raid5.c 116.4 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4
/*
 * raid5.c : Multiple Devices driver for Linux
 *	   Copyright (C) 1996, 1997 Ingo Molnar, Miguel de Icaza, Gadi Oxman
 *	   Copyright (C) 1999, 2000 Ingo Molnar
5
 *	   Copyright (C) 2002, 2003 H. Peter Anvin
L
Linus Torvalds 已提交
6
 *
7 8 9
 * RAID-4/5/6 management functions.
 * Thanks to Penguin Computing for making the RAID-6 development possible
 * by donating a test server!
L
Linus Torvalds 已提交
10 11 12 13 14 15 16 17 18 19 20
 *
 * 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.
 */

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
/*
 * BITMAP UNPLUGGING:
 *
 * The sequencing for updating the bitmap reliably is a little
 * subtle (and I got it wrong the first time) so it deserves some
 * explanation.
 *
 * We group bitmap updates into batches.  Each batch has a number.
 * We may write out several batches at once, but that isn't very important.
 * conf->bm_write is the number of the last batch successfully written.
 * conf->bm_flush is the number of the last batch that was closed to
 *    new additions.
 * When we discover that we will need to write to any block in a stripe
 * (in add_stripe_bio) we update the in-memory bitmap and record in sh->bm_seq
 * the number of the batch it will be in. This is bm_flush+1.
 * When we are ready to do a write, if that batch hasn't been written yet,
 *   we plug the array and queue the stripe for later.
 * When an unplug happens, we increment bm_flush, thus closing the current
 *   batch.
 * When we notice that bm_flush > bm_write, we write out all pending updates
 * to the bitmap, and advance bm_write to where bm_flush was.
 * This may occasionally write a bit out twice, but is sure never to
 * miss any bits.
 */
L
Linus Torvalds 已提交
45 46 47 48 49

#include <linux/module.h>
#include <linux/slab.h>
#include <linux/highmem.h>
#include <linux/bitops.h>
50
#include <linux/kthread.h>
L
Linus Torvalds 已提交
51
#include <asm/atomic.h>
52
#include "raid6.h"
L
Linus Torvalds 已提交
53

54 55
#include <linux/raid/bitmap.h>

L
Linus Torvalds 已提交
56 57 58 59 60 61 62 63 64
/*
 * Stripe cache
 */

#define NR_STRIPES		256
#define STRIPE_SIZE		PAGE_SIZE
#define STRIPE_SHIFT		(PAGE_SHIFT - 9)
#define STRIPE_SECTORS		(STRIPE_SIZE>>9)
#define	IO_THRESHOLD		1
65
#define NR_HASH			(PAGE_SIZE / sizeof(struct hlist_head))
L
Linus Torvalds 已提交
66 67
#define HASH_MASK		(NR_HASH - 1)

68
#define stripe_hash(conf, sect)	(&((conf)->stripe_hashtbl[((sect) >> STRIPE_SHIFT) & HASH_MASK]))
L
Linus Torvalds 已提交
69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96

/* bio's attached to a stripe+device for I/O are linked together in bi_sector
 * order without overlap.  There may be several bio's per stripe+device, and
 * a bio could span several devices.
 * When walking this list for a particular stripe+device, we must never proceed
 * beyond a bio that extends past this device, as the next bio might no longer
 * be valid.
 * This macro is used to determine the 'next' bio in the list, given the sector
 * of the current stripe+device
 */
#define r5_next_bio(bio, sect) ( ( (bio)->bi_sector + ((bio)->bi_size>>9) < sect + STRIPE_SECTORS) ? (bio)->bi_next : NULL)
/*
 * The following can be used to debug the driver
 */
#define RAID5_DEBUG	0
#define RAID5_PARANOIA	1
#if RAID5_PARANOIA && defined(CONFIG_SMP)
# define CHECK_DEVLOCK() assert_spin_locked(&conf->device_lock)
#else
# define CHECK_DEVLOCK()
#endif

#define PRINTK(x...) ((void)(RAID5_DEBUG && printk(x)))
#if RAID5_DEBUG
#define inline
#define __inline__
#endif

97 98 99 100 101 102 103 104 105 106
#if !RAID6_USE_EMPTY_ZERO_PAGE
/* In .bss so it's zeroed */
const char raid6_empty_zero_page[PAGE_SIZE] __attribute__((aligned(256)));
#endif

static inline int raid6_next_disk(int disk, int raid_disks)
{
	disk++;
	return (disk < raid_disks) ? disk : 0;
}
L
Linus Torvalds 已提交
107 108
static void print_raid5_conf (raid5_conf_t *conf);

109
static void __release_stripe(raid5_conf_t *conf, struct stripe_head *sh)
L
Linus Torvalds 已提交
110 111
{
	if (atomic_dec_and_test(&sh->count)) {
112 113
		BUG_ON(!list_empty(&sh->lru));
		BUG_ON(atomic_read(&conf->active_stripes)==0);
L
Linus Torvalds 已提交
114
		if (test_bit(STRIPE_HANDLE, &sh->state)) {
115
			if (test_bit(STRIPE_DELAYED, &sh->state)) {
L
Linus Torvalds 已提交
116
				list_add_tail(&sh->lru, &conf->delayed_list);
117 118
				blk_plug_device(conf->mddev->queue);
			} else if (test_bit(STRIPE_BIT_DELAY, &sh->state) &&
119
				   sh->bm_seq - conf->seq_write > 0) {
120
				list_add_tail(&sh->lru, &conf->bitmap_list);
121 122
				blk_plug_device(conf->mddev->queue);
			} else {
123
				clear_bit(STRIPE_BIT_DELAY, &sh->state);
L
Linus Torvalds 已提交
124
				list_add_tail(&sh->lru, &conf->handle_list);
125
			}
L
Linus Torvalds 已提交
126 127 128 129 130 131 132 133
			md_wakeup_thread(conf->mddev->thread);
		} else {
			if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) {
				atomic_dec(&conf->preread_active_stripes);
				if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD)
					md_wakeup_thread(conf->mddev->thread);
			}
			atomic_dec(&conf->active_stripes);
134 135
			if (!test_bit(STRIPE_EXPANDING, &sh->state)) {
				list_add_tail(&sh->lru, &conf->inactive_list);
L
Linus Torvalds 已提交
136
				wake_up(&conf->wait_for_stripe);
137 138
				if (conf->retry_read_aligned)
					md_wakeup_thread(conf->mddev->thread);
139
			}
L
Linus Torvalds 已提交
140 141 142 143 144 145 146
		}
	}
}
static void release_stripe(struct stripe_head *sh)
{
	raid5_conf_t *conf = sh->raid_conf;
	unsigned long flags;
147

L
Linus Torvalds 已提交
148 149 150 151 152
	spin_lock_irqsave(&conf->device_lock, flags);
	__release_stripe(conf, sh);
	spin_unlock_irqrestore(&conf->device_lock, flags);
}

153
static inline void remove_hash(struct stripe_head *sh)
L
Linus Torvalds 已提交
154 155 156
{
	PRINTK("remove_hash(), stripe %llu\n", (unsigned long long)sh->sector);

157
	hlist_del_init(&sh->hash);
L
Linus Torvalds 已提交
158 159
}

160
static inline void insert_hash(raid5_conf_t *conf, struct stripe_head *sh)
L
Linus Torvalds 已提交
161
{
162
	struct hlist_head *hp = stripe_hash(conf, sh->sector);
L
Linus Torvalds 已提交
163 164 165 166

	PRINTK("insert_hash(), stripe %llu\n", (unsigned long long)sh->sector);

	CHECK_DEVLOCK();
167
	hlist_add_head(&sh->hash, hp);
L
Linus Torvalds 已提交
168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
}


/* find an idle stripe, make sure it is unhashed, and return it. */
static struct stripe_head *get_free_stripe(raid5_conf_t *conf)
{
	struct stripe_head *sh = NULL;
	struct list_head *first;

	CHECK_DEVLOCK();
	if (list_empty(&conf->inactive_list))
		goto out;
	first = conf->inactive_list.next;
	sh = list_entry(first, struct stripe_head, lru);
	list_del_init(first);
	remove_hash(sh);
	atomic_inc(&conf->active_stripes);
out:
	return sh;
}

static void shrink_buffers(struct stripe_head *sh, int num)
{
	struct page *p;
	int i;

	for (i=0; i<num ; i++) {
		p = sh->dev[i].page;
		if (!p)
			continue;
		sh->dev[i].page = NULL;
199
		put_page(p);
L
Linus Torvalds 已提交
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219
	}
}

static int grow_buffers(struct stripe_head *sh, int num)
{
	int i;

	for (i=0; i<num; i++) {
		struct page *page;

		if (!(page = alloc_page(GFP_KERNEL))) {
			return 1;
		}
		sh->dev[i].page = page;
	}
	return 0;
}

static void raid5_build_block (struct stripe_head *sh, int i);

220
static void init_stripe(struct stripe_head *sh, sector_t sector, int pd_idx, int disks)
L
Linus Torvalds 已提交
221 222
{
	raid5_conf_t *conf = sh->raid_conf;
223
	int i;
L
Linus Torvalds 已提交
224

225 226
	BUG_ON(atomic_read(&sh->count) != 0);
	BUG_ON(test_bit(STRIPE_HANDLE, &sh->state));
L
Linus Torvalds 已提交
227 228 229 230 231 232
	
	CHECK_DEVLOCK();
	PRINTK("init_stripe called, stripe %llu\n", 
		(unsigned long long)sh->sector);

	remove_hash(sh);
233

L
Linus Torvalds 已提交
234 235 236 237
	sh->sector = sector;
	sh->pd_idx = pd_idx;
	sh->state = 0;

238 239 240
	sh->disks = disks;

	for (i = sh->disks; i--; ) {
L
Linus Torvalds 已提交
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
		struct r5dev *dev = &sh->dev[i];

		if (dev->toread || dev->towrite || dev->written ||
		    test_bit(R5_LOCKED, &dev->flags)) {
			printk("sector=%llx i=%d %p %p %p %d\n",
			       (unsigned long long)sh->sector, i, dev->toread,
			       dev->towrite, dev->written,
			       test_bit(R5_LOCKED, &dev->flags));
			BUG();
		}
		dev->flags = 0;
		raid5_build_block(sh, i);
	}
	insert_hash(conf, sh);
}

257
static struct stripe_head *__find_stripe(raid5_conf_t *conf, sector_t sector, int disks)
L
Linus Torvalds 已提交
258 259
{
	struct stripe_head *sh;
260
	struct hlist_node *hn;
L
Linus Torvalds 已提交
261 262 263

	CHECK_DEVLOCK();
	PRINTK("__find_stripe, sector %llu\n", (unsigned long long)sector);
264
	hlist_for_each_entry(sh, hn, stripe_hash(conf, sector), hash)
265
		if (sh->sector == sector && sh->disks == disks)
L
Linus Torvalds 已提交
266 267 268 269 270 271 272 273
			return sh;
	PRINTK("__stripe %llu not in cache\n", (unsigned long long)sector);
	return NULL;
}

static void unplug_slaves(mddev_t *mddev);
static void raid5_unplug_device(request_queue_t *q);

274 275
static struct stripe_head *get_active_stripe(raid5_conf_t *conf, sector_t sector, int disks,
					     int pd_idx, int noblock)
L
Linus Torvalds 已提交
276 277 278 279 280 281 282 283
{
	struct stripe_head *sh;

	PRINTK("get_stripe, sector %llu\n", (unsigned long long)sector);

	spin_lock_irq(&conf->device_lock);

	do {
284 285 286
		wait_event_lock_irq(conf->wait_for_stripe,
				    conf->quiesce == 0,
				    conf->device_lock, /* nothing */);
287
		sh = __find_stripe(conf, sector, disks);
L
Linus Torvalds 已提交
288 289 290 291 292 293 294 295 296
		if (!sh) {
			if (!conf->inactive_blocked)
				sh = get_free_stripe(conf);
			if (noblock && sh == NULL)
				break;
			if (!sh) {
				conf->inactive_blocked = 1;
				wait_event_lock_irq(conf->wait_for_stripe,
						    !list_empty(&conf->inactive_list) &&
297 298
						    (atomic_read(&conf->active_stripes)
						     < (conf->max_nr_stripes *3/4)
L
Linus Torvalds 已提交
299 300
						     || !conf->inactive_blocked),
						    conf->device_lock,
301
						    raid5_unplug_device(conf->mddev->queue)
L
Linus Torvalds 已提交
302 303 304
					);
				conf->inactive_blocked = 0;
			} else
305
				init_stripe(sh, sector, pd_idx, disks);
L
Linus Torvalds 已提交
306 307
		} else {
			if (atomic_read(&sh->count)) {
308
			  BUG_ON(!list_empty(&sh->lru));
L
Linus Torvalds 已提交
309 310 311
			} else {
				if (!test_bit(STRIPE_HANDLE, &sh->state))
					atomic_inc(&conf->active_stripes);
312 313
				if (list_empty(&sh->lru) &&
				    !test_bit(STRIPE_EXPANDING, &sh->state))
314 315
					BUG();
				list_del_init(&sh->lru);
L
Linus Torvalds 已提交
316 317 318 319 320 321 322 323 324 325 326
			}
		}
	} while (sh == NULL);

	if (sh)
		atomic_inc(&sh->count);

	spin_unlock_irq(&conf->device_lock);
	return sh;
}

327
static int grow_one_stripe(raid5_conf_t *conf)
L
Linus Torvalds 已提交
328 329
{
	struct stripe_head *sh;
330 331 332 333 334 335 336 337 338 339 340 341
	sh = kmem_cache_alloc(conf->slab_cache, GFP_KERNEL);
	if (!sh)
		return 0;
	memset(sh, 0, sizeof(*sh) + (conf->raid_disks-1)*sizeof(struct r5dev));
	sh->raid_conf = conf;
	spin_lock_init(&sh->lock);

	if (grow_buffers(sh, conf->raid_disks)) {
		shrink_buffers(sh, conf->raid_disks);
		kmem_cache_free(conf->slab_cache, sh);
		return 0;
	}
342
	sh->disks = conf->raid_disks;
343 344 345 346 347 348 349 350 351 352
	/* we just created an active stripe so... */
	atomic_set(&sh->count, 1);
	atomic_inc(&conf->active_stripes);
	INIT_LIST_HEAD(&sh->lru);
	release_stripe(sh);
	return 1;
}

static int grow_stripes(raid5_conf_t *conf, int num)
{
353
	struct kmem_cache *sc;
L
Linus Torvalds 已提交
354 355
	int devs = conf->raid_disks;

356 357 358 359
	sprintf(conf->cache_name[0], "raid5/%s", mdname(conf->mddev));
	sprintf(conf->cache_name[1], "raid5/%s-alt", mdname(conf->mddev));
	conf->active_name = 0;
	sc = kmem_cache_create(conf->cache_name[conf->active_name],
L
Linus Torvalds 已提交
360 361 362 363 364
			       sizeof(struct stripe_head)+(devs-1)*sizeof(struct r5dev),
			       0, 0, NULL, NULL);
	if (!sc)
		return 1;
	conf->slab_cache = sc;
365
	conf->pool_size = devs;
366
	while (num--)
367
		if (!grow_one_stripe(conf))
L
Linus Torvalds 已提交
368 369 370
			return 1;
	return 0;
}
371 372

#ifdef CONFIG_MD_RAID5_RESHAPE
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
static int resize_stripes(raid5_conf_t *conf, int newsize)
{
	/* Make all the stripes able to hold 'newsize' devices.
	 * New slots in each stripe get 'page' set to a new page.
	 *
	 * This happens in stages:
	 * 1/ create a new kmem_cache and allocate the required number of
	 *    stripe_heads.
	 * 2/ gather all the old stripe_heads and tranfer the pages across
	 *    to the new stripe_heads.  This will have the side effect of
	 *    freezing the array as once all stripe_heads have been collected,
	 *    no IO will be possible.  Old stripe heads are freed once their
	 *    pages have been transferred over, and the old kmem_cache is
	 *    freed when all stripes are done.
	 * 3/ reallocate conf->disks to be suitable bigger.  If this fails,
	 *    we simple return a failre status - no need to clean anything up.
	 * 4/ allocate new pages for the new slots in the new stripe_heads.
	 *    If this fails, we don't bother trying the shrink the
	 *    stripe_heads down again, we just leave them as they are.
	 *    As each stripe_head is processed the new one is released into
	 *    active service.
	 *
	 * Once step2 is started, we cannot afford to wait for a write,
	 * so we use GFP_NOIO allocations.
	 */
	struct stripe_head *osh, *nsh;
	LIST_HEAD(newstripes);
	struct disk_info *ndisks;
	int err = 0;
402
	struct kmem_cache *sc;
403 404 405 406 407
	int i;

	if (newsize <= conf->pool_size)
		return 0; /* never bother to shrink */

408 409
	md_allow_write(conf->mddev);

410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447
	/* Step 1 */
	sc = kmem_cache_create(conf->cache_name[1-conf->active_name],
			       sizeof(struct stripe_head)+(newsize-1)*sizeof(struct r5dev),
			       0, 0, NULL, NULL);
	if (!sc)
		return -ENOMEM;

	for (i = conf->max_nr_stripes; i; i--) {
		nsh = kmem_cache_alloc(sc, GFP_KERNEL);
		if (!nsh)
			break;

		memset(nsh, 0, sizeof(*nsh) + (newsize-1)*sizeof(struct r5dev));

		nsh->raid_conf = conf;
		spin_lock_init(&nsh->lock);

		list_add(&nsh->lru, &newstripes);
	}
	if (i) {
		/* didn't get enough, give up */
		while (!list_empty(&newstripes)) {
			nsh = list_entry(newstripes.next, struct stripe_head, lru);
			list_del(&nsh->lru);
			kmem_cache_free(sc, nsh);
		}
		kmem_cache_destroy(sc);
		return -ENOMEM;
	}
	/* Step 2 - Must use GFP_NOIO now.
	 * OK, we have enough stripes, start collecting inactive
	 * stripes and copying them over
	 */
	list_for_each_entry(nsh, &newstripes, lru) {
		spin_lock_irq(&conf->device_lock);
		wait_event_lock_irq(conf->wait_for_stripe,
				    !list_empty(&conf->inactive_list),
				    conf->device_lock,
448
				    unplug_slaves(conf->mddev)
449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494
			);
		osh = get_free_stripe(conf);
		spin_unlock_irq(&conf->device_lock);
		atomic_set(&nsh->count, 1);
		for(i=0; i<conf->pool_size; i++)
			nsh->dev[i].page = osh->dev[i].page;
		for( ; i<newsize; i++)
			nsh->dev[i].page = NULL;
		kmem_cache_free(conf->slab_cache, osh);
	}
	kmem_cache_destroy(conf->slab_cache);

	/* Step 3.
	 * At this point, we are holding all the stripes so the array
	 * is completely stalled, so now is a good time to resize
	 * conf->disks.
	 */
	ndisks = kzalloc(newsize * sizeof(struct disk_info), GFP_NOIO);
	if (ndisks) {
		for (i=0; i<conf->raid_disks; i++)
			ndisks[i] = conf->disks[i];
		kfree(conf->disks);
		conf->disks = ndisks;
	} else
		err = -ENOMEM;

	/* Step 4, return new stripes to service */
	while(!list_empty(&newstripes)) {
		nsh = list_entry(newstripes.next, struct stripe_head, lru);
		list_del_init(&nsh->lru);
		for (i=conf->raid_disks; i < newsize; i++)
			if (nsh->dev[i].page == NULL) {
				struct page *p = alloc_page(GFP_NOIO);
				nsh->dev[i].page = p;
				if (!p)
					err = -ENOMEM;
			}
		release_stripe(nsh);
	}
	/* critical section pass, GFP_NOIO no longer needed */

	conf->slab_cache = sc;
	conf->active_name = 1-conf->active_name;
	conf->pool_size = newsize;
	return err;
}
495
#endif
L
Linus Torvalds 已提交
496

497
static int drop_one_stripe(raid5_conf_t *conf)
L
Linus Torvalds 已提交
498 499 500
{
	struct stripe_head *sh;

501 502 503 504 505
	spin_lock_irq(&conf->device_lock);
	sh = get_free_stripe(conf);
	spin_unlock_irq(&conf->device_lock);
	if (!sh)
		return 0;
506
	BUG_ON(atomic_read(&sh->count));
507
	shrink_buffers(sh, conf->pool_size);
508 509 510 511 512 513 514 515 516 517
	kmem_cache_free(conf->slab_cache, sh);
	atomic_dec(&conf->active_stripes);
	return 1;
}

static void shrink_stripes(raid5_conf_t *conf)
{
	while (drop_one_stripe(conf))
		;

N
NeilBrown 已提交
518 519
	if (conf->slab_cache)
		kmem_cache_destroy(conf->slab_cache);
L
Linus Torvalds 已提交
520 521 522
	conf->slab_cache = NULL;
}

