Blame view
drivers/md/raid1.c
76.9 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 3 4 5 6 7 8 9 10 11 |
/* * raid1.c : Multiple Devices driver for Linux * * Copyright (C) 1999, 2000, 2001 Ingo Molnar, Red Hat * * Copyright (C) 1996, 1997, 1998 Ingo Molnar, Miguel de Icaza, Gadi Oxman * * RAID-1 management functions. * * Better read-balancing code written by Mika Kuoppala <miku@iki.fi>, 2000 * |
96de0e252 Convert files to ... |
12 |
* Fixes to reconstruction by Jakob Østergaard" <jakob@ostenfeld.dk> |
1da177e4c Linux-2.6.12-rc2 |
13 14 |
* Various fixes by Neil Brown <neilb@cse.unsw.edu.au> * |
191ea9b2c [PATCH] md: raid1... |
15 16 17 18 19 20 21 22 23 |
* Changes by Peter T. Breuer <ptb@it.uc3m.es> 31/1/2003 to support * bitmapped intelligence in resync: * * - bitmap marked during normal i/o * - bitmap used to skip nondirty blocks during sync * * Additions to bitmap code, (C) 2003-2004 Paul Clements, SteelEye Technology: * - persistent bitmap code * |
1da177e4c Linux-2.6.12-rc2 |
24 25 26 27 28 29 30 31 32 |
* 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. */ |
5a0e3ad6a include cleanup: ... |
33 |
#include <linux/slab.h> |
255707274 md: build failure... |
34 |
#include <linux/delay.h> |
bff61975b md: move lots of ... |
35 |
#include <linux/blkdev.h> |
056075c76 md: Add module.h ... |
36 |
#include <linux/module.h> |
bff61975b md: move lots of ... |
37 |
#include <linux/seq_file.h> |
8bda470e8 md/raid: use prin... |
38 |
#include <linux/ratelimit.h> |
43b2e5d86 md: move md_k.h f... |
39 |
#include "md.h" |
ef740c372 md: move headers ... |
40 41 |
#include "raid1.h" #include "bitmap.h" |
191ea9b2c [PATCH] md: raid1... |
42 |
|
1da177e4c Linux-2.6.12-rc2 |
43 44 45 46 |
/* * Number of guaranteed r1bios in case of extreme VM load: */ #define NR_RAID1_BIOS 256 |
34db0cd60 md: add proper wr... |
47 48 49 50 51 |
/* When there are this many requests queue to be written by * the raid1 thread, we become 'congested' to provide back-pressure * for writeback. */ static int max_queued_requests = 1024; |
1da177e4c Linux-2.6.12-rc2 |
52 |
|
e80963604 md/raid1: typedef... |
53 54 |
static void allow_barrier(struct r1conf *conf); static void lower_barrier(struct r1conf *conf); |
1da177e4c Linux-2.6.12-rc2 |
55 |
|
dd0fc66fb [PATCH] gfp flags... |
56 |
static void * r1bio_pool_alloc(gfp_t gfp_flags, void *data) |
1da177e4c Linux-2.6.12-rc2 |
57 58 |
{ struct pool_info *pi = data; |
9f2c9d12b md: remove typede... |
59 |
int size = offsetof(struct r1bio, bios[pi->raid_disks]); |
1da177e4c Linux-2.6.12-rc2 |
60 61 |
/* allocate a r1bio with room for raid_disks entries in the bios array */ |
7eaceacca block: remove per... |
62 |
return kzalloc(size, gfp_flags); |
1da177e4c Linux-2.6.12-rc2 |
63 64 65 66 67 68 69 70 71 72 73 74 |
} static void r1bio_pool_free(void *r1_bio, void *data) { kfree(r1_bio); } #define RESYNC_BLOCK_SIZE (64*1024) //#define RESYNC_BLOCK_SIZE PAGE_SIZE #define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9) #define RESYNC_PAGES ((RESYNC_BLOCK_SIZE + PAGE_SIZE-1) / PAGE_SIZE) #define RESYNC_WINDOW (2048*1024) |
dd0fc66fb [PATCH] gfp flags... |
75 |
static void * r1buf_pool_alloc(gfp_t gfp_flags, void *data) |
1da177e4c Linux-2.6.12-rc2 |
76 77 78 |
{ struct pool_info *pi = data; struct page *page; |
9f2c9d12b md: remove typede... |
79 |
struct r1bio *r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
80 81 82 83 |
struct bio *bio; int i, j; r1_bio = r1bio_pool_alloc(gfp_flags, pi); |
7eaceacca block: remove per... |
84 |
if (!r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
85 |
return NULL; |
1da177e4c Linux-2.6.12-rc2 |
86 87 88 89 90 |
/* * Allocate bios : 1 for reading, n-1 for writing */ for (j = pi->raid_disks ; j-- ; ) { |
6746557f0 md: use bio_kmall... |
91 |
bio = bio_kmalloc(gfp_flags, RESYNC_PAGES); |
1da177e4c Linux-2.6.12-rc2 |
92 93 94 95 96 97 |
if (!bio) goto out_free_bio; r1_bio->bios[j] = bio; } /* * Allocate RESYNC_PAGES data pages and attach them to |
d11c171e6 [PATCH] md: allow... |
98 99 100 |
* the first bio. * If this is a user-requested check/repair, allocate * RESYNC_PAGES for each bio. |
1da177e4c Linux-2.6.12-rc2 |
101 |
*/ |
d11c171e6 [PATCH] md: allow... |
102 103 104 105 106 107 108 109 110 111 112 113 |
if (test_bit(MD_RECOVERY_REQUESTED, &pi->mddev->recovery)) j = pi->raid_disks; else j = 1; while(j--) { bio = r1_bio->bios[j]; for (i = 0; i < RESYNC_PAGES; i++) { page = alloc_page(gfp_flags); if (unlikely(!page)) goto out_free_pages; bio->bi_io_vec[i].bv_page = page; |
303a0e11d md/raid1 - don't ... |
114 |
bio->bi_vcnt = i+1; |
d11c171e6 [PATCH] md: allow... |
115 116 117 118 119 120 121 122 |
} } /* If not user-requests, copy the page pointers to all bios */ if (!test_bit(MD_RECOVERY_REQUESTED, &pi->mddev->recovery)) { for (i=0; i<RESYNC_PAGES ; i++) for (j=1; j<pi->raid_disks; j++) r1_bio->bios[j]->bi_io_vec[i].bv_page = r1_bio->bios[0]->bi_io_vec[i].bv_page; |
1da177e4c Linux-2.6.12-rc2 |
123 124 125 126 127 128 129 |
} r1_bio->master_bio = NULL; return r1_bio; out_free_pages: |
303a0e11d md/raid1 - don't ... |
130 131 132 |
for (j=0 ; j < pi->raid_disks; j++) for (i=0; i < r1_bio->bios[j]->bi_vcnt ; i++) put_page(r1_bio->bios[j]->bi_io_vec[i].bv_page); |
d11c171e6 [PATCH] md: allow... |
133 |
j = -1; |
1da177e4c Linux-2.6.12-rc2 |
134 |
out_free_bio: |
8f19ccb2f md/raid1: Allocat... |
135 |
while (++j < pi->raid_disks) |
1da177e4c Linux-2.6.12-rc2 |
136 137 138 139 140 141 142 143 |
bio_put(r1_bio->bios[j]); r1bio_pool_free(r1_bio, data); return NULL; } static void r1buf_pool_free(void *__r1_bio, void *data) { struct pool_info *pi = data; |
d11c171e6 [PATCH] md: allow... |
144 |
int i,j; |
9f2c9d12b md: remove typede... |
145 |
struct r1bio *r1bio = __r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
146 |
|
d11c171e6 [PATCH] md: allow... |
147 148 149 150 151 |
for (i = 0; i < RESYNC_PAGES; i++) for (j = pi->raid_disks; j-- ;) { if (j == 0 || r1bio->bios[j]->bi_io_vec[i].bv_page != r1bio->bios[0]->bi_io_vec[i].bv_page) |
1345b1d8a [PATCH] md: defin... |
152 |
safe_put_page(r1bio->bios[j]->bi_io_vec[i].bv_page); |
d11c171e6 [PATCH] md: allow... |
153 |
} |
1da177e4c Linux-2.6.12-rc2 |
154 155 156 157 158 |
for (i=0 ; i < pi->raid_disks; i++) bio_put(r1bio->bios[i]); r1bio_pool_free(r1bio, data); } |
e80963604 md/raid1: typedef... |
159 |
static void put_all_bios(struct r1conf *conf, struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
160 161 |
{ int i; |
8f19ccb2f md/raid1: Allocat... |
162 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
1da177e4c Linux-2.6.12-rc2 |
163 |
struct bio **bio = r1_bio->bios + i; |
4367af556 md/raid1: clear b... |
164 |
if (!BIO_SPECIAL(*bio)) |
1da177e4c Linux-2.6.12-rc2 |
165 166 167 168 |
bio_put(*bio); *bio = NULL; } } |
9f2c9d12b md: remove typede... |
169 |
static void free_r1bio(struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
170 |
{ |
e80963604 md/raid1: typedef... |
171 |
struct r1conf *conf = r1_bio->mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
172 |
|
1da177e4c Linux-2.6.12-rc2 |
173 174 175 |
put_all_bios(conf, r1_bio); mempool_free(r1_bio, conf->r1bio_pool); } |
9f2c9d12b md: remove typede... |
176 |
static void put_buf(struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
177 |
{ |
e80963604 md/raid1: typedef... |
178 |
struct r1conf *conf = r1_bio->mddev->private; |
3e198f782 [PATCH] md: tidyu... |
179 |
int i; |
8f19ccb2f md/raid1: Allocat... |
180 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
3e198f782 [PATCH] md: tidyu... |
181 182 183 184 |
struct bio *bio = r1_bio->bios[i]; if (bio->bi_end_io) rdev_dec_pending(conf->mirrors[i].rdev, r1_bio->mddev); } |
1da177e4c Linux-2.6.12-rc2 |
185 186 |
mempool_free(r1_bio, conf->r1buf_pool); |
17999be4a [PATCH] md: impro... |
187 |
lower_barrier(conf); |
1da177e4c Linux-2.6.12-rc2 |
188 |
} |
9f2c9d12b md: remove typede... |
189 |
static void reschedule_retry(struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
190 191 |
{ unsigned long flags; |
fd01b88c7 md: remove typede... |
192 |
struct mddev *mddev = r1_bio->mddev; |
e80963604 md/raid1: typedef... |
193 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
194 195 196 |
spin_lock_irqsave(&conf->device_lock, flags); list_add(&r1_bio->retry_list, &conf->retry_list); |
ddaf22aba [PATCH] md: attem... |
197 |
conf->nr_queued ++; |
1da177e4c Linux-2.6.12-rc2 |
198 |
spin_unlock_irqrestore(&conf->device_lock, flags); |
17999be4a [PATCH] md: impro... |
199 |
wake_up(&conf->wait_barrier); |
1da177e4c Linux-2.6.12-rc2 |
200 201 202 203 204 205 206 207 |
md_wakeup_thread(mddev->thread); } /* * raid_end_bio_io() is called when we have finished servicing a mirrored * operation and are ready to return a success/failure code to the buffer * cache layer. */ |
9f2c9d12b md: remove typede... |
208 |
static void call_bio_endio(struct r1bio *r1_bio) |
d2eb35acf md/raid1: avoid r... |
209 210 211 |
{ struct bio *bio = r1_bio->master_bio; int done; |
e80963604 md/raid1: typedef... |
212 |
struct r1conf *conf = r1_bio->mddev->private; |
d2eb35acf md/raid1: avoid r... |
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 |
if (bio->bi_phys_segments) { unsigned long flags; spin_lock_irqsave(&conf->device_lock, flags); bio->bi_phys_segments--; done = (bio->bi_phys_segments == 0); spin_unlock_irqrestore(&conf->device_lock, flags); } else done = 1; if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) clear_bit(BIO_UPTODATE, &bio->bi_flags); if (done) { bio_endio(bio, 0); /* * Wake up any possible resync thread that waits for the device * to go idle. */ allow_barrier(conf); } } |
9f2c9d12b md: remove typede... |
234 |
static void raid_end_bio_io(struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
235 236 |
{ struct bio *bio = r1_bio->master_bio; |
4b6d287f6 [PATCH] md: add w... |
237 238 |
/* if nobody has done the final endio yet, do it now */ if (!test_and_set_bit(R1BIO_Returned, &r1_bio->state)) { |
36a4e1fe0 md: remove PRINTK... |
239 240 241 242 243 244 |
pr_debug("raid1: sync end %s on sectors %llu-%llu ", (bio_data_dir(bio) == WRITE) ? "write" : "read", (unsigned long long) bio->bi_sector, (unsigned long long) bio->bi_sector + (bio->bi_size >> 9) - 1); |
4b6d287f6 [PATCH] md: add w... |
245 |
|
d2eb35acf md/raid1: avoid r... |
246 |
call_bio_endio(r1_bio); |
4b6d287f6 [PATCH] md: add w... |
247 |
} |
1da177e4c Linux-2.6.12-rc2 |
248 249 250 251 252 253 |
free_r1bio(r1_bio); } /* * Update disk head position estimator based on IRQ completion info. */ |
9f2c9d12b md: remove typede... |
254 |
static inline void update_head_pos(int disk, struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
255 |
{ |
e80963604 md/raid1: typedef... |
256 |
struct r1conf *conf = r1_bio->mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
257 258 259 260 |
conf->mirrors[disk].head_position = r1_bio->sector + (r1_bio->sectors); } |
ba3ae3bee md/raid1: factor ... |
261 262 263 |
/* * Find the disk number which triggered given bio */ |
9f2c9d12b md: remove typede... |
264 |
static int find_bio_disk(struct r1bio *r1_bio, struct bio *bio) |
ba3ae3bee md/raid1: factor ... |
265 266 |
{ int mirror; |
301946364 md/raid1: Replac... |
267 268 |
struct r1conf *conf = r1_bio->mddev->private; int raid_disks = conf->raid_disks; |
ba3ae3bee md/raid1: factor ... |
269 |
|
8f19ccb2f md/raid1: Allocat... |
270 |
for (mirror = 0; mirror < raid_disks * 2; mirror++) |
ba3ae3bee md/raid1: factor ... |
271 272 |
if (r1_bio->bios[mirror] == bio) break; |
8f19ccb2f md/raid1: Allocat... |
273 |
BUG_ON(mirror == raid_disks * 2); |
ba3ae3bee md/raid1: factor ... |
274 275 276 277 |
update_head_pos(mirror, r1_bio); return mirror; } |
6712ecf8f Drop 'size' argum... |
278 |
static void raid1_end_read_request(struct bio *bio, int error) |
1da177e4c Linux-2.6.12-rc2 |
279 280 |
{ int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); |
9f2c9d12b md: remove typede... |
281 |
struct r1bio *r1_bio = bio->bi_private; |
1da177e4c Linux-2.6.12-rc2 |
282 |
int mirror; |
e80963604 md/raid1: typedef... |
283 |
struct r1conf *conf = r1_bio->mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
284 |
|
1da177e4c Linux-2.6.12-rc2 |
285 286 287 288 |
mirror = r1_bio->read_disk; /* * this branch is our 'one mirror IO has finished' event handler: */ |
ddaf22aba [PATCH] md: attem... |
289 |
update_head_pos(mirror, r1_bio); |
dd00a99e7 md: avoid a possi... |
290 291 292 293 294 295 |
if (uptodate) set_bit(R1BIO_Uptodate, &r1_bio->state); else { /* If all other devices have failed, we want to return * the error upwards rather than fail the last device. * Here we redefine "uptodate" to mean "Don't want to retry" |
1da177e4c Linux-2.6.12-rc2 |
296 |
*/ |
dd00a99e7 md: avoid a possi... |
297 298 299 300 301 302 303 304 |
unsigned long flags; spin_lock_irqsave(&conf->device_lock, flags); if (r1_bio->mddev->degraded == conf->raid_disks || (r1_bio->mddev->degraded == conf->raid_disks-1 && !test_bit(Faulty, &conf->mirrors[mirror].rdev->flags))) uptodate = 1; spin_unlock_irqrestore(&conf->device_lock, flags); } |
1da177e4c Linux-2.6.12-rc2 |
305 |
|
dd00a99e7 md: avoid a possi... |
306 |
if (uptodate) |
1da177e4c Linux-2.6.12-rc2 |
307 |
raid_end_bio_io(r1_bio); |
dd00a99e7 md: avoid a possi... |
308 |
else { |
1da177e4c Linux-2.6.12-rc2 |
309 310 311 312 |
/* * oops, read error: */ char b[BDEVNAME_SIZE]; |
8bda470e8 md/raid: use prin... |
313 314 315 316 317 318 319 320 |
printk_ratelimited( KERN_ERR "md/raid1:%s: %s: " "rescheduling sector %llu ", mdname(conf->mddev), bdevname(conf->mirrors[mirror].rdev->bdev, b), (unsigned long long)r1_bio->sector); |
d2eb35acf md/raid1: avoid r... |
321 |
set_bit(R1BIO_ReadError, &r1_bio->state); |
1da177e4c Linux-2.6.12-rc2 |
322 323 324 325 |
reschedule_retry(r1_bio); } rdev_dec_pending(conf->mirrors[mirror].rdev, conf->mddev); |
1da177e4c Linux-2.6.12-rc2 |
326 |
} |
9f2c9d12b md: remove typede... |
327 |
static void close_write(struct r1bio *r1_bio) |
cd5ff9a16 md/raid1: Handle... |
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 |
{ /* it really is the end of this request */ if (test_bit(R1BIO_BehindIO, &r1_bio->state)) { /* free extra copy of the data pages */ int i = r1_bio->behind_page_count; while (i--) safe_put_page(r1_bio->behind_bvecs[i].bv_page); kfree(r1_bio->behind_bvecs); r1_bio->behind_bvecs = NULL; } /* clear the bitmap if all writes complete successfully */ bitmap_endwrite(r1_bio->mddev->bitmap, r1_bio->sector, r1_bio->sectors, !test_bit(R1BIO_Degraded, &r1_bio->state), test_bit(R1BIO_BehindIO, &r1_bio->state)); md_write_end(r1_bio->mddev); } |
9f2c9d12b md: remove typede... |
345 |
static void r1_bio_write_done(struct r1bio *r1_bio) |
4e78064f4 md: Fix possible ... |
346 |
{ |
cd5ff9a16 md/raid1: Handle... |
347 348 349 350 351 352 353 |
if (!atomic_dec_and_test(&r1_bio->remaining)) return; if (test_bit(R1BIO_WriteError, &r1_bio->state)) reschedule_retry(r1_bio); else { close_write(r1_bio); |
4367af556 md/raid1: clear b... |
354 355 356 357 |
if (test_bit(R1BIO_MadeGood, &r1_bio->state)) reschedule_retry(r1_bio); else raid_end_bio_io(r1_bio); |
4e78064f4 md: Fix possible ... |
358 359 |
} } |
6712ecf8f Drop 'size' argum... |
360 |
static void raid1_end_write_request(struct bio *bio, int error) |
1da177e4c Linux-2.6.12-rc2 |
361 362 |
