); if (s.syncing) handle_failed_sync(conf, sh, &s); } |
474af965f
|
3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 |
/* * might be able to return some write requests if the parity blocks * are safe, or on a failed drive */ pdev = &sh->dev[sh->pd_idx]; s.p_failed = (s.failed >= 1 && s.failed_num[0] == sh->pd_idx) || (s.failed >= 2 && s.failed_num[1] == sh->pd_idx); qdev = &sh->dev[sh->qd_idx]; s.q_failed = (s.failed >= 1 && s.failed_num[0] == sh->qd_idx) || (s.failed >= 2 && s.failed_num[1] == sh->qd_idx) || conf->level < 6; if (s.written && (s.p_failed || ((test_bit(R5_Insync, &pdev->flags) && !test_bit(R5_LOCKED, &pdev->flags) && test_bit(R5_UPTODATE, &pdev->flags)))) && (s.q_failed || ((test_bit(R5_Insync, &qdev->flags) && !test_bit(R5_LOCKED, &qdev->flags) && test_bit(R5_UPTODATE, &qdev->flags))))) handle_stripe_clean_event(conf, sh, disks, &s.return_bi); /* Now we might consider reading some blocks, either to check/generate * parity, or to satisfy requests * or to load a block that is being partially written. */ if (s.to_read || s.non_overwrite || (conf->level == 6 && s.to_write && s.failed) || (s.syncing && (s.uptodate + s.compute < disks)) || s.expanding) handle_stripe_fill(sh, &s, disks); |
84789554e
|
3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 |
/* Now we check to see if any write operations have recently * completed */ prexor = 0; if (sh->reconstruct_state == reconstruct_state_prexor_drain_result) prexor = 1; if (sh->reconstruct_state == reconstruct_state_drain_result || sh->reconstruct_state == reconstruct_state_prexor_drain_result) { sh->reconstruct_state = reconstruct_state_idle; /* All the 'written' buffers and the parity block are ready to * be written back to disk */ BUG_ON(!test_bit(R5_UPTODATE, &sh->dev[sh->pd_idx].flags)); BUG_ON(sh->qd_idx >= 0 && !test_bit(R5_UPTODATE, &sh->dev[sh->qd_idx].flags)); for (i = disks; i--; ) { struct r5dev *dev = &sh->dev[i]; if (test_bit(R5_LOCKED, &dev->flags) && (i == sh->pd_idx || i == sh->qd_idx || dev->written)) { pr_debug("Writing block %d ", i); set_bit(R5_Wantwrite, &dev->flags); if (prexor) continue; if (!test_bit(R5_Insync, &dev->flags) || ((i == sh->pd_idx || i == sh->qd_idx) && s.failed == 0)) set_bit(STRIPE_INSYNC, &sh->state); } } if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) s.dec_preread_active = 1; } /* Now to consider new write requests and what else, if anything * should be read. We do not handle new writes when: * 1/ A 'write' operation (copy+xor) is already in flight. * 2/ A 'check' operation is in flight, as it may clobber the parity * block. */ if (s.to_write && !sh->reconstruct_state && !sh->check_state) handle_stripe_dirtying(conf, sh, &s, disks); /* maybe we need to check and possibly fix the parity for this stripe * Any reads will already have been scheduled, so we just see if enough * data is available. The parity check is held off while parity * dependent operations are in flight. */ if (sh->check_state || (s.syncing && s.locked == 0 && !test_bit(STRIPE_COMPUTE_RUN, &sh->state) && !test_bit(STRIPE_INSYNC, &sh->state))) { if (conf->level == 6) handle_parity_checks6(conf, sh, &s, disks); else handle_parity_checks5(conf, sh, &s, disks); } |
c5a310006
|
3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 |
if (s.syncing && s.locked == 0 && test_bit(STRIPE_INSYNC, &sh->state)) { md_done_sync(conf->mddev, STRIPE_SECTORS, 1); clear_bit(STRIPE_SYNCING, &sh->state); } /* If the failed drives are just a ReadError, then we might need * to progress the repair/check process */ if (s.failed <= conf->max_degraded && !conf->mddev->ro) for (i = 0; i < s.failed; i++) { struct r5dev *dev = &sh->dev[s.failed_num[i]]; if (test_bit(R5_ReadError, &dev->flags) && !test_bit(R5_LOCKED, &dev->flags) && test_bit(R5_UPTODATE, &dev->flags) ) { if (!test_bit(R5_ReWrite, &dev->flags)) { set_bit(R5_Wantwrite, &dev->flags); set_bit(R5_ReWrite, &dev->flags); set_bit(R5_LOCKED, &dev->flags); s.locked++; } else { /* let's read it back */ set_bit(R5_Wantread, &dev->flags); set_bit(R5_LOCKED, &dev->flags); s.locked++; } } } |
3687c0618
|
3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 |
/* Finish reconstruct operations initiated by the expansion process */ if (sh->reconstruct_state == reconstruct_state_result) { struct stripe_head *sh_src = get_active_stripe(conf, sh->sector, 1, 1, 1); if (sh_src && test_bit(STRIPE_EXPAND_SOURCE, &sh_src->state)) { /* sh cannot be written until sh_src has been read. * so arrange for sh to be delayed a little */ set_bit(STRIPE_DELAYED, &sh->state); set_bit(STRIPE_HANDLE, &sh->state); if (!test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh_src->state)) atomic_inc(&conf->preread_active_stripes); release_stripe(sh_src); goto finish; } if (sh_src) release_stripe(sh_src); sh->reconstruct_state = reconstruct_state_idle; clear_bit(STRIPE_EXPANDING, &sh->state); for (i = conf->raid_disks; i--; ) { set_bit(R5_Wantwrite, &sh->dev[i].flags); set_bit(R5_LOCKED, &sh->dev[i].flags); s.locked++; } } |
f416885ef
|
3265 |
|
3687c0618
|
3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 |
if (s.expanded && test_bit(STRIPE_EXPANDING, &sh->state) && !sh->reconstruct_state) { /* Need to write out all blocks after computing parity */ sh->disks = conf->raid_disks; stripe_set_idx(sh->sector, conf, 0, sh); schedule_reconstruction(sh, &s, 1, 1); } else if (s.expanded && !sh->reconstruct_state && s.locked == 0) { clear_bit(STRIPE_EXPAND_READY, &sh->state); atomic_dec(&conf->reshape_stripes); wake_up(&conf->wait_for_overlap); md_done_sync(conf->mddev, STRIPE_SECTORS, 1); } if (s.expanding && s.locked == 0 && !test_bit(STRIPE_COMPUTE_RUN, &sh->state)) handle_stripe_expansion(conf, sh); |
16a53ecc3
|
3282 |
|
3687c0618
|
3283 |
finish: |
6bfe0b499
|
3284 |
/* wait for this device to become unblocked */ |
43220aa0f
|
3285 |
if (conf->mddev->external && unlikely(s.blocked_rdev)) |
c5709ef6a
|
3286 |
md_wait_for_blocked_rdev(s.blocked_rdev, conf->mddev); |
6bfe0b499
|
3287 |
|
bc2607f39
|
3288 3289 |
if (s.handle_bad_blocks) for (i = disks; i--; ) { |
3cb030020
|
3290 |
struct md_rdev *rdev; |
bc2607f39
|
3291 3292 3293 3294 3295 3296 3297 3298 3299 |
struct r5dev *dev = &sh->dev[i]; if (test_and_clear_bit(R5_WriteError, &dev->flags)) { /* We own a safe reference to the rdev */ rdev = conf->disks[i].rdev; if (!rdev_set_badblocks(rdev, sh->sector, STRIPE_SECTORS, 0)) md_error(conf->mddev, rdev); rdev_dec_pending(rdev, conf->mddev); } |
b84db560e
|
3300 3301 3302 3303 3304 3305 |
if (test_and_clear_bit(R5_MadeGood, &dev->flags)) { rdev = conf->disks[i].rdev; rdev_clear_badblocks(rdev, sh->sector, STRIPE_SECTORS); rdev_dec_pending(rdev, conf->mddev); } |
bc2607f39
|
3306 |
} |
6c0069c0a
|
3307 3308 |
if (s.ops_request) raid_run_ops(sh, s.ops_request); |
f0e43bcde
|
3309 |
ops_run_io(sh, &s); |
16a53ecc3
|
3310 |
|
c5709ef6a
|
3311 |
if (s.dec_preread_active) { |
729a18663
|
3312 |
/* We delay this until after ops_run_io so that if make_request |
e9c7469bb
|
3313 |
* is waiting on a flush, it won't continue until the writes |
729a18663
|
3314 3315 3316 3317 3318 3319 3320 |
* have actually been submitted. */ atomic_dec(&conf->preread_active_stripes); if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD) md_wakeup_thread(conf->mddev->thread); } |
c5709ef6a
|
3321 |
return_io(s.return_bi); |
16a53ecc3
|
3322 |
|
257a4b42a
|
3323 |
clear_bit_unlock(STRIPE_ACTIVE, &sh->state); |
16a53ecc3
|
3324 |
} |
d1688a6d5
|
3325 |
static void raid5_activate_delayed(struct r5conf *conf) |
16a53ecc3
|
3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 |
{ if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD) { while (!list_empty(&conf->delayed_list)) { struct list_head *l = conf->delayed_list.next; struct stripe_head *sh; sh = list_entry(l, struct stripe_head, lru); list_del_init(l); clear_bit(STRIPE_DELAYED, &sh->state); if (!test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) atomic_inc(&conf->preread_active_stripes); |
8b3e6cdc5
|
3336 |
list_add_tail(&sh->lru, &conf->hold_list); |
16a53ecc3
|
3337 |
} |
482c08349
|
3338 |
} |
16a53ecc3
|
3339 |
} |
d1688a6d5
|
3340 |
static void activate_bit_delay(struct r5conf *conf) |
16a53ecc3
|
3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 |
{ /* device_lock is held */ struct list_head head; list_add(&head, &conf->bitmap_list); list_del_init(&conf->bitmap_list); while (!list_empty(&head)) { struct stripe_head *sh = list_entry(head.next, struct stripe_head, lru); list_del_init(&sh->lru); atomic_inc(&sh->count); __release_stripe(conf, sh); } } |
fd01b88c7
|
3353 |
int md_raid5_congested(struct mddev *mddev, int bits) |
f022b2fdd
|
3354 |
{ |
d1688a6d5
|
3355 |
struct r5conf *conf = mddev->private; |
f022b2fdd
|
3356 3357 3358 3359 |
/* No difference between reads and writes. Just check * how busy the stripe_cache is */ |
3fa841d7e
|
3360 |
|
f022b2fdd
|
3361 3362 3363 3364 3365 3366 3367 3368 3369 |
if (conf->inactive_blocked) return 1; if (conf->quiesce) return 1; if (list_empty_careful(&conf->inactive_list)) return 1; return 0; } |
11d8a6e37
|
3370 3371 3372 3373 |
EXPORT_SYMBOL_GPL(md_raid5_congested); static int raid5_congested(void *data, int bits) { |
fd01b88c7
|
3374 |
struct mddev *mddev = data; |
11d8a6e37
|
3375 3376 3377 3378 |
return mddev_congested(mddev, bits) || md_raid5_congested(mddev, bits); } |
f022b2fdd
|
3379 |
|
23032a0eb
|
3380 3381 3382 |
/* We want read requests to align with chunks where possible, * but write requests don't need to. */ |
cc371e66e
|
3383 3384 3385 |
static int raid5_mergeable_bvec(struct request_queue *q, struct bvec_merge_data *bvm, struct bio_vec *biovec) |
23032a0eb
|
3386 |
{ |
fd01b88c7
|
3387 |
struct mddev *mddev = q->queuedata; |
cc371e66e
|
3388 |
sector_t sector = bvm->bi_sector + get_start_sect(bvm->bi_bdev); |
23032a0eb
|
3389 |
int max; |
9d8f03636
|
3390 |
unsigned int chunk_sectors = mddev->chunk_sectors; |
cc371e66e
|
3391 |
unsigned int bio_sectors = bvm->bi_size >> 9; |
23032a0eb
|
3392 |
|
cc371e66e
|
3393 |
if ((bvm->bi_rw & 1) == WRITE) |
23032a0eb
|
3394 |
return biovec->bv_len; /* always allow writes to be mergeable */ |
664e7c413
|
3395 3396 |
if (mddev->new_chunk_sectors < mddev->chunk_sectors) chunk_sectors = mddev->new_chunk_sectors; |
23032a0eb
|
3397 3398 3399 3400 3401 3402 3403 |
max = (chunk_sectors - ((sector & (chunk_sectors - 1)) + bio_sectors)) << 9; if (max < 0) max = 0; if (max <= biovec->bv_len && bio_sectors == 0) return biovec->bv_len; else return max; } |
f679623f5
|
3404 |
|
fd01b88c7
|
3405 |
static int in_chunk_boundary(struct mddev *mddev, struct bio *bio) |
f679623f5
|
3406 3407 |
{ sector_t sector = bio->bi_sector + get_start_sect(bio->bi_bdev); |
9d8f03636
|
3408 |
unsigned int chunk_sectors = mddev->chunk_sectors; |
f679623f5
|
3409 |
unsigned int bio_sectors = bio->bi_size >> 9; |
664e7c413
|
3410 3411 |
if (mddev->new_chunk_sectors < mddev->chunk_sectors) chunk_sectors = mddev->new_chunk_sectors; |
f679623f5
|
3412 3413 3414 3415 3416 |
return chunk_sectors >= ((sector & (chunk_sectors - 1)) + bio_sectors); } /* |
46031f9a3
|
3417 3418 3419 |
* add bio to the retry LIFO ( in O(1) ... we are in interrupt ) * later sampled by raid5d. */ |
d1688a6d5
|
3420 |
static void add_bio_to_retry(struct bio *bi,struct r5conf *conf) |
46031f9a3
|
3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 |
{ unsigned long flags; spin_lock_irqsave(&conf->device_lock, flags); bi->bi_next = conf->retry_read_aligned_list; conf->retry_read_aligned_list = bi; spin_unlock_irqrestore(&conf->device_lock, flags); md_wakeup_thread(conf->mddev->thread); } |
d1688a6d5
|
3432 |
static struct bio *remove_bio_from_retry(struct r5conf *conf) |
46031f9a3
|
3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 |
{ struct bio *bi; bi = conf->retry_read_aligned; if (bi) { conf->retry_read_aligned = NULL; return bi; } bi = conf->retry_read_aligned_list; if(bi) { |
387bb1737
|
3443 |
conf->retry_read_aligned_list = bi->bi_next; |
46031f9a3
|
3444 |
bi->bi_next = NULL; |
960e739d9
|
3445 3446 3447 3448 |
/* * this sets the active strip count to 1 and the processed * strip count to zero (upper 8 bits) */ |
46031f9a3
|
3449 |