523
static int raid5_end_read_request(struct bio * bi, unsigned int bytes_done,
L
Linus Torvalds 已提交
524 525 526 527
				   int error)
{
 	struct stripe_head *sh = bi->bi_private;
	raid5_conf_t *conf = sh->raid_conf;
528
	int disks = sh->disks, i;
L
Linus Torvalds 已提交
529
	int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags);
530 531
	char b[BDEVNAME_SIZE];
	mdk_rdev_t *rdev;
L
Linus Torvalds 已提交
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549

	if (bi->bi_size)
		return 1;

	for (i=0 ; i<disks; i++)
		if (bi == &sh->dev[i].req)
			break;

	PRINTK("end_read_request %llu/%d, count: %d, uptodate %d.\n", 
		(unsigned long long)sh->sector, i, atomic_read(&sh->count), 
		uptodate);
	if (i == disks) {
		BUG();
		return 0;
	}

	if (uptodate) {
		set_bit(R5_UPTODATE, &sh->dev[i].flags);
550
		if (test_bit(R5_ReadError, &sh->dev[i].flags)) {
551 552 553 554 555
			rdev = conf->disks[i].rdev;
			printk(KERN_INFO "raid5:%s: read error corrected (%lu sectors at %llu on %s)\n",
			       mdname(conf->mddev), STRIPE_SECTORS,
			       (unsigned long long)sh->sector + rdev->data_offset,
			       bdevname(rdev->bdev, b));
556 557 558
			clear_bit(R5_ReadError, &sh->dev[i].flags);
			clear_bit(R5_ReWrite, &sh->dev[i].flags);
		}
559 560
		if (atomic_read(&conf->disks[i].rdev->read_errors))
			atomic_set(&conf->disks[i].rdev->read_errors, 0);
L
Linus Torvalds 已提交
561
	} else {
562
		const char *bdn = bdevname(conf->disks[i].rdev->bdev, b);
563
		int retry = 0;
564 565
		rdev = conf->disks[i].rdev;

L
Linus Torvalds 已提交
566
		clear_bit(R5_UPTODATE, &sh->dev[i].flags);
567
		atomic_inc(&rdev->read_errors);
568
		if (conf->mddev->degraded)
569 570 571 572
			printk(KERN_WARNING "raid5:%s: read error not correctable (sector %llu on %s).\n",
			       mdname(conf->mddev),
			       (unsigned long long)sh->sector + rdev->data_offset,
			       bdn);
573
		else if (test_bit(R5_ReWrite, &sh->dev[i].flags))
574
			/* Oh, no!!! */
575 576 577 578 579
			printk(KERN_WARNING "raid5:%s: read error NOT corrected!! (sector %llu on %s).\n",
			       mdname(conf->mddev),
			       (unsigned long long)sh->sector + rdev->data_offset,
			       bdn);
		else if (atomic_read(&rdev->read_errors)
580
			 > conf->max_nr_stripes)
N
NeilBrown 已提交
581
			printk(KERN_WARNING
582 583
			       "raid5:%s: Too many read errors, failing device %s.\n",
			       mdname(conf->mddev), bdn);
584 585 586 587 588
		else
			retry = 1;
		if (retry)
			set_bit(R5_ReadError, &sh->dev[i].flags);
		else {
589 590
			clear_bit(R5_ReadError, &sh->dev[i].flags);
			clear_bit(R5_ReWrite, &sh->dev[i].flags);
591
			md_error(conf->mddev, rdev);
592
		}
L
Linus Torvalds 已提交
593 594 595 596 597 598 599 600 601 602 603 604 605
	}
	rdev_dec_pending(conf->disks[i].rdev, conf->mddev);
	clear_bit(R5_LOCKED, &sh->dev[i].flags);
	set_bit(STRIPE_HANDLE, &sh->state);
	release_stripe(sh);
	return 0;
}

static int raid5_end_write_request (struct bio *bi, unsigned int bytes_done,
				    int error)
{
 	struct stripe_head *sh = bi->bi_private;
	raid5_conf_t *conf = sh->raid_conf;
606
	int disks = sh->disks, i;
L
Linus Torvalds 已提交
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630
	int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags);

	if (bi->bi_size)
		return 1;

	for (i=0 ; i<disks; i++)
		if (bi == &sh->dev[i].req)
			break;

	PRINTK("end_write_request %llu/%d, count %d, uptodate: %d.\n", 
		(unsigned long long)sh->sector, i, atomic_read(&sh->count),
		uptodate);
	if (i == disks) {
		BUG();
		return 0;
	}

	if (!uptodate)
		md_error(conf->mddev, conf->disks[i].rdev);

	rdev_dec_pending(conf->disks[i].rdev, conf->mddev);
	
	clear_bit(R5_LOCKED, &sh->dev[i].flags);
	set_bit(STRIPE_HANDLE, &sh->state);
631
	release_stripe(sh);
L
Linus Torvalds 已提交
632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653
	return 0;
}


static sector_t compute_blocknr(struct stripe_head *sh, int i);
	
static void raid5_build_block (struct stripe_head *sh, int i)
{
	struct r5dev *dev = &sh->dev[i];

	bio_init(&dev->req);
	dev->req.bi_io_vec = &dev->vec;
	dev->req.bi_vcnt++;
	dev->req.bi_max_vecs++;
	dev->vec.bv_page = dev->page;
	dev->vec.bv_len = STRIPE_SIZE;
	dev->vec.bv_offset = 0;

	dev->req.bi_sector = sh->sector;
	dev->req.bi_private = sh;

	dev->flags = 0;
654
	dev->sector = compute_blocknr(sh, i);
L
Linus Torvalds 已提交
655 656 657 658 659 660 661 662
}

static void error(mddev_t *mddev, mdk_rdev_t *rdev)
{
	char b[BDEVNAME_SIZE];
	raid5_conf_t *conf = (raid5_conf_t *) mddev->private;
	PRINTK("raid5: error called\n");

663
	if (!test_bit(Faulty, &rdev->flags)) {
664
		set_bit(MD_CHANGE_DEVS, &mddev->flags);
665 666 667
		if (test_and_clear_bit(In_sync, &rdev->flags)) {
			unsigned long flags;
			spin_lock_irqsave(&conf->device_lock, flags);
L
Linus Torvalds 已提交
668
			mddev->degraded++;
669
			spin_unlock_irqrestore(&conf->device_lock, flags);
L
Linus Torvalds 已提交
670 671 672 673 674
			/*
			 * if recovery was running, make sure it aborts.
			 */
			set_bit(MD_RECOVERY_ERR, &mddev->recovery);
		}
675
		set_bit(Faulty, &rdev->flags);
L
Linus Torvalds 已提交
676 677 678
		printk (KERN_ALERT
			"raid5: Disk failure on %s, disabling device."
			" Operation continuing on %d devices\n",
679
			bdevname(rdev->bdev,b), conf->raid_disks - mddev->degraded);
L
Linus Torvalds 已提交
680
	}
681
}
L
Linus Torvalds 已提交
682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718

/*
 * Input: a 'big' sector number,
 * Output: index of the data and parity disk, and the sector # in them.
 */
static sector_t raid5_compute_sector(sector_t r_sector, unsigned int raid_disks,
			unsigned int data_disks, unsigned int * dd_idx,
			unsigned int * pd_idx, raid5_conf_t *conf)
{
	long stripe;
	unsigned long chunk_number;
	unsigned int chunk_offset;
	sector_t new_sector;
	int sectors_per_chunk = conf->chunk_size >> 9;

	/* First compute the information on this sector */

	/*
	 * Compute the chunk number and the sector offset inside the chunk
	 */
	chunk_offset = sector_div(r_sector, sectors_per_chunk);
	chunk_number = r_sector;
	BUG_ON(r_sector != chunk_number);

	/*
	 * Compute the stripe number
	 */
	stripe = chunk_number / data_disks;

	/*
	 * Compute the data disk and parity disk indexes inside the stripe
	 */
	*dd_idx = chunk_number % data_disks;

	/*
	 * Select the parity disk based on the user selected algorithm.
	 */
719 720
	switch(conf->level) {
	case 4:
L
Linus Torvalds 已提交
721
		*pd_idx = data_disks;
722 723 724
		break;
	case 5:
		switch (conf->algorithm) {
L
Linus Torvalds 已提交
725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
		case ALGORITHM_LEFT_ASYMMETRIC:
			*pd_idx = data_disks - stripe % raid_disks;
			if (*dd_idx >= *pd_idx)
				(*dd_idx)++;
			break;
		case ALGORITHM_RIGHT_ASYMMETRIC:
			*pd_idx = stripe % raid_disks;
			if (*dd_idx >= *pd_idx)
				(*dd_idx)++;
			break;
		case ALGORITHM_LEFT_SYMMETRIC:
			*pd_idx = data_disks - stripe % raid_disks;
			*dd_idx = (*pd_idx + 1 + *dd_idx) % raid_disks;
			break;
		case ALGORITHM_RIGHT_SYMMETRIC:
			*pd_idx = stripe % raid_disks;
			*dd_idx = (*pd_idx + 1 + *dd_idx) % raid_disks;
			break;
		default:
N
NeilBrown 已提交
744
			printk(KERN_ERR "raid5: unsupported algorithm %d\n",
L
Linus Torvalds 已提交
745
				conf->algorithm);
746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
		}
		break;
	case 6:

		/**** FIX THIS ****/
		switch (conf->algorithm) {
		case ALGORITHM_LEFT_ASYMMETRIC:
			*pd_idx = raid_disks - 1 - (stripe % raid_disks);
			if (*pd_idx == raid_disks-1)
				(*dd_idx)++; 	/* Q D D D P */
			else if (*dd_idx >= *pd_idx)
				(*dd_idx) += 2; /* D D P Q D */
			break;
		case ALGORITHM_RIGHT_ASYMMETRIC:
			*pd_idx = stripe % raid_disks;
			if (*pd_idx == raid_disks-1)
				(*dd_idx)++; 	/* Q D D D P */
			else if (*dd_idx >= *pd_idx)
				(*dd_idx) += 2; /* D D P Q D */
			break;
		case ALGORITHM_LEFT_SYMMETRIC:
			*pd_idx = raid_disks - 1 - (stripe % raid_disks);
			*dd_idx = (*pd_idx + 2 + *dd_idx) % raid_disks;
			break;
		case ALGORITHM_RIGHT_SYMMETRIC:
			*pd_idx = stripe % raid_disks;
			*dd_idx = (*pd_idx + 2 + *dd_idx) % raid_disks;
			break;
		default:
			printk (KERN_CRIT "raid6: unsupported algorithm %d\n",
				conf->algorithm);
		}
		break;
L
Linus Torvalds 已提交
779 780 781 782 783 784 785 786 787 788 789 790 791
	}

	/*
	 * Finally, compute the new sector number
	 */
	new_sector = (sector_t)stripe * sectors_per_chunk + chunk_offset;
	return new_sector;
}


static sector_t compute_blocknr(struct stripe_head *sh, int i)
{
	raid5_conf_t *conf = sh->raid_conf;
792 793
	int raid_disks = sh->disks;
	int data_disks = raid_disks - conf->max_degraded;
L
Linus Torvalds 已提交
794 795 796 797 798 799 800
	sector_t new_sector = sh->sector, check;
	int sectors_per_chunk = conf->chunk_size >> 9;
	sector_t stripe;
	int chunk_offset;
	int chunk_number, dummy1, dummy2, dd_idx = i;
	sector_t r_sector;

801

L
Linus Torvalds 已提交
802 803 804 805
	chunk_offset = sector_div(new_sector, sectors_per_chunk);
	stripe = new_sector;
	BUG_ON(new_sector != stripe);

806 807 808 809 810 811
	if (i == sh->pd_idx)
		return 0;
	switch(conf->level) {
	case 4: break;
	case 5:
		switch (conf->algorithm) {
L
Linus Torvalds 已提交
812 813 814 815 816 817 818 819 820 821 822 823
		case ALGORITHM_LEFT_ASYMMETRIC:
		case ALGORITHM_RIGHT_ASYMMETRIC:
			if (i > sh->pd_idx)
				i--;
			break;
		case ALGORITHM_LEFT_SYMMETRIC:
		case ALGORITHM_RIGHT_SYMMETRIC:
			if (i < sh->pd_idx)
				i += raid_disks;
			i -= (sh->pd_idx + 1);
			break;
		default:
N
NeilBrown 已提交
824
			printk(KERN_ERR "raid5: unsupported algorithm %d\n",
825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851
			       conf->algorithm);
		}
		break;
	case 6:
		if (i == raid6_next_disk(sh->pd_idx, raid_disks))
			return 0; /* It is the Q disk */
		switch (conf->algorithm) {
		case ALGORITHM_LEFT_ASYMMETRIC:
		case ALGORITHM_RIGHT_ASYMMETRIC:
		  	if (sh->pd_idx == raid_disks-1)
				i--; 	/* Q D D D P */
			else if (i > sh->pd_idx)
				i -= 2; /* D D P Q D */
			break;
		case ALGORITHM_LEFT_SYMMETRIC:
		case ALGORITHM_RIGHT_SYMMETRIC:
			if (sh->pd_idx == raid_disks-1)
				i--; /* Q D D D P */
			else {
				/* D D P Q D */
				if (i < sh->pd_idx)
					i += raid_disks;
				i -= (sh->pd_idx + 2);
			}
			break;
		default:
			printk (KERN_CRIT "raid6: unsupported algorithm %d\n",
L
Linus Torvalds 已提交
852
				conf->algorithm);
853 854
		}
		break;
L
Linus Torvalds 已提交
855 856 857 858 859 860 861
	}

	chunk_number = stripe * data_disks + i;
	r_sector = (sector_t)chunk_number * sectors_per_chunk + chunk_offset;

	check = raid5_compute_sector (r_sector, raid_disks, data_disks, &dummy1, &dummy2, conf);
	if (check != sh->sector || dummy1 != dd_idx || dummy2 != sh->pd_idx) {
N
NeilBrown 已提交
862
		printk(KERN_ERR "compute_blocknr: map not correct\n");
L
Linus Torvalds 已提交
863 864 865 866 867 868 869 870
		return 0;
	}
	return r_sector;
}



/*
871 872 873 874 875
 * Copy data between a page in the stripe cache, and one or more bion
 * The page could align with the middle of the bio, or there could be
 * several bion, each with several bio_vecs, which cover part of the page
 * Multiple bion are linked together on bi_next.  There may be extras
 * at the end of this list.  We ignore them.
L
Linus Torvalds 已提交
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903
 */
static void copy_data(int frombio, struct bio *bio,
		     struct page *page,
		     sector_t sector)
{
	char *pa = page_address(page);
	struct bio_vec *bvl;
	int i;
	int page_offset;

	if (bio->bi_sector >= sector)
		page_offset = (signed)(bio->bi_sector - sector) * 512;
	else
		page_offset = (signed)(sector - bio->bi_sector) * -512;
	bio_for_each_segment(bvl, bio, i) {
		int len = bio_iovec_idx(bio,i)->bv_len;
		int clen;
		int b_offset = 0;

		if (page_offset < 0) {
			b_offset = -page_offset;
			page_offset += b_offset;
			len -= b_offset;
		}

		if (len > 0 && page_offset + len > STRIPE_SIZE)
			clen = STRIPE_SIZE - page_offset;
		else clen = len;
904

L
Linus Torvalds 已提交
905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928
		if (clen > 0) {
			char *ba = __bio_kmap_atomic(bio, i, KM_USER0);
			if (frombio)
				memcpy(pa+page_offset, ba+b_offset, clen);
			else
				memcpy(ba+b_offset, pa+page_offset, clen);
			__bio_kunmap_atomic(ba, KM_USER0);
		}
		if (clen < len) /* hit end of page */
			break;
		page_offset +=  len;
	}
}

#define check_xor() 	do { 						\
			   if (count == MAX_XOR_BLOCKS) {		\
				xor_block(count, STRIPE_SIZE, ptr);	\
				count = 1;				\
			   }						\
			} while(0)


static void compute_block(struct stripe_head *sh, int dd_idx)
{
929
	int i, count, disks = sh->disks;
L
Linus Torvalds 已提交
930 931 932 933 934 935 936 937 938 939 940 941 942 943 944
	void *ptr[MAX_XOR_BLOCKS], *p;

	PRINTK("compute_block, stripe %llu, idx %d\n", 
		(unsigned long long)sh->sector, dd_idx);

	ptr[0] = page_address(sh->dev[dd_idx].page);
	memset(ptr[0], 0, STRIPE_SIZE);
	count = 1;
	for (i = disks ; i--; ) {
		if (i == dd_idx)
			continue;
		p = page_address(sh->dev[i].page);
		if (test_bit(R5_UPTODATE, &sh->dev[i].flags))
			ptr[count++] = p;
		else
N
NeilBrown 已提交
945
			printk(KERN_ERR "compute_block() %d, stripe %llu, %d"
L
Linus Torvalds 已提交
946 947 948 949 950 951 952 953 954 955
				" not present\n", dd_idx,
				(unsigned long long)sh->sector, i);

		check_xor();
	}
	if (count != 1)
		xor_block(count, STRIPE_SIZE, ptr);
	set_bit(R5_UPTODATE, &sh->dev[dd_idx].flags);
}

956
static void compute_parity5(struct stripe_head *sh, int method)
L
Linus Torvalds 已提交
957 958
{
	raid5_conf_t *conf = sh->raid_conf;
959
	int i, pd_idx = sh->pd_idx, disks = sh->disks, count;
L
Linus Torvalds 已提交
960 961 962
	void *ptr[MAX_XOR_BLOCKS];
	struct bio *chosen;

963
	PRINTK("compute_parity5, stripe %llu, method %d\n",
L
Linus Torvalds 已提交
964 965 966 967 968 969
		(unsigned long long)sh->sector, method);

	count = 1;
	ptr[0] = page_address(sh->dev[pd_idx].page);
	switch(method) {
	case READ_MODIFY_WRITE:
970
		BUG_ON(!test_bit(R5_UPTODATE, &sh->dev[pd_idx].flags));
L
Linus Torvalds 已提交
971 972 973 974 975 976 977 978 979 980 981 982
		for (i=disks ; i-- ;) {
			if (i==pd_idx)
				continue;
			if (sh->dev[i].towrite &&
			    test_bit(R5_UPTODATE, &sh->dev[i].flags)) {
				ptr[count++] = page_address(sh->dev[i].page);
				chosen = sh->dev[i].towrite;
				sh->dev[i].towrite = NULL;

				if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
					wake_up(&conf->wait_for_overlap);

983
				BUG_ON(sh->dev[i].written);
L
Linus Torvalds 已提交
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998
				sh->dev[i].written = chosen;
				check_xor();
			}
		}
		break;
	case RECONSTRUCT_WRITE:
		memset(ptr[0], 0, STRIPE_SIZE);
		for (i= disks; i-- ;)
			if (i!=pd_idx && sh->dev[i].towrite) {
				chosen = sh->dev[i].towrite;
				sh->dev[i].towrite = NULL;

				if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
					wake_up(&conf->wait_for_overlap);

999
				BUG_ON(sh->dev[i].written);
L
Linus Torvalds 已提交
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
				sh->dev[i].written = chosen;
			}
		break;
	case CHECK_PARITY:
		break;
	}
	if (count>1) {
		xor_block(count, STRIPE_SIZE, ptr);
		count = 1;
	}
	
	for (i = disks; i--;)
		if (sh->dev[i].written) {
			sector_t sector = sh->dev[i].sector;
			struct bio *wbi = sh->dev[i].written;
			while (wbi && wbi->bi_sector < sector + STRIPE_SECTORS) {
				copy_data(1, wbi, sh->dev[i].page, sector);
				wbi = r5_next_bio(wbi, sector);
			}

			set_bit(R5_LOCKED, &sh->dev[i].flags);
			set_bit(R5_UPTODATE, &sh->dev[i].flags);
		}

	switch(method) {
	case RECONSTRUCT_WRITE:
	case CHECK_PARITY:
		for (i=disks; i--;)
			if (i != pd_idx) {
				ptr[count++] = page_address(sh->dev[i].page);
				check_xor();
			}
		break;
	case READ_MODIFY_WRITE:
		for (i = disks; i--;)
			if (sh->dev[i].written) {
				ptr[count++] = page_address(sh->dev[i].page);
				check_xor();
			}
	}
	if (count != 1)
		xor_block(count, STRIPE_SIZE, ptr);
	
	if (method != CHECK_PARITY) {
		set_bit(R5_UPTODATE, &sh->dev[pd_idx].flags);
		set_bit(R5_LOCKED,   &sh->dev[pd_idx].flags);
	} else
		clear_bit(R5_UPTODATE, &sh->dev[pd_idx].flags);
}

1050 1051 1052
static void compute_parity6(struct stripe_head *sh, int method)
{
	raid6_conf_t *conf = sh->raid_conf;
1053
	int i, pd_idx = sh->pd_idx, qd_idx, d0_idx, disks = sh->disks, count;
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
	struct bio *chosen;
	/**** FIX THIS: This could be very bad if disks is close to 256 ****/
	void *ptrs[disks];

	qd_idx = raid6_next_disk(pd_idx, disks);
	d0_idx = raid6_next_disk(qd_idx, disks);