{ int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); |
9f2c9d12b md: remove typede... |
363 |
struct r1bio *r1_bio = bio->bi_private; |
a9701a304 [PATCH] md: suppo... |
364 |
int mirror, behind = test_bit(R1BIO_BehindIO, &r1_bio->state); |
e80963604 md/raid1: typedef... |
365 |
struct r1conf *conf = r1_bio->mddev->private; |
04b857f74 [PATCH] md: Fix s... |
366 |
struct bio *to_put = NULL; |
1da177e4c Linux-2.6.12-rc2 |
367 |
|
ba3ae3bee md/raid1: factor ... |
368 |
mirror = find_bio_disk(r1_bio, bio); |
1da177e4c Linux-2.6.12-rc2 |
369 |
|
e9c7469bb md: implment REQ_... |
370 371 372 |
/* * 'one mirror IO has finished' event handler: */ |
e9c7469bb md: implment REQ_... |
373 |
if (!uptodate) { |
cd5ff9a16 md/raid1: Handle... |
374 375 |
set_bit(WriteErrorSeen, &conf->mirrors[mirror].rdev->flags); |
19d671695 md/raid1: Mark de... |
376 377 378 379 |
if (!test_and_set_bit(WantReplacement, &conf->mirrors[mirror].rdev->flags)) set_bit(MD_RECOVERY_NEEDED, & conf->mddev->recovery); |
cd5ff9a16 md/raid1: Handle... |
380 |
set_bit(R1BIO_WriteError, &r1_bio->state); |
4367af556 md/raid1: clear b... |
381 |
} else { |
1da177e4c Linux-2.6.12-rc2 |
382 |
/* |
e9c7469bb md: implment REQ_... |
383 384 385 386 387 388 389 390 |
* Set R1BIO_Uptodate in our master bio, so that we * will return a good error code for to the higher * levels even if IO on some other mirrored buffer * fails. * * The 'master' represents the composite IO operation * to user-side. So if something waits for IO, then it * will wait for the 'master' bio. |
1da177e4c Linux-2.6.12-rc2 |
391 |
*/ |
4367af556 md/raid1: clear b... |
392 393 |
sector_t first_bad; int bad_sectors; |
cd5ff9a16 md/raid1: Handle... |
394 395 |
r1_bio->bios[mirror] = NULL; to_put = bio; |
e9c7469bb md: implment REQ_... |
396 |
set_bit(R1BIO_Uptodate, &r1_bio->state); |
4367af556 md/raid1: clear b... |
397 398 399 400 401 402 403 404 |
/* Maybe we can clear some bad blocks. */ if (is_badblock(conf->mirrors[mirror].rdev, r1_bio->sector, r1_bio->sectors, &first_bad, &bad_sectors)) { r1_bio->bios[mirror] = IO_MADE_GOOD; set_bit(R1BIO_MadeGood, &r1_bio->state); } } |
e9c7469bb md: implment REQ_... |
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 |
if (behind) { if (test_bit(WriteMostly, &conf->mirrors[mirror].rdev->flags)) atomic_dec(&r1_bio->behind_remaining); /* * In behind mode, we ACK the master bio once the I/O * has safely reached all non-writemostly * disks. Setting the Returned bit ensures that this * gets done only once -- we don't ever want to return * -EIO here, instead we'll wait */ if (atomic_read(&r1_bio->behind_remaining) >= (atomic_read(&r1_bio->remaining)-1) && test_bit(R1BIO_Uptodate, &r1_bio->state)) { /* Maybe we can return now */ if (!test_and_set_bit(R1BIO_Returned, &r1_bio->state)) { struct bio *mbio = r1_bio->master_bio; |
36a4e1fe0 md: remove PRINTK... |
421 422 423 424 425 426 |
pr_debug("raid1: behind end write sectors" " %llu-%llu ", (unsigned long long) mbio->bi_sector, (unsigned long long) mbio->bi_sector + (mbio->bi_size >> 9) - 1); |
d2eb35acf md/raid1: avoid r... |
427 |
call_bio_endio(r1_bio); |
4b6d287f6 [PATCH] md: add w... |
428 429 430 |
} } } |
4367af556 md/raid1: clear b... |
431 432 433 |
if (r1_bio->bios[mirror] == NULL) rdev_dec_pending(conf->mirrors[mirror].rdev, conf->mddev); |
e9c7469bb md: implment REQ_... |
434 |
|
1da177e4c Linux-2.6.12-rc2 |
435 |
/* |
1da177e4c Linux-2.6.12-rc2 |
436 437 438 |
* Let's see if all mirrored write operations have finished * already. */ |
af6d7b760 md/raid1: improve... |
439 |
r1_bio_write_done(r1_bio); |
c70810b32 [PATCH] md: refor... |
440 |
|
04b857f74 [PATCH] md: Fix s... |
441 442 |
if (to_put) bio_put(to_put); |
1da177e4c Linux-2.6.12-rc2 |
443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 |
} /* * This routine returns the disk from which the requested read should * be done. There is a per-array 'next expected sequential IO' sector * number - if this matches on the next IO then we use the last disk. * There is also a per-disk 'last know head position' sector that is * maintained from IRQ contexts, both the normal and the resync IO * completion handlers update this position correctly. If there is no * perfect sequential match then we pick the disk whose head is closest. * * If there are 2 mirrors in the same 2 devices, performance degrades * because position is mirror, not device based. * * The rdev for the device selected will have nr_pending incremented. */ |
e80963604 md/raid1: typedef... |
460 |
static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sectors) |
1da177e4c Linux-2.6.12-rc2 |
461 |
{ |
af3a2cd6b md: Fix read bala... |
462 |
const sector_t this_sector = r1_bio->sector; |
d2eb35acf md/raid1: avoid r... |
463 464 |
int sectors; int best_good_sectors; |
f3ac8bf7c md: tidy up devic... |
465 |
int start_disk; |
76073054c md/raid1: clean u... |
466 |
int best_disk; |
f3ac8bf7c md: tidy up devic... |
467 |
int i; |
76073054c md/raid1: clean u... |
468 |
sector_t best_dist; |
3cb030020 md: removing type... |
469 |
struct md_rdev *rdev; |
f3ac8bf7c md: tidy up devic... |
470 |
int choose_first; |
1da177e4c Linux-2.6.12-rc2 |
471 472 473 |
rcu_read_lock(); /* |
8ddf9efe6 [PATCH] md: suppo... |
474 |
* Check if we can balance. We can balance on the whole |
1da177e4c Linux-2.6.12-rc2 |
475 476 477 478 |
* device if no resync is going on, or below the resync window. * We take the first readable disk when above the resync window. */ retry: |
d2eb35acf md/raid1: avoid r... |
479 |
sectors = r1_bio->sectors; |
76073054c md/raid1: clean u... |
480 481 |
best_disk = -1; best_dist = MaxSector; |
d2eb35acf md/raid1: avoid r... |
482 |
best_good_sectors = 0; |
1da177e4c Linux-2.6.12-rc2 |
483 484 |
if (conf->mddev->recovery_cp < MaxSector && (this_sector + sectors >= conf->next_resync)) { |
f3ac8bf7c md: tidy up devic... |
485 486 487 488 489 |
choose_first = 1; start_disk = 0; } else { choose_first = 0; start_disk = conf->last_used; |
1da177e4c Linux-2.6.12-rc2 |
490 |
} |
8f19ccb2f md/raid1: Allocat... |
491 |
for (i = 0 ; i < conf->raid_disks * 2 ; i++) { |
76073054c md/raid1: clean u... |
492 |
sector_t dist; |
d2eb35acf md/raid1: avoid r... |
493 494 |
sector_t first_bad; int bad_sectors; |
f3ac8bf7c md: tidy up devic... |
495 496 497 498 499 500 501 |
int disk = start_disk + i; if (disk >= conf->raid_disks) disk -= conf->raid_disks; rdev = rcu_dereference(conf->mirrors[disk].rdev); if (r1_bio->bios[disk] == IO_BLOCKED || rdev == NULL |
76073054c md/raid1: clean u... |
502 |
|| test_bit(Faulty, &rdev->flags)) |
f3ac8bf7c md: tidy up devic... |
503 |
continue; |
76073054c md/raid1: clean u... |
504 505 |
if (!test_bit(In_sync, &rdev->flags) && rdev->recovery_offset < this_sector + sectors) |
1da177e4c Linux-2.6.12-rc2 |
506 |
continue; |
76073054c md/raid1: clean u... |
507 508 509 |
if (test_bit(WriteMostly, &rdev->flags)) { /* Don't balance among write-mostly, just * use the first as a last resort */ |
307729c8b md/raid1: perform... |
510 511 512 513 514 515 516 517 518 |
if (best_disk < 0) { if (is_badblock(rdev, this_sector, sectors, &first_bad, &bad_sectors)) { if (first_bad < this_sector) /* Cannot use this */ continue; best_good_sectors = first_bad - this_sector; } else best_good_sectors = sectors; |
76073054c md/raid1: clean u... |
519 |
best_disk = disk; |
307729c8b md/raid1: perform... |
520 |
} |
76073054c md/raid1: clean u... |
521 522 523 524 525 |
continue; } /* This is a reasonable device to use. It might * even be best. */ |
d2eb35acf md/raid1: avoid r... |
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 |
if (is_badblock(rdev, this_sector, sectors, &first_bad, &bad_sectors)) { if (best_dist < MaxSector) /* already have a better device */ continue; if (first_bad <= this_sector) { /* cannot read here. If this is the 'primary' * device, then we must not read beyond * bad_sectors from another device.. */ bad_sectors -= (this_sector - first_bad); if (choose_first && sectors > bad_sectors) sectors = bad_sectors; if (best_good_sectors > sectors) best_good_sectors = sectors; } else { sector_t good_sectors = first_bad - this_sector; if (good_sectors > best_good_sectors) { best_good_sectors = good_sectors; best_disk = disk; } if (choose_first) break; } continue; } else best_good_sectors = sectors; |
76073054c md/raid1: clean u... |
554 555 556 557 558 559 560 561 |
dist = abs(this_sector - conf->mirrors[disk].head_position); if (choose_first /* Don't change to another disk for sequential reads */ || conf->next_seq_sect == this_sector || dist == 0 /* If device is idle, use it */ || atomic_read(&rdev->nr_pending) == 0) { best_disk = disk; |
1da177e4c Linux-2.6.12-rc2 |
562 563 |
break; } |
76073054c md/raid1: clean u... |
564 565 566 |
if (dist < best_dist) { best_dist = dist; best_disk = disk; |
1da177e4c Linux-2.6.12-rc2 |
567 |
} |
f3ac8bf7c md: tidy up devic... |
568 |
} |
1da177e4c Linux-2.6.12-rc2 |
569 |
|
76073054c md/raid1: clean u... |
570 571 |
if (best_disk >= 0) { rdev = rcu_dereference(conf->mirrors[best_disk].rdev); |
8ddf9efe6 [PATCH] md: suppo... |
572 573 574 |
if (!rdev) goto retry; atomic_inc(&rdev->nr_pending); |
76073054c md/raid1: clean u... |
575 |
if (test_bit(Faulty, &rdev->flags)) { |
1da177e4c Linux-2.6.12-rc2 |
576 577 578 |
/* cannot risk returning a device that failed * before we inc'ed nr_pending */ |
03c902e17 [PATCH] md: fix r... |
579 |
rdev_dec_pending(rdev, conf->mddev); |
1da177e4c Linux-2.6.12-rc2 |
580 581 |
goto retry; } |
d2eb35acf md/raid1: avoid r... |
582 |
sectors = best_good_sectors; |
8ddf9efe6 [PATCH] md: suppo... |
583 |
conf->next_seq_sect = this_sector + sectors; |
76073054c md/raid1: clean u... |
584 |
conf->last_used = best_disk; |
1da177e4c Linux-2.6.12-rc2 |
585 586 |
} rcu_read_unlock(); |
d2eb35acf md/raid1: avoid r... |
587 |
*max_sectors = sectors; |
1da177e4c Linux-2.6.12-rc2 |
588 |
|
76073054c md/raid1: clean u... |
589 |
return best_disk; |
1da177e4c Linux-2.6.12-rc2 |
590 |
} |
fd01b88c7 md: remove typede... |
591 |
int md_raid1_congested(struct mddev *mddev, int bits) |
0d1292282 [PATCH] md: defin... |
592 |
{ |
e80963604 md/raid1: typedef... |
593 |
struct r1conf *conf = mddev->private; |
0d1292282 [PATCH] md: defin... |
594 |
int i, ret = 0; |
34db0cd60 md: add proper wr... |
595 596 597 |
if ((bits & (1 << BDI_async_congested)) && conf->pending_count >= max_queued_requests) return 1; |
0d1292282 [PATCH] md: defin... |
598 |
rcu_read_lock(); |
301946364 md/raid1: Replac... |
599 |
for (i = 0; i < conf->raid_disks; i++) { |
3cb030020 md: removing type... |
600 |
struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); |
0d1292282 [PATCH] md: defin... |
601 |
if (rdev && !test_bit(Faulty, &rdev->flags)) { |
165125e1e [BLOCK] Get rid o... |
602 |
struct request_queue *q = bdev_get_queue(rdev->bdev); |
0d1292282 [PATCH] md: defin... |
603 |
|
1ed7242e5 MD: raid1 changes... |
604 |
BUG_ON(!q); |
0d1292282 [PATCH] md: defin... |
605 606 607 |
/* Note the '|| 1' - when read_balance prefers * non-congested targets, it can be removed */ |
91a9e99d7 md/raid1: fix bui... |
608 |
if ((bits & (1<<BDI_async_congested)) || 1) |
0d1292282 [PATCH] md: defin... |
609 610 611 612 613 614 615 616 |
ret |= bdi_congested(&q->backing_dev_info, bits); else ret &= bdi_congested(&q->backing_dev_info, bits); } } rcu_read_unlock(); return ret; } |
1ed7242e5 MD: raid1 changes... |
617 |
EXPORT_SYMBOL_GPL(md_raid1_congested); |
0d1292282 [PATCH] md: defin... |
618 |
|
1ed7242e5 MD: raid1 changes... |
619 620 |
static int raid1_congested(void *data, int bits) { |
fd01b88c7 md: remove typede... |
621 |
struct mddev *mddev = data; |
1ed7242e5 MD: raid1 changes... |
622 623 624 625 |
return mddev_congested(mddev, bits) || md_raid1_congested(mddev, bits); } |
0d1292282 [PATCH] md: defin... |
626 |
|
e80963604 md/raid1: typedef... |
627 |
static void flush_pending_writes(struct r1conf *conf) |
a35e63efa md: fix deadlock ... |
628 629 630 |
{ /* Any writes that have been queued but are awaiting * bitmap updates get flushed here. |
a35e63efa md: fix deadlock ... |
631 |
*/ |
a35e63efa md: fix deadlock ... |
632 633 634 635 636 |
spin_lock_irq(&conf->device_lock); if (conf->pending_bio_list.head) { struct bio *bio; bio = bio_list_get(&conf->pending_bio_list); |
34db0cd60 md: add proper wr... |
637 |
conf->pending_count = 0; |
a35e63efa md: fix deadlock ... |
638 639 640 641 |
spin_unlock_irq(&conf->device_lock); /* flush any pending bitmap writes to * disk before proceeding w/ I/O */ bitmap_unplug(conf->mddev->bitmap); |
34db0cd60 md: add proper wr... |
642 |
wake_up(&conf->wait_barrier); |
a35e63efa md: fix deadlock ... |
643 644 645 646 647 648 649 |
while (bio) { /* submit pending writes */ struct bio *next = bio->bi_next; bio->bi_next = NULL; generic_make_request(bio); bio = next; } |
a35e63efa md: fix deadlock ... |
650 651 |
} else spin_unlock_irq(&conf->device_lock); |
7eaceacca block: remove per... |
652 |
} |
17999be4a [PATCH] md: impro... |
653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 |
/* Barriers.... * Sometimes we need to suspend IO while we do something else, * either some resync/recovery, or reconfigure the array. * To do this we raise a 'barrier'. * The 'barrier' is a counter that can be raised multiple times * to count how many activities are happening which preclude * normal IO. * We can only raise the barrier if there is no pending IO. * i.e. if nr_pending == 0. * We choose only to raise the barrier if no-one is waiting for the * barrier to go down. This means that as soon as an IO request * is ready, no other operations which require a barrier will start * until the IO request has had a chance. * * So: regular IO calls 'wait_barrier'. When that returns there * is no backgroup IO happening, It must arrange to call * allow_barrier when it has finished its IO. * backgroup IO calls must call raise_barrier. Once that returns * there is no normal IO happeing. It must arrange to call * lower_barrier when the particular background IO completes. |
1da177e4c Linux-2.6.12-rc2 |
673 674 |
*/ #define RESYNC_DEPTH 32 |
e80963604 md/raid1: typedef... |
675 |
static void raise_barrier(struct r1conf *conf) |
1da177e4c Linux-2.6.12-rc2 |
676 677 |
{ spin_lock_irq(&conf->resync_lock); |
17999be4a [PATCH] md: impro... |
678 679 680 |
/* Wait until no block IO is waiting */ wait_event_lock_irq(conf->wait_barrier, !conf->nr_waiting, |
c3b328ac8 md: fix up raid1/... |
681 |
conf->resync_lock, ); |
17999be4a [PATCH] md: impro... |
682 683 684 |
/* block any new IO from starting */ conf->barrier++; |
046abeede md/raid1: fix som... |
685 |
/* Now wait for all pending IO to complete */ |
17999be4a [PATCH] md: impro... |
686 687 |
wait_event_lock_irq(conf->wait_barrier, !conf->nr_pending && conf->barrier < RESYNC_DEPTH, |
c3b328ac8 md: fix up raid1/... |
688 |
conf->resync_lock, ); |
17999be4a [PATCH] md: impro... |
689 690 691 |
spin_unlock_irq(&conf->resync_lock); } |
e80963604 md/raid1: typedef... |
692 |
static void lower_barrier(struct r1conf *conf) |
17999be4a [PATCH] md: impro... |
693 694 |
{ unsigned long flags; |
709ae4879 md/raid1: add tak... |
695 |
BUG_ON(conf->barrier <= 0); |
17999be4a [PATCH] md: impro... |
696 697 698 699 700 |
spin_lock_irqsave(&conf->resync_lock, flags); conf->barrier--; spin_unlock_irqrestore(&conf->resync_lock, flags); wake_up(&conf->wait_barrier); } |