bi->bi_phys_segments = 1; /* biased count of active stripes */ |
46031f9a3
|
3450 3451 3452 3453 3454 3455 3456 |
} return bi; } /* |
f679623f5
|
3457 3458 3459 3460 3461 |
* The "raid5_align_endio" should check if the read succeeded and if it * did, call bio_endio on the original bio (having bio_put the new bio * first). * If the read failed.. */ |
6712ecf8f
|
3462 |
static void raid5_align_endio(struct bio *bi, int error) |
f679623f5
|
3463 3464 |
{ struct bio* raid_bi = bi->bi_private; |
fd01b88c7
|
3465 |
struct mddev *mddev; |
d1688a6d5
|
3466 |
struct r5conf *conf; |
46031f9a3
|
3467 |
int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags); |
3cb030020
|
3468 |
struct md_rdev *rdev; |
46031f9a3
|
3469 |
|
f679623f5
|
3470 |
bio_put(bi); |
46031f9a3
|
3471 |
|
46031f9a3
|
3472 3473 |
rdev = (void*)raid_bi->bi_next; raid_bi->bi_next = NULL; |
2b7f22284
|
3474 3475 |
mddev = rdev->mddev; conf = mddev->private; |
46031f9a3
|
3476 3477 3478 3479 |
rdev_dec_pending(rdev, conf->mddev); if (!error && uptodate) { |
6712ecf8f
|
3480 |
bio_endio(raid_bi, 0); |
46031f9a3
|
3481 3482 |
if (atomic_dec_and_test(&conf->active_aligned_reads)) wake_up(&conf->wait_for_stripe); |
6712ecf8f
|
3483 |
return; |
46031f9a3
|
3484 |
} |
45b4233ca
|
3485 3486 |
pr_debug("raid5_align_endio : io error...handing IO for a retry "); |
46031f9a3
|
3487 3488 |
add_bio_to_retry(raid_bi, conf); |
f679623f5
|
3489 |
} |
387bb1737
|
3490 3491 |
static int bio_fits_rdev(struct bio *bi) { |
165125e1e
|
3492 |
struct request_queue *q = bdev_get_queue(bi->bi_bdev); |
387bb1737
|
3493 |
|
ae03bf639
|
3494 |
if ((bi->bi_size>>9) > queue_max_sectors(q)) |
387bb1737
|
3495 3496 |
return 0; blk_recount_segments(q, bi); |
8a78362c4
|
3497 |
if (bi->bi_phys_segments > queue_max_segments(q)) |
387bb1737
|
3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 |
return 0; if (q->merge_bvec_fn) /* it's too hard to apply the merge_bvec_fn at this stage, * just just give up */ return 0; return 1; } |
fd01b88c7
|
3508 |
static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio) |
f679623f5
|
3509 |
{ |
d1688a6d5
|
3510 |
struct r5conf *conf = mddev->private; |
8553fe7ec
|
3511 |
int dd_idx; |
f679623f5
|
3512 |
struct bio* align_bi; |
3cb030020
|
3513 |
struct md_rdev *rdev; |
f679623f5
|
3514 3515 |
if (!in_chunk_boundary(mddev, raid_bio)) { |
45b4233ca
|
3516 3517 |
pr_debug("chunk_aligned_read : non aligned "); |
f679623f5
|
3518 3519 3520 |
return 0; } /* |
a167f6632
|
3521 |
* use bio_clone_mddev to make a copy of the bio |
f679623f5
|
3522 |
*/ |
a167f6632
|
3523 |
align_bi = bio_clone_mddev(raid_bio, GFP_NOIO, mddev); |
f679623f5
|
3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 |
if (!align_bi) return 0; /* * set bi_end_io to a new function, and set bi_private to the * original bio. */ align_bi->bi_end_io = raid5_align_endio; align_bi->bi_private = raid_bio; /* * compute position */ |
112bf8970
|
3535 3536 |
align_bi->bi_sector = raid5_compute_sector(conf, raid_bio->bi_sector, 0, |
911d4ee85
|
3537 |
&dd_idx, NULL); |
f679623f5
|
3538 3539 3540 3541 |
rcu_read_lock(); rdev = rcu_dereference(conf->disks[dd_idx].rdev); if (rdev && test_bit(In_sync, &rdev->flags)) { |
31c176ecd
|
3542 3543 |
sector_t first_bad; int bad_sectors; |
f679623f5
|
3544 3545 |
atomic_inc(&rdev->nr_pending); rcu_read_unlock(); |
46031f9a3
|
3546 3547 3548 3549 |
raid_bio->bi_next = (void*)rdev; align_bi->bi_bdev = rdev->bdev; align_bi->bi_flags &= ~(1 << BIO_SEG_VALID); align_bi->bi_sector += rdev->data_offset; |
31c176ecd
|
3550 3551 3552 3553 |
if (!bio_fits_rdev(align_bi) || is_badblock(rdev, align_bi->bi_sector, align_bi->bi_size>>9, &first_bad, &bad_sectors)) { /* too big in some way, or has a known bad block */ |
387bb1737
|
3554 3555 3556 3557 |
bio_put(align_bi); rdev_dec_pending(rdev, mddev); return 0; } |
46031f9a3
|
3558 3559 3560 3561 3562 3563 |
spin_lock_irq(&conf->device_lock); wait_event_lock_irq(conf->wait_for_stripe, conf->quiesce == 0, conf->device_lock, /* nothing */); atomic_inc(&conf->active_aligned_reads); spin_unlock_irq(&conf->device_lock); |
f679623f5
|
3564 3565 3566 3567 |
generic_make_request(align_bi); return 1; } else { rcu_read_unlock(); |
46031f9a3
|
3568 |
bio_put(align_bi); |
f679623f5
|
3569 3570 3571 |
return 0; } } |
8b3e6cdc5
|
3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 |
/* __get_priority_stripe - get the next stripe to process * * Full stripe writes are allowed to pass preread active stripes up until * the bypass_threshold is exceeded. In general the bypass_count * increments when the handle_list is handled before the hold_list; however, it * will not be incremented when STRIPE_IO_STARTED is sampled set signifying a * stripe with in flight i/o. The bypass_count will be reset when the * head of the hold_list has changed, i.e. the head was promoted to the * handle_list. */ |
d1688a6d5
|
3582 |
static struct stripe_head *__get_priority_stripe(struct r5conf *conf) |
8b3e6cdc5
|
3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 |
{ struct stripe_head *sh; pr_debug("%s: handle: %s hold: %s full_writes: %d bypass_count: %d ", __func__, list_empty(&conf->handle_list) ? "empty" : "busy", list_empty(&conf->hold_list) ? "empty" : "busy", atomic_read(&conf->pending_full_writes), conf->bypass_count); if (!list_empty(&conf->handle_list)) { sh = list_entry(conf->handle_list.next, typeof(*sh), lru); if (list_empty(&conf->hold_list)) conf->bypass_count = 0; else if (!test_bit(STRIPE_IO_STARTED, &sh->state)) { if (conf->hold_list.next == conf->last_hold) conf->bypass_count++; else { conf->last_hold = conf->hold_list.next; conf->bypass_count -= conf->bypass_threshold; if (conf->bypass_count < 0) conf->bypass_count = 0; } } } else if (!list_empty(&conf->hold_list) && ((conf->bypass_threshold && conf->bypass_count > conf->bypass_threshold) || atomic_read(&conf->pending_full_writes) == 0)) { sh = list_entry(conf->hold_list.next, typeof(*sh), lru); conf->bypass_count -= conf->bypass_threshold; if (conf->bypass_count < 0) conf->bypass_count = 0; } else return NULL; list_del_init(&sh->lru); atomic_inc(&sh->count); BUG_ON(atomic_read(&sh->count) != 1); return sh; } |
f679623f5
|
3625 |
|
b4fdcb02f
|
3626 |
static void make_request(struct mddev *mddev, struct bio * bi) |
1da177e4c
|
3627 |
{ |
d1688a6d5
|
3628 |
struct r5conf *conf = mddev->private; |
911d4ee85
|
3629 |
int dd_idx; |
1da177e4c
|
3630 3631 3632 |
sector_t new_sector; sector_t logical_sector, last_sector; struct stripe_head *sh; |
a362357b6
|
3633 |
const int rw = bio_data_dir(bi); |
490773268
|
3634 |
int remaining; |
7c13edc87
|
3635 |
int plugged; |
1da177e4c
|
3636 |
|
e9c7469bb
|
3637 3638 |
if (unlikely(bi->bi_rw & REQ_FLUSH)) { md_flush_request(mddev, bi); |
5a7bbad27
|
3639 |
return; |
e5dcdd80a
|
3640 |
} |
3d310eb7b
|
3641 |
md_write_start(mddev, bi); |
06d91a5fe
|
3642 |
|
802ba064c
|
3643 |
if (rw == READ && |
524886151
|
3644 |
mddev->reshape_position == MaxSector && |
21a52c6d0
|
3645 |
chunk_aligned_read(mddev,bi)) |
5a7bbad27
|
3646 |
return; |
524886151
|
3647 |
|
1da177e4c
|
3648 3649 3650 3651 |
logical_sector = bi->bi_sector & ~((sector_t)STRIPE_SECTORS-1); last_sector = bi->bi_sector + (bi->bi_size>>9); bi->bi_next = NULL; bi->bi_phys_segments = 1; /* over-loaded to count active stripes */ |
06d91a5fe
|
3652 |
|
7c13edc87
|
3653 |
plugged = mddev_check_plugged(mddev); |
1da177e4c
|
3654 3655 |
for (;logical_sector < last_sector; logical_sector += STRIPE_SECTORS) { DEFINE_WAIT(w); |
16a53ecc3
|
3656 |
int disks, data_disks; |
b5663ba40
|
3657 |
int previous; |
b578d55fd
|
3658 |
|
7ecaa1e6a
|
3659 |
retry: |
b5663ba40
|
3660 |
previous = 0; |
b0f9ec047
|
3661 |
disks = conf->raid_disks; |
b578d55fd
|
3662 |
prepare_to_wait(&conf->wait_for_overlap, &w, TASK_UNINTERRUPTIBLE); |
b0f9ec047
|
3663 |
if (unlikely(conf->reshape_progress != MaxSector)) { |
fef9c61fd
|
3664 |
/* spinlock is needed as reshape_progress may be |
df8e7f763
|
3665 3666 |
* 64bit on a 32bit platform, and so it might be * possible to see a half-updated value |
aeb878b09
|
3667 |
* Of course reshape_progress could change after |
df8e7f763
|
3668 3669 3670 3671 |
* the lock is dropped, so once we get a reference * to the stripe that we think it is, we will have * to check again. */ |
7ecaa1e6a
|
3672 |
spin_lock_irq(&conf->device_lock); |
fef9c61fd
|
3673 3674 3675 |
if (mddev->delta_disks < 0 ? logical_sector < conf->reshape_progress : logical_sector >= conf->reshape_progress) { |
7ecaa1e6a
|
3676 |
disks = conf->previous_raid_disks; |
b5663ba40
|
3677 3678 |
previous = 1; } else { |
fef9c61fd
|
3679 3680 3681 |
if (mddev->delta_disks < 0 ? logical_sector < conf->reshape_safe : logical_sector >= conf->reshape_safe) { |
b578d55fd
|
3682 3683 3684 3685 3686 |
spin_unlock_irq(&conf->device_lock); schedule(); goto retry; } } |
7ecaa1e6a
|
3687 3688 |
spin_unlock_irq(&conf->device_lock); } |
16a53ecc3
|
3689 |
data_disks = disks - conf->max_degraded; |
112bf8970
|
3690 3691 |
new_sector = raid5_compute_sector(conf, logical_sector, previous, |
911d4ee85
|
3692 |
&dd_idx, NULL); |
0c55e0225
|
3693 3694 |
pr_debug("raid456: make_request, sector %llu logical %llu ", |
1da177e4c
|
3695 3696 |
(unsigned long long)new_sector, (unsigned long long)logical_sector); |
b5663ba40
|
3697 |
sh = get_active_stripe(conf, new_sector, previous, |
a8c906ca3
|
3698 |
(bi->bi_rw&RWA_MASK), 0); |
1da177e4c
|
3699 |
if (sh) { |
b0f9ec047
|
3700 |
if (unlikely(previous)) { |
7ecaa1e6a
|
3701 |
/* expansion might have moved on while waiting for a |
df8e7f763
|
3702 3703 3704 3705 3706 3707 |
* stripe, so we must do the range check again. * Expansion could still move past after this * test, but as we are holding a reference to * 'sh', we know that if that happens, * STRIPE_EXPANDING will get set and the expansion * won't proceed until we finish with the stripe. |
7ecaa1e6a
|
3708 3709 3710 |
*/ int must_retry = 0; spin_lock_irq(&conf->device_lock); |
b0f9ec047
|
3711 3712 3713 |
if (mddev->delta_disks < 0 ? logical_sector >= conf->reshape_progress : logical_sector < conf->reshape_progress) |
7ecaa1e6a
|
3714 3715 3716 3717 3718 |
/* mismatch, need to try again */ must_retry = 1; spin_unlock_irq(&conf->device_lock); if (must_retry) { release_stripe(sh); |
7a3ab9089
|
3719 |
schedule(); |
7ecaa1e6a
|
3720 3721 3722 |
goto retry; } } |
e62e58a5f
|
3723 |
|
ffd96e35c
|
3724 |
if (rw == WRITE && |
a5c308d4d
|
3725 |
logical_sector >= mddev->suspend_lo && |
e464eafdb
|
3726 3727 |
logical_sector < mddev->suspend_hi) { release_stripe(sh); |
e62e58a5f
|
3728 3729 3730 3731 3732 3733 3734 3735 3736 3737 |
/* As the suspend_* range is controlled by * userspace, we want an interruptible * wait. */ flush_signals(current); prepare_to_wait(&conf->wait_for_overlap, &w, TASK_INTERRUPTIBLE); if (logical_sector >= mddev->suspend_lo && logical_sector < mddev->suspend_hi) schedule(); |
e464eafdb
|
3738 3739 |
goto retry; } |
7ecaa1e6a
|
3740 3741 |
if (test_bit(STRIPE_EXPANDING, &sh->state) || |
ffd96e35c
|
3742 |
!add_stripe_bio(sh, bi, dd_idx, rw)) { |
7ecaa1e6a
|
3743 3744 |
/* Stripe is busy expanding or * add failed due to overlap. Flush everything |
1da177e4c
|
3745 3746 |
* and wait a while */ |
482c08349
|
3747 |
md_wakeup_thread(mddev->thread); |
1da177e4c
|
3748 3749 3750 3751 3752 |
release_stripe(sh); schedule(); goto retry; } finish_wait(&conf->wait_for_overlap, &w); |
6ed3003c1
|
3753 3754 |
set_bit(STRIPE_HANDLE, &sh->state); clear_bit(STRIPE_DELAYED, &sh->state); |
e9c7469bb
|
3755 |
if ((bi->bi_rw & REQ_SYNC) && |
729a18663
|
3756 3757 |
!test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) atomic_inc(&conf->preread_active_stripes); |
1da177e4c
|
3758 |
release_stripe(sh); |
1da177e4c
|
3759 3760 3761 3762 3763 3764 3765 3766 |
} else { /* cannot get stripe for read-ahead, just give-up */ clear_bit(BIO_UPTODATE, &bi->bi_flags); finish_wait(&conf->wait_for_overlap, &w); break; } } |
7c13edc87
|
3767 3768 |
if (!plugged) md_wakeup_thread(mddev->thread); |
1da177e4c
|
3769 |
spin_lock_irq(&conf->device_lock); |
960e739d9
|
3770 |
remaining = raid5_dec_bi_phys_segments(bi); |
f6344757a
|
3771 3772 |
spin_unlock_irq(&conf->device_lock); if (remaining == 0) { |