	PRINTK("compute_parity, stripe %llu, method %d\n",
		(unsigned long long)sh->sector, method);

	switch(method) {
	case READ_MODIFY_WRITE:
		BUG();		/* READ_MODIFY_WRITE N/A for RAID-6 */
	case RECONSTRUCT_WRITE:
		for (i= disks; i-- ;)
			if ( i != pd_idx && i != qd_idx && sh->dev[i].towrite ) {
				chosen = sh->dev[i].towrite;
				sh->dev[i].towrite = NULL;

				if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
					wake_up(&conf->wait_for_overlap);

E
Eric Sesterhenn 已提交
1076
				BUG_ON(sh->dev[i].written);
1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
				sh->dev[i].written = chosen;
			}
		break;
	case CHECK_PARITY:
		BUG();		/* Not implemented yet */
	}

	for (i = disks; i--;)
		if (sh->dev[i].written) {
			sector_t sector = sh->dev[i].sector;
			struct bio *wbi = sh->dev[i].written;
			while (wbi && wbi->bi_sector < sector + STRIPE_SECTORS) {
				copy_data(1, wbi, sh->dev[i].page, sector);
				wbi = r5_next_bio(wbi, sector);
			}

			set_bit(R5_LOCKED, &sh->dev[i].flags);
			set_bit(R5_UPTODATE, &sh->dev[i].flags);
		}

//	switch(method) {
//	case RECONSTRUCT_WRITE:
//	case CHECK_PARITY:
//	case UPDATE_PARITY:
		/* Note that unlike RAID-5, the ordering of the disks matters greatly. */
		/* FIX: Is this ordering of drives even remotely optimal? */
		count = 0;
		i = d0_idx;
		do {
			ptrs[count++] = page_address(sh->dev[i].page);
			if (count <= disks-2 && !test_bit(R5_UPTODATE, &sh->dev[i].flags))
				printk("block %d/%d not uptodate on parity calc\n", i,count);
			i = raid6_next_disk(i, disks);
		} while ( i != d0_idx );
//		break;
//	}

	raid6_call.gen_syndrome(disks, STRIPE_SIZE, ptrs);

	switch(method) {
	case RECONSTRUCT_WRITE:
		set_bit(R5_UPTODATE, &sh->dev[pd_idx].flags);
		set_bit(R5_UPTODATE, &sh->dev[qd_idx].flags);
		set_bit(R5_LOCKED,   &sh->dev[pd_idx].flags);
		set_bit(R5_LOCKED,   &sh->dev[qd_idx].flags);
		break;
	case UPDATE_PARITY:
		set_bit(R5_UPTODATE, &sh->dev[pd_idx].flags);
		set_bit(R5_UPTODATE, &sh->dev[qd_idx].flags);
		break;
	}
}


/* Compute one missing block */
static void compute_block_1(struct stripe_head *sh, int dd_idx, int nozero)
{
1134
	int i, count, disks = sh->disks;
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171
	void *ptr[MAX_XOR_BLOCKS], *p;
	int pd_idx = sh->pd_idx;
	int qd_idx = raid6_next_disk(pd_idx, disks);

	PRINTK("compute_block_1, stripe %llu, idx %d\n",
		(unsigned long long)sh->sector, dd_idx);

	if ( dd_idx == qd_idx ) {
		/* We're actually computing the Q drive */
		compute_parity6(sh, UPDATE_PARITY);
	} else {
		ptr[0] = page_address(sh->dev[dd_idx].page);
		if (!nozero) memset(ptr[0], 0, STRIPE_SIZE);
		count = 1;
		for (i = disks ; i--; ) {
			if (i == dd_idx || i == qd_idx)
				continue;
			p = page_address(sh->dev[i].page);
			if (test_bit(R5_UPTODATE, &sh->dev[i].flags))
				ptr[count++] = p;
			else
				printk("compute_block() %d, stripe %llu, %d"
				       " not present\n", dd_idx,
				       (unsigned long long)sh->sector, i);

			check_xor();
		}
		if (count != 1)
			xor_block(count, STRIPE_SIZE, ptr);
		if (!nozero) set_bit(R5_UPTODATE, &sh->dev[dd_idx].flags);
		else clear_bit(R5_UPTODATE, &sh->dev[dd_idx].flags);
	}
}

/* Compute two missing blocks */
static void compute_block_2(struct stripe_head *sh, int dd_idx1, int dd_idx2)
{
1172
	int i, count, disks = sh->disks;
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
	int pd_idx = sh->pd_idx;
	int qd_idx = raid6_next_disk(pd_idx, disks);
	int d0_idx = raid6_next_disk(qd_idx, disks);
	int faila, failb;

	/* faila and failb are disk numbers relative to d0_idx */
	/* pd_idx become disks-2 and qd_idx become disks-1 */
	faila = (dd_idx1 < d0_idx) ? dd_idx1+(disks-d0_idx) : dd_idx1-d0_idx;
	failb = (dd_idx2 < d0_idx) ? dd_idx2+(disks-d0_idx) : dd_idx2-d0_idx;

	BUG_ON(faila == failb);
	if ( failb < faila ) { int tmp = faila; faila = failb; failb = tmp; }

	PRINTK("compute_block_2, stripe %llu, idx %d,%d (%d,%d)\n",
	       (unsigned long long)sh->sector, dd_idx1, dd_idx2, faila, failb);

	if ( failb == disks-1 ) {
		/* Q disk is one of the missing disks */
		if ( faila == disks-2 ) {
			/* Missing P+Q, just recompute */
			compute_parity6(sh, UPDATE_PARITY);
			return;
		} else {
			/* We're missing D+Q; recompute D from P */
			compute_block_1(sh, (dd_idx1 == qd_idx) ? dd_idx2 : dd_idx1, 0);
			compute_parity6(sh, UPDATE_PARITY); /* Is this necessary? */
			return;
		}
	}

	/* We're missing D+P or D+D; build pointer table */
	{
		/**** FIX THIS: This could be very bad if disks is close to 256 ****/
		void *ptrs[disks];

		count = 0;
		i = d0_idx;
		do {
			ptrs[count++] = page_address(sh->dev[i].page);
			i = raid6_next_disk(i, disks);
			if (i != dd_idx1 && i != dd_idx2 &&
			    !test_bit(R5_UPTODATE, &sh->dev[i].flags))
				printk("compute_2 with missing block %d/%d\n", count, i);
		} while ( i != d0_idx );

		if ( failb == disks-2 ) {
			/* We're missing D+P. */
			raid6_datap_recov(disks, STRIPE_SIZE, faila, ptrs);
		} else {
			/* We're missing D+D. */
			raid6_2data_recov(disks, STRIPE_SIZE, faila, failb, ptrs);
		}

		/* Both the above update both missing blocks */
		set_bit(R5_UPTODATE, &sh->dev[dd_idx1].flags);
		set_bit(R5_UPTODATE, &sh->dev[dd_idx2].flags);
	}
}



L
Linus Torvalds 已提交
1234 1235
/*
 * Each stripe/dev can have one or more bion attached.
1236
 * toread/towrite point to the first in a chain.
L
Linus Torvalds 已提交
1237 1238 1239 1240 1241 1242
 * The bi_next chain must be in order.
 */
static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, int forwrite)
{
	struct bio **bip;
	raid5_conf_t *conf = sh->raid_conf;
1243
	int firstwrite=0;
L
Linus Torvalds 已提交
1244 1245 1246 1247 1248 1249 1250 1251

	PRINTK("adding bh b#%llu to stripe s#%llu\n",
		(unsigned long long)bi->bi_sector,
		(unsigned long long)sh->sector);


	spin_lock(&sh->lock);
	spin_lock_irq(&conf->device_lock);
1252
	if (forwrite) {
L
Linus Torvalds 已提交
1253
		bip = &sh->dev[dd_idx].towrite;
1254 1255 1256
		if (*bip == NULL && sh->dev[dd_idx].written == NULL)
			firstwrite = 1;
	} else
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262 1263 1264 1265
		bip = &sh->dev[dd_idx].toread;
	while (*bip && (*bip)->bi_sector < bi->bi_sector) {
		if ((*bip)->bi_sector + ((*bip)->bi_size >> 9) > bi->bi_sector)
			goto overlap;
		bip = & (*bip)->bi_next;
	}
	if (*bip && (*bip)->bi_sector < bi->bi_sector + ((bi->bi_size)>>9))
		goto overlap;

1266
	BUG_ON(*bip && bi->bi_next && (*bip) != bi->bi_next);
L
Linus Torvalds 已提交
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277
	if (*bip)
		bi->bi_next = *bip;
	*bip = bi;
	bi->bi_phys_segments ++;
	spin_unlock_irq(&conf->device_lock);
	spin_unlock(&sh->lock);

	PRINTK("added bi b#%llu to stripe s#%llu, disk %d.\n",
		(unsigned long long)bi->bi_sector,
		(unsigned long long)sh->sector, dd_idx);

1278 1279 1280
	if (conf->mddev->bitmap && firstwrite) {
		bitmap_startwrite(conf->mddev->bitmap, sh->sector,
				  STRIPE_SECTORS, 0);
1281
		sh->bm_seq = conf->seq_flush+1;
1282 1283 1284
		set_bit(STRIPE_BIT_DELAY, &sh->state);
	}

L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
	if (forwrite) {
		/* check if page is covered */
		sector_t sector = sh->dev[dd_idx].sector;
		for (bi=sh->dev[dd_idx].towrite;
		     sector < sh->dev[dd_idx].sector + STRIPE_SECTORS &&
			     bi && bi->bi_sector <= sector;
		     bi = r5_next_bio(bi, sh->dev[dd_idx].sector)) {
			if (bi->bi_sector + (bi->bi_size>>9) >= sector)
				sector = bi->bi_sector + (bi->bi_size>>9);
		}
		if (sector >= sh->dev[dd_idx].sector + STRIPE_SECTORS)
			set_bit(R5_OVERWRITE, &sh->dev[dd_idx].flags);
	}
	return 1;

 overlap:
	set_bit(R5_Overlap, &sh->dev[dd_idx].flags);
	spin_unlock_irq(&conf->device_lock);
	spin_unlock(&sh->lock);
	return 0;
}

1307 1308
static void end_reshape(raid5_conf_t *conf);

1309 1310 1311 1312 1313 1314 1315
static int page_is_zero(struct page *p)
{
	char *a = page_address(p);
	return ((*(u32*)a) == 0 &&
		memcmp(a, a+4, STRIPE_SIZE-4)==0);
}

1316 1317 1318 1319
static int stripe_to_pdidx(sector_t stripe, raid5_conf_t *conf, int disks)
{
	int sectors_per_chunk = conf->chunk_size >> 9;
	int pd_idx, dd_idx;
1320 1321
	int chunk_offset = sector_div(stripe, sectors_per_chunk);

1322 1323 1324 1325
	raid5_compute_sector(stripe * (disks - conf->max_degraded)
			     *sectors_per_chunk + chunk_offset,
			     disks, disks - conf->max_degraded,
			     &dd_idx, &pd_idx, conf);
1326 1327 1328
	return pd_idx;
}

L
Linus Torvalds 已提交
1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347

/*
 * handle_stripe - do things to a stripe.
 *
 * We lock the stripe and then examine the state of various bits
 * to see what needs to be done.
 * Possible results:
 *    return some read request which now have data
 *    return some write requests which are safely on disc
 *    schedule a read on some buffers
 *    schedule a write of some buffers
 *    return confirmation of parity correctness
 *
 * Parity calculations are done inside the stripe lock
 * buffers are taken off read_list or write_list, and bh_cache buffers
 * get BH_Lock set before the stripe lock is released.
 *
 */
 
1348
static void handle_stripe5(struct stripe_head *sh)
L
Linus Torvalds 已提交
1349 1350
{
	raid5_conf_t *conf = sh->raid_conf;
1351
	int disks = sh->disks;
L
Linus Torvalds 已提交
1352 1353 1354
	struct bio *return_bi= NULL;
	struct bio *bi;
	int i;
1355
	int syncing, expanding, expanded;
L
Linus Torvalds 已提交
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369
	int locked=0, uptodate=0, to_read=0, to_write=0, failed=0, written=0;
	int non_overwrite = 0;
	int failed_num=0;
	struct r5dev *dev;

	PRINTK("handling stripe %llu, cnt=%d, pd_idx=%d\n",
		(unsigned long long)sh->sector, atomic_read(&sh->count),
		sh->pd_idx);

	spin_lock(&sh->lock);
	clear_bit(STRIPE_HANDLE, &sh->state);
	clear_bit(STRIPE_DELAYED, &sh->state);

	syncing = test_bit(STRIPE_SYNCING, &sh->state);
1370 1371
	expanding = test_bit(STRIPE_EXPAND_SOURCE, &sh->state);
	expanded = test_bit(STRIPE_EXPAND_READY, &sh->state);
L
Linus Torvalds 已提交
1372 1373
	/* Now to look around and see what can be done */

1374
	rcu_read_lock();
L
Linus Torvalds 已提交
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416
	for (i=disks; i--; ) {
		mdk_rdev_t *rdev;
		dev = &sh->dev[i];
		clear_bit(R5_Insync, &dev->flags);

		PRINTK("check %d: state 0x%lx read %p write %p written %p\n",
			i, dev->flags, dev->toread, dev->towrite, dev->written);
		/* maybe we can reply to a read */
		if (test_bit(R5_UPTODATE, &dev->flags) && dev->toread) {
			struct bio *rbi, *rbi2;
			PRINTK("Return read for disc %d\n", i);
			spin_lock_irq(&conf->device_lock);
			rbi = dev->toread;
			dev->toread = NULL;
			if (test_and_clear_bit(R5_Overlap, &dev->flags))
				wake_up(&conf->wait_for_overlap);
			spin_unlock_irq(&conf->device_lock);
			while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) {
				copy_data(0, rbi, dev->page, dev->sector);
				rbi2 = r5_next_bio(rbi, dev->sector);
				spin_lock_irq(&conf->device_lock);
				if (--rbi->bi_phys_segments == 0) {
					rbi->bi_next = return_bi;
					return_bi = rbi;
				}
				spin_unlock_irq(&conf->device_lock);
				rbi = rbi2;
			}
		}

		/* now count some things */
		if (test_bit(R5_LOCKED, &dev->flags)) locked++;
		if (test_bit(R5_UPTODATE, &dev->flags)) uptodate++;

		
		if (dev->toread) to_read++;
		if (dev->towrite) {
			to_write++;
			if (!test_bit(R5_OVERWRITE, &dev->flags))
				non_overwrite++;
		}
		if (dev->written) written++;
1417
		rdev = rcu_dereference(conf->disks[i].rdev);
1418
		if (!rdev || !test_bit(In_sync, &rdev->flags)) {
N
NeilBrown 已提交
1419
			/* The ReadError flag will just be confusing now */
1420 1421 1422
			clear_bit(R5_ReadError, &dev->flags);
			clear_bit(R5_ReWrite, &dev->flags);
		}
1423
		if (!rdev || !test_bit(In_sync, &rdev->flags)
1424
		    || test_bit(R5_ReadError, &dev->flags)) {
L
Linus Torvalds 已提交
1425 1426 1427 1428 1429
			failed++;
			failed_num = i;
		} else
			set_bit(R5_Insync, &dev->flags);
	}
1430
	rcu_read_unlock();
L
Linus Torvalds 已提交
1431 1432 1433 1434 1435 1436 1437 1438
	PRINTK("locked=%d uptodate=%d to_read=%d"
		" to_write=%d failed=%d failed_num=%d\n",
		locked, uptodate, to_read, to_write, failed, failed_num);
	/* check if the array has lost two devices and, if so, some requests might
	 * need to be failed
	 */
	if (failed > 1 && to_read+to_write+written) {
		for (i=disks; i--; ) {
1439
			int bitmap_end = 0;
1440 1441

			if (test_bit(R5_ReadError, &sh->dev[i].flags)) {
1442 1443 1444
				mdk_rdev_t *rdev;
				rcu_read_lock();
				rdev = rcu_dereference(conf->disks[i].rdev);
1445
				if (rdev && test_bit(In_sync, &rdev->flags))
1446 1447
					/* multiple read failures in one stripe */
					md_error(conf->mddev, rdev);
1448
				rcu_read_unlock();
1449 1450
			}

1451
			spin_lock_irq(&conf->device_lock);
L
Linus Torvalds 已提交
1452 1453 1454
			/* fail all writes first */
			bi = sh->dev[i].towrite;
			sh->dev[i].towrite = NULL;
1455
			if (bi) { to_write--; bitmap_end = 1; }
L
Linus Torvalds 已提交
1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472

			if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
				wake_up(&conf->wait_for_overlap);

			while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS){
				struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector);
				clear_bit(BIO_UPTODATE, &bi->bi_flags);
				if (--bi->bi_phys_segments == 0) {
					md_write_end(conf->mddev);
					bi->bi_next = return_bi;
					return_bi = bi;
				}
				bi = nextbi;
			}
			/* and fail all 'written' */
			bi = sh->dev[i].written;
			sh->dev[i].written = NULL;
1473
			if (bi) bitmap_end = 1;
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
			while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS) {
				struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector);
				clear_bit(BIO_UPTODATE, &bi->bi_flags);
				if (--bi->bi_phys_segments == 0) {
					md_write_end(conf->mddev);
					bi->bi_next = return_bi;
					return_bi = bi;
				}
				bi = bi2;
			}

			/* fail any reads if this device is non-operational */
1486 1487
			if (!test_bit(R5_Insync, &sh->dev[i].flags) ||
			    test_bit(R5_ReadError, &sh->dev[i].flags)) {
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
				bi = sh->dev[i].toread;
				sh->dev[i].toread = NULL;
				if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
					wake_up(&conf->wait_for_overlap);
				if (bi) to_read--;
				while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS){
					struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector);
					clear_bit(BIO_UPTODATE, &bi->bi_flags);
					if (--bi->bi_phys_segments == 0) {
						bi->bi_next = return_bi;
						return_bi = bi;
					}
					bi = nextbi;
				}
			}
1503 1504 1505 1506
			spin_unlock_irq(&conf->device_lock);
			if (bitmap_end)
				bitmap_endwrite(conf->mddev->bitmap, sh->sector,
						STRIPE_SECTORS, 0, 0);
L
Linus Torvalds 已提交
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534
		}
	}
	if (failed > 1 && syncing) {
		md_done_sync(conf->mddev, STRIPE_SECTORS,0);
		clear_bit(STRIPE_SYNCING, &sh->state);
		syncing = 0;
	}

	/* might be able to return some write requests if the parity block
	 * is safe, or on a failed drive
	 */
	dev = &sh->dev[sh->pd_idx];
	if ( written &&
	     ( (test_bit(R5_Insync, &dev->flags) && !test_bit(R5_LOCKED, &dev->flags) &&
		test_bit(R5_UPTODATE, &dev->flags))
	       || (failed == 1 && failed_num == sh->pd_idx))
	    ) {
	    /* any written block on an uptodate or failed drive can be returned.
	     * Note that if we 'wrote' to a failed drive, it will be UPTODATE, but 
	     * never LOCKED, so we don't need to test 'failed' directly.
	     */
	    for (i=disks; i--; )
		if (sh->dev[i].written) {
		    dev = &sh->dev[i];
		    if (!test_bit(R5_LOCKED, &dev->flags) &&
			 test_bit(R5_UPTODATE, &dev->flags) ) {
			/* We can return any write requests */
			    struct bio *wbi, *wbi2;
1535
			    int bitmap_end = 0;
L
Linus Torvalds 已提交
1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
			    PRINTK("Return write for disc %d\n", i);
			    spin_lock_irq(&conf->device_lock);
			    wbi = dev->written;
			    dev->written = NULL;
			    while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) {
				    wbi2 = r5_next_bio(wbi, dev->sector);
				    if (--wbi->bi_phys_segments == 0) {
					    md_write_end(conf->mddev);
					    wbi->bi_next = return_bi;
					    return_bi = wbi;
				    }
				    wbi = wbi2;
			    }
1549 1550
			    if (dev->towrite == NULL)
				    bitmap_end = 1;
L
Linus Torvalds 已提交
1551
			    spin_unlock_irq(&conf->device_lock);
1552 1553 1554 1555
			    if (bitmap_end)
				    bitmap_endwrite(conf->mddev->bitmap, sh->sector,
						    STRIPE_SECTORS,
						    !test_bit(STRIPE_DEGRADED, &sh->state), 0);
L
Linus Torvalds 已提交
1556 1557 1558 1559 1560 1561 1562 1563
		    }
		}
	}

