md.c 210.6 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
/*
   md.c : Multiple Devices driver for Linux
	  Copyright (C) 1998, 1999, 2000 Ingo Molnar

     completely rewritten, based on the MD driver code from Marc Zyngier

   Changes:

   - RAID-1/RAID-5 extensions by Miguel de Icaza, Gadi Oxman, Ingo Molnar
   - RAID-6 extensions by H. Peter Anvin <hpa@zytor.com>
   - boot support for linear and striped mode by Harald Hoyer <HarryH@Royal.Net>
   - kerneld support by Boris Tobotras <boris@xtalk.msk.su>
   - kmod support by: Cyrus Durgin
   - RAID0 bugfixes: Mark Anthony Lisher <markal@iname.com>
   - Devfs support by Richard Gooch <rgooch@atnf.csiro.au>

   - lots of fixes and improvements to the RAID1/RAID5 and generic
     RAID code (such as request based resynchronization):

     Neil Brown <neilb@cse.unsw.edu.au>.

22 23 24
   - persistent bitmap code
     Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.

L
Linus Torvalds 已提交
25 26 27 28 29 30 31 32 33 34
   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.
*/

35
#include <linux/kthread.h>
36
#include <linux/blkdev.h>
L
Linus Torvalds 已提交
37
#include <linux/sysctl.h>
38
#include <linux/seq_file.h>
39
#include <linux/mutex.h>
L
Linus Torvalds 已提交
40
#include <linux/buffer_head.h> /* for invalidate_bdev */
41
#include <linux/poll.h>
42
#include <linux/ctype.h>
43
#include <linux/string.h>
44 45 46 47
#include <linux/hdreg.h>
#include <linux/proc_fs.h>
#include <linux/random.h>
#include <linux/reboot.h>
48
#include <linux/file.h>
49
#include <linux/compat.h>
50
#include <linux/delay.h>
51 52
#include <linux/raid/md_p.h>
#include <linux/raid/md_u.h>
53
#include <linux/slab.h>
54
#include "md.h"
55
#include "bitmap.h"
L
Linus Torvalds 已提交
56 57 58 59 60

#define DEBUG 0
#define dprintk(x...) ((void)(DEBUG && printk(x)))

#ifndef MODULE
61
static void autostart_arrays(int part);
L
Linus Torvalds 已提交
62 63
#endif

64
static LIST_HEAD(pers_list);
L
Linus Torvalds 已提交
65 66
static DEFINE_SPINLOCK(pers_lock);

67 68
static void md_print_devices(void);

69
static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
70 71
static struct workqueue_struct *md_wq;
static struct workqueue_struct *md_misc_wq;
72

73 74
#define MD_BUG(x...) { printk("md: bug in file %s, line %d\n", __FILE__, __LINE__); md_print_devices(); }

75 76 77 78 79 80
/*
 * Default number of read corrections we'll attempt on an rdev
 * before ejecting it from the array. We divide the read error
 * count by 2 for every hour elapsed between read errors.
 */
#define MD_DEFAULT_MAX_CORRECTED_READ_ERRORS 20
L
Linus Torvalds 已提交
81 82 83 84
/*
 * Current RAID-1,4,5 parallel reconstruction 'guaranteed speed limit'
 * is 1000 KB/sec, so the extra system load does not show up that much.
 * Increase it if you want to have more _guaranteed_ speed. Note that
85
 * the RAID driver will use the maximum available bandwidth if the IO
L
Linus Torvalds 已提交
86 87 88 89 90
 * subsystem is idle. There is also an 'absolute maximum' reconstruction
 * speed limit - in case reconstruction slows down your system despite
 * idle IO detection.
 *
 * you can change it via /proc/sys/dev/raid/speed_limit_min and _max.
91
 * or /sys/block/mdX/md/sync_speed_{min,max}
L
Linus Torvalds 已提交
92 93 94 95
 */

static int sysctl_speed_limit_min = 1000;
static int sysctl_speed_limit_max = 200000;
96 97 98 99 100 101 102 103 104 105 106
static inline int speed_min(mddev_t *mddev)
{
	return mddev->sync_speed_min ?
		mddev->sync_speed_min : sysctl_speed_limit_min;
}

static inline int speed_max(mddev_t *mddev)
{
	return mddev->sync_speed_max ?
		mddev->sync_speed_max : sysctl_speed_limit_max;
}
L
Linus Torvalds 已提交
107 108 109 110 111 112 113 114

static struct ctl_table_header *raid_table_header;

static ctl_table raid_table[] = {
	{
		.procname	= "speed_limit_min",
		.data		= &sysctl_speed_limit_min,
		.maxlen		= sizeof(int),
115
		.mode		= S_IRUGO|S_IWUSR,
116
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
117 118 119 120 121
	},
	{
		.procname	= "speed_limit_max",
		.data		= &sysctl_speed_limit_max,
		.maxlen		= sizeof(int),
122
		.mode		= S_IRUGO|S_IWUSR,
123
		.proc_handler	= proc_dointvec,
L
Linus Torvalds 已提交
124
	},
125
	{ }
L
Linus Torvalds 已提交
126 127 128 129 130 131
};

static ctl_table raid_dir_table[] = {
	{
		.procname	= "raid",
		.maxlen		= 0,
132
		.mode		= S_IRUGO|S_IXUGO,
L
Linus Torvalds 已提交
133 134
		.child		= raid_table,
	},
135
	{ }
L
Linus Torvalds 已提交
136 137 138 139 140 141 142 143 144
};

static ctl_table raid_root_table[] = {
	{
		.procname	= "dev",
		.maxlen		= 0,
		.mode		= 0555,
		.child		= raid_dir_table,
	},
145
	{  }
L
Linus Torvalds 已提交
146 147
};

148
static const struct block_device_operations md_fops;
L
Linus Torvalds 已提交
149

150 151
static int start_readonly;

152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
/* bio_clone_mddev
 * like bio_clone, but with a local bio set
 */

static void mddev_bio_destructor(struct bio *bio)
{
	mddev_t *mddev, **mddevp;

	mddevp = (void*)bio;
	mddev = mddevp[-1];

	bio_free(bio, mddev->bio_set);
}

struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
			    mddev_t *mddev)
{
	struct bio *b;
	mddev_t **mddevp;

	if (!mddev || !mddev->bio_set)
		return bio_alloc(gfp_mask, nr_iovecs);

	b = bio_alloc_bioset(gfp_mask, nr_iovecs,
			     mddev->bio_set);
	if (!b)
		return NULL;
	mddevp = (void*)b;
	mddevp[-1] = mddev;
	b->bi_destructor = mddev_bio_destructor;
	return b;
}
EXPORT_SYMBOL_GPL(bio_alloc_mddev);

struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
			    mddev_t *mddev)
{
	struct bio *b;
	mddev_t **mddevp;

	if (!mddev || !mddev->bio_set)
		return bio_clone(bio, gfp_mask);

	b = bio_alloc_bioset(gfp_mask, bio->bi_max_vecs,
			     mddev->bio_set);
	if (!b)
		return NULL;
	mddevp = (void*)b;
	mddevp[-1] = mddev;
	b->bi_destructor = mddev_bio_destructor;
	__bio_clone(b, bio);
	if (bio_integrity(bio)) {
		int ret;

		ret = bio_integrity_clone(b, bio, gfp_mask, mddev->bio_set);

		if (ret < 0) {
			bio_put(b);
			return NULL;
		}
	}

	return b;
}
EXPORT_SYMBOL_GPL(bio_clone_mddev);

218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
void md_trim_bio(struct bio *bio, int offset, int size)
{
	/* 'bio' is a cloned bio which we need to trim to match
	 * the given offset and size.
	 * This requires adjusting bi_sector, bi_size, and bi_io_vec
	 */
	int i;
	struct bio_vec *bvec;
	int sofar = 0;

	size <<= 9;
	if (offset == 0 && size == bio->bi_size)
		return;

	bio->bi_sector += offset;
	bio->bi_size = size;
	offset <<= 9;
	clear_bit(BIO_SEG_VALID, &bio->bi_flags);

	while (bio->bi_idx < bio->bi_vcnt &&
	       bio->bi_io_vec[bio->bi_idx].bv_len <= offset) {
		/* remove this whole bio_vec */
		offset -= bio->bi_io_vec[bio->bi_idx].bv_len;
		bio->bi_idx++;
	}
	if (bio->bi_idx < bio->bi_vcnt) {
		bio->bi_io_vec[bio->bi_idx].bv_offset += offset;
		bio->bi_io_vec[bio->bi_idx].bv_len -= offset;
	}
	/* avoid any complications with bi_idx being non-zero*/
	if (bio->bi_idx) {
		memmove(bio->bi_io_vec, bio->bi_io_vec+bio->bi_idx,
			(bio->bi_vcnt - bio->bi_idx) * sizeof(struct bio_vec));
		bio->bi_vcnt -= bio->bi_idx;
		bio->bi_idx = 0;
	}
	/* Make sure vcnt and last bv are not too big */
	bio_for_each_segment(bvec, bio, i) {
		if (sofar + bvec->bv_len > size)
			bvec->bv_len = size - sofar;
		if (bvec->bv_len == 0) {
			bio->bi_vcnt = i;
			break;
		}
		sofar += bvec->bv_len;
	}
}
EXPORT_SYMBOL_GPL(md_trim_bio);

267 268 269 270 271 272 273 274 275 276
/*
 * We have a system wide 'event count' that is incremented
 * on any 'interesting' event, and readers of /proc/mdstat
 * can use 'poll' or 'select' to find out when the event
 * count increases.
 *
 * Events are:
 *  start array, stop array, error, add device, remove device,
 *  start build, activate spare
 */
277
static DECLARE_WAIT_QUEUE_HEAD(md_event_waiters);
278
static atomic_t md_event_count;
279
void md_new_event(mddev_t *mddev)
280 281 282 283
{
	atomic_inc(&md_event_count);
	wake_up(&md_event_waiters);
}
284
EXPORT_SYMBOL_GPL(md_new_event);
285

286 287 288
/* Alternate version that can be called from interrupts
 * when calling sysfs_notify isn't needed.
 */
289
static void md_new_event_inintr(mddev_t *mddev)
290 291 292 293 294
{
	atomic_inc(&md_event_count);
	wake_up(&md_event_waiters);
}

L
Linus Torvalds 已提交
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
/*
 * Enables to iterate over all existing md arrays
 * all_mddevs_lock protects this list.
 */
static LIST_HEAD(all_mddevs);
static DEFINE_SPINLOCK(all_mddevs_lock);


/*
 * iterates through all used mddevs in the system.
 * We take care to grab the all_mddevs_lock whenever navigating
 * the list, and to always hold a refcount when unlocked.
 * Any code which breaks out of this loop while own
 * a reference to the current mddev and must mddev_put it.
 */
310
#define for_each_mddev(mddev,tmp)					\
L
Linus Torvalds 已提交
311 312 313 314 315 316 317 318 319 320 321 322 323 324 325
									\
	for (({ spin_lock(&all_mddevs_lock); 				\
		tmp = all_mddevs.next;					\
		mddev = NULL;});					\
	     ({ if (tmp != &all_mddevs)					\
			mddev_get(list_entry(tmp, mddev_t, all_mddevs));\
		spin_unlock(&all_mddevs_lock);				\
		if (mddev) mddev_put(mddev);				\
		mddev = list_entry(tmp, mddev_t, all_mddevs);		\
		tmp != &all_mddevs;});					\
	     ({ spin_lock(&all_mddevs_lock);				\
		tmp = tmp->next;})					\
		)


326 327 328 329 330 331 332
/* Rather than calling directly into the personality make_request function,
 * IO requests come here first so that we can check if the device is
 * being suspended pending a reconfiguration.
 * We hold a refcount over the call to ->make_request.  By the time that
 * call has finished, the bio has been linked into some internal structure
 * and so is visible to ->quiesce(), so we don't need the refcount any more.
 */
333
static void md_make_request(struct request_queue *q, struct bio *bio)
L
Linus Torvalds 已提交
334
{
335
	const int rw = bio_data_dir(bio);
336
	mddev_t *mddev = q->queuedata;
337
	int cpu;
338
	unsigned int sectors;
339

340 341
	if (mddev == NULL || mddev->pers == NULL
	    || !mddev->ready) {
342
		bio_io_error(bio);
343
		return;
344
	}
345
	smp_rmb(); /* Ensure implications of  'active' are visible */
346
	rcu_read_lock();
347
	if (mddev->suspended) {
348 349 350 351
		DEFINE_WAIT(__wait);
		for (;;) {
			prepare_to_wait(&mddev->sb_wait, &__wait,
					TASK_UNINTERRUPTIBLE);
352
			if (!mddev->suspended)
353 354 355 356 357 358 359 360 361
				break;
			rcu_read_unlock();
			schedule();
			rcu_read_lock();
		}
		finish_wait(&mddev->sb_wait, &__wait);
	}
	atomic_inc(&mddev->active_io);
	rcu_read_unlock();
362

363 364 365 366 367
	/*
	 * save the sectors now since our bio can
	 * go away inside make_request
	 */
	sectors = bio_sectors(bio);
368
	mddev->pers->make_request(mddev, bio);
369 370 371

	cpu = part_stat_lock();
	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
372
	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw], sectors);
373 374
	part_stat_unlock();

375 376 377 378
	if (atomic_dec_and_test(&mddev->active_io) && mddev->suspended)
		wake_up(&mddev->sb_wait);
}

379 380 381 382 383 384
/* mddev_suspend makes sure no new requests are submitted
 * to the device, and that any requests that have been submitted
 * are completely handled.
 * Once ->stop is called and completes, the module will be completely
 * unused.
 */
385
void mddev_suspend(mddev_t *mddev)
386 387 388 389 390 391 392
{
	BUG_ON(mddev->suspended);
	mddev->suspended = 1;
	synchronize_rcu();
	wait_event(mddev->sb_wait, atomic_read(&mddev->active_io) == 0);
	mddev->pers->quiesce(mddev, 1);
}
393
EXPORT_SYMBOL_GPL(mddev_suspend);
394

395
void mddev_resume(mddev_t *mddev)
396 397 398 399
{
	mddev->suspended = 0;
	wake_up(&mddev->sb_wait);
	mddev->pers->quiesce(mddev, 0);
400 401 402

	md_wakeup_thread(mddev->thread);
	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */
L
Linus Torvalds 已提交
403
}
404
EXPORT_SYMBOL_GPL(mddev_resume);
L
Linus Torvalds 已提交
405

406 407 408 409 410 411
int mddev_congested(mddev_t *mddev, int bits)
{
	return mddev->suspended;
}
EXPORT_SYMBOL(mddev_congested);

412
/*
413
 * Generic flush handling for md
414 415
 */

416
static void md_end_flush(struct bio *bio, int err)
417 418 419 420 421 422 423
{
	mdk_rdev_t *rdev = bio->bi_private;
	mddev_t *mddev = rdev->mddev;

	rdev_dec_pending(rdev, mddev);

	if (atomic_dec_and_test(&mddev->flush_pending)) {
424
		/* The pre-request flush has finished */
425
		queue_work(md_wq, &mddev->flush_work);
426 427 428 429
	}
	bio_put(bio);
}

430 431
static void md_submit_flush_data(struct work_struct *ws);

432
static void submit_flushes(struct work_struct *ws)
433
{
434
	mddev_t *mddev = container_of(ws, mddev_t, flush_work);
435 436
	mdk_rdev_t *rdev;

437 438
	INIT_WORK(&mddev->flush_work, md_submit_flush_data);
	atomic_set(&mddev->flush_pending, 1);
439 440 441 442 443 444 445 446 447 448 449 450
	rcu_read_lock();
	list_for_each_entry_rcu(rdev, &mddev->disks, same_set)
		if (rdev->raid_disk >= 0 &&
		    !test_bit(Faulty, &rdev->flags)) {
			/* Take two references, one is dropped
			 * when request finishes, one after
			 * we reclaim rcu_read_lock
			 */
			struct bio *bi;
			atomic_inc(&rdev->nr_pending);
			atomic_inc(&rdev->nr_pending);
			rcu_read_unlock();
451
			bi = bio_alloc_mddev(GFP_KERNEL, 0, mddev);
452
			bi->bi_end_io = md_end_flush;
453 454 455
			bi->bi_private = rdev;
			bi->bi_bdev = rdev->bdev;
			atomic_inc(&mddev->flush_pending);
456
			submit_bio(WRITE_FLUSH, bi);
457 458 459 460
			rcu_read_lock();
			rdev_dec_pending(rdev, mddev);
		}
	rcu_read_unlock();
461 462
	if (atomic_dec_and_test(&mddev->flush_pending))
		queue_work(md_wq, &mddev->flush_work);
463 464
}

465
static void md_submit_flush_data(struct work_struct *ws)
466
{
467 468
	mddev_t *mddev = container_of(ws, mddev_t, flush_work);
	struct bio *bio = mddev->flush_bio;
469

470
	if (bio->bi_size == 0)
471 472 473
		/* an empty barrier - all done */
		bio_endio(bio, 0);
	else {
474
		bio->bi_rw &= ~REQ_FLUSH;
475
		mddev->pers->make_request(mddev, bio);
476
	}
477 478 479

	mddev->flush_bio = NULL;
	wake_up(&mddev->sb_wait);
480 481
}

482
void md_flush_request(mddev_t *mddev, struct bio *bio)
483 484 485
{
	spin_lock_irq(&mddev->write_lock);
	wait_event_lock_irq(mddev->sb_wait,
486
			    !mddev->flush_bio,
487
			    mddev->write_lock, /*nothing*/);
488
	mddev->flush_bio = bio;
489 490
	spin_unlock_irq(&mddev->write_lock);

491 492
	INIT_WORK(&mddev->flush_work, submit_flushes);
	queue_work(md_wq, &mddev->flush_work);
493
}
494
EXPORT_SYMBOL(md_flush_request);
495

496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
/* Support for plugging.
 * This mirrors the plugging support in request_queue, but does not
 * require having a whole queue or request structures.
 * We allocate an md_plug_cb for each md device and each thread it gets
 * plugged on.  This links tot the private plug_handle structure in the
 * personality data where we keep a count of the number of outstanding
 * plugs so other code can see if a plug is active.
 */
struct md_plug_cb {
	struct blk_plug_cb cb;
	mddev_t *mddev;
};

static void plugger_unplug(struct blk_plug_cb *cb)
{
	struct md_plug_cb *mdcb = container_of(cb, struct md_plug_cb, cb);
	if (atomic_dec_and_test(&mdcb->mddev->plug_cnt))
		md_wakeup_thread(mdcb->mddev->thread);
	kfree(mdcb);
}

/* Check that an unplug wakeup will come shortly.
 * If not, wakeup the md thread immediately
 */
int mddev_check_plugged(mddev_t *mddev)
{
	struct blk_plug *plug = current->plug;
	struct md_plug_cb *mdcb;

	if (!plug)
		return 0;

	list_for_each_entry(mdcb, &plug->cb_list, cb.list) {
		if (mdcb->cb.callback == plugger_unplug &&
		    mdcb->mddev == mddev) {
			/* Already on the list, move to top */
			if (mdcb != list_first_entry(&plug->cb_list,
						    struct md_plug_cb,
						    cb.list))
				list_move(&mdcb->cb.list, &plug->cb_list);
			return 1;
		}
	}
	/* Not currently on the callback list */
	mdcb = kmalloc(sizeof(*mdcb), GFP_ATOMIC);
	if (!mdcb)
		return 0;

	mdcb->mddev = mddev;
	mdcb->cb.callback = plugger_unplug;
	atomic_inc(&mddev->plug_cnt);
	list_add(&mdcb->cb.list, &plug->cb_list);
	return 1;
}
EXPORT_SYMBOL_GPL(mddev_check_plugged);
551

L
Linus Torvalds 已提交
552 553 554 555 556 557
static inline mddev_t *mddev_get(mddev_t *mddev)
{
	atomic_inc(&mddev->active);
	return mddev;
}

558
static void mddev_delayed_delete(struct work_struct *ws);
559

L
Linus Torvalds 已提交
560 561
static void mddev_put(mddev_t *mddev)
{
562 563
	struct bio_set *bs = NULL;

L
Linus Torvalds 已提交
564 565
	if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
		return;
566
	if (!mddev->raid_disks && list_empty(&mddev->disks) &&
567 568 569
	    mddev->ctime == 0 && !mddev->hold_active) {
		/* Array is not configured at all, and not held active,
		 * so destroy it */
L
Linus Torvalds 已提交
570
		list_del(&mddev->all_mddevs);
571 572
		bs = mddev->bio_set;
		mddev->bio_set = NULL;
573
		if (mddev->gendisk) {
574 575 576 577
			/* We did a probe so need to clean up.  Call
			 * queue_work inside the spinlock so that
			 * flush_workqueue() after mddev_find will
			 * succeed in waiting for the work to be done.
578 579
			 */
			INIT_WORK(&mddev->del_work, mddev_delayed_delete);
580
			queue_work(md_misc_wq, &mddev->del_work);
581 582 583 584
		} else
			kfree(mddev);
	}
	spin_unlock(&all_mddevs_lock);
585 586
	if (bs)
		bioset_free(bs);
L
Linus Torvalds 已提交
587 588
}

589
void mddev_init(mddev_t *mddev)
590 591 592 593 594 595 596 597 598 599
{
	mutex_init(&mddev->open_mutex);
	mutex_init(&mddev->reconfig_mutex);
	mutex_init(&mddev->bitmap_info.mutex);
	INIT_LIST_HEAD(&mddev->disks);
	INIT_LIST_HEAD(&mddev->all_mddevs);
	init_timer(&mddev->safemode_timer);
	atomic_set(&mddev->active, 1);
	atomic_set(&mddev->openers, 0);
	atomic_set(&mddev->active_io, 0);
600
	atomic_set(&mddev->plug_cnt, 0);
601 602 603 604 605 606 607 608 609
	spin_lock_init(&mddev->write_lock);
	atomic_set(&mddev->flush_pending, 0);
	init_waitqueue_head(&mddev->sb_wait);
	init_waitqueue_head(&mddev->recovery_wait);
	mddev->reshape_position = MaxSector;
	mddev->resync_min = 0;
	mddev->resync_max = MaxSector;
	mddev->level = LEVEL_NONE;
}
610
EXPORT_SYMBOL_GPL(mddev_init);
611

L
Linus Torvalds 已提交
612 613 614 615
static mddev_t * mddev_find(dev_t unit)
{
	mddev_t *mddev, *new = NULL;

616 617 618
	if (unit && MAJOR(unit) != MD_MAJOR)
		unit &= ~((1<<MdpMinorShift)-1);

L
Linus Torvalds 已提交
619 620
 retry:
	spin_lock(&all_mddevs_lock);
621 622 623 624 625 626 627 628 629 630 631 632

	if (unit) {
		list_for_each_entry(mddev, &all_mddevs, all_mddevs)
			if (mddev->unit == unit) {
				mddev_get(mddev);
				spin_unlock(&all_mddevs_lock);
				kfree(new);
				return mddev;
			}

		if (new) {
			list_add(&new->all_mddevs, &all_mddevs);
L
Linus Torvalds 已提交
633
			spin_unlock(&all_mddevs_lock);
634 635
			new->hold_active = UNTIL_IOCTL;
			return new;
L
Linus Torvalds 已提交
636
		}
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
	} else if (new) {
		/* find an unused unit number */
		static int next_minor = 512;
		int start = next_minor;
		int is_free = 0;
		int dev = 0;
		while (!is_free) {
			dev = MKDEV(MD_MAJOR, next_minor);
			next_minor++;
			if (next_minor > MINORMASK)
				next_minor = 0;
			if (next_minor == start) {
				/* Oh dear, all in use. */
				spin_unlock(&all_mddevs_lock);
				kfree(new);
				return NULL;
			}
				
			is_free = 1;
			list_for_each_entry(mddev, &all_mddevs, all_mddevs)
				if (mddev->unit == dev) {
					is_free = 0;
					break;
				}
		}
		new->unit = dev;
		new->md_minor = MINOR(dev);
		new->hold_active = UNTIL_STOP;
L
Linus Torvalds 已提交
665 666 667 668 669 670
		list_add(&new->all_mddevs, &all_mddevs);
		spin_unlock(&all_mddevs_lock);
		return new;
	}
	spin_unlock(&all_mddevs_lock);