1da177e4c
|
3773 |
|
16a53ecc3
|
3774 |
if ( rw == WRITE ) |
1da177e4c
|
3775 |
md_write_end(mddev); |
6712ecf8f
|
3776 |
|
0e13fe23a
|
3777 |
bio_endio(bi, 0); |
1da177e4c
|
3778 |
} |
1da177e4c
|
3779 |
} |
fd01b88c7
|
3780 |
static sector_t raid5_size(struct mddev *mddev, sector_t sectors, int raid_disks); |
b522adcde
|
3781 |
|
fd01b88c7
|
3782 |
static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr, int *skipped) |
1da177e4c
|
3783 |
{ |
52c03291a
|
3784 3785 3786 3787 3788 3789 3790 3791 3792 |
/* reshaping is quite different to recovery/resync so it is * handled quite separately ... here. * * On each call to sync_request, we gather one chunk worth of * destination stripes and flag them as expanding. * Then we find all the source stripes and request reads. * As the reads complete, handle_stripe will copy the data * into the destination stripe and release that stripe. */ |
d1688a6d5
|
3793 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
3794 |
struct stripe_head *sh; |
ccfcc3c10
|
3795 |
sector_t first_sector, last_sector; |
f416885ef
|
3796 3797 3798 |
int raid_disks = conf->previous_raid_disks; int data_disks = raid_disks - conf->max_degraded; int new_data_disks = conf->raid_disks - conf->max_degraded; |
52c03291a
|
3799 3800 |
int i; int dd_idx; |
c8f517c44
|
3801 |
sector_t writepos, readpos, safepos; |
ec32a2bd3
|
3802 |
sector_t stripe_addr; |
7a6613810
|
3803 |
int reshape_sectors; |
ab69ae12c
|
3804 |
struct list_head stripes; |
52c03291a
|
3805 |
|
fef9c61fd
|
3806 3807 3808 3809 3810 3811 |
if (sector_nr == 0) { /* If restarting in the middle, skip the initial sectors */ if (mddev->delta_disks < 0 && conf->reshape_progress < raid5_size(mddev, 0, 0)) { sector_nr = raid5_size(mddev, 0, 0) - conf->reshape_progress; |
a639755cf
|
3812 |
} else if (mddev->delta_disks >= 0 && |
fef9c61fd
|
3813 3814 |
conf->reshape_progress > 0) sector_nr = conf->reshape_progress; |
f416885ef
|
3815 |
sector_div(sector_nr, new_data_disks); |
fef9c61fd
|
3816 |
if (sector_nr) { |
8dee72114
|
3817 3818 |
mddev->curr_resync_completed = sector_nr; sysfs_notify(&mddev->kobj, NULL, "sync_completed"); |
fef9c61fd
|
3819 3820 3821 |
*skipped = 1; return sector_nr; } |
52c03291a
|
3822 |
} |
7a6613810
|
3823 3824 3825 3826 |
/* We need to process a full chunk at a time. * If old and new chunk sizes differ, we need to process the * largest of these */ |
664e7c413
|
3827 3828 |
if (mddev->new_chunk_sectors > mddev->chunk_sectors) reshape_sectors = mddev->new_chunk_sectors; |
7a6613810
|
3829 |
else |
9d8f03636
|
3830 |
reshape_sectors = mddev->chunk_sectors; |
7a6613810
|
3831 |
|
52c03291a
|
3832 3833 3834 3835 3836 |
/* we update the metadata when there is more than 3Meg * in the block range (that is rather arbitrary, should * probably be time based) or when the data about to be * copied would over-write the source of the data at * the front of the range. |
fef9c61fd
|
3837 3838 |
* i.e. one new_stripe along from reshape_progress new_maps * to after where reshape_safe old_maps to |
52c03291a
|
3839 |
*/ |
fef9c61fd
|
3840 |
writepos = conf->reshape_progress; |
f416885ef
|
3841 |
sector_div(writepos, new_data_disks); |
c8f517c44
|
3842 3843 |
readpos = conf->reshape_progress; sector_div(readpos, data_disks); |
fef9c61fd
|
3844 |
safepos = conf->reshape_safe; |
f416885ef
|
3845 |
sector_div(safepos, data_disks); |
fef9c61fd
|
3846 |
if (mddev->delta_disks < 0) { |
ed37d83e6
|
3847 |
writepos -= min_t(sector_t, reshape_sectors, writepos); |
c8f517c44
|
3848 |
readpos += reshape_sectors; |
7a6613810
|
3849 |
safepos += reshape_sectors; |
fef9c61fd
|
3850 |
} else { |
7a6613810
|
3851 |
writepos += reshape_sectors; |
ed37d83e6
|
3852 3853 |
readpos -= min_t(sector_t, reshape_sectors, readpos); safepos -= min_t(sector_t, reshape_sectors, safepos); |
fef9c61fd
|
3854 |
} |
52c03291a
|
3855 |
|
c8f517c44
|
3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 |
/* 'writepos' is the most advanced device address we might write. * 'readpos' is the least advanced device address we might read. * 'safepos' is the least address recorded in the metadata as having * been reshaped. * If 'readpos' is behind 'writepos', then there is no way that we can * ensure safety in the face of a crash - that must be done by userspace * making a backup of the data. So in that case there is no particular * rush to update metadata. * Otherwise if 'safepos' is behind 'writepos', then we really need to * update the metadata to advance 'safepos' to match 'readpos' so that * we can be safe in the event of a crash. * So we insist on updating metadata if safepos is behind writepos and * readpos is beyond writepos. * In any case, update the metadata every 10 seconds. * Maybe that number should be configurable, but I'm not sure it is * worth it.... maybe it could be a multiple of safemode_delay??? */ |
fef9c61fd
|
3873 |
if ((mddev->delta_disks < 0 |
c8f517c44
|
3874 3875 3876 |
? (safepos > writepos && readpos < writepos) : (safepos < writepos && readpos > writepos)) || time_after(jiffies, conf->reshape_checkpoint + 10*HZ)) { |
52c03291a
|
3877 3878 3879 |
/* Cannot proceed until we've updated the superblock... */ wait_event(conf->wait_for_overlap, atomic_read(&conf->reshape_stripes)==0); |
fef9c61fd
|
3880 |
mddev->reshape_position = conf->reshape_progress; |
75d3da43c
|
3881 |
mddev->curr_resync_completed = sector_nr; |
c8f517c44
|
3882 |
conf->reshape_checkpoint = jiffies; |
850b2b420
|
3883 |
set_bit(MD_CHANGE_DEVS, &mddev->flags); |
52c03291a
|
3884 |
md_wakeup_thread(mddev->thread); |
850b2b420
|
3885 |
wait_event(mddev->sb_wait, mddev->flags == 0 || |
52c03291a
|
3886 3887 |
kthread_should_stop()); spin_lock_irq(&conf->device_lock); |
fef9c61fd
|
3888 |
conf->reshape_safe = mddev->reshape_position; |
52c03291a
|
3889 3890 |
spin_unlock_irq(&conf->device_lock); wake_up(&conf->wait_for_overlap); |
acb180b0e
|
3891 |
sysfs_notify(&mddev->kobj, NULL, "sync_completed"); |
52c03291a
|
3892 |
} |
ec32a2bd3
|
3893 3894 3895 3896 |
if (mddev->delta_disks < 0) { BUG_ON(conf->reshape_progress == 0); stripe_addr = writepos; BUG_ON((mddev->dev_sectors & |
7a6613810
|
3897 3898 |
~((sector_t)reshape_sectors - 1)) - reshape_sectors - stripe_addr |
ec32a2bd3
|
3899 3900 |
!= sector_nr); } else { |
7a6613810
|
3901 |
BUG_ON(writepos != sector_nr + reshape_sectors); |
ec32a2bd3
|
3902 3903 |
stripe_addr = sector_nr; } |
ab69ae12c
|
3904 |
INIT_LIST_HEAD(&stripes); |
7a6613810
|
3905 |
for (i = 0; i < reshape_sectors; i += STRIPE_SECTORS) { |
52c03291a
|
3906 |
int j; |
a9f326ebf
|
3907 |
int skipped_disk = 0; |
a8c906ca3
|
3908 |
sh = get_active_stripe(conf, stripe_addr+i, 0, 0, 1); |
52c03291a
|
3909 3910 3911 3912 3913 3914 3915 3916 3917 |
set_bit(STRIPE_EXPANDING, &sh->state); atomic_inc(&conf->reshape_stripes); /* If any of this stripe is beyond the end of the old * array, then we need to zero those blocks */ for (j=sh->disks; j--;) { sector_t s; if (j == sh->pd_idx) continue; |
f416885ef
|
3918 |
if (conf->level == 6 && |
d0dabf7e5
|
3919 |
j == sh->qd_idx) |
f416885ef
|
3920 |
continue; |
784052ecc
|
3921 |
s = compute_blocknr(sh, j, 0); |
b522adcde
|
3922 |
if (s < raid5_size(mddev, 0, 0)) { |
a9f326ebf
|
3923 |
skipped_disk = 1; |
52c03291a
|
3924 3925 3926 3927 3928 3929 |
continue; } memset(page_address(sh->dev[j].page), 0, STRIPE_SIZE); set_bit(R5_Expanded, &sh->dev[j].flags); set_bit(R5_UPTODATE, &sh->dev[j].flags); } |
a9f326ebf
|
3930 |
if (!skipped_disk) { |
52c03291a
|
3931 3932 3933 |
set_bit(STRIPE_EXPAND_READY, &sh->state); set_bit(STRIPE_HANDLE, &sh->state); } |
ab69ae12c
|
3934 |
list_add(&sh->lru, &stripes); |
52c03291a
|
3935 3936 |
} spin_lock_irq(&conf->device_lock); |
fef9c61fd
|
3937 |
if (mddev->delta_disks < 0) |
7a6613810
|
3938 |
conf->reshape_progress -= reshape_sectors * new_data_disks; |
fef9c61fd
|
3939 |
else |
7a6613810
|
3940 |
conf->reshape_progress += reshape_sectors * new_data_disks; |
52c03291a
|
3941 3942 3943 3944 3945 3946 |
spin_unlock_irq(&conf->device_lock); /* Ok, those stripe are ready. We can start scheduling * reads on the source stripes. * The source stripes are determined by mapping the first and last * block on the destination stripes. */ |
52c03291a
|
3947 |
first_sector = |
ec32a2bd3
|
3948 |
raid5_compute_sector(conf, stripe_addr*(new_data_disks), |
911d4ee85
|
3949 |
1, &dd_idx, NULL); |
52c03291a
|
3950 |
last_sector = |
0e6e0271a
|
3951 |
raid5_compute_sector(conf, ((stripe_addr+reshape_sectors) |
09c9e5fa1
|
3952 |
* new_data_disks - 1), |
911d4ee85
|
3953 |
1, &dd_idx, NULL); |
58c0fed40
|
3954 3955 |
if (last_sector >= mddev->dev_sectors) last_sector = mddev->dev_sectors - 1; |
52c03291a
|
3956 |
while (first_sector <= last_sector) { |
a8c906ca3
|
3957 |
sh = get_active_stripe(conf, first_sector, 1, 0, 1); |
52c03291a
|
3958 3959 3960 3961 3962 |
set_bit(STRIPE_EXPAND_SOURCE, &sh->state); set_bit(STRIPE_HANDLE, &sh->state); release_stripe(sh); first_sector += STRIPE_SECTORS; } |
ab69ae12c
|
3963 3964 3965 3966 3967 3968 3969 3970 |
/* Now that the sources are clearly marked, we can release * the destination stripes */ while (!list_empty(&stripes)) { sh = list_entry(stripes.next, struct stripe_head, lru); list_del_init(&sh->lru); release_stripe(sh); } |
c62072777
|
3971 3972 3973 |
/* If this takes us to the resync_max point where we have to pause, * then we need to write out the superblock. */ |
7a6613810
|
3974 |
sector_nr += reshape_sectors; |
c03f6a196
|
3975 3976 |
if ((sector_nr - mddev->curr_resync_completed) * 2 >= mddev->resync_max - mddev->curr_resync_completed) { |
c62072777
|
3977 3978 3979 |
/* Cannot proceed until we've updated the superblock... */ wait_event(conf->wait_for_overlap, atomic_read(&conf->reshape_stripes) == 0); |
fef9c61fd
|
3980 |
mddev->reshape_position = conf->reshape_progress; |
75d3da43c
|
3981 |
mddev->curr_resync_completed = sector_nr; |
c8f517c44
|
3982 |
conf->reshape_checkpoint = jiffies; |
c62072777
|
3983 3984 3985 3986 3987 3988 |
set_bit(MD_CHANGE_DEVS, &mddev->flags); md_wakeup_thread(mddev->thread); wait_event(mddev->sb_wait, !test_bit(MD_CHANGE_DEVS, &mddev->flags) || kthread_should_stop()); spin_lock_irq(&conf->device_lock); |
fef9c61fd
|
3989 |
conf->reshape_safe = mddev->reshape_position; |
c62072777
|
3990 3991 |
spin_unlock_irq(&conf->device_lock); wake_up(&conf->wait_for_overlap); |
acb180b0e
|
3992 |
sysfs_notify(&mddev->kobj, NULL, "sync_completed"); |
c62072777
|
3993 |
} |
7a6613810
|
3994 |
return reshape_sectors; |
52c03291a
|
3995 3996 3997 |
} /* FIXME go_faster isn't used */ |
fd01b88c7
|
3998 |
static inline sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipped, int go_faster) |
52c03291a
|
3999 |
{ |
d1688a6d5
|
4000 |
struct r5conf *conf = mddev->private; |
52c03291a
|
4001 |
struct stripe_head *sh; |
58c0fed40
|
4002 |
sector_t max_sector = mddev->dev_sectors; |
57dab0bdf
|
4003 |
sector_t sync_blocks; |
16a53ecc3
|
4004 4005 |
int still_degraded = 0; int i; |
1da177e4c
|
4006 |
|
72626685d
|
4007 |
if (sector_nr >= max_sector) { |
1da177e4c
|
4008 |
/* just being told to finish up .. nothing much to do */ |
cea9c2280
|
4009 |
|
292695531
|
4010 4011 4012 4013 |
if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) { end_reshape(conf); return 0; } |
72626685d
|
4014 4015 4016 4017 |
if (mddev->curr_resync < max_sector) /* aborted */ bitmap_end_sync(mddev->bitmap, mddev->curr_resync, &sync_blocks, 1); |
16a53ecc3
|
4018 |
else /* completed sync */ |
72626685d
|
4019 4020 |
conf->fullsync = 0; bitmap_close_sync(mddev->bitmap); |
1da177e4c
|
4021 4022 |
return 0; } |
ccfcc3c10
|
4023 |
|
64bd660b5
|
4024 4025 |
/* Allow raid5_quiesce to complete */ wait_event(conf->wait_for_overlap, conf->quiesce != 2); |
52c03291a
|
4026 4027 |
if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) return reshape_request(mddev, sector_nr, skipped); |
f67055780
|
4028 |
|
c62072777
|
4029 4030 4031 4032 4033 |
/* No need to check resync_max as we never do more than one * stripe, and as resync_max will always be on a chunk boundary, * if the check in md_do_sync didn't fire, there is no chance * of overstepping resync_max here */ |
16a53ecc3
|
4034 |