	/* Now we might consider reading some blocks, either to check/generate
	 * parity, or to satisfy requests
	 * or to load a block that is being partially written.
	 */
1564
	if (to_read || non_overwrite || (syncing && (uptodate < disks)) || expanding) {
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569 1570
		for (i=disks; i--;) {
			dev = &sh->dev[i];
			if (!test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) &&
			    (dev->toread ||
			     (dev->towrite && !test_bit(R5_OVERWRITE, &dev->flags)) ||
			     syncing ||
1571
			     expanding ||
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 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 1656 1657 1658 1659 1660 1661
			     (failed && (sh->dev[failed_num].toread ||
					 (sh->dev[failed_num].towrite && !test_bit(R5_OVERWRITE, &sh->dev[failed_num].flags))))
				    )
				) {
				/* we would like to get this block, possibly
				 * by computing it, but we might not be able to
				 */
				if (uptodate == disks-1) {
					PRINTK("Computing block %d\n", i);
					compute_block(sh, i);
					uptodate++;
				} else if (test_bit(R5_Insync, &dev->flags)) {
					set_bit(R5_LOCKED, &dev->flags);
					set_bit(R5_Wantread, &dev->flags);
					locked++;
					PRINTK("Reading block %d (sync=%d)\n", 
						i, syncing);
				}
			}
		}
		set_bit(STRIPE_HANDLE, &sh->state);
	}

	/* now to consider writing and what else, if anything should be read */
	if (to_write) {
		int rmw=0, rcw=0;
		for (i=disks ; i--;) {
			/* would I have to read this buffer for read_modify_write */
			dev = &sh->dev[i];
			if ((dev->towrite || i == sh->pd_idx) &&
			    (!test_bit(R5_LOCKED, &dev->flags) 
				    ) &&
			    !test_bit(R5_UPTODATE, &dev->flags)) {
				if (test_bit(R5_Insync, &dev->flags)
/*				    && !(!mddev->insync && i == sh->pd_idx) */
					)
					rmw++;
				else rmw += 2*disks;  /* cannot read it */
			}
			/* Would I have to read this buffer for reconstruct_write */
			if (!test_bit(R5_OVERWRITE, &dev->flags) && i != sh->pd_idx &&
			    (!test_bit(R5_LOCKED, &dev->flags) 
				    ) &&
			    !test_bit(R5_UPTODATE, &dev->flags)) {
				if (test_bit(R5_Insync, &dev->flags)) rcw++;
				else rcw += 2*disks;
			}
		}
		PRINTK("for sector %llu, rmw=%d rcw=%d\n", 
			(unsigned long long)sh->sector, rmw, rcw);
		set_bit(STRIPE_HANDLE, &sh->state);
		if (rmw < rcw && rmw > 0)
			/* prefer read-modify-write, but need to get some data */
			for (i=disks; i--;) {
				dev = &sh->dev[i];
				if ((dev->towrite || i == sh->pd_idx) &&
				    !test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) &&
				    test_bit(R5_Insync, &dev->flags)) {
					if (test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))
					{
						PRINTK("Read_old block %d for r-m-w\n", i);
						set_bit(R5_LOCKED, &dev->flags);
						set_bit(R5_Wantread, &dev->flags);
						locked++;
					} else {
						set_bit(STRIPE_DELAYED, &sh->state);
						set_bit(STRIPE_HANDLE, &sh->state);
					}
				}
			}
		if (rcw <= rmw && rcw > 0)
			/* want reconstruct write, but need to get some data */
			for (i=disks; i--;) {
				dev = &sh->dev[i];
				if (!test_bit(R5_OVERWRITE, &dev->flags) && i != sh->pd_idx &&
				    !test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) &&
				    test_bit(R5_Insync, &dev->flags)) {
					if (test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))
					{
						PRINTK("Read_old block %d for Reconstruct\n", i);
						set_bit(R5_LOCKED, &dev->flags);
						set_bit(R5_Wantread, &dev->flags);
						locked++;
					} else {
						set_bit(STRIPE_DELAYED, &sh->state);
						set_bit(STRIPE_HANDLE, &sh->state);
					}
				}
			}
		/* now if nothing is locked, and if we have enough data, we can start a write request */
1662 1663
		if (locked == 0 && (rcw == 0 ||rmw == 0) &&
		    !test_bit(STRIPE_BIT_DELAY, &sh->state)) {
L
Linus Torvalds 已提交
1664
			PRINTK("Computing parity...\n");
1665
			compute_parity5(sh, rcw==0 ? RECONSTRUCT_WRITE : READ_MODIFY_WRITE);
L
Linus Torvalds 已提交
1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688
			/* now every locked buffer is ready to be written */
			for (i=disks; i--;)
				if (test_bit(R5_LOCKED, &sh->dev[i].flags)) {
					PRINTK("Writing block %d\n", i);
					locked++;
					set_bit(R5_Wantwrite, &sh->dev[i].flags);
					if (!test_bit(R5_Insync, &sh->dev[i].flags)
					    || (i==sh->pd_idx && failed == 0))
						set_bit(STRIPE_INSYNC, &sh->state);
				}
			if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) {
				atomic_dec(&conf->preread_active_stripes);
				if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD)
					md_wakeup_thread(conf->mddev->thread);
			}
		}
	}

	/* maybe we need to check and possibly fix the parity for this stripe
	 * Any reads will already have been scheduled, so we just see if enough data
	 * is available
	 */
	if (syncing && locked == 0 &&
N
NeilBrown 已提交
1689
	    !test_bit(STRIPE_INSYNC, &sh->state)) {
L
Linus Torvalds 已提交
1690 1691
		set_bit(STRIPE_HANDLE, &sh->state);
		if (failed == 0) {
1692
			BUG_ON(uptodate != disks);
1693
			compute_parity5(sh, CHECK_PARITY);
L
Linus Torvalds 已提交
1694
			uptodate--;
1695
			if (page_is_zero(sh->dev[sh->pd_idx].page)) {
L
Linus Torvalds 已提交
1696 1697
				/* parity is correct (on disc, not in buffer any more) */
				set_bit(STRIPE_INSYNC, &sh->state);
1698 1699 1700 1701 1702
			} else {
				conf->mddev->resync_mismatches += STRIPE_SECTORS;
				if (test_bit(MD_RECOVERY_CHECK, &conf->mddev->recovery))
					/* don't try to repair!! */
					set_bit(STRIPE_INSYNC, &sh->state);
N
NeilBrown 已提交
1703 1704 1705 1706
				else {
					compute_block(sh, sh->pd_idx);
					uptodate++;
				}
L
Linus Torvalds 已提交
1707 1708 1709
			}
		}
		if (!test_bit(STRIPE_INSYNC, &sh->state)) {
N
NeilBrown 已提交
1710
			/* either failed parity check, or recovery is happening */
L
Linus Torvalds 已提交
1711 1712 1713
			if (failed==0)
				failed_num = sh->pd_idx;
			dev = &sh->dev[failed_num];
N
NeilBrown 已提交
1714 1715 1716
			BUG_ON(!test_bit(R5_UPTODATE, &dev->flags));
			BUG_ON(uptodate != disks);

L
Linus Torvalds 已提交
1717 1718
			set_bit(R5_LOCKED, &dev->flags);
			set_bit(R5_Wantwrite, &dev->flags);
1719
			clear_bit(STRIPE_DEGRADED, &sh->state);
L
Linus Torvalds 已提交
1720 1721 1722 1723 1724 1725 1726 1727
			locked++;
			set_bit(STRIPE_INSYNC, &sh->state);
		}
	}
	if (syncing && locked == 0 && test_bit(STRIPE_INSYNC, &sh->state)) {
		md_done_sync(conf->mddev, STRIPE_SECTORS,1);
		clear_bit(STRIPE_SYNCING, &sh->state);
	}
1728 1729 1730 1731

	/* If the failed drive is just a ReadError, then we might need to progress
	 * the repair/check process
	 */
1732 1733
	if (failed == 1 && ! conf->mddev->ro &&
	    test_bit(R5_ReadError, &sh->dev[failed_num].flags)
1734 1735 1736 1737 1738 1739 1740 1741
	    && !test_bit(R5_LOCKED, &sh->dev[failed_num].flags)
	    && test_bit(R5_UPTODATE, &sh->dev[failed_num].flags)
		) {
		dev = &sh->dev[failed_num];
		if (!test_bit(R5_ReWrite, &dev->flags)) {
			set_bit(R5_Wantwrite, &dev->flags);
			set_bit(R5_ReWrite, &dev->flags);
			set_bit(R5_LOCKED, &dev->flags);
1742
			locked++;
1743 1744 1745 1746
		} else {
			/* let's read it back */
			set_bit(R5_Wantread, &dev->flags);
			set_bit(R5_LOCKED, &dev->flags);
1747
			locked++;
1748 1749 1750
		}
	}

1751 1752 1753 1754
	if (expanded && test_bit(STRIPE_EXPANDING, &sh->state)) {
		/* Need to write out all blocks after computing parity */
		sh->disks = conf->raid_disks;
		sh->pd_idx = stripe_to_pdidx(sh->sector, conf, conf->raid_disks);
1755
		compute_parity5(sh, RECONSTRUCT_WRITE);
1756 1757 1758 1759 1760 1761 1762 1763
		for (i= conf->raid_disks; i--;) {
			set_bit(R5_LOCKED, &sh->dev[i].flags);
			locked++;
			set_bit(R5_Wantwrite, &sh->dev[i].flags);
		}
		clear_bit(STRIPE_EXPANDING, &sh->state);
	} else if (expanded) {
		clear_bit(STRIPE_EXPAND_READY, &sh->state);
1764
		atomic_dec(&conf->reshape_stripes);
1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812
		wake_up(&conf->wait_for_overlap);
		md_done_sync(conf->mddev, STRIPE_SECTORS, 1);
	}

	if (expanding && locked == 0) {
		/* We have read all the blocks in this stripe and now we need to
		 * copy some of them into a target stripe for expand.
		 */
		clear_bit(STRIPE_EXPAND_SOURCE, &sh->state);
		for (i=0; i< sh->disks; i++)
			if (i != sh->pd_idx) {
				int dd_idx, pd_idx, j;
				struct stripe_head *sh2;

				sector_t bn = compute_blocknr(sh, i);
				sector_t s = raid5_compute_sector(bn, conf->raid_disks,
								  conf->raid_disks-1,
								  &dd_idx, &pd_idx, conf);
				sh2 = get_active_stripe(conf, s, conf->raid_disks, pd_idx, 1);
				if (sh2 == NULL)
					/* so far only the early blocks of this stripe
					 * have been requested.  When later blocks
					 * get requested, we will try again
					 */
					continue;
				if(!test_bit(STRIPE_EXPANDING, &sh2->state) ||
				   test_bit(R5_Expanded, &sh2->dev[dd_idx].flags)) {
					/* must have already done this block */
					release_stripe(sh2);
					continue;
				}
				memcpy(page_address(sh2->dev[dd_idx].page),
				       page_address(sh->dev[i].page),
				       STRIPE_SIZE);
				set_bit(R5_Expanded, &sh2->dev[dd_idx].flags);
				set_bit(R5_UPTODATE, &sh2->dev[dd_idx].flags);
				for (j=0; j<conf->raid_disks; j++)
					if (j != sh2->pd_idx &&
					    !test_bit(R5_Expanded, &sh2->dev[j].flags))
						break;
				if (j == conf->raid_disks) {
					set_bit(STRIPE_EXPAND_READY, &sh2->state);
					set_bit(STRIPE_HANDLE, &sh2->state);
				}
				release_stripe(sh2);
			}
	}

L
Linus Torvalds 已提交
1813 1814 1815 1816 1817 1818 1819 1820
	spin_unlock(&sh->lock);

	while ((bi=return_bi)) {
		int bytes = bi->bi_size;

		return_bi = bi->bi_next;
		bi->bi_next = NULL;
		bi->bi_size = 0;
1821 1822 1823
		bi->bi_end_io(bi, bytes,
			      test_bit(BIO_UPTODATE, &bi->bi_flags)
			        ? 0 : -EIO);
L
Linus Torvalds 已提交
1824 1825 1826 1827 1828 1829
	}
	for (i=disks; i-- ;) {
		int rw;
		struct bio *bi;
		mdk_rdev_t *rdev;
		if (test_and_clear_bit(R5_Wantwrite, &sh->dev[i].flags))
1830
			rw = WRITE;
L
Linus Torvalds 已提交
1831
		else if (test_and_clear_bit(R5_Wantread, &sh->dev[i].flags))
1832
			rw = READ;
L
Linus Torvalds 已提交
1833 1834 1835 1836 1837 1838
		else
			continue;
 
		bi = &sh->dev[i].req;
 
		bi->bi_rw = rw;
1839
		if (rw == WRITE)
L
Linus Torvalds 已提交
1840 1841 1842 1843 1844
			bi->bi_end_io = raid5_end_write_request;
		else
			bi->bi_end_io = raid5_end_read_request;
 
		rcu_read_lock();
1845
		rdev = rcu_dereference(conf->disks[i].rdev);
1846
		if (rdev && test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
1847 1848 1849 1850 1851 1852
			rdev = NULL;
		if (rdev)
			atomic_inc(&rdev->nr_pending);
		rcu_read_unlock();
 
		if (rdev) {
1853
			if (syncing || expanding || expanded)
L
Linus Torvalds 已提交
1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869
				md_sync_acct(rdev->bdev, STRIPE_SECTORS);

			bi->bi_bdev = rdev->bdev;
			PRINTK("for %llu schedule op %ld on disc %d\n",
				(unsigned long long)sh->sector, bi->bi_rw, i);
			atomic_inc(&sh->count);
			bi->bi_sector = sh->sector + rdev->data_offset;
			bi->bi_flags = 1 << BIO_UPTODATE;
			bi->bi_vcnt = 1;	
			bi->bi_max_vecs = 1;
			bi->bi_idx = 0;
			bi->bi_io_vec = &sh->dev[i].vec;
			bi->bi_io_vec[0].bv_len = STRIPE_SIZE;
			bi->bi_io_vec[0].bv_offset = 0;
			bi->bi_size = STRIPE_SIZE;
			bi->bi_next = NULL;
1870 1871 1872
			if (rw == WRITE &&
			    test_bit(R5_ReWrite, &sh->dev[i].flags))
				atomic_add(STRIPE_SECTORS, &rdev->corrected_errors);
L
Linus Torvalds 已提交
1873 1874
			generic_make_request(bi);
		} else {
1875
			if (rw == WRITE)
1876
				set_bit(STRIPE_DEGRADED, &sh->state);
L
Linus Torvalds 已提交
1877 1878 1879 1880 1881 1882 1883 1884
			PRINTK("skip op %ld on disc %d for sector %llu\n",
				bi->bi_rw, i, (unsigned long long)sh->sector);
			clear_bit(R5_LOCKED, &sh->dev[i].flags);
			set_bit(STRIPE_HANDLE, &sh->state);
		}
	}
}

1885
static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
L
Linus Torvalds 已提交
1886
{
1887
	raid6_conf_t *conf = sh->raid_conf;
1888
	int disks = sh->disks;
1889 1890 1891
	struct bio *return_bi= NULL;
	struct bio *bi;
	int i;
1892
	int syncing, expanding, expanded;
1893 1894 1895 1896 1897 1898 1899
	int locked=0, uptodate=0, to_read=0, to_write=0, failed=0, written=0;
	int non_overwrite = 0;
	int failed_num[2] = {0, 0};
	struct r5dev *dev, *pdev, *qdev;
	int pd_idx = sh->pd_idx;
	int qd_idx = raid6_next_disk(pd_idx, disks);
	int p_failed, q_failed;
L
Linus Torvalds 已提交
1900

1901 1902 1903
	PRINTK("handling stripe %llu, state=%#lx cnt=%d, pd_idx=%d, qd_idx=%d\n",
	       (unsigned long long)sh->sector, sh->state, atomic_read(&sh->count),
	       pd_idx, qd_idx);
1904

1905 1906 1907 1908 1909
	spin_lock(&sh->lock);
	clear_bit(STRIPE_HANDLE, &sh->state);
	clear_bit(STRIPE_DELAYED, &sh->state);

	syncing = test_bit(STRIPE_SYNCING, &sh->state);
1910 1911
	expanding = test_bit(STRIPE_EXPAND_SOURCE, &sh->state);
	expanded = test_bit(STRIPE_EXPAND_READY, &sh->state);
1912
	/* Now to look around and see what can be done */
L
Linus Torvalds 已提交
1913 1914

	rcu_read_lock();
1915 1916 1917 1918
	for (i=disks; i--; ) {
		mdk_rdev_t *rdev;
		dev = &sh->dev[i];
		clear_bit(R5_Insync, &dev->flags);
L
Linus Torvalds 已提交
1919

1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
		PRINTK("check %d: state 0x%lx read %p write %p written %p\n",
			i, dev->flags, dev->toread, dev->towrite, dev->written);
		/* maybe we can reply to a read */
		if (test_bit(R5_UPTODATE, &dev->flags) && dev->toread) {
			struct bio *rbi, *rbi2;
			PRINTK("Return read for disc %d\n", i);
			spin_lock_irq(&conf->device_lock);
			rbi = dev->toread;
			dev->toread = NULL;
			if (test_and_clear_bit(R5_Overlap, &dev->flags))
				wake_up(&conf->wait_for_overlap);
			spin_unlock_irq(&conf->device_lock);
			while (rbi && rbi->bi_sector < dev->sector + STRIPE_SECTORS) {
				copy_data(0, rbi, dev->page, dev->sector);
				rbi2 = r5_next_bio(rbi, dev->sector);
				spin_lock_irq(&conf->device_lock);
				if (--rbi->bi_phys_segments == 0) {
					rbi->bi_next = return_bi;
					return_bi = rbi;
				}
				spin_unlock_irq(&conf->device_lock);
				rbi = rbi2;
			}
		}
L
Linus Torvalds 已提交
1944

1945 1946 1947
		/* now count some things */
		if (test_bit(R5_LOCKED, &dev->flags)) locked++;
		if (test_bit(R5_UPTODATE, &dev->flags)) uptodate++;
L
Linus Torvalds 已提交
1948

1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961

		if (dev->toread) to_read++;
		if (dev->towrite) {
			to_write++;
			if (!test_bit(R5_OVERWRITE, &dev->flags))
				non_overwrite++;
		}
		if (dev->written) written++;
		rdev = rcu_dereference(conf->disks[i].rdev);
		if (!rdev || !test_bit(In_sync, &rdev->flags)) {
			/* The ReadError flag will just be confusing now */
			clear_bit(R5_ReadError, &dev->flags);
			clear_bit(R5_ReWrite, &dev->flags);
L
Linus Torvalds 已提交
1962
		}
1963 1964 1965 1966 1967 1968 1969
		if (!rdev || !test_bit(In_sync, &rdev->flags)
		    || test_bit(R5_ReadError, &dev->flags)) {
			if ( failed < 2 )
				failed_num[failed] = i;
			failed++;
		} else
			set_bit(R5_Insync, &dev->flags);
L
Linus Torvalds 已提交
1970 1971
	}
	rcu_read_unlock();
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981
	PRINTK("locked=%d uptodate=%d to_read=%d"
	       " to_write=%d failed=%d failed_num=%d,%d\n",
	       locked, uptodate, to_read, to_write, failed,
	       failed_num[0], failed_num[1]);
	/* check if the array has lost >2 devices and, if so, some requests might
	 * need to be failed
	 */
	if (failed > 2 && to_read+to_write+written) {
		for (i=disks; i--; ) {
			int bitmap_end = 0;
L
Linus Torvalds 已提交
1982

1983 1984 1985 1986 1987 1988 1989 1990 1991
			if (test_bit(R5_ReadError, &sh->dev[i].flags)) {
				mdk_rdev_t *rdev;
				rcu_read_lock();
				rdev = rcu_dereference(conf->disks[i].rdev);
				if (rdev && test_bit(In_sync, &rdev->flags))
					/* multiple read failures in one stripe */
					md_error(conf->mddev, rdev);
				rcu_read_unlock();
			}
L
Linus Torvalds 已提交
1992

1993 1994 1995 1996 1997
			spin_lock_irq(&conf->device_lock);
			/* fail all writes first */
			bi = sh->dev[i].towrite;
			sh->dev[i].towrite = NULL;
			if (bi) { to_write--; bitmap_end = 1; }
L
Linus Torvalds 已提交
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 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
			if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
				wake_up(&conf->wait_for_overlap);

			while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS){
				struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector);
				clear_bit(BIO_UPTODATE, &bi->bi_flags);
				if (--bi->bi_phys_segments == 0) {
					md_write_end(conf->mddev);
					bi->bi_next = return_bi;
					return_bi = bi;
				}
				bi = nextbi;
			}
			/* and fail all 'written' */
			bi = sh->dev[i].written;
			sh->dev[i].written = NULL;
			if (bi) bitmap_end = 1;
			while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS) {
				struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector);
				clear_bit(BIO_UPTODATE, &bi->bi_flags);
				if (--bi->bi_phys_segments == 0) {
					md_write_end(conf->mddev);
					bi->bi_next = return_bi;
					return_bi = bi;
				}
				bi = bi2;
			}