671
	new = kzalloc(sizeof(*new), GFP_KERNEL);
L
Linus Torvalds 已提交
672 673 674 675 676 677 678 679 680
	if (!new)
		return NULL;

	new->unit = unit;
	if (MAJOR(unit) == MD_MAJOR)
		new->md_minor = MINOR(unit);
	else
		new->md_minor = MINOR(unit) >> MdpMinorShift;

681
	mddev_init(new);
L
Linus Torvalds 已提交
682 683 684 685 686 687

	goto retry;
}

static inline int mddev_lock(mddev_t * mddev)
{
688
	return mutex_lock_interruptible(&mddev->reconfig_mutex);
L
Linus Torvalds 已提交
689 690
}

691 692 693 694 695
static inline int mddev_is_locked(mddev_t *mddev)
{
	return mutex_is_locked(&mddev->reconfig_mutex);
}

L
Linus Torvalds 已提交
696 697
static inline int mddev_trylock(mddev_t * mddev)
{
698
	return mutex_trylock(&mddev->reconfig_mutex);
L
Linus Torvalds 已提交
699 700
}

701 702
static struct attribute_group md_redundancy_group;

703
static void mddev_unlock(mddev_t * mddev)
L
Linus Torvalds 已提交
704
{
705
	if (mddev->to_remove) {
706 707 708 709
		/* These cannot be removed under reconfig_mutex as
		 * an access to the files will try to take reconfig_mutex
		 * while holding the file unremovable, which leads to
		 * a deadlock.
710 711 712 713 714 715 716
		 * So hold set sysfs_active while the remove in happeing,
		 * and anything else which might set ->to_remove or my
		 * otherwise change the sysfs namespace will fail with
		 * -EBUSY if sysfs_active is still set.
		 * We set sysfs_active under reconfig_mutex and elsewhere
		 * test it under the same mutex to ensure its correct value
		 * is seen.
717
		 */
718 719
		struct attribute_group *to_remove = mddev->to_remove;
		mddev->to_remove = NULL;
720
		mddev->sysfs_active = 1;
721 722
		mutex_unlock(&mddev->reconfig_mutex);

723 724 725 726 727 728 729 730 731 732
		if (mddev->kobj.sd) {
			if (to_remove != &md_redundancy_group)
				sysfs_remove_group(&mddev->kobj, to_remove);
			if (mddev->pers == NULL ||
			    mddev->pers->sync_request == NULL) {
				sysfs_remove_group(&mddev->kobj, &md_redundancy_group);
				if (mddev->sysfs_action)
					sysfs_put(mddev->sysfs_action);
				mddev->sysfs_action = NULL;
			}
733
		}
734
		mddev->sysfs_active = 0;
735 736
	} else
		mutex_unlock(&mddev->reconfig_mutex);
L
Linus Torvalds 已提交
737

738
	md_wakeup_thread(mddev->thread);
L
Linus Torvalds 已提交
739 740
}

741
static mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr)
L
Linus Torvalds 已提交
742
{
743
	mdk_rdev_t *rdev;
L
Linus Torvalds 已提交
744

745
	list_for_each_entry(rdev, &mddev->disks, same_set)
L
Linus Torvalds 已提交
746 747
		if (rdev->desc_nr == nr)
			return rdev;
748

L
Linus Torvalds 已提交
749 750 751 752 753 754 755
	return NULL;
}

static mdk_rdev_t * find_rdev(mddev_t * mddev, dev_t dev)
{
	mdk_rdev_t *rdev;

756
	list_for_each_entry(rdev, &mddev->disks, same_set)
L
Linus Torvalds 已提交
757 758
		if (rdev->bdev->bd_dev == dev)
			return rdev;
759

L
Linus Torvalds 已提交
760 761 762
	return NULL;
}

763
static struct mdk_personality *find_pers(int level, char *clevel)
764 765
{
	struct mdk_personality *pers;
766 767
	list_for_each_entry(pers, &pers_list, list) {
		if (level != LEVEL_NONE && pers->level == level)
768
			return pers;
769 770 771
		if (strcmp(pers->name, clevel)==0)
			return pers;
	}
772 773 774
	return NULL;
}

775
/* return the offset of the super block in 512byte sectors */
776
static inline sector_t calc_dev_sboffset(mdk_rdev_t *rdev)
L
Linus Torvalds 已提交
777
{
778
	sector_t num_sectors = i_size_read(rdev->bdev->bd_inode) / 512;
779
	return MD_NEW_SIZE_SECTORS(num_sectors);
L
Linus Torvalds 已提交
780 781 782 783 784 785 786 787 788 789
}

static int alloc_disk_sb(mdk_rdev_t * rdev)
{
	if (rdev->sb_page)
		MD_BUG();

	rdev->sb_page = alloc_page(GFP_KERNEL);
	if (!rdev->sb_page) {
		printk(KERN_ALERT "md: out of memory.\n");
790
		return -ENOMEM;
L
Linus Torvalds 已提交
791 792 793 794 795 796 797 798
	}

	return 0;
}

static void free_disk_sb(mdk_rdev_t * rdev)
{
	if (rdev->sb_page) {
799
		put_page(rdev->sb_page);
L
Linus Torvalds 已提交
800 801
		rdev->sb_loaded = 0;
		rdev->sb_page = NULL;
802
		rdev->sb_start = 0;
803
		rdev->sectors = 0;
L
Linus Torvalds 已提交
804
	}
805 806 807 808
	if (rdev->bb_page) {
		put_page(rdev->bb_page);
		rdev->bb_page = NULL;
	}
L
Linus Torvalds 已提交
809 810 811
}


812
static void super_written(struct bio *bio, int error)
813 814
{
	mdk_rdev_t *rdev = bio->bi_private;
815
	mddev_t *mddev = rdev->mddev;
816

817 818 819 820
	if (error || !test_bit(BIO_UPTODATE, &bio->bi_flags)) {
		printk("md: super_written gets error=%d, uptodate=%d\n",
		       error, test_bit(BIO_UPTODATE, &bio->bi_flags));
		WARN_ON(test_bit(BIO_UPTODATE, &bio->bi_flags));
821
		md_error(mddev, rdev);
822
	}
823

824 825
	if (atomic_dec_and_test(&mddev->pending_writes))
		wake_up(&mddev->sb_wait);
N
Neil Brown 已提交
826
	bio_put(bio);
827 828 829 830 831 832 833 834 835 836 837
}

void md_super_write(mddev_t *mddev, mdk_rdev_t *rdev,
		   sector_t sector, int size, struct page *page)
{
	/* write first size bytes of page to sector of rdev
	 * Increment mddev->pending_writes before returning
	 * and decrement it on completion, waking up sb_wait
	 * if zero is reached.
	 * If an error occurred, call md_error
	 */
838
	struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, mddev);
839

840
	bio->bi_bdev = rdev->meta_bdev ? rdev->meta_bdev : rdev->bdev;
841 842 843 844
	bio->bi_sector = sector;
	bio_add_page(bio, page, size, 0);
	bio->bi_private = rdev;
	bio->bi_end_io = super_written;
845

846
	atomic_inc(&mddev->pending_writes);
J
Jens Axboe 已提交
847
	submit_bio(REQ_WRITE | REQ_SYNC | REQ_FLUSH | REQ_FUA, bio);
848 849 850 851
}

void md_super_wait(mddev_t *mddev)
{
852
	/* wait for all superblock writes that were scheduled to complete */
853 854 855 856 857 858 859 860
	DEFINE_WAIT(wq);
	for(;;) {
		prepare_to_wait(&mddev->sb_wait, &wq, TASK_UNINTERRUPTIBLE);
		if (atomic_read(&mddev->pending_writes)==0)
			break;
		schedule();
	}
	finish_wait(&mddev->sb_wait, &wq);
861 862
}

863
static void bi_complete(struct bio *bio, int error)
L
Linus Torvalds 已提交
864 865 866 867
{
	complete((struct completion*)bio->bi_private);
}

868
int sync_page_io(mdk_rdev_t *rdev, sector_t sector, int size,
869
		 struct page *page, int rw, bool metadata_op)
L
Linus Torvalds 已提交
870
{
871
	struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, rdev->mddev);
L
Linus Torvalds 已提交
872 873 874
	struct completion event;
	int ret;

J
Jens Axboe 已提交
875
	rw |= REQ_SYNC;
L
Linus Torvalds 已提交
876

877 878
	bio->bi_bdev = (metadata_op && rdev->meta_bdev) ?
		rdev->meta_bdev : rdev->bdev;
879 880 881 882
	if (metadata_op)
		bio->bi_sector = sector + rdev->sb_start;
	else
		bio->bi_sector = sector + rdev->data_offset;
L
Linus Torvalds 已提交
883 884 885 886 887 888 889 890 891 892 893
	bio_add_page(bio, page, size, 0);
	init_completion(&event);
	bio->bi_private = &event;
	bio->bi_end_io = bi_complete;
	submit_bio(rw, bio);
	wait_for_completion(&event);

	ret = test_bit(BIO_UPTODATE, &bio->bi_flags);
	bio_put(bio);
	return ret;
}
894
EXPORT_SYMBOL_GPL(sync_page_io);
L
Linus Torvalds 已提交
895

896
static int read_disk_sb(mdk_rdev_t * rdev, int size)
L
Linus Torvalds 已提交
897 898 899 900 901 902 903 904 905 906
{
	char b[BDEVNAME_SIZE];
	if (!rdev->sb_page) {
		MD_BUG();
		return -EINVAL;
	}
	if (rdev->sb_loaded)
		return 0;


907
	if (!sync_page_io(rdev, 0, size, rdev->sb_page, READ, true))
L
Linus Torvalds 已提交
908 909 910 911 912 913 914 915 916 917 918 919
		goto fail;
	rdev->sb_loaded = 1;
	return 0;

fail:
	printk(KERN_WARNING "md: disabled device %s, could not read superblock.\n",
		bdevname(rdev->bdev,b));
	return -EINVAL;
}

static int uuid_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
A
Andre Noll 已提交
920 921 922 923
	return 	sb1->set_uuid0 == sb2->set_uuid0 &&
		sb1->set_uuid1 == sb2->set_uuid1 &&
		sb1->set_uuid2 == sb2->set_uuid2 &&
		sb1->set_uuid3 == sb2->set_uuid3;
L
Linus Torvalds 已提交
924 925 926 927 928 929 930 931 932 933 934 935
}

static int sb_equal(mdp_super_t *sb1, mdp_super_t *sb2)
{
	int ret;
	mdp_super_t *tmp1, *tmp2;

	tmp1 = kmalloc(sizeof(*tmp1),GFP_KERNEL);
	tmp2 = kmalloc(sizeof(*tmp2),GFP_KERNEL);

	if (!tmp1 || !tmp2) {
		ret = 0;
936
		printk(KERN_INFO "md.c sb_equal(): failed to allocate memory!\n");
L
Linus Torvalds 已提交
937 938 939 940 941 942 943 944 945 946 947 948
		goto abort;
	}

	*tmp1 = *sb1;
	*tmp2 = *sb2;

	/*
	 * nr_disks is not constant
	 */
	tmp1->nr_disks = 0;
	tmp2->nr_disks = 0;

A
Andre Noll 已提交
949
	ret = (memcmp(tmp1, tmp2, MD_SB_GENERIC_CONSTANT_WORDS * 4) == 0);
L
Linus Torvalds 已提交
950
abort:
951 952
	kfree(tmp1);
	kfree(tmp2);
L
Linus Torvalds 已提交
953 954 955
	return ret;
}

956 957 958 959 960 961 962

static u32 md_csum_fold(u32 csum)
{
	csum = (csum & 0xffff) + (csum >> 16);
	return (csum & 0xffff) + (csum >> 16);
}

L
Linus Torvalds 已提交
963 964
static unsigned int calc_sb_csum(mdp_super_t * sb)
{
965 966 967
	u64 newcsum = 0;
	u32 *sb32 = (u32*)sb;
	int i;
L
Linus Torvalds 已提交
968 969 970 971
	unsigned int disk_csum, csum;

	disk_csum = sb->sb_csum;
	sb->sb_csum = 0;
972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988

	for (i = 0; i < MD_SB_BYTES/4 ; i++)
		newcsum += sb32[i];
	csum = (newcsum & 0xffffffff) + (newcsum>>32);


#ifdef CONFIG_ALPHA
	/* This used to use csum_partial, which was wrong for several
	 * reasons including that different results are returned on
	 * different architectures.  It isn't critical that we get exactly
	 * the same return value as before (we always csum_fold before
	 * testing, and that removes any differences).  However as we
	 * know that csum_partial always returned a 16bit value on
	 * alphas, do a fold to maximise conformity to previous behaviour.
	 */
	sb->sb_csum = md_csum_fold(disk_csum);
#else
L
Linus Torvalds 已提交
989
	sb->sb_csum = disk_csum;
990
#endif
L
Linus Torvalds 已提交
991 992 993 994 995 996 997 998 999 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
	return csum;
}


/*
 * Handle superblock details.
 * We want to be able to handle multiple superblock formats
 * so we have a common interface to them all, and an array of
 * different handlers.
 * We rely on user-space to write the initial superblock, and support
 * reading and updating of superblocks.
 * Interface methods are:
 *   int load_super(mdk_rdev_t *dev, mdk_rdev_t *refdev, int minor_version)
 *      loads and validates a superblock on dev.
 *      if refdev != NULL, compare superblocks on both devices
 *    Return:
 *      0 - dev has a superblock that is compatible with refdev
 *      1 - dev has a superblock that is compatible and newer than refdev
 *          so dev should be used as the refdev in future
 *     -EINVAL superblock incompatible or invalid
 *     -othererror e.g. -EIO
 *
 *   int validate_super(mddev_t *mddev, mdk_rdev_t *dev)
 *      Verify that dev is acceptable into mddev.
 *       The first time, mddev->raid_disks will be 0, and data from
 *       dev should be merged in.  Subsequent calls check that dev
 *       is new enough.  Return 0 or -EINVAL
 *
 *   void sync_super(mddev_t *mddev, mdk_rdev_t *dev)
 *     Update the superblock for rdev with data in mddev
 *     This does not write to disc.
 *
 */

struct super_type  {
1026 1027 1028 1029 1030 1031 1032
	char		    *name;
	struct module	    *owner;
	int		    (*load_super)(mdk_rdev_t *rdev, mdk_rdev_t *refdev,
					  int minor_version);
	int		    (*validate_super)(mddev_t *mddev, mdk_rdev_t *rdev);
	void		    (*sync_super)(mddev_t *mddev, mdk_rdev_t *rdev);
	unsigned long long  (*rdev_size_change)(mdk_rdev_t *rdev,
1033
						sector_t num_sectors);
L
Linus Torvalds 已提交
1034 1035
};

1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
/*
 * Check that the given mddev has no bitmap.
 *
 * This function is called from the run method of all personalities that do not
 * support bitmaps. It prints an error message and returns non-zero if mddev
 * has a bitmap. Otherwise, it returns 0.
 *
 */
int md_check_no_bitmap(mddev_t *mddev)
{
1046
	if (!mddev->bitmap_info.file && !mddev->bitmap_info.offset)
1047 1048 1049 1050 1051 1052 1053
		return 0;
	printk(KERN_ERR "%s: bitmaps are not supported for %s\n",
		mdname(mddev), mddev->pers->name);
	return 1;
}
EXPORT_SYMBOL(md_check_no_bitmap);

L
Linus Torvalds 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
/*
 * load_super for 0.90.0 
 */
static int super_90_load(mdk_rdev_t *rdev, mdk_rdev_t *refdev, int minor_version)
{
	char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
	mdp_super_t *sb;
	int ret;

	/*
1064
	 * Calculate the position of the superblock (512byte sectors),
L
Linus Torvalds 已提交
1065 1066 1067 1068
	 * it's at the end of the disk.
	 *
	 * It also happens to be a multiple of 4Kb.
	 */
1069
	rdev->sb_start = calc_dev_sboffset(rdev);
L
Linus Torvalds 已提交
1070

1071
	ret = read_disk_sb(rdev, MD_SB_BYTES);
L
Linus Torvalds 已提交
1072 1073 1074 1075 1076
	if (ret) return ret;

	ret = -EINVAL;

	bdevname(rdev->bdev, b);
1077
	sb = page_address(rdev->sb_page);
L
Linus Torvalds 已提交
1078 1079 1080 1081 1082 1083 1084 1085

	if (sb->md_magic != MD_SB_MAGIC) {
		printk(KERN_ERR "md: invalid raid superblock magic on %s\n",
		       b);
		goto abort;
	}

	if (sb->major_version != 0 ||
1086 1087
	    sb->minor_version < 90 ||
	    sb->minor_version > 91) {
L
Linus Torvalds 已提交
1088 1089 1090 1091 1092 1093 1094 1095 1096
		printk(KERN_WARNING "Bad version number %d.%d on %s\n",
			sb->major_version, sb->minor_version,
			b);
		goto abort;
	}

	if (sb->raid_disks <= 0)
		goto abort;

1097
	if (md_csum_fold(calc_sb_csum(sb)) != md_csum_fold(sb->sb_csum)) {
L
Linus Torvalds 已提交
1098 1099 1100 1101 1102 1103 1104
		printk(KERN_WARNING "md: invalid superblock checksum on %s\n",
			b);
		goto abort;
	}

	rdev->preferred_minor = sb->md_minor;
	rdev->data_offset = 0;
1105
	rdev->sb_size = MD_SB_BYTES;
1106
	rdev->badblocks.shift = -1;
L
Linus Torvalds 已提交
1107 1108 1109 1110 1111 1112

	if (sb->level == LEVEL_MULTIPATH)
		rdev->desc_nr = -1;
	else
		rdev->desc_nr = sb->this_disk.number;

1113
	if (!refdev) {
L
Linus Torvalds 已提交
1114
		ret = 1;
1115
	} else {
L
Linus Torvalds 已提交
1116
		__u64 ev1, ev2;
1117
		mdp_super_t *refsb = page_address(refdev->sb_page);
L
Linus Torvalds 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135
		if (!uuid_equal(refsb, sb)) {
			printk(KERN_WARNING "md: %s has different UUID to %s\n",
				b, bdevname(refdev->bdev,b2));
			goto abort;
		}
		if (!sb_equal(refsb, sb)) {
			printk(KERN_WARNING "md: %s has same UUID"
			       " but different superblock to %s\n",
			       b, bdevname(refdev->bdev, b2));
			goto abort;
		}
		ev1 = md_event(sb);
		ev2 = md_event(refsb);
		if (ev1 > ev2)
			ret = 1;
		else 
			ret = 0;
	}
1136
	rdev->sectors = rdev->sb_start;
L
Linus Torvalds 已提交
1137

1138
	if (rdev->sectors < sb->size * 2 && sb->level > 1)
1139 1140 1141
		/* "this cannot possibly happen" ... */
		ret = -EINVAL;

L
Linus Torvalds 已提交
1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
 abort:
	return ret;
}

/*
 * validate_super for 0.90.0
 */
static int super_90_validate(mddev_t *mddev, mdk_rdev_t *rdev)
{
	mdp_disk_t *desc;
1152
	mdp_super_t *sb = page_address(rdev->sb_page);
1153
	__u64 ev1 = md_event(sb);
L
Linus Torvalds 已提交
1154

1155
	rdev->raid_disk = -1;
1156 1157 1158 1159
	clear_bit(Faulty, &rdev->flags);
	clear_bit(In_sync, &rdev->flags);
	clear_bit(WriteMostly, &rdev->flags);

L
Linus Torvalds 已提交
1160 1161 1162 1163
	if (mddev->raid_disks == 0) {
		mddev->major_version = 0;
		mddev->minor_version = sb->minor_version;
		mddev->patch_version = sb->patch_version;
1164
		mddev->external = 0;
1165
		mddev->chunk_sectors = sb->chunk_size >> 9;
L
Linus Torvalds 已提交
1166 1167 1168
		mddev->ctime = sb->ctime;
		mddev->utime = sb->utime;
		mddev->level = sb->level;
1169
		mddev->clevel[0] = 0;
L
Linus Torvalds 已提交
1170 1171
		mddev->layout = sb->layout;
		mddev->raid_disks = sb->raid_disks;
1172
		mddev->dev_sectors = sb->size * 2;
1173
		mddev->events = ev1;
1174 1175
		mddev->bitmap_info.offset = 0;
		mddev->bitmap_info.default_offset = MD_SB_BYTES >> 9;
L
Linus Torvalds 已提交
1176

1177 1178 1179 1180 1181
		if (mddev->minor_version >= 91) {
			mddev->reshape_position = sb->reshape_position;
			mddev->delta_disks = sb->delta_disks;
			mddev->new_level = sb->new_level;
			mddev->new_layout = sb->new_layout;
1182
			mddev->new_chunk_sectors = sb->new_chunk >> 9;
1183 1184 1185 1186 1187
		} else {
			mddev->reshape_position = MaxSector;
			mddev->delta_disks = 0;
			mddev->new_level = mddev->level;
			mddev->new_layout = mddev->layout;
1188
			mddev->new_chunk_sectors = mddev->chunk_sectors;
1189 1190
		}

L
Linus Torvalds 已提交
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206
		if (sb->state & (1<<MD_SB_CLEAN))
			mddev->recovery_cp = MaxSector;
		else {
			if (sb->events_hi == sb->cp_events_hi && 
				sb->events_lo == sb->cp_events_lo) {
				mddev->recovery_cp = sb->recovery_cp;
			} else
				mddev->recovery_cp = 0;
		}

		memcpy(mddev->uuid+0, &sb->set_uuid0, 4);
		memcpy(mddev->uuid+4, &sb->set_uuid1, 4);
		memcpy(mddev->uuid+8, &sb->set_uuid2, 4);
		memcpy(mddev->uuid+12,&sb->set_uuid3, 4);

		mddev->max_disks = MD_SB_DISKS;
1207 1208

		if (sb->state & (1<<MD_SB_BITMAP_PRESENT) &&
1209 1210 1211
		    mddev->bitmap_info.file == NULL)
			mddev->bitmap_info.offset =
				mddev->bitmap_info.default_offset;
1212

1213
	} else if (mddev->pers == NULL) {
1214 1215
		/* Insist on good event counter while assembling, except
		 * for spares (which don't need an event count) */
L
Linus Torvalds 已提交
1216
		++ev1;
1217 1218 1219 1220
		if (sb->disks[rdev->desc_nr].state & (
			    (1<<MD_DISK_SYNC) | (1 << MD_DISK_ACTIVE)))
			if (ev1 < mddev->events) 
				return -EINVAL;
1221 1222 1223 1224 1225 1226
	} else if (mddev->bitmap) {
		/* if adding to array with a bitmap, then we can accept an
		 * older device ... but not too old.
		 */
		if (ev1 < mddev->bitmap->events_cleared)
			return 0;
1227 1228 1229 1230 1231
	} else {
		if (ev1 < mddev->events)
			/* just a hot-add of a new device, leave raid_disk at -1 */
			return 0;
	}
1232