/* if there is too many failed drives and we are trying |
1da177e4c
|
4035 4036 4037 |
* to resync, then assert that we are finished, because there is * nothing we can do. */ |
3285edf15
|
4038 |
if (mddev->degraded >= conf->max_degraded && |
16a53ecc3
|
4039 |
test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) { |
58c0fed40
|
4040 |
sector_t rv = mddev->dev_sectors - sector_nr; |
57afd89f9
|
4041 |
*skipped = 1; |
1da177e4c
|
4042 4043 |
return rv; } |
72626685d
|
4044 |
if (!bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, 1) && |
3855ad9f3
|
4045 |
!test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) && |
72626685d
|
4046 4047 4048 4049 4050 4051 |
!conf->fullsync && sync_blocks >= STRIPE_SECTORS) { /* we can skip this block, and probably more */ sync_blocks /= STRIPE_SECTORS; *skipped = 1; return sync_blocks * STRIPE_SECTORS; /* keep things rounded to whole stripes */ } |
1da177e4c
|
4052 |
|
b47490c9b
|
4053 4054 |
bitmap_cond_end_sync(mddev->bitmap, sector_nr); |
a8c906ca3
|
4055 |
sh = get_active_stripe(conf, sector_nr, 0, 1, 0); |
1da177e4c
|
4056 |
if (sh == NULL) { |
a8c906ca3
|
4057 |
sh = get_active_stripe(conf, sector_nr, 0, 0, 0); |
1da177e4c
|
4058 |
/* make sure we don't swamp the stripe cache if someone else |
16a53ecc3
|
4059 |
* is trying to get access |
1da177e4c
|
4060 |
*/ |
66c006a55
|
4061 |
schedule_timeout_uninterruptible(1); |
1da177e4c
|
4062 |
} |
16a53ecc3
|
4063 4064 4065 4066 |
/* Need to check if array will still be degraded after recovery/resync * We don't need to check the 'failed' flag as when that gets set, * recovery aborts. */ |
f001a70cd
|
4067 |
for (i = 0; i < conf->raid_disks; i++) |
16a53ecc3
|
4068 4069 4070 4071 |
if (conf->disks[i].rdev == NULL) still_degraded = 1; bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, still_degraded); |
83206d66b
|
4072 |
set_bit(STRIPE_SYNC_REQUESTED, &sh->state); |
1da177e4c
|
4073 |
|
1442577bf
|
4074 |
handle_stripe(sh); |
1da177e4c
|
4075 4076 4077 4078 |
release_stripe(sh); return STRIPE_SECTORS; } |
d1688a6d5
|
4079 |
static int retry_aligned_read(struct r5conf *conf, struct bio *raid_bio) |
46031f9a3
|
4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 |
{ /* We may not be able to submit a whole bio at once as there * may not be enough stripe_heads available. * We cannot pre-allocate enough stripe_heads as we may need * more than exist in the cache (if we allow ever large chunks). * So we do one stripe head at a time and record in * ->bi_hw_segments how many have been done. * * We *know* that this entire raid_bio is in one chunk, so * it will be only one 'dd_idx' and only need one call to raid5_compute_sector. */ struct stripe_head *sh; |
911d4ee85
|
4092 |
int dd_idx; |
46031f9a3
|
4093 4094 4095 4096 4097 4098 |
sector_t sector, logical_sector, last_sector; int scnt = 0; int remaining; int handled = 0; logical_sector = raid_bio->bi_sector & ~((sector_t)STRIPE_SECTORS-1); |
112bf8970
|
4099 |
sector = raid5_compute_sector(conf, logical_sector, |
911d4ee85
|
4100 |
0, &dd_idx, NULL); |
46031f9a3
|
4101 4102 4103 |
last_sector = raid_bio->bi_sector + (raid_bio->bi_size>>9); for (; logical_sector < last_sector; |
387bb1737
|
4104 4105 4106 |
logical_sector += STRIPE_SECTORS, sector += STRIPE_SECTORS, scnt++) { |
46031f9a3
|
4107 |
|
960e739d9
|
4108 |
if (scnt < raid5_bi_hw_segments(raid_bio)) |
46031f9a3
|
4109 4110 |
/* already done this stripe */ continue; |
a8c906ca3
|
4111 |
sh = get_active_stripe(conf, sector, 0, 1, 0); |
46031f9a3
|
4112 4113 4114 |
if (!sh) { /* failed to get a stripe - must wait */ |
960e739d9
|
4115 |
raid5_set_bi_hw_segments(raid_bio, scnt); |
46031f9a3
|
4116 4117 4118 4119 4120 |
conf->retry_read_aligned = raid_bio; return handled; } set_bit(R5_ReadError, &sh->dev[dd_idx].flags); |
387bb1737
|
4121 4122 |
if (!add_stripe_bio(sh, raid_bio, dd_idx, 0)) { release_stripe(sh); |
960e739d9
|
4123 |
raid5_set_bi_hw_segments(raid_bio, scnt); |
387bb1737
|
4124 4125 4126 |
conf->retry_read_aligned = raid_bio; return handled; } |
36d1c6476
|
4127 |
handle_stripe(sh); |
46031f9a3
|
4128 4129 4130 4131 |
release_stripe(sh); handled++; } spin_lock_irq(&conf->device_lock); |
960e739d9
|
4132 |
remaining = raid5_dec_bi_phys_segments(raid_bio); |
46031f9a3
|
4133 |
spin_unlock_irq(&conf->device_lock); |
0e13fe23a
|
4134 4135 |
if (remaining == 0) bio_endio(raid_bio, 0); |
46031f9a3
|
4136 4137 4138 4139 |
if (atomic_dec_and_test(&conf->active_aligned_reads)) wake_up(&conf->wait_for_stripe); return handled; } |
46031f9a3
|
4140 |
|
1da177e4c
|
4141 4142 4143 4144 4145 4146 4147 |
/* * This is our raid5 kernel thread. * * We scan the hash table for stripes which can be handled now. * During the scan, completed stripes are saved for us by the interrupt * handler, so that they will not have to wait for our next wakeup. */ |
fd01b88c7
|
4148 |
static void raid5d(struct mddev *mddev) |
1da177e4c
|
4149 4150 |
{ struct stripe_head *sh; |
d1688a6d5
|
4151 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
4152 |
int handled; |
e1dfa0a29
|
4153 |
struct blk_plug plug; |
1da177e4c
|
4154 |
|
45b4233ca
|
4155 4156 |
pr_debug("+++ raid5d active "); |
1da177e4c
|
4157 4158 |
md_check_recovery(mddev); |
1da177e4c
|
4159 |
|
e1dfa0a29
|
4160 |
blk_start_plug(&plug); |
1da177e4c
|
4161 4162 4163 |
handled = 0; spin_lock_irq(&conf->device_lock); while (1) { |
46031f9a3
|
4164 |
struct bio *bio; |
1da177e4c
|
4165 |
|
7c13edc87
|
4166 4167 4168 4169 |
if (atomic_read(&mddev->plug_cnt) == 0 && !list_empty(&conf->bitmap_list)) { /* Now is a good time to flush some bitmap updates */ conf->seq_flush++; |
700e432d8
|
4170 |
spin_unlock_irq(&conf->device_lock); |
72626685d
|
4171 |
bitmap_unplug(mddev->bitmap); |
700e432d8
|
4172 |
spin_lock_irq(&conf->device_lock); |
7c13edc87
|
4173 |
conf->seq_write = conf->seq_flush; |
72626685d
|
4174 4175 |
activate_bit_delay(conf); } |
7c13edc87
|
4176 4177 |
if (atomic_read(&mddev->plug_cnt) == 0) raid5_activate_delayed(conf); |
72626685d
|
4178 |
|
46031f9a3
|
4179 4180 4181 4182 4183 4184 4185 4186 4187 |
while ((bio = remove_bio_from_retry(conf))) { int ok; spin_unlock_irq(&conf->device_lock); ok = retry_aligned_read(conf, bio); spin_lock_irq(&conf->device_lock); if (!ok) break; handled++; } |
8b3e6cdc5
|
4188 |
sh = __get_priority_stripe(conf); |
c9f21aaff
|
4189 |
if (!sh) |
1da177e4c
|
4190 |
break; |
1da177e4c
|
4191 4192 4193 |
spin_unlock_irq(&conf->device_lock); handled++; |
417b8d4ac
|
4194 4195 4196 |
handle_stripe(sh); release_stripe(sh); cond_resched(); |
1da177e4c
|
4197 |
|
de393cdea
|
4198 4199 |
if (mddev->flags & ~(1<<MD_CHANGE_PENDING)) md_check_recovery(mddev); |
1da177e4c
|
4200 4201 |
spin_lock_irq(&conf->device_lock); } |
45b4233ca
|
4202 4203 |
pr_debug("%d stripes handled ", handled); |
1da177e4c
|
4204 4205 |
spin_unlock_irq(&conf->device_lock); |
c9f21aaff
|
4206 |
async_tx_issue_pending_all(); |
e1dfa0a29
|
4207 |
blk_finish_plug(&plug); |
1da177e4c
|
4208 |
|
45b4233ca
|
4209 4210 |
pr_debug("--- raid5d inactive "); |
1da177e4c
|
4211 |
} |
3f294f4fb
|
4212 |
static ssize_t |
fd01b88c7
|
4213 |
raid5_show_stripe_cache_size(struct mddev *mddev, char *page) |
3f294f4fb
|
4214 |
{ |
d1688a6d5
|
4215 |
struct r5conf *conf = mddev->private; |
96de1e663
|
4216 4217 4218 4219 4220 |
if (conf) return sprintf(page, "%d ", conf->max_nr_stripes); else return 0; |
3f294f4fb
|
4221 |
} |
c41d4ac40
|
4222 |
int |
fd01b88c7
|
4223 |
raid5_set_cache_size(struct mddev *mddev, int size) |
3f294f4fb
|
4224 |
{ |
d1688a6d5
|
4225 |
struct r5conf *conf = mddev->private; |
b5470dc5f
|
4226 |
int err; |
c41d4ac40
|
4227 |
if (size <= 16 || size > 32768) |
3f294f4fb
|
4228 |
return -EINVAL; |
c41d4ac40
|
4229 |
while (size < conf->max_nr_stripes) { |
3f294f4fb
|
4230 4231 4232 4233 4234 |
if (drop_one_stripe(conf)) conf->max_nr_stripes--; else break; } |
b5470dc5f
|
4235 4236 4237 |
err = md_allow_write(mddev); if (err) return err; |
c41d4ac40
|
4238 |
while (size > conf->max_nr_stripes) { |
3f294f4fb
|
4239 4240 4241 4242 |
if (grow_one_stripe(conf)) conf->max_nr_stripes++; else break; } |
c41d4ac40
|
4243 4244 4245 4246 4247 |
return 0; } EXPORT_SYMBOL(raid5_set_cache_size); static ssize_t |
fd01b88c7
|
4248 |
raid5_store_stripe_cache_size(struct mddev *mddev, const char *page, size_t len) |
c41d4ac40
|
4249 |
{ |
d1688a6d5
|
4250 |
struct r5conf *conf = mddev->private; |
c41d4ac40
|
4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 |
unsigned long new; int err; if (len >= PAGE_SIZE) return -EINVAL; if (!conf) return -ENODEV; if (strict_strtoul(page, 10, &new)) return -EINVAL; err = raid5_set_cache_size(mddev, new); if (err) return err; |
3f294f4fb
|
4264 4265 |
return len; } |
007583c92
|
4266 |
|
96de1e663
|
4267 4268 4269 4270 |
static struct md_sysfs_entry raid5_stripecache_size = __ATTR(stripe_cache_size, S_IRUGO | S_IWUSR, raid5_show_stripe_cache_size, raid5_store_stripe_cache_size); |
3f294f4fb
|
4271 4272 |
static ssize_t |
fd01b88c7
|
4273 |
raid5_show_preread_threshold(struct mddev *mddev, char *page) |
8b3e6cdc5
|
4274 |
{ |
d1688a6d5
|
4275 |
struct r5conf *conf = mddev->private; |
8b3e6cdc5
|
4276 4277 4278 4279 4280 4281 4282 4283 |
if (conf) return sprintf(page, "%d ", conf->bypass_threshold); else return 0; } static ssize_t |
fd01b88c7
|
4284 |
raid5_store_preread_threshold(struct mddev *mddev, const char *page, size_t len) |
8b3e6cdc5
|
4285 |
{ |
d1688a6d5
|
4286 |
struct r5conf *conf = mddev->private; |
4ef197d87
|
4287 |
unsigned long new; |
8b3e6cdc5
|
4288 4289 4290 4291 |
if (len >= PAGE_SIZE) return -EINVAL; if (!conf) return -ENODEV; |
4ef197d87
|
4292 |
if (strict_strtoul(page, 10, &new)) |
8b3e6cdc5
|
4293 |
return -EINVAL; |
4ef197d87
|
4294 |
if (new > conf->max_nr_stripes) |
8b3e6cdc5
|
4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 |
return -EINVAL; conf->bypass_threshold = new; return len; } static struct md_sysfs_entry raid5_preread_bypass_threshold = __ATTR(preread_bypass_threshold, S_IRUGO | S_IWUSR, raid5_show_preread_threshold, raid5_store_preread_threshold); static ssize_t |
fd01b88c7
|
4307 |
stripe_cache_active_show(struct mddev *mddev, char *page) |
3f294f4fb
|
4308 |
{ |
d1688a6d5
|
4309 |
struct r5conf *conf = mddev->private; |
96de1e663
|
4310 4311 4312 4313 4314 |
if (conf) return sprintf(page, "%d ", atomic_read(&conf->active_stripes)); else return 0; |
3f294f4fb
|
4315 |
} |
96de1e663
|
4316 4317 |
static struct md_sysfs_entry raid5_stripecache_active = __ATTR_RO(stripe_cache_active); |
3f294f4fb
|
4318 |
|
007583c92
|
4319 |
static struct attribute *raid5_attrs[] = { |
3f294f4fb
|
4320 4321 |
&raid5_stripecache_size.attr, &raid5_stripecache_active.attr, |
8b3e6cdc5
|
4322 |
&raid5_preread_bypass_threshold.attr, |
3f294f4fb
|
4323 4324 |
NULL, }; |
007583c92
|
4325 4326 4327 |
static struct attribute_group raid5_attrs_group = { .name = NULL, .attrs = raid5_attrs, |
3f294f4fb
|
4328 |
}; |
80c3a6ce4
|
4329 |
static sector_t |
fd01b88c7
|
4330 |
raid5_size(struct mddev *mddev, sector_t sectors, int raid_disks) |
80c3a6ce4
|
4331 |
{ |
d1688a6d5
|
4332 |
struct r5conf *conf = mddev->private; |
80c3a6ce4
|
4333 4334 4335 |
if (!sectors) sectors = mddev->dev_sectors; |
5e5e3e78e
|
4336 |
if (!raid_disks) |
7ec054783
|
4337 |
/* size is defined by the smallest of previous and new size */ |
5e5e3e78e
|
4338 |
raid_disks = min(conf->raid_disks, conf->previous_raid_disks); |
80c3a6ce4
|
4339 |
|
9d8f03636
|
4340 |
sectors &= ~((sector_t)mddev->chunk_sectors - 1); |
664e7c413
|
4341 |
sectors &= ~((sector_t)mddev->new_chunk_sectors - 1); |
80c3a6ce4
|
4342 4343 |
return sectors * (raid_disks - conf->max_degraded); } |
d1688a6d5
|
4344 |
static void raid5_free_percpu(struct r5conf *conf) |
36d1c6476
|
4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 |
{ struct raid5_percpu *percpu; unsigned long cpu; if (!conf->percpu) return; get_online_cpus(); for_each_possible_cpu(cpu) { percpu = per_cpu_ptr(conf->percpu, cpu); safe_put_page(percpu->spare_page); |
d6f38f31f
|
4356 |
kfree(percpu->scribble); |
36d1c6476
|
4357 4358 4359 4360 4361 4362 4363 4364 |
} #ifdef CONFIG_HOTPLUG_CPU unregister_cpu_notifier(&conf->cpu_notify); #endif put_online_cpus(); free_percpu(conf->percpu); } |