			/* fail any reads if this device is non-operational */
			if (!test_bit(R5_Insync, &sh->dev[i].flags) ||
			    test_bit(R5_ReadError, &sh->dev[i].flags)) {
				bi = sh->dev[i].toread;
				sh->dev[i].toread = NULL;
				if (test_and_clear_bit(R5_Overlap, &sh->dev[i].flags))
					wake_up(&conf->wait_for_overlap);
				if (bi) to_read--;
				while (bi && bi->bi_sector < sh->dev[i].sector + STRIPE_SECTORS){
					struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector);
					clear_bit(BIO_UPTODATE, &bi->bi_flags);
					if (--bi->bi_phys_segments == 0) {
						bi->bi_next = return_bi;
						return_bi = bi;
					}
					bi = nextbi;
				}
			}
			spin_unlock_irq(&conf->device_lock);
			if (bitmap_end)
				bitmap_endwrite(conf->mddev->bitmap, sh->sector,
						STRIPE_SECTORS, 0, 0);
		}
	}
	if (failed > 2 && syncing) {
		md_done_sync(conf->mddev, STRIPE_SECTORS,0);
		clear_bit(STRIPE_SYNCING, &sh->state);
		syncing = 0;
	}

	/*
	 * might be able to return some write requests if the parity blocks
	 * are safe, or on a failed drive
	 */
	pdev = &sh->dev[pd_idx];
	p_failed = (failed >= 1 && failed_num[0] == pd_idx)
		|| (failed >= 2 && failed_num[1] == pd_idx);
	qdev = &sh->dev[qd_idx];
	q_failed = (failed >= 1 && failed_num[0] == qd_idx)
		|| (failed >= 2 && failed_num[1] == qd_idx);

	if ( written &&
	     ( p_failed || ((test_bit(R5_Insync, &pdev->flags)
			     && !test_bit(R5_LOCKED, &pdev->flags)
			     && test_bit(R5_UPTODATE, &pdev->flags))) ) &&
	     ( q_failed || ((test_bit(R5_Insync, &qdev->flags)
			     && !test_bit(R5_LOCKED, &qdev->flags)
			     && test_bit(R5_UPTODATE, &qdev->flags))) ) ) {
		/* any written block on an uptodate or failed drive can be
		 * returned.  Note that if we 'wrote' to a failed drive,
		 * it will be UPTODATE, but never LOCKED, so we don't need
		 * to test 'failed' directly.
		 */
		for (i=disks; i--; )
			if (sh->dev[i].written) {
				dev = &sh->dev[i];
				if (!test_bit(R5_LOCKED, &dev->flags) &&
				    test_bit(R5_UPTODATE, &dev->flags) ) {
					/* We can return any write requests */
					int bitmap_end = 0;
					struct bio *wbi, *wbi2;
					PRINTK("Return write for stripe %llu disc %d\n",
					       (unsigned long long)sh->sector, i);
					spin_lock_irq(&conf->device_lock);
					wbi = dev->written;
					dev->written = NULL;
					while (wbi && wbi->bi_sector < dev->sector + STRIPE_SECTORS) {
						wbi2 = r5_next_bio(wbi, dev->sector);
						if (--wbi->bi_phys_segments == 0) {
							md_write_end(conf->mddev);
							wbi->bi_next = return_bi;
							return_bi = wbi;
						}
						wbi = wbi2;
					}
					if (dev->towrite == NULL)
						bitmap_end = 1;
					spin_unlock_irq(&conf->device_lock);
					if (bitmap_end)
						bitmap_endwrite(conf->mddev->bitmap, sh->sector,
								STRIPE_SECTORS,
								!test_bit(STRIPE_DEGRADED, &sh->state), 0);
				}
			}
	}

	/* Now we might consider reading some blocks, either to check/generate
	 * parity, or to satisfy requests
	 * or to load a block that is being partially written.
	 */
2117 2118
	if (to_read || non_overwrite || (to_write && failed) ||
	    (syncing && (uptodate < disks)) || expanding) {
2119 2120 2121 2122 2123 2124
		for (i=disks; i--;) {
			dev = &sh->dev[i];
			if (!test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) &&
			    (dev->toread ||
			     (dev->towrite && !test_bit(R5_OVERWRITE, &dev->flags)) ||
			     syncing ||
2125
			     expanding ||
2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 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 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 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 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359
			     (failed >= 1 && (sh->dev[failed_num[0]].toread || to_write)) ||
			     (failed >= 2 && (sh->dev[failed_num[1]].toread || to_write))
				    )
				) {
				/* we would like to get this block, possibly
				 * by computing it, but we might not be able to
				 */
				if (uptodate == disks-1) {
					PRINTK("Computing stripe %llu block %d\n",
					       (unsigned long long)sh->sector, i);
					compute_block_1(sh, i, 0);
					uptodate++;
				} else if ( uptodate == disks-2 && failed >= 2 ) {
					/* Computing 2-failure is *very* expensive; only do it if failed >= 2 */
					int other;
					for (other=disks; other--;) {
						if ( other == i )
							continue;
						if ( !test_bit(R5_UPTODATE, &sh->dev[other].flags) )
							break;
					}
					BUG_ON(other < 0);
					PRINTK("Computing stripe %llu blocks %d,%d\n",
					       (unsigned long long)sh->sector, i, other);
					compute_block_2(sh, i, other);
					uptodate += 2;
				} else if (test_bit(R5_Insync, &dev->flags)) {
					set_bit(R5_LOCKED, &dev->flags);
					set_bit(R5_Wantread, &dev->flags);
					locked++;
					PRINTK("Reading block %d (sync=%d)\n",
						i, syncing);
				}
			}
		}
		set_bit(STRIPE_HANDLE, &sh->state);
	}

	/* now to consider writing and what else, if anything should be read */
	if (to_write) {
		int rcw=0, must_compute=0;
		for (i=disks ; i--;) {
			dev = &sh->dev[i];
			/* Would I have to read this buffer for reconstruct_write */
			if (!test_bit(R5_OVERWRITE, &dev->flags)
			    && i != pd_idx && i != qd_idx
			    && (!test_bit(R5_LOCKED, &dev->flags)
				    ) &&
			    !test_bit(R5_UPTODATE, &dev->flags)) {
				if (test_bit(R5_Insync, &dev->flags)) rcw++;
				else {
					PRINTK("raid6: must_compute: disk %d flags=%#lx\n", i, dev->flags);
					must_compute++;
				}
			}
		}
		PRINTK("for sector %llu, rcw=%d, must_compute=%d\n",
		       (unsigned long long)sh->sector, rcw, must_compute);
		set_bit(STRIPE_HANDLE, &sh->state);

		if (rcw > 0)
			/* want reconstruct write, but need to get some data */
			for (i=disks; i--;) {
				dev = &sh->dev[i];
				if (!test_bit(R5_OVERWRITE, &dev->flags)
				    && !(failed == 0 && (i == pd_idx || i == qd_idx))
				    && !test_bit(R5_LOCKED, &dev->flags) && !test_bit(R5_UPTODATE, &dev->flags) &&
				    test_bit(R5_Insync, &dev->flags)) {
					if (test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))
					{
						PRINTK("Read_old stripe %llu block %d for Reconstruct\n",
						       (unsigned long long)sh->sector, i);
						set_bit(R5_LOCKED, &dev->flags);
						set_bit(R5_Wantread, &dev->flags);
						locked++;
					} else {
						PRINTK("Request delayed stripe %llu block %d for Reconstruct\n",
						       (unsigned long long)sh->sector, i);
						set_bit(STRIPE_DELAYED, &sh->state);
						set_bit(STRIPE_HANDLE, &sh->state);
					}
				}
			}
		/* now if nothing is locked, and if we have enough data, we can start a write request */
		if (locked == 0 && rcw == 0 &&
		    !test_bit(STRIPE_BIT_DELAY, &sh->state)) {
			if ( must_compute > 0 ) {
				/* We have failed blocks and need to compute them */
				switch ( failed ) {
				case 0:	BUG();
				case 1: compute_block_1(sh, failed_num[0], 0); break;
				case 2: compute_block_2(sh, failed_num[0], failed_num[1]); break;
				default: BUG();	/* This request should have been failed? */
				}
			}

			PRINTK("Computing parity for stripe %llu\n", (unsigned long long)sh->sector);
			compute_parity6(sh, RECONSTRUCT_WRITE);
			/* now every locked buffer is ready to be written */
			for (i=disks; i--;)
				if (test_bit(R5_LOCKED, &sh->dev[i].flags)) {
					PRINTK("Writing stripe %llu block %d\n",
					       (unsigned long long)sh->sector, i);
					locked++;
					set_bit(R5_Wantwrite, &sh->dev[i].flags);
				}
			/* after a RECONSTRUCT_WRITE, the stripe MUST be in-sync */
			set_bit(STRIPE_INSYNC, &sh->state);

			if (test_and_clear_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) {
				atomic_dec(&conf->preread_active_stripes);
				if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD)
					md_wakeup_thread(conf->mddev->thread);
			}
		}
	}

	/* maybe we need to check and possibly fix the parity for this stripe
	 * Any reads will already have been scheduled, so we just see if enough data
	 * is available
	 */
	if (syncing && locked == 0 && !test_bit(STRIPE_INSYNC, &sh->state)) {
		int update_p = 0, update_q = 0;
		struct r5dev *dev;

		set_bit(STRIPE_HANDLE, &sh->state);

		BUG_ON(failed>2);
		BUG_ON(uptodate < disks);
		/* Want to check and possibly repair P and Q.
		 * However there could be one 'failed' device, in which
		 * case we can only check one of them, possibly using the
		 * other to generate missing data
		 */

		/* If !tmp_page, we cannot do the calculations,
		 * but as we have set STRIPE_HANDLE, we will soon be called
		 * by stripe_handle with a tmp_page - just wait until then.
		 */
		if (tmp_page) {
			if (failed == q_failed) {
				/* The only possible failed device holds 'Q', so it makes
				 * sense to check P (If anything else were failed, we would
				 * have used P to recreate it).
				 */
				compute_block_1(sh, pd_idx, 1);
				if (!page_is_zero(sh->dev[pd_idx].page)) {
					compute_block_1(sh,pd_idx,0);
					update_p = 1;
				}
			}
			if (!q_failed && failed < 2) {
				/* q is not failed, and we didn't use it to generate
				 * anything, so it makes sense to check it
				 */
				memcpy(page_address(tmp_page),
				       page_address(sh->dev[qd_idx].page),
				       STRIPE_SIZE);
				compute_parity6(sh, UPDATE_PARITY);
				if (memcmp(page_address(tmp_page),
					   page_address(sh->dev[qd_idx].page),
					   STRIPE_SIZE)!= 0) {
					clear_bit(STRIPE_INSYNC, &sh->state);
					update_q = 1;
				}
			}
			if (update_p || update_q) {
				conf->mddev->resync_mismatches += STRIPE_SECTORS;
				if (test_bit(MD_RECOVERY_CHECK, &conf->mddev->recovery))
					/* don't try to repair!! */
					update_p = update_q = 0;
			}

			/* now write out any block on a failed drive,
			 * or P or Q if they need it
			 */

			if (failed == 2) {
				dev = &sh->dev[failed_num[1]];
				locked++;
				set_bit(R5_LOCKED, &dev->flags);
				set_bit(R5_Wantwrite, &dev->flags);
			}
			if (failed >= 1) {
				dev = &sh->dev[failed_num[0]];
				locked++;
				set_bit(R5_LOCKED, &dev->flags);
				set_bit(R5_Wantwrite, &dev->flags);
			}

			if (update_p) {
				dev = &sh->dev[pd_idx];
				locked ++;
				set_bit(R5_LOCKED, &dev->flags);
				set_bit(R5_Wantwrite, &dev->flags);
			}
			if (update_q) {
				dev = &sh->dev[qd_idx];
				locked++;
				set_bit(R5_LOCKED, &dev->flags);
				set_bit(R5_Wantwrite, &dev->flags);
			}
			clear_bit(STRIPE_DEGRADED, &sh->state);

			set_bit(STRIPE_INSYNC, &sh->state);
		}
	}

	if (syncing && locked == 0 && test_bit(STRIPE_INSYNC, &sh->state)) {
		md_done_sync(conf->mddev, STRIPE_SECTORS,1);
		clear_bit(STRIPE_SYNCING, &sh->state);
	}

	/* If the failed drives are just a ReadError, then we might need
	 * to progress the repair/check process
	 */
	if (failed <= 2 && ! conf->mddev->ro)
		for (i=0; i<failed;i++) {
			dev = &sh->dev[failed_num[i]];
			if (test_bit(R5_ReadError, &dev->flags)
			    && !test_bit(R5_LOCKED, &dev->flags)
			    && test_bit(R5_UPTODATE, &dev->flags)
				) {
				if (!test_bit(R5_ReWrite, &dev->flags)) {
					set_bit(R5_Wantwrite, &dev->flags);
					set_bit(R5_ReWrite, &dev->flags);
					set_bit(R5_LOCKED, &dev->flags);
				} else {
					/* let's read it back */
					set_bit(R5_Wantread, &dev->flags);
					set_bit(R5_LOCKED, &dev->flags);
				}
			}
		}
2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432

	if (expanded && test_bit(STRIPE_EXPANDING, &sh->state)) {
		/* Need to write out all blocks after computing P&Q */
		sh->disks = conf->raid_disks;
		sh->pd_idx = stripe_to_pdidx(sh->sector, conf,
					     conf->raid_disks);
		compute_parity6(sh, RECONSTRUCT_WRITE);
		for (i = conf->raid_disks ; i-- ;  ) {
			set_bit(R5_LOCKED, &sh->dev[i].flags);
			locked++;
			set_bit(R5_Wantwrite, &sh->dev[i].flags);
		}
		clear_bit(STRIPE_EXPANDING, &sh->state);
	} else if (expanded) {
		clear_bit(STRIPE_EXPAND_READY, &sh->state);
		atomic_dec(&conf->reshape_stripes);
		wake_up(&conf->wait_for_overlap);
		md_done_sync(conf->mddev, STRIPE_SECTORS, 1);
	}

	if (expanding && locked == 0) {
		/* We have read all the blocks in this stripe and now we need to
		 * copy some of them into a target stripe for expand.
		 */
		clear_bit(STRIPE_EXPAND_SOURCE, &sh->state);
		for (i = 0; i < sh->disks ; i++)
			if (i != pd_idx && i != qd_idx) {
				int dd_idx2, pd_idx2, j;
				struct stripe_head *sh2;

				sector_t bn = compute_blocknr(sh, i);
				sector_t s = raid5_compute_sector(
					bn, conf->raid_disks,
					conf->raid_disks - conf->max_degraded,
					&dd_idx2, &pd_idx2, conf);
				sh2 = get_active_stripe(conf, s,
							conf->raid_disks,
						       pd_idx2, 1);
				if (sh2 == NULL)
					/* so for only the early blocks of
					 * this stripe have been requests.
					 * When later blocks get requests, we
					 * will try again
					 */
					continue;
				if (!test_bit(STRIPE_EXPANDING, &sh2->state) ||
				    test_bit(R5_Expanded,
					     &sh2->dev[dd_idx2].flags)) {
					/* must have already done this block */
					release_stripe(sh2);
					continue;
				}
				memcpy(page_address(sh2->dev[dd_idx2].page),
				       page_address(sh->dev[i].page),
				       STRIPE_SIZE);
				set_bit(R5_Expanded, &sh2->dev[dd_idx2].flags);
				set_bit(R5_UPTODATE, &sh2->dev[dd_idx2].flags);
				for (j = 0 ; j < conf->raid_disks ; j++)
					if (j != sh2->pd_idx &&
					    j != raid6_next_disk(sh2->pd_idx,
							   sh2->disks) &&
					    !test_bit(R5_Expanded,
						      &sh2->dev[j].flags))
						break;
				if (j == conf->raid_disks) {
					set_bit(STRIPE_EXPAND_READY,
						&sh2->state);
					set_bit(STRIPE_HANDLE, &sh2->state);
				}
				release_stripe(sh2);
			}
	}

2433 2434 2435 2436 2437 2438 2439 2440
	spin_unlock(&sh->lock);

	while ((bi=return_bi)) {
		int bytes = bi->bi_size;

		return_bi = bi->bi_next;
		bi->bi_next = NULL;
		bi->bi_size = 0;
2441 2442 2443
		bi->bi_end_io(bi, bytes,
			      test_bit(BIO_UPTODATE, &bi->bi_flags)
			        ? 0 : -EIO);
2444 2445 2446 2447 2448 2449
	}
	for (i=disks; i-- ;) {
		int rw;
		struct bio *bi;
		mdk_rdev_t *rdev;
		if (test_and_clear_bit(R5_Wantwrite, &sh->dev[i].flags))
2450
			rw = WRITE;
2451
		else if (test_and_clear_bit(R5_Wantread, &sh->dev[i].flags))
2452
			rw = READ;
2453 2454 2455 2456 2457 2458
		else
			continue;

		bi = &sh->dev[i].req;

		bi->bi_rw = rw;
2459
		if (rw == WRITE)
2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472
			bi->bi_end_io = raid5_end_write_request;
		else
			bi->bi_end_io = raid5_end_read_request;

		rcu_read_lock();
		rdev = rcu_dereference(conf->disks[i].rdev);
		if (rdev && test_bit(Faulty, &rdev->flags))
			rdev = NULL;
		if (rdev)
			atomic_inc(&rdev->nr_pending);
		rcu_read_unlock();

		if (rdev) {
2473
			if (syncing || expanding || expanded)
2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494
				md_sync_acct(rdev->bdev, STRIPE_SECTORS);

			bi->bi_bdev = rdev->bdev;
			PRINTK("for %llu schedule op %ld on disc %d\n",
				(unsigned long long)sh->sector, bi->bi_rw, i);
			atomic_inc(&sh->count);
			bi->bi_sector = sh->sector + rdev->data_offset;
			bi->bi_flags = 1 << BIO_UPTODATE;
			bi->bi_vcnt = 1;
			bi->bi_max_vecs = 1;
			bi->bi_idx = 0;
			bi->bi_io_vec = &sh->dev[i].vec;
			bi->bi_io_vec[0].bv_len = STRIPE_SIZE;
			bi->bi_io_vec[0].bv_offset = 0;
			bi->bi_size = STRIPE_SIZE;
			bi->bi_next = NULL;
			if (rw == WRITE &&
			    test_bit(R5_ReWrite, &sh->dev[i].flags))
				atomic_add(STRIPE_SECTORS, &rdev->corrected_errors);
			generic_make_request(bi);
		} else {
2495
			if (rw == WRITE)
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579
				set_bit(STRIPE_DEGRADED, &sh->state);
			PRINTK("skip op %ld on disc %d for sector %llu\n",
				bi->bi_rw, i, (unsigned long long)sh->sector);
			clear_bit(R5_LOCKED, &sh->dev[i].flags);
			set_bit(STRIPE_HANDLE, &sh->state);
		}
	}
}

static void handle_stripe(struct stripe_head *sh, struct page *tmp_page)
{
	if (sh->raid_conf->level == 6)
		handle_stripe6(sh, tmp_page);
	else
		handle_stripe5(sh);
}



static void raid5_activate_delayed(raid5_conf_t *conf)
{
	if (atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD) {
		while (!list_empty(&conf->delayed_list)) {
			struct list_head *l = conf->delayed_list.next;
			struct stripe_head *sh;
			sh = list_entry(l, struct stripe_head, lru);
			list_del_init(l);
			clear_bit(STRIPE_DELAYED, &sh->state);
			if (!test_and_set_bit(STRIPE_PREREAD_ACTIVE, &sh->state))
				atomic_inc(&conf->preread_active_stripes);
			list_add_tail(&sh->lru, &conf->handle_list);
		}
	}
}

static void activate_bit_delay(raid5_conf_t *conf)
{
	/* device_lock is held */
	struct list_head head;
	list_add(&head, &conf->bitmap_list);
	list_del_init(&conf->bitmap_list);
	while (!list_empty(&head)) {
		struct stripe_head *sh = list_entry(head.next, struct stripe_head, lru);
		list_del_init(&sh->lru);
		atomic_inc(&sh->count);
		__release_stripe(conf, sh);
	}
}

static void unplug_slaves(mddev_t *mddev)
{
	raid5_conf_t *conf = mddev_to_conf(mddev);
	int i;

	rcu_read_lock();
	for (i=0; i<mddev->raid_disks; i++) {
		mdk_rdev_t *rdev = rcu_dereference(conf->disks[i].rdev);
		if (rdev && !test_bit(Faulty, &rdev->flags) && atomic_read(&rdev->nr_pending)) {
			request_queue_t *r_queue = bdev_get_queue(rdev->bdev);

			atomic_inc(&rdev->nr_pending);
			rcu_read_unlock();

			if (r_queue->unplug_fn)
				r_queue->unplug_fn(r_queue);

			rdev_dec_pending(rdev, mddev);
			rcu_read_lock();
		}
	}
	rcu_read_unlock();
}

static void raid5_unplug_device(request_queue_t *q)
{
	mddev_t *mddev = q->queuedata;
	raid5_conf_t *conf = mddev_to_conf(mddev);
	unsigned long flags;

	spin_lock_irqsave(&conf->device_lock, flags);

	if (blk_remove_plug(q)) {
		conf->seq_flush++;
		raid5_activate_delayed(conf);
2580
	}
L
Linus Torvalds 已提交
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596
	md_wakeup_thread(mddev->thread);

	spin_unlock_irqrestore(&conf->device_lock, flags);

	unplug_slaves(mddev);
}

static int raid5_issue_flush(request_queue_t *q, struct gendisk *disk,
			     sector_t *error_sector)
{
	mddev_t *mddev = q->queuedata;
	raid5_conf_t *conf = mddev_to_conf(mddev);
	int i, ret = 0;

	rcu_read_lock();
	for (i=0; i<mddev->raid_disks && ret == 0; i++) {
2597
		mdk_rdev_t *rdev = rcu_dereference(conf->disks[i].rdev);
2598
		if (rdev && !test_bit(Faulty, &rdev->flags)) {
L
Linus Torvalds 已提交
2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617
			struct block_device *bdev = rdev->bdev;
			request_queue_t *r_queue = bdev_get_queue(bdev);

			if (!r_queue->issue_flush_fn)
				ret = -EOPNOTSUPP;
			else {
				atomic_inc(&rdev->nr_pending);
				rcu_read_unlock();
				ret = r_queue->issue_flush_fn(r_queue, bdev->bd_disk,
							      error_sector);
				rdev_dec_pending(rdev, mddev);
				rcu_read_lock();
			}
		}
	}
	rcu_read_unlock();
	return ret;
}