e80963604 md/raid1: typedef... |
701 |
static void wait_barrier(struct r1conf *conf) |
17999be4a [PATCH] md: impro... |
702 703 704 705 706 707 |
{ spin_lock_irq(&conf->resync_lock); if (conf->barrier) { conf->nr_waiting++; wait_event_lock_irq(conf->wait_barrier, !conf->barrier, conf->resync_lock, |
c3b328ac8 md: fix up raid1/... |
708 |
); |
17999be4a [PATCH] md: impro... |
709 |
conf->nr_waiting--; |
1da177e4c Linux-2.6.12-rc2 |
710 |
} |
17999be4a [PATCH] md: impro... |
711 |
conf->nr_pending++; |
1da177e4c Linux-2.6.12-rc2 |
712 713 |
spin_unlock_irq(&conf->resync_lock); } |
e80963604 md/raid1: typedef... |
714 |
static void allow_barrier(struct r1conf *conf) |
17999be4a [PATCH] md: impro... |
715 716 717 718 719 720 721 |
{ unsigned long flags; spin_lock_irqsave(&conf->resync_lock, flags); conf->nr_pending--; spin_unlock_irqrestore(&conf->resync_lock, flags); wake_up(&conf->wait_barrier); } |
e80963604 md/raid1: typedef... |
722 |
static void freeze_array(struct r1conf *conf) |
ddaf22aba [PATCH] md: attem... |
723 724 725 726 |
{ /* stop syncio and normal IO and wait for everything to * go quite. * We increment barrier and nr_waiting, and then |
1c830532f md: fix possible ... |
727 728 729 730 731 732 733 734 |
* wait until nr_pending match nr_queued+1 * This is called in the context of one normal IO request * that has failed. Thus any sync request that might be pending * will be blocked by nr_pending, and we need to wait for * pending IO requests to complete or be queued for re-try. * Thus the number queued (nr_queued) plus this request (1) * must match the number of pending IOs (nr_pending) before * we continue. |
ddaf22aba [PATCH] md: attem... |
735 736 737 738 739 |
*/ spin_lock_irq(&conf->resync_lock); conf->barrier++; conf->nr_waiting++; wait_event_lock_irq(conf->wait_barrier, |
1c830532f md: fix possible ... |
740 |
conf->nr_pending == conf->nr_queued+1, |
ddaf22aba [PATCH] md: attem... |
741 |
conf->resync_lock, |
c3b328ac8 md: fix up raid1/... |
742 |
flush_pending_writes(conf)); |
ddaf22aba [PATCH] md: attem... |
743 744 |
spin_unlock_irq(&conf->resync_lock); } |
e80963604 md/raid1: typedef... |
745 |
static void unfreeze_array(struct r1conf *conf) |
ddaf22aba [PATCH] md: attem... |
746 747 748 749 750 751 752 753 |
{ /* reverse the effect of the freeze */ spin_lock_irq(&conf->resync_lock); conf->barrier--; conf->nr_waiting--; wake_up(&conf->wait_barrier); spin_unlock_irq(&conf->resync_lock); } |
17999be4a [PATCH] md: impro... |
754 |
|
4e78064f4 md: Fix possible ... |
755 |
/* duplicate the data pages for behind I/O |
4e78064f4 md: Fix possible ... |
756 |
*/ |
9f2c9d12b md: remove typede... |
757 |
static void alloc_behind_pages(struct bio *bio, struct r1bio *r1_bio) |
4b6d287f6 [PATCH] md: add w... |
758 759 760 |
{ int i; struct bio_vec *bvec; |
2ca68f5ed md/raid1: store b... |
761 |
struct bio_vec *bvecs = kzalloc(bio->bi_vcnt * sizeof(struct bio_vec), |
4b6d287f6 [PATCH] md: add w... |
762 |
GFP_NOIO); |
2ca68f5ed md/raid1: store b... |
763 |
if (unlikely(!bvecs)) |
af6d7b760 md/raid1: improve... |
764 |
return; |
4b6d287f6 [PATCH] md: add w... |
765 |
|
4b6d287f6 [PATCH] md: add w... |
766 |
bio_for_each_segment(bvec, bio, i) { |
2ca68f5ed md/raid1: store b... |
767 768 769 |
bvecs[i] = *bvec; bvecs[i].bv_page = alloc_page(GFP_NOIO); if (unlikely(!bvecs[i].bv_page)) |
4b6d287f6 [PATCH] md: add w... |
770 |
goto do_sync_io; |
2ca68f5ed md/raid1: store b... |
771 772 773 |
memcpy(kmap(bvecs[i].bv_page) + bvec->bv_offset, kmap(bvec->bv_page) + bvec->bv_offset, bvec->bv_len); kunmap(bvecs[i].bv_page); |
4b6d287f6 [PATCH] md: add w... |
774 775 |
kunmap(bvec->bv_page); } |
2ca68f5ed md/raid1: store b... |
776 |
r1_bio->behind_bvecs = bvecs; |
af6d7b760 md/raid1: improve... |
777 778 779 |
r1_bio->behind_page_count = bio->bi_vcnt; set_bit(R1BIO_BehindIO, &r1_bio->state); return; |
4b6d287f6 [PATCH] md: add w... |
780 781 |
do_sync_io: |
af6d7b760 md/raid1: improve... |
782 |
for (i = 0; i < bio->bi_vcnt; i++) |
2ca68f5ed md/raid1: store b... |
783 784 785 |
if (bvecs[i].bv_page) put_page(bvecs[i].bv_page); kfree(bvecs); |
36a4e1fe0 md: remove PRINTK... |
786 787 |
pr_debug("%dB behind alloc failed, doing sync I/O ", bio->bi_size); |
4b6d287f6 [PATCH] md: add w... |
788 |
} |
b4fdcb02f Merge branch 'for... |
789 |
static void make_request(struct mddev *mddev, struct bio * bio) |
1da177e4c Linux-2.6.12-rc2 |
790 |
{ |
e80963604 md/raid1: typedef... |
791 |
struct r1conf *conf = mddev->private; |
0f6d02d58 md: remove typede... |
792 |
struct mirror_info *mirror; |
9f2c9d12b md: remove typede... |
793 |
struct r1bio *r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
794 |
struct bio *read_bio; |
1f68f0c4b md/raid1: avoid w... |
795 |
int i, disks; |
84255d101 md: fix possible ... |
796 |
struct bitmap *bitmap; |
191ea9b2c [PATCH] md: raid1... |
797 |
unsigned long flags; |
a362357b6 [BLOCK] Unify the... |
798 |
const int rw = bio_data_dir(bio); |
2c7d46ec1 md raid-1/10 Fix ... |
799 |
const unsigned long do_sync = (bio->bi_rw & REQ_SYNC); |
e9c7469bb md: implment REQ_... |
800 |
const unsigned long do_flush_fua = (bio->bi_rw & (REQ_FLUSH | REQ_FUA)); |
3cb030020 md: removing type... |
801 |
struct md_rdev *blocked_rdev; |
c3b328ac8 md: fix up raid1/... |
802 |
int plugged; |
1f68f0c4b md/raid1: avoid w... |
803 804 805 |
int first_clone; int sectors_handled; int max_sectors; |
191ea9b2c [PATCH] md: raid1... |
806 |
|
1da177e4c Linux-2.6.12-rc2 |
807 808 809 810 811 |
/* * Register the new request and wait if the reconstruction * thread has put up a bar for new requests. * Continue immediately if no resync is active currently. */ |
62de608da [PATCH] md: Impro... |
812 |
|
3d310eb7b [PATCH] md: fix d... |
813 |
md_write_start(mddev, bio); /* wait on superblock update early */ |
6eef4b21f md: add honouring... |
814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 |
if (bio_data_dir(bio) == WRITE && bio->bi_sector + bio->bi_size/512 > mddev->suspend_lo && bio->bi_sector < mddev->suspend_hi) { /* As the suspend_* range is controlled by * userspace, we want an interruptible * wait. */ DEFINE_WAIT(w); for (;;) { flush_signals(current); prepare_to_wait(&conf->wait_barrier, &w, TASK_INTERRUPTIBLE); if (bio->bi_sector + bio->bi_size/512 <= mddev->suspend_lo || bio->bi_sector >= mddev->suspend_hi) break; schedule(); } finish_wait(&conf->wait_barrier, &w); } |
62de608da [PATCH] md: Impro... |
833 |
|
17999be4a [PATCH] md: impro... |
834 |
wait_barrier(conf); |
1da177e4c Linux-2.6.12-rc2 |
835 |
|
84255d101 md: fix possible ... |
836 |
bitmap = mddev->bitmap; |
1da177e4c Linux-2.6.12-rc2 |
837 838 839 840 841 842 843 844 845 |
/* * make_request() can abort the operation when READA is being * used and no empty request is available. * */ r1_bio = mempool_alloc(conf->r1bio_pool, GFP_NOIO); r1_bio->master_bio = bio; r1_bio->sectors = bio->bi_size >> 9; |
191ea9b2c [PATCH] md: raid1... |
846 |
r1_bio->state = 0; |
1da177e4c Linux-2.6.12-rc2 |
847 848 |
r1_bio->mddev = mddev; r1_bio->sector = bio->bi_sector; |
d2eb35acf md/raid1: avoid r... |
849 850 851 852 853 854 855 856 857 |
/* We might need to issue multiple reads to different * devices if there are bad blocks around, so we keep * track of the number of reads in bio->bi_phys_segments. * If this is 0, there is only one r1_bio and no locking * will be needed when requests complete. If it is * non-zero, then it is the number of not-completed requests. */ bio->bi_phys_segments = 0; clear_bit(BIO_SEG_VALID, &bio->bi_flags); |
a362357b6 [BLOCK] Unify the... |
858 |
if (rw == READ) { |
1da177e4c Linux-2.6.12-rc2 |
859 860 861 |
/* * read balancing logic: */ |
d2eb35acf md/raid1: avoid r... |
862 863 864 865 |
int rdisk; read_again: rdisk = read_balance(conf, r1_bio, &max_sectors); |
1da177e4c Linux-2.6.12-rc2 |
866 867 868 869 |
if (rdisk < 0) { /* couldn't find anywhere to read from */ raid_end_bio_io(r1_bio); |
5a7bbad27 block: remove sup... |
870 |
return; |
1da177e4c Linux-2.6.12-rc2 |
871 872 |
} mirror = conf->mirrors + rdisk; |
e555190d8 md/raid1: delay r... |
873 874 875 876 877 878 879 880 881 |
if (test_bit(WriteMostly, &mirror->rdev->flags) && bitmap) { /* Reading from a write-mostly device must * take care not to over-take any writes * that are 'behind' */ wait_event(bitmap->behind_wait, atomic_read(&bitmap->behind_writes) == 0); } |
1da177e4c Linux-2.6.12-rc2 |
882 |
r1_bio->read_disk = rdisk; |
a167f6632 md: use separate ... |
883 |
read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev); |
d2eb35acf md/raid1: avoid r... |
884 885 |
md_trim_bio(read_bio, r1_bio->sector - bio->bi_sector, max_sectors); |
1da177e4c Linux-2.6.12-rc2 |
886 887 888 889 890 891 |
r1_bio->bios[rdisk] = read_bio; read_bio->bi_sector = r1_bio->sector + mirror->rdev->data_offset; read_bio->bi_bdev = mirror->rdev->bdev; read_bio->bi_end_io = raid1_end_read_request; |
7b6d91dae block: unify flag... |
892 |
read_bio->bi_rw = READ | do_sync; |
1da177e4c Linux-2.6.12-rc2 |
893 |
read_bio->bi_private = r1_bio; |
d2eb35acf md/raid1: avoid r... |
894 895 896 897 |
if (max_sectors < r1_bio->sectors) { /* could not read all from this device, so we will * need another r1_bio. */ |
d2eb35acf md/raid1: avoid r... |
898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 |
sectors_handled = (r1_bio->sector + max_sectors - bio->bi_sector); r1_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (bio->bi_phys_segments == 0) bio->bi_phys_segments = 2; else bio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); /* Cannot call generic_make_request directly * as that will be queued in __make_request * and subsequent mempool_alloc might block waiting * for it. So hand bio over to raid1d. */ reschedule_retry(r1_bio); r1_bio = mempool_alloc(conf->r1bio_pool, GFP_NOIO); r1_bio->master_bio = bio; r1_bio->sectors = (bio->bi_size >> 9) - sectors_handled; r1_bio->state = 0; r1_bio->mddev = mddev; r1_bio->sector = bio->bi_sector + sectors_handled; goto read_again; } else generic_make_request(read_bio); |
5a7bbad27 block: remove sup... |
925 |
return; |
1da177e4c Linux-2.6.12-rc2 |
926 927 928 929 930 |
} /* * WRITE: */ |
34db0cd60 md: add proper wr... |
931 932 933 934 935 |
if (conf->pending_count >= max_queued_requests) { md_wakeup_thread(mddev->thread); wait_event(conf->wait_barrier, conf->pending_count < max_queued_requests); } |
1f68f0c4b md/raid1: avoid w... |
936 |
/* first select target devices under rcu_lock and |
1da177e4c Linux-2.6.12-rc2 |
937 938 |
* inc refcount on their rdev. Record them by setting * bios[x] to bio |
1f68f0c4b md/raid1: avoid w... |
939 940 941 942 943 944 |
* If there are known/acknowledged bad blocks on any device on * which we have seen a write error, we want to avoid writing those * blocks. * This potentially requires several writes to write around * the bad blocks. Each set of writes gets it's own r1bio * with a set of bios attached. |
1da177e4c Linux-2.6.12-rc2 |
945 |
*/ |
c3b328ac8 md: fix up raid1/... |
946 |
plugged = mddev_check_plugged(mddev); |
8f19ccb2f md/raid1: Allocat... |
947 |
disks = conf->raid_disks * 2; |
6bfe0b499 md: support block... |
948 949 |
retry_write: blocked_rdev = NULL; |
1da177e4c Linux-2.6.12-rc2 |
950 |
rcu_read_lock(); |
1f68f0c4b md/raid1: avoid w... |
951 |
max_sectors = r1_bio->sectors; |
1da177e4c Linux-2.6.12-rc2 |
952 |
for (i = 0; i < disks; i++) { |
3cb030020 md: removing type... |
953 |
struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); |
6bfe0b499 md: support block... |
954 955 956 957 958 |
if (rdev && unlikely(test_bit(Blocked, &rdev->flags))) { atomic_inc(&rdev->nr_pending); blocked_rdev = rdev; break; } |
1f68f0c4b md/raid1: avoid w... |
959 960 |
r1_bio->bios[i] = NULL; if (!rdev || test_bit(Faulty, &rdev->flags)) { |
8f19ccb2f md/raid1: Allocat... |
961 962 |
if (i < conf->raid_disks) set_bit(R1BIO_Degraded, &r1_bio->state); |
1f68f0c4b md/raid1: avoid w... |
963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 |
continue; } atomic_inc(&rdev->nr_pending); if (test_bit(WriteErrorSeen, &rdev->flags)) { sector_t first_bad; int bad_sectors; int is_bad; is_bad = is_badblock(rdev, r1_bio->sector, max_sectors, &first_bad, &bad_sectors); if (is_bad < 0) { /* mustn't write here until the bad block is * acknowledged*/ set_bit(BlockedBadBlocks, &rdev->flags); blocked_rdev = rdev; break; } if (is_bad && first_bad <= r1_bio->sector) { /* Cannot write here at all */ bad_sectors -= (r1_bio->sector - first_bad); if (bad_sectors < max_sectors) /* mustn't write more than bad_sectors * to other devices yet */ max_sectors = bad_sectors; |
03c902e17 [PATCH] md: fix r... |
990 |
rdev_dec_pending(rdev, mddev); |
1f68f0c4b md/raid1: avoid w... |
991 992 993 994 995 996 997 998 999 1000 1001 |
/* We don't set R1BIO_Degraded as that * only applies if the disk is * missing, so it might be re-added, * and we want to know to recover this * chunk. * In this case the device is here, * and the fact that this chunk is not * in-sync is recorded in the bad * block log */ continue; |
964147d5c md/raid1: fix cou... |
1002 |
} |
1f68f0c4b md/raid1: avoid w... |
1003 1004 1005 1006 1007 1008 1009 |
if (is_bad) { int good_sectors = first_bad - r1_bio->sector; if (good_sectors < max_sectors) max_sectors = good_sectors; } } r1_bio->bios[i] = bio; |
1da177e4c Linux-2.6.12-rc2 |
1010 1011 |
} rcu_read_unlock(); |
6bfe0b499 md: support block... |
1012 1013 1014 1015 1016 1017 1018 |
if (unlikely(blocked_rdev)) { /* Wait for this device to become unblocked */ int j; for (j = 0; j < i; j++) if (r1_bio->bios[j]) rdev_dec_pending(conf->mirrors[j].rdev, mddev); |
1f68f0c4b md/raid1: avoid w... |
1019 |
r1_bio->state = 0; |
6bfe0b499 md: support block... |
1020 1021 1022 1023 1024 |
allow_barrier(conf); md_wait_for_blocked_rdev(blocked_rdev, mddev); wait_barrier(conf); goto retry_write; } |
1f68f0c4b md/raid1: avoid w... |
1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 |
if (max_sectors < r1_bio->sectors) { /* We are splitting this write into multiple parts, so * we need to prepare for allocating another r1_bio. */ r1_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (bio->bi_phys_segments == 0) bio->bi_phys_segments = 2; else bio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); |
191ea9b2c [PATCH] md: raid1... |
1036 |
} |
1f68f0c4b md/raid1: avoid w... |
1037 |
sectors_handled = r1_bio->sector + max_sectors - bio->bi_sector; |
4b6d287f6 [PATCH] md: add w... |
1038 |
|
4e78064f4 md: Fix possible ... |
1039 |
atomic_set(&r1_bio->remaining, 1); |
4b6d287f6 [PATCH] md: add w... |
1040 |
atomic_set(&r1_bio->behind_remaining, 0); |
06d91a5fe [PATCH] md: impro... |
1041 |
|
1f68f0c4b md/raid1: avoid w... |
1042 |
first_clone = 1; |
1da177e4c Linux-2.6.12-rc2 |
1043 1044 1045 1046 |
for (i = 0; i < disks; i++) { struct bio *mbio; if (!r1_bio->bios[i]) continue; |
a167f6632 md: use separate ... |
1047 |
mbio = bio_clone_mddev(bio, GFP_NOIO, mddev); |
1f68f0c4b md/raid1: avoid w... |
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 |