d1688a6d5
|
4365 |
static void free_conf(struct r5conf *conf) |
95fc17aac
|
4366 4367 |
{ shrink_stripes(conf); |
36d1c6476
|
4368 |
raid5_free_percpu(conf); |
95fc17aac
|
4369 4370 4371 4372 |
kfree(conf->disks); kfree(conf->stripe_hashtbl); kfree(conf); } |
36d1c6476
|
4373 4374 4375 4376 |
#ifdef CONFIG_HOTPLUG_CPU static int raid456_cpu_notify(struct notifier_block *nfb, unsigned long action, void *hcpu) { |
d1688a6d5
|
4377 |
struct r5conf *conf = container_of(nfb, struct r5conf, cpu_notify); |
36d1c6476
|
4378 4379 4380 4381 4382 4383 |
long cpu = (long)hcpu; struct raid5_percpu *percpu = per_cpu_ptr(conf->percpu, cpu); switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: |
d6f38f31f
|
4384 |
if (conf->level == 6 && !percpu->spare_page) |
36d1c6476
|
4385 |
percpu->spare_page = alloc_page(GFP_KERNEL); |
d6f38f31f
|
4386 4387 4388 4389 4390 4391 4392 |
if (!percpu->scribble) percpu->scribble = kmalloc(conf->scribble_len, GFP_KERNEL); if (!percpu->scribble || (conf->level == 6 && !percpu->spare_page)) { safe_put_page(percpu->spare_page); kfree(percpu->scribble); |
36d1c6476
|
4393 4394 4395 |
pr_err("%s: failed memory allocation for cpu%ld ", __func__, cpu); |
55af6bb50
|
4396 |
return notifier_from_errno(-ENOMEM); |
36d1c6476
|
4397 4398 4399 4400 4401 |
} break; case CPU_DEAD: case CPU_DEAD_FROZEN: safe_put_page(percpu->spare_page); |
d6f38f31f
|
4402 |
kfree(percpu->scribble); |
36d1c6476
|
4403 |
percpu->spare_page = NULL; |
d6f38f31f
|
4404 |
percpu->scribble = NULL; |
36d1c6476
|
4405 4406 4407 4408 4409 4410 4411 |
break; default: break; } return NOTIFY_OK; } #endif |
d1688a6d5
|
4412 |
static int raid5_alloc_percpu(struct r5conf *conf) |
36d1c6476
|
4413 4414 4415 |
{ unsigned long cpu; struct page *spare_page; |
a29d8b8e2
|
4416 |
struct raid5_percpu __percpu *allcpus; |
d6f38f31f
|
4417 |
void *scribble; |
36d1c6476
|
4418 |
int err; |
36d1c6476
|
4419 4420 4421 4422 4423 4424 4425 4426 |
allcpus = alloc_percpu(struct raid5_percpu); if (!allcpus) return -ENOMEM; conf->percpu = allcpus; get_online_cpus(); err = 0; for_each_present_cpu(cpu) { |
d6f38f31f
|
4427 4428 4429 4430 4431 4432 4433 4434 |
if (conf->level == 6) { spare_page = alloc_page(GFP_KERNEL); if (!spare_page) { err = -ENOMEM; break; } per_cpu_ptr(conf->percpu, cpu)->spare_page = spare_page; } |
5e5e3e78e
|
4435 |
scribble = kmalloc(conf->scribble_len, GFP_KERNEL); |
d6f38f31f
|
4436 |
if (!scribble) { |
36d1c6476
|
4437 4438 4439 |
err = -ENOMEM; break; } |
d6f38f31f
|
4440 |
per_cpu_ptr(conf->percpu, cpu)->scribble = scribble; |
36d1c6476
|
4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 |
} #ifdef CONFIG_HOTPLUG_CPU conf->cpu_notify.notifier_call = raid456_cpu_notify; conf->cpu_notify.priority = 0; if (err == 0) err = register_cpu_notifier(&conf->cpu_notify); #endif put_online_cpus(); return err; } |
d1688a6d5
|
4452 |
static struct r5conf *setup_conf(struct mddev *mddev) |
1da177e4c
|
4453 |
{ |
d1688a6d5
|
4454 |
struct r5conf *conf; |
5e5e3e78e
|
4455 |
int raid_disk, memory, max_disks; |
3cb030020
|
4456 |
struct md_rdev *rdev; |
1da177e4c
|
4457 |
struct disk_info *disk; |
1da177e4c
|
4458 |
|
91adb5647
|
4459 4460 4461 |
if (mddev->new_level != 5 && mddev->new_level != 4 && mddev->new_level != 6) { |
0c55e0225
|
4462 4463 |
printk(KERN_ERR "md/raid:%s: raid level not set to 4/5/6 (%d) ", |
91adb5647
|
4464 4465 |
mdname(mddev), mddev->new_level); return ERR_PTR(-EIO); |
1da177e4c
|
4466 |
} |
91adb5647
|
4467 4468 4469 4470 |
if ((mddev->new_level == 5 && !algorithm_valid_raid5(mddev->new_layout)) || (mddev->new_level == 6 && !algorithm_valid_raid6(mddev->new_layout))) { |
0c55e0225
|
4471 4472 |
printk(KERN_ERR "md/raid:%s: layout %d not supported ", |
91adb5647
|
4473 4474 |
mdname(mddev), mddev->new_layout); return ERR_PTR(-EIO); |
99c0fb5f9
|
4475 |
} |
91adb5647
|
4476 |
if (mddev->new_level == 6 && mddev->raid_disks < 4) { |
0c55e0225
|
4477 4478 |
printk(KERN_ERR "md/raid:%s: not enough configured devices (%d, minimum 4) ", |
91adb5647
|
4479 4480 |
mdname(mddev), mddev->raid_disks); return ERR_PTR(-EINVAL); |
4bbf3771c
|
4481 |
} |
664e7c413
|
4482 4483 4484 |
if (!mddev->new_chunk_sectors || (mddev->new_chunk_sectors << 9) % PAGE_SIZE || !is_power_of_2(mddev->new_chunk_sectors)) { |
0c55e0225
|
4485 4486 4487 |
printk(KERN_ERR "md/raid:%s: invalid chunk size %d ", mdname(mddev), mddev->new_chunk_sectors << 9); |
91adb5647
|
4488 |
return ERR_PTR(-EINVAL); |
f67055780
|
4489 |
} |
d1688a6d5
|
4490 |
conf = kzalloc(sizeof(struct r5conf), GFP_KERNEL); |
91adb5647
|
4491 |
if (conf == NULL) |
1da177e4c
|
4492 |
goto abort; |
f5efd45ae
|
4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 |
spin_lock_init(&conf->device_lock); init_waitqueue_head(&conf->wait_for_stripe); init_waitqueue_head(&conf->wait_for_overlap); INIT_LIST_HEAD(&conf->handle_list); INIT_LIST_HEAD(&conf->hold_list); INIT_LIST_HEAD(&conf->delayed_list); INIT_LIST_HEAD(&conf->bitmap_list); INIT_LIST_HEAD(&conf->inactive_list); atomic_set(&conf->active_stripes, 0); atomic_set(&conf->preread_active_stripes, 0); atomic_set(&conf->active_aligned_reads, 0); conf->bypass_threshold = BYPASS_THRESHOLD; |
d890fa2b0
|
4505 |
conf->recovery_disabled = mddev->recovery_disabled - 1; |
91adb5647
|
4506 4507 4508 4509 4510 |
conf->raid_disks = mddev->raid_disks; if (mddev->reshape_position == MaxSector) conf->previous_raid_disks = mddev->raid_disks; else |
f67055780
|
4511 |
conf->previous_raid_disks = mddev->raid_disks - mddev->delta_disks; |
5e5e3e78e
|
4512 4513 |
max_disks = max(conf->raid_disks, conf->previous_raid_disks); conf->scribble_len = scribble_len(max_disks); |
f67055780
|
4514 |
|
5e5e3e78e
|
4515 |
conf->disks = kzalloc(max_disks * sizeof(struct disk_info), |
b55e6bfcd
|
4516 4517 4518 |
GFP_KERNEL); if (!conf->disks) goto abort; |
9ffae0cf3
|
4519 |
|
1da177e4c
|
4520 |
conf->mddev = mddev; |
fccddba06
|
4521 |
if ((conf->stripe_hashtbl = kzalloc(PAGE_SIZE, GFP_KERNEL)) == NULL) |
1da177e4c
|
4522 |
goto abort; |
1da177e4c
|
4523 |
|
36d1c6476
|
4524 4525 4526 |
conf->level = mddev->new_level; if (raid5_alloc_percpu(conf) != 0) goto abort; |
0c55e0225
|
4527 4528 |
pr_debug("raid456: run(%s) called. ", mdname(mddev)); |
1da177e4c
|
4529 |
|
159ec1fc0
|
4530 |
list_for_each_entry(rdev, &mddev->disks, same_set) { |
1da177e4c
|
4531 |
raid_disk = rdev->raid_disk; |
5e5e3e78e
|
4532 |
if (raid_disk >= max_disks |
1da177e4c
|
4533 4534 4535 4536 4537 |
|| raid_disk < 0) continue; disk = conf->disks + raid_disk; disk->rdev = rdev; |
b2d444d7a
|
4538 |
if (test_bit(In_sync, &rdev->flags)) { |
1da177e4c
|
4539 |
char b[BDEVNAME_SIZE]; |
0c55e0225
|
4540 4541 4542 4543 |
printk(KERN_INFO "md/raid:%s: device %s operational as raid" " disk %d ", mdname(mddev), bdevname(rdev->bdev, b), raid_disk); |
d6b212f4b
|
4544 |
} else if (rdev->saved_raid_disk != raid_disk) |
8c2e870a6
|
4545 4546 |
/* Cannot rely on bitmap to complete recovery */ conf->fullsync = 1; |
1da177e4c
|
4547 |
} |
09c9e5fa1
|
4548 |
conf->chunk_sectors = mddev->new_chunk_sectors; |
91adb5647
|
4549 |
conf->level = mddev->new_level; |
16a53ecc3
|
4550 4551 4552 4553 |
if (conf->level == 6) conf->max_degraded = 2; else conf->max_degraded = 1; |
91adb5647
|
4554 |
conf->algorithm = mddev->new_layout; |
1da177e4c
|
4555 |
conf->max_nr_stripes = NR_STRIPES; |
fef9c61fd
|
4556 |
conf->reshape_progress = mddev->reshape_position; |
e183eaedd
|
4557 |
if (conf->reshape_progress != MaxSector) { |
09c9e5fa1
|
4558 |
conf->prev_chunk_sectors = mddev->chunk_sectors; |
e183eaedd
|
4559 4560 |
conf->prev_algo = mddev->layout; } |
1da177e4c
|
4561 |
|
91adb5647
|
4562 |
memory = conf->max_nr_stripes * (sizeof(struct stripe_head) + |
5e5e3e78e
|
4563 |
max_disks * ((sizeof(struct bio) + PAGE_SIZE))) / 1024; |
91adb5647
|
4564 4565 |
if (grow_stripes(conf, conf->max_nr_stripes)) { printk(KERN_ERR |
0c55e0225
|
4566 4567 4568 |
"md/raid:%s: couldn't allocate %dkB for buffers ", mdname(mddev), memory); |
91adb5647
|
4569 4570 |
goto abort; } else |
0c55e0225
|
4571 4572 4573 |
printk(KERN_INFO "md/raid:%s: allocated %dkB ", mdname(mddev), memory); |
1da177e4c
|
4574 |
|
0da3c6194
|
4575 |
conf->thread = md_register_thread(raid5d, mddev, NULL); |
91adb5647
|
4576 4577 |
if (!conf->thread) { printk(KERN_ERR |
0c55e0225
|
4578 4579 |
"md/raid:%s: couldn't allocate thread. ", |
91adb5647
|
4580 |
mdname(mddev)); |
16a53ecc3
|
4581 4582 |
goto abort; } |
91adb5647
|
4583 4584 4585 4586 4587 |
return conf; abort: if (conf) { |
95fc17aac
|
4588 |
free_conf(conf); |
91adb5647
|
4589 4590 4591 4592 |
return ERR_PTR(-EIO); } else return ERR_PTR(-ENOMEM); } |
c148ffdcd
|
4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 |
static int only_parity(int raid_disk, int algo, int raid_disks, int max_degraded) { switch (algo) { case ALGORITHM_PARITY_0: if (raid_disk < max_degraded) return 1; break; case ALGORITHM_PARITY_N: if (raid_disk >= raid_disks - max_degraded) return 1; break; case ALGORITHM_PARITY_0_6: if (raid_disk == 0 || raid_disk == raid_disks - 1) return 1; break; case ALGORITHM_LEFT_ASYMMETRIC_6: case ALGORITHM_RIGHT_ASYMMETRIC_6: case ALGORITHM_LEFT_SYMMETRIC_6: case ALGORITHM_RIGHT_SYMMETRIC_6: if (raid_disk == raid_disks - 1) return 1; } return 0; } |
fd01b88c7
|
4619 |
static int run(struct mddev *mddev) |
91adb5647
|
4620 |
{ |
d1688a6d5
|
4621 |
struct r5conf *conf; |
9f7c22200
|
4622 |
int working_disks = 0; |
c148ffdcd
|
4623 |
int dirty_parity_disks = 0; |
3cb030020
|
4624 |
struct md_rdev *rdev; |
c148ffdcd
|
4625 |
sector_t reshape_offset = 0; |
91adb5647
|
4626 |
|
8c6ac868b
|
4627 |
if (mddev->recovery_cp != MaxSector) |
0c55e0225
|
4628 |
printk(KERN_NOTICE "md/raid:%s: not clean" |
8c6ac868b
|
4629 4630 4631 |
" -- starting background reconstruction ", mdname(mddev)); |
91adb5647
|
4632 4633 4634 4635 4636 4637 4638 4639 |
if (mddev->reshape_position != MaxSector) { /* Check that we can continue the reshape. * Currently only disks can change, it must * increase, and we must be past the point where * a stripe over-writes itself */ sector_t here_new, here_old; int old_disks; |
18b003349
|
4640 |
int max_degraded = (mddev->level == 6 ? 2 : 1); |
91adb5647
|
4641 |
|
88ce4930e
|
4642 |
if (mddev->new_level != mddev->level) { |
0c55e0225
|
4643 |
printk(KERN_ERR "md/raid:%s: unsupported reshape " |
91adb5647
|
4644 4645 4646 4647 4648 |
"required - aborting. ", mdname(mddev)); return -EINVAL; } |
91adb5647
|
4649 4650 4651 4652 4653 4654 |
old_disks = mddev->raid_disks - mddev->delta_disks; /* reshape_position must be on a new-stripe boundary, and one * further up in new geometry must map after here in old * geometry. */ here_new = mddev->reshape_position; |
664e7c413
|
4655 |
if (sector_div(here_new, mddev->new_chunk_sectors * |
91adb5647
|
4656 |
(mddev->raid_disks - max_degraded))) { |
0c55e0225
|
4657 4658 4659 |
printk(KERN_ERR "md/raid:%s: reshape_position not " "on a stripe boundary ", mdname(mddev)); |
91adb5647
|
4660 4661 |
return -EINVAL; } |
c148ffdcd
|
4662 |
reshape_offset = here_new * mddev->new_chunk_sectors; |
91adb5647
|
4663 4664 |
/* here_new is the stripe we will write to */ here_old = mddev->reshape_position; |
9d8f03636
|
4665 |
sector_div(here_old, mddev->chunk_sectors * |
91adb5647
|
4666 4667 4668 |
(old_disks-max_degraded)); /* here_old is the first stripe that we might need to read * from */ |
67ac6011d
|
4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 |
if (mddev->delta_disks == 0) { /* We cannot be sure it is safe to start an in-place * reshape. It is only safe if user-space if monitoring * and taking constant backups. * mdadm always starts a situation like this in * readonly mode so it can take control before * allowing any writes. So just check for that. */ if ((here_new * mddev->new_chunk_sectors != here_old * mddev->chunk_sectors) || mddev->ro == 0) { |
0c55e0225
|
4680 4681 4682 4683 |
printk(KERN_ERR "md/raid:%s: in-place reshape must be started" " in read-only mode - aborting ", mdname(mddev)); |
67ac6011d
|
4684 4685 4686 4687 4688 4689 4690 |