L
Linus Torvalds 已提交
1233 1234 1235 1236
	if (mddev->level != LEVEL_MULTIPATH) {
		desc = sb->disks + rdev->desc_nr;

		if (desc->state & (1<<MD_DISK_FAULTY))
1237
			set_bit(Faulty, &rdev->flags);
1238 1239
		else if (desc->state & (1<<MD_DISK_SYNC) /* &&
			    desc->raid_disk < mddev->raid_disks */) {
1240
			set_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
1241
			rdev->raid_disk = desc->raid_disk;
1242 1243 1244 1245 1246 1247 1248 1249
		} else if (desc->state & (1<<MD_DISK_ACTIVE)) {
			/* active but not in sync implies recovery up to
			 * reshape position.  We don't know exactly where
			 * that is, so set to zero for now */
			if (mddev->minor_version >= 91) {
				rdev->recovery_offset = 0;
				rdev->raid_disk = desc->raid_disk;
			}
L
Linus Torvalds 已提交
1250
		}
1251 1252
		if (desc->state & (1<<MD_DISK_WRITEMOSTLY))
			set_bit(WriteMostly, &rdev->flags);
1253
	} else /* MULTIPATH are always insync */
1254
		set_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
	return 0;
}

/*
 * sync_super for 0.90.0
 */
static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev)
{
	mdp_super_t *sb;
	mdk_rdev_t *rdev2;
	int next_spare = mddev->raid_disks;
1266

L
Linus Torvalds 已提交
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280

	/* make rdev->sb match mddev data..
	 *
	 * 1/ zero out disks
	 * 2/ Add info for each disk, keeping track of highest desc_nr (next_spare);
	 * 3/ any empty disks < next_spare become removed
	 *
	 * disks[0] gets initialised to REMOVED because
	 * we cannot be sure from other fields if it has
	 * been initialised or not.
	 */
	int i;
	int active=0, working=0,failed=0,spare=0,nr_disks=0;

1281 1282
	rdev->sb_size = MD_SB_BYTES;

1283
	sb = page_address(rdev->sb_page);
L
Linus Torvalds 已提交
1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297

	memset(sb, 0, sizeof(*sb));

	sb->md_magic = MD_SB_MAGIC;
	sb->major_version = mddev->major_version;
	sb->patch_version = mddev->patch_version;
	sb->gvalid_words  = 0; /* ignored */
	memcpy(&sb->set_uuid0, mddev->uuid+0, 4);
	memcpy(&sb->set_uuid1, mddev->uuid+4, 4);
	memcpy(&sb->set_uuid2, mddev->uuid+8, 4);
	memcpy(&sb->set_uuid3, mddev->uuid+12,4);

	sb->ctime = mddev->ctime;
	sb->level = mddev->level;
1298
	sb->size = mddev->dev_sectors / 2;
L
Linus Torvalds 已提交
1299 1300
	sb->raid_disks = mddev->raid_disks;
	sb->md_minor = mddev->md_minor;
1301
	sb->not_persistent = 0;
L
Linus Torvalds 已提交
1302 1303 1304 1305 1306
	sb->utime = mddev->utime;
	sb->state = 0;
	sb->events_hi = (mddev->events>>32);
	sb->events_lo = (u32)mddev->events;

1307 1308 1309 1310 1311 1312 1313 1314
	if (mddev->reshape_position == MaxSector)
		sb->minor_version = 90;
	else {
		sb->minor_version = 91;
		sb->reshape_position = mddev->reshape_position;
		sb->new_level = mddev->new_level;
		sb->delta_disks = mddev->delta_disks;
		sb->new_layout = mddev->new_layout;
1315
		sb->new_chunk = mddev->new_chunk_sectors << 9;
1316 1317
	}
	mddev->minor_version = sb->minor_version;
L
Linus Torvalds 已提交
1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
	if (mddev->in_sync)
	{
		sb->recovery_cp = mddev->recovery_cp;
		sb->cp_events_hi = (mddev->events>>32);
		sb->cp_events_lo = (u32)mddev->events;
		if (mddev->recovery_cp == MaxSector)
			sb->state = (1<< MD_SB_CLEAN);
	} else
		sb->recovery_cp = 0;

	sb->layout = mddev->layout;
1329
	sb->chunk_size = mddev->chunk_sectors << 9;
L
Linus Torvalds 已提交
1330

1331
	if (mddev->bitmap && mddev->bitmap_info.file == NULL)
1332 1333
		sb->state |= (1<<MD_SB_BITMAP_PRESENT);

L
Linus Torvalds 已提交
1334
	sb->disks[0].state = (1<<MD_DISK_REMOVED);
1335
	list_for_each_entry(rdev2, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
1336
		mdp_disk_t *d;
1337
		int desc_nr;
1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350
		int is_active = test_bit(In_sync, &rdev2->flags);

		if (rdev2->raid_disk >= 0 &&
		    sb->minor_version >= 91)
			/* we have nowhere to store the recovery_offset,
			 * but if it is not below the reshape_position,
			 * we can piggy-back on that.
			 */
			is_active = 1;
		if (rdev2->raid_disk < 0 ||
		    test_bit(Faulty, &rdev2->flags))
			is_active = 0;
		if (is_active)
1351
			desc_nr = rdev2->raid_disk;
L
Linus Torvalds 已提交
1352
		else
1353
			desc_nr = next_spare++;
1354
		rdev2->desc_nr = desc_nr;
L
Linus Torvalds 已提交
1355 1356 1357 1358 1359
		d = &sb->disks[rdev2->desc_nr];
		nr_disks++;
		d->number = rdev2->desc_nr;
		d->major = MAJOR(rdev2->bdev->bd_dev);
		d->minor = MINOR(rdev2->bdev->bd_dev);
1360
		if (is_active)
L
Linus Torvalds 已提交
1361 1362 1363
			d->raid_disk = rdev2->raid_disk;
		else
			d->raid_disk = rdev2->desc_nr; /* compatibility */
1364
		if (test_bit(Faulty, &rdev2->flags))
L
Linus Torvalds 已提交
1365
			d->state = (1<<MD_DISK_FAULTY);
1366
		else if (is_active) {
L
Linus Torvalds 已提交
1367
			d->state = (1<<MD_DISK_ACTIVE);
1368 1369
			if (test_bit(In_sync, &rdev2->flags))
				d->state |= (1<<MD_DISK_SYNC);
L
Linus Torvalds 已提交
1370 1371 1372 1373 1374 1375 1376
			active++;
			working++;
		} else {
			d->state = 0;
			spare++;
			working++;
		}
1377 1378
		if (test_bit(WriteMostly, &rdev2->flags))
			d->state |= (1<<MD_DISK_WRITEMOSTLY);
L
Linus Torvalds 已提交
1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
	}
	/* now set the "removed" and "faulty" bits on any missing devices */
	for (i=0 ; i < mddev->raid_disks ; i++) {
		mdp_disk_t *d = &sb->disks[i];
		if (d->state == 0 && d->number == 0) {
			d->number = i;
			d->raid_disk = i;
			d->state = (1<<MD_DISK_REMOVED);
			d->state |= (1<<MD_DISK_FAULTY);
			failed++;
		}
	}
	sb->nr_disks = nr_disks;
	sb->active_disks = active;
	sb->working_disks = working;
	sb->failed_disks = failed;
	sb->spare_disks = spare;

	sb->this_disk = sb->disks[rdev->desc_nr];
	sb->sb_csum = calc_sb_csum(sb);
}

1401 1402 1403 1404
/*
 * rdev_size_change for 0.90.0
 */
static unsigned long long
1405
super_90_rdev_size_change(mdk_rdev_t *rdev, sector_t num_sectors)
1406
{
1407
	if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
1408
		return 0; /* component must fit device */
1409
	if (rdev->mddev->bitmap_info.offset)
1410
		return 0; /* can't move bitmap */
1411
	rdev->sb_start = calc_dev_sboffset(rdev);
1412 1413
	if (!num_sectors || num_sectors > rdev->sb_start)
		num_sectors = rdev->sb_start;
1414
	md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
1415 1416
		       rdev->sb_page);
	md_super_wait(rdev->mddev);
1417
	return num_sectors;
1418 1419 1420
}


L
Linus Torvalds 已提交
1421 1422 1423 1424
/*
 * version 1 superblock
 */

1425
static __le32 calc_sb_1_csum(struct mdp_superblock_1 * sb)
L
Linus Torvalds 已提交
1426
{
1427 1428
	__le32 disk_csum;
	u32 csum;
L
Linus Torvalds 已提交
1429 1430
	unsigned long long newcsum;
	int size = 256 + le32_to_cpu(sb->max_dev)*2;
1431
	__le32 *isuper = (__le32*)sb;
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436 1437 1438 1439 1440
	int i;

	disk_csum = sb->sb_csum;
	sb->sb_csum = 0;
	newcsum = 0;
	for (i=0; size>=4; size -= 4 )
		newcsum += le32_to_cpu(*isuper++);

	if (size == 2)
1441
		newcsum += le16_to_cpu(*(__le16*) isuper);
L
Linus Torvalds 已提交
1442 1443 1444 1445 1446 1447

	csum = (newcsum & 0xffffffff) + (newcsum >> 32);
	sb->sb_csum = disk_csum;
	return cpu_to_le32(csum);
}

1448 1449
static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors,
			    int acknowledged);
L
Linus Torvalds 已提交
1450 1451 1452 1453
static int super_1_load(mdk_rdev_t *rdev, mdk_rdev_t *refdev, int minor_version)
{
	struct mdp_superblock_1 *sb;
	int ret;
1454
	sector_t sb_start;
L
Linus Torvalds 已提交
1455
	char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
1456
	int bmask;
L
Linus Torvalds 已提交
1457 1458

	/*
1459
	 * Calculate the position of the superblock in 512byte sectors.
L
Linus Torvalds 已提交
1460 1461 1462 1463 1464 1465 1466 1467
	 * It is always aligned to a 4K boundary and
	 * depeding on minor_version, it can be:
	 * 0: At least 8K, but less than 12K, from end of device
	 * 1: At start of device
	 * 2: 4K from start of device.
	 */
	switch(minor_version) {
	case 0:
1468
		sb_start = i_size_read(rdev->bdev->bd_inode) >> 9;
1469 1470
		sb_start -= 8*2;
		sb_start &= ~(sector_t)(4*2-1);
L
Linus Torvalds 已提交
1471 1472
		break;
	case 1:
1473
		sb_start = 0;
L
Linus Torvalds 已提交
1474 1475
		break;
	case 2:
1476
		sb_start = 8;
L
Linus Torvalds 已提交
1477 1478 1479 1480
		break;
	default:
		return -EINVAL;
	}
1481
	rdev->sb_start = sb_start;
L
Linus Torvalds 已提交
1482

1483 1484 1485 1486
	/* superblock is rarely larger than 1K, but it can be larger,
	 * and it is safe to read 4k, so we do that
	 */
	ret = read_disk_sb(rdev, 4096);
L
Linus Torvalds 已提交
1487 1488 1489
	if (ret) return ret;


1490
	sb = page_address(rdev->sb_page);
L
Linus Torvalds 已提交
1491 1492 1493 1494

	if (sb->magic != cpu_to_le32(MD_SB_MAGIC) ||
	    sb->major_version != cpu_to_le32(1) ||
	    le32_to_cpu(sb->max_dev) > (4096-256)/2 ||
1495
	    le64_to_cpu(sb->super_offset) != rdev->sb_start ||
1496
	    (le32_to_cpu(sb->feature_map) & ~MD_FEATURE_ALL) != 0)
L
Linus Torvalds 已提交
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508
		return -EINVAL;

	if (calc_sb_1_csum(sb) != sb->sb_csum) {
		printk("md: invalid superblock checksum on %s\n",
			bdevname(rdev->bdev,b));
		return -EINVAL;
	}
	if (le64_to_cpu(sb->data_size) < 10) {
		printk("md: data_size too small on %s\n",
		       bdevname(rdev->bdev,b));
		return -EINVAL;
	}
1509

L
Linus Torvalds 已提交
1510 1511
	rdev->preferred_minor = 0xffff;
	rdev->data_offset = le64_to_cpu(sb->data_offset);
1512
	atomic_set(&rdev->corrected_errors, le32_to_cpu(sb->cnt_corrected_read));
L
Linus Torvalds 已提交
1513

1514
	rdev->sb_size = le32_to_cpu(sb->max_dev) * 2 + 256;
1515
	bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
1516
	if (rdev->sb_size & bmask)
1517 1518 1519
		rdev->sb_size = (rdev->sb_size | bmask) + 1;

	if (minor_version
1520
	    && rdev->data_offset < sb_start + (rdev->sb_size/512))
1521
		return -EINVAL;
1522

1523 1524 1525 1526 1527
	if (sb->level == cpu_to_le32(LEVEL_MULTIPATH))
		rdev->desc_nr = -1;
	else
		rdev->desc_nr = le32_to_cpu(sb->dev_number);

1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
	if (!rdev->bb_page) {
		rdev->bb_page = alloc_page(GFP_KERNEL);
		if (!rdev->bb_page)
			return -ENOMEM;
	}
	if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BAD_BLOCKS) &&
	    rdev->badblocks.count == 0) {
		/* need to load the bad block list.
		 * Currently we limit it to one page.
		 */
		s32 offset;
		sector_t bb_sector;
		u64 *bbp;
		int i;
		int sectors = le16_to_cpu(sb->bblog_size);
		if (sectors > (PAGE_SIZE / 512))
			return -EINVAL;
		offset = le32_to_cpu(sb->bblog_offset);
		if (offset == 0)
			return -EINVAL;
		bb_sector = (long long)offset;
		if (!sync_page_io(rdev, bb_sector, sectors << 9,
				  rdev->bb_page, READ, true))
			return -EIO;
		bbp = (u64 *)page_address(rdev->bb_page);
		rdev->badblocks.shift = sb->bblog_shift;
		for (i = 0 ; i < (sectors << (9-3)) ; i++, bbp++) {
			u64 bb = le64_to_cpu(*bbp);
			int count = bb & (0x3ff);
			u64 sector = bb >> 10;
			sector <<= sb->bblog_shift;
			count <<= sb->bblog_shift;
			if (bb + 1 == 0)
				break;
			if (md_set_badblocks(&rdev->badblocks,
					     sector, count, 1) == 0)
				return -EINVAL;
		}
	} else if (sb->bblog_offset == 0)
		rdev->badblocks.shift = -1;

1569
	if (!refdev) {
1570
		ret = 1;
1571
	} else {
L
Linus Torvalds 已提交
1572
		__u64 ev1, ev2;
1573
		struct mdp_superblock_1 *refsb = page_address(refdev->sb_page);
L
Linus Torvalds 已提交
1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588

		if (memcmp(sb->set_uuid, refsb->set_uuid, 16) != 0 ||
		    sb->level != refsb->level ||
		    sb->layout != refsb->layout ||
		    sb->chunksize != refsb->chunksize) {
			printk(KERN_WARNING "md: %s has strangely different"
				" superblock to %s\n",
				bdevname(rdev->bdev,b),
				bdevname(refdev->bdev,b2));
			return -EINVAL;
		}
		ev1 = le64_to_cpu(sb->events);
		ev2 = le64_to_cpu(refsb->events);

		if (ev1 > ev2)
1589 1590 1591
			ret = 1;
		else
			ret = 0;
L
Linus Torvalds 已提交
1592
	}
1593
	if (minor_version)
1594
		rdev->sectors = (i_size_read(rdev->bdev->bd_inode) >> 9) -
1595
			le64_to_cpu(sb->data_offset);
L
Linus Torvalds 已提交
1596
	else
1597 1598
		rdev->sectors = rdev->sb_start;
	if (rdev->sectors < le64_to_cpu(sb->data_size))
L
Linus Torvalds 已提交
1599
		return -EINVAL;
1600 1601
	rdev->sectors = le64_to_cpu(sb->data_size);
	if (le64_to_cpu(sb->size) > rdev->sectors)
1602
		return -EINVAL;
1603
	return ret;
L
Linus Torvalds 已提交
1604 1605 1606 1607
}

static int super_1_validate(mddev_t *mddev, mdk_rdev_t *rdev)
{
1608
	struct mdp_superblock_1 *sb = page_address(rdev->sb_page);
1609
	__u64 ev1 = le64_to_cpu(sb->events);
L
Linus Torvalds 已提交
1610

1611
	rdev->raid_disk = -1;
1612 1613 1614 1615
	clear_bit(Faulty, &rdev->flags);
	clear_bit(In_sync, &rdev->flags);
	clear_bit(WriteMostly, &rdev->flags);

L
Linus Torvalds 已提交
1616 1617 1618
	if (mddev->raid_disks == 0) {
		mddev->major_version = 1;
		mddev->patch_version = 0;
1619
		mddev->external = 0;
1620
		mddev->chunk_sectors = le32_to_cpu(sb->chunksize);
L
Linus Torvalds 已提交
1621 1622 1623
		mddev->ctime = le64_to_cpu(sb->ctime) & ((1ULL << 32)-1);
		mddev->utime = le64_to_cpu(sb->utime) & ((1ULL << 32)-1);
		mddev->level = le32_to_cpu(sb->level);
1624
		mddev->clevel[0] = 0;
L
Linus Torvalds 已提交
1625 1626
		mddev->layout = le32_to_cpu(sb->layout);
		mddev->raid_disks = le32_to_cpu(sb->raid_disks);
1627
		mddev->dev_sectors = le64_to_cpu(sb->size);
1628
		mddev->events = ev1;
1629 1630
		mddev->bitmap_info.offset = 0;
		mddev->bitmap_info.default_offset = 1024 >> 9;
L
Linus Torvalds 已提交
1631 1632 1633 1634 1635
		
		mddev->recovery_cp = le64_to_cpu(sb->resync_offset);
		memcpy(mddev->uuid, sb->set_uuid, 16);

		mddev->max_disks =  (4096-256)/2;
1636

1637
		if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_BITMAP_OFFSET) &&
1638 1639 1640
		    mddev->bitmap_info.file == NULL )
			mddev->bitmap_info.offset =
				(__s32)le32_to_cpu(sb->bitmap_offset);
1641

1642 1643 1644 1645 1646
		if ((le32_to_cpu(sb->feature_map) & MD_FEATURE_RESHAPE_ACTIVE)) {
			mddev->reshape_position = le64_to_cpu(sb->reshape_position);
			mddev->delta_disks = le32_to_cpu(sb->delta_disks);
			mddev->new_level = le32_to_cpu(sb->new_level);
			mddev->new_layout = le32_to_cpu(sb->new_layout);
1647
			mddev->new_chunk_sectors = le32_to_cpu(sb->new_chunk);
1648 1649 1650 1651 1652
		} else {
			mddev->reshape_position = MaxSector;
			mddev->delta_disks = 0;
			mddev->new_level = mddev->level;
			mddev->new_layout = mddev->layout;
1653
			mddev->new_chunk_sectors = mddev->chunk_sectors;
1654 1655
		}

1656
	} else if (mddev->pers == NULL) {
1657 1658
		/* Insist of good event counter while assembling, except for
		 * spares (which don't need an event count) */
L
Linus Torvalds 已提交
1659
		++ev1;
1660 1661 1662 1663 1664
		if (rdev->desc_nr >= 0 &&
		    rdev->desc_nr < le32_to_cpu(sb->max_dev) &&
		    le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < 0xfffe)
			if (ev1 < mddev->events)
				return -EINVAL;
1665 1666 1667 1668 1669 1670
	} else if (mddev->bitmap) {
		/* If adding to array with a bitmap, then we can accept an
		 * older device, but not too old.
		 */
		if (ev1 < mddev->bitmap->events_cleared)
			return 0;
1671 1672 1673 1674 1675
	} else {
		if (ev1 < mddev->events)
			/* just a hot-add of a new device, leave raid_disk at -1 */
			return 0;
	}
L
Linus Torvalds 已提交
1676 1677
	if (mddev->level != LEVEL_MULTIPATH) {
		int role;
1678 1679 1680 1681 1682 1683
		if (rdev->desc_nr < 0 ||
		    rdev->desc_nr >= le32_to_cpu(sb->max_dev)) {
			role = 0xffff;
			rdev->desc_nr = -1;
		} else
			role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]);
L
Linus Torvalds 已提交
1684 1685 1686 1687
		switch(role) {
		case 0xffff: /* spare */
			break;
		case 0xfffe: /* faulty */
1688
			set_bit(Faulty, &rdev->flags);
L
Linus Torvalds 已提交
1689 1690
			break;
		default:
1691 1692 1693 1694 1695
			if ((le32_to_cpu(sb->feature_map) &
			     MD_FEATURE_RECOVERY_OFFSET))
				rdev->recovery_offset = le64_to_cpu(sb->recovery_offset);
			else
				set_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
1696 1697 1698
			rdev->raid_disk = role;
			break;
		}
1699 1700
		if (sb->devflags & WriteMostly1)
			set_bit(WriteMostly, &rdev->flags);
1701
	} else /* MULTIPATH are always insync */
1702
		set_bit(In_sync, &rdev->flags);
1703

L
Linus Torvalds 已提交
1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
	return 0;
}

static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev)
{
	struct mdp_superblock_1 *sb;
	mdk_rdev_t *rdev2;
	int max_dev, i;
	/* make rdev->sb match mddev and rdev data. */

1714
	sb = page_address(rdev->sb_page);
L
Linus Torvalds 已提交
1715 1716 1717

	sb->feature_map = 0;
	sb->pad0 = 0;
1718
	sb->recovery_offset = cpu_to_le64(0);
L
Linus Torvalds 已提交
1719 1720 1721 1722 1723 1724 1725 1726 1727 1728
	memset(sb->pad1, 0, sizeof(sb->pad1));
	memset(sb->pad3, 0, sizeof(sb->pad3));

	sb->utime = cpu_to_le64((__u64)mddev->utime);
	sb->events = cpu_to_le64(mddev->events);
	if (mddev->in_sync)
		sb->resync_offset = cpu_to_le64(mddev->recovery_cp);
	else
		sb->resync_offset = cpu_to_le64(0);

1729
	sb->cnt_corrected_read = cpu_to_le32(atomic_read(&rdev->corrected_errors));
1730

1731
	sb->raid_disks = cpu_to_le32(mddev->raid_disks);
1732
	sb->size = cpu_to_le64(mddev->dev_sectors);
1733
	sb->chunksize = cpu_to_le32(mddev->chunk_sectors);
1734 1735
	sb->level = cpu_to_le32(mddev->level);
	sb->layout = cpu_to_le32(mddev->layout);
1736

1737 1738
	if (mddev->bitmap && mddev->bitmap_info.file == NULL) {
		sb->bitmap_offset = cpu_to_le32((__u32)mddev->bitmap_info.offset);
1739
		sb->feature_map = cpu_to_le32(MD_FEATURE_BITMAP_OFFSET);
1740
	}
1741 1742

	if (rdev->raid_disk >= 0 &&
1743
	    !test_bit(In_sync, &rdev->flags)) {
1744 1745 1746 1747
		sb->feature_map |=
			cpu_to_le32(MD_FEATURE_RECOVERY_OFFSET);
		sb->recovery_offset =
			cpu_to_le64(rdev->recovery_offset);
1748 1749
	}

1750 1751 1752 1753 1754 1755
	if (mddev->reshape_position != MaxSector) {
		sb->feature_map |= cpu_to_le32(MD_FEATURE_RESHAPE_ACTIVE);
		sb->reshape_position = cpu_to_le64(mddev->reshape_position);
		sb->new_layout = cpu_to_le32(mddev->new_layout);
		sb->delta_disks = cpu_to_le32(mddev->delta_disks);
		sb->new_level = cpu_to_le32(mddev->new_level);
1756
		sb->new_chunk = cpu_to_le32(mddev->new_chunk_sectors);
1757
	}
1758