md_trim_bio(mbio, r1_bio->sector - bio->bi_sector, max_sectors); if (first_clone) { /* do behind I/O ? * Not if there are too many, or cannot * allocate memory, or a reader on WriteMostly * is waiting for behind writes to flush */ if (bitmap && (atomic_read(&bitmap->behind_writes) < mddev->bitmap_info.max_write_behind) && !waitqueue_active(&bitmap->behind_wait)) alloc_behind_pages(mbio, r1_bio); bitmap_startwrite(bitmap, r1_bio->sector, r1_bio->sectors, test_bit(R1BIO_BehindIO, &r1_bio->state)); first_clone = 0; } |
2ca68f5ed md/raid1: store b... |
1067 |
if (r1_bio->behind_bvecs) { |
4b6d287f6 [PATCH] md: add w... |
1068 1069 1070 1071 1072 1073 1074 |
struct bio_vec *bvec; int j; /* Yes, I really want the '__' version so that * we clear any unused pointer in the io_vec, rather * than leave them unchanged. This is important * because when we come to free the pages, we won't |
046abeede md/raid1: fix som... |
1075 |
* know the original bi_idx, so we just free |
4b6d287f6 [PATCH] md: add w... |
1076 1077 1078 |
* them all */ __bio_for_each_segment(bvec, mbio, j, 0) |
2ca68f5ed md/raid1: store b... |
1079 |
bvec->bv_page = r1_bio->behind_bvecs[j].bv_page; |
4b6d287f6 [PATCH] md: add w... |
1080 1081 1082 |
if (test_bit(WriteMostly, &conf->mirrors[i].rdev->flags)) atomic_inc(&r1_bio->behind_remaining); } |
1f68f0c4b md/raid1: avoid w... |
1083 1084 1085 1086 1087 1088 1089 1090 |
r1_bio->bios[i] = mbio; mbio->bi_sector = (r1_bio->sector + conf->mirrors[i].rdev->data_offset); mbio->bi_bdev = conf->mirrors[i].rdev->bdev; mbio->bi_end_io = raid1_end_write_request; mbio->bi_rw = WRITE | do_flush_fua | do_sync; mbio->bi_private = r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
1091 |
atomic_inc(&r1_bio->remaining); |
4e78064f4 md: Fix possible ... |
1092 1093 |
spin_lock_irqsave(&conf->device_lock, flags); bio_list_add(&conf->pending_bio_list, mbio); |
34db0cd60 md: add proper wr... |
1094 |
conf->pending_count++; |
4e78064f4 md: Fix possible ... |
1095 |
spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1096 |
} |
079fa166a md/raid1,10: Remo... |
1097 1098 1099 |
/* Mustn't call r1_bio_write_done before this next test, * as it could result in the bio being freed. */ |
1f68f0c4b md/raid1: avoid w... |
1100 |
if (sectors_handled < (bio->bi_size >> 9)) { |
079fa166a md/raid1,10: Remo... |
1101 |
r1_bio_write_done(r1_bio); |
1f68f0c4b md/raid1: avoid w... |
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 |
/* We need another r1_bio. It has already been counted * in bio->bi_phys_segments */ r1_bio = mempool_alloc(conf->r1bio_pool, GFP_NOIO); r1_bio->master_bio = bio; r1_bio->sectors = (bio->bi_size >> 9) - sectors_handled; r1_bio->state = 0; r1_bio->mddev = mddev; r1_bio->sector = bio->bi_sector + sectors_handled; goto retry_write; } |
079fa166a md/raid1,10: Remo... |
1113 1114 1115 1116 |
r1_bio_write_done(r1_bio); /* In case raid1d snuck in to freeze_array */ wake_up(&conf->wait_barrier); |
c3b328ac8 md: fix up raid1/... |
1117 |
if (do_sync || !bitmap || !plugged) |
e3881a681 [PATCH] md: pass ... |
1118 |
md_wakeup_thread(mddev->thread); |
1da177e4c Linux-2.6.12-rc2 |
1119 |
} |
fd01b88c7 md: remove typede... |
1120 |
static void status(struct seq_file *seq, struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
1121 |
{ |
e80963604 md/raid1: typedef... |
1122 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
1123 1124 1125 |
int i; seq_printf(seq, " [%d/%d] [", conf->raid_disks, |
11ce99e62 [PATCH] md: Remov... |
1126 |
conf->raid_disks - mddev->degraded); |
ddac7c7e3 [PATCH] md: Fix i... |
1127 1128 |
rcu_read_lock(); for (i = 0; i < conf->raid_disks; i++) { |
3cb030020 md: removing type... |
1129 |
struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); |
1da177e4c Linux-2.6.12-rc2 |
1130 |
seq_printf(seq, "%s", |
ddac7c7e3 [PATCH] md: Fix i... |
1131 1132 1133 |
rdev && test_bit(In_sync, &rdev->flags) ? "U" : "_"); } rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
1134 1135 |
seq_printf(seq, "]"); } |
fd01b88c7 md: remove typede... |
1136 |
static void error(struct mddev *mddev, struct md_rdev *rdev) |
1da177e4c Linux-2.6.12-rc2 |
1137 1138 |
{ char b[BDEVNAME_SIZE]; |
e80963604 md/raid1: typedef... |
1139 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
1140 1141 1142 1143 1144 1145 1146 |
/* * If it is not operational, then we have already marked it as dead * else if it is the last working disks, ignore the error, let the * next level up know. * else mark the drive as failed */ |
b2d444d7a [PATCH] md: conve... |
1147 |
if (test_bit(In_sync, &rdev->flags) |
4044ba58d md: don't retry r... |
1148 |
&& (conf->raid_disks - mddev->degraded) == 1) { |
1da177e4c Linux-2.6.12-rc2 |
1149 1150 |
/* * Don't fail the drive, act as though we were just a |
4044ba58d md: don't retry r... |
1151 1152 1153 |
* normal single drive. * However don't try a recovery from this drive as * it is very likely to fail. |
1da177e4c Linux-2.6.12-rc2 |
1154 |
*/ |
5389042ff md: change manage... |
1155 |
conf->recovery_disabled = mddev->recovery_disabled; |
1da177e4c Linux-2.6.12-rc2 |
1156 |
return; |
4044ba58d md: don't retry r... |
1157 |
} |
de393cdea md: make it easie... |
1158 |
set_bit(Blocked, &rdev->flags); |
c04be0aa8 [PATCH] md: Impro... |
1159 1160 1161 |
if (test_and_clear_bit(In_sync, &rdev->flags)) { unsigned long flags; spin_lock_irqsave(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1162 |
mddev->degraded++; |
dd00a99e7 md: avoid a possi... |
1163 |
set_bit(Faulty, &rdev->flags); |
c04be0aa8 [PATCH] md: Impro... |
1164 |
spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1165 1166 1167 |
/* * if recovery is running, make sure it aborts. */ |
dfc706450 md: restart recov... |
1168 |
set_bit(MD_RECOVERY_INTR, &mddev->recovery); |
dd00a99e7 md: avoid a possi... |
1169 1170 |
} else set_bit(Faulty, &rdev->flags); |
850b2b420 [PATCH] md: repla... |
1171 |
set_bit(MD_CHANGE_DEVS, &mddev->flags); |
067032bc6 md: Fix single pr... |
1172 1173 1174 1175 1176 |
printk(KERN_ALERT "md/raid1:%s: Disk failure on %s, disabling device. " "md/raid1:%s: Operation continuing on %d devices. ", |
9dd1e2faf md/raid1: improve... |
1177 1178 |
mdname(mddev), bdevname(rdev->bdev, b), mdname(mddev), conf->raid_disks - mddev->degraded); |
1da177e4c Linux-2.6.12-rc2 |
1179 |
} |
e80963604 md/raid1: typedef... |
1180 |
static void print_conf(struct r1conf *conf) |
1da177e4c Linux-2.6.12-rc2 |
1181 1182 |
{ int i; |
1da177e4c Linux-2.6.12-rc2 |
1183 |
|
9dd1e2faf md/raid1: improve... |
1184 1185 |
printk(KERN_DEBUG "RAID1 conf printout: "); |
1da177e4c Linux-2.6.12-rc2 |
1186 |
if (!conf) { |
9dd1e2faf md/raid1: improve... |
1187 1188 |
printk(KERN_DEBUG "(!conf) "); |
1da177e4c Linux-2.6.12-rc2 |
1189 1190 |
return; } |
9dd1e2faf md/raid1: improve... |
1191 1192 |
printk(KERN_DEBUG " --- wd:%d rd:%d ", conf->raid_disks - conf->mddev->degraded, |
1da177e4c Linux-2.6.12-rc2 |
1193 |
conf->raid_disks); |
ddac7c7e3 [PATCH] md: Fix i... |
1194 |
rcu_read_lock(); |
1da177e4c Linux-2.6.12-rc2 |
1195 1196 |
for (i = 0; i < conf->raid_disks; i++) { char b[BDEVNAME_SIZE]; |
3cb030020 md: removing type... |
1197 |
struct md_rdev *rdev = rcu_dereference(conf->mirrors[i].rdev); |
ddac7c7e3 [PATCH] md: Fix i... |
1198 |
if (rdev) |
9dd1e2faf md/raid1: improve... |
1199 1200 |
printk(KERN_DEBUG " disk %d, wo:%d, o:%d, dev:%s ", |
ddac7c7e3 [PATCH] md: Fix i... |
1201 1202 1203 |
i, !test_bit(In_sync, &rdev->flags), !test_bit(Faulty, &rdev->flags), bdevname(rdev->bdev,b)); |
1da177e4c Linux-2.6.12-rc2 |
1204 |
} |
ddac7c7e3 [PATCH] md: Fix i... |
1205 |
rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
1206 |
} |
e80963604 md/raid1: typedef... |
1207 |
static void close_sync(struct r1conf *conf) |
1da177e4c Linux-2.6.12-rc2 |
1208 |
{ |
17999be4a [PATCH] md: impro... |
1209 1210 |
wait_barrier(conf); allow_barrier(conf); |
1da177e4c Linux-2.6.12-rc2 |
1211 1212 1213 1214 |
mempool_destroy(conf->r1buf_pool); conf->r1buf_pool = NULL; } |
fd01b88c7 md: remove typede... |
1215 |
static int raid1_spare_active(struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
1216 1217 |
{ int i; |
e80963604 md/raid1: typedef... |
1218 |
struct r1conf *conf = mddev->private; |
6b9656205 md: provide appro... |
1219 1220 |
int count = 0; unsigned long flags; |
1da177e4c Linux-2.6.12-rc2 |
1221 1222 1223 |
/* * Find all failed disks within the RAID1 configuration |
ddac7c7e3 [PATCH] md: Fix i... |
1224 1225 |
* and mark them readable. * Called under mddev lock, so rcu protection not needed. |
1da177e4c Linux-2.6.12-rc2 |
1226 1227 |
*/ for (i = 0; i < conf->raid_disks; i++) { |
3cb030020 md: removing type... |
1228 |
struct md_rdev *rdev = conf->mirrors[i].rdev; |
8c7a2c2bc md/raid1: handle ... |
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 |
struct md_rdev *repl = conf->mirrors[conf->raid_disks + i].rdev; if (repl && repl->recovery_offset == MaxSector && !test_bit(Faulty, &repl->flags) && !test_and_set_bit(In_sync, &repl->flags)) { /* replacement has just become active */ if (!rdev || !test_and_clear_bit(In_sync, &rdev->flags)) count++; if (rdev) { /* Replaced device not technically * faulty, but we need to be sure * it gets removed and never re-added */ set_bit(Faulty, &rdev->flags); sysfs_notify_dirent_safe( rdev->sysfs_state); } } |
ddac7c7e3 [PATCH] md: Fix i... |
1248 1249 |
if (rdev && !test_bit(Faulty, &rdev->flags) |
c04be0aa8 [PATCH] md: Impro... |
1250 |
&& !test_and_set_bit(In_sync, &rdev->flags)) { |
6b9656205 md: provide appro... |
1251 |
count++; |
654e8b5ab MD: raid1 s/sysfs... |
1252 |
sysfs_notify_dirent_safe(rdev->sysfs_state); |
1da177e4c Linux-2.6.12-rc2 |
1253 1254 |
} } |
6b9656205 md: provide appro... |
1255 1256 1257 |
spin_lock_irqsave(&conf->device_lock, flags); mddev->degraded -= count; spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1258 1259 |
print_conf(conf); |
6b9656205 md: provide appro... |
1260 |
return count; |
1da177e4c Linux-2.6.12-rc2 |
1261 |
} |
fd01b88c7 md: remove typede... |
1262 |
static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev) |
1da177e4c Linux-2.6.12-rc2 |
1263 |
{ |
e80963604 md/raid1: typedef... |
1264 |
struct r1conf *conf = mddev->private; |
199050ea1 rationalise retur... |
1265 |
int err = -EEXIST; |
41158c7eb [PATCH] md: optim... |
1266 |
int mirror = 0; |
0f6d02d58 md: remove typede... |
1267 |
struct mirror_info *p; |
6c2fce2ef Support adding a ... |
1268 |
int first = 0; |
301946364 md/raid1: Replac... |
1269 |
int last = conf->raid_disks - 1; |
1da177e4c Linux-2.6.12-rc2 |
1270 |
|
5389042ff md: change manage... |
1271 1272 |
if (mddev->recovery_disabled == conf->recovery_disabled) return -EBUSY; |
6c2fce2ef Support adding a ... |
1273 1274 |
if (rdev->raid_disk >= 0) first = last = rdev->raid_disk; |
7ef449d1e md/raid1: If ther... |
1275 1276 1277 |
for (mirror = first; mirror <= last; mirror++) { p = conf->mirrors+mirror; if (!p->rdev) { |
1da177e4c Linux-2.6.12-rc2 |
1278 |
|
8f6c2e4b3 md: Use new topol... |
1279 1280 |
disk_stack_limits(mddev->gendisk, rdev->bdev, rdev->data_offset << 9); |
627a2d3c2 md: deal with mer... |
1281 1282 1283 1284 1285 |
/* as we don't honour merge_bvec_fn, we must * never risk violating it, so limit * ->max_segments to one lying with a single * page, as a one page request is never in * violation. |
1da177e4c Linux-2.6.12-rc2 |
1286 |
*/ |
627a2d3c2 md: deal with mer... |
1287 1288 1289 1290 1291 |
if (rdev->bdev->bd_disk->queue->merge_bvec_fn) { blk_queue_max_segments(mddev->queue, 1); blk_queue_segment_boundary(mddev->queue, PAGE_CACHE_SIZE - 1); } |
1da177e4c Linux-2.6.12-rc2 |
1292 1293 1294 |
p->head_position = 0; rdev->raid_disk = mirror; |
199050ea1 rationalise retur... |
1295 |
err = 0; |
6aea114a7 [PATCH] md: fix -... |
1296 1297 1298 1299 |
/* As all devices are equivalent, we don't need a full recovery * if this was recently any drive of the array */ if (rdev->saved_raid_disk < 0) |
41158c7eb [PATCH] md: optim... |
1300 |
conf->fullsync = 1; |
d6065f7bf [PATCH] md: provi... |
1301 |
rcu_assign_pointer(p->rdev, rdev); |
1da177e4c Linux-2.6.12-rc2 |
1302 1303 |
break; } |
7ef449d1e md/raid1: If ther... |
1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 |
if (test_bit(WantReplacement, &p->rdev->flags) && p[conf->raid_disks].rdev == NULL) { /* Add this device as a replacement */ clear_bit(In_sync, &rdev->flags); set_bit(Replacement, &rdev->flags); rdev->raid_disk = mirror; err = 0; conf->fullsync = 1; rcu_assign_pointer(p[conf->raid_disks].rdev, rdev); break; } } |
ac5e7113e md: Push down dat... |
1316 |
md_integrity_add_rdev(rdev, mddev); |
1da177e4c Linux-2.6.12-rc2 |
1317 |
print_conf(conf); |
199050ea1 rationalise retur... |
1318 |
return err; |
1da177e4c Linux-2.6.12-rc2 |
1319 |
} |
b8321b68d md: change hot_re... |
1320 |
static int raid1_remove_disk(struct mddev *mddev, struct md_rdev *rdev) |
1da177e4c Linux-2.6.12-rc2 |
1321 |
{ |
e80963604 md/raid1: typedef... |
1322 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
1323 |
int err = 0; |
b8321b68d md: change hot_re... |
1324 |
int number = rdev->raid_disk; |
0f6d02d58 md: remove typede... |
1325 |
struct mirror_info *p = conf->mirrors+ number; |
1da177e4c Linux-2.6.12-rc2 |
1326 |
|
b014f14c8 md/raid1: Allow a... |
1327 1328 |
if (rdev != p->rdev) p = conf->mirrors + conf->raid_disks + number; |
1da177e4c Linux-2.6.12-rc2 |
1329 |
print_conf(conf); |
b8321b68d md: change hot_re... |
1330 |
if (rdev == p->rdev) { |
b2d444d7a [PATCH] md: conve... |
1331 |
if (test_bit(In_sync, &rdev->flags) || |
1da177e4c Linux-2.6.12-rc2 |
1332 1333 1334 1335 |
atomic_read(&rdev->nr_pending)) { err = -EBUSY; goto abort; } |
046abeede md/raid1: fix som... |
1336 |
/* Only remove non-faulty devices if recovery |
dfc706450 md: restart recov... |
1337 1338 1339 |
* is not possible. */ if (!test_bit(Faulty, &rdev->flags) && |
5389042ff md: change manage... |
1340 |
mddev->recovery_disabled != conf->recovery_disabled && |
dfc706450 md: restart recov... |
1341 1342 1343 1344 |
mddev->degraded < conf->raid_disks) { err = -EBUSY; goto abort; } |
1da177e4c Linux-2.6.12-rc2 |
1345 |
p->rdev = NULL; |
fbd568a3e [PATCH] Change sy... |
1346 |
synchronize_rcu(); |
1da177e4c Linux-2.6.12-rc2 |
1347 1348 1349 1350 |
if (atomic_read(&rdev->nr_pending)) { /* lost the race, try later */ err = -EBUSY; p->rdev = rdev; |
ac5e7113e md: Push down dat... |
1351 |
goto abort; |
8c7a2c2bc md/raid1: handle ... |
1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 |
} else if (conf->mirrors[conf->raid_disks + number].rdev) { /* We just removed a device that is being replaced. * Move down the replacement. We drain all IO before * doing this to avoid confusion. */ struct md_rdev *repl = conf->mirrors[conf->raid_disks + number].rdev; raise_barrier(conf); clear_bit(Replacement, &repl->flags); p->rdev = repl; conf->mirrors[conf->raid_disks + number].rdev = NULL; lower_barrier(conf); clear_bit(WantReplacement, &rdev->flags); } else |
b014f14c8 md/raid1: Allow a... |
1366 |
clear_bit(WantReplacement, &rdev->flags); |
a91a2785b block: Require su... |
1367 |
err = md_integrity_register(mddev); |
1da177e4c Linux-2.6.12-rc2 |
1368 1369 1370 1371 1372 1373 |
} abort: print_conf(conf); return err; } |
6712ecf8f Drop 'size' argum... |
1374 |
static void end_sync_read(struct bio *bio, int error) |
1da177e4c Linux-2.6.12-rc2 |
1375 |
{ |
9f2c9d12b md: remove typede... |
1376 |
struct r1bio *r1_bio = bio->bi_private; |
1da177e4c Linux-2.6.12-rc2 |
1377 |
|
0fc280f60 md/raid1/ avoid b... |
1378 |