return -EINVAL; } } else if (mddev->delta_disks < 0 ? (here_new * mddev->new_chunk_sectors <= here_old * mddev->chunk_sectors) : (here_new * mddev->new_chunk_sectors >= here_old * mddev->chunk_sectors)) { |
91adb5647
|
4691 |
/* Reading from the same stripe as writing to - bad */ |
0c55e0225
|
4692 4693 4694 4695 |
printk(KERN_ERR "md/raid:%s: reshape_position too early for " "auto-recovery - aborting. ", mdname(mddev)); |
91adb5647
|
4696 4697 |
return -EINVAL; } |
0c55e0225
|
4698 4699 4700 |
printk(KERN_INFO "md/raid:%s: reshape will continue ", mdname(mddev)); |
91adb5647
|
4701 4702 4703 4704 |
/* OK, we should be able to continue; */ } else { BUG_ON(mddev->level != mddev->new_level); BUG_ON(mddev->layout != mddev->new_layout); |
664e7c413
|
4705 |
BUG_ON(mddev->chunk_sectors != mddev->new_chunk_sectors); |
91adb5647
|
4706 |
BUG_ON(mddev->delta_disks != 0); |
1da177e4c
|
4707 |
} |
91adb5647
|
4708 |
|
245f46c2c
|
4709 4710 4711 4712 |
if (mddev->private == NULL) conf = setup_conf(mddev); else conf = mddev->private; |
91adb5647
|
4713 4714 4715 4716 4717 4718 4719 4720 4721 4722 |
if (IS_ERR(conf)) return PTR_ERR(conf); mddev->thread = conf->thread; conf->thread = NULL; mddev->private = conf; /* * 0 for a fully functional array, 1 or 2 for a degraded array. */ |
c148ffdcd
|
4723 4724 4725 |
list_for_each_entry(rdev, &mddev->disks, same_set) { if (rdev->raid_disk < 0) continue; |
2f1158824
|
4726 |
if (test_bit(In_sync, &rdev->flags)) { |
91adb5647
|
4727 |
working_disks++; |
2f1158824
|
4728 4729 |
continue; } |
c148ffdcd
|
4730 4731 4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 |
/* This disc is not fully in-sync. However if it * just stored parity (beyond the recovery_offset), * when we don't need to be concerned about the * array being dirty. * When reshape goes 'backwards', we never have * partially completed devices, so we only need * to worry about reshape going forwards. */ /* Hack because v0.91 doesn't store recovery_offset properly. */ if (mddev->major_version == 0 && mddev->minor_version > 90) rdev->recovery_offset = reshape_offset; |
c148ffdcd
|
4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 |
if (rdev->recovery_offset < reshape_offset) { /* We need to check old and new layout */ if (!only_parity(rdev->raid_disk, conf->algorithm, conf->raid_disks, conf->max_degraded)) continue; } if (!only_parity(rdev->raid_disk, conf->prev_algo, conf->previous_raid_disks, conf->max_degraded)) continue; dirty_parity_disks++; } |
91adb5647
|
4758 |
|
5e5e3e78e
|
4759 4760 |
mddev->degraded = (max(conf->raid_disks, conf->previous_raid_disks) - working_disks); |
91adb5647
|
4761 |
|
674806d62
|
4762 |
if (has_failed(conf)) { |
0c55e0225
|
4763 |
printk(KERN_ERR "md/raid:%s: not enough operational devices" |
1da177e4c
|
4764 4765 |
" (%d/%d failed) ", |
02c2de8cc
|
4766 |
mdname(mddev), mddev->degraded, conf->raid_disks); |
1da177e4c
|
4767 4768 |
goto abort; } |
91adb5647
|
4769 |
/* device size must be a multiple of chunk size */ |
9d8f03636
|
4770 |
mddev->dev_sectors &= ~(mddev->chunk_sectors - 1); |
91adb5647
|
4771 |
mddev->resync_max_sectors = mddev->dev_sectors; |
c148ffdcd
|
4772 |
if (mddev->degraded > dirty_parity_disks && |
1da177e4c
|
4773 |
mddev->recovery_cp != MaxSector) { |
6ff8d8ec0
|
4774 4775 |
if (mddev->ok_start_degraded) printk(KERN_WARNING |
0c55e0225
|
4776 4777 4778 |
"md/raid:%s: starting dirty degraded array" " - data corruption possible. ", |
6ff8d8ec0
|
4779 4780 4781 |
mdname(mddev)); else { printk(KERN_ERR |
0c55e0225
|
4782 4783 |
"md/raid:%s: cannot start dirty degraded array. ", |
6ff8d8ec0
|
4784 4785 4786 |
mdname(mddev)); goto abort; } |
1da177e4c
|
4787 |
} |
1da177e4c
|
4788 |
if (mddev->degraded == 0) |
0c55e0225
|
4789 4790 4791 |
printk(KERN_INFO "md/raid:%s: raid level %d active with %d out of %d" " devices, algorithm %d ", mdname(mddev), conf->level, |
e183eaedd
|
4792 4793 |
mddev->raid_disks-mddev->degraded, mddev->raid_disks, mddev->new_layout); |
1da177e4c
|
4794 |
else |
0c55e0225
|
4795 4796 4797 4798 4799 4800 |
printk(KERN_ALERT "md/raid:%s: raid level %d active with %d" " out of %d devices, algorithm %d ", mdname(mddev), conf->level, mddev->raid_disks - mddev->degraded, mddev->raid_disks, mddev->new_layout); |
1da177e4c
|
4801 4802 |
print_raid5_conf(conf); |
fef9c61fd
|
4803 |
if (conf->reshape_progress != MaxSector) { |
fef9c61fd
|
4804 |
conf->reshape_safe = conf->reshape_progress; |
f67055780
|
4805 4806 4807 4808 4809 4810 |
atomic_set(&conf->reshape_stripes, 0); clear_bit(MD_RECOVERY_SYNC, &mddev->recovery); clear_bit(MD_RECOVERY_CHECK, &mddev->recovery); set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery); set_bit(MD_RECOVERY_RUNNING, &mddev->recovery); mddev->sync_thread = md_register_thread(md_do_sync, mddev, |
0da3c6194
|
4811 |
"reshape"); |
f67055780
|
4812 |
} |
1da177e4c
|
4813 4814 |
/* Ok, everything is just fine now */ |
a64c876fd
|
4815 4816 |
if (mddev->to_remove == &raid5_attrs_group) mddev->to_remove = NULL; |
00bcb4ac7
|
4817 4818 |
else if (mddev->kobj.sd && sysfs_create_group(&mddev->kobj, &raid5_attrs_group)) |
5e55e2f5f
|
4819 |
printk(KERN_WARNING |
4a5add499
|
4820 4821 |
"raid5: failed to create sysfs attributes for %s ", |
5e55e2f5f
|
4822 |
mdname(mddev)); |
4a5add499
|
4823 |
md_set_array_sectors(mddev, raid5_size(mddev, 0, 0)); |
7a5febe9f
|
4824 |
|
4a5add499
|
4825 |
if (mddev->queue) { |
9f7c22200
|
4826 |
int chunk_size; |
4a5add499
|
4827 4828 4829 4830 4831 4832 4833 4834 4835 |
/* read-ahead size must cover two whole stripes, which * is 2 * (datadisks) * chunksize where 'n' is the * number of raid devices */ int data_disks = conf->previous_raid_disks - conf->max_degraded; int stripe = data_disks * ((mddev->chunk_sectors << 9) / PAGE_SIZE); if (mddev->queue->backing_dev_info.ra_pages < 2 * stripe) mddev->queue->backing_dev_info.ra_pages = 2 * stripe; |
91adb5647
|
4836 |
|
4a5add499
|
4837 |
blk_queue_merge_bvec(mddev->queue, raid5_mergeable_bvec); |
f022b2fdd
|
4838 |
|
11d8a6e37
|
4839 4840 |
mddev->queue->backing_dev_info.congested_data = mddev; mddev->queue->backing_dev_info.congested_fn = raid5_congested; |
7a5febe9f
|
4841 |
|
9f7c22200
|
4842 4843 4844 4845 |
chunk_size = mddev->chunk_sectors << 9; blk_queue_io_min(mddev->queue, chunk_size); blk_queue_io_opt(mddev->queue, chunk_size * (conf->raid_disks - conf->max_degraded)); |
8f6c2e4b3
|
4846 |
|
9f7c22200
|
4847 4848 4849 4850 |
list_for_each_entry(rdev, &mddev->disks, same_set) disk_stack_limits(mddev->gendisk, rdev->bdev, rdev->data_offset << 9); } |
23032a0eb
|
4851 |
|
1da177e4c
|
4852 4853 |
return 0; abort: |
01f96c0a9
|
4854 |
md_unregister_thread(&mddev->thread); |
e4f869d9d
|
4855 4856 |
print_raid5_conf(conf); free_conf(conf); |
1da177e4c
|
4857 |
mddev->private = NULL; |
0c55e0225
|
4858 4859 |
printk(KERN_ALERT "md/raid:%s: failed to run raid set. ", mdname(mddev)); |
1da177e4c
|
4860 4861 |
return -EIO; } |
fd01b88c7
|
4862 |
static int stop(struct mddev *mddev) |
1da177e4c
|
4863 |
{ |
d1688a6d5
|
4864 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
4865 |
|
01f96c0a9
|
4866 |
md_unregister_thread(&mddev->thread); |
11d8a6e37
|
4867 4868 |
if (mddev->queue) mddev->queue->backing_dev_info.congested_fn = NULL; |
95fc17aac
|
4869 |
free_conf(conf); |
a64c876fd
|
4870 4871 |
mddev->private = NULL; mddev->to_remove = &raid5_attrs_group; |
1da177e4c
|
4872 4873 |
return 0; } |
fd01b88c7
|
4874 |
static void status(struct seq_file *seq, struct mddev *mddev) |
1da177e4c
|
4875 |
{ |
d1688a6d5
|
4876 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
4877 |
int i; |
9d8f03636
|
4878 4879 |
seq_printf(seq, " level %d, %dk chunk, algorithm %d", mddev->level, mddev->chunk_sectors / 2, mddev->layout); |
02c2de8cc
|
4880 |
seq_printf (seq, " [%d/%d] [", conf->raid_disks, conf->raid_disks - mddev->degraded); |
1da177e4c
|
4881 4882 4883 |
for (i = 0; i < conf->raid_disks; i++) seq_printf (seq, "%s", conf->disks[i].rdev && |
b2d444d7a
|
4884 |
test_bit(In_sync, &conf->disks[i].rdev->flags) ? "U" : "_"); |
1da177e4c
|
4885 |
seq_printf (seq, "]"); |
1da177e4c
|
4886 |
} |
d1688a6d5
|
4887 |
static void print_raid5_conf (struct r5conf *conf) |
1da177e4c
|
4888 4889 4890 |
{ int i; struct disk_info *tmp; |
0c55e0225
|
4891 4892 |
printk(KERN_DEBUG "RAID conf printout: "); |
1da177e4c
|
4893 4894 4895 4896 4897 |
if (!conf) { printk("(conf==NULL) "); return; } |
0c55e0225
|
4898 4899 4900 4901 |
printk(KERN_DEBUG " --- level:%d rd:%d wd:%d ", conf->level, conf->raid_disks, conf->raid_disks - conf->mddev->degraded); |
1da177e4c
|
4902 4903 4904 4905 4906 |
for (i = 0; i < conf->raid_disks; i++) { char b[BDEVNAME_SIZE]; tmp = conf->disks + i; if (tmp->rdev) |
0c55e0225
|
4907 4908 4909 4910 |
printk(KERN_DEBUG " disk %d, o:%d, dev:%s ", i, !test_bit(Faulty, &tmp->rdev->flags), bdevname(tmp->rdev->bdev, b)); |
1da177e4c
|
4911 4912 |
} } |
fd01b88c7
|
4913 |
static int raid5_spare_active(struct mddev *mddev) |
1da177e4c
|
4914 4915 |
{ int i; |
d1688a6d5
|
4916 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
4917 |
struct disk_info *tmp; |
6b9656205
|
4918 4919 |
int count = 0; unsigned long flags; |
1da177e4c
|
4920 4921 4922 4923 |
for (i = 0; i < conf->raid_disks; i++) { tmp = conf->disks + i; if (tmp->rdev |
70fffd0bf
|
4924 |
&& tmp->rdev->recovery_offset == MaxSector |
b2d444d7a
|
4925 |
&& !test_bit(Faulty, &tmp->rdev->flags) |
c04be0aa8
|
4926 |
&& !test_and_set_bit(In_sync, &tmp->rdev->flags)) { |
6b9656205
|
4927 |
count++; |
43c73ca43
|
4928 |
sysfs_notify_dirent_safe(tmp->rdev->sysfs_state); |
1da177e4c
|
4929 4930 |
} } |
6b9656205
|
4931 4932 4933 |
spin_lock_irqsave(&conf->device_lock, flags); mddev->degraded -= count; spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c
|
4934 |
print_raid5_conf(conf); |
6b9656205
|
4935 |
return count; |
1da177e4c
|
4936 |
} |
fd01b88c7
|
4937 |
static int raid5_remove_disk(struct mddev *mddev, int number) |
1da177e4c
|
4938 |
{ |
d1688a6d5
|
4939 |
struct r5conf *conf = mddev->private; |
1da177e4c
|
4940 |
int err = 0; |
3cb030020
|
4941 |
struct md_rdev *rdev; |
1da177e4c
|
4942 4943 4944 4945 4946 |
struct disk_info *p = conf->disks + number; print_raid5_conf(conf); rdev = p->rdev; if (rdev) { |
ec32a2bd3
|
4947 4948 4949 |
if (number >= conf->raid_disks && conf->reshape_progress == MaxSector) clear_bit(In_sync, &rdev->flags); |
b2d444d7a
|
4950 |
if (test_bit(In_sync, &rdev->flags) || |
1da177e4c
|
4951 4952 4953 4954 |
atomic_read(&rdev->nr_pending)) { err = -EBUSY; goto abort; } |
dfc706450
|
4955 4956 4957 4958 |
/* Only remove non-faulty devices if recovery * isn't possible. */ if (!test_bit(Faulty, &rdev->flags) && |
7f0da59bd
|
4959 |
mddev->recovery_disabled != conf->recovery_disabled && |
674806d62
|
4960 |
!has_failed(conf) && |
ec32a2bd3
|
4961 |
number < conf->raid_disks) { |
dfc706450
|
4962 4963 4964 |
err = -EBUSY; goto abort; } |
1da177e4c
|
4965 |
p->rdev = NULL; |
fbd568a3e
|
4966 |
synchronize_rcu(); |
1da177e4c
|
4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 |
if (atomic_read(&rdev->nr_pending)) { /* lost the race, try later */ err = -EBUSY; p->rdev = rdev; } } abort: print_raid5_conf(conf); return err; } |
fd01b88c7
|
4978 |
static int raid5_add_disk(struct mddev *mddev, struct md_rdev *rdev) |
1da177e4c
|
4979 |
{ |
d1688a6d5
|
4980 |
struct r5conf *conf = mddev->private; |
199050ea1
|
4981 |
int err = -EEXIST; |
1da177e4c
|
4982 4983 |
int disk; struct disk_info *p; |
6c2fce2ef
|
4984 4985 |
int first = 0; int last = conf->raid_disks - 1; |
1da177e4c
|
4986 |
|
7f0da59bd
|
4987 4988 |
if (mddev->recovery_disabled == conf->recovery_disabled) return -EBUSY; |
674806d62
|
4989 |
if (has_failed(conf)) |
1da177e4c
|
4990 |
/* no point adding a device */ |
199050ea1
|
4991 |
return -EINVAL; |
1da177e4c
|
4992 |
|
6c2fce2ef
|
4993 4994 |
if (rdev->raid_disk >= 0) first = last = rdev->raid_disk; |
1da177e4c
|
4995 4996 |
/* |
16a53ecc3
|
4997 4998 |
* find the disk ... but prefer rdev->saved_raid_disk * if possible. |
1da177e4c
|
4999 |
*/ |
16a53ecc3
|
5000 |
if (rdev->saved_raid_disk >= 0 && |
6c2fce2ef
|
5001 |
rdev->saved_raid_disk >= first && |
16a53ecc3
|
5002 5003 5004 |
conf->disks[rdev->saved_raid_disk].rdev == NULL) disk = rdev->saved_raid_disk; else |