2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635
static int raid5_congested(void *data, int bits)
{
	mddev_t *mddev = data;
	raid5_conf_t *conf = mddev_to_conf(mddev);

	/* No difference between reads and writes.  Just check
	 * how busy the stripe_cache is
	 */
	if (conf->inactive_blocked)
		return 1;
	if (conf->quiesce)
		return 1;
	if (list_empty_careful(&conf->inactive_list))
		return 1;

	return 0;
}

2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646
/* We want read requests to align with chunks where possible,
 * but write requests don't need to.
 */
static int raid5_mergeable_bvec(request_queue_t *q, struct bio *bio, struct bio_vec *biovec)
{
	mddev_t *mddev = q->queuedata;
	sector_t sector = bio->bi_sector + get_start_sect(bio->bi_bdev);
	int max;
	unsigned int chunk_sectors = mddev->chunk_size >> 9;
	unsigned int bio_sectors = bio->bi_size >> 9;

2647
	if (bio_data_dir(bio) == WRITE)
2648 2649 2650 2651 2652 2653 2654 2655 2656 2657
		return biovec->bv_len; /* always allow writes to be mergeable */

	max =  (chunk_sectors - ((sector & (chunk_sectors - 1)) + bio_sectors)) << 9;
	if (max < 0) max = 0;
	if (max <= biovec->bv_len && bio_sectors == 0)
		return biovec->bv_len;
	else
		return max;
}

2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668

static int in_chunk_boundary(mddev_t *mddev, struct bio *bio)
{
	sector_t sector = bio->bi_sector + get_start_sect(bio->bi_bdev);
	unsigned int chunk_sectors = mddev->chunk_size >> 9;
	unsigned int bio_sectors = bio->bi_size >> 9;

	return  chunk_sectors >=
		((sector & (chunk_sectors - 1)) + bio_sectors);
}

2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697
/*
 *  add bio to the retry LIFO  ( in O(1) ... we are in interrupt )
 *  later sampled by raid5d.
 */
static void add_bio_to_retry(struct bio *bi,raid5_conf_t *conf)
{
	unsigned long flags;

	spin_lock_irqsave(&conf->device_lock, flags);

	bi->bi_next = conf->retry_read_aligned_list;
	conf->retry_read_aligned_list = bi;

	spin_unlock_irqrestore(&conf->device_lock, flags);
	md_wakeup_thread(conf->mddev->thread);
}


static struct bio *remove_bio_from_retry(raid5_conf_t *conf)
{
	struct bio *bi;

	bi = conf->retry_read_aligned;
	if (bi) {
		conf->retry_read_aligned = NULL;
		return bi;
	}
	bi = conf->retry_read_aligned_list;
	if(bi) {
2698
		conf->retry_read_aligned_list = bi->bi_next;
2699 2700 2701 2702 2703 2704 2705 2706 2707
		bi->bi_next = NULL;
		bi->bi_phys_segments = 1; /* biased count of active stripes */
		bi->bi_hw_segments = 0; /* count of processed stripes */
	}

	return bi;
}


2708 2709 2710 2711 2712 2713
/*
 *  The "raid5_align_endio" should check if the read succeeded and if it
 *  did, call bio_endio on the original bio (having bio_put the new bio
 *  first).
 *  If the read failed..
 */
2714
static int raid5_align_endio(struct bio *bi, unsigned int bytes, int error)
2715 2716
{
	struct bio* raid_bi  = bi->bi_private;
2717 2718 2719 2720 2721
	mddev_t *mddev;
	raid5_conf_t *conf;
	int uptodate = test_bit(BIO_UPTODATE, &bi->bi_flags);
	mdk_rdev_t *rdev;

2722 2723 2724
	if (bi->bi_size)
		return 1;
	bio_put(bi);
2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743

	mddev = raid_bi->bi_bdev->bd_disk->queue->queuedata;
	conf = mddev_to_conf(mddev);
	rdev = (void*)raid_bi->bi_next;
	raid_bi->bi_next = NULL;

	rdev_dec_pending(rdev, conf->mddev);

	if (!error && uptodate) {
		bio_endio(raid_bi, bytes, 0);
		if (atomic_dec_and_test(&conf->active_aligned_reads))
			wake_up(&conf->wait_for_stripe);
		return 0;
	}


	PRINTK("raid5_align_endio : io error...handing IO for a retry\n");

	add_bio_to_retry(raid_bi, conf);
2744 2745 2746
	return 0;
}

2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767
static int bio_fits_rdev(struct bio *bi)
{
	request_queue_t *q = bdev_get_queue(bi->bi_bdev);

	if ((bi->bi_size>>9) > q->max_sectors)
		return 0;
	blk_recount_segments(q, bi);
	if (bi->bi_phys_segments > q->max_phys_segments ||
	    bi->bi_hw_segments > q->max_hw_segments)
		return 0;

	if (q->merge_bvec_fn)
		/* it's too hard to apply the merge_bvec_fn at this stage,
		 * just just give up
		 */
		return 0;

	return 1;
}


2768 2769 2770 2771 2772
static int chunk_aligned_read(request_queue_t *q, struct bio * raid_bio)
{
	mddev_t *mddev = q->queuedata;
	raid5_conf_t *conf = mddev_to_conf(mddev);
	const unsigned int raid_disks = conf->raid_disks;
2773
	const unsigned int data_disks = raid_disks - conf->max_degraded;
2774 2775 2776 2777 2778
	unsigned int dd_idx, pd_idx;
	struct bio* align_bi;
	mdk_rdev_t *rdev;

	if (!in_chunk_boundary(mddev, raid_bio)) {
2779
		PRINTK("chunk_aligned_read : non aligned\n");
2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808
		return 0;
	}
	/*
 	 * use bio_clone to make a copy of the bio
	 */
	align_bi = bio_clone(raid_bio, GFP_NOIO);
	if (!align_bi)
		return 0;
	/*
	 *   set bi_end_io to a new function, and set bi_private to the
	 *     original bio.
	 */
	align_bi->bi_end_io  = raid5_align_endio;
	align_bi->bi_private = raid_bio;
	/*
	 *	compute position
	 */
	align_bi->bi_sector =  raid5_compute_sector(raid_bio->bi_sector,
					raid_disks,
					data_disks,
					&dd_idx,
					&pd_idx,
					conf);

	rcu_read_lock();
	rdev = rcu_dereference(conf->disks[dd_idx].rdev);
	if (rdev && test_bit(In_sync, &rdev->flags)) {
		atomic_inc(&rdev->nr_pending);
		rcu_read_unlock();
2809 2810 2811 2812 2813
		raid_bio->bi_next = (void*)rdev;
		align_bi->bi_bdev =  rdev->bdev;
		align_bi->bi_flags &= ~(1 << BIO_SEG_VALID);
		align_bi->bi_sector += rdev->data_offset;

2814 2815 2816 2817 2818 2819 2820
		if (!bio_fits_rdev(align_bi)) {
			/* too big in some way */
			bio_put(align_bi);
			rdev_dec_pending(rdev, mddev);
			return 0;
		}

2821 2822 2823 2824 2825 2826 2827
		spin_lock_irq(&conf->device_lock);
		wait_event_lock_irq(conf->wait_for_stripe,
				    conf->quiesce == 0,
				    conf->device_lock, /* nothing */);
		atomic_inc(&conf->active_aligned_reads);
		spin_unlock_irq(&conf->device_lock);

2828 2829 2830 2831
		generic_make_request(align_bi);
		return 1;
	} else {
		rcu_read_unlock();
2832
		bio_put(align_bi);
2833 2834 2835 2836 2837
		return 0;
	}
}


2838
static int make_request(request_queue_t *q, struct bio * bi)
L
Linus Torvalds 已提交
2839 2840 2841 2842 2843 2844 2845
{
	mddev_t *mddev = q->queuedata;
	raid5_conf_t *conf = mddev_to_conf(mddev);
	unsigned int dd_idx, pd_idx;
	sector_t new_sector;
	sector_t logical_sector, last_sector;
	struct stripe_head *sh;
2846
	const int rw = bio_data_dir(bi);
2847
	int remaining;
L
Linus Torvalds 已提交
2848

2849 2850 2851 2852 2853
	if (unlikely(bio_barrier(bi))) {
		bio_endio(bi, bi->bi_size, -EOPNOTSUPP);
		return 0;
	}

2854
	md_write_start(mddev, bi);
2855

2856 2857
	disk_stat_inc(mddev->gendisk, ios[rw]);
	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bi));
L
Linus Torvalds 已提交
2858

2859
	if (rw == READ &&
2860 2861 2862 2863
	     mddev->reshape_position == MaxSector &&
	     chunk_aligned_read(q,bi))
            	return 0;

L
Linus Torvalds 已提交
2864 2865 2866 2867
	logical_sector = bi->bi_sector & ~((sector_t)STRIPE_SECTORS-1);
	last_sector = bi->bi_sector + (bi->bi_size>>9);
	bi->bi_next = NULL;
	bi->bi_phys_segments = 1;	/* over-loaded to count active stripes */
2868

L
Linus Torvalds 已提交
2869 2870
	for (;logical_sector < last_sector; logical_sector += STRIPE_SECTORS) {
		DEFINE_WAIT(w);
2871
		int disks, data_disks;
2872

2873
	retry:
2874
		prepare_to_wait(&conf->wait_for_overlap, &w, TASK_UNINTERRUPTIBLE);
2875 2876 2877
		if (likely(conf->expand_progress == MaxSector))
			disks = conf->raid_disks;
		else {
2878 2879 2880 2881 2882 2883 2884 2885
			/* spinlock is needed as expand_progress may be
			 * 64bit on a 32bit platform, and so it might be
			 * possible to see a half-updated value
			 * Ofcourse expand_progress could change after
			 * the lock is dropped, so once we get a reference
			 * to the stripe that we think it is, we will have
			 * to check again.
			 */
2886 2887 2888 2889
			spin_lock_irq(&conf->device_lock);
			disks = conf->raid_disks;
			if (logical_sector >= conf->expand_progress)
				disks = conf->previous_raid_disks;
2890 2891 2892 2893 2894 2895 2896
			else {
				if (logical_sector >= conf->expand_lo) {
					spin_unlock_irq(&conf->device_lock);
					schedule();
					goto retry;
				}
			}
2897 2898
			spin_unlock_irq(&conf->device_lock);
		}
2899 2900 2901
		data_disks = disks - conf->max_degraded;

 		new_sector = raid5_compute_sector(logical_sector, disks, data_disks,
2902
						  &dd_idx, &pd_idx, conf);
L
Linus Torvalds 已提交
2903 2904 2905 2906
		PRINTK("raid5: make_request, sector %llu logical %llu\n",
			(unsigned long long)new_sector, 
			(unsigned long long)logical_sector);

2907
		sh = get_active_stripe(conf, new_sector, disks, pd_idx, (bi->bi_rw&RWA_MASK));
L
Linus Torvalds 已提交
2908
		if (sh) {
2909 2910
			if (unlikely(conf->expand_progress != MaxSector)) {
				/* expansion might have moved on while waiting for a
2911 2912 2913 2914 2915 2916
				 * stripe, so we must do the range check again.
				 * Expansion could still move past after this
				 * test, but as we are holding a reference to
				 * 'sh', we know that if that happens,
				 *  STRIPE_EXPANDING will get set and the expansion
				 * won't proceed until we finish with the stripe.
2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929
				 */
				int must_retry = 0;
				spin_lock_irq(&conf->device_lock);
				if (logical_sector <  conf->expand_progress &&
				    disks == conf->previous_raid_disks)
					/* mismatch, need to try again */
					must_retry = 1;
				spin_unlock_irq(&conf->device_lock);
				if (must_retry) {
					release_stripe(sh);
					goto retry;
				}
			}
2930 2931 2932 2933 2934 2935 2936 2937 2938
			/* FIXME what if we get a false positive because these
			 * are being updated.
			 */
			if (logical_sector >= mddev->suspend_lo &&
			    logical_sector < mddev->suspend_hi) {
				release_stripe(sh);
				schedule();
				goto retry;
			}
2939 2940 2941 2942 2943

			if (test_bit(STRIPE_EXPANDING, &sh->state) ||
			    !add_stripe_bio(sh, bi, dd_idx, (bi->bi_rw&RW_MASK))) {
				/* Stripe is busy expanding or
				 * add failed due to overlap.  Flush everything
L
Linus Torvalds 已提交
2944 2945 2946 2947 2948 2949 2950 2951
				 * and wait a while
				 */
				raid5_unplug_device(mddev->queue);
				release_stripe(sh);
				schedule();
				goto retry;
			}
			finish_wait(&conf->wait_for_overlap, &w);
2952
			handle_stripe(sh, NULL);
L
Linus Torvalds 已提交
2953 2954 2955 2956 2957 2958 2959 2960 2961 2962
			release_stripe(sh);
		} else {
			/* cannot get stripe for read-ahead, just give-up */
			clear_bit(BIO_UPTODATE, &bi->bi_flags);
			finish_wait(&conf->wait_for_overlap, &w);
			break;
		}
			
	}
	spin_lock_irq(&conf->device_lock);
2963 2964 2965
	remaining = --bi->bi_phys_segments;
	spin_unlock_irq(&conf->device_lock);
	if (remaining == 0) {
L
Linus Torvalds 已提交
2966 2967
		int bytes = bi->bi_size;

2968
		if ( rw == WRITE )
L
Linus Torvalds 已提交
2969 2970
			md_write_end(mddev);
		bi->bi_size = 0;
2971 2972 2973
		bi->bi_end_io(bi, bytes,
			      test_bit(BIO_UPTODATE, &bi->bi_flags)
			        ? 0 : -EIO);
L
Linus Torvalds 已提交
2974 2975 2976 2977
	}
	return 0;
}

2978
static sector_t reshape_request(mddev_t *mddev, sector_t sector_nr, int *skipped)
L
Linus Torvalds 已提交
2979
{
2980 2981 2982 2983 2984 2985 2986 2987 2988
	/* reshaping is quite different to recovery/resync so it is
	 * handled quite separately ... here.
	 *
	 * On each call to sync_request, we gather one chunk worth of
	 * destination stripes and flag them as expanding.
	 * Then we find all the source stripes and request reads.
	 * As the reads complete, handle_stripe will copy the data
	 * into the destination stripe and release that stripe.
	 */
L
Linus Torvalds 已提交
2989 2990
	raid5_conf_t *conf = (raid5_conf_t *) mddev->private;
	struct stripe_head *sh;
2991 2992
	int pd_idx;
	sector_t first_sector, last_sector;
2993 2994 2995
	int raid_disks = conf->previous_raid_disks;
	int data_disks = raid_disks - conf->max_degraded;
	int new_data_disks = conf->raid_disks - conf->max_degraded;
2996 2997 2998 2999 3000 3001 3002 3003
	int i;
	int dd_idx;
	sector_t writepos, safepos, gap;

	if (sector_nr == 0 &&
	    conf->expand_progress != 0) {
		/* restarting in the middle, skip the initial sectors */
		sector_nr = conf->expand_progress;
3004
		sector_div(sector_nr, new_data_disks);
3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017
		*skipped = 1;
		return sector_nr;
	}

	/* we update the metadata when there is more than 3Meg
	 * in the block range (that is rather arbitrary, should
	 * probably be time based) or when the data about to be
	 * copied would over-write the source of the data at
	 * the front of the range.
	 * i.e. one new_stripe forward from expand_progress new_maps
	 * to after where expand_lo old_maps to
	 */
	writepos = conf->expand_progress +
3018 3019
		conf->chunk_size/512*(new_data_disks);
	sector_div(writepos, new_data_disks);
3020
	safepos = conf->expand_lo;
3021
	sector_div(safepos, data_disks);
3022 3023 3024
	gap = conf->expand_progress - conf->expand_lo;

	if (writepos >= safepos ||
3025
	    gap > (new_data_disks)*3000*2 /*3Meg*/) {
3026 3027 3028 3029
		/* Cannot proceed until we've updated the superblock... */
		wait_event(conf->wait_for_overlap,
			   atomic_read(&conf->reshape_stripes)==0);
		mddev->reshape_position = conf->expand_progress;
3030
		set_bit(MD_CHANGE_DEVS, &mddev->flags);
3031
		md_wakeup_thread(mddev->thread);
3032
		wait_event(mddev->sb_wait, mddev->flags == 0 ||
3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054
			   kthread_should_stop());
		spin_lock_irq(&conf->device_lock);
		conf->expand_lo = mddev->reshape_position;
		spin_unlock_irq(&conf->device_lock);
		wake_up(&conf->wait_for_overlap);
	}

	for (i=0; i < conf->chunk_size/512; i+= STRIPE_SECTORS) {
		int j;
		int skipped = 0;
		pd_idx = stripe_to_pdidx(sector_nr+i, conf, conf->raid_disks);
		sh = get_active_stripe(conf, sector_nr+i,
				       conf->raid_disks, pd_idx, 0);
		set_bit(STRIPE_EXPANDING, &sh->state);
		atomic_inc(&conf->reshape_stripes);
		/* If any of this stripe is beyond the end of the old
		 * array, then we need to zero those blocks
		 */
		for (j=sh->disks; j--;) {
			sector_t s;
			if (j == sh->pd_idx)
				continue;
3055 3056 3057
			if (conf->level == 6 &&
			    j == raid6_next_disk(sh->pd_idx, sh->disks))
				continue;
3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081
			s = compute_blocknr(sh, j);
			if (s < (mddev->array_size<<1)) {
				skipped = 1;
				continue;
			}
			memset(page_address(sh->dev[j].page), 0, STRIPE_SIZE);
			set_bit(R5_Expanded, &sh->dev[j].flags);
			set_bit(R5_UPTODATE, &sh->dev[j].flags);
		}
		if (!skipped) {
			set_bit(STRIPE_EXPAND_READY, &sh->state);
			set_bit(STRIPE_HANDLE, &sh->state);
		}
		release_stripe(sh);
	}
	spin_lock_irq(&conf->device_lock);
	conf->expand_progress = (sector_nr + i)*(conf->raid_disks-1);
	spin_unlock_irq(&conf->device_lock);
	/* Ok, those stripe are ready. We can start scheduling
	 * reads on the source stripes.
	 * The source stripes are determined by mapping the first and last
	 * block on the destination stripes.
	 */
	first_sector =
3082
		raid5_compute_sector(sector_nr*(new_data_disks),
3083 3084 3085 3086
				     raid_disks, data_disks,
				     &dd_idx, &pd_idx, conf);
	last_sector =
		raid5_compute_sector((sector_nr+conf->chunk_size/512)
3087
				     *(new_data_disks) -1,
3088 3089 3090 3091 3092
				     raid_disks, data_disks,
				     &dd_idx, &pd_idx, conf);
	if (last_sector >= (mddev->size<<1))
		last_sector = (mddev->size<<1)-1;
	while (first_sector <= last_sector) {
3093 3094
		pd_idx = stripe_to_pdidx(first_sector, conf,
					 conf->previous_raid_disks);
3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110
		sh = get_active_stripe(conf, first_sector,
				       conf->previous_raid_disks, pd_idx, 0);
		set_bit(STRIPE_EXPAND_SOURCE, &sh->state);
		set_bit(STRIPE_HANDLE, &sh->state);
		release_stripe(sh);
		first_sector += STRIPE_SECTORS;
	}
	return conf->chunk_size>>9;
}

/* FIXME go_faster isn't used */
static inline sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, int go_faster)
{
	raid5_conf_t *conf = (raid5_conf_t *) mddev->private;
	struct stripe_head *sh;
	int pd_idx;
L
Linus Torvalds 已提交
3111
	int raid_disks = conf->raid_disks;
3112 3113
	sector_t max_sector = mddev->size << 1;
	int sync_blocks;
3114 3115
	int still_degraded = 0;
	int i;
L
Linus Torvalds 已提交
3116