update_head_pos(r1_bio->read_disk, r1_bio); |
ba3ae3bee md/raid1: factor ... |
1379 |
|
1da177e4c Linux-2.6.12-rc2 |
1380 1381 1382 1383 1384 |
/* * we have read a block, now it needs to be re-written, * or re-read if the read failed. * We don't do much here, just schedule handling by raid1d */ |
69382e853 [PATCH] md: bette... |
1385 |
if (test_bit(BIO_UPTODATE, &bio->bi_flags)) |
1da177e4c Linux-2.6.12-rc2 |
1386 |
set_bit(R1BIO_Uptodate, &r1_bio->state); |
d11c171e6 [PATCH] md: allow... |
1387 1388 1389 |
if (atomic_dec_and_test(&r1_bio->remaining)) reschedule_retry(r1_bio); |
1da177e4c Linux-2.6.12-rc2 |
1390 |
} |
6712ecf8f Drop 'size' argum... |
1391 |
static void end_sync_write(struct bio *bio, int error) |
1da177e4c Linux-2.6.12-rc2 |
1392 1393 |
{ int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); |
9f2c9d12b md: remove typede... |
1394 |
struct r1bio *r1_bio = bio->bi_private; |
fd01b88c7 md: remove typede... |
1395 |
struct mddev *mddev = r1_bio->mddev; |
e80963604 md/raid1: typedef... |
1396 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
1397 |
int mirror=0; |
4367af556 md/raid1: clear b... |
1398 1399 |
sector_t first_bad; int bad_sectors; |
1da177e4c Linux-2.6.12-rc2 |
1400 |
|
ba3ae3bee md/raid1: factor ... |
1401 |
mirror = find_bio_disk(r1_bio, bio); |
6b1117d50 [PATCH] md: Don't... |
1402 |
if (!uptodate) { |
57dab0bdf md: use sector_t ... |
1403 |
sector_t sync_blocks = 0; |
6b1117d50 [PATCH] md: Don't... |
1404 1405 1406 1407 |
sector_t s = r1_bio->sector; long sectors_to_go = r1_bio->sectors; /* make sure these bits doesn't get cleared. */ do { |
5e3db645f [PATCH] md: fix u... |
1408 |
bitmap_end_sync(mddev->bitmap, s, |
6b1117d50 [PATCH] md: Don't... |
1409 1410 1411 1412 |
&sync_blocks, 1); s += sync_blocks; sectors_to_go -= sync_blocks; } while (sectors_to_go > 0); |
d8f05d299 md/raid1: record ... |
1413 1414 |
set_bit(WriteErrorSeen, &conf->mirrors[mirror].rdev->flags); |
19d671695 md/raid1: Mark de... |
1415 1416 1417 1418 |
if (!test_and_set_bit(WantReplacement, &conf->mirrors[mirror].rdev->flags)) set_bit(MD_RECOVERY_NEEDED, & mddev->recovery); |
d8f05d299 md/raid1: record ... |
1419 |
set_bit(R1BIO_WriteError, &r1_bio->state); |
4367af556 md/raid1: clear b... |
1420 1421 1422 |
} else if (is_badblock(conf->mirrors[mirror].rdev, r1_bio->sector, r1_bio->sectors, |
3a9f28a51 md/raid1: improve... |
1423 1424 1425 1426 1427 1428 |
&first_bad, &bad_sectors) && !is_badblock(conf->mirrors[r1_bio->read_disk].rdev, r1_bio->sector, r1_bio->sectors, &first_bad, &bad_sectors) ) |
4367af556 md/raid1: clear b... |
1429 |
set_bit(R1BIO_MadeGood, &r1_bio->state); |
e3b9703e2 [PATCH] md: yet a... |
1430 |
|
1da177e4c Linux-2.6.12-rc2 |
1431 |
if (atomic_dec_and_test(&r1_bio->remaining)) { |
4367af556 md/raid1: clear b... |
1432 |
int s = r1_bio->sectors; |
d8f05d299 md/raid1: record ... |
1433 1434 |
if (test_bit(R1BIO_MadeGood, &r1_bio->state) || test_bit(R1BIO_WriteError, &r1_bio->state)) |
4367af556 md/raid1: clear b... |
1435 1436 1437 1438 1439 |
reschedule_retry(r1_bio); else { put_buf(r1_bio); md_done_sync(mddev, s, uptodate); } |
1da177e4c Linux-2.6.12-rc2 |
1440 |
} |
1da177e4c Linux-2.6.12-rc2 |
1441 |
} |
3cb030020 md: removing type... |
1442 |
static int r1_sync_page_io(struct md_rdev *rdev, sector_t sector, |
d8f05d299 md/raid1: record ... |
1443 1444 1445 1446 1447 |
int sectors, struct page *page, int rw) { if (sync_page_io(rdev, sector, sectors << 9, page, rw, false)) /* success */ return 1; |
19d671695 md/raid1: Mark de... |
1448 |
if (rw == WRITE) { |
d8f05d299 md/raid1: record ... |
1449 |
set_bit(WriteErrorSeen, &rdev->flags); |
19d671695 md/raid1: Mark de... |
1450 1451 1452 1453 1454 |
if (!test_and_set_bit(WantReplacement, &rdev->flags)) set_bit(MD_RECOVERY_NEEDED, & rdev->mddev->recovery); } |
d8f05d299 md/raid1: record ... |
1455 1456 1457 1458 1459 |
/* need to record an error - either for the block or the device */ if (!rdev_set_badblocks(rdev, sector, sectors, 0)) md_error(rdev->mddev, rdev); return 0; } |
9f2c9d12b md: remove typede... |
1460 |
static int fix_sync_read_error(struct r1bio *r1_bio) |
1da177e4c Linux-2.6.12-rc2 |
1461 |
{ |
a68e58703 md/raid1: split o... |
1462 1463 1464 1465 1466 1467 1468 |
/* Try some synchronous reads of other devices to get * good data, much like with normal read errors. Only * read into the pages we already have so we don't * need to re-issue the read request. * We don't need to freeze the array, because being in an * active sync request, there is no normal IO, and * no overlapping syncs. |
06f603851 md/raid1: avoid r... |
1469 1470 1471 |
* We don't need to check is_badblock() again as we * made sure that anything with a bad block in range * will have bi_end_io clear. |
a68e58703 md/raid1: split o... |
1472 |
*/ |
fd01b88c7 md: remove typede... |
1473 |
struct mddev *mddev = r1_bio->mddev; |
e80963604 md/raid1: typedef... |
1474 |
struct r1conf *conf = mddev->private; |
a68e58703 md/raid1: split o... |
1475 1476 1477 1478 1479 1480 1481 1482 1483 |
struct bio *bio = r1_bio->bios[r1_bio->read_disk]; sector_t sect = r1_bio->sector; int sectors = r1_bio->sectors; int idx = 0; while(sectors) { int s = sectors; int d = r1_bio->read_disk; int success = 0; |
3cb030020 md: removing type... |
1484 |
struct md_rdev *rdev; |
78d7f5f72 md/raid1: tidy up... |
1485 |
int start; |
a68e58703 md/raid1: split o... |
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 |
if (s > (PAGE_SIZE>>9)) s = PAGE_SIZE >> 9; do { if (r1_bio->bios[d]->bi_end_io == end_sync_read) { /* No rcu protection needed here devices * can only be removed when no resync is * active, and resync is currently active */ rdev = conf->mirrors[d].rdev; |
9d3d80113 md/raid1: move rd... |
1496 |
if (sync_page_io(rdev, sect, s<<9, |
a68e58703 md/raid1: split o... |
1497 1498 1499 1500 1501 1502 1503 |
bio->bi_io_vec[idx].bv_page, READ, false)) { success = 1; break; } } d++; |
8f19ccb2f md/raid1: Allocat... |
1504 |
if (d == conf->raid_disks * 2) |
a68e58703 md/raid1: split o... |
1505 1506 |
d = 0; } while (!success && d != r1_bio->read_disk); |
78d7f5f72 md/raid1: tidy up... |
1507 |
if (!success) { |
a68e58703 md/raid1: split o... |
1508 |
char b[BDEVNAME_SIZE]; |
3a9f28a51 md/raid1: improve... |
1509 1510 1511 1512 1513 1514 |
int abort = 0; /* Cannot read from anywhere, this block is lost. * Record a bad block on each device. If that doesn't * work just disable and interrupt the recovery. * Don't fail devices as that won't really help. */ |
a68e58703 md/raid1: split o... |
1515 1516 1517 1518 1519 1520 |
printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O read error" " for block %llu ", mdname(mddev), bdevname(bio->bi_bdev, b), (unsigned long long)r1_bio->sector); |
8f19ccb2f md/raid1: Allocat... |
1521 |
for (d = 0; d < conf->raid_disks * 2; d++) { |
3a9f28a51 md/raid1: improve... |
1522 1523 1524 1525 1526 1527 1528 |
rdev = conf->mirrors[d].rdev; if (!rdev || test_bit(Faulty, &rdev->flags)) continue; if (!rdev_set_badblocks(rdev, sect, s, 0)) abort = 1; } if (abort) { |
d890fa2b0 md: Fix some bugs... |
1529 1530 |
conf->recovery_disabled = mddev->recovery_disabled; |
3a9f28a51 md/raid1: improve... |
1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 |
set_bit(MD_RECOVERY_INTR, &mddev->recovery); md_done_sync(mddev, r1_bio->sectors, 0); put_buf(r1_bio); return 0; } /* Try next page */ sectors -= s; sect += s; idx++; continue; |
d11c171e6 [PATCH] md: allow... |
1541 |
} |
78d7f5f72 md/raid1: tidy up... |
1542 1543 1544 1545 1546 |
start = d; /* write it back and re-read */ while (d != r1_bio->read_disk) { if (d == 0) |
8f19ccb2f md/raid1: Allocat... |
1547 |
d = conf->raid_disks * 2; |
78d7f5f72 md/raid1: tidy up... |
1548 1549 1550 1551 |
d--; if (r1_bio->bios[d]->bi_end_io != end_sync_read) continue; rdev = conf->mirrors[d].rdev; |
d8f05d299 md/raid1: record ... |
1552 1553 1554 |
if (r1_sync_page_io(rdev, sect, s, bio->bi_io_vec[idx].bv_page, WRITE) == 0) { |
78d7f5f72 md/raid1: tidy up... |
1555 1556 |
r1_bio->bios[d]->bi_end_io = NULL; rdev_dec_pending(rdev, mddev); |
9d3d80113 md/raid1: move rd... |
1557 |
} |
78d7f5f72 md/raid1: tidy up... |
1558 1559 1560 1561 |
} d = start; while (d != r1_bio->read_disk) { if (d == 0) |
8f19ccb2f md/raid1: Allocat... |
1562 |
d = conf->raid_disks * 2; |
78d7f5f72 md/raid1: tidy up... |
1563 1564 1565 1566 |
d--; if (r1_bio->bios[d]->bi_end_io != end_sync_read) continue; rdev = conf->mirrors[d].rdev; |
d8f05d299 md/raid1: record ... |
1567 1568 1569 |
if (r1_sync_page_io(rdev, sect, s, bio->bi_io_vec[idx].bv_page, READ) != 0) |
9d3d80113 md/raid1: move rd... |
1570 |
atomic_add(s, &rdev->corrected_errors); |
78d7f5f72 md/raid1: tidy up... |
1571 |
} |
a68e58703 md/raid1: split o... |
1572 1573 1574 1575 |
sectors -= s; sect += s; idx ++; } |
78d7f5f72 md/raid1: tidy up... |
1576 |
set_bit(R1BIO_Uptodate, &r1_bio->state); |
7ca78d57d md/raid1: try fix... |
1577 |
set_bit(BIO_UPTODATE, &bio->bi_flags); |
a68e58703 md/raid1: split o... |
1578 1579 |
return 1; } |
9f2c9d12b md: remove typede... |
1580 |
static int process_checks(struct r1bio *r1_bio) |
a68e58703 md/raid1: split o... |
1581 1582 1583 1584 1585 1586 1587 1588 |
{ /* We have read all readable devices. If we haven't * got the block, then there is no hope left. * If we have, then we want to do a comparison * and skip the write if everything is the same. * If any blocks failed to read, then we need to * attempt an over-write */ |
fd01b88c7 md: remove typede... |
1589 |
struct mddev *mddev = r1_bio->mddev; |
e80963604 md/raid1: typedef... |
1590 |
struct r1conf *conf = mddev->private; |
a68e58703 md/raid1: split o... |
1591 1592 |
int primary; int i; |
8f19ccb2f md/raid1: Allocat... |
1593 |
for (primary = 0; primary < conf->raid_disks * 2; primary++) |
a68e58703 md/raid1: split o... |
1594 1595 1596 1597 1598 1599 1600 |
if (r1_bio->bios[primary]->bi_end_io == end_sync_read && test_bit(BIO_UPTODATE, &r1_bio->bios[primary]->bi_flags)) { r1_bio->bios[primary]->bi_end_io = NULL; rdev_dec_pending(conf->mirrors[primary].rdev, mddev); break; } r1_bio->read_disk = primary; |
8f19ccb2f md/raid1: Allocat... |
1601 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
78d7f5f72 md/raid1: tidy up... |
1602 1603 1604 1605 1606 |
int j; int vcnt = r1_bio->sectors >> (PAGE_SHIFT- 9); struct bio *pbio = r1_bio->bios[primary]; struct bio *sbio = r1_bio->bios[i]; int size; |
a68e58703 md/raid1: split o... |
1607 |
|
78d7f5f72 md/raid1: tidy up... |
1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 |
if (r1_bio->bios[i]->bi_end_io != end_sync_read) continue; if (test_bit(BIO_UPTODATE, &sbio->bi_flags)) { for (j = vcnt; j-- ; ) { struct page *p, *s; p = pbio->bi_io_vec[j].bv_page; s = sbio->bi_io_vec[j].bv_page; if (memcmp(page_address(p), page_address(s), PAGE_SIZE)) break; |
69382e853 [PATCH] md: bette... |
1620 |
} |
78d7f5f72 md/raid1: tidy up... |
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 |
} else j = 0; if (j >= 0) mddev->resync_mismatches += r1_bio->sectors; if (j < 0 || (test_bit(MD_RECOVERY_CHECK, &mddev->recovery) && test_bit(BIO_UPTODATE, &sbio->bi_flags))) { /* No need to write to this device. */ sbio->bi_end_io = NULL; rdev_dec_pending(conf->mirrors[i].rdev, mddev); continue; } /* fixup the bio for reuse */ sbio->bi_vcnt = vcnt; sbio->bi_size = r1_bio->sectors << 9; sbio->bi_idx = 0; sbio->bi_phys_segments = 0; sbio->bi_flags &= ~(BIO_POOL_MASK - 1); sbio->bi_flags |= 1 << BIO_UPTODATE; sbio->bi_next = NULL; sbio->bi_sector = r1_bio->sector + conf->mirrors[i].rdev->data_offset; sbio->bi_bdev = conf->mirrors[i].rdev->bdev; size = sbio->bi_size; for (j = 0; j < vcnt ; j++) { struct bio_vec *bi; bi = &sbio->bi_io_vec[j]; bi->bv_offset = 0; if (size > PAGE_SIZE) bi->bv_len = PAGE_SIZE; else bi->bv_len = size; size -= PAGE_SIZE; memcpy(page_address(bi->bv_page), page_address(pbio->bi_io_vec[j].bv_page), PAGE_SIZE); |
69382e853 [PATCH] md: bette... |
1656 |
} |
78d7f5f72 md/raid1: tidy up... |
1657 |
} |
a68e58703 md/raid1: split o... |
1658 1659 |
return 0; } |
9f2c9d12b md: remove typede... |
1660 |
static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio) |
a68e58703 md/raid1: split o... |
1661 |
{ |
e80963604 md/raid1: typedef... |
1662 |
struct r1conf *conf = mddev->private; |
a68e58703 md/raid1: split o... |
1663 |
int i; |
8f19ccb2f md/raid1: Allocat... |
1664 |
int disks = conf->raid_disks * 2; |
a68e58703 md/raid1: split o... |
1665 1666 1667 |
struct bio *bio, *wbio; bio = r1_bio->bios[r1_bio->read_disk]; |
a68e58703 md/raid1: split o... |
1668 1669 1670 1671 |
if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) /* ouch - failed to read all of that. */ if (!fix_sync_read_error(r1_bio)) return; |
7ca78d57d md/raid1: try fix... |
1672 1673 1674 1675 |
if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) if (process_checks(r1_bio) < 0) return; |
d11c171e6 [PATCH] md: allow... |
1676 1677 1678 |
/* * schedule writes */ |
1da177e4c Linux-2.6.12-rc2 |
1679 1680 1681 |
atomic_set(&r1_bio->remaining, 1); for (i = 0; i < disks ; i++) { wbio = r1_bio->bios[i]; |
3e198f782 [PATCH] md: tidyu... |
1682 1683 1684 1685 |
if (wbio->bi_end_io == NULL || (wbio->bi_end_io == end_sync_read && (i == r1_bio->read_disk || !test_bit(MD_RECOVERY_SYNC, &mddev->recovery)))) |
1da177e4c Linux-2.6.12-rc2 |
1686 |
continue; |
3e198f782 [PATCH] md: tidyu... |
1687 1688 |
wbio->bi_rw = WRITE; wbio->bi_end_io = end_sync_write; |
1da177e4c Linux-2.6.12-rc2 |
1689 1690 |
atomic_inc(&r1_bio->remaining); md_sync_acct(conf->mirrors[i].rdev->bdev, wbio->bi_size >> 9); |
191ea9b2c [PATCH] md: raid1... |
1691 |
|
1da177e4c Linux-2.6.12-rc2 |
1692 1693 1694 1695 |
generic_make_request(wbio); } if (atomic_dec_and_test(&r1_bio->remaining)) { |
191ea9b2c [PATCH] md: raid1... |
1696 |
/* if we're here, all write(s) have completed, so clean up */ |
1da177e4c Linux-2.6.12-rc2 |
1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 |
md_done_sync(mddev, r1_bio->sectors, 1); put_buf(r1_bio); } } /* * This is a kernel thread which: * * 1. Retries failed read operations on working mirrors. * 2. Updates the raid superblock when problems encounter. |
d2eb35acf md/raid1: avoid r... |
1707 |
* 3. Performs writes following reads for array synchronising. |
1da177e4c Linux-2.6.12-rc2 |
1708 |
*/ |
e80963604 md/raid1: typedef... |
1709 |
static void fix_read_error(struct r1conf *conf, int read_disk, |
867868fb5 [PATCH] md: Facto... |
1710 1711 |
sector_t sect, int sectors) { |
fd01b88c7 md: remove typede... |
1712 |
struct mddev *mddev = conf->mddev; |
867868fb5 [PATCH] md: Facto... |
1713 1714 1715 1716 1717 |
while(sectors) { int s = sectors; int d = read_disk; int success = 0; int start; |
3cb030020 md: removing type... |
1718 |
struct md_rdev *rdev; |
867868fb5 [PATCH] md: Facto... |
1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 |
if (s > (PAGE_SIZE>>9)) s = PAGE_SIZE >> 9; do { /* Note: no rcu protection needed here * as this is synchronous in the raid1d thread * which is the thread that might remove * a device. If raid1d ever becomes multi-threaded.... */ |
d2eb35acf md/raid1: avoid r... |
1729 1730 |
sector_t first_bad; int bad_sectors; |
867868fb5 [PATCH] md: Facto... |
1731 1732 1733 |
rdev = conf->mirrors[d].rdev; if (rdev && test_bit(In_sync, &rdev->flags) && |