6c2fce2ef
|
5005 5006 |
disk = first; for ( ; disk <= last ; disk++) |
1da177e4c
|
5007 |
if ((p=conf->disks + disk)->rdev == NULL) { |
b2d444d7a
|
5008 |
clear_bit(In_sync, &rdev->flags); |
1da177e4c
|
5009 |
rdev->raid_disk = disk; |
199050ea1
|
5010 |
err = 0; |
72626685d
|
5011 5012 |
if (rdev->saved_raid_disk != disk) conf->fullsync = 1; |
d6065f7bf
|
5013 |
rcu_assign_pointer(p->rdev, rdev); |
1da177e4c
|
5014 5015 5016 |
break; } print_raid5_conf(conf); |
199050ea1
|
5017 |
return err; |
1da177e4c
|
5018 |
} |
fd01b88c7
|
5019 |
static int raid5_resize(struct mddev *mddev, sector_t sectors) |
1da177e4c
|
5020 5021 5022 5023 5024 5025 5026 5027 |
{ /* no resync is happening, and there is enough space * on all devices, so we can resize. * We need to make sure resync covers any new space. * If the array is shrinking we should possibly wait until * any io in the removed space completes, but it hardly seems * worth it. */ |
9d8f03636
|
5028 |
sectors &= ~((sector_t)mddev->chunk_sectors - 1); |
1f403624b
|
5029 5030 |
md_set_array_sectors(mddev, raid5_size(mddev, sectors, mddev->raid_disks)); |
b522adcde
|
5031 5032 5033 |
if (mddev->array_sectors > raid5_size(mddev, sectors, mddev->raid_disks)) return -EINVAL; |
f233ea5c9
|
5034 |
set_capacity(mddev->gendisk, mddev->array_sectors); |
449aad3e2
|
5035 |
revalidate_disk(mddev->gendisk); |
b098636cf
|
5036 5037 |
if (sectors > mddev->dev_sectors && mddev->recovery_cp > mddev->dev_sectors) { |
58c0fed40
|
5038 |
mddev->recovery_cp = mddev->dev_sectors; |
1da177e4c
|
5039 5040 |
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); } |
58c0fed40
|
5041 |
mddev->dev_sectors = sectors; |
4b5c7ae83
|
5042 |
mddev->resync_max_sectors = sectors; |
1da177e4c
|
5043 5044 |
return 0; } |
fd01b88c7
|
5045 |
static int check_stripe_cache(struct mddev *mddev) |
01ee22b49
|
5046 5047 5048 5049 5050 5051 5052 5053 5054 |
{ /* Can only proceed if there are plenty of stripe_heads. * We need a minimum of one full stripe,, and for sensible progress * it is best to have about 4 times that. * If we require 4 times, then the default 256 4K stripe_heads will * allow for chunk sizes up to 256K, which is probably OK. * If the chunk size is greater, user-space should request more * stripe_heads first. */ |
d1688a6d5
|
5055 |
struct r5conf *conf = mddev->private; |
01ee22b49
|
5056 5057 5058 5059 |
if (((mddev->chunk_sectors << 9) / STRIPE_SIZE) * 4 > conf->max_nr_stripes || ((mddev->new_chunk_sectors << 9) / STRIPE_SIZE) * 4 > conf->max_nr_stripes) { |
0c55e0225
|
5060 5061 5062 |
printk(KERN_WARNING "md/raid:%s: reshape: not enough stripes. Needed %lu ", mdname(mddev), |
01ee22b49
|
5063 5064 5065 5066 5067 5068 |
((max(mddev->chunk_sectors, mddev->new_chunk_sectors) << 9) / STRIPE_SIZE)*4); return 0; } return 1; } |
fd01b88c7
|
5069 |
static int check_reshape(struct mddev *mddev) |
292695531
|
5070 |
{ |
d1688a6d5
|
5071 |
struct r5conf *conf = mddev->private; |
292695531
|
5072 |
|
88ce4930e
|
5073 5074 |
if (mddev->delta_disks == 0 && mddev->new_layout == mddev->layout && |
664e7c413
|
5075 |
mddev->new_chunk_sectors == mddev->chunk_sectors) |
50ac168a6
|
5076 |
return 0; /* nothing to do */ |
dba034eef
|
5077 5078 5079 |
if (mddev->bitmap) /* Cannot grow a bitmap yet */ return -EBUSY; |
674806d62
|
5080 |
if (has_failed(conf)) |
ec32a2bd3
|
5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092 5093 |
return -EINVAL; if (mddev->delta_disks < 0) { /* We might be able to shrink, but the devices must * be made bigger first. * For raid6, 4 is the minimum size. * Otherwise 2 is the minimum */ int min = 2; if (mddev->level == 6) min = 4; if (mddev->raid_disks + mddev->delta_disks < min) return -EINVAL; } |
292695531
|
5094 |
|
01ee22b49
|
5095 |
if (!check_stripe_cache(mddev)) |
292695531
|
5096 |
return -ENOSPC; |
292695531
|
5097 |
|
ec32a2bd3
|
5098 |
return resize_stripes(conf, conf->raid_disks + mddev->delta_disks); |
63c70c4f3
|
5099 |
} |
fd01b88c7
|
5100 |
static int raid5_start_reshape(struct mddev *mddev) |
63c70c4f3
|
5101 |
{ |
d1688a6d5
|
5102 |
struct r5conf *conf = mddev->private; |
3cb030020
|
5103 |
struct md_rdev *rdev; |
63c70c4f3
|
5104 |
int spares = 0; |
c04be0aa8
|
5105 |
unsigned long flags; |
63c70c4f3
|
5106 |
|
f416885ef
|
5107 |
if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) |
63c70c4f3
|
5108 |
return -EBUSY; |
01ee22b49
|
5109 5110 |
if (!check_stripe_cache(mddev)) return -ENOSPC; |
159ec1fc0
|
5111 |
list_for_each_entry(rdev, &mddev->disks, same_set) |
469518a34
|
5112 5113 |
if (!test_bit(In_sync, &rdev->flags) && !test_bit(Faulty, &rdev->flags)) |
292695531
|
5114 |
spares++; |
63c70c4f3
|
5115 |
|
f416885ef
|
5116 |
if (spares - mddev->degraded < mddev->delta_disks - conf->max_degraded) |
292695531
|
5117 5118 5119 5120 |
/* Not enough devices even to make a degraded array * of that size */ return -EINVAL; |
ec32a2bd3
|
5121 5122 5123 5124 5125 5126 |
/* Refuse to reduce size of the array. Any reductions in * array size must be through explicit setting of array_size * attribute. */ if (raid5_size(mddev, 0, conf->raid_disks + mddev->delta_disks) < mddev->array_sectors) { |
0c55e0225
|
5127 |
printk(KERN_ERR "md/raid:%s: array size must be reduced " |
ec32a2bd3
|
5128 5129 5130 5131 |
"before number of disks ", mdname(mddev)); return -EINVAL; } |
f67055780
|
5132 |
atomic_set(&conf->reshape_stripes, 0); |
292695531
|
5133 5134 |
spin_lock_irq(&conf->device_lock); conf->previous_raid_disks = conf->raid_disks; |
63c70c4f3
|
5135 |
conf->raid_disks += mddev->delta_disks; |
09c9e5fa1
|
5136 5137 |
conf->prev_chunk_sectors = conf->chunk_sectors; conf->chunk_sectors = mddev->new_chunk_sectors; |
88ce4930e
|
5138 5139 |
conf->prev_algo = conf->algorithm; conf->algorithm = mddev->new_layout; |
fef9c61fd
|
5140 5141 5142 5143 5144 |
if (mddev->delta_disks < 0) conf->reshape_progress = raid5_size(mddev, 0, 0); else conf->reshape_progress = 0; conf->reshape_safe = conf->reshape_progress; |
86b42c713
|
5145 |
conf->generation++; |
292695531
|
5146 5147 5148 5149 |
spin_unlock_irq(&conf->device_lock); /* Add some new drives, as many as will fit. * We know there are enough to make the newly sized array work. |
3424bf6a7
|
5150 5151 5152 5153 |
* Don't add devices if we are reducing the number of * devices in the array. This is because it is not possible * to correctly record the "partially reconstructed" state of * such devices during the reshape and confusion could result. |
292695531
|
5154 |
*/ |
87a8dec91
|
5155 5156 5157 5158 5159 5160 |
if (mddev->delta_disks >= 0) { int added_devices = 0; list_for_each_entry(rdev, &mddev->disks, same_set) if (rdev->raid_disk < 0 && !test_bit(Faulty, &rdev->flags)) { if (raid5_add_disk(mddev, rdev) == 0) { |
87a8dec91
|
5161 5162 5163 5164 5165 5166 |
if (rdev->raid_disk >= conf->previous_raid_disks) { set_bit(In_sync, &rdev->flags); added_devices++; } else rdev->recovery_offset = 0; |
36fad858a
|
5167 5168 |
if (sysfs_link_rdev(mddev, rdev)) |
87a8dec91
|
5169 |
/* Failure here is OK */; |
50da08409
|
5170 |
} |
87a8dec91
|
5171 5172 5173 5174 5175 5176 |
} else if (rdev->raid_disk >= conf->previous_raid_disks && !test_bit(Faulty, &rdev->flags)) { /* This is a spare that was manually added */ set_bit(In_sync, &rdev->flags); added_devices++; } |
292695531
|
5177 |
|
87a8dec91
|
5178 5179 5180 5181 |
/* When a reshape changes the number of devices, * ->degraded is measured against the larger of the * pre and post number of devices. */ |
ec32a2bd3
|
5182 |
spin_lock_irqsave(&conf->device_lock, flags); |
9eb07c259
|
5183 |
mddev->degraded += (conf->raid_disks - conf->previous_raid_disks) |
ec32a2bd3
|
5184 5185 5186 |
- added_devices; spin_unlock_irqrestore(&conf->device_lock, flags); } |
63c70c4f3
|
5187 |
mddev->raid_disks = conf->raid_disks; |
e516402c0
|
5188 |
mddev->reshape_position = conf->reshape_progress; |
850b2b420
|
5189 |
set_bit(MD_CHANGE_DEVS, &mddev->flags); |
f67055780
|
5190 |
|
292695531
|
5191 5192 5193 5194 5195 |
clear_bit(MD_RECOVERY_SYNC, &mddev->recovery); clear_bit(MD_RECOVERY_CHECK, &mddev->recovery); set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery); set_bit(MD_RECOVERY_RUNNING, &mddev->recovery); mddev->sync_thread = md_register_thread(md_do_sync, mddev, |
0da3c6194
|
5196 |
"reshape"); |
292695531
|
5197 5198 5199 5200 |
if (!mddev->sync_thread) { mddev->recovery = 0; spin_lock_irq(&conf->device_lock); mddev->raid_disks = conf->raid_disks = conf->previous_raid_disks; |
fef9c61fd
|
5201 |
conf->reshape_progress = MaxSector; |
292695531
|
5202 5203 5204 |
spin_unlock_irq(&conf->device_lock); return -EAGAIN; } |
c8f517c44
|
5205 |
conf->reshape_checkpoint = jiffies; |
292695531
|
5206 5207 5208 5209 |
md_wakeup_thread(mddev->sync_thread); md_new_event(mddev); return 0; } |
292695531
|
5210 |
|
ec32a2bd3
|
5211 5212 5213 |
/* This is called from the reshape thread and should make any * changes needed in 'conf' */ |
d1688a6d5
|
5214 |
static void end_reshape(struct r5conf *conf) |
292695531
|
5215 |
{ |
292695531
|
5216 |
|
f67055780
|
5217 |
if (!test_bit(MD_RECOVERY_INTR, &conf->mddev->recovery)) { |
f67055780
|
5218 |
|
f67055780
|
5219 |
spin_lock_irq(&conf->device_lock); |
cea9c2280
|
5220 |
conf->previous_raid_disks = conf->raid_disks; |
fef9c61fd
|
5221 |
conf->reshape_progress = MaxSector; |
f67055780
|
5222 |
spin_unlock_irq(&conf->device_lock); |
b0f9ec047
|
5223 |
wake_up(&conf->wait_for_overlap); |
16a53ecc3
|
5224 5225 5226 5227 |
/* read-ahead size must cover two whole stripes, which is * 2 * (datadisks) * chunksize where 'n' is the number of raid devices */ |
4a5add499
|
5228 |
if (conf->mddev->queue) { |
cea9c2280
|
5229 |
int data_disks = conf->raid_disks - conf->max_degraded; |
09c9e5fa1
|
5230 |
int stripe = data_disks * ((conf->chunk_sectors << 9) |
cea9c2280
|
5231 |
/ PAGE_SIZE); |
16a53ecc3
|
5232 5233 5234 |
if (conf->mddev->queue->backing_dev_info.ra_pages < 2 * stripe) conf->mddev->queue->backing_dev_info.ra_pages = 2 * stripe; } |
292695531
|
5235 |
} |
292695531
|
5236 |
} |
ec32a2bd3
|
5237 5238 5239 |
/* This is called from the raid5d thread with mddev_lock held. * It makes config changes to the device. */ |
fd01b88c7
|
5240 |
static void raid5_finish_reshape(struct mddev *mddev) |
cea9c2280
|
5241 |
{ |
d1688a6d5
|
5242 |
struct r5conf *conf = mddev->private; |
cea9c2280
|
5243 5244 |
if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) { |
ec32a2bd3
|
5245 5246 5247 |
if (mddev->delta_disks > 0) { md_set_array_sectors(mddev, raid5_size(mddev, 0, 0)); set_capacity(mddev->gendisk, mddev->array_sectors); |
449aad3e2
|
5248 |
revalidate_disk(mddev->gendisk); |
ec32a2bd3
|
5249 5250 |
} else { int d; |
ec32a2bd3
|
5251 5252 5253 5254 5255 5256 5257 5258 |
mddev->degraded = conf->raid_disks; for (d = 0; d < conf->raid_disks ; d++) if (conf->disks[d].rdev && test_bit(In_sync, &conf->disks[d].rdev->flags)) mddev->degraded--; for (d = conf->raid_disks ; d < conf->raid_disks - mddev->delta_disks; |
1a67dde0a
|
5259 |
d++) { |
3cb030020
|
5260 |
struct md_rdev *rdev = conf->disks[d].rdev; |
1a67dde0a
|
5261 |
if (rdev && raid5_remove_disk(mddev, d) == 0) { |
36fad858a
|
5262 |
sysfs_unlink_rdev(mddev, rdev); |
1a67dde0a
|
5263 5264 5265 |
rdev->raid_disk = -1; } } |
cea9c2280
|
5266 |
} |
88ce4930e
|
5267 |
mddev->layout = conf->algorithm; |
09c9e5fa1
|
5268 |
mddev->chunk_sectors = conf->chunk_sectors; |
ec32a2bd3
|
5269 5270 |
mddev->reshape_position = MaxSector; mddev->delta_disks = 0; |
cea9c2280
|
5271 5272 |
} } |
fd01b88c7
|
5273 |
static void raid5_quiesce(struct mddev *mddev, int state) |
72626685d
|
5274 |
{ |
d1688a6d5
|
5275 |
struct r5conf *conf = mddev->private; |
72626685d
|
5276 5277 |
switch(state) { |
e464eafdb
|
5278 5279 5280 |
case 2: /* resume for a suspend */ wake_up(&conf->wait_for_overlap); break; |
72626685d
|
5281 5282 |
case 1: /* stop all writes */ spin_lock_irq(&conf->device_lock); |
64bd660b5
|
5283 5284 5285 5286 |
/* '2' tells resync/reshape to pause so that all * active stripes can drain */ conf->quiesce = 2; |
72626685d
|
5287 |
wait_event_lock_irq(conf->wait_for_stripe, |
46031f9a3
|
5288 5289 |
atomic_read(&conf->active_stripes) == 0 && atomic_read(&conf->active_aligned_reads) == 0, |