1759 1760 1761 1762 1763 1764 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
	if (rdev->badblocks.count == 0)
		/* Nothing to do for bad blocks*/ ;
	else if (sb->bblog_offset == 0)
		/* Cannot record bad blocks on this device */
		md_error(mddev, rdev);
	else {
		struct badblocks *bb = &rdev->badblocks;
		u64 *bbp = (u64 *)page_address(rdev->bb_page);
		u64 *p = bb->page;
		sb->feature_map |= cpu_to_le32(MD_FEATURE_BAD_BLOCKS);
		if (bb->changed) {
			unsigned seq;

retry:
			seq = read_seqbegin(&bb->lock);

			memset(bbp, 0xff, PAGE_SIZE);

			for (i = 0 ; i < bb->count ; i++) {
				u64 internal_bb = *p++;
				u64 store_bb = ((BB_OFFSET(internal_bb) << 10)
						| BB_LEN(internal_bb));
				*bbp++ = cpu_to_le64(store_bb);
			}
			if (read_seqretry(&bb->lock, seq))
				goto retry;

			bb->sector = (rdev->sb_start +
				      (int)le32_to_cpu(sb->bblog_offset));
			bb->size = le16_to_cpu(sb->bblog_size);
			bb->changed = 0;
		}
	}

L
Linus Torvalds 已提交
1793
	max_dev = 0;
1794
	list_for_each_entry(rdev2, &mddev->disks, same_set)
L
Linus Torvalds 已提交
1795 1796
		if (rdev2->desc_nr+1 > max_dev)
			max_dev = rdev2->desc_nr+1;
1797

1798 1799
	if (max_dev > le32_to_cpu(sb->max_dev)) {
		int bmask;
1800
		sb->max_dev = cpu_to_le32(max_dev);
1801 1802 1803 1804
		rdev->sb_size = max_dev * 2 + 256;
		bmask = queue_logical_block_size(rdev->bdev->bd_disk->queue)-1;
		if (rdev->sb_size & bmask)
			rdev->sb_size = (rdev->sb_size | bmask) + 1;
1805 1806 1807
	} else
		max_dev = le32_to_cpu(sb->max_dev);

L
Linus Torvalds 已提交
1808 1809 1810
	for (i=0; i<max_dev;i++)
		sb->dev_roles[i] = cpu_to_le16(0xfffe);
	
1811
	list_for_each_entry(rdev2, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
1812
		i = rdev2->desc_nr;
1813
		if (test_bit(Faulty, &rdev2->flags))
L
Linus Torvalds 已提交
1814
			sb->dev_roles[i] = cpu_to_le16(0xfffe);
1815
		else if (test_bit(In_sync, &rdev2->flags))
L
Linus Torvalds 已提交
1816
			sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
1817
		else if (rdev2->raid_disk >= 0)
1818
			sb->dev_roles[i] = cpu_to_le16(rdev2->raid_disk);
L
Linus Torvalds 已提交
1819 1820 1821 1822 1823 1824 1825
		else
			sb->dev_roles[i] = cpu_to_le16(0xffff);
	}

	sb->sb_csum = calc_sb_1_csum(sb);
}

1826
static unsigned long long
1827
super_1_rdev_size_change(mdk_rdev_t *rdev, sector_t num_sectors)
1828 1829
{
	struct mdp_superblock_1 *sb;
1830
	sector_t max_sectors;
1831
	if (num_sectors && num_sectors < rdev->mddev->dev_sectors)
1832
		return 0; /* component must fit device */
1833
	if (rdev->sb_start < rdev->data_offset) {
1834
		/* minor versions 1 and 2; superblock before data */
1835
		max_sectors = i_size_read(rdev->bdev->bd_inode) >> 9;
1836 1837 1838
		max_sectors -= rdev->data_offset;
		if (!num_sectors || num_sectors > max_sectors)
			num_sectors = max_sectors;
1839
	} else if (rdev->mddev->bitmap_info.offset) {
1840 1841 1842 1843
		/* minor version 0 with bitmap we can't move */
		return 0;
	} else {
		/* minor version 0; superblock after data */
1844
		sector_t sb_start;
1845
		sb_start = (i_size_read(rdev->bdev->bd_inode) >> 9) - 8*2;
1846
		sb_start &= ~(sector_t)(4*2 - 1);
1847
		max_sectors = rdev->sectors + sb_start - rdev->sb_start;
1848 1849
		if (!num_sectors || num_sectors > max_sectors)
			num_sectors = max_sectors;
1850
		rdev->sb_start = sb_start;
1851
	}
1852
	sb = page_address(rdev->sb_page);
1853
	sb->data_size = cpu_to_le64(num_sectors);
1854
	sb->super_offset = rdev->sb_start;
1855
	sb->sb_csum = calc_sb_1_csum(sb);
1856
	md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
1857 1858
		       rdev->sb_page);
	md_super_wait(rdev->mddev);
1859
	return num_sectors;
1860
}
L
Linus Torvalds 已提交
1861

1862
static struct super_type super_types[] = {
L
Linus Torvalds 已提交
1863 1864 1865
	[0] = {
		.name	= "0.90.0",
		.owner	= THIS_MODULE,
1866 1867 1868 1869
		.load_super	    = super_90_load,
		.validate_super	    = super_90_validate,
		.sync_super	    = super_90_sync,
		.rdev_size_change   = super_90_rdev_size_change,
L
Linus Torvalds 已提交
1870 1871 1872 1873
	},
	[1] = {
		.name	= "md-1",
		.owner	= THIS_MODULE,
1874 1875 1876 1877
		.load_super	    = super_1_load,
		.validate_super	    = super_1_validate,
		.sync_super	    = super_1_sync,
		.rdev_size_change   = super_1_rdev_size_change,
L
Linus Torvalds 已提交
1878 1879 1880
	},
};

1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892
static void sync_super(mddev_t *mddev, mdk_rdev_t *rdev)
{
	if (mddev->sync_super) {
		mddev->sync_super(mddev, rdev);
		return;
	}

	BUG_ON(mddev->major_version >= ARRAY_SIZE(super_types));

	super_types[mddev->major_version].sync_super(mddev, rdev);
}

L
Linus Torvalds 已提交
1893 1894
static int match_mddev_units(mddev_t *mddev1, mddev_t *mddev2)
{
1895
	mdk_rdev_t *rdev, *rdev2;
L
Linus Torvalds 已提交
1896

1897 1898 1899
	rcu_read_lock();
	rdev_for_each_rcu(rdev, mddev1)
		rdev_for_each_rcu(rdev2, mddev2)
1900
			if (rdev->bdev->bd_contains ==
1901 1902
			    rdev2->bdev->bd_contains) {
				rcu_read_unlock();
1903
				return 1;
1904 1905
			}
	rcu_read_unlock();
L
Linus Torvalds 已提交
1906 1907 1908 1909 1910
	return 0;
}

static LIST_HEAD(pending_raid_disks);

1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923
/*
 * Try to register data integrity profile for an mddev
 *
 * This is called when an array is started and after a disk has been kicked
 * from the array. It only succeeds if all working and active component devices
 * are integrity capable with matching profiles.
 */
int md_integrity_register(mddev_t *mddev)
{
	mdk_rdev_t *rdev, *reference = NULL;

	if (list_empty(&mddev->disks))
		return 0; /* nothing to do */
1924 1925
	if (!mddev->gendisk || blk_get_integrity(mddev->gendisk))
		return 0; /* shouldn't register, or already is */
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941
	list_for_each_entry(rdev, &mddev->disks, same_set) {
		/* skip spares and non-functional disks */
		if (test_bit(Faulty, &rdev->flags))
			continue;
		if (rdev->raid_disk < 0)
			continue;
		if (!reference) {
			/* Use the first rdev as the reference */
			reference = rdev;
			continue;
		}
		/* does this rdev's profile match the reference profile? */
		if (blk_integrity_compare(reference->bdev->bd_disk,
				rdev->bdev->bd_disk) < 0)
			return -EINVAL;
	}
1942 1943
	if (!reference || !bdev_get_integrity(reference->bdev))
		return 0;
1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
	/*
	 * All component devices are integrity capable and have matching
	 * profiles, register the common profile for the md device.
	 */
	if (blk_integrity_register(mddev->gendisk,
			bdev_get_integrity(reference->bdev)) != 0) {
		printk(KERN_ERR "md: failed to register integrity for %s\n",
			mdname(mddev));
		return -EINVAL;
	}
1954 1955 1956 1957 1958 1959
	printk(KERN_NOTICE "md: data integrity enabled on %s\n", mdname(mddev));
	if (bioset_integrity_create(mddev->bio_set, BIO_POOL_SIZE)) {
		printk(KERN_ERR "md: failed to create integrity pool for %s\n",
		       mdname(mddev));
		return -EINVAL;
	}
1960 1961 1962 1963 1964 1965
	return 0;
}
EXPORT_SYMBOL(md_integrity_register);

/* Disable data integrity if non-capable/non-matching disk is being added */
void md_integrity_add_rdev(mdk_rdev_t *rdev, mddev_t *mddev)
1966 1967
{
	struct blk_integrity *bi_rdev = bdev_get_integrity(rdev->bdev);
1968
	struct blk_integrity *bi_mddev = blk_get_integrity(mddev->gendisk);
1969

1970
	if (!bi_mddev) /* nothing to do */
1971
		return;
1972
	if (rdev->raid_disk < 0) /* skip spares */
1973
		return;
1974 1975 1976 1977 1978
	if (bi_rdev && blk_integrity_compare(mddev->gendisk,
					     rdev->bdev->bd_disk) >= 0)
		return;
	printk(KERN_NOTICE "disabling data integrity on %s\n", mdname(mddev));
	blk_integrity_unregister(mddev->gendisk);
1979
}
1980
EXPORT_SYMBOL(md_integrity_add_rdev);
1981

L
Linus Torvalds 已提交
1982 1983
static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev)
{
1984
	char b[BDEVNAME_SIZE];
1985
	struct kobject *ko;
1986
	char *s;
1987
	int err;
L
Linus Torvalds 已提交
1988 1989 1990 1991 1992

	if (rdev->mddev) {
		MD_BUG();
		return -EINVAL;
	}
1993 1994 1995 1996 1997

	/* prevent duplicates */
	if (find_rdev(mddev, rdev->bdev->bd_dev))
		return -EEXIST;

1998 1999 2000
	/* make sure rdev->sectors exceeds mddev->dev_sectors */
	if (rdev->sectors && (mddev->dev_sectors == 0 ||
			rdev->sectors < mddev->dev_sectors)) {
2001 2002 2003 2004 2005 2006 2007 2008
		if (mddev->pers) {
			/* Cannot change size, so fail
			 * If mddev->level <= 0, then we don't care
			 * about aligning sizes (e.g. linear)
			 */
			if (mddev->level > 0)
				return -ENOSPC;
		} else
2009
			mddev->dev_sectors = rdev->sectors;
2010
	}
L
Linus Torvalds 已提交
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025

	/* Verify rdev->desc_nr is unique.
	 * If it is -1, assign a free number, else
	 * check number is not in use
	 */
	if (rdev->desc_nr < 0) {
		int choice = 0;
		if (mddev->pers) choice = mddev->raid_disks;
		while (find_rdev_nr(mddev, choice))
			choice++;
		rdev->desc_nr = choice;
	} else {
		if (find_rdev_nr(mddev, rdev->desc_nr))
			return -EBUSY;
	}
2026 2027 2028 2029 2030
	if (mddev->max_disks && rdev->desc_nr >= mddev->max_disks) {
		printk(KERN_WARNING "md: %s: array is limited to %d devices\n",
		       mdname(mddev), mddev->max_disks);
		return -EBUSY;
	}
2031
	bdevname(rdev->bdev,b);
2032
	while ( (s=strchr(b, '/')) != NULL)
2033
		*s = '!';
2034

L
Linus Torvalds 已提交
2035
	rdev->mddev = mddev;
2036
	printk(KERN_INFO "md: bind<%s>\n", b);
2037

2038
	if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
2039
		goto fail;
2040

2041
	ko = &part_to_dev(rdev->bdev->bd_part)->kobj;
2042 2043 2044
	if (sysfs_create_link(&rdev->kobj, ko, "block"))
		/* failure here is OK */;
	rdev->sysfs_state = sysfs_get_dirent_safe(rdev->kobj.sd, "state");
2045

2046
	list_add_rcu(&rdev->same_set, &mddev->disks);
2047
	bd_link_disk_holder(rdev->bdev, mddev->gendisk);
2048 2049

	/* May as well allow recovery to be retried once */
2050
	mddev->recovery_disabled++;
2051

L
Linus Torvalds 已提交
2052
	return 0;
2053 2054 2055 2056 2057

 fail:
	printk(KERN_WARNING "md: failed to register dev-%s for %s\n",
	       b, mdname(mddev));
	return err;
L
Linus Torvalds 已提交
2058 2059
}

2060
static void md_delayed_delete(struct work_struct *ws)
2061 2062 2063
{
	mdk_rdev_t *rdev = container_of(ws, mdk_rdev_t, del_work);
	kobject_del(&rdev->kobj);
2064
	kobject_put(&rdev->kobj);
2065 2066
}

L
Linus Torvalds 已提交
2067 2068 2069 2070 2071 2072 2073
static void unbind_rdev_from_array(mdk_rdev_t * rdev)
{
	char b[BDEVNAME_SIZE];
	if (!rdev->mddev) {
		MD_BUG();
		return;
	}
2074
	bd_unlink_disk_holder(rdev->bdev, rdev->mddev->gendisk);
2075
	list_del_rcu(&rdev->same_set);
L
Linus Torvalds 已提交
2076 2077
	printk(KERN_INFO "md: unbind<%s>\n", bdevname(rdev->bdev,b));
	rdev->mddev = NULL;
2078
	sysfs_remove_link(&rdev->kobj, "block");
2079 2080
	sysfs_put(rdev->sysfs_state);
	rdev->sysfs_state = NULL;
2081 2082 2083
	kfree(rdev->badblocks.page);
	rdev->badblocks.count = 0;
	rdev->badblocks.page = NULL;
2084
	/* We need to delay this, otherwise we can deadlock when
2085 2086
	 * writing to 'remove' to "dev/state".  We also need
	 * to delay it due to rcu usage.
2087
	 */
2088
	synchronize_rcu();
2089 2090
	INIT_WORK(&rdev->del_work, md_delayed_delete);
	kobject_get(&rdev->kobj);
2091
	queue_work(md_misc_wq, &rdev->del_work);
L
Linus Torvalds 已提交
2092 2093 2094 2095 2096 2097 2098
}

/*
 * prevent the device from being mounted, repartitioned or
 * otherwise reused by a RAID array (or any other kernel
 * subsystem), by bd_claiming the device.
 */
2099
static int lock_rdev(mdk_rdev_t *rdev, dev_t dev, int shared)
L
Linus Torvalds 已提交
2100 2101 2102 2103 2104
{
	int err = 0;
	struct block_device *bdev;
	char b[BDEVNAME_SIZE];

2105 2106
	bdev = blkdev_get_by_dev(dev, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
				 shared ? (mdk_rdev_t *)lock_rdev : rdev);
L
Linus Torvalds 已提交
2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121
	if (IS_ERR(bdev)) {
		printk(KERN_ERR "md: could not open %s.\n",
			__bdevname(dev, b));
		return PTR_ERR(bdev);
	}
	rdev->bdev = bdev;
	return err;
}

static void unlock_rdev(mdk_rdev_t *rdev)
{
	struct block_device *bdev = rdev->bdev;
	rdev->bdev = NULL;
	if (!bdev)
		MD_BUG();
2122
	blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
L
Linus Torvalds 已提交
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135
}

void md_autodetect_dev(dev_t dev);

static void export_rdev(mdk_rdev_t * rdev)
{
	char b[BDEVNAME_SIZE];
	printk(KERN_INFO "md: export_rdev(%s)\n",
		bdevname(rdev->bdev,b));
	if (rdev->mddev)
		MD_BUG();
	free_disk_sb(rdev);
#ifndef MODULE
2136 2137
	if (test_bit(AutoDetected, &rdev->flags))
		md_autodetect_dev(rdev->bdev->bd_dev);
L
Linus Torvalds 已提交
2138 2139
#endif
	unlock_rdev(rdev);
2140
	kobject_put(&rdev->kobj);
L
Linus Torvalds 已提交
2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
}

static void kick_rdev_from_array(mdk_rdev_t * rdev)
{
	unbind_rdev_from_array(rdev);
	export_rdev(rdev);
}

static void export_array(mddev_t *mddev)
{
2151
	mdk_rdev_t *rdev, *tmp;
L
Linus Torvalds 已提交
2152

2153
	rdev_for_each(rdev, tmp, mddev) {
L
Linus Torvalds 已提交
2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171
		if (!rdev->mddev) {
			MD_BUG();
			continue;
		}
		kick_rdev_from_array(rdev);
	}
	if (!list_empty(&mddev->disks))
		MD_BUG();
	mddev->raid_disks = 0;
	mddev->major_version = 0;
}

static void print_desc(mdp_disk_t *desc)
{
	printk(" DISK<N:%d,(%d,%d),R:%d,S:%d>\n", desc->number,
		desc->major,desc->minor,desc->raid_disk,desc->state);
}

2172
static void print_sb_90(mdp_super_t *sb)
L
Linus Torvalds 已提交
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
{
	int i;

	printk(KERN_INFO 
		"md:  SB: (V:%d.%d.%d) ID:<%08x.%08x.%08x.%08x> CT:%08x\n",
		sb->major_version, sb->minor_version, sb->patch_version,
		sb->set_uuid0, sb->set_uuid1, sb->set_uuid2, sb->set_uuid3,
		sb->ctime);
	printk(KERN_INFO "md:     L%d S%08d ND:%d RD:%d md%d LO:%d CS:%d\n",
		sb->level, sb->size, sb->nr_disks, sb->raid_disks,
		sb->md_minor, sb->layout, sb->chunk_size);
	printk(KERN_INFO "md:     UT:%08x ST:%d AD:%d WD:%d"
		" FD:%d SD:%d CSUM:%08x E:%08lx\n",
		sb->utime, sb->state, sb->active_disks, sb->working_disks,
		sb->failed_disks, sb->spare_disks,
		sb->sb_csum, (unsigned long)sb->events_lo);

	printk(KERN_INFO);
	for (i = 0; i < MD_SB_DISKS; i++) {
		mdp_disk_t *desc;

		desc = sb->disks + i;
		if (desc->number || desc->major || desc->minor ||
		    desc->raid_disk || (desc->state && (desc->state != 4))) {
			printk("     D %2d: ", i);
			print_desc(desc);
		}
	}
	printk(KERN_INFO "md:     THIS: ");
	print_desc(&sb->this_disk);
2203
}
L
Linus Torvalds 已提交
2204

2205 2206 2207 2208 2209
static void print_sb_1(struct mdp_superblock_1 *sb)
{
	__u8 *uuid;

	uuid = sb->set_uuid;
2210
	printk(KERN_INFO
2211
	       "md:  SB: (V:%u) (F:0x%08x) Array-ID:<%pU>\n"
2212
	       "md:    Name: \"%s\" CT:%llu\n",
2213 2214
		le32_to_cpu(sb->major_version),
		le32_to_cpu(sb->feature_map),
2215
		uuid,
2216 2217 2218 2219 2220
		sb->set_name,
		(unsigned long long)le64_to_cpu(sb->ctime)
		       & MD_SUPERBLOCK_1_TIME_SEC_MASK);

	uuid = sb->device_uuid;
2221 2222
	printk(KERN_INFO
	       "md:       L%u SZ%llu RD:%u LO:%u CS:%u DO:%llu DS:%llu SO:%llu"
2223
			" RO:%llu\n"
2224
	       "md:     Dev:%08x UUID: %pU\n"
2225 2226
	       "md:       (F:0x%08x) UT:%llu Events:%llu ResyncOffset:%llu CSUM:0x%08x\n"
	       "md:         (MaxDev:%u) \n",
2227 2228 2229 2230 2231 2232 2233 2234 2235 2236
		le32_to_cpu(sb->level),
		(unsigned long long)le64_to_cpu(sb->size),
		le32_to_cpu(sb->raid_disks),
		le32_to_cpu(sb->layout),
		le32_to_cpu(sb->chunksize),
		(unsigned long long)le64_to_cpu(sb->data_offset),
		(unsigned long long)le64_to_cpu(sb->data_size),
		(unsigned long long)le64_to_cpu(sb->super_offset),
		(unsigned long long)le64_to_cpu(sb->recovery_offset),
		le32_to_cpu(sb->dev_number),
2237
		uuid,
2238 2239 2240 2241 2242 2243 2244
		sb->devflags,
		(unsigned long long)le64_to_cpu(sb->utime) & MD_SUPERBLOCK_1_TIME_SEC_MASK,
		(unsigned long long)le64_to_cpu(sb->events),
		(unsigned long long)le64_to_cpu(sb->resync_offset),
		le32_to_cpu(sb->sb_csum),
		le32_to_cpu(sb->max_dev)
		);
L
Linus Torvalds 已提交
2245 2246
}

2247
static void print_rdev(mdk_rdev_t *rdev, int major_version)
L
Linus Torvalds 已提交
2248 2249
{
	char b[BDEVNAME_SIZE];
2250 2251
	printk(KERN_INFO "md: rdev %s, Sect:%08llu F:%d S:%d DN:%u\n",
		bdevname(rdev->bdev, b), (unsigned long long)rdev->sectors,
2252 2253
	        test_bit(Faulty, &rdev->flags), test_bit(In_sync, &rdev->flags),
	        rdev->desc_nr);
L
Linus Torvalds 已提交
2254
	if (rdev->sb_loaded) {
2255 2256 2257
		printk(KERN_INFO "md: rdev superblock (MJ:%d):\n", major_version);
		switch (major_version) {
		case 0:
2258
			print_sb_90(page_address(rdev->sb_page));
2259 2260
			break;
		case 1:
2261
			print_sb_1(page_address(rdev->sb_page));
2262 2263
			break;
		}
L
Linus Torvalds 已提交
2264 2265 2266 2267
	} else
		printk(KERN_INFO "md: no rdev superblock!\n");
}

2268
static void md_print_devices(void)
L
Linus Torvalds 已提交
2269
{
2270
	struct list_head *tmp;
L
Linus Torvalds 已提交
2271 2272 2273 2274 2275 2276 2277 2278
	mdk_rdev_t *rdev;
	mddev_t *mddev;
	char b[BDEVNAME_SIZE];

	printk("\n");
	printk("md:	**********************************\n");
	printk("md:	* <COMPLETE RAID STATE PRINTOUT> *\n");
	printk("md:	**********************************\n");
2279
	for_each_mddev(mddev, tmp) {
L
Linus Torvalds 已提交
2280

2281 2282 2283 2284
		if (mddev->bitmap)
			bitmap_print_sb(mddev->bitmap);
		else
			printk("%s: ", mdname(mddev));
2285
		list_for_each_entry(rdev, &mddev->disks, same_set)
L
Linus Torvalds 已提交
2286 2287 2288
			printk("<%s>", bdevname(rdev->bdev,b));
		printk("\n");

2289
		list_for_each_entry(rdev, &mddev->disks, same_set)
2290
			print_rdev(rdev, mddev->major_version);
L
Linus Torvalds 已提交
2291 2292 2293 2294 2295 2296
	}
	printk("md:	**********************************\n");
	printk("\n");
}