d2eb35acf md/raid1: avoid r... |
1734 1735 |
is_badblock(rdev, sect, s, &first_bad, &bad_sectors) == 0 && |
ccebd4c41 md-new-param-to_s... |
1736 1737 |
sync_page_io(rdev, sect, s<<9, conf->tmppage, READ, false)) |
867868fb5 [PATCH] md: Facto... |
1738 1739 1740 |
success = 1; else { d++; |
8f19ccb2f md/raid1: Allocat... |
1741 |
if (d == conf->raid_disks * 2) |
867868fb5 [PATCH] md: Facto... |
1742 1743 1744 1745 1746 |
d = 0; } } while (!success && d != read_disk); if (!success) { |
d8f05d299 md/raid1: record ... |
1747 |
/* Cannot read from anywhere - mark it bad */ |
3cb030020 md: removing type... |
1748 |
struct md_rdev *rdev = conf->mirrors[read_disk].rdev; |
d8f05d299 md/raid1: record ... |
1749 1750 |
if (!rdev_set_badblocks(rdev, sect, s, 0)) md_error(mddev, rdev); |
867868fb5 [PATCH] md: Facto... |
1751 1752 1753 1754 1755 1756 |
break; } /* write it back and re-read */ start = d; while (d != read_disk) { if (d==0) |
8f19ccb2f md/raid1: Allocat... |
1757 |
d = conf->raid_disks * 2; |
867868fb5 [PATCH] md: Facto... |
1758 1759 1760 |
d--; rdev = conf->mirrors[d].rdev; if (rdev && |
d8f05d299 md/raid1: record ... |
1761 1762 1763 |
test_bit(In_sync, &rdev->flags)) r1_sync_page_io(rdev, sect, s, conf->tmppage, WRITE); |
867868fb5 [PATCH] md: Facto... |
1764 1765 1766 1767 1768 |
} d = start; while (d != read_disk) { char b[BDEVNAME_SIZE]; if (d==0) |
8f19ccb2f md/raid1: Allocat... |
1769 |
d = conf->raid_disks * 2; |
867868fb5 [PATCH] md: Facto... |
1770 1771 1772 1773 |
d--; rdev = conf->mirrors[d].rdev; if (rdev && test_bit(In_sync, &rdev->flags)) { |
d8f05d299 md/raid1: record ... |
1774 1775 |
if (r1_sync_page_io(rdev, sect, s, conf->tmppage, READ)) { |
867868fb5 [PATCH] md: Facto... |
1776 1777 |
atomic_add(s, &rdev->corrected_errors); printk(KERN_INFO |
9dd1e2faf md/raid1: improve... |
1778 |
"md/raid1:%s: read error corrected " |
867868fb5 [PATCH] md: Facto... |
1779 1780 1781 |
"(%d sectors at %llu on %s) ", mdname(mddev), s, |
969b755aa [PATCH] md: fix p... |
1782 1783 |
(unsigned long long)(sect + rdev->data_offset), |
867868fb5 [PATCH] md: Facto... |
1784 1785 1786 1787 1788 1789 1790 1791 |
bdevname(rdev->bdev, b)); } } } sectors -= s; sect += s; } } |
cd5ff9a16 md/raid1: Handle... |
1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 |
static void bi_complete(struct bio *bio, int error) { complete((struct completion *)bio->bi_private); } static int submit_bio_wait(int rw, struct bio *bio) { struct completion event; rw |= REQ_SYNC; init_completion(&event); bio->bi_private = &event; bio->bi_end_io = bi_complete; submit_bio(rw, bio); wait_for_completion(&event); return test_bit(BIO_UPTODATE, &bio->bi_flags); } |
9f2c9d12b md: remove typede... |
1810 |
static int narrow_write_error(struct r1bio *r1_bio, int i) |
cd5ff9a16 md/raid1: Handle... |
1811 |
{ |
fd01b88c7 md: remove typede... |
1812 |
struct mddev *mddev = r1_bio->mddev; |
e80963604 md/raid1: typedef... |
1813 |
struct r1conf *conf = mddev->private; |
3cb030020 md: removing type... |
1814 |
struct md_rdev *rdev = conf->mirrors[i].rdev; |
cd5ff9a16 md/raid1: Handle... |
1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 |
int vcnt, idx; struct bio_vec *vec; /* bio has the data to be written to device 'i' where * we just recently had a write error. * We repeatedly clone the bio and trim down to one block, * then try the write. Where the write fails we record * a bad block. * It is conceivable that the bio doesn't exactly align with * blocks. We must handle this somehow. * * We currently own a reference on the rdev. */ int block_sectors; sector_t sector; int sectors; int sect_to_write = r1_bio->sectors; int ok = 1; if (rdev->badblocks.shift < 0) return 0; block_sectors = 1 << rdev->badblocks.shift; sector = r1_bio->sector; sectors = ((sector + block_sectors) & ~(sector_t)(block_sectors - 1)) - sector; if (test_bit(R1BIO_BehindIO, &r1_bio->state)) { vcnt = r1_bio->behind_page_count; vec = r1_bio->behind_bvecs; idx = 0; while (vec[idx].bv_page == NULL) idx++; } else { vcnt = r1_bio->master_bio->bi_vcnt; vec = r1_bio->master_bio->bi_io_vec; idx = r1_bio->master_bio->bi_idx; } while (sect_to_write) { struct bio *wbio; if (sectors > sect_to_write) sectors = sect_to_write; /* Write at 'sector' for 'sectors'*/ wbio = bio_alloc_mddev(GFP_NOIO, vcnt, mddev); memcpy(wbio->bi_io_vec, vec, vcnt * sizeof(struct bio_vec)); wbio->bi_sector = r1_bio->sector; wbio->bi_rw = WRITE; wbio->bi_vcnt = vcnt; wbio->bi_size = r1_bio->sectors << 9; wbio->bi_idx = idx; md_trim_bio(wbio, sector - r1_bio->sector, sectors); wbio->bi_sector += rdev->data_offset; wbio->bi_bdev = rdev->bdev; if (submit_bio_wait(WRITE, wbio) == 0) /* failure! */ ok = rdev_set_badblocks(rdev, sector, sectors, 0) && ok; bio_put(wbio); sect_to_write -= sectors; sector += sectors; sectors = block_sectors; } return ok; } |
e80963604 md/raid1: typedef... |
1885 |
static void handle_sync_write_finished(struct r1conf *conf, struct r1bio *r1_bio) |
62096bce2 md/raid1: factor ... |
1886 1887 1888 |
{ int m; int s = r1_bio->sectors; |
8f19ccb2f md/raid1: Allocat... |
1889 |
for (m = 0; m < conf->raid_disks * 2 ; m++) { |
3cb030020 md: removing type... |
1890 |
struct md_rdev *rdev = conf->mirrors[m].rdev; |
62096bce2 md/raid1: factor ... |
1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 |
struct bio *bio = r1_bio->bios[m]; if (bio->bi_end_io == NULL) continue; if (test_bit(BIO_UPTODATE, &bio->bi_flags) && test_bit(R1BIO_MadeGood, &r1_bio->state)) { rdev_clear_badblocks(rdev, r1_bio->sector, s); } if (!test_bit(BIO_UPTODATE, &bio->bi_flags) && test_bit(R1BIO_WriteError, &r1_bio->state)) { if (!rdev_set_badblocks(rdev, r1_bio->sector, s, 0)) md_error(conf->mddev, rdev); } } put_buf(r1_bio); md_done_sync(conf->mddev, s, 1); } |
e80963604 md/raid1: typedef... |
1907 |
static void handle_write_finished(struct r1conf *conf, struct r1bio *r1_bio) |
62096bce2 md/raid1: factor ... |
1908 1909 |
{ int m; |
8f19ccb2f md/raid1: Allocat... |
1910 |
for (m = 0; m < conf->raid_disks * 2 ; m++) |
62096bce2 md/raid1: factor ... |
1911 |
if (r1_bio->bios[m] == IO_MADE_GOOD) { |
3cb030020 md: removing type... |
1912 |
struct md_rdev *rdev = conf->mirrors[m].rdev; |
62096bce2 md/raid1: factor ... |
1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 |
rdev_clear_badblocks(rdev, r1_bio->sector, r1_bio->sectors); rdev_dec_pending(rdev, conf->mddev); } else if (r1_bio->bios[m] != NULL) { /* This drive got a write error. We need to * narrow down and record precise write * errors. */ if (!narrow_write_error(r1_bio, m)) { md_error(conf->mddev, conf->mirrors[m].rdev); /* an I/O failed, we can't clear the bitmap */ set_bit(R1BIO_Degraded, &r1_bio->state); } rdev_dec_pending(conf->mirrors[m].rdev, conf->mddev); } if (test_bit(R1BIO_WriteError, &r1_bio->state)) close_write(r1_bio); raid_end_bio_io(r1_bio); } |
e80963604 md/raid1: typedef... |
1935 |
static void handle_read_error(struct r1conf *conf, struct r1bio *r1_bio) |
62096bce2 md/raid1: factor ... |
1936 1937 1938 |
{ int disk; int max_sectors; |
fd01b88c7 md: remove typede... |
1939 |
struct mddev *mddev = conf->mddev; |
62096bce2 md/raid1: factor ... |
1940 1941 |
struct bio *bio; char b[BDEVNAME_SIZE]; |
3cb030020 md: removing type... |
1942 |
struct md_rdev *rdev; |
62096bce2 md/raid1: factor ... |
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 |
clear_bit(R1BIO_ReadError, &r1_bio->state); /* we got a read error. Maybe the drive is bad. Maybe just * the block and we can fix it. * We freeze all other IO, and try reading the block from * other devices. When we find one, we re-write * and check it that fixes the read error. * This is all done synchronously while the array is * frozen */ if (mddev->ro == 0) { freeze_array(conf); fix_read_error(conf, r1_bio->read_disk, r1_bio->sector, r1_bio->sectors); unfreeze_array(conf); } else md_error(mddev, conf->mirrors[r1_bio->read_disk].rdev); bio = r1_bio->bios[r1_bio->read_disk]; bdevname(bio->bi_bdev, b); read_more: disk = read_balance(conf, r1_bio, &max_sectors); if (disk == -1) { printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O" " read error for block %llu ", mdname(mddev), b, (unsigned long long)r1_bio->sector); raid_end_bio_io(r1_bio); } else { const unsigned long do_sync = r1_bio->master_bio->bi_rw & REQ_SYNC; if (bio) { r1_bio->bios[r1_bio->read_disk] = mddev->ro ? IO_BLOCKED : NULL; bio_put(bio); } r1_bio->read_disk = disk; bio = bio_clone_mddev(r1_bio->master_bio, GFP_NOIO, mddev); md_trim_bio(bio, r1_bio->sector - bio->bi_sector, max_sectors); r1_bio->bios[r1_bio->read_disk] = bio; rdev = conf->mirrors[disk].rdev; printk_ratelimited(KERN_ERR "md/raid1:%s: redirecting sector %llu" " to other mirror: %s ", mdname(mddev), (unsigned long long)r1_bio->sector, bdevname(rdev->bdev, b)); bio->bi_sector = r1_bio->sector + rdev->data_offset; bio->bi_bdev = rdev->bdev; bio->bi_end_io = raid1_end_read_request; bio->bi_rw = READ | do_sync; bio->bi_private = r1_bio; if (max_sectors < r1_bio->sectors) { /* Drat - have to split this up more */ struct bio *mbio = r1_bio->master_bio; int sectors_handled = (r1_bio->sector + max_sectors - mbio->bi_sector); r1_bio->sectors = max_sectors; spin_lock_irq(&conf->device_lock); if (mbio->bi_phys_segments == 0) mbio->bi_phys_segments = 2; else mbio->bi_phys_segments++; spin_unlock_irq(&conf->device_lock); generic_make_request(bio); bio = NULL; r1_bio = mempool_alloc(conf->r1bio_pool, GFP_NOIO); r1_bio->master_bio = mbio; r1_bio->sectors = (mbio->bi_size >> 9) - sectors_handled; r1_bio->state = 0; set_bit(R1BIO_ReadError, &r1_bio->state); r1_bio->mddev = mddev; r1_bio->sector = mbio->bi_sector + sectors_handled; goto read_more; } else generic_make_request(bio); } } |
fd01b88c7 md: remove typede... |
2026 |
static void raid1d(struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
2027 |
{ |
9f2c9d12b md: remove typede... |
2028 |
struct r1bio *r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
2029 |
unsigned long flags; |
e80963604 md/raid1: typedef... |
2030 |
struct r1conf *conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
2031 |
struct list_head *head = &conf->retry_list; |
e1dfa0a29 md: use new plugg... |
2032 |
struct blk_plug plug; |
1da177e4c Linux-2.6.12-rc2 |
2033 2034 |
md_check_recovery(mddev); |
e1dfa0a29 md: use new plugg... |
2035 2036 |
blk_start_plug(&plug); |
1da177e4c Linux-2.6.12-rc2 |
2037 |
for (;;) { |
191ea9b2c [PATCH] md: raid1... |
2038 |
|
c3b328ac8 md: fix up raid1/... |
2039 2040 |
if (atomic_read(&mddev->plug_cnt) == 0) flush_pending_writes(conf); |
191ea9b2c [PATCH] md: raid1... |
2041 |
|
a35e63efa md: fix deadlock ... |
2042 2043 2044 |
spin_lock_irqsave(&conf->device_lock, flags); if (list_empty(head)) { spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
2045 |
break; |
a35e63efa md: fix deadlock ... |
2046 |
} |
9f2c9d12b md: remove typede... |
2047 |
r1_bio = list_entry(head->prev, struct r1bio, retry_list); |
1da177e4c Linux-2.6.12-rc2 |
2048 |
list_del(head->prev); |
ddaf22aba [PATCH] md: attem... |
2049 |
conf->nr_queued--; |
1da177e4c Linux-2.6.12-rc2 |
2050 2051 2052 |
spin_unlock_irqrestore(&conf->device_lock, flags); mddev = r1_bio->mddev; |
070ec55d0 md: remove mddev_... |
2053 |
conf = mddev->private; |
4367af556 md/raid1: clear b... |
2054 |
if (test_bit(R1BIO_IsSync, &r1_bio->state)) { |
d8f05d299 md/raid1: record ... |
2055 |
if (test_bit(R1BIO_MadeGood, &r1_bio->state) || |
62096bce2 md/raid1: factor ... |
2056 2057 2058 |
test_bit(R1BIO_WriteError, &r1_bio->state)) handle_sync_write_finished(conf, r1_bio); else |
4367af556 md/raid1: clear b... |
2059 |
sync_request_write(mddev, r1_bio); |
cd5ff9a16 md/raid1: Handle... |
2060 |
} else if (test_bit(R1BIO_MadeGood, &r1_bio->state) || |
62096bce2 md/raid1: factor ... |
2061 2062 2063 2064 2065 |
test_bit(R1BIO_WriteError, &r1_bio->state)) handle_write_finished(conf, r1_bio); else if (test_bit(R1BIO_ReadError, &r1_bio->state)) handle_read_error(conf, r1_bio); else |
d2eb35acf md/raid1: avoid r... |
2066 2067 2068 2069 |
/* just a partial read to be scheduled from separate * context */ generic_make_request(r1_bio->bios[r1_bio->read_disk]); |
62096bce2 md/raid1: factor ... |
2070 |
|
1d9d52416 md/raid1/raid10: ... |
2071 |
cond_resched(); |
de393cdea md: make it easie... |
2072 2073 |
if (mddev->flags & ~(1<<MD_CHANGE_PENDING)) md_check_recovery(mddev); |
1da177e4c Linux-2.6.12-rc2 |
2074 |
} |
e1dfa0a29 md: use new plugg... |
2075 |
blk_finish_plug(&plug); |
1da177e4c Linux-2.6.12-rc2 |
2076 |
} |
e80963604 md/raid1: typedef... |
2077 |
static int init_resync(struct r1conf *conf) |
1da177e4c Linux-2.6.12-rc2 |
2078 2079 2080 2081 |
{ int buffs; buffs = RESYNC_WINDOW / RESYNC_BLOCK_SIZE; |
9e77c485f BUG_ON() Conversi... |
2082 |
BUG_ON(conf->r1buf_pool); |
1da177e4c Linux-2.6.12-rc2 |
2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 |
conf->r1buf_pool = mempool_create(buffs, r1buf_pool_alloc, r1buf_pool_free, conf->poolinfo); if (!conf->r1buf_pool) return -ENOMEM; conf->next_resync = 0; return 0; } /* * perform a "sync" on one "block" * * We need to make sure that no normal I/O request - particularly write * requests - conflict with active sync requests. * * This is achieved by tracking pending requests and a 'barrier' concept * that can be installed to exclude normal IO requests. */ |
fd01b88c7 md: remove typede... |
2100 |
static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipped, int go_faster) |
1da177e4c Linux-2.6.12-rc2 |
2101 |
{ |
e80963604 md/raid1: typedef... |
2102 |
struct r1conf *conf = mddev->private; |
9f2c9d12b md: remove typede... |
2103 |
struct r1bio *r1_bio; |
1da177e4c Linux-2.6.12-rc2 |
2104 2105 |
struct bio *bio; sector_t max_sector, nr_sectors; |
3e198f782 [PATCH] md: tidyu... |
2106 |
int disk = -1; |
1da177e4c Linux-2.6.12-rc2 |
2107 |
int i; |
3e198f782 [PATCH] md: tidyu... |
2108 2109 |
int wonly = -1; int write_targets = 0, read_targets = 0; |
57dab0bdf md: use sector_t ... |
2110 |
sector_t sync_blocks; |
e3b9703e2 [PATCH] md: yet a... |
2111 |
int still_degraded = 0; |
06f603851 md/raid1: avoid r... |
2112 2113 |
int good_sectors = RESYNC_SECTORS; int min_bad = 0; /* number of sectors that are bad in all devices */ |
1da177e4c Linux-2.6.12-rc2 |
2114 2115 2116 |
if (!conf->r1buf_pool) if (init_resync(conf)) |
57afd89f9 [PATCH] md: impro... |
2117 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
2118 |
|
58c0fed40 md: Make mddev->s... |
2119 |
max_sector = mddev->dev_sectors; |
1da177e4c Linux-2.6.12-rc2 |
2120 |
if (sector_nr >= max_sector) { |
191ea9b2c [PATCH] md: raid1... |
2121 2122 2123 2124 2125 |
/* If we aborted, we need to abort the * sync on the 'current' bitmap chunk (there will * only be one in raid1 resync. * We can find the current addess in mddev->curr_resync */ |
6a806c510 [PATCH] md/raid1:... |
2126 2127 |
if (mddev->curr_resync < max_sector) /* aborted */ bitmap_end_sync(mddev->bitmap, mddev->curr_resync, |
191ea9b2c [PATCH] md: raid1... |
2128 |
&sync_blocks, 1); |
6a806c510 [PATCH] md/raid1:... |
2129 |
else /* completed sync */ |
191ea9b2c [PATCH] md: raid1... |
2130 |
conf->fullsync = 0; |
6a806c510 [PATCH] md/raid1:... |
2131 2132 |
bitmap_close_sync(mddev->bitmap); |
1da177e4c Linux-2.6.12-rc2 |
2133 2134 2135 |
close_sync(conf); return 0; } |
07d84d109 [PATCH] md: Allow... |