3117
	if (sector_nr >= max_sector) {
L
Linus Torvalds 已提交
3118 3119
		/* just being told to finish up .. nothing much to do */
		unplug_slaves(mddev);
3120 3121 3122 3123
		if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) {
			end_reshape(conf);
			return 0;
		}
3124 3125 3126 3127

		if (mddev->curr_resync < max_sector) /* aborted */
			bitmap_end_sync(mddev->bitmap, mddev->curr_resync,
					&sync_blocks, 1);
3128
		else /* completed sync */
3129 3130 3131
			conf->fullsync = 0;
		bitmap_close_sync(mddev->bitmap);

L
Linus Torvalds 已提交
3132 3133
		return 0;
	}
3134

3135 3136
	if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
		return reshape_request(mddev, sector_nr, skipped);
3137

3138
	/* if there is too many failed drives and we are trying
L
Linus Torvalds 已提交
3139 3140 3141
	 * to resync, then assert that we are finished, because there is
	 * nothing we can do.
	 */
3142
	if (mddev->degraded >= conf->max_degraded &&
3143
	    test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
3144 3145
		sector_t rv = (mddev->size << 1) - sector_nr;
		*skipped = 1;
L
Linus Torvalds 已提交
3146 3147
		return rv;
	}
3148
	if (!bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, 1) &&
3149
	    !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) &&
3150 3151 3152 3153 3154 3155
	    !conf->fullsync && sync_blocks >= STRIPE_SECTORS) {
		/* we can skip this block, and probably more */
		sync_blocks /= STRIPE_SECTORS;
		*skipped = 1;
		return sync_blocks * STRIPE_SECTORS; /* keep things rounded to whole stripes */
	}
L
Linus Torvalds 已提交
3156

3157
	pd_idx = stripe_to_pdidx(sector_nr, conf, raid_disks);
3158
	sh = get_active_stripe(conf, sector_nr, raid_disks, pd_idx, 1);
L
Linus Torvalds 已提交
3159
	if (sh == NULL) {
3160
		sh = get_active_stripe(conf, sector_nr, raid_disks, pd_idx, 0);
L
Linus Torvalds 已提交
3161
		/* make sure we don't swamp the stripe cache if someone else
3162
		 * is trying to get access
L
Linus Torvalds 已提交
3163
		 */
3164
		schedule_timeout_uninterruptible(1);
L
Linus Torvalds 已提交
3165
	}
3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
	/* Need to check if array will still be degraded after recovery/resync
	 * We don't need to check the 'failed' flag as when that gets set,
	 * recovery aborts.
	 */
	for (i=0; i<mddev->raid_disks; i++)
		if (conf->disks[i].rdev == NULL)
			still_degraded = 1;

	bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, still_degraded);

	spin_lock(&sh->lock);
L
Linus Torvalds 已提交
3177 3178 3179 3180
	set_bit(STRIPE_SYNCING, &sh->state);
	clear_bit(STRIPE_INSYNC, &sh->state);
	spin_unlock(&sh->lock);

3181
	handle_stripe(sh, NULL);
L
Linus Torvalds 已提交
3182 3183 3184 3185 3186
	release_stripe(sh);

	return STRIPE_SECTORS;
}

3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215
static int  retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
{
	/* We may not be able to submit a whole bio at once as there
	 * may not be enough stripe_heads available.
	 * We cannot pre-allocate enough stripe_heads as we may need
	 * more than exist in the cache (if we allow ever large chunks).
	 * So we do one stripe head at a time and record in
	 * ->bi_hw_segments how many have been done.
	 *
	 * We *know* that this entire raid_bio is in one chunk, so
	 * it will be only one 'dd_idx' and only need one call to raid5_compute_sector.
	 */
	struct stripe_head *sh;
	int dd_idx, pd_idx;
	sector_t sector, logical_sector, last_sector;
	int scnt = 0;
	int remaining;
	int handled = 0;

	logical_sector = raid_bio->bi_sector & ~((sector_t)STRIPE_SECTORS-1);
	sector = raid5_compute_sector(	logical_sector,
					conf->raid_disks,
					conf->raid_disks - conf->max_degraded,
					&dd_idx,
					&pd_idx,
					conf);
	last_sector = raid_bio->bi_sector + (raid_bio->bi_size>>9);

	for (; logical_sector < last_sector;
3216 3217 3218
	     logical_sector += STRIPE_SECTORS,
		     sector += STRIPE_SECTORS,
		     scnt++) {
3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233

		if (scnt < raid_bio->bi_hw_segments)
			/* already done this stripe */
			continue;

		sh = get_active_stripe(conf, sector, conf->raid_disks, pd_idx, 1);

		if (!sh) {
			/* failed to get a stripe - must wait */
			raid_bio->bi_hw_segments = scnt;
			conf->retry_read_aligned = raid_bio;
			return handled;
		}

		set_bit(R5_ReadError, &sh->dev[dd_idx].flags);
3234 3235 3236 3237 3238 3239 3240
		if (!add_stripe_bio(sh, raid_bio, dd_idx, 0)) {
			release_stripe(sh);
			raid_bio->bi_hw_segments = scnt;
			conf->retry_read_aligned = raid_bio;
			return handled;
		}

3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251
		handle_stripe(sh, NULL);
		release_stripe(sh);
		handled++;
	}
	spin_lock_irq(&conf->device_lock);
	remaining = --raid_bio->bi_phys_segments;
	spin_unlock_irq(&conf->device_lock);
	if (remaining == 0) {
		int bytes = raid_bio->bi_size;

		raid_bio->bi_size = 0;
3252 3253 3254
		raid_bio->bi_end_io(raid_bio, bytes,
			      test_bit(BIO_UPTODATE, &raid_bio->bi_flags)
			        ? 0 : -EIO);
3255 3256 3257 3258 3259 3260 3261 3262
	}
	if (atomic_dec_and_test(&conf->active_aligned_reads))
		wake_up(&conf->wait_for_stripe);
	return handled;
}



L
Linus Torvalds 已提交
3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283
/*
 * This is our raid5 kernel thread.
 *
 * We scan the hash table for stripes which can be handled now.
 * During the scan, completed stripes are saved for us by the interrupt
 * handler, so that they will not have to wait for our next wakeup.
 */
static void raid5d (mddev_t *mddev)
{
	struct stripe_head *sh;
	raid5_conf_t *conf = mddev_to_conf(mddev);
	int handled;

	PRINTK("+++ raid5d active\n");

	md_check_recovery(mddev);

	handled = 0;
	spin_lock_irq(&conf->device_lock);
	while (1) {
		struct list_head *first;
3284
		struct bio *bio;
L
Linus Torvalds 已提交
3285

3286
		if (conf->seq_flush != conf->seq_write) {
3287
			int seq = conf->seq_flush;
3288
			spin_unlock_irq(&conf->device_lock);
3289
			bitmap_unplug(mddev->bitmap);
3290
			spin_lock_irq(&conf->device_lock);
3291 3292 3293 3294
			conf->seq_write = seq;
			activate_bit_delay(conf);
		}

L
Linus Torvalds 已提交
3295 3296 3297 3298 3299 3300
		if (list_empty(&conf->handle_list) &&
		    atomic_read(&conf->preread_active_stripes) < IO_THRESHOLD &&
		    !blk_queue_plugged(mddev->queue) &&
		    !list_empty(&conf->delayed_list))
			raid5_activate_delayed(conf);

3301 3302 3303 3304 3305 3306 3307 3308 3309 3310
		while ((bio = remove_bio_from_retry(conf))) {
			int ok;
			spin_unlock_irq(&conf->device_lock);
			ok = retry_aligned_read(conf, bio);
			spin_lock_irq(&conf->device_lock);
			if (!ok)
				break;
			handled++;
		}

L
Linus Torvalds 已提交
3311 3312 3313 3314 3315 3316 3317 3318
		if (list_empty(&conf->handle_list))
			break;

		first = conf->handle_list.next;
		sh = list_entry(first, struct stripe_head, lru);

		list_del_init(first);
		atomic_inc(&sh->count);
3319
		BUG_ON(atomic_read(&sh->count)!= 1);
L
Linus Torvalds 已提交
3320 3321 3322
		spin_unlock_irq(&conf->device_lock);
		
		handled++;
3323
		handle_stripe(sh, conf->spare_page);
L
Linus Torvalds 已提交
3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336
		release_stripe(sh);

		spin_lock_irq(&conf->device_lock);
	}
	PRINTK("%d stripes handled\n", handled);

	spin_unlock_irq(&conf->device_lock);

	unplug_slaves(mddev);

	PRINTK("--- raid5d inactive\n");
}

3337
static ssize_t
3338
raid5_show_stripe_cache_size(mddev_t *mddev, char *page)
3339
{
3340
	raid5_conf_t *conf = mddev_to_conf(mddev);
3341 3342 3343 3344
	if (conf)
		return sprintf(page, "%d\n", conf->max_nr_stripes);
	else
		return 0;
3345 3346 3347
}

static ssize_t
3348
raid5_store_stripe_cache_size(mddev_t *mddev, const char *page, size_t len)
3349
{
3350
	raid5_conf_t *conf = mddev_to_conf(mddev);
3351 3352 3353 3354
	char *end;
	int new;
	if (len >= PAGE_SIZE)
		return -EINVAL;
3355 3356
	if (!conf)
		return -ENODEV;
3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368

	new = simple_strtoul(page, &end, 10);
	if (!*page || (*end && *end != '\n') )
		return -EINVAL;
	if (new <= 16 || new > 32768)
		return -EINVAL;
	while (new < conf->max_nr_stripes) {
		if (drop_one_stripe(conf))
			conf->max_nr_stripes--;
		else
			break;
	}
3369
	md_allow_write(mddev);
3370 3371 3372 3373 3374 3375 3376
	while (new > conf->max_nr_stripes) {
		if (grow_one_stripe(conf))
			conf->max_nr_stripes++;
		else break;
	}
	return len;
}
3377

3378 3379 3380 3381
static struct md_sysfs_entry
raid5_stripecache_size = __ATTR(stripe_cache_size, S_IRUGO | S_IWUSR,
				raid5_show_stripe_cache_size,
				raid5_store_stripe_cache_size);
3382 3383

static ssize_t
3384
stripe_cache_active_show(mddev_t *mddev, char *page)
3385
{
3386
	raid5_conf_t *conf = mddev_to_conf(mddev);
3387 3388 3389 3390
	if (conf)
		return sprintf(page, "%d\n", atomic_read(&conf->active_stripes));
	else
		return 0;
3391 3392
}

3393 3394
static struct md_sysfs_entry
raid5_stripecache_active = __ATTR_RO(stripe_cache_active);
3395

3396
static struct attribute *raid5_attrs[] =  {
3397 3398 3399 3400
	&raid5_stripecache_size.attr,
	&raid5_stripecache_active.attr,
	NULL,
};
3401 3402 3403
static struct attribute_group raid5_attrs_group = {
	.name = NULL,
	.attrs = raid5_attrs,
3404 3405
};

3406
static int run(mddev_t *mddev)
L
Linus Torvalds 已提交
3407 3408 3409 3410 3411 3412
{
	raid5_conf_t *conf;
	int raid_disk, memory;
	mdk_rdev_t *rdev;
	struct disk_info *disk;
	struct list_head *tmp;
3413
	int working_disks = 0;
L
Linus Torvalds 已提交
3414

3415 3416
	if (mddev->level != 5 && mddev->level != 4 && mddev->level != 6) {
		printk(KERN_ERR "raid5: %s: raid level not set to 4/5/6 (%d)\n",
N
NeilBrown 已提交
3417
		       mdname(mddev), mddev->level);
L
Linus Torvalds 已提交
3418 3419 3420
		return -EIO;
	}

3421 3422 3423 3424 3425 3426 3427 3428
	if (mddev->reshape_position != MaxSector) {
		/* Check that we can continue the reshape.
		 * Currently only disks can change, it must
		 * increase, and we must be past the point where
		 * a stripe over-writes itself
		 */
		sector_t here_new, here_old;
		int old_disks;
3429
		int max_degraded = (mddev->level == 5 ? 1 : 2);
3430 3431 3432 3433

		if (mddev->new_level != mddev->level ||
		    mddev->new_layout != mddev->layout ||
		    mddev->new_chunk != mddev->chunk_size) {
3434 3435
			printk(KERN_ERR "raid5: %s: unsupported reshape "
			       "required - aborting.\n",
3436 3437 3438 3439
			       mdname(mddev));
			return -EINVAL;
		}
		if (mddev->delta_disks <= 0) {
3440 3441
			printk(KERN_ERR "raid5: %s: unsupported reshape "
			       "(reduce disks) required - aborting.\n",
3442 3443 3444 3445 3446
			       mdname(mddev));
			return -EINVAL;
		}
		old_disks = mddev->raid_disks - mddev->delta_disks;
		/* reshape_position must be on a new-stripe boundary, and one
3447 3448
		 * further up in new geometry must map after here in old
		 * geometry.
3449 3450
		 */
		here_new = mddev->reshape_position;
3451 3452 3453 3454
		if (sector_div(here_new, (mddev->chunk_size>>9)*
			       (mddev->raid_disks - max_degraded))) {
			printk(KERN_ERR "raid5: reshape_position not "
			       "on a stripe boundary\n");
3455 3456 3457 3458
			return -EINVAL;
		}
		/* here_new is the stripe we will write to */
		here_old = mddev->reshape_position;
3459 3460 3461 3462
		sector_div(here_old, (mddev->chunk_size>>9)*
			   (old_disks-max_degraded));
		/* here_old is the first stripe that we might need to read
		 * from */
3463 3464
		if (here_new >= here_old) {
			/* Reading from the same stripe as writing to - bad */
3465 3466
			printk(KERN_ERR "raid5: reshape_position too early for "
			       "auto-recovery - aborting.\n");
3467 3468 3469 3470 3471 3472 3473
			return -EINVAL;
		}
		printk(KERN_INFO "raid5: reshape will continue\n");
		/* OK, we should be able to continue; */
	}


3474
	mddev->private = kzalloc(sizeof (raid5_conf_t), GFP_KERNEL);
L
Linus Torvalds 已提交
3475 3476
	if ((conf = mddev->private) == NULL)
		goto abort;
3477 3478 3479 3480 3481 3482 3483 3484
	if (mddev->reshape_position == MaxSector) {
		conf->previous_raid_disks = conf->raid_disks = mddev->raid_disks;
	} else {
		conf->raid_disks = mddev->raid_disks;
		conf->previous_raid_disks = mddev->raid_disks - mddev->delta_disks;
	}

	conf->disks = kzalloc(conf->raid_disks * sizeof(struct disk_info),
3485 3486 3487
			      GFP_KERNEL);
	if (!conf->disks)
		goto abort;
3488

L
Linus Torvalds 已提交
3489 3490
	conf->mddev = mddev;

3491
	if ((conf->stripe_hashtbl = kzalloc(PAGE_SIZE, GFP_KERNEL)) == NULL)
L
Linus Torvalds 已提交
3492 3493
		goto abort;

3494 3495 3496 3497 3498
	if (mddev->level == 6) {
		conf->spare_page = alloc_page(GFP_KERNEL);
		if (!conf->spare_page)
			goto abort;
	}
L
Linus Torvalds 已提交
3499 3500 3501 3502 3503
	spin_lock_init(&conf->device_lock);
	init_waitqueue_head(&conf->wait_for_stripe);
	init_waitqueue_head(&conf->wait_for_overlap);
	INIT_LIST_HEAD(&conf->handle_list);
	INIT_LIST_HEAD(&conf->delayed_list);
3504
	INIT_LIST_HEAD(&conf->bitmap_list);
L
Linus Torvalds 已提交
3505 3506 3507
	INIT_LIST_HEAD(&conf->inactive_list);
	atomic_set(&conf->active_stripes, 0);
	atomic_set(&conf->preread_active_stripes, 0);
3508
	atomic_set(&conf->active_aligned_reads, 0);
L
Linus Torvalds 已提交
3509 3510 3511 3512 3513

	PRINTK("raid5: run(%s) called.\n", mdname(mddev));

	ITERATE_RDEV(mddev,rdev,tmp) {
		raid_disk = rdev->raid_disk;
3514
		if (raid_disk >= conf->raid_disks
L
Linus Torvalds 已提交
3515 3516 3517 3518 3519 3520
		    || raid_disk < 0)
			continue;
		disk = conf->disks + raid_disk;

		disk->rdev = rdev;

3521
		if (test_bit(In_sync, &rdev->flags)) {
L
Linus Torvalds 已提交
3522 3523 3524 3525
			char b[BDEVNAME_SIZE];
			printk(KERN_INFO "raid5: device %s operational as raid"
				" disk %d\n", bdevname(rdev->bdev,b),
				raid_disk);
3526
			working_disks++;
L
Linus Torvalds 已提交
3527 3528 3529 3530
		}
	}

	/*
3531
	 * 0 for a fully functional array, 1 or 2 for a degraded array.
L
Linus Torvalds 已提交
3532
	 */
3533
	mddev->degraded = conf->raid_disks - working_disks;
L
Linus Torvalds 已提交
3534 3535 3536
	conf->mddev = mddev;
	conf->chunk_size = mddev->chunk_size;
	conf->level = mddev->level;
3537 3538 3539 3540
	if (conf->level == 6)
		conf->max_degraded = 2;
	else
		conf->max_degraded = 1;
L
Linus Torvalds 已提交
3541 3542
	conf->algorithm = mddev->layout;
	conf->max_nr_stripes = NR_STRIPES;
3543
	conf->expand_progress = mddev->reshape_position;
L
Linus Torvalds 已提交
3544 3545 3546

	/* device size must be a multiple of chunk size */
	mddev->size &= ~(mddev->chunk_size/1024 -1);
3547
	mddev->resync_max_sectors = mddev->size << 1;
L
Linus Torvalds 已提交
3548

3549 3550 3551 3552 3553
	if (conf->level == 6 && conf->raid_disks < 4) {
		printk(KERN_ERR "raid6: not enough configured devices for %s (%d, minimum 4)\n",
		       mdname(mddev), conf->raid_disks);
		goto abort;
	}
L
Linus Torvalds 已提交
3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564
	if (!conf->chunk_size || conf->chunk_size % 4) {
		printk(KERN_ERR "raid5: invalid chunk size %d for %s\n",
			conf->chunk_size, mdname(mddev));
		goto abort;
	}
	if (conf->algorithm > ALGORITHM_RIGHT_SYMMETRIC) {
		printk(KERN_ERR 
			"raid5: unsupported parity algorithm %d for %s\n",
			conf->algorithm, mdname(mddev));
		goto abort;
	}
3565
	if (mddev->degraded > conf->max_degraded) {
L
Linus Torvalds 已提交
3566 3567
		printk(KERN_ERR "raid5: not enough operational devices for %s"
			" (%d/%d failed)\n",
3568
			mdname(mddev), mddev->degraded, conf->raid_disks);
L
Linus Torvalds 已提交
3569 3570 3571
		goto abort;
	}

3572
	if (mddev->degraded > 0 &&
L
Linus Torvalds 已提交
3573
	    mddev->recovery_cp != MaxSector) {
3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584
		if (mddev->ok_start_degraded)
			printk(KERN_WARNING
			       "raid5: starting dirty degraded array: %s"
			       "- data corruption possible.\n",
			       mdname(mddev));
		else {
			printk(KERN_ERR
			       "raid5: cannot start dirty degraded array for %s\n",
			       mdname(mddev));
			goto abort;
		}
L
Linus Torvalds 已提交
3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595
	}

	{
		mddev->thread = md_register_thread(raid5d, mddev, "%s_raid5");
		if (!mddev->thread) {
			printk(KERN_ERR 
				"raid5: couldn't allocate thread for %s\n",
				mdname(mddev));
			goto abort;
		}
	}
3596
	memory = conf->max_nr_stripes * (sizeof(struct stripe_head) +
L
Linus Torvalds 已提交
3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620
		 conf->raid_disks * ((sizeof(struct bio) + PAGE_SIZE))) / 1024;
	if (grow_stripes(conf, conf->max_nr_stripes)) {
		printk(KERN_ERR 
			"raid5: couldn't allocate %dkB for buffers\n", memory);
		shrink_stripes(conf);
		md_unregister_thread(mddev->thread);
		goto abort;
	} else
		printk(KERN_INFO "raid5: allocated %dkB for %s\n",
			memory, mdname(mddev));

	if (mddev->degraded == 0)
		printk("raid5: raid level %d set %s active with %d out of %d"
			" devices, algorithm %d\n", conf->level, mdname(mddev), 
			mddev->raid_disks-mddev->degraded, mddev->raid_disks,
			conf->algorithm);
	else
		printk(KERN_ALERT "raid5: raid level %d set %s active with %d"
			" out of %d devices, algorithm %d\n", conf->level,
			mdname(mddev), mddev->raid_disks - mddev->degraded,
			mddev->raid_disks, conf->algorithm);

	print_raid5_conf(conf);

3621 3622
	if (conf->expand_progress != MaxSector) {
		printk("...ok start reshape thread\n");
3623
		conf->expand_lo = conf->expand_progress;
3624 3625 3626 3627 3628 3629 3630 3631 3632
		atomic_set(&conf->reshape_stripes, 0);
		clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
		clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
		set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
		set_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
		mddev->sync_thread = md_register_thread(md_do_sync, mddev,
							"%s_reshape");
	}