2297
static void sync_sbs(mddev_t * mddev, int nospares)
L
Linus Torvalds 已提交
2298
{
2299 2300 2301 2302 2303 2304
	/* Update each superblock (in-memory image), but
	 * if we are allowed to, skip spares which already
	 * have the right event counter, or have one earlier
	 * (which would mean they aren't being marked as dirty
	 * with the rest of the array)
	 */
L
Linus Torvalds 已提交
2305
	mdk_rdev_t *rdev;
2306
	list_for_each_entry(rdev, &mddev->disks, same_set) {
2307 2308 2309 2310 2311 2312 2313
		if (rdev->sb_events == mddev->events ||
		    (nospares &&
		     rdev->raid_disk < 0 &&
		     rdev->sb_events+1 == mddev->events)) {
			/* Don't update this superblock */
			rdev->sb_loaded = 2;
		} else {
2314
			sync_super(mddev, rdev);
2315 2316
			rdev->sb_loaded = 1;
		}
L
Linus Torvalds 已提交
2317 2318 2319
	}
}

2320
static void md_update_sb(mddev_t * mddev, int force_change)
L
Linus Torvalds 已提交
2321 2322
{
	mdk_rdev_t *rdev;
2323
	int sync_req;
2324
	int nospares = 0;
2325
	int any_badblocks_changed = 0;
L
Linus Torvalds 已提交
2326 2327

repeat:
2328 2329 2330 2331 2332 2333 2334 2335 2336
	/* First make sure individual recovery_offsets are correct */
	list_for_each_entry(rdev, &mddev->disks, same_set) {
		if (rdev->raid_disk >= 0 &&
		    mddev->delta_disks >= 0 &&
		    !test_bit(In_sync, &rdev->flags) &&
		    mddev->curr_resync_completed > rdev->recovery_offset)
				rdev->recovery_offset = mddev->curr_resync_completed;

	}	
2337
	if (!mddev->persistent) {
2338
		clear_bit(MD_CHANGE_CLEAN, &mddev->flags);
2339
		clear_bit(MD_CHANGE_DEVS, &mddev->flags);
2340
		if (!mddev->external) {
2341
			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
2342 2343 2344 2345 2346 2347 2348 2349 2350 2351
			list_for_each_entry(rdev, &mddev->disks, same_set) {
				if (rdev->badblocks.changed) {
					md_ack_all_badblocks(&rdev->badblocks);
					md_error(mddev, rdev);
				}
				clear_bit(Blocked, &rdev->flags);
				clear_bit(BlockedBadBlocks, &rdev->flags);
				wake_up(&rdev->blocked_wait);
			}
		}
2352 2353 2354 2355
		wake_up(&mddev->sb_wait);
		return;
	}

2356
	spin_lock_irq(&mddev->write_lock);
2357

2358 2359
	mddev->utime = get_seconds();

2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370
	if (test_and_clear_bit(MD_CHANGE_DEVS, &mddev->flags))
		force_change = 1;
	if (test_and_clear_bit(MD_CHANGE_CLEAN, &mddev->flags))
		/* just a clean<-> dirty transition, possibly leave spares alone,
		 * though if events isn't the right even/odd, we will have to do
		 * spares after all
		 */
		nospares = 1;
	if (force_change)
		nospares = 0;
	if (mddev->degraded)
2371 2372 2373 2374 2375 2376 2377 2378 2379
		/* If the array is degraded, then skipping spares is both
		 * dangerous and fairly pointless.
		 * Dangerous because a device that was removed from the array
		 * might have a event_count that still looks up-to-date,
		 * so it can be re-added without a resync.
		 * Pointless because if there are any spares to skip,
		 * then a recovery will happen and soon that array won't
		 * be degraded any more and the spare can go back to sleep then.
		 */
2380
		nospares = 0;
2381

2382
	sync_req = mddev->in_sync;
2383 2384 2385

	/* If this is just a dirty<->clean transition, and the array is clean
	 * and 'events' is odd, we can roll back to the previous clean state */
2386
	if (nospares
2387
	    && (mddev->in_sync && mddev->recovery_cp == MaxSector)
2388 2389
	    && mddev->can_decrease_events
	    && mddev->events != 1) {
2390
		mddev->events--;
2391 2392
		mddev->can_decrease_events = 0;
	} else {
2393 2394
		/* otherwise we have to go forward and ... */
		mddev->events ++;
2395
		mddev->can_decrease_events = nospares;
2396
	}
L
Linus Torvalds 已提交
2397 2398 2399 2400 2401 2402 2403 2404 2405 2406

	if (!mddev->events) {
		/*
		 * oops, this 64-bit counter should never wrap.
		 * Either we are in around ~1 trillion A.C., assuming
		 * 1 reboot per second, or we have a bug:
		 */
		MD_BUG();
		mddev->events --;
	}
2407

2408
	list_for_each_entry(rdev, &mddev->disks, same_set) {
2409 2410
		if (rdev->badblocks.changed)
			any_badblocks_changed++;
2411 2412 2413
		if (test_bit(Faulty, &rdev->flags))
			set_bit(FaultRecorded, &rdev->flags);
	}
2414

2415
	sync_sbs(mddev, nospares);
2416
	spin_unlock_irq(&mddev->write_lock);
L
Linus Torvalds 已提交
2417 2418 2419 2420 2421

	dprintk(KERN_INFO 
		"md: updating %s RAID superblock on device (in sync %d)\n",
		mdname(mddev),mddev->in_sync);

2422
	bitmap_update_sb(mddev->bitmap);
2423
	list_for_each_entry(rdev, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
2424 2425
		char b[BDEVNAME_SIZE];
		dprintk(KERN_INFO "md: ");
2426 2427
		if (rdev->sb_loaded != 1)
			continue; /* no noise on spare devices */
2428
		if (test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
2429 2430 2431
			dprintk("(skipping faulty ");

		dprintk("%s ", bdevname(rdev->bdev,b));
2432
		if (!test_bit(Faulty, &rdev->flags)) {
2433
			md_super_write(mddev,rdev,
2434
				       rdev->sb_start, rdev->sb_size,
2435 2436 2437
				       rdev->sb_page);
			dprintk(KERN_INFO "(write) %s's sb offset: %llu\n",
				bdevname(rdev->bdev,b),
2438
				(unsigned long long)rdev->sb_start);
2439
			rdev->sb_events = mddev->events;
2440 2441 2442 2443 2444 2445 2446
			if (rdev->badblocks.size) {
				md_super_write(mddev, rdev,
					       rdev->badblocks.sector,
					       rdev->badblocks.size << 9,
					       rdev->bb_page);
				rdev->badblocks.size = 0;
			}
2447

L
Linus Torvalds 已提交
2448 2449
		} else
			dprintk(")\n");
2450
		if (mddev->level == LEVEL_MULTIPATH)
L
Linus Torvalds 已提交
2451 2452 2453
			/* only need to write one superblock... */
			break;
	}
2454
	md_super_wait(mddev);
2455
	/* if there was a failure, MD_CHANGE_DEVS was set, and we re-write super */
2456

2457
	spin_lock_irq(&mddev->write_lock);
2458 2459
	if (mddev->in_sync != sync_req ||
	    test_bit(MD_CHANGE_DEVS, &mddev->flags)) {
2460
		/* have to write it out again */
2461
		spin_unlock_irq(&mddev->write_lock);
2462 2463
		goto repeat;
	}
2464
	clear_bit(MD_CHANGE_PENDING, &mddev->flags);
2465
	spin_unlock_irq(&mddev->write_lock);
2466
	wake_up(&mddev->sb_wait);
2467 2468
	if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
		sysfs_notify(&mddev->kobj, NULL, "sync_completed");
2469

2470 2471 2472 2473 2474
	list_for_each_entry(rdev, &mddev->disks, same_set) {
		if (test_and_clear_bit(FaultRecorded, &rdev->flags))
			clear_bit(Blocked, &rdev->flags);

		if (any_badblocks_changed)
2475
			md_ack_all_badblocks(&rdev->badblocks);
2476 2477 2478
		clear_bit(BlockedBadBlocks, &rdev->flags);
		wake_up(&rdev->blocked_wait);
	}
L
Linus Torvalds 已提交
2479 2480
}

2481
/* words written to sysfs files may, or may not, be \n terminated.
2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500
 * We want to accept with case. For this we use cmd_match.
 */
static int cmd_match(const char *cmd, const char *str)
{
	/* See if cmd, written into a sysfs file, matches
	 * str.  They must either be the same, or cmd can
	 * have a trailing newline
	 */
	while (*cmd && *str && *cmd == *str) {
		cmd++;
		str++;
	}
	if (*cmd == '\n')
		cmd++;
	if (*str || *cmd)
		return 0;
	return 1;
}

2501 2502 2503 2504 2505 2506 2507
struct rdev_sysfs_entry {
	struct attribute attr;
	ssize_t (*show)(mdk_rdev_t *, char *);
	ssize_t (*store)(mdk_rdev_t *, const char *, size_t);
};

static ssize_t
2508
state_show(mdk_rdev_t *rdev, char *page)
2509 2510
{
	char *sep = "";
2511
	size_t len = 0;
2512

2513 2514
	if (test_bit(Faulty, &rdev->flags) ||
	    rdev->badblocks.unacked_exist) {
2515 2516 2517
		len+= sprintf(page+len, "%sfaulty",sep);
		sep = ",";
	}
2518
	if (test_bit(In_sync, &rdev->flags)) {
2519 2520 2521
		len += sprintf(page+len, "%sin_sync",sep);
		sep = ",";
	}
2522 2523 2524 2525
	if (test_bit(WriteMostly, &rdev->flags)) {
		len += sprintf(page+len, "%swrite_mostly",sep);
		sep = ",";
	}
2526 2527
	if (test_bit(Blocked, &rdev->flags) ||
	    rdev->badblocks.unacked_exist) {
2528 2529 2530
		len += sprintf(page+len, "%sblocked", sep);
		sep = ",";
	}
2531 2532
	if (!test_bit(Faulty, &rdev->flags) &&
	    !test_bit(In_sync, &rdev->flags)) {
2533 2534 2535
		len += sprintf(page+len, "%sspare", sep);
		sep = ",";
	}
2536 2537 2538 2539
	if (test_bit(WriteErrorSeen, &rdev->flags)) {
		len += sprintf(page+len, "%swrite_error", sep);
		sep = ",";
	}
2540 2541 2542
	return len+sprintf(page+len, "\n");
}

2543 2544 2545 2546
static ssize_t
state_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
	/* can write
2547
	 *  faulty  - simulates an error
2548
	 *  remove  - disconnects the device
2549 2550
	 *  writemostly - sets write_mostly
	 *  -writemostly - clears write_mostly
2551 2552
	 *  blocked - sets the Blocked flags
	 *  -blocked - clears the Blocked and possibly simulates an error
2553
	 *  insync - sets Insync providing device isn't active
2554 2555
	 *  write_error - sets WriteErrorSeen
	 *  -write_error - clears WriteErrorSeen
2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566
	 */
	int err = -EINVAL;
	if (cmd_match(buf, "faulty") && rdev->mddev->pers) {
		md_error(rdev->mddev, rdev);
		err = 0;
	} else if (cmd_match(buf, "remove")) {
		if (rdev->raid_disk >= 0)
			err = -EBUSY;
		else {
			mddev_t *mddev = rdev->mddev;
			kick_rdev_from_array(rdev);
2567 2568
			if (mddev->pers)
				md_update_sb(mddev, 1);
2569 2570 2571
			md_new_event(mddev);
			err = 0;
		}
2572 2573 2574 2575 2576
	} else if (cmd_match(buf, "writemostly")) {
		set_bit(WriteMostly, &rdev->flags);
		err = 0;
	} else if (cmd_match(buf, "-writemostly")) {
		clear_bit(WriteMostly, &rdev->flags);
2577 2578 2579 2580 2581
		err = 0;
	} else if (cmd_match(buf, "blocked")) {
		set_bit(Blocked, &rdev->flags);
		err = 0;
	} else if (cmd_match(buf, "-blocked")) {
2582 2583 2584 2585 2586 2587 2588
		if (!test_bit(Faulty, &rdev->flags) &&
		    test_bit(BlockedBadBlocks, &rdev->flags)) {
			/* metadata handler doesn't understand badblocks,
			 * so we need to fail the device
			 */
			md_error(rdev->mddev, rdev);
		}
2589
		clear_bit(Blocked, &rdev->flags);
2590
		clear_bit(BlockedBadBlocks, &rdev->flags);
2591 2592 2593 2594
		wake_up(&rdev->blocked_wait);
		set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
		md_wakeup_thread(rdev->mddev->thread);

2595 2596 2597
		err = 0;
	} else if (cmd_match(buf, "insync") && rdev->raid_disk == -1) {
		set_bit(In_sync, &rdev->flags);
2598
		err = 0;
2599 2600 2601 2602 2603 2604
	} else if (cmd_match(buf, "write_error")) {
		set_bit(WriteErrorSeen, &rdev->flags);
		err = 0;
	} else if (cmd_match(buf, "-write_error")) {
		clear_bit(WriteErrorSeen, &rdev->flags);
		err = 0;
2605
	}
2606 2607
	if (!err)
		sysfs_notify_dirent_safe(rdev->sysfs_state);
2608 2609
	return err ? err : len;
}
2610 2611
static struct rdev_sysfs_entry rdev_state =
__ATTR(state, S_IRUGO|S_IWUSR, state_show, state_store);
2612

2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630
static ssize_t
errors_show(mdk_rdev_t *rdev, char *page)
{
	return sprintf(page, "%d\n", atomic_read(&rdev->corrected_errors));
}

static ssize_t
errors_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
	char *e;
	unsigned long n = simple_strtoul(buf, &e, 10);
	if (*buf && (*e == 0 || *e == '\n')) {
		atomic_set(&rdev->corrected_errors, n);
		return len;
	}
	return -EINVAL;
}
static struct rdev_sysfs_entry rdev_errors =
2631
__ATTR(errors, S_IRUGO|S_IWUSR, errors_show, errors_store);
2632

2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645
static ssize_t
slot_show(mdk_rdev_t *rdev, char *page)
{
	if (rdev->raid_disk < 0)
		return sprintf(page, "none\n");
	else
		return sprintf(page, "%d\n", rdev->raid_disk);
}

static ssize_t
slot_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
	char *e;
2646
	int err;
2647 2648 2649 2650 2651
	int slot = simple_strtoul(buf, &e, 10);
	if (strncmp(buf, "none", 4)==0)
		slot = -1;
	else if (e==buf || (*e && *e!= '\n'))
		return -EINVAL;
2652
	if (rdev->mddev->pers && slot == -1) {
2653 2654 2655 2656 2657 2658 2659 2660 2661 2662
		/* Setting 'slot' on an active array requires also
		 * updating the 'rd%d' link, and communicating
		 * with the personality with ->hot_*_disk.
		 * For now we only support removing
		 * failed/spare devices.  This normally happens automatically,
		 * but not when the metadata is externally managed.
		 */
		if (rdev->raid_disk == -1)
			return -EEXIST;
		/* personality does all needed checks */
2663
		if (rdev->mddev->pers->hot_remove_disk == NULL)
2664 2665 2666 2667 2668
			return -EINVAL;
		err = rdev->mddev->pers->
			hot_remove_disk(rdev->mddev, rdev->raid_disk);
		if (err)
			return err;
2669
		sysfs_unlink_rdev(rdev->mddev, rdev);
2670
		rdev->raid_disk = -1;
2671 2672
		set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
		md_wakeup_thread(rdev->mddev->thread);
2673 2674 2675
	} else if (rdev->mddev->pers) {
		mdk_rdev_t *rdev2;
		/* Activating a spare .. or possibly reactivating
2676
		 * if we ever get bitmaps working here.
2677 2678 2679 2680 2681
		 */

		if (rdev->raid_disk != -1)
			return -EBUSY;

2682 2683 2684
		if (test_bit(MD_RECOVERY_RUNNING, &rdev->mddev->recovery))
			return -EBUSY;

2685 2686 2687
		if (rdev->mddev->pers->hot_add_disk == NULL)
			return -EINVAL;

2688
		list_for_each_entry(rdev2, &rdev->mddev->disks, same_set)
2689 2690 2691
			if (rdev2->raid_disk == slot)
				return -EEXIST;

2692 2693 2694 2695
		if (slot >= rdev->mddev->raid_disks &&
		    slot >= rdev->mddev->raid_disks + rdev->mddev->delta_disks)
			return -ENOSPC;

2696 2697 2698 2699 2700 2701 2702
		rdev->raid_disk = slot;
		if (test_bit(In_sync, &rdev->flags))
			rdev->saved_raid_disk = slot;
		else
			rdev->saved_raid_disk = -1;
		err = rdev->mddev->pers->
			hot_add_disk(rdev->mddev, rdev);
2703
		if (err) {
2704 2705
			rdev->raid_disk = -1;
			return err;
2706
		} else
2707
			sysfs_notify_dirent_safe(rdev->sysfs_state);
2708
		if (sysfs_link_rdev(rdev->mddev, rdev))
2709
			/* failure here is OK */;
2710
		/* don't wakeup anyone, leave that to userspace. */
2711
	} else {
2712 2713
		if (slot >= rdev->mddev->raid_disks &&
		    slot >= rdev->mddev->raid_disks + rdev->mddev->delta_disks)
2714 2715 2716
			return -ENOSPC;
		rdev->raid_disk = slot;
		/* assume it is working */
2717 2718
		clear_bit(Faulty, &rdev->flags);
		clear_bit(WriteMostly, &rdev->flags);
2719
		set_bit(In_sync, &rdev->flags);
2720
		sysfs_notify_dirent_safe(rdev->sysfs_state);
2721
	}
2722 2723 2724 2725 2726
	return len;
}


static struct rdev_sysfs_entry rdev_slot =
2727
__ATTR(slot, S_IRUGO|S_IWUSR, slot_show, slot_store);
2728

2729 2730 2731
static ssize_t
offset_show(mdk_rdev_t *rdev, char *page)
{
2732
	return sprintf(page, "%llu\n", (unsigned long long)rdev->data_offset);
2733 2734 2735 2736 2737 2738 2739 2740 2741
}

static ssize_t
offset_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
	char *e;
	unsigned long long offset = simple_strtoull(buf, &e, 10);
	if (e==buf || (*e && *e != '\n'))
		return -EINVAL;
2742
	if (rdev->mddev->pers && rdev->raid_disk >= 0)
2743
		return -EBUSY;
2744
	if (rdev->sectors && rdev->mddev->external)
2745 2746 2747
		/* Must set offset before size, so overlap checks
		 * can be sane */
		return -EBUSY;
2748 2749 2750 2751 2752
	rdev->data_offset = offset;
	return len;
}

static struct rdev_sysfs_entry rdev_offset =
2753
__ATTR(offset, S_IRUGO|S_IWUSR, offset_show, offset_store);
2754

2755 2756 2757
static ssize_t
rdev_size_show(mdk_rdev_t *rdev, char *page)
{
2758
	return sprintf(page, "%llu\n", (unsigned long long)rdev->sectors / 2);
2759 2760
}

2761 2762 2763 2764 2765 2766 2767 2768 2769 2770
static int overlaps(sector_t s1, sector_t l1, sector_t s2, sector_t l2)
{
	/* check if two start/length pairs overlap */
	if (s1+l1 <= s2)
		return 0;
	if (s2+l2 <= s1)
		return 0;
	return 1;
}

2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789
static int strict_blocks_to_sectors(const char *buf, sector_t *sectors)
{
	unsigned long long blocks;
	sector_t new;

	if (strict_strtoull(buf, 10, &blocks) < 0)
		return -EINVAL;

	if (blocks & 1ULL << (8 * sizeof(blocks) - 1))
		return -EINVAL; /* sector conversion overflow */

	new = blocks * 2;
	if (new != blocks * 2)
		return -EINVAL; /* unsigned long long to sector_t overflow */

	*sectors = new;
	return 0;
}

2790 2791 2792
static ssize_t
rdev_size_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
2793
	mddev_t *my_mddev = rdev->mddev;
2794
	sector_t oldsectors = rdev->sectors;
2795
	sector_t sectors;
2796

2797
	if (strict_blocks_to_sectors(buf, &sectors) < 0)
2798
		return -EINVAL;
2799
	if (my_mddev->pers && rdev->raid_disk >= 0) {
2800
		if (my_mddev->persistent) {
2801 2802 2803
			sectors = super_types[my_mddev->major_version].
				rdev_size_change(rdev, sectors);
			if (!sectors)
2804
				return -EBUSY;
2805
		} else if (!sectors)
2806
			sectors = (i_size_read(rdev->bdev->bd_inode) >> 9) -
2807
				rdev->data_offset;
2808
	}
2809
	if (sectors < my_mddev->dev_sectors)
2810
		return -EINVAL; /* component must fit device */
2811

2812 2813
	rdev->sectors = sectors;
	if (sectors > oldsectors && my_mddev->external) {
2814 2815
		/* need to check that all other rdevs with the same ->bdev
		 * do not overlap.  We need to unlock the mddev to avoid
2816
		 * a deadlock.  We have already changed rdev->sectors, and if
2817 2818 2819 2820
		 * we have to change it back, we will have the lock again.
		 */
		mddev_t *mddev;
		int overlap = 0;
2821
		struct list_head *tmp;
2822

2823
		mddev_unlock(my_mddev);
2824
		for_each_mddev(mddev, tmp) {
2825 2826 2827
			mdk_rdev_t *rdev2;

			mddev_lock(mddev);
2828
			list_for_each_entry(rdev2, &mddev->disks, same_set)
2829 2830 2831 2832 2833
				if (rdev->bdev == rdev2->bdev &&
				    rdev != rdev2 &&
				    overlaps(rdev->data_offset, rdev->sectors,
					     rdev2->data_offset,
					     rdev2->sectors)) {
2834 2835 2836 2837 2838 2839 2840 2841 2842
					overlap = 1;
					break;
				}
			mddev_unlock(mddev);
			if (overlap) {
				mddev_put(mddev);
				break;
			}
		}
2843
		mddev_lock(my_mddev);
2844 2845 2846
		if (overlap) {
			/* Someone else could have slipped in a size
			 * change here, but doing so is just silly.
2847
			 * We put oldsectors back because we *know* it is
2848 2849 2850
			 * safe, and trust userspace not to race with
			 * itself
			 */
2851
			rdev->sectors = oldsectors;
2852 2853 2854
			return -EBUSY;
		}
	}
2855 2856 2857 2858
	return len;
}

static struct rdev_sysfs_entry rdev_size =
2859
__ATTR(size, S_IRUGO|S_IWUSR, rdev_size_show, rdev_size_store);
2860

2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896

static ssize_t recovery_start_show(mdk_rdev_t *rdev, char *page)
{
	unsigned long long recovery_start = rdev->recovery_offset;

	if (test_bit(In_sync, &rdev->flags) ||
	    recovery_start == MaxSector)
		return sprintf(page, "none\n");

	return sprintf(page, "%llu\n", recovery_start);
}

static ssize_t recovery_start_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
	unsigned long long recovery_start;

	if (cmd_match(buf, "none"))
		recovery_start = MaxSector;
	else if (strict_strtoull(buf, 10, &recovery_start))
		return -EINVAL;

	if (rdev->mddev->pers &&
	    rdev->raid_disk >= 0)
		return -EBUSY;

	rdev->recovery_offset = recovery_start;
	if (recovery_start == MaxSector)
		set_bit(In_sync, &rdev->flags);
	else
		clear_bit(In_sync, &rdev->flags);
	return len;
}

static struct rdev_sysfs_entry rdev_recovery_start =
__ATTR(recovery_start, S_IRUGO|S_IWUSR, recovery_start_show, recovery_start_store);