72626685d
|
5290 |
conf->device_lock, /* nothing */); |
64bd660b5
|
5291 |
conf->quiesce = 1; |
72626685d
|
5292 |
spin_unlock_irq(&conf->device_lock); |
64bd660b5
|
5293 5294 |
/* allow reshape to continue */ wake_up(&conf->wait_for_overlap); |
72626685d
|
5295 5296 5297 5298 5299 5300 |
break; case 0: /* re-enable writes */ spin_lock_irq(&conf->device_lock); conf->quiesce = 0; wake_up(&conf->wait_for_stripe); |
e464eafdb
|
5301 |
wake_up(&conf->wait_for_overlap); |
72626685d
|
5302 5303 5304 |
spin_unlock_irq(&conf->device_lock); break; } |
72626685d
|
5305 |
} |
b15c2e57f
|
5306 |
|
d562b0c43
|
5307 |
|
fd01b88c7
|
5308 |
static void *raid45_takeover_raid0(struct mddev *mddev, int level) |
54071b380
|
5309 |
{ |
e373ab109
|
5310 |
struct r0conf *raid0_conf = mddev->private; |
d76c8420c
|
5311 |
sector_t sectors; |
54071b380
|
5312 |
|
f1b29bcae
|
5313 |
/* for raid0 takeover only one zone is supported */ |
e373ab109
|
5314 |
if (raid0_conf->nr_strip_zones > 1) { |
0c55e0225
|
5315 5316 5317 |
printk(KERN_ERR "md/raid:%s: cannot takeover raid0 with more than one zone. ", mdname(mddev)); |
f1b29bcae
|
5318 5319 |
return ERR_PTR(-EINVAL); } |
e373ab109
|
5320 5321 |
sectors = raid0_conf->strip_zone[0].zone_end; sector_div(sectors, raid0_conf->strip_zone[0].nb_dev); |
3b71bd933
|
5322 |
mddev->dev_sectors = sectors; |
f1b29bcae
|
5323 |
mddev->new_level = level; |
54071b380
|
5324 5325 5326 5327 5328 5329 5330 5331 5332 |
mddev->new_layout = ALGORITHM_PARITY_N; mddev->new_chunk_sectors = mddev->chunk_sectors; mddev->raid_disks += 1; mddev->delta_disks = 1; /* make sure it will be not marked as dirty */ mddev->recovery_cp = MaxSector; return setup_conf(mddev); } |
fd01b88c7
|
5333 |
static void *raid5_takeover_raid1(struct mddev *mddev) |
d562b0c43
|
5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 |
{ int chunksect; if (mddev->raid_disks != 2 || mddev->degraded > 1) return ERR_PTR(-EINVAL); /* Should check if there are write-behind devices? */ chunksect = 64*2; /* 64K by default */ /* The array must be an exact multiple of chunksize */ while (chunksect && (mddev->array_sectors & (chunksect-1))) chunksect >>= 1; if ((chunksect<<9) < STRIPE_SIZE) /* array size does not allow a suitable chunk size */ return ERR_PTR(-EINVAL); mddev->new_level = 5; mddev->new_layout = ALGORITHM_LEFT_SYMMETRIC; |
664e7c413
|
5355 |
mddev->new_chunk_sectors = chunksect; |
d562b0c43
|
5356 5357 5358 |
return setup_conf(mddev); } |
fd01b88c7
|
5359 |
static void *raid5_takeover_raid6(struct mddev *mddev) |
fc9739c6d
|
5360 5361 5362 5363 5364 5365 5366 5367 5368 5369 5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 |
{ int new_layout; switch (mddev->layout) { case ALGORITHM_LEFT_ASYMMETRIC_6: new_layout = ALGORITHM_LEFT_ASYMMETRIC; break; case ALGORITHM_RIGHT_ASYMMETRIC_6: new_layout = ALGORITHM_RIGHT_ASYMMETRIC; break; case ALGORITHM_LEFT_SYMMETRIC_6: new_layout = ALGORITHM_LEFT_SYMMETRIC; break; case ALGORITHM_RIGHT_SYMMETRIC_6: new_layout = ALGORITHM_RIGHT_SYMMETRIC; break; case ALGORITHM_PARITY_0_6: new_layout = ALGORITHM_PARITY_0; break; case ALGORITHM_PARITY_N: new_layout = ALGORITHM_PARITY_N; break; default: return ERR_PTR(-EINVAL); } mddev->new_level = 5; mddev->new_layout = new_layout; mddev->delta_disks = -1; mddev->raid_disks -= 1; return setup_conf(mddev); } |
d562b0c43
|
5391 |
|
fd01b88c7
|
5392 |
static int raid5_check_reshape(struct mddev *mddev) |
b35460352
|
5393 |
{ |
88ce4930e
|
5394 5395 5396 5397 |
/* For a 2-drive array, the layout and chunk size can be changed * immediately as not restriping is needed. * For larger arrays we record the new value - after validation * to be used by a reshape pass. |
b35460352
|
5398 |
*/ |
d1688a6d5
|
5399 |
struct r5conf *conf = mddev->private; |
597a711b6
|
5400 |
int new_chunk = mddev->new_chunk_sectors; |
b35460352
|
5401 |
|
597a711b6
|
5402 |
if (mddev->new_layout >= 0 && !algorithm_valid_raid5(mddev->new_layout)) |
b35460352
|
5403 5404 |
return -EINVAL; if (new_chunk > 0) { |
0ba459d26
|
5405 |
if (!is_power_of_2(new_chunk)) |
b35460352
|
5406 |
return -EINVAL; |
597a711b6
|
5407 |
if (new_chunk < (PAGE_SIZE>>9)) |
b35460352
|
5408 |
return -EINVAL; |
597a711b6
|
5409 |
if (mddev->array_sectors & (new_chunk-1)) |
b35460352
|
5410 5411 5412 5413 5414 |
/* not factor of array size */ return -EINVAL; } /* They look valid */ |
88ce4930e
|
5415 |
if (mddev->raid_disks == 2) { |
597a711b6
|
5416 5417 5418 5419 |
/* can make the change immediately */ if (mddev->new_layout >= 0) { conf->algorithm = mddev->new_layout; mddev->layout = mddev->new_layout; |
88ce4930e
|
5420 5421 |
} if (new_chunk > 0) { |
597a711b6
|
5422 5423 |
conf->chunk_sectors = new_chunk ; mddev->chunk_sectors = new_chunk; |
88ce4930e
|
5424 5425 5426 |
} set_bit(MD_CHANGE_DEVS, &mddev->flags); md_wakeup_thread(mddev->thread); |
b35460352
|
5427 |
} |
50ac168a6
|
5428 |
return check_reshape(mddev); |
88ce4930e
|
5429 |
} |
fd01b88c7
|
5430 |
static int raid6_check_reshape(struct mddev *mddev) |
88ce4930e
|
5431 |
{ |
597a711b6
|
5432 |
int new_chunk = mddev->new_chunk_sectors; |
50ac168a6
|
5433 |
|
597a711b6
|
5434 |
if (mddev->new_layout >= 0 && !algorithm_valid_raid6(mddev->new_layout)) |
88ce4930e
|
5435 |
return -EINVAL; |
b35460352
|
5436 |
if (new_chunk > 0) { |
0ba459d26
|
5437 |
if (!is_power_of_2(new_chunk)) |
88ce4930e
|
5438 |
return -EINVAL; |
597a711b6
|
5439 |
if (new_chunk < (PAGE_SIZE >> 9)) |
88ce4930e
|
5440 |
return -EINVAL; |
597a711b6
|
5441 |
if (mddev->array_sectors & (new_chunk-1)) |
88ce4930e
|
5442 5443 |
/* not factor of array size */ return -EINVAL; |
b35460352
|
5444 |
} |
88ce4930e
|
5445 5446 |
/* They look valid */ |
50ac168a6
|
5447 |
return check_reshape(mddev); |
b35460352
|
5448 |
} |
fd01b88c7
|
5449 |
static void *raid5_takeover(struct mddev *mddev) |
d562b0c43
|
5450 5451 |
{ /* raid5 can take over: |
f1b29bcae
|
5452 |
* raid0 - if there is only one strip zone - make it a raid4 layout |
d562b0c43
|
5453 5454 5455 |
* raid1 - if there are two drives. We need to know the chunk size * raid4 - trivial - just use a raid4 layout. * raid6 - Providing it is a *_6 layout |
d562b0c43
|
5456 |
*/ |
f1b29bcae
|
5457 5458 |
if (mddev->level == 0) return raid45_takeover_raid0(mddev, 5); |
d562b0c43
|
5459 5460 |
if (mddev->level == 1) return raid5_takeover_raid1(mddev); |
e9d4758f6
|
5461 5462 5463 5464 5465 |
if (mddev->level == 4) { mddev->new_layout = ALGORITHM_PARITY_N; mddev->new_level = 5; return setup_conf(mddev); } |
fc9739c6d
|
5466 5467 |
if (mddev->level == 6) return raid5_takeover_raid6(mddev); |
d562b0c43
|
5468 5469 5470 |
return ERR_PTR(-EINVAL); } |
fd01b88c7
|
5471 |
static void *raid4_takeover(struct mddev *mddev) |
a78d38a1a
|
5472 |
{ |
f1b29bcae
|
5473 5474 5475 |
/* raid4 can take over: * raid0 - if there is only one strip zone * raid5 - if layout is right |
a78d38a1a
|
5476 |
*/ |
f1b29bcae
|
5477 5478 |
if (mddev->level == 0) return raid45_takeover_raid0(mddev, 4); |
a78d38a1a
|
5479 5480 5481 5482 5483 5484 5485 5486 |
if (mddev->level == 5 && mddev->layout == ALGORITHM_PARITY_N) { mddev->new_layout = 0; mddev->new_level = 4; return setup_conf(mddev); } return ERR_PTR(-EINVAL); } |
d562b0c43
|
5487 |
|
84fc4b56d
|
5488 |
static struct md_personality raid5_personality; |
245f46c2c
|
5489 |
|
fd01b88c7
|
5490 |
static void *raid6_takeover(struct mddev *mddev) |
245f46c2c
|
5491 5492 5493 5494 5495 5496 5497 5498 5499 5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 |
{ /* Currently can only take over a raid5. We map the * personality to an equivalent raid6 personality * with the Q block at the end. */ int new_layout; if (mddev->pers != &raid5_personality) return ERR_PTR(-EINVAL); if (mddev->degraded > 1) return ERR_PTR(-EINVAL); if (mddev->raid_disks > 253) return ERR_PTR(-EINVAL); if (mddev->raid_disks < 3) return ERR_PTR(-EINVAL); switch (mddev->layout) { case ALGORITHM_LEFT_ASYMMETRIC: new_layout = ALGORITHM_LEFT_ASYMMETRIC_6; break; case ALGORITHM_RIGHT_ASYMMETRIC: new_layout = ALGORITHM_RIGHT_ASYMMETRIC_6; break; case ALGORITHM_LEFT_SYMMETRIC: new_layout = ALGORITHM_LEFT_SYMMETRIC_6; break; case ALGORITHM_RIGHT_SYMMETRIC: new_layout = ALGORITHM_RIGHT_SYMMETRIC_6; break; case ALGORITHM_PARITY_0: new_layout = ALGORITHM_PARITY_0_6; break; case ALGORITHM_PARITY_N: new_layout = ALGORITHM_PARITY_N; break; default: return ERR_PTR(-EINVAL); } mddev->new_level = 6; mddev->new_layout = new_layout; mddev->delta_disks = 1; mddev->raid_disks += 1; return setup_conf(mddev); } |
84fc4b56d
|
5535 |
static struct md_personality raid6_personality = |
16a53ecc3
|
5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 |
{ .name = "raid6", .level = 6, .owner = THIS_MODULE, .make_request = make_request, .run = run, .stop = stop, .status = status, .error_handler = error, .hot_add_disk = raid5_add_disk, .hot_remove_disk= raid5_remove_disk, .spare_active = raid5_spare_active, .sync_request = sync_request, .resize = raid5_resize, |
80c3a6ce4
|
5550 |
.size = raid5_size, |
50ac168a6
|
5551 |
.check_reshape = raid6_check_reshape, |
f416885ef
|
5552 |
.start_reshape = raid5_start_reshape, |
cea9c2280
|
5553 |
.finish_reshape = raid5_finish_reshape, |
16a53ecc3
|
5554 |
.quiesce = raid5_quiesce, |
245f46c2c
|
5555 |
.takeover = raid6_takeover, |
16a53ecc3
|
5556 |
}; |
84fc4b56d
|
5557 |
static struct md_personality raid5_personality = |
1da177e4c
|
5558 5559 |
{ .name = "raid5", |
2604b703b
|
5560 |
.level = 5, |
1da177e4c
|
5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 |
.owner = THIS_MODULE, .make_request = make_request, .run = run, .stop = stop, .status = status, .error_handler = error, .hot_add_disk = raid5_add_disk, .hot_remove_disk= raid5_remove_disk, .spare_active = raid5_spare_active, .sync_request = sync_request, .resize = raid5_resize, |
80c3a6ce4
|
5572 |
.size = raid5_size, |
63c70c4f3
|
5573 5574 |
.check_reshape = raid5_check_reshape, .start_reshape = raid5_start_reshape, |
cea9c2280
|
5575 |
.finish_reshape = raid5_finish_reshape, |
72626685d
|
5576 |
.quiesce = raid5_quiesce, |
d562b0c43
|
5577 |
.takeover = raid5_takeover, |
1da177e4c
|
5578 |
}; |
84fc4b56d
|
5579 |
static struct md_personality raid4_personality = |
1da177e4c
|
5580 |
{ |
2604b703b
|
5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 |
.name = "raid4", .level = 4, .owner = THIS_MODULE, .make_request = make_request, .run = run, .stop = stop, .status = status, .error_handler = error, .hot_add_disk = raid5_add_disk, .hot_remove_disk= raid5_remove_disk, .spare_active = raid5_spare_active, .sync_request = sync_request, .resize = raid5_resize, |
80c3a6ce4
|
5594 |
.size = raid5_size, |
3d37890ba
|
5595 5596 |
.check_reshape = raid5_check_reshape, .start_reshape = raid5_start_reshape, |
cea9c2280
|
5597 |
.finish_reshape = raid5_finish_reshape, |
2604b703b
|
5598 |
.quiesce = raid5_quiesce, |
a78d38a1a
|
5599 |
.takeover = raid4_takeover, |
2604b703b
|
5600 5601 5602 5603 |
}; static int __init raid5_init(void) { |
16a53ecc3
|
5604 |
register_md_personality(&raid6_personality); |
2604b703b
|
5605 5606 5607 |
register_md_personality(&raid5_personality); register_md_personality(&raid4_personality); return 0; |
1da177e4c
|
5608 |
} |
2604b703b
|
5609 |
static void raid5_exit(void) |
1da177e4c
|
5610 |
{ |
16a53ecc3
|
5611 |
unregister_md_personality(&raid6_personality); |
2604b703b
|
5612 5613 |
unregister_md_personality(&raid5_personality); unregister_md_personality(&raid4_personality); |
1da177e4c
|
5614 5615 5616 5617 5618 |
} module_init(raid5_init); module_exit(raid5_exit); MODULE_LICENSE("GPL"); |
0efb9e619
|
5619 |
MODULE_DESCRIPTION("RAID4/5/6 (striping with parity) personality for MD"); |
1da177e4c
|
5620 |
MODULE_ALIAS("md-personality-4"); /* RAID5 */ |
d9d166c2a
|
5621 5622 |
MODULE_ALIAS("md-raid5"); MODULE_ALIAS("md-raid4"); |
2604b703b
|
5623 5624 |
MODULE_ALIAS("md-level-5"); MODULE_ALIAS("md-level-4"); |
16a53ecc3
|
5625 5626 5627 5628 5629 5630 5631 |
MODULE_ALIAS("md-personality-8"); /* RAID6 */ MODULE_ALIAS("md-raid6"); MODULE_ALIAS("md-level-6"); /* This used to be two separate modules, they were: */ MODULE_ALIAS("raid5"); MODULE_ALIAS("raid6"); |