2136 2137 |
if (mddev->bitmap == NULL && mddev->recovery_cp == MaxSector && |
6394cca54 [PATCH] md: fix r... |
2138 |
!test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) && |
07d84d109 [PATCH] md: Allow... |
2139 2140 2141 2142 |
conf->fullsync == 0) { *skipped = 1; return max_sector - sector_nr; } |
6394cca54 [PATCH] md: fix r... |
2143 2144 2145 |
/* before building a request, check if we can skip these blocks.. * This call the bitmap_start_sync doesn't actually record anything */ |
e3b9703e2 [PATCH] md: yet a... |
2146 |
if (!bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, 1) && |
e5de485f0 [PATCH] md: make ... |
2147 |
!conf->fullsync && !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) { |
191ea9b2c [PATCH] md: raid1... |
2148 2149 2150 2151 |
/* We can skip this block, and probably several more */ *skipped = 1; return sync_blocks; } |
1da177e4c Linux-2.6.12-rc2 |
2152 |
/* |
17999be4a [PATCH] md: impro... |
2153 2154 2155 |
* If there is non-resync activity waiting for a turn, * and resync is going fast enough, * then let it though before starting on this new sync request. |
1da177e4c Linux-2.6.12-rc2 |
2156 |
*/ |
17999be4a [PATCH] md: impro... |
2157 |
if (!go_faster && conf->nr_waiting) |
1da177e4c Linux-2.6.12-rc2 |
2158 |
msleep_interruptible(1000); |
17999be4a [PATCH] md: impro... |
2159 |
|
b47490c9b md: Update md bit... |
2160 |
bitmap_cond_end_sync(mddev->bitmap, sector_nr); |
1c4588e9c md/raid1: perform... |
2161 |
r1_bio = mempool_alloc(conf->r1buf_pool, GFP_NOIO); |
17999be4a [PATCH] md: impro... |
2162 2163 2164 |
raise_barrier(conf); conf->next_resync = sector_nr; |
1da177e4c Linux-2.6.12-rc2 |
2165 |
|
3e198f782 [PATCH] md: tidyu... |
2166 |
rcu_read_lock(); |
1da177e4c Linux-2.6.12-rc2 |
2167 |
/* |
3e198f782 [PATCH] md: tidyu... |
2168 2169 2170 2171 2172 2173 |
* If we get a correctably read error during resync or recovery, * we might want to read from a different device. So we * flag all drives that could conceivably be read from for READ, * and any others (which will be non-In_sync devices) for WRITE. * If a read fails, we try reading from something else for which READ * is OK. |
1da177e4c Linux-2.6.12-rc2 |
2174 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
2175 |
|
1da177e4c Linux-2.6.12-rc2 |
2176 2177 |
r1_bio->mddev = mddev; r1_bio->sector = sector_nr; |
191ea9b2c [PATCH] md: raid1... |
2178 |
r1_bio->state = 0; |
1da177e4c Linux-2.6.12-rc2 |
2179 |
set_bit(R1BIO_IsSync, &r1_bio->state); |
1da177e4c Linux-2.6.12-rc2 |
2180 |
|
8f19ccb2f md/raid1: Allocat... |
2181 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
3cb030020 md: removing type... |
2182 |
struct md_rdev *rdev; |
1da177e4c Linux-2.6.12-rc2 |
2183 2184 2185 2186 |
bio = r1_bio->bios[i]; /* take from bio_init */ bio->bi_next = NULL; |
db8d9d359 md/raid1: minor b... |
2187 |
bio->bi_flags &= ~(BIO_POOL_MASK-1); |
1da177e4c Linux-2.6.12-rc2 |
2188 |
bio->bi_flags |= 1 << BIO_UPTODATE; |
802ba064c [PATCH] md: Don't... |
2189 |
bio->bi_rw = READ; |
1da177e4c Linux-2.6.12-rc2 |
2190 2191 2192 |
bio->bi_vcnt = 0; bio->bi_idx = 0; bio->bi_phys_segments = 0; |
1da177e4c Linux-2.6.12-rc2 |
2193 2194 2195 |
bio->bi_size = 0; bio->bi_end_io = NULL; bio->bi_private = NULL; |
3e198f782 [PATCH] md: tidyu... |
2196 2197 |
rdev = rcu_dereference(conf->mirrors[i].rdev); if (rdev == NULL || |
06f603851 md/raid1: avoid r... |
2198 |
test_bit(Faulty, &rdev->flags)) { |
8f19ccb2f md/raid1: Allocat... |
2199 2200 |
if (i < conf->raid_disks) still_degraded = 1; |
3e198f782 [PATCH] md: tidyu... |
2201 |
} else if (!test_bit(In_sync, &rdev->flags)) { |
1da177e4c Linux-2.6.12-rc2 |
2202 2203 2204 |
bio->bi_rw = WRITE; bio->bi_end_io = end_sync_write; write_targets ++; |
3e198f782 [PATCH] md: tidyu... |
2205 2206 |
} else { /* may need to read from here */ |
06f603851 md/raid1: avoid r... |
2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 |
sector_t first_bad = MaxSector; int bad_sectors; if (is_badblock(rdev, sector_nr, good_sectors, &first_bad, &bad_sectors)) { if (first_bad > sector_nr) good_sectors = first_bad - sector_nr; else { bad_sectors -= (sector_nr - first_bad); if (min_bad == 0 || min_bad > bad_sectors) min_bad = bad_sectors; } } if (sector_nr < first_bad) { if (test_bit(WriteMostly, &rdev->flags)) { if (wonly < 0) wonly = i; } else { if (disk < 0) disk = i; } bio->bi_rw = READ; bio->bi_end_io = end_sync_read; read_targets++; |
3e198f782 [PATCH] md: tidyu... |
2232 |
} |
3e198f782 [PATCH] md: tidyu... |
2233 |
} |
06f603851 md/raid1: avoid r... |
2234 2235 2236 2237 2238 2239 |
if (bio->bi_end_io) { atomic_inc(&rdev->nr_pending); bio->bi_sector = sector_nr + rdev->data_offset; bio->bi_bdev = rdev->bdev; bio->bi_private = r1_bio; } |
1da177e4c Linux-2.6.12-rc2 |
2240 |
} |
3e198f782 [PATCH] md: tidyu... |
2241 2242 2243 2244 |
rcu_read_unlock(); if (disk < 0) disk = wonly; r1_bio->read_disk = disk; |
191ea9b2c [PATCH] md: raid1... |
2245 |
|
06f603851 md/raid1: avoid r... |
2246 2247 2248 2249 2250 |
if (read_targets == 0 && min_bad > 0) { /* These sectors are bad on all InSync devices, so we * need to mark them bad on all write targets */ int ok = 1; |
8f19ccb2f md/raid1: Allocat... |
2251 |
for (i = 0 ; i < conf->raid_disks * 2 ; i++) |
06f603851 md/raid1: avoid r... |
2252 |
if (r1_bio->bios[i]->bi_end_io == end_sync_write) { |
3cb030020 md: removing type... |
2253 |
struct md_rdev *rdev = |
06f603851 md/raid1: avoid r... |
2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 |
rcu_dereference(conf->mirrors[i].rdev); ok = rdev_set_badblocks(rdev, sector_nr, min_bad, 0 ) && ok; } set_bit(MD_CHANGE_DEVS, &mddev->flags); *skipped = 1; put_buf(r1_bio); if (!ok) { /* Cannot record the badblocks, so need to * abort the resync. * If there are multiple read targets, could just * fail the really bad ones ??? */ conf->recovery_disabled = mddev->recovery_disabled; set_bit(MD_RECOVERY_INTR, &mddev->recovery); return 0; } else return min_bad; } if (min_bad > 0 && min_bad < good_sectors) { /* only resync enough to reach the next bad->good * transition */ good_sectors = min_bad; } |
3e198f782 [PATCH] md: tidyu... |
2281 2282 2283 2284 2285 |
if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) && read_targets > 0) /* extra read targets are also write targets */ write_targets += read_targets-1; if (write_targets == 0 || read_targets == 0) { |
1da177e4c Linux-2.6.12-rc2 |
2286 2287 2288 |
/* There is nowhere to write, so all non-sync * drives must be failed - so we are finished */ |
57afd89f9 [PATCH] md: impro... |
2289 2290 |
sector_t rv = max_sector - sector_nr; *skipped = 1; |
1da177e4c Linux-2.6.12-rc2 |
2291 |
put_buf(r1_bio); |
1da177e4c Linux-2.6.12-rc2 |
2292 2293 |
return rv; } |
c62072777 md: allow a maxim... |
2294 2295 |
if (max_sector > mddev->resync_max) max_sector = mddev->resync_max; /* Don't do IO beyond here */ |
06f603851 md/raid1: avoid r... |
2296 2297 |
if (max_sector > sector_nr + good_sectors) max_sector = sector_nr + good_sectors; |
1da177e4c Linux-2.6.12-rc2 |
2298 |
nr_sectors = 0; |
289e99e8e [PATCH] md: initi... |
2299 |
sync_blocks = 0; |
1da177e4c Linux-2.6.12-rc2 |
2300 2301 2302 2303 2304 2305 2306 |
do { struct page *page; int len = PAGE_SIZE; if (sector_nr + (len>>9) > max_sector) len = (max_sector - sector_nr) << 9; if (len == 0) break; |
6a806c510 [PATCH] md/raid1:... |
2307 2308 |
if (sync_blocks == 0) { if (!bitmap_start_sync(mddev->bitmap, sector_nr, |
e5de485f0 [PATCH] md: make ... |
2309 2310 2311 |
&sync_blocks, still_degraded) && !conf->fullsync && !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) |
6a806c510 [PATCH] md/raid1:... |
2312 |
break; |
9e77c485f BUG_ON() Conversi... |
2313 |
BUG_ON(sync_blocks < (PAGE_SIZE>>9)); |
7571ae887 md/raid1: avoid ... |
2314 |
if ((len >> 9) > sync_blocks) |
6a806c510 [PATCH] md/raid1:... |
2315 |
len = sync_blocks<<9; |
ab7a30c70 [PATCH] md: fix b... |
2316 |
} |
191ea9b2c [PATCH] md: raid1... |
2317 |
|
8f19ccb2f md/raid1: Allocat... |
2318 |
for (i = 0 ; i < conf->raid_disks * 2; i++) { |
1da177e4c Linux-2.6.12-rc2 |
2319 2320 |
bio = r1_bio->bios[i]; if (bio->bi_end_io) { |
d11c171e6 [PATCH] md: allow... |
2321 |
page = bio->bi_io_vec[bio->bi_vcnt].bv_page; |
1da177e4c Linux-2.6.12-rc2 |
2322 2323 |
if (bio_add_page(bio, page, len, 0) == 0) { /* stop here */ |
d11c171e6 [PATCH] md: allow... |
2324 |
bio->bi_io_vec[bio->bi_vcnt].bv_page = page; |
1da177e4c Linux-2.6.12-rc2 |
2325 2326 2327 |
while (i > 0) { i--; bio = r1_bio->bios[i]; |
6a806c510 [PATCH] md/raid1:... |
2328 2329 |
if (bio->bi_end_io==NULL) continue; |
1da177e4c Linux-2.6.12-rc2 |
2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 |
/* remove last page from this bio */ bio->bi_vcnt--; bio->bi_size -= len; bio->bi_flags &= ~(1<< BIO_SEG_VALID); } goto bio_full; } } } nr_sectors += len>>9; sector_nr += len>>9; |
191ea9b2c [PATCH] md: raid1... |
2341 |
sync_blocks -= (len>>9); |
1da177e4c Linux-2.6.12-rc2 |
2342 2343 |
} while (r1_bio->bios[disk]->bi_vcnt < RESYNC_PAGES); bio_full: |
1da177e4c Linux-2.6.12-rc2 |
2344 |
r1_bio->sectors = nr_sectors; |
d11c171e6 [PATCH] md: allow... |
2345 2346 2347 2348 2349 |
/* For a user-requested sync, we read all readable devices and do a * compare */ if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) { atomic_set(&r1_bio->remaining, read_targets); |
8f19ccb2f md/raid1: Allocat... |
2350 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
d11c171e6 [PATCH] md: allow... |
2351 2352 |
bio = r1_bio->bios[i]; if (bio->bi_end_io == end_sync_read) { |
ddac7c7e3 [PATCH] md: Fix i... |
2353 |
md_sync_acct(bio->bi_bdev, nr_sectors); |
d11c171e6 [PATCH] md: allow... |
2354 2355 2356 2357 2358 2359 |
generic_make_request(bio); } } } else { atomic_set(&r1_bio->remaining, 1); bio = r1_bio->bios[r1_bio->read_disk]; |
ddac7c7e3 [PATCH] md: Fix i... |
2360 |
md_sync_acct(bio->bi_bdev, nr_sectors); |
d11c171e6 [PATCH] md: allow... |
2361 |
generic_make_request(bio); |
1da177e4c Linux-2.6.12-rc2 |
2362 |
|
d11c171e6 [PATCH] md: allow... |
2363 |
} |
1da177e4c Linux-2.6.12-rc2 |
2364 2365 |
return nr_sectors; } |
fd01b88c7 md: remove typede... |
2366 |
static sector_t raid1_size(struct mddev *mddev, sector_t sectors, int raid_disks) |
80c3a6ce4 md: add 'size' as... |
2367 2368 2369 2370 2371 2372 |
{ if (sectors) return sectors; return mddev->dev_sectors; } |
e80963604 md/raid1: typedef... |
2373 |
static struct r1conf *setup_conf(struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
2374 |
{ |
e80963604 md/raid1: typedef... |
2375 |
struct r1conf *conf; |
709ae4879 md/raid1: add tak... |
2376 |
int i; |
0f6d02d58 md: remove typede... |
2377 |
struct mirror_info *disk; |
3cb030020 md: removing type... |
2378 |
struct md_rdev *rdev; |
709ae4879 md/raid1: add tak... |
2379 |
int err = -ENOMEM; |
1da177e4c Linux-2.6.12-rc2 |
2380 |
|
e80963604 md/raid1: typedef... |
2381 |
conf = kzalloc(sizeof(struct r1conf), GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
2382 |
if (!conf) |
709ae4879 md/raid1: add tak... |
2383 |
goto abort; |
1da177e4c Linux-2.6.12-rc2 |
2384 |
|
8f19ccb2f md/raid1: Allocat... |
2385 2386 |
conf->mirrors = kzalloc(sizeof(struct mirror_info) * mddev->raid_disks * 2, |
1da177e4c Linux-2.6.12-rc2 |
2387 2388 |
GFP_KERNEL); if (!conf->mirrors) |
709ae4879 md/raid1: add tak... |
2389 |
goto abort; |
1da177e4c Linux-2.6.12-rc2 |
2390 |
|
ddaf22aba [PATCH] md: attem... |
2391 2392 |
conf->tmppage = alloc_page(GFP_KERNEL); if (!conf->tmppage) |
709ae4879 md/raid1: add tak... |
2393 |
goto abort; |
ddaf22aba [PATCH] md: attem... |
2394 |
|
709ae4879 md/raid1: add tak... |
2395 |
conf->poolinfo = kzalloc(sizeof(*conf->poolinfo), GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
2396 |
if (!conf->poolinfo) |
709ae4879 md/raid1: add tak... |
2397 |
goto abort; |
8f19ccb2f md/raid1: Allocat... |
2398 |
conf->poolinfo->raid_disks = mddev->raid_disks * 2; |
1da177e4c Linux-2.6.12-rc2 |
2399 2400 2401 2402 |
conf->r1bio_pool = mempool_create(NR_RAID1_BIOS, r1bio_pool_alloc, r1bio_pool_free, conf->poolinfo); if (!conf->r1bio_pool) |
709ae4879 md/raid1: add tak... |
2403 |
goto abort; |
ed9bfdf1a md: raid1/raid10:... |
2404 |
conf->poolinfo->mddev = mddev; |
1da177e4c Linux-2.6.12-rc2 |
2405 |
|
c19d57980 md/raid1: recogni... |
2406 |
err = -EINVAL; |
e7e72bf64 Remove blkdev war... |
2407 |
spin_lock_init(&conf->device_lock); |
159ec1fc0 md: use list_for_... |
2408 |
list_for_each_entry(rdev, &mddev->disks, same_set) { |
709ae4879 md/raid1: add tak... |
2409 |
int disk_idx = rdev->raid_disk; |
1da177e4c Linux-2.6.12-rc2 |
2410 2411 2412 |
if (disk_idx >= mddev->raid_disks || disk_idx < 0) continue; |
c19d57980 md/raid1: recogni... |
2413 2414 2415 2416 |
if (test_bit(Replacement, &rdev->flags)) disk = conf->mirrors + conf->raid_disks + disk_idx; else disk = conf->mirrors + disk_idx; |
1da177e4c Linux-2.6.12-rc2 |
2417 |
|
c19d57980 md/raid1: recogni... |
2418 2419 |
if (disk->rdev) goto abort; |
1da177e4c Linux-2.6.12-rc2 |
2420 |
disk->rdev = rdev; |
1da177e4c Linux-2.6.12-rc2 |
2421 2422 |
disk->head_position = 0; |
1da177e4c Linux-2.6.12-rc2 |
2423 2424 2425 |
} conf->raid_disks = mddev->raid_disks; conf->mddev = mddev; |
1da177e4c Linux-2.6.12-rc2 |
2426 |
INIT_LIST_HEAD(&conf->retry_list); |
1da177e4c Linux-2.6.12-rc2 |
2427 2428 |
spin_lock_init(&conf->resync_lock); |
17999be4a [PATCH] md: impro... |
2429 |
init_waitqueue_head(&conf->wait_barrier); |
1da177e4c Linux-2.6.12-rc2 |
2430 |
|
191ea9b2c [PATCH] md: raid1... |
2431 |
bio_list_init(&conf->pending_bio_list); |
34db0cd60 md: add proper wr... |
2432 |
conf->pending_count = 0; |
d890fa2b0 md: Fix some bugs... |
2433 |
conf->recovery_disabled = mddev->recovery_disabled - 1; |
191ea9b2c [PATCH] md: raid1... |
2434 |
|
c19d57980 md/raid1: recogni... |
2435 |
err = -EIO; |
709ae4879 md/raid1: add tak... |
2436 |
conf->last_used = -1; |
8f19ccb2f md/raid1: Allocat... |
2437 |
for (i = 0; i < conf->raid_disks * 2; i++) { |
1da177e4c Linux-2.6.12-rc2 |
2438 2439 |
disk = conf->mirrors + i; |
c19d57980 md/raid1: recogni... |
2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 |
if (i < conf->raid_disks && disk[conf->raid_disks].rdev) { /* This slot has a replacement. */ if (!disk->rdev) { /* No original, just make the replacement * a recovering spare */ disk->rdev = disk[conf->raid_disks].rdev; disk[conf->raid_disks].rdev = NULL; } else if (!test_bit(In_sync, &disk->rdev->flags)) /* Original is not in_sync - bad */ goto abort; } |
5fd6c1dce [PATCH] md: allow... |
2454 2455 |
if (!disk->rdev || !test_bit(In_sync, &disk->rdev->flags)) { |
1da177e4c Linux-2.6.12-rc2 |
2456 |
disk->head_position = 0; |
918f02383 md: make sure a r... |
2457 2458 |
if (disk->rdev) conf->fullsync = 1; |
709ae4879 md/raid1: add tak... |
2459 2460 2461 2462 2463 2464 |
} else if (conf->last_used < 0) /* * The first working device is used as a * starting point to read balancing. */ conf->last_used = i; |
1da177e4c Linux-2.6.12-rc2 |
2465 |
} |
709ae4879 md/raid1: add tak... |
2466 |
|
709ae4879 md/raid1: add tak... |
2467 |
if (conf->last_used < 0) { |
9dd1e2faf md/raid1: improve... |
2468 2469 |
printk(KERN_ERR "md/raid1:%s: no operational mirrors ", |
709ae4879 md/raid1: add tak... |
2470 2471 2472 2473 2474 2475 2476 |