2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908

static ssize_t
badblocks_show(struct badblocks *bb, char *page, int unack);
static ssize_t
badblocks_store(struct badblocks *bb, const char *page, size_t len, int unack);

static ssize_t bb_show(mdk_rdev_t *rdev, char *page)
{
	return badblocks_show(&rdev->badblocks, page, 0);
}
static ssize_t bb_store(mdk_rdev_t *rdev, const char *page, size_t len)
{
2909 2910 2911 2912 2913
	int rv = badblocks_store(&rdev->badblocks, page, len, 0);
	/* Maybe that ack was all we needed */
	if (test_and_clear_bit(BlockedBadBlocks, &rdev->flags))
		wake_up(&rdev->blocked_wait);
	return rv;
2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929
}
static struct rdev_sysfs_entry rdev_bad_blocks =
__ATTR(bad_blocks, S_IRUGO|S_IWUSR, bb_show, bb_store);


static ssize_t ubb_show(mdk_rdev_t *rdev, char *page)
{
	return badblocks_show(&rdev->badblocks, page, 1);
}
static ssize_t ubb_store(mdk_rdev_t *rdev, const char *page, size_t len)
{
	return badblocks_store(&rdev->badblocks, page, len, 1);
}
static struct rdev_sysfs_entry rdev_unack_bad_blocks =
__ATTR(unacknowledged_bad_blocks, S_IRUGO|S_IWUSR, ubb_show, ubb_store);

2930 2931
static struct attribute *rdev_default_attrs[] = {
	&rdev_state.attr,
2932
	&rdev_errors.attr,
2933
	&rdev_slot.attr,
2934
	&rdev_offset.attr,
2935
	&rdev_size.attr,
2936
	&rdev_recovery_start.attr,
2937 2938
	&rdev_bad_blocks.attr,
	&rdev_unack_bad_blocks.attr,
2939 2940 2941 2942 2943 2944 2945
	NULL,
};
static ssize_t
rdev_attr_show(struct kobject *kobj, struct attribute *attr, char *page)
{
	struct rdev_sysfs_entry *entry = container_of(attr, struct rdev_sysfs_entry, attr);
	mdk_rdev_t *rdev = container_of(kobj, mdk_rdev_t, kobj);
2946 2947
	mddev_t *mddev = rdev->mddev;
	ssize_t rv;
2948 2949 2950

	if (!entry->show)
		return -EIO;
2951 2952 2953 2954 2955 2956 2957 2958 2959 2960

	rv = mddev ? mddev_lock(mddev) : -EBUSY;
	if (!rv) {
		if (rdev->mddev == NULL)
			rv = -EBUSY;
		else
			rv = entry->show(rdev, page);
		mddev_unlock(mddev);
	}
	return rv;
2961 2962 2963 2964 2965 2966 2967 2968
}

static ssize_t
rdev_attr_store(struct kobject *kobj, struct attribute *attr,
	      const char *page, size_t length)
{
	struct rdev_sysfs_entry *entry = container_of(attr, struct rdev_sysfs_entry, attr);
	mdk_rdev_t *rdev = container_of(kobj, mdk_rdev_t, kobj);
2969 2970
	ssize_t rv;
	mddev_t *mddev = rdev->mddev;
2971 2972 2973

	if (!entry->store)
		return -EIO;
2974 2975
	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;
2976
	rv = mddev ? mddev_lock(mddev): -EBUSY;
2977
	if (!rv) {
2978 2979 2980 2981
		if (rdev->mddev == NULL)
			rv = -EBUSY;
		else
			rv = entry->store(rdev, page, length);
2982
		mddev_unlock(mddev);
2983 2984
	}
	return rv;
2985 2986 2987 2988 2989 2990 2991
}

static void rdev_free(struct kobject *ko)
{
	mdk_rdev_t *rdev = container_of(ko, mdk_rdev_t, kobj);
	kfree(rdev);
}
2992
static const struct sysfs_ops rdev_sysfs_ops = {
2993 2994 2995 2996 2997 2998 2999 3000 3001
	.show		= rdev_attr_show,
	.store		= rdev_attr_store,
};
static struct kobj_type rdev_ktype = {
	.release	= rdev_free,
	.sysfs_ops	= &rdev_sysfs_ops,
	.default_attrs	= rdev_default_attrs,
};

3002
int md_rdev_init(mdk_rdev_t *rdev)
N
NeilBrown 已提交
3003 3004 3005 3006 3007 3008 3009 3010 3011
{
	rdev->desc_nr = -1;
	rdev->saved_raid_disk = -1;
	rdev->raid_disk = -1;
	rdev->flags = 0;
	rdev->data_offset = 0;
	rdev->sb_events = 0;
	rdev->last_read_error.tv_sec  = 0;
	rdev->last_read_error.tv_nsec = 0;
3012 3013
	rdev->sb_loaded = 0;
	rdev->bb_page = NULL;
N
NeilBrown 已提交
3014 3015 3016 3017 3018 3019
	atomic_set(&rdev->nr_pending, 0);
	atomic_set(&rdev->read_errors, 0);
	atomic_set(&rdev->corrected_errors, 0);

	INIT_LIST_HEAD(&rdev->same_set);
	init_waitqueue_head(&rdev->blocked_wait);
3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032

	/* Add space to store bad block list.
	 * This reserves the space even on arrays where it cannot
	 * be used - I wonder if that matters
	 */
	rdev->badblocks.count = 0;
	rdev->badblocks.shift = 0;
	rdev->badblocks.page = kmalloc(PAGE_SIZE, GFP_KERNEL);
	seqlock_init(&rdev->badblocks.lock);
	if (rdev->badblocks.page == NULL)
		return -ENOMEM;

	return 0;
N
NeilBrown 已提交
3033 3034
}
EXPORT_SYMBOL_GPL(md_rdev_init);
L
Linus Torvalds 已提交
3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051
/*
 * Import a device. If 'super_format' >= 0, then sanity check the superblock
 *
 * mark the device faulty if:
 *
 *   - the device is nonexistent (zero size)
 *   - the device has no valid superblock
 *
 * a faulty rdev _never_ has rdev->sb set.
 */
static mdk_rdev_t *md_import_device(dev_t newdev, int super_format, int super_minor)
{
	char b[BDEVNAME_SIZE];
	int err;
	mdk_rdev_t *rdev;
	sector_t size;

3052
	rdev = kzalloc(sizeof(*rdev), GFP_KERNEL);
L
Linus Torvalds 已提交
3053 3054 3055 3056 3057
	if (!rdev) {
		printk(KERN_ERR "md: could not alloc mem for new device!\n");
		return ERR_PTR(-ENOMEM);
	}

3058 3059 3060 3061 3062
	err = md_rdev_init(rdev);
	if (err)
		goto abort_free;
	err = alloc_disk_sb(rdev);
	if (err)
L
Linus Torvalds 已提交
3063 3064
		goto abort_free;

3065
	err = lock_rdev(rdev, newdev, super_format == -2);
L
Linus Torvalds 已提交
3066 3067 3068
	if (err)
		goto abort_free;

3069
	kobject_init(&rdev->kobj, &rdev_ktype);
3070

3071
	size = i_size_read(rdev->bdev->bd_inode) >> BLOCK_SIZE_BITS;
L
Linus Torvalds 已提交
3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083
	if (!size) {
		printk(KERN_WARNING 
			"md: %s has zero or unknown size, marking faulty!\n",
			bdevname(rdev->bdev,b));
		err = -EINVAL;
		goto abort_free;
	}

	if (super_format >= 0) {
		err = super_types[super_format].
			load_super(rdev, NULL, super_minor);
		if (err == -EINVAL) {
3084 3085 3086 3087 3088
			printk(KERN_WARNING
				"md: %s does not have a valid v%d.%d "
			       "superblock, not importing!\n",
				bdevname(rdev->bdev,b),
			       super_format, super_minor);
L
Linus Torvalds 已提交
3089 3090 3091 3092 3093 3094 3095 3096 3097
			goto abort_free;
		}
		if (err < 0) {
			printk(KERN_WARNING 
				"md: could not read %s's sb, not importing!\n",
				bdevname(rdev->bdev,b));
			goto abort_free;
		}
	}
3098 3099 3100
	if (super_format == -1)
		/* hot-add for 0.90, or non-persistent: so no badblocks */
		rdev->badblocks.shift = -1;
3101

L
Linus Torvalds 已提交
3102 3103 3104
	return rdev;

abort_free:
3105 3106 3107
	if (rdev->bdev)
		unlock_rdev(rdev);
	free_disk_sb(rdev);
3108
	kfree(rdev->badblocks.page);
L
Linus Torvalds 已提交
3109 3110 3111 3112 3113 3114 3115 3116 3117
	kfree(rdev);
	return ERR_PTR(err);
}

/*
 * Check a full RAID array for plausibility
 */


3118
static void analyze_sbs(mddev_t * mddev)
L
Linus Torvalds 已提交
3119 3120
{
	int i;
3121
	mdk_rdev_t *rdev, *freshest, *tmp;
L
Linus Torvalds 已提交
3122 3123 3124
	char b[BDEVNAME_SIZE];

	freshest = NULL;
3125
	rdev_for_each(rdev, tmp, mddev)
L
Linus Torvalds 已提交
3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145
		switch (super_types[mddev->major_version].
			load_super(rdev, freshest, mddev->minor_version)) {
		case 1:
			freshest = rdev;
			break;
		case 0:
			break;
		default:
			printk( KERN_ERR \
				"md: fatal superblock inconsistency in %s"
				" -- removing from array\n", 
				bdevname(rdev->bdev,b));
			kick_rdev_from_array(rdev);
		}


	super_types[mddev->major_version].
		validate_super(mddev, freshest);

	i = 0;
3146
	rdev_for_each(rdev, tmp, mddev) {
3147 3148 3149
		if (mddev->max_disks &&
		    (rdev->desc_nr >= mddev->max_disks ||
		     i > mddev->max_disks)) {
3150 3151 3152 3153 3154 3155 3156
			printk(KERN_WARNING
			       "md: %s: %s: only %d devices permitted\n",
			       mdname(mddev), bdevname(rdev->bdev, b),
			       mddev->max_disks);
			kick_rdev_from_array(rdev);
			continue;
		}
L
Linus Torvalds 已提交
3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168
		if (rdev != freshest)
			if (super_types[mddev->major_version].
			    validate_super(mddev, rdev)) {
				printk(KERN_WARNING "md: kicking non-fresh %s"
					" from array!\n",
					bdevname(rdev->bdev,b));
				kick_rdev_from_array(rdev);
				continue;
			}
		if (mddev->level == LEVEL_MULTIPATH) {
			rdev->desc_nr = i++;
			rdev->raid_disk = rdev->desc_nr;
3169
			set_bit(In_sync, &rdev->flags);
3170
		} else if (rdev->raid_disk >= (mddev->raid_disks - min(0, mddev->delta_disks))) {
3171 3172
			rdev->raid_disk = -1;
			clear_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
3173 3174 3175 3176
		}
	}
}

3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217
/* Read a fixed-point number.
 * Numbers in sysfs attributes should be in "standard" units where
 * possible, so time should be in seconds.
 * However we internally use a a much smaller unit such as 
 * milliseconds or jiffies.
 * This function takes a decimal number with a possible fractional
 * component, and produces an integer which is the result of
 * multiplying that number by 10^'scale'.
 * all without any floating-point arithmetic.
 */
int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale)
{
	unsigned long result = 0;
	long decimals = -1;
	while (isdigit(*cp) || (*cp == '.' && decimals < 0)) {
		if (*cp == '.')
			decimals = 0;
		else if (decimals < scale) {
			unsigned int value;
			value = *cp - '0';
			result = result * 10 + value;
			if (decimals >= 0)
				decimals++;
		}
		cp++;
	}
	if (*cp == '\n')
		cp++;
	if (*cp)
		return -EINVAL;
	if (decimals < 0)
		decimals = 0;
	while (decimals < scale) {
		result *= 10;
		decimals ++;
	}
	*res = result;
	return 0;
}


3218 3219
static void md_safemode_timeout(unsigned long data);

3220 3221 3222 3223 3224 3225 3226 3227 3228 3229
static ssize_t
safe_delay_show(mddev_t *mddev, char *page)
{
	int msec = (mddev->safemode_delay*1000)/HZ;
	return sprintf(page, "%d.%03d\n", msec/1000, msec%1000);
}
static ssize_t
safe_delay_store(mddev_t *mddev, const char *cbuf, size_t len)
{
	unsigned long msec;
3230

3231
	if (strict_strtoul_scaled(cbuf, &msec, 3) < 0)
3232 3233 3234 3235
		return -EINVAL;
	if (msec == 0)
		mddev->safemode_delay = 0;
	else {
3236
		unsigned long old_delay = mddev->safemode_delay;
3237 3238 3239
		mddev->safemode_delay = (msec*HZ)/1000;
		if (mddev->safemode_delay == 0)
			mddev->safemode_delay = 1;
3240 3241
		if (mddev->safemode_delay < old_delay)
			md_safemode_timeout((unsigned long)mddev);
3242 3243 3244 3245
	}
	return len;
}
static struct md_sysfs_entry md_safe_delay =
3246
__ATTR(safe_mode_delay, S_IRUGO|S_IWUSR,safe_delay_show, safe_delay_store);
3247

3248
static ssize_t
3249
level_show(mddev_t *mddev, char *page)
3250
{
3251
	struct mdk_personality *p = mddev->pers;
3252
	if (p)
3253
		return sprintf(page, "%s\n", p->name);
3254 3255 3256 3257 3258 3259
	else if (mddev->clevel[0])
		return sprintf(page, "%s\n", mddev->clevel);
	else if (mddev->level != LEVEL_NONE)
		return sprintf(page, "%d\n", mddev->level);
	else
		return 0;
3260 3261
}

3262 3263 3264
static ssize_t
level_store(mddev_t *mddev, const char *buf, size_t len)
{
3265
	char clevel[16];
3266
	ssize_t rv = len;
3267
	struct mdk_personality *pers;
3268
	long level;
3269
	void *priv;
3270
	mdk_rdev_t *rdev;
3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290

	if (mddev->pers == NULL) {
		if (len == 0)
			return 0;
		if (len >= sizeof(mddev->clevel))
			return -ENOSPC;
		strncpy(mddev->clevel, buf, len);
		if (mddev->clevel[len-1] == '\n')
			len--;
		mddev->clevel[len] = 0;
		mddev->level = LEVEL_NONE;
		return rv;
	}

	/* request to change the personality.  Need to ensure:
	 *  - array is not engaged in resync/recovery/reshape
	 *  - old personality can be suspended
	 *  - new personality will access other array.
	 */

3291 3292 3293
	if (mddev->sync_thread ||
	    mddev->reshape_position != MaxSector ||
	    mddev->sysfs_active)
3294
		return -EBUSY;
3295 3296 3297 3298 3299 3300 3301 3302

	if (!mddev->pers->quiesce) {
		printk(KERN_WARNING "md: %s: %s does not support online personality change\n",
		       mdname(mddev), mddev->pers->name);
		return -EINVAL;
	}

	/* Now find the new personality */
3303
	if (len == 0 || len >= sizeof(clevel))
3304
		return -EINVAL;
3305 3306
	strncpy(clevel, buf, len);
	if (clevel[len-1] == '\n')
3307
		len--;
3308 3309 3310
	clevel[len] = 0;
	if (strict_strtol(clevel, 10, &level))
		level = LEVEL_NONE;
3311

3312 3313
	if (request_module("md-%s", clevel) != 0)
		request_module("md-level-%s", clevel);
3314
	spin_lock(&pers_lock);
3315
	pers = find_pers(level, clevel);
3316 3317
	if (!pers || !try_module_get(pers->owner)) {
		spin_unlock(&pers_lock);
3318
		printk(KERN_WARNING "md: personality %s not loaded\n", clevel);
3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330
		return -EINVAL;
	}
	spin_unlock(&pers_lock);

	if (pers == mddev->pers) {
		/* Nothing to do! */
		module_put(pers->owner);
		return rv;
	}
	if (!pers->takeover) {
		module_put(pers->owner);
		printk(KERN_WARNING "md: %s: %s does not support personality takeover\n",
3331
		       mdname(mddev), clevel);
3332 3333 3334
		return -EINVAL;
	}

3335 3336 3337
	list_for_each_entry(rdev, &mddev->disks, same_set)
		rdev->new_raid_disk = rdev->raid_disk;

3338 3339 3340 3341 3342 3343 3344
	/* ->takeover must set new_* and/or delta_disks
	 * if it succeeds, and may set them when it fails.
	 */
	priv = pers->takeover(mddev);
	if (IS_ERR(priv)) {
		mddev->new_level = mddev->level;
		mddev->new_layout = mddev->layout;
3345
		mddev->new_chunk_sectors = mddev->chunk_sectors;
3346 3347 3348 3349
		mddev->raid_disks -= mddev->delta_disks;
		mddev->delta_disks = 0;
		module_put(pers->owner);
		printk(KERN_WARNING "md: %s: %s would not accept array\n",
3350
		       mdname(mddev), clevel);
3351 3352 3353 3354 3355 3356
		return PTR_ERR(priv);
	}

	/* Looks like we have a winner */
	mddev_suspend(mddev);
	mddev->pers->stop(mddev);
3357 3358 3359 3360 3361 3362 3363 3364
	
	if (mddev->pers->sync_request == NULL &&
	    pers->sync_request != NULL) {
		/* need to add the md_redundancy_group */
		if (sysfs_create_group(&mddev->kobj, &md_redundancy_group))
			printk(KERN_WARNING
			       "md: cannot register extra attributes for %s\n",
			       mdname(mddev));
3365
		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, NULL, "sync_action");
3366 3367 3368 3369 3370 3371 3372 3373
	}		
	if (mddev->pers->sync_request != NULL &&
	    pers->sync_request == NULL) {
		/* need to remove the md_redundancy_group */
		if (mddev->to_remove == NULL)
			mddev->to_remove = &md_redundancy_group;
	}

3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387
	if (mddev->pers->sync_request == NULL &&
	    mddev->external) {
		/* We are converting from a no-redundancy array
		 * to a redundancy array and metadata is managed
		 * externally so we need to be sure that writes
		 * won't block due to a need to transition
		 *      clean->dirty
		 * until external management is started.
		 */
		mddev->in_sync = 0;
		mddev->safemode_delay = 0;
		mddev->safemode = 0;
	}

3388 3389 3390
	list_for_each_entry(rdev, &mddev->disks, same_set) {
		if (rdev->raid_disk < 0)
			continue;
3391
		if (rdev->new_raid_disk >= mddev->raid_disks)
3392 3393 3394
			rdev->new_raid_disk = -1;
		if (rdev->new_raid_disk == rdev->raid_disk)
			continue;
3395
		sysfs_unlink_rdev(mddev, rdev);
3396 3397 3398 3399 3400 3401 3402 3403
	}
	list_for_each_entry(rdev, &mddev->disks, same_set) {
		if (rdev->raid_disk < 0)
			continue;
		if (rdev->new_raid_disk == rdev->raid_disk)
			continue;
		rdev->raid_disk = rdev->new_raid_disk;
		if (rdev->raid_disk < 0)
3404
			clear_bit(In_sync, &rdev->flags);
3405
		else {
3406 3407 3408 3409
			if (sysfs_link_rdev(mddev, rdev))
				printk(KERN_WARNING "md: cannot register rd%d"
				       " for %s after level change\n",
				       rdev->raid_disk, mdname(mddev));
3410
		}
3411 3412 3413
	}

	module_put(mddev->pers->owner);
3414 3415 3416 3417 3418
	mddev->pers = pers;
	mddev->private = priv;
	strlcpy(mddev->clevel, pers->name, sizeof(mddev->clevel));
	mddev->level = mddev->new_level;
	mddev->layout = mddev->new_layout;
3419
	mddev->chunk_sectors = mddev->new_chunk_sectors;
3420
	mddev->delta_disks = 0;
3421
	mddev->degraded = 0;
3422 3423 3424 3425 3426 3427 3428
	if (mddev->pers->sync_request == NULL) {
		/* this is now an array without redundancy, so
		 * it must always be in_sync
		 */
		mddev->in_sync = 1;
		del_timer_sync(&mddev->safemode_timer);
	}
3429 3430 3431 3432 3433
	pers->run(mddev);
	mddev_resume(mddev);
	set_bit(MD_CHANGE_DEVS, &mddev->flags);
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
3434
	sysfs_notify(&mddev->kobj, NULL, "level");
3435
	md_new_event(mddev);
3436 3437 3438 3439
	return rv;
}

static struct md_sysfs_entry md_level =
3440
__ATTR(level, S_IRUGO|S_IWUSR, level_show, level_store);
3441

3442 3443 3444 3445 3446

static ssize_t
layout_show(mddev_t *mddev, char *page)
{
	/* just a number, not meaningful for all levels */
3447 3448 3449 3450
	if (mddev->reshape_position != MaxSector &&
	    mddev->layout != mddev->new_layout)
		return sprintf(page, "%d (%d)\n",
			       mddev->new_layout, mddev->layout);
3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462
	return sprintf(page, "%d\n", mddev->layout);
}

static ssize_t
layout_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long n = simple_strtoul(buf, &e, 10);

	if (!*buf || (*e && *e != '\n'))
		return -EINVAL;

3463 3464
	if (mddev->pers) {
		int err;
3465
		if (mddev->pers->check_reshape == NULL)
3466
			return -EBUSY;
3467
		mddev->new_layout = n;
3468
		err = mddev->pers->check_reshape(mddev);
3469 3470
		if (err) {
			mddev->new_layout = mddev->layout;
3471
			return err;
3472
		}
3473
	} else {
3474
		mddev->new_layout = n;
3475 3476 3477
		if (mddev->reshape_position == MaxSector)
			mddev->layout = n;
	}
3478 3479 3480
	return len;
}
static struct md_sysfs_entry md_layout =
3481
__ATTR(layout, S_IRUGO|S_IWUSR, layout_show, layout_store);
3482 3483


3484
static ssize_t
3485
raid_disks_show(mddev_t *mddev, char *page)
3486
{
3487 3488
	if (mddev->raid_disks == 0)
		return 0;
3489 3490 3491 3492
	if (mddev->reshape_position != MaxSector &&
	    mddev->delta_disks != 0)
		return sprintf(page, "%d (%d)\n", mddev->raid_disks,
			       mddev->raid_disks - mddev->delta_disks);
3493 3494 3495
	return sprintf(page, "%d\n", mddev->raid_disks);
}

3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509
static int update_raid_disks(mddev_t *mddev, int raid_disks);

static ssize_t
raid_disks_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	int rv = 0;
	unsigned long n = simple_strtoul(buf, &e, 10);

	if (!*buf || (*e && *e != '\n'))
		return -EINVAL;

	if (mddev->pers)
		rv = update_raid_disks(mddev, n);
3510 3511 3512 3513 3514
	else if (mddev->reshape_position != MaxSector) {
		int olddisks = mddev->raid_disks - mddev->delta_disks;
		mddev->delta_disks = n - olddisks;
		mddev->raid_disks = n;
	} else
3515 3516 3517 3518
		mddev->raid_disks = n;
	return rv ? rv : len;
}
static struct md_sysfs_entry md_raid_disks =
3519
__ATTR(raid_disks, S_IRUGO|S_IWUSR, raid_disks_show, raid_disks_store);
3520

3521 3522 3523
static ssize_t
chunk_size_show(mddev_t *mddev, char *page)
{
3524
	if (mddev->reshape_position != MaxSector &&
3525 3526 3527
	    mddev->chunk_sectors != mddev->new_chunk_sectors)
		return sprintf(page, "%d (%d)\n",
			       mddev->new_chunk_sectors << 9,
3528 3529
			       mddev->chunk_sectors << 9);
	return sprintf(page, "%d\n", mddev->chunk_sectors << 9);
3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540
}

static ssize_t
chunk_size_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long n = simple_strtoul(buf, &e, 10);

	if (!*buf || (*e && *e != '\n'))
		return -EINVAL;