L
Linus Torvalds 已提交
3633
	/* read-ahead size must cover two whole stripes, which is
3634
	 * 2 * (datadisks) * chunksize where 'n' is the number of raid devices
L
Linus Torvalds 已提交
3635 3636
	 */
	{
3637 3638
		int data_disks = conf->previous_raid_disks - conf->max_degraded;
		int stripe = data_disks *
3639
			(mddev->chunk_size / PAGE_SIZE);
L
Linus Torvalds 已提交
3640 3641 3642 3643 3644
		if (mddev->queue->backing_dev_info.ra_pages < 2 * stripe)
			mddev->queue->backing_dev_info.ra_pages = 2 * stripe;
	}

	/* Ok, everything is just fine now */
3645
	sysfs_create_group(&mddev->kobj, &raid5_attrs_group);
3646 3647 3648

	mddev->queue->unplug_fn = raid5_unplug_device;
	mddev->queue->issue_flush_fn = raid5_issue_flush;
3649 3650 3651
	mddev->queue->backing_dev_info.congested_fn = raid5_congested;
	mddev->queue->backing_dev_info.congested_data = mddev;

3652 3653
	mddev->array_size =  mddev->size * (conf->previous_raid_disks -
					    conf->max_degraded);
3654

3655 3656
	blk_queue_merge_bvec(mddev->queue, raid5_mergeable_bvec);

L
Linus Torvalds 已提交
3657 3658 3659 3660
	return 0;
abort:
	if (conf) {
		print_raid5_conf(conf);
3661
		safe_put_page(conf->spare_page);
3662
		kfree(conf->disks);
3663
		kfree(conf->stripe_hashtbl);
L
Linus Torvalds 已提交
3664 3665 3666 3667 3668 3669 3670 3671 3672
		kfree(conf);
	}
	mddev->private = NULL;
	printk(KERN_ALERT "raid5: failed to run raid set %s\n", mdname(mddev));
	return -EIO;
}



3673
static int stop(mddev_t *mddev)
L
Linus Torvalds 已提交
3674 3675 3676 3677 3678 3679
{
	raid5_conf_t *conf = (raid5_conf_t *) mddev->private;

	md_unregister_thread(mddev->thread);
	mddev->thread = NULL;
	shrink_stripes(conf);
3680
	kfree(conf->stripe_hashtbl);
L
Linus Torvalds 已提交
3681
	blk_sync_queue(mddev->queue); /* the unplug fn references 'conf'*/
3682
	sysfs_remove_group(&mddev->kobj, &raid5_attrs_group);
3683
	kfree(conf->disks);
3684
	kfree(conf);
L
Linus Torvalds 已提交
3685 3686 3687 3688 3689
	mddev->private = NULL;
	return 0;
}

#if RAID5_DEBUG
3690
static void print_sh (struct seq_file *seq, struct stripe_head *sh)
L
Linus Torvalds 已提交
3691 3692 3693
{
	int i;

3694 3695 3696 3697 3698
	seq_printf(seq, "sh %llu, pd_idx %d, state %ld.\n",
		   (unsigned long long)sh->sector, sh->pd_idx, sh->state);
	seq_printf(seq, "sh %llu,  count %d.\n",
		   (unsigned long long)sh->sector, atomic_read(&sh->count));
	seq_printf(seq, "sh %llu, ", (unsigned long long)sh->sector);
3699
	for (i = 0; i < sh->disks; i++) {
3700 3701
		seq_printf(seq, "(cache%d: %p %ld) ",
			   i, sh->dev[i].page, sh->dev[i].flags);
L
Linus Torvalds 已提交
3702
	}
3703
	seq_printf(seq, "\n");
L
Linus Torvalds 已提交
3704 3705
}

3706
static void printall (struct seq_file *seq, raid5_conf_t *conf)
L
Linus Torvalds 已提交
3707 3708
{
	struct stripe_head *sh;
3709
	struct hlist_node *hn;
L
Linus Torvalds 已提交
3710 3711 3712 3713
	int i;

	spin_lock_irq(&conf->device_lock);
	for (i = 0; i < NR_HASH; i++) {
3714
		hlist_for_each_entry(sh, hn, &conf->stripe_hashtbl[i], hash) {
L
Linus Torvalds 已提交
3715 3716
			if (sh->raid_conf != conf)
				continue;
3717
			print_sh(seq, sh);
L
Linus Torvalds 已提交
3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
		}
	}
	spin_unlock_irq(&conf->device_lock);
}
#endif

static void status (struct seq_file *seq, mddev_t *mddev)
{
	raid5_conf_t *conf = (raid5_conf_t *) mddev->private;
	int i;

	seq_printf (seq, " level %d, %dk chunk, algorithm %d", mddev->level, mddev->chunk_size >> 10, mddev->layout);
3730
	seq_printf (seq, " [%d/%d] [", conf->raid_disks, conf->raid_disks - mddev->degraded);
L
Linus Torvalds 已提交
3731 3732 3733
	for (i = 0; i < conf->raid_disks; i++)
		seq_printf (seq, "%s",
			       conf->disks[i].rdev &&
3734
			       test_bit(In_sync, &conf->disks[i].rdev->flags) ? "U" : "_");
L
Linus Torvalds 已提交
3735 3736
	seq_printf (seq, "]");
#if RAID5_DEBUG
3737 3738
	seq_printf (seq, "\n");
	printall(seq, conf);
L
Linus Torvalds 已提交
3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751
#endif
}

static void print_raid5_conf (raid5_conf_t *conf)
{
	int i;
	struct disk_info *tmp;

	printk("RAID5 conf printout:\n");
	if (!conf) {
		printk("(conf==NULL)\n");
		return;
	}
3752 3753
	printk(" --- rd:%d wd:%d\n", conf->raid_disks,
		 conf->raid_disks - conf->mddev->degraded);
L
Linus Torvalds 已提交
3754 3755 3756 3757 3758 3759

	for (i = 0; i < conf->raid_disks; i++) {
		char b[BDEVNAME_SIZE];
		tmp = conf->disks + i;
		if (tmp->rdev)
		printk(" disk %d, o:%d, dev:%s\n",
3760
			i, !test_bit(Faulty, &tmp->rdev->flags),
L
Linus Torvalds 已提交
3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773
			bdevname(tmp->rdev->bdev,b));
	}
}

static int raid5_spare_active(mddev_t *mddev)
{
	int i;
	raid5_conf_t *conf = mddev->private;
	struct disk_info *tmp;

	for (i = 0; i < conf->raid_disks; i++) {
		tmp = conf->disks + i;
		if (tmp->rdev
3774
		    && !test_bit(Faulty, &tmp->rdev->flags)
3775 3776 3777
		    && !test_and_set_bit(In_sync, &tmp->rdev->flags)) {
			unsigned long flags;
			spin_lock_irqsave(&conf->device_lock, flags);
L
Linus Torvalds 已提交
3778
			mddev->degraded--;
3779
			spin_unlock_irqrestore(&conf->device_lock, flags);
L
Linus Torvalds 已提交
3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795
		}
	}
	print_raid5_conf(conf);
	return 0;
}

static int raid5_remove_disk(mddev_t *mddev, int number)
{
	raid5_conf_t *conf = mddev->private;
	int err = 0;
	mdk_rdev_t *rdev;
	struct disk_info *p = conf->disks + number;

	print_raid5_conf(conf);
	rdev = p->rdev;
	if (rdev) {
3796
		if (test_bit(In_sync, &rdev->flags) ||
L
Linus Torvalds 已提交
3797 3798 3799 3800 3801
		    atomic_read(&rdev->nr_pending)) {
			err = -EBUSY;
			goto abort;
		}
		p->rdev = NULL;
3802
		synchronize_rcu();
L
Linus Torvalds 已提交
3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821
		if (atomic_read(&rdev->nr_pending)) {
			/* lost the race, try later */
			err = -EBUSY;
			p->rdev = rdev;
		}
	}
abort:

	print_raid5_conf(conf);
	return err;
}

static int raid5_add_disk(mddev_t *mddev, mdk_rdev_t *rdev)
{
	raid5_conf_t *conf = mddev->private;
	int found = 0;
	int disk;
	struct disk_info *p;

3822
	if (mddev->degraded > conf->max_degraded)
L
Linus Torvalds 已提交
3823 3824 3825 3826
		/* no point adding a device */
		return 0;

	/*
3827 3828
	 * find the disk ... but prefer rdev->saved_raid_disk
	 * if possible.
L
Linus Torvalds 已提交
3829
	 */
3830 3831 3832 3833 3834 3835
	if (rdev->saved_raid_disk >= 0 &&
	    conf->disks[rdev->saved_raid_disk].rdev == NULL)
		disk = rdev->saved_raid_disk;
	else
		disk = 0;
	for ( ; disk < conf->raid_disks; disk++)
L
Linus Torvalds 已提交
3836
		if ((p=conf->disks + disk)->rdev == NULL) {
3837
			clear_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
3838 3839
			rdev->raid_disk = disk;
			found = 1;
3840 3841
			if (rdev->saved_raid_disk != disk)
				conf->fullsync = 1;
3842
			rcu_assign_pointer(p->rdev, rdev);
L
Linus Torvalds 已提交
3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857
			break;
		}
	print_raid5_conf(conf);
	return found;
}

static int raid5_resize(mddev_t *mddev, sector_t sectors)
{
	/* 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.
	 */
3858 3859
	raid5_conf_t *conf = mddev_to_conf(mddev);

L
Linus Torvalds 已提交
3860
	sectors &= ~((sector_t)mddev->chunk_size/512 - 1);
3861
	mddev->array_size = (sectors * (mddev->raid_disks-conf->max_degraded))>>1;
L
Linus Torvalds 已提交
3862 3863 3864 3865 3866 3867 3868
	set_capacity(mddev->gendisk, mddev->array_size << 1);
	mddev->changed = 1;
	if (sectors/2  > mddev->size && mddev->recovery_cp == MaxSector) {
		mddev->recovery_cp = mddev->size << 1;
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	}
	mddev->size = sectors /2;
3869
	mddev->resync_max_sectors = sectors;
L
Linus Torvalds 已提交
3870 3871 3872
	return 0;
}

3873
#ifdef CONFIG_MD_RAID5_RESHAPE
3874
static int raid5_check_reshape(mddev_t *mddev)
3875 3876 3877 3878
{
	raid5_conf_t *conf = mddev_to_conf(mddev);
	int err;

3879 3880 3881 3882
	if (mddev->delta_disks < 0 ||
	    mddev->new_level != mddev->level)
		return -EINVAL; /* Cannot shrink array or change level yet */
	if (mddev->delta_disks == 0)
3883 3884 3885 3886 3887 3888 3889 3890 3891 3892
		return 0; /* nothing to do */

	/* Can only proceed if there are plenty of stripe_heads.
	 * We need a minimum of one full stripe,, and for sensible progress
	 * it is best to have about 4 times that.
	 * If we require 4 times, then the default 256 4K stripe_heads will
	 * allow for chunk sizes up to 256K, which is probably OK.
	 * If the chunk size is greater, user-space should request more
	 * stripe_heads first.
	 */
3893 3894
	if ((mddev->chunk_size / STRIPE_SIZE) * 4 > conf->max_nr_stripes ||
	    (mddev->new_chunk / STRIPE_SIZE) * 4 > conf->max_nr_stripes) {
3895 3896 3897 3898 3899
		printk(KERN_WARNING "raid5: reshape: not enough stripes.  Needed %lu\n",
		       (mddev->chunk_size / STRIPE_SIZE)*4);
		return -ENOSPC;
	}

3900 3901 3902 3903
	err = resize_stripes(conf, conf->raid_disks + mddev->delta_disks);
	if (err)
		return err;

3904 3905
	if (mddev->degraded > conf->max_degraded)
		return -EINVAL;
3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916
	/* looks like we might be able to manage this */
	return 0;
}

static int raid5_start_reshape(mddev_t *mddev)
{
	raid5_conf_t *conf = mddev_to_conf(mddev);
	mdk_rdev_t *rdev;
	struct list_head *rtmp;
	int spares = 0;
	int added_devices = 0;
3917
	unsigned long flags;
3918

3919
	if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
3920 3921
		return -EBUSY;

3922 3923 3924 3925
	ITERATE_RDEV(mddev, rdev, rtmp)
		if (rdev->raid_disk < 0 &&
		    !test_bit(Faulty, &rdev->flags))
			spares++;
3926

3927
	if (spares - mddev->degraded < mddev->delta_disks - conf->max_degraded)
3928 3929 3930 3931 3932
		/* Not enough devices even to make a degraded array
		 * of that size
		 */
		return -EINVAL;

3933
	atomic_set(&conf->reshape_stripes, 0);
3934 3935
	spin_lock_irq(&conf->device_lock);
	conf->previous_raid_disks = conf->raid_disks;
3936
	conf->raid_disks += mddev->delta_disks;
3937
	conf->expand_progress = 0;
3938
	conf->expand_lo = 0;
3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950
	spin_unlock_irq(&conf->device_lock);

	/* Add some new drives, as many as will fit.
	 * We know there are enough to make the newly sized array work.
	 */
	ITERATE_RDEV(mddev, rdev, rtmp)
		if (rdev->raid_disk < 0 &&
		    !test_bit(Faulty, &rdev->flags)) {
			if (raid5_add_disk(mddev, rdev)) {
				char nm[20];
				set_bit(In_sync, &rdev->flags);
				added_devices++;
3951
				rdev->recovery_offset = 0;
3952 3953 3954 3955 3956 3957
				sprintf(nm, "rd%d", rdev->raid_disk);
				sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
			} else
				break;
		}

3958
	spin_lock_irqsave(&conf->device_lock, flags);
3959
	mddev->degraded = (conf->raid_disks - conf->previous_raid_disks) - added_devices;
3960
	spin_unlock_irqrestore(&conf->device_lock, flags);
3961
	mddev->raid_disks = conf->raid_disks;
3962
	mddev->reshape_position = 0;
3963
	set_bit(MD_CHANGE_DEVS, &mddev->flags);
3964

3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988
	clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
	clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
	set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
	set_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
	mddev->sync_thread = md_register_thread(md_do_sync, mddev,
						"%s_reshape");
	if (!mddev->sync_thread) {
		mddev->recovery = 0;
		spin_lock_irq(&conf->device_lock);
		mddev->raid_disks = conf->raid_disks = conf->previous_raid_disks;
		conf->expand_progress = MaxSector;
		spin_unlock_irq(&conf->device_lock);
		return -EAGAIN;
	}
	md_wakeup_thread(mddev->sync_thread);
	md_new_event(mddev);
	return 0;
}
#endif

static void end_reshape(raid5_conf_t *conf)
{
	struct block_device *bdev;

3989
	if (!test_bit(MD_RECOVERY_INTR, &conf->mddev->recovery)) {
3990 3991
		conf->mddev->array_size = conf->mddev->size *
			(conf->raid_disks - conf->max_degraded);
3992 3993 3994 3995 3996 3997
		set_capacity(conf->mddev->gendisk, conf->mddev->array_size << 1);
		conf->mddev->changed = 1;

		bdev = bdget_disk(conf->mddev->gendisk, 0);
		if (bdev) {
			mutex_lock(&bdev->bd_inode->i_mutex);
3998
			i_size_write(bdev->bd_inode, (loff_t)conf->mddev->array_size << 10);
3999 4000 4001 4002 4003 4004 4005
			mutex_unlock(&bdev->bd_inode->i_mutex);
			bdput(bdev);
		}
		spin_lock_irq(&conf->device_lock);
		conf->expand_progress = MaxSector;
		spin_unlock_irq(&conf->device_lock);
		conf->mddev->reshape_position = MaxSector;
4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016

		/* read-ahead size must cover two whole stripes, which is
		 * 2 * (datadisks) * chunksize where 'n' is the number of raid devices
		 */
		{
			int data_disks = conf->previous_raid_disks - conf->max_degraded;
			int stripe = data_disks *
				(conf->mddev->chunk_size / PAGE_SIZE);
			if (conf->mddev->queue->backing_dev_info.ra_pages < 2 * stripe)
				conf->mddev->queue->backing_dev_info.ra_pages = 2 * stripe;
		}
4017 4018 4019
	}
}

4020 4021 4022 4023 4024
static void raid5_quiesce(mddev_t *mddev, int state)
{
	raid5_conf_t *conf = mddev_to_conf(mddev);

	switch(state) {
4025 4026 4027 4028
	case 2: /* resume for a suspend */
		wake_up(&conf->wait_for_overlap);
		break;

4029 4030 4031 4032
	case 1: /* stop all writes */
		spin_lock_irq(&conf->device_lock);
		conf->quiesce = 1;
		wait_event_lock_irq(conf->wait_for_stripe,
4033 4034
				    atomic_read(&conf->active_stripes) == 0 &&
				    atomic_read(&conf->active_aligned_reads) == 0,
4035 4036 4037 4038 4039 4040 4041 4042
				    conf->device_lock, /* nothing */);
		spin_unlock_irq(&conf->device_lock);
		break;

	case 0: /* re-enable writes */
		spin_lock_irq(&conf->device_lock);
		conf->quiesce = 0;
		wake_up(&conf->wait_for_stripe);
4043
		wake_up(&conf->wait_for_overlap);
4044 4045 4046 4047
		spin_unlock_irq(&conf->device_lock);
		break;
	}
}
4048

4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063
static struct mdk_personality raid6_personality =
{
	.name		= "raid6",
	.level		= 6,
	.owner		= THIS_MODULE,
	.make_request	= make_request,
	.run		= run,
	.stop		= stop,
	.status		= status,
	.error_handler	= error,
	.hot_add_disk	= raid5_add_disk,
	.hot_remove_disk= raid5_remove_disk,
	.spare_active	= raid5_spare_active,
	.sync_request	= sync_request,
	.resize		= raid5_resize,
4064 4065 4066 4067
#ifdef CONFIG_MD_RAID5_RESHAPE
	.check_reshape	= raid5_check_reshape,
	.start_reshape  = raid5_start_reshape,
#endif
4068 4069
	.quiesce	= raid5_quiesce,
};
4070
static struct mdk_personality raid5_personality =
L
Linus Torvalds 已提交
4071 4072
{
	.name		= "raid5",
4073
	.level		= 5,
L
Linus Torvalds 已提交
4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084
	.owner		= THIS_MODULE,
	.make_request	= make_request,
	.run		= run,
	.stop		= stop,
	.status		= status,
	.error_handler	= error,
	.hot_add_disk	= raid5_add_disk,
	.hot_remove_disk= raid5_remove_disk,
	.spare_active	= raid5_spare_active,
	.sync_request	= sync_request,
	.resize		= raid5_resize,
4085
#ifdef CONFIG_MD_RAID5_RESHAPE
4086 4087
	.check_reshape	= raid5_check_reshape,
	.start_reshape  = raid5_start_reshape,
4088
#endif
4089
	.quiesce	= raid5_quiesce,
L
Linus Torvalds 已提交
4090 4091
};

4092
static struct mdk_personality raid4_personality =
L
Linus Torvalds 已提交
4093
{
4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111
	.name		= "raid4",
	.level		= 4,
	.owner		= THIS_MODULE,
	.make_request	= make_request,
	.run		= run,
	.stop		= stop,
	.status		= status,
	.error_handler	= error,
	.hot_add_disk	= raid5_add_disk,
	.hot_remove_disk= raid5_remove_disk,
	.spare_active	= raid5_spare_active,
	.sync_request	= sync_request,
	.resize		= raid5_resize,
	.quiesce	= raid5_quiesce,
};

static int __init raid5_init(void)
{
4112 4113 4114 4115 4116 4117
	int e;

	e = raid6_select_algo();
	if ( e )
		return e;
	register_md_personality(&raid6_personality);
4118 4119 4120
	register_md_personality(&raid5_personality);
	register_md_personality(&raid4_personality);
	return 0;
L
Linus Torvalds 已提交
4121 4122
}

4123
static void raid5_exit(void)
L
Linus Torvalds 已提交
4124
{
4125
	unregister_md_personality(&raid6_personality);
4126 4127
	unregister_md_personality(&raid5_personality);
	unregister_md_personality(&raid4_personality);
L
Linus Torvalds 已提交
4128 4129 4130 4131 4132 4133
}

module_init(raid5_init);
module_exit(raid5_exit);
MODULE_LICENSE("GPL");
MODULE_ALIAS("md-personality-4"); /* RAID5 */
4134 4135
MODULE_ALIAS("md-raid5");
MODULE_ALIAS("md-raid4");
4136 4137
MODULE_ALIAS("md-level-5");
MODULE_ALIAS("md-level-4");
4138 4139 4140 4141 4142 4143 4144
MODULE_ALIAS("md-personality-8"); /* RAID6 */
MODULE_ALIAS("md-raid6");
MODULE_ALIAS("md-level-6");

/* This used to be two separate modules, they were: */
MODULE_ALIAS("raid5");
MODULE_ALIAS("raid6");