mdname(mddev)); goto abort; } err = -ENOMEM; conf->thread = md_register_thread(raid1d, mddev, NULL); if (!conf->thread) { printk(KERN_ERR |
9dd1e2faf md/raid1: improve... |
2477 2478 |
"md/raid1:%s: couldn't allocate thread ", |
709ae4879 md/raid1: add tak... |
2479 2480 |
mdname(mddev)); goto abort; |
11ce99e62 [PATCH] md: Remov... |
2481 |
} |
1da177e4c Linux-2.6.12-rc2 |
2482 |
|
709ae4879 md/raid1: add tak... |
2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 |
return conf; abort: if (conf) { if (conf->r1bio_pool) mempool_destroy(conf->r1bio_pool); kfree(conf->mirrors); safe_put_page(conf->tmppage); kfree(conf->poolinfo); kfree(conf); } return ERR_PTR(err); } |
fd01b88c7 md: remove typede... |
2496 |
static int run(struct mddev *mddev) |
709ae4879 md/raid1: add tak... |
2497 |
{ |
e80963604 md/raid1: typedef... |
2498 |
struct r1conf *conf; |
709ae4879 md/raid1: add tak... |
2499 |
int i; |
3cb030020 md: removing type... |
2500 |
struct md_rdev *rdev; |
709ae4879 md/raid1: add tak... |
2501 2502 |
if (mddev->level != 1) { |
9dd1e2faf md/raid1: improve... |
2503 2504 |
printk(KERN_ERR "md/raid1:%s: raid level not set to mirroring (%d) ", |
709ae4879 md/raid1: add tak... |
2505 2506 2507 2508 |
mdname(mddev), mddev->level); return -EIO; } if (mddev->reshape_position != MaxSector) { |
9dd1e2faf md/raid1: improve... |
2509 2510 |
printk(KERN_ERR "md/raid1:%s: reshape_position set but not supported ", |
709ae4879 md/raid1: add tak... |
2511 2512 2513 |
mdname(mddev)); return -EIO; } |
1da177e4c Linux-2.6.12-rc2 |
2514 |
/* |
709ae4879 md/raid1: add tak... |
2515 2516 2517 |
* copy the already verified devices into our private RAID1 * bookkeeping area. [whatever we allocate in run(), * should be freed in stop()] |
1da177e4c Linux-2.6.12-rc2 |
2518 |
*/ |
709ae4879 md/raid1: add tak... |
2519 2520 2521 2522 |
if (mddev->private == NULL) conf = setup_conf(mddev); else conf = mddev->private; |
1da177e4c Linux-2.6.12-rc2 |
2523 |
|
709ae4879 md/raid1: add tak... |
2524 2525 |
if (IS_ERR(conf)) return PTR_ERR(conf); |
1da177e4c Linux-2.6.12-rc2 |
2526 |
|
709ae4879 md/raid1: add tak... |
2527 |
list_for_each_entry(rdev, &mddev->disks, same_set) { |
1ed7242e5 MD: raid1 changes... |
2528 2529 |
if (!mddev->gendisk) continue; |
709ae4879 md/raid1: add tak... |
2530 2531 2532 |
disk_stack_limits(mddev->gendisk, rdev->bdev, rdev->data_offset << 9); /* as we don't honour merge_bvec_fn, we must never risk |
627a2d3c2 md: deal with mer... |
2533 2534 |
* violating it, so limit ->max_segments to 1 lying within * a single page, as a one page request is never in violation. |
709ae4879 md/raid1: add tak... |
2535 |
*/ |
627a2d3c2 md: deal with mer... |
2536 2537 2538 2539 2540 |
if (rdev->bdev->bd_disk->queue->merge_bvec_fn) { blk_queue_max_segments(mddev->queue, 1); blk_queue_segment_boundary(mddev->queue, PAGE_CACHE_SIZE - 1); } |
1da177e4c Linux-2.6.12-rc2 |
2541 |
} |
191ea9b2c [PATCH] md: raid1... |
2542 |
|
709ae4879 md/raid1: add tak... |
2543 2544 2545 2546 2547 2548 2549 2550 2551 |
mddev->degraded = 0; for (i=0; i < conf->raid_disks; i++) if (conf->mirrors[i].rdev == NULL || !test_bit(In_sync, &conf->mirrors[i].rdev->flags) || test_bit(Faulty, &conf->mirrors[i].rdev->flags)) mddev->degraded++; if (conf->raid_disks - mddev->degraded == 1) mddev->recovery_cp = MaxSector; |
8c6ac868b md: Push down rec... |
2552 |
if (mddev->recovery_cp != MaxSector) |
9dd1e2faf md/raid1: improve... |
2553 |
printk(KERN_NOTICE "md/raid1:%s: not clean" |
8c6ac868b md: Push down rec... |
2554 2555 2556 |
" -- starting background reconstruction ", mdname(mddev)); |
1da177e4c Linux-2.6.12-rc2 |
2557 |
printk(KERN_INFO |
9dd1e2faf md/raid1: improve... |
2558 2559 |
"md/raid1:%s: active with %d out of %d mirrors ", |
1da177e4c Linux-2.6.12-rc2 |
2560 2561 |
mdname(mddev), mddev->raid_disks - mddev->degraded, mddev->raid_disks); |
709ae4879 md/raid1: add tak... |
2562 |
|
1da177e4c Linux-2.6.12-rc2 |
2563 2564 2565 |
/* * Ok, everything is just fine now */ |
709ae4879 md/raid1: add tak... |
2566 2567 2568 |
mddev->thread = conf->thread; conf->thread = NULL; mddev->private = conf; |
1f403624b md: centralize ->... |
2569 |
md_set_array_sectors(mddev, raid1_size(mddev, 0, 0)); |
1da177e4c Linux-2.6.12-rc2 |
2570 |
|
1ed7242e5 MD: raid1 changes... |
2571 2572 2573 2574 |
if (mddev->queue) { mddev->queue->backing_dev_info.congested_fn = raid1_congested; mddev->queue->backing_dev_info.congested_data = mddev; } |
a91a2785b block: Require su... |
2575 |
return md_integrity_register(mddev); |
1da177e4c Linux-2.6.12-rc2 |
2576 |
} |
fd01b88c7 md: remove typede... |
2577 |
static int stop(struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
2578 |
{ |
e80963604 md/raid1: typedef... |
2579 |
struct r1conf *conf = mddev->private; |
4b6d287f6 [PATCH] md: add w... |
2580 |
struct bitmap *bitmap = mddev->bitmap; |
4b6d287f6 [PATCH] md: add w... |
2581 2582 |
/* wait for behind writes to complete */ |
e555190d8 md/raid1: delay r... |
2583 |
if (bitmap && atomic_read(&bitmap->behind_writes) > 0) { |
9dd1e2faf md/raid1: improve... |
2584 2585 2586 |
printk(KERN_INFO "md/raid1:%s: behind writes in progress - waiting to stop. ", mdname(mddev)); |
4b6d287f6 [PATCH] md: add w... |
2587 |
/* need to kick something here to make sure I/O goes? */ |
e555190d8 md/raid1: delay r... |
2588 2589 |
wait_event(bitmap->behind_wait, atomic_read(&bitmap->behind_writes) == 0); |
4b6d287f6 [PATCH] md: add w... |
2590 |
} |
1da177e4c Linux-2.6.12-rc2 |
2591 |
|
409c57f38 md: enable suspen... |
2592 2593 |
raise_barrier(conf); lower_barrier(conf); |
01f96c0a9 md: Avoid waking ... |
2594 |
md_unregister_thread(&mddev->thread); |
1da177e4c Linux-2.6.12-rc2 |
2595 2596 |
if (conf->r1bio_pool) mempool_destroy(conf->r1bio_pool); |
990a8baf5 [PATCH] md: remov... |
2597 2598 |
kfree(conf->mirrors); kfree(conf->poolinfo); |
1da177e4c Linux-2.6.12-rc2 |
2599 2600 2601 2602 |
kfree(conf); mddev->private = NULL; return 0; } |
fd01b88c7 md: remove typede... |
2603 |
static int raid1_resize(struct mddev *mddev, sector_t sectors) |
1da177e4c Linux-2.6.12-rc2 |
2604 2605 2606 2607 2608 2609 2610 2611 |
{ /* 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. */ |
1f403624b md: centralize ->... |
2612 |
md_set_array_sectors(mddev, raid1_size(mddev, sectors, 0)); |
b522adcde md: 'array_size' ... |
2613 2614 |
if (mddev->array_sectors > raid1_size(mddev, sectors, 0)) return -EINVAL; |
f233ea5c9 md: Make mddev->a... |
2615 |
set_capacity(mddev->gendisk, mddev->array_sectors); |
449aad3e2 md: Use revalidat... |
2616 |
revalidate_disk(mddev->gendisk); |
b522adcde md: 'array_size' ... |
2617 |
if (sectors > mddev->dev_sectors && |
b098636cf md: allow resync_... |
2618 |
mddev->recovery_cp > mddev->dev_sectors) { |
58c0fed40 md: Make mddev->s... |
2619 |
mddev->recovery_cp = mddev->dev_sectors; |
1da177e4c Linux-2.6.12-rc2 |
2620 2621 |
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); } |
b522adcde md: 'array_size' ... |
2622 |
mddev->dev_sectors = sectors; |
4b5c7ae83 [PATCH] md: when ... |
2623 |
mddev->resync_max_sectors = sectors; |
1da177e4c Linux-2.6.12-rc2 |
2624 2625 |
return 0; } |
fd01b88c7 md: remove typede... |
2626 |
static int raid1_reshape(struct mddev *mddev) |
1da177e4c Linux-2.6.12-rc2 |
2627 2628 2629 2630 2631 2632 2633 2634 |
{ /* We need to: * 1/ resize the r1bio_pool * 2/ resize conf->mirrors * * We allocate a new r1bio_pool if we can. * Then raise a device barrier and wait until all IO stops. * Then resize conf->mirrors and swap in the new r1bio pool. |
6ea9c07c6 [PATCH] md: cause... |
2635 2636 2637 |
* * At the same time, we "pack" the devices so that all the missing * devices have the higher raid_disk numbers. |
1da177e4c Linux-2.6.12-rc2 |
2638 2639 2640 |
*/ mempool_t *newpool, *oldpool; struct pool_info *newpoolinfo; |
0f6d02d58 md: remove typede... |
2641 |
struct mirror_info *newmirrors; |
e80963604 md/raid1: typedef... |
2642 |
struct r1conf *conf = mddev->private; |
63c70c4f3 [PATCH] md: Split... |
2643 |
int cnt, raid_disks; |
c04be0aa8 [PATCH] md: Impro... |
2644 |
unsigned long flags; |
b5470dc5f md: resolve exter... |
2645 |
int d, d2, err; |
1da177e4c Linux-2.6.12-rc2 |
2646 |
|
63c70c4f3 [PATCH] md: Split... |
2647 |
/* Cannot change chunk_size, layout, or level */ |
664e7c413 md: Convert mddev... |
2648 |
if (mddev->chunk_sectors != mddev->new_chunk_sectors || |
63c70c4f3 [PATCH] md: Split... |
2649 2650 |
mddev->layout != mddev->new_layout || mddev->level != mddev->new_level) { |
664e7c413 md: Convert mddev... |
2651 |
mddev->new_chunk_sectors = mddev->chunk_sectors; |
63c70c4f3 [PATCH] md: Split... |
2652 2653 2654 2655 |
mddev->new_layout = mddev->layout; mddev->new_level = mddev->level; return -EINVAL; } |
b5470dc5f md: resolve exter... |
2656 2657 2658 |
err = md_allow_write(mddev); if (err) return err; |
2a2275d63 [PATCH] md: fix p... |
2659 |
|
63c70c4f3 [PATCH] md: Split... |
2660 |
raid_disks = mddev->raid_disks + mddev->delta_disks; |
6ea9c07c6 [PATCH] md: cause... |
2661 2662 2663 2664 2665 2666 |
if (raid_disks < conf->raid_disks) { cnt=0; for (d= 0; d < conf->raid_disks; d++) if (conf->mirrors[d].rdev) cnt++; if (cnt > raid_disks) |
1da177e4c Linux-2.6.12-rc2 |
2667 |
return -EBUSY; |
6ea9c07c6 [PATCH] md: cause... |
2668 |
} |
1da177e4c Linux-2.6.12-rc2 |
2669 2670 2671 2672 2673 |
newpoolinfo = kmalloc(sizeof(*newpoolinfo), GFP_KERNEL); if (!newpoolinfo) return -ENOMEM; newpoolinfo->mddev = mddev; |
8f19ccb2f md/raid1: Allocat... |
2674 |
newpoolinfo->raid_disks = raid_disks * 2; |
1da177e4c Linux-2.6.12-rc2 |
2675 2676 2677 2678 2679 2680 2681 |
newpool = mempool_create(NR_RAID1_BIOS, r1bio_pool_alloc, r1bio_pool_free, newpoolinfo); if (!newpool) { kfree(newpoolinfo); return -ENOMEM; } |
8f19ccb2f md/raid1: Allocat... |
2682 2683 |
newmirrors = kzalloc(sizeof(struct mirror_info) * raid_disks * 2, GFP_KERNEL); |
1da177e4c Linux-2.6.12-rc2 |
2684 2685 2686 2687 2688 |
if (!newmirrors) { kfree(newpoolinfo); mempool_destroy(newpool); return -ENOMEM; } |
1da177e4c Linux-2.6.12-rc2 |
2689 |
|
17999be4a [PATCH] md: impro... |
2690 |
raise_barrier(conf); |
1da177e4c Linux-2.6.12-rc2 |
2691 2692 2693 2694 |
/* ok, everything is stopped */ oldpool = conf->r1bio_pool; conf->r1bio_pool = newpool; |
6ea9c07c6 [PATCH] md: cause... |
2695 |
|
a88aa7865 md: correctly upd... |
2696 |
for (d = d2 = 0; d < conf->raid_disks; d++) { |
3cb030020 md: removing type... |
2697 |
struct md_rdev *rdev = conf->mirrors[d].rdev; |
a88aa7865 md: correctly upd... |
2698 |
if (rdev && rdev->raid_disk != d2) { |
36fad858a md: introduce lin... |
2699 |
sysfs_unlink_rdev(mddev, rdev); |
a88aa7865 md: correctly upd... |
2700 |
rdev->raid_disk = d2; |
36fad858a md: introduce lin... |
2701 2702 |
sysfs_unlink_rdev(mddev, rdev); if (sysfs_link_rdev(mddev, rdev)) |
a88aa7865 md: correctly upd... |
2703 |
printk(KERN_WARNING |
36fad858a md: introduce lin... |
2704 2705 2706 |
"md/raid1:%s: cannot register rd%d ", mdname(mddev), rdev->raid_disk); |
6ea9c07c6 [PATCH] md: cause... |
2707 |
} |
a88aa7865 md: correctly upd... |
2708 2709 2710 |
if (rdev) newmirrors[d2++].rdev = rdev; } |
1da177e4c Linux-2.6.12-rc2 |
2711 2712 2713 2714 |
kfree(conf->mirrors); conf->mirrors = newmirrors; kfree(conf->poolinfo); conf->poolinfo = newpoolinfo; |
c04be0aa8 [PATCH] md: Impro... |
2715 |
spin_lock_irqsave(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
2716 |
mddev->degraded += (raid_disks - conf->raid_disks); |
c04be0aa8 [PATCH] md: Impro... |
2717 |
spin_unlock_irqrestore(&conf->device_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
2718 |
conf->raid_disks = mddev->raid_disks = raid_disks; |
63c70c4f3 [PATCH] md: Split... |
2719 |
mddev->delta_disks = 0; |
1da177e4c Linux-2.6.12-rc2 |
2720 |
|
6ea9c07c6 [PATCH] md: cause... |
2721 |
conf->last_used = 0; /* just make sure it is in-range */ |
17999be4a [PATCH] md: impro... |
2722 |
lower_barrier(conf); |
1da177e4c Linux-2.6.12-rc2 |
2723 2724 2725 2726 2727 2728 2729 |
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); md_wakeup_thread(mddev->thread); mempool_destroy(oldpool); return 0; } |
fd01b88c7 md: remove typede... |
2730 |
static void raid1_quiesce(struct mddev *mddev, int state) |
36fa30636 [PATCH] md: all h... |
2731 |
{ |
e80963604 md/raid1: typedef... |
2732 |
struct r1conf *conf = mddev->private; |
36fa30636 [PATCH] md: all h... |
2733 2734 |
switch(state) { |
6eef4b21f md: add honouring... |
2735 2736 2737 |
case 2: /* wake for suspend */ wake_up(&conf->wait_barrier); break; |
9e6603da9 [PATCH] md: raid1... |
2738 |
case 1: |
17999be4a [PATCH] md: impro... |
2739 |
raise_barrier(conf); |
36fa30636 [PATCH] md: all h... |
2740 |
break; |
9e6603da9 [PATCH] md: raid1... |
2741 |
case 0: |
17999be4a [PATCH] md: impro... |
2742 |
lower_barrier(conf); |
36fa30636 [PATCH] md: all h... |
2743 2744 |
break; } |
36fa30636 [PATCH] md: all h... |
2745 |
} |
fd01b88c7 md: remove typede... |
2746 |
static void *raid1_takeover(struct mddev *mddev) |
709ae4879 md/raid1: add tak... |
2747 2748 2749 2750 2751 |
{ /* raid1 can take over: * raid5 with 2 devices, any layout or chunk size */ if (mddev->level == 5 && mddev->raid_disks == 2) { |
e80963604 md/raid1: typedef... |
2752 |
struct r1conf *conf; |
709ae4879 md/raid1: add tak... |
2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 |
mddev->new_level = 1; mddev->new_layout = 0; mddev->new_chunk_sectors = 0; conf = setup_conf(mddev); if (!IS_ERR(conf)) conf->barrier = 1; return conf; } return ERR_PTR(-EINVAL); } |
1da177e4c Linux-2.6.12-rc2 |
2763 |
|
84fc4b56d md: rename "mdk_p... |
2764 |
static struct md_personality raid1_personality = |
1da177e4c Linux-2.6.12-rc2 |
2765 2766 |
{ .name = "raid1", |
2604b703b [PATCH] md: remov... |
2767 |
.level = 1, |
1da177e4c Linux-2.6.12-rc2 |
2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 |
.owner = THIS_MODULE, .make_request = make_request, .run = run, .stop = stop, .status = status, .error_handler = error, .hot_add_disk = raid1_add_disk, .hot_remove_disk= raid1_remove_disk, .spare_active = raid1_spare_active, .sync_request = sync_request, .resize = raid1_resize, |
80c3a6ce4 md: add 'size' as... |
2779 |
.size = raid1_size, |
63c70c4f3 [PATCH] md: Split... |
2780 |
.check_reshape = raid1_reshape, |
36fa30636 [PATCH] md: all h... |
2781 |
.quiesce = raid1_quiesce, |
709ae4879 md/raid1: add tak... |
2782 |
.takeover = raid1_takeover, |
1da177e4c Linux-2.6.12-rc2 |
2783 2784 2785 2786 |
}; static int __init raid_init(void) { |
2604b703b [PATCH] md: remov... |
2787 |
return register_md_personality(&raid1_personality); |
1da177e4c Linux-2.6.12-rc2 |
2788 2789 2790 2791 |
} static void raid_exit(void) { |
2604b703b [PATCH] md: remov... |
2792 |
unregister_md_personality(&raid1_personality); |
1da177e4c Linux-2.6.12-rc2 |
2793 2794 2795 2796 2797 |
} module_init(raid_init); module_exit(raid_exit); MODULE_LICENSE("GPL"); |
0efb9e619 md: add MODULE_DE... |
2798 |
MODULE_DESCRIPTION("RAID1 (mirroring) personality for MD"); |
1da177e4c Linux-2.6.12-rc2 |
2799 |
MODULE_ALIAS("md-personality-3"); /* RAID1 */ |
d9d166c2a [PATCH] md: allow... |
2800 |
MODULE_ALIAS("md-raid1"); |
2604b703b [PATCH] md: remov... |
2801 |
MODULE_ALIAS("md-level-1"); |
34db0cd60 md: add proper wr... |
2802 2803 |
module_param(max_queued_requests, int, S_IRUGO|S_IWUSR); |