3541 3542
	if (mddev->pers) {
		int err;
3543
		if (mddev->pers->check_reshape == NULL)
3544
			return -EBUSY;
3545
		mddev->new_chunk_sectors = n >> 9;
3546
		err = mddev->pers->check_reshape(mddev);
3547 3548
		if (err) {
			mddev->new_chunk_sectors = mddev->chunk_sectors;
3549
			return err;
3550
		}
3551
	} else {
3552
		mddev->new_chunk_sectors = n >> 9;
3553
		if (mddev->reshape_position == MaxSector)
3554
			mddev->chunk_sectors = n >> 9;
3555
	}
3556 3557 3558
	return len;
}
static struct md_sysfs_entry md_chunk_size =
3559
__ATTR(chunk_size, S_IRUGO|S_IWUSR, chunk_size_show, chunk_size_store);
3560

3561 3562 3563
static ssize_t
resync_start_show(mddev_t *mddev, char *page)
{
3564 3565
	if (mddev->recovery_cp == MaxSector)
		return sprintf(page, "none\n");
3566 3567 3568 3569 3570 3571 3572 3573 3574
	return sprintf(page, "%llu\n", (unsigned long long)mddev->recovery_cp);
}

static ssize_t
resync_start_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long long n = simple_strtoull(buf, &e, 10);

3575
	if (mddev->pers && !test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))
3576
		return -EBUSY;
3577 3578 3579
	if (cmd_match(buf, "none"))
		n = MaxSector;
	else if (!*buf || (*e && *e != '\n'))
3580 3581 3582 3583 3584 3585
		return -EINVAL;

	mddev->recovery_cp = n;
	return len;
}
static struct md_sysfs_entry md_resync_start =
3586
__ATTR(resync_start, S_IRUGO|S_IWUSR, resync_start_show, resync_start_store);
3587

3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599
/*
 * The array state can be:
 *
 * clear
 *     No devices, no size, no level
 *     Equivalent to STOP_ARRAY ioctl
 * inactive
 *     May have some settings, but array is not active
 *        all IO results in error
 *     When written, doesn't tear down array, but just stops it
 * suspended (not supported yet)
 *     All IO requests will block. The array can be reconfigured.
3600
 *     Writing this, if accepted, will block until array is quiescent
3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625
 * readonly
 *     no resync can happen.  no superblocks get written.
 *     write requests fail
 * read-auto
 *     like readonly, but behaves like 'clean' on a write request.
 *
 * clean - no pending writes, but otherwise active.
 *     When written to inactive array, starts without resync
 *     If a write request arrives then
 *       if metadata is known, mark 'dirty' and switch to 'active'.
 *       if not known, block and switch to write-pending
 *     If written to an active array that has pending writes, then fails.
 * active
 *     fully active: IO and resync can be happening.
 *     When written to inactive array, starts with resync
 *
 * write-pending
 *     clean, but writes are blocked waiting for 'active' to be written.
 *
 * active-idle
 *     like active, but no writes have been seen for a while (100msec).
 *
 */
enum array_state { clear, inactive, suspended, readonly, read_auto, clean, active,
		   write_pending, active_idle, bad_word};
3626
static char *array_states[] = {
3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654
	"clear", "inactive", "suspended", "readonly", "read-auto", "clean", "active",
	"write-pending", "active-idle", NULL };

static int match_word(const char *word, char **list)
{
	int n;
	for (n=0; list[n]; n++)
		if (cmd_match(word, list[n]))
			break;
	return n;
}

static ssize_t
array_state_show(mddev_t *mddev, char *page)
{
	enum array_state st = inactive;

	if (mddev->pers)
		switch(mddev->ro) {
		case 1:
			st = readonly;
			break;
		case 2:
			st = read_auto;
			break;
		case 0:
			if (mddev->in_sync)
				st = clean;
3655
			else if (test_bit(MD_CHANGE_PENDING, &mddev->flags))
3656
				st = write_pending;
3657 3658 3659 3660 3661 3662 3663 3664
			else if (mddev->safemode)
				st = active_idle;
			else
				st = active;
		}
	else {
		if (list_empty(&mddev->disks) &&
		    mddev->raid_disks == 0 &&
3665
		    mddev->dev_sectors == 0)
3666 3667 3668 3669 3670 3671 3672
			st = clear;
		else
			st = inactive;
	}
	return sprintf(page, "%s\n", array_states[st]);
}

3673
static int do_md_stop(mddev_t * mddev, int ro, int is_open);
3674
static int md_set_readonly(mddev_t * mddev, int is_open);
3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687
static int do_md_run(mddev_t * mddev);
static int restart_array(mddev_t *mddev);

static ssize_t
array_state_store(mddev_t *mddev, const char *buf, size_t len)
{
	int err = -EINVAL;
	enum array_state st = match_word(buf, array_states);
	switch(st) {
	case bad_word:
		break;
	case clear:
		/* stopping an active array */
3688
		if (atomic_read(&mddev->openers) > 0)
3689
			return -EBUSY;
3690
		err = do_md_stop(mddev, 0, 0);
3691 3692 3693 3694
		break;
	case inactive:
		/* stopping an active array */
		if (mddev->pers) {
3695
			if (atomic_read(&mddev->openers) > 0)
3696
				return -EBUSY;
3697
			err = do_md_stop(mddev, 2, 0);
3698 3699
		} else
			err = 0; /* already inactive */
3700 3701 3702 3703 3704
		break;
	case suspended:
		break; /* not supported yet */
	case readonly:
		if (mddev->pers)
3705
			err = md_set_readonly(mddev, 0);
3706 3707
		else {
			mddev->ro = 1;
3708
			set_disk_ro(mddev->gendisk, 1);
3709 3710 3711 3712 3713
			err = do_md_run(mddev);
		}
		break;
	case read_auto:
		if (mddev->pers) {
3714
			if (mddev->ro == 0)
3715
				err = md_set_readonly(mddev, 0);
3716
			else if (mddev->ro == 1)
3717 3718 3719 3720 3721
				err = restart_array(mddev);
			if (err == 0) {
				mddev->ro = 2;
				set_disk_ro(mddev->gendisk, 0);
			}
3722 3723 3724 3725 3726 3727 3728 3729 3730 3731
		} else {
			mddev->ro = 2;
			err = do_md_run(mddev);
		}
		break;
	case clean:
		if (mddev->pers) {
			restart_array(mddev);
			spin_lock_irq(&mddev->write_lock);
			if (atomic_read(&mddev->writes_pending) == 0) {
3732 3733
				if (mddev->in_sync == 0) {
					mddev->in_sync = 1;
3734 3735
					if (mddev->safemode == 1)
						mddev->safemode = 0;
3736
					set_bit(MD_CHANGE_CLEAN, &mddev->flags);
3737 3738 3739 3740
				}
				err = 0;
			} else
				err = -EBUSY;
3741
			spin_unlock_irq(&mddev->write_lock);
3742 3743
		} else
			err = -EINVAL;
3744 3745 3746 3747
		break;
	case active:
		if (mddev->pers) {
			restart_array(mddev);
3748
			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
3749 3750 3751 3752
			wake_up(&mddev->sb_wait);
			err = 0;
		} else {
			mddev->ro = 0;
3753
			set_disk_ro(mddev->gendisk, 0);
3754 3755 3756 3757 3758 3759 3760 3761 3762 3763
			err = do_md_run(mddev);
		}
		break;
	case write_pending:
	case active_idle:
		/* these cannot be set */
		break;
	}
	if (err)
		return err;
3764
	else {
3765
		sysfs_notify_dirent_safe(mddev->sysfs_state);
3766
		return len;
3767
	}
3768
}
3769 3770
static struct md_sysfs_entry md_array_state =
__ATTR(array_state, S_IRUGO|S_IWUSR, array_state_show, array_state_store);
3771

3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794
static ssize_t
max_corrected_read_errors_show(mddev_t *mddev, char *page) {
	return sprintf(page, "%d\n",
		       atomic_read(&mddev->max_corr_read_errors));
}

static ssize_t
max_corrected_read_errors_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long n = simple_strtoul(buf, &e, 10);

	if (*buf && (*e == 0 || *e == '\n')) {
		atomic_set(&mddev->max_corr_read_errors, n);
		return len;
	}
	return -EINVAL;
}

static struct md_sysfs_entry max_corr_read_errors =
__ATTR(max_read_errors, S_IRUGO|S_IWUSR, max_corrected_read_errors_show,
	max_corrected_read_errors_store);

3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839
static ssize_t
null_show(mddev_t *mddev, char *page)
{
	return -EINVAL;
}

static ssize_t
new_dev_store(mddev_t *mddev, const char *buf, size_t len)
{
	/* buf must be %d:%d\n? giving major and minor numbers */
	/* The new device is added to the array.
	 * If the array has a persistent superblock, we read the
	 * superblock to initialise info and check validity.
	 * Otherwise, only checking done is that in bind_rdev_to_array,
	 * which mainly checks size.
	 */
	char *e;
	int major = simple_strtoul(buf, &e, 10);
	int minor;
	dev_t dev;
	mdk_rdev_t *rdev;
	int err;

	if (!*buf || *e != ':' || !e[1] || e[1] == '\n')
		return -EINVAL;
	minor = simple_strtoul(e+1, &e, 10);
	if (*e && *e != '\n')
		return -EINVAL;
	dev = MKDEV(major, minor);
	if (major != MAJOR(dev) ||
	    minor != MINOR(dev))
		return -EOVERFLOW;


	if (mddev->persistent) {
		rdev = md_import_device(dev, mddev->major_version,
					mddev->minor_version);
		if (!IS_ERR(rdev) && !list_empty(&mddev->disks)) {
			mdk_rdev_t *rdev0 = list_entry(mddev->disks.next,
						       mdk_rdev_t, same_set);
			err = super_types[mddev->major_version]
				.load_super(rdev, rdev0, mddev->minor_version);
			if (err < 0)
				goto out;
		}
3840 3841 3842
	} else if (mddev->external)
		rdev = md_import_device(dev, -2, -1);
	else
3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854
		rdev = md_import_device(dev, -1, -1);

	if (IS_ERR(rdev))
		return PTR_ERR(rdev);
	err = bind_rdev_to_array(rdev, mddev);
 out:
	if (err)
		export_rdev(rdev);
	return err ? err : len;
}

static struct md_sysfs_entry md_new_device =
3855
__ATTR(new_dev, S_IWUSR, null_show, new_dev_store);
3856

3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875
static ssize_t
bitmap_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *end;
	unsigned long chunk, end_chunk;

	if (!mddev->bitmap)
		goto out;
	/* buf should be <chunk> <chunk> ... or <chunk>-<chunk> ... (range) */
	while (*buf) {
		chunk = end_chunk = simple_strtoul(buf, &end, 0);
		if (buf == end) break;
		if (*end == '-') { /* range */
			buf = end + 1;
			end_chunk = simple_strtoul(buf, &end, 0);
			if (buf == end) break;
		}
		if (*end && !isspace(*end)) break;
		bitmap_dirty_bits(mddev->bitmap, chunk, end_chunk);
3876
		buf = skip_spaces(end);
3877 3878 3879 3880 3881 3882 3883 3884 3885
	}
	bitmap_unplug(mddev->bitmap); /* flush the bits to disk */
out:
	return len;
}

static struct md_sysfs_entry md_bitmap =
__ATTR(bitmap_set_bits, S_IWUSR, null_show, bitmap_store);

3886 3887 3888
static ssize_t
size_show(mddev_t *mddev, char *page)
{
3889 3890
	return sprintf(page, "%llu\n",
		(unsigned long long)mddev->dev_sectors / 2);
3891 3892
}

3893
static int update_size(mddev_t *mddev, sector_t num_sectors);
3894 3895 3896 3897 3898 3899 3900 3901

static ssize_t
size_store(mddev_t *mddev, const char *buf, size_t len)
{
	/* If array is inactive, we can reduce the component size, but
	 * not increase it (except from 0).
	 * If array is active, we can try an on-line resize
	 */
3902 3903
	sector_t sectors;
	int err = strict_blocks_to_sectors(buf, &sectors);
3904

3905 3906
	if (err < 0)
		return err;
3907
	if (mddev->pers) {
3908
		err = update_size(mddev, sectors);
3909
		md_update_sb(mddev, 1);
3910
	} else {
3911 3912 3913
		if (mddev->dev_sectors == 0 ||
		    mddev->dev_sectors > sectors)
			mddev->dev_sectors = sectors;
3914 3915 3916 3917 3918 3919 3920
		else
			err = -ENOSPC;
	}
	return err ? err : len;
}

static struct md_sysfs_entry md_size =
3921
__ATTR(component_size, S_IRUGO|S_IWUSR, size_show, size_store);
3922

3923 3924

/* Metdata version.
3925 3926 3927
 * This is one of
 *   'none' for arrays with no metadata (good luck...)
 *   'external' for arrays with externally managed metadata,
3928 3929 3930 3931 3932 3933 3934 3935
 * or N.M for internally known formats
 */
static ssize_t
metadata_show(mddev_t *mddev, char *page)
{
	if (mddev->persistent)
		return sprintf(page, "%d.%d\n",
			       mddev->major_version, mddev->minor_version);
3936 3937
	else if (mddev->external)
		return sprintf(page, "external:%s\n", mddev->metadata_type);
3938 3939 3940 3941 3942 3943 3944 3945 3946
	else
		return sprintf(page, "none\n");
}

static ssize_t
metadata_store(mddev_t *mddev, const char *buf, size_t len)
{
	int major, minor;
	char *e;
3947 3948 3949 3950 3951 3952 3953
	/* Changing the details of 'external' metadata is
	 * always permitted.  Otherwise there must be
	 * no devices attached to the array.
	 */
	if (mddev->external && strncmp(buf, "external:", 9) == 0)
		;
	else if (!list_empty(&mddev->disks))
3954 3955 3956 3957
		return -EBUSY;

	if (cmd_match(buf, "none")) {
		mddev->persistent = 0;
3958 3959 3960 3961 3962 3963
		mddev->external = 0;
		mddev->major_version = 0;
		mddev->minor_version = 90;
		return len;
	}
	if (strncmp(buf, "external:", 9) == 0) {
3964
		size_t namelen = len-9;
3965 3966 3967 3968 3969 3970 3971 3972
		if (namelen >= sizeof(mddev->metadata_type))
			namelen = sizeof(mddev->metadata_type)-1;
		strncpy(mddev->metadata_type, buf+9, namelen);
		mddev->metadata_type[namelen] = 0;
		if (namelen && mddev->metadata_type[namelen-1] == '\n')
			mddev->metadata_type[--namelen] = 0;
		mddev->persistent = 0;
		mddev->external = 1;
3973 3974 3975 3976 3977 3978 3979 3980 3981
		mddev->major_version = 0;
		mddev->minor_version = 90;
		return len;
	}
	major = simple_strtoul(buf, &e, 10);
	if (e==buf || *e != '.')
		return -EINVAL;
	buf = e+1;
	minor = simple_strtoul(buf, &e, 10);
3982
	if (e==buf || (*e && *e != '\n') )
3983
		return -EINVAL;
3984
	if (major >= ARRAY_SIZE(super_types) || super_types[major].name == NULL)
3985 3986 3987 3988
		return -ENOENT;
	mddev->major_version = major;
	mddev->minor_version = minor;
	mddev->persistent = 1;
3989
	mddev->external = 0;
3990 3991 3992 3993
	return len;
}

static struct md_sysfs_entry md_metadata =
3994
__ATTR(metadata_version, S_IRUGO|S_IWUSR, metadata_show, metadata_store);
3995

3996
static ssize_t
3997
action_show(mddev_t *mddev, char *page)
3998
{
3999
	char *type = "idle";
4000 4001 4002
	if (test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))
		type = "frozen";
	else if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) ||
4003
	    (!mddev->ro && test_bit(MD_RECOVERY_NEEDED, &mddev->recovery))) {
4004 4005 4006
		if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
			type = "reshape";
		else if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
4007 4008 4009 4010 4011 4012
			if (!test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
				type = "resync";
			else if (test_bit(MD_RECOVERY_CHECK, &mddev->recovery))
				type = "check";
			else
				type = "repair";
4013
		} else if (test_bit(MD_RECOVERY_RECOVER, &mddev->recovery))
4014 4015 4016 4017 4018
			type = "recover";
	}
	return sprintf(page, "%s\n", type);
}

4019 4020
static void reap_sync_thread(mddev_t *mddev);

4021
static ssize_t
4022
action_store(mddev_t *mddev, const char *page, size_t len)
4023
{
4024 4025 4026
	if (!mddev->pers || !mddev->pers->sync_request)
		return -EINVAL;

4027 4028 4029 4030 4031 4032
	if (cmd_match(page, "frozen"))
		set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
	else
		clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);

	if (cmd_match(page, "idle") || cmd_match(page, "frozen")) {
4033 4034
		if (mddev->sync_thread) {
			set_bit(MD_RECOVERY_INTR, &mddev->recovery);
4035
			reap_sync_thread(mddev);
4036
		}
4037 4038
	} else if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) ||
		   test_bit(MD_RECOVERY_NEEDED, &mddev->recovery))
4039
		return -EBUSY;
4040 4041 4042 4043
	else if (cmd_match(page, "resync"))
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	else if (cmd_match(page, "recover")) {
		set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
4044
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
4045
	} else if (cmd_match(page, "reshape")) {
4046 4047 4048 4049 4050 4051
		int err;
		if (mddev->pers->start_reshape == NULL)
			return -EINVAL;
		err = mddev->pers->start_reshape(mddev);
		if (err)
			return err;
4052
		sysfs_notify(&mddev->kobj, NULL, "degraded");
4053
	} else {
4054
		if (cmd_match(page, "check"))
4055
			set_bit(MD_RECOVERY_CHECK, &mddev->recovery);
4056
		else if (!cmd_match(page, "repair"))
4057 4058 4059 4060
			return -EINVAL;
		set_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
		set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
	}
4061
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
4062
	md_wakeup_thread(mddev->thread);
4063
	sysfs_notify_dirent_safe(mddev->sysfs_action);
4064 4065 4066
	return len;
}

4067
static ssize_t
4068
mismatch_cnt_show(mddev_t *mddev, char *page)
4069 4070 4071 4072 4073
{
	return sprintf(page, "%llu\n",
		       (unsigned long long) mddev->resync_mismatches);
}

4074 4075
static struct md_sysfs_entry md_scan_mode =
__ATTR(sync_action, S_IRUGO|S_IWUSR, action_show, action_store);
4076

4077

4078
static struct md_sysfs_entry md_mismatches = __ATTR_RO(mismatch_cnt);
4079

4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131
static ssize_t
sync_min_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%d (%s)\n", speed_min(mddev),
		       mddev->sync_speed_min ? "local": "system");
}

static ssize_t
sync_min_store(mddev_t *mddev, const char *buf, size_t len)
{
	int min;
	char *e;
	if (strncmp(buf, "system", 6)==0) {
		mddev->sync_speed_min = 0;
		return len;
	}
	min = simple_strtoul(buf, &e, 10);
	if (buf == e || (*e && *e != '\n') || min <= 0)
		return -EINVAL;
	mddev->sync_speed_min = min;
	return len;
}

static struct md_sysfs_entry md_sync_min =
__ATTR(sync_speed_min, S_IRUGO|S_IWUSR, sync_min_show, sync_min_store);

static ssize_t
sync_max_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%d (%s)\n", speed_max(mddev),
		       mddev->sync_speed_max ? "local": "system");
}

static ssize_t
sync_max_store(mddev_t *mddev, const char *buf, size_t len)
{
	int max;
	char *e;
	if (strncmp(buf, "system", 6)==0) {
		mddev->sync_speed_max = 0;
		return len;
	}
	max = simple_strtoul(buf, &e, 10);
	if (buf == e || (*e && *e != '\n') || max <= 0)
		return -EINVAL;
	mddev->sync_speed_max = max;
	return len;
}

static struct md_sysfs_entry md_sync_max =
__ATTR(sync_speed_max, S_IRUGO|S_IWUSR, sync_max_show, sync_max_store);

4132 4133 4134 4135 4136 4137
static ssize_t
degraded_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%d\n", mddev->degraded);
}
static struct md_sysfs_entry md_degraded = __ATTR_RO(degraded);
4138

4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168
static ssize_t
sync_force_parallel_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%d\n", mddev->parallel_resync);
}

static ssize_t
sync_force_parallel_store(mddev_t *mddev, const char *buf, size_t len)
{
	long n;

	if (strict_strtol(buf, 10, &n))
		return -EINVAL;

	if (n != 0 && n != 1)
		return -EINVAL;

	mddev->parallel_resync = n;

	if (mddev->sync_thread)
		wake_up(&resync_wait);

	return len;
}

/* force parallel resync, even with shared block devices */
static struct md_sysfs_entry md_sync_force_parallel =
__ATTR(sync_force_parallel, S_IRUGO|S_IWUSR,
       sync_force_parallel_show, sync_force_parallel_store);

4169 4170 4171 4172
static ssize_t
sync_speed_show(mddev_t *mddev, char *page)
{
	unsigned long resync, dt, db;
4173 4174
	if (mddev->curr_resync == 0)
		return sprintf(page, "none\n");
4175 4176
	resync = mddev->curr_mark_cnt - atomic_read(&mddev->recovery_active);
	dt = (jiffies - mddev->resync_mark) / HZ;
4177
	if (!dt) dt++;
4178 4179
	db = resync - mddev->resync_mark_cnt;
	return sprintf(page, "%lu\n", db/dt/2); /* K/sec */
4180 4181
}

4182
static struct md_sysfs_entry md_sync_speed = __ATTR_RO(sync_speed);
4183 4184 4185 4186

static ssize_t
sync_completed_show(mddev_t *mddev, char *page)
{
4187
	unsigned long long max_sectors, resync;
4188

4189 4190 4191
	if (!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
		return sprintf(page, "none\n");

4192
	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
4193
		max_sectors = mddev->resync_max_sectors;
4194
	else
4195
		max_sectors = mddev->dev_sectors;
4196

4197
	resync = mddev->curr_resync_completed;
4198
	return sprintf(page, "%llu / %llu\n", resync, max_sectors);
4199 4200
}

4201
static struct md_sysfs_entry md_sync_completed = __ATTR_RO(sync_completed);
4202

4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220
static ssize_t
min_sync_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%llu\n",
		       (unsigned long long)mddev->resync_min);
}
static ssize_t
min_sync_store(mddev_t *mddev, const char *buf, size_t len)
{
	unsigned long long min;
	if (strict_strtoull(buf, 10, &min))
		return -EINVAL;
	if (min > mddev->resync_max)
		return -EINVAL;
	if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
		return -EBUSY;

	/* Must be a multiple of chunk_size */
4221
	if (mddev->chunk_sectors) {
4222
		sector_t temp = min;
4223
		if (sector_div(temp, mddev->chunk_sectors))
4224 4225 4226 4227 4228 4229 4230 4231 4232 4233
			return -EINVAL;
	}
	mddev->resync_min = min;

	return len;
}

static struct md_sysfs_entry md_min_sync =
__ATTR(sync_min, S_IRUGO|S_IWUSR, min_sync_show, min_sync_store);

4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248
static ssize_t
max_sync_show(mddev_t *mddev, char *page)
{
	if (mddev->resync_max == MaxSector)
		return sprintf(page, "max\n");
	else
		return sprintf(page, "%llu\n",
			       (unsigned long long)mddev->resync_max);
}
static ssize_t
max_sync_store(mddev_t *mddev, const char *buf, size_t len)
{
	if (strncmp(buf, "max", 3) == 0)
		mddev->resync_max = MaxSector;
	else {
4249 4250 4251 4252
		unsigned long long max;
		if (strict_strtoull(buf, 10, &max))
			return -EINVAL;
		if (max < mddev->resync_min)
4253 4254
			return -EINVAL;
		if (max < mddev->resync_max &&
4255
		    mddev->ro == 0 &&
4256 4257 4258 4259
		    test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
			return -EBUSY;

		/* Must be a multiple of chunk_size */
4260
		if (mddev->chunk_sectors) {
4261
			sector_t temp = max;
4262
			if (sector_div(temp, mddev->chunk_sectors))
4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273
				return -EINVAL;
		}
		mddev->resync_max = max;
	}
	wake_up(&mddev->recovery_wait);
	return len;
}

static struct md_sysfs_entry md_max_sync =
__ATTR(sync_max, S_IRUGO|S_IWUSR, max_sync_show, max_sync_store);

4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284
static ssize_t
suspend_lo_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%llu\n", (unsigned long long)mddev->suspend_lo);
}

static ssize_t
suspend_lo_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long long new = simple_strtoull(buf, &e, 10);
4285
	unsigned long long old = mddev->suspend_lo;
4286

4287 4288
	if (mddev->pers == NULL || 
	    mddev->pers->quiesce == NULL)
4289 4290 4291
		return -EINVAL;
	if (buf == e || (*e && *e != '\n'))
		return -EINVAL;
4292 4293 4294 4295

	mddev->suspend_lo = new;
	if (new >= old)
		/* Shrinking suspended region */
4296
		mddev->pers->quiesce(mddev, 2);
4297 4298 4299 4300 4301 4302
	else {
		/* Expanding suspended region - need to wait */
		mddev->pers->quiesce(mddev, 1);
		mddev->pers->quiesce(mddev, 0);
	}
	return len;
4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318
}
static struct md_sysfs_entry md_suspend_lo =
__ATTR(suspend_lo, S_IRUGO|S_IWUSR, suspend_lo_show, suspend_lo_store);


static ssize_t
suspend_hi_show(mddev_t *mddev, char *page)
{
	return sprintf(page, "%llu\n", (unsigned long long)mddev->suspend_hi);
}

static ssize_t
suspend_hi_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long long new = simple_strtoull(buf, &e, 10);
4319
	unsigned long long old = mddev->suspend_hi;
4320

4321 4322
	if (mddev->pers == NULL ||
	    mddev->pers->quiesce == NULL)
4323 4324 4325
		return -EINVAL;
	if (buf == e || (*e && *e != '\n'))
		return -EINVAL;
4326 4327 4328 4329 4330 4331 4332

	mddev->suspend_hi = new;
	if (new <= old)
		/* Shrinking suspended region */
		mddev->pers->quiesce(mddev, 2);
	else {
		/* Expanding suspended region - need to wait */
4333 4334
		mddev->pers->quiesce(mddev, 1);
		mddev->pers->quiesce(mddev, 0);
4335 4336
	}
	return len;
4337 4338 4339 4340
}
static struct md_sysfs_entry md_suspend_hi =
__ATTR(suspend_hi, S_IRUGO|S_IWUSR, suspend_hi_show, suspend_hi_store);

4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363
static ssize_t
reshape_position_show(mddev_t *mddev, char *page)
{
	if (mddev->reshape_position != MaxSector)
		return sprintf(page, "%llu\n",
			       (unsigned long long)mddev->reshape_position);
	strcpy(page, "none\n");
	return 5;
}

static ssize_t
reshape_position_store(mddev_t *mddev, const char *buf, size_t len)
{
	char *e;
	unsigned long long new = simple_strtoull(buf, &e, 10);
	if (mddev->pers)
		return -EBUSY;
	if (buf == e || (*e && *e != '\n'))
		return -EINVAL;
	mddev->reshape_position = new;
	mddev->delta_disks = 0;
	mddev->new_level = mddev->level;
	mddev->new_layout = mddev->layout;
4364
	mddev->new_chunk_sectors = mddev->chunk_sectors;
4365 4366 4367 4368 4369 4370 4371
	return len;
}

static struct md_sysfs_entry md_reshape_position =
__ATTR(reshape_position, S_IRUGO|S_IWUSR, reshape_position_show,
       reshape_position_store);

4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397
static ssize_t
array_size_show(mddev_t *mddev, char *page)
{
	if (mddev->external_size)
		return sprintf(page, "%llu\n",
			       (unsigned long long)mddev->array_sectors/2);
	else
		return sprintf(page, "default\n");
}

static ssize_t
array_size_store(mddev_t *mddev, const char *buf, size_t len)
{
	sector_t sectors;

	if (strncmp(buf, "default", 7) == 0) {
		if (mddev->pers)
			sectors = mddev->pers->size(mddev, 0, 0);
		else
			sectors = mddev->array_sectors;

		mddev->external_size = 0;
	} else {
		if (strict_blocks_to_sectors(buf, &sectors) < 0)
			return -EINVAL;
		if (mddev->pers && mddev->pers->size(mddev, 0, 0) < sectors)
4398
			return -E2BIG;
4399 4400 4401 4402 4403

		mddev->external_size = 1;
	}

	mddev->array_sectors = sectors;
4404 4405
	if (mddev->pers) {
		set_capacity(mddev->gendisk, mddev->array_sectors);
4406
		revalidate_disk(mddev->gendisk);
4407
	}
4408 4409 4410 4411 4412 4413
	return len;
}

static struct md_sysfs_entry md_array_size =
__ATTR(array_size, S_IRUGO|S_IWUSR, array_size_show,
       array_size_store);
4414

4415 4416
static struct attribute *md_default_attrs[] = {
	&md_level.attr,
4417
	&md_layout.attr,
4418
	&md_raid_disks.attr,
4419
	&md_chunk_size.attr,
4420
	&md_size.attr,
4421
	&md_resync_start.attr,
4422
	&md_metadata.attr,
4423
	&md_new_device.attr,
4424
	&md_safe_delay.attr,
4425
	&md_array_state.attr,
4426
	&md_reshape_position.attr,
4427
	&md_array_size.attr,
4428
	&max_corr_read_errors.attr,
4429 4430 4431 4432
	NULL,
};

static struct attribute *md_redundancy_attrs[] = {
4433
	&md_scan_mode.attr,
4434
	&md_mismatches.attr,
4435 4436 4437
	&md_sync_min.attr,
	&md_sync_max.attr,
	&md_sync_speed.attr,
4438
	&md_sync_force_parallel.attr,
4439
	&md_sync_completed.attr,
4440
	&md_min_sync.attr,
4441
	&md_max_sync.attr,
4442 4443
	&md_suspend_lo.attr,
	&md_suspend_hi.attr,
4444
	&md_bitmap.attr,
4445
	&md_degraded.attr,
4446 4447
	NULL,
};
4448 4449 4450 4451 4452
static struct attribute_group md_redundancy_group = {
	.name = NULL,
	.attrs = md_redundancy_attrs,
};

4453 4454 4455 4456 4457 4458

static ssize_t
md_attr_show(struct kobject *kobj, struct attribute *attr, char *page)
{
	struct md_sysfs_entry *entry = container_of(attr, struct md_sysfs_entry, attr);
	mddev_t *mddev = container_of(kobj, struct mddev_s, kobj);
4459
	ssize_t rv;
4460 4461 4462

	if (!entry->show)
		return -EIO;
I
Ingo Molnar 已提交
4463 4464 4465 4466 4467
	rv = mddev_lock(mddev);
	if (!rv) {
		rv = entry->show(mddev, page);
		mddev_unlock(mddev);
	}
4468
	return rv;
4469 4470 4471 4472 4473 4474 4475 4476
}

static ssize_t
md_attr_store(struct kobject *kobj, struct attribute *attr,
	      const char *page, size_t length)
{
	struct md_sysfs_entry *entry = container_of(attr, struct md_sysfs_entry, attr);
	mddev_t *mddev = container_of(kobj, struct mddev_s, kobj);
4477
	ssize_t rv;
4478 4479 4480

	if (!entry->store)
		return -EIO;
4481 4482
	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;
I
Ingo Molnar 已提交
4483
	rv = mddev_lock(mddev);
4484 4485
	if (mddev->hold_active == UNTIL_IOCTL)
		mddev->hold_active = 0;
I
Ingo Molnar 已提交
4486 4487 4488 4489
	if (!rv) {
		rv = entry->store(mddev, page, length);
		mddev_unlock(mddev);
	}
4490
	return rv;
4491 4492 4493 4494 4495
}

static void md_free(struct kobject *ko)
{
	mddev_t *mddev = container_of(ko, mddev_t, kobj);
4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506

	if (mddev->sysfs_state)
		sysfs_put(mddev->sysfs_state);

	if (mddev->gendisk) {
		del_gendisk(mddev->gendisk);
		put_disk(mddev->gendisk);
	}
	if (mddev->queue)
		blk_cleanup_queue(mddev->queue);

4507 4508 4509
	kfree(mddev);
}

4510
static const struct sysfs_ops md_sysfs_ops = {
4511 4512 4513 4514 4515 4516 4517 4518 4519
	.show	= md_attr_show,
	.store	= md_attr_store,
};
static struct kobj_type md_ktype = {
	.release	= md_free,
	.sysfs_ops	= &md_sysfs_ops,
	.default_attrs	= md_default_attrs,
};

L
Linus Torvalds 已提交
4520 4521
int mdp_major = 0;

4522 4523 4524 4525
static void mddev_delayed_delete(struct work_struct *ws)
{
	mddev_t *mddev = container_of(ws, mddev_t, del_work);

4526
	sysfs_remove_group(&mddev->kobj, &md_bitmap_group);
4527 4528 4529 4530
	kobject_del(&mddev->kobj);
	kobject_put(&mddev->kobj);
}

4531
static int md_alloc(dev_t dev, char *name)
L
Linus Torvalds 已提交
4532
{
4533
	static DEFINE_MUTEX(disks_mutex);
L
Linus Torvalds 已提交
4534 4535
	mddev_t *mddev = mddev_find(dev);
	struct gendisk *disk;
4536 4537 4538
	int partitioned;
	int shift;
	int unit;
4539
	int error;
L
Linus Torvalds 已提交
4540 4541

	if (!mddev)
4542 4543 4544 4545 4546
		return -ENODEV;

	partitioned = (MAJOR(mddev->unit) != MD_MAJOR);
	shift = partitioned ? MdpMinorShift : 0;
	unit = MINOR(mddev->unit) >> shift;
L
Linus Torvalds 已提交
4547

4548 4549
	/* wait for any previous instance of this device to be
	 * completely removed (mddev_delayed_delete).
4550
	 */
4551
	flush_workqueue(md_misc_wq);
4552

4553
	mutex_lock(&disks_mutex);
4554 4555 4556
	error = -EEXIST;
	if (mddev->gendisk)
		goto abort;
4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567

	if (name) {
		/* Need to ensure that 'name' is not a duplicate.
		 */
		mddev_t *mddev2;
		spin_lock(&all_mddevs_lock);

		list_for_each_entry(mddev2, &all_mddevs, all_mddevs)
			if (mddev2->gendisk &&
			    strcmp(mddev2->gendisk->disk_name, name) == 0) {
				spin_unlock(&all_mddevs_lock);
4568
				goto abort;
4569 4570
			}
		spin_unlock(&all_mddevs_lock);
L
Linus Torvalds 已提交
4571
	}
4572

4573
	error = -ENOMEM;
4574
	mddev->queue = blk_alloc_queue(GFP_KERNEL);
4575 4576
	if (!mddev->queue)
		goto abort;
4577 4578 4579
	mddev->queue->queuedata = mddev;

	blk_queue_make_request(mddev->queue, md_make_request);
4580

L
Linus Torvalds 已提交
4581 4582
	disk = alloc_disk(1 << shift);
	if (!disk) {
4583 4584
		blk_cleanup_queue(mddev->queue);
		mddev->queue = NULL;
4585
		goto abort;
L
Linus Torvalds 已提交
4586
	}
4587
	disk->major = MAJOR(mddev->unit);
L
Linus Torvalds 已提交
4588
	disk->first_minor = unit << shift;
4589 4590 4591
	if (name)
		strcpy(disk->disk_name, name);
	else if (partitioned)
L
Linus Torvalds 已提交
4592
		sprintf(disk->disk_name, "md_d%d", unit);
4593
	else
L
Linus Torvalds 已提交
4594 4595 4596 4597
		sprintf(disk->disk_name, "md%d", unit);
	disk->fops = &md_fops;
	disk->private_data = mddev;
	disk->queue = mddev->queue;
4598
	blk_queue_flush(mddev->queue, REQ_FLUSH | REQ_FUA);
4599
	/* Allow extended partitions.  This makes the
4600
	 * 'mdp' device redundant, but we can't really
4601 4602 4603
	 * remove it now.
	 */
	disk->flags |= GENHD_FL_EXT_DEVT;
L
Linus Torvalds 已提交
4604
	mddev->gendisk = disk;
4605 4606 4607 4608 4609 4610
	/* As soon as we call add_disk(), another thread could get
	 * through to md_open, so make sure it doesn't get too far
	 */
	mutex_lock(&mddev->open_mutex);
	add_disk(disk);

4611 4612
	error = kobject_init_and_add(&mddev->kobj, &md_ktype,
				     &disk_to_dev(disk)->kobj, "%s", "md");
4613 4614 4615 4616
	if (error) {
		/* This isn't possible, but as kobject_init_and_add is marked
		 * __must_check, we must do something with the result
		 */
4617 4618
		printk(KERN_WARNING "md: cannot register %s/md - name in use\n",
		       disk->disk_name);
4619 4620
		error = 0;
	}
4621 4622
	if (mddev->kobj.sd &&
	    sysfs_create_group(&mddev->kobj, &md_bitmap_group))
4623
		printk(KERN_DEBUG "pointless warning\n");
4624
	mutex_unlock(&mddev->open_mutex);
4625 4626
 abort:
	mutex_unlock(&disks_mutex);
4627
	if (!error && mddev->kobj.sd) {
4628
		kobject_uevent(&mddev->kobj, KOBJ_ADD);
4629
		mddev->sysfs_state = sysfs_get_dirent_safe(mddev->kobj.sd, "array_state");
4630
	}
4631
	mddev_put(mddev);
4632
	return error;
4633 4634 4635 4636 4637
}

static struct kobject *md_probe(dev_t dev, int *part, void *data)
{
	md_alloc(dev, NULL);
L
Linus Torvalds 已提交
4638 4639 4640
	return NULL;
}

4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659
static int add_named_array(const char *val, struct kernel_param *kp)
{
	/* val must be "md_*" where * is not all digits.
	 * We allocate an array with a large free minor number, and
	 * set the name to val.  val must not already be an active name.
	 */
	int len = strlen(val);
	char buf[DISK_NAME_LEN];

	while (len && val[len-1] == '\n')
		len--;
	if (len >= DISK_NAME_LEN)
		return -E2BIG;
	strlcpy(buf, val, len+1);
	if (strncmp(buf, "md_", 3) != 0)
		return -EINVAL;
	return md_alloc(0, buf);
}

L
Linus Torvalds 已提交
4660 4661 4662 4663
static void md_safemode_timeout(unsigned long data)
{
	mddev_t *mddev = (mddev_t *) data;

4664 4665 4666
	if (!atomic_read(&mddev->writes_pending)) {
		mddev->safemode = 1;
		if (mddev->external)
4667
			sysfs_notify_dirent_safe(mddev->sysfs_state);
4668
	}
L
Linus Torvalds 已提交
4669 4670 4671
	md_wakeup_thread(mddev->thread);
}

4672
static int start_dirty_degraded;
L
Linus Torvalds 已提交
4673

4674
int md_run(mddev_t *mddev)
L
Linus Torvalds 已提交
4675
{
4676
	int err;
L
Linus Torvalds 已提交
4677
	mdk_rdev_t *rdev;
4678
	struct mdk_personality *pers;
L
Linus Torvalds 已提交
4679

4680 4681
	if (list_empty(&mddev->disks))
		/* cannot run an array with no devices.. */
L
Linus Torvalds 已提交
4682 4683 4684 4685
		return -EINVAL;

	if (mddev->pers)
		return -EBUSY;
4686 4687 4688
	/* Cannot run until previous stop completes properly */
	if (mddev->sysfs_active)
		return -EBUSY;
4689

L
Linus Torvalds 已提交
4690 4691 4692
	/*
	 * Analyze all RAID superblock(s)
	 */
4693 4694 4695
	if (!mddev->raid_disks) {
		if (!mddev->persistent)
			return -EINVAL;
4696
		analyze_sbs(mddev);
4697
	}
L
Linus Torvalds 已提交
4698

4699 4700 4701 4702
	if (mddev->level != LEVEL_NONE)
		request_module("md-level-%d", mddev->level);
	else if (mddev->clevel[0])
		request_module("md-%s", mddev->clevel);
L
Linus Torvalds 已提交
4703 4704 4705 4706 4707 4708

	/*
	 * Drop all container device buffers, from now on
	 * the only valid external interface is through the md
	 * device.
	 */
4709
	list_for_each_entry(rdev, &mddev->disks, same_set) {
4710
		if (test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
4711 4712
			continue;
		sync_blockdev(rdev->bdev);
4713
		invalidate_bdev(rdev->bdev);
4714 4715 4716

		/* perform some consistency tests on the device.
		 * We don't want the data to overlap the metadata,
4717
		 * Internal Bitmap issues have been handled elsewhere.
4718
		 */
4719 4720 4721
		if (rdev->meta_bdev) {
			/* Nothing to check */;
		} else if (rdev->data_offset < rdev->sb_start) {
4722 4723
			if (mddev->dev_sectors &&
			    rdev->data_offset + mddev->dev_sectors
4724
			    > rdev->sb_start) {
4725 4726 4727 4728 4729
				printk("md: %s: data overlaps metadata\n",
				       mdname(mddev));
				return -EINVAL;
			}
		} else {
4730
			if (rdev->sb_start + rdev->sb_size/512
4731 4732 4733 4734 4735 4736
			    > rdev->data_offset) {
				printk("md: %s: metadata overlaps data\n",
				       mdname(mddev));
				return -EINVAL;
			}
		}
4737
		sysfs_notify_dirent_safe(rdev->sysfs_state);
L
Linus Torvalds 已提交
4738 4739
	}

4740
	if (mddev->bio_set == NULL)
4741 4742
		mddev->bio_set = bioset_create(BIO_POOL_SIZE,
					       sizeof(mddev_t *));
4743

L
Linus Torvalds 已提交
4744
	spin_lock(&pers_lock);
4745
	pers = find_pers(mddev->level, mddev->clevel);
4746
	if (!pers || !try_module_get(pers->owner)) {
L
Linus Torvalds 已提交
4747
		spin_unlock(&pers_lock);
4748 4749 4750 4751 4752 4753
		if (mddev->level != LEVEL_NONE)
			printk(KERN_WARNING "md: personality for level %d is not loaded!\n",
			       mddev->level);
		else
			printk(KERN_WARNING "md: personality for level %s is not loaded!\n",
			       mddev->clevel);
L
Linus Torvalds 已提交
4754 4755
		return -EINVAL;
	}
4756
	mddev->pers = pers;
L
Linus Torvalds 已提交
4757
	spin_unlock(&pers_lock);
4758 4759 4760 4761
	if (mddev->level != pers->level) {
		mddev->level = pers->level;
		mddev->new_level = pers->level;
	}
4762
	strlcpy(mddev->clevel, pers->name, sizeof(mddev->clevel));
L
Linus Torvalds 已提交
4763

4764
	if (mddev->reshape_position != MaxSector &&
4765
	    pers->start_reshape == NULL) {
4766 4767 4768 4769 4770 4771
		/* This personality cannot handle reshaping... */
		mddev->pers = NULL;
		module_put(pers->owner);
		return -EINVAL;
	}

4772 4773 4774 4775 4776 4777 4778
	if (pers->sync_request) {
		/* Warn if this is a potentially silly
		 * configuration.
		 */
		char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
		mdk_rdev_t *rdev2;
		int warned = 0;
4779 4780 4781

		list_for_each_entry(rdev, &mddev->disks, same_set)
			list_for_each_entry(rdev2, &mddev->disks, same_set) {
4782 4783 4784 4785 4786 4787 4788 4789 4790 4791 4792 4793 4794
				if (rdev < rdev2 &&
				    rdev->bdev->bd_contains ==
				    rdev2->bdev->bd_contains) {
					printk(KERN_WARNING
					       "%s: WARNING: %s appears to be"
					       " on the same physical disk as"
					       " %s.\n",
					       mdname(mddev),
					       bdevname(rdev->bdev,b),
					       bdevname(rdev2->bdev,b2));
					warned = 1;
				}
			}
4795

4796 4797 4798 4799 4800 4801
		if (warned)
			printk(KERN_WARNING
			       "True protection against single-disk"
			       " failure might be compromised.\n");
	}

4802
	mddev->recovery = 0;
4803 4804 4805
	/* may be over-ridden by personality */
	mddev->resync_max_sectors = mddev->dev_sectors;

4806
	mddev->ok_start_degraded = start_dirty_degraded;
L
Linus Torvalds 已提交
4807

4808
	if (start_readonly && mddev->ro == 0)
4809 4810
		mddev->ro = 2; /* read-only, but switch on first write */

4811
	err = mddev->pers->run(mddev);
4812 4813
	if (err)
		printk(KERN_ERR "md: pers->run() failed ...\n");
4814 4815 4816 4817 4818 4819 4820 4821 4822 4823 4824
	else if (mddev->pers->size(mddev, 0, 0) < mddev->array_sectors) {
		WARN_ONCE(!mddev->external_size, "%s: default size too small,"
			  " but 'external_size' not in effect?\n", __func__);
		printk(KERN_ERR
		       "md: invalid array_size %llu > default size %llu\n",
		       (unsigned long long)mddev->array_sectors / 2,
		       (unsigned long long)mddev->pers->size(mddev, 0, 0) / 2);
		err = -EINVAL;
		mddev->pers->stop(mddev);
	}
	if (err == 0 && mddev->pers->sync_request) {
4825 4826 4827 4828 4829 4830 4831
		err = bitmap_create(mddev);
		if (err) {
			printk(KERN_ERR "%s: failed to create bitmap (%d)\n",
			       mdname(mddev), err);
			mddev->pers->stop(mddev);
		}
	}
L
Linus Torvalds 已提交
4832 4833 4834
	if (err) {
		module_put(mddev->pers->owner);
		mddev->pers = NULL;
4835 4836
		bitmap_destroy(mddev);
		return err;
L
Linus Torvalds 已提交
4837
	}
4838
	if (mddev->pers->sync_request) {
4839 4840
		if (mddev->kobj.sd &&
		    sysfs_create_group(&mddev->kobj, &md_redundancy_group))
4841 4842 4843
			printk(KERN_WARNING
			       "md: cannot register extra attributes for %s\n",
			       mdname(mddev));
4844
		mddev->sysfs_action = sysfs_get_dirent_safe(mddev->kobj.sd, "sync_action");
4845
	} else if (mddev->ro == 2) /* auto-readonly not meaningful */
4846 4847
		mddev->ro = 0;

L
Linus Torvalds 已提交
4848
 	atomic_set(&mddev->writes_pending,0);
4849 4850
	atomic_set(&mddev->max_corr_read_errors,
		   MD_DEFAULT_MAX_CORRECTED_READ_ERRORS);
L
Linus Torvalds 已提交
4851 4852 4853
	mddev->safemode = 0;
	mddev->safemode_timer.function = md_safemode_timeout;
	mddev->safemode_timer.data = (unsigned long) mddev;
4854
	mddev->safemode_delay = (200 * HZ)/1000 +1; /* 200 msec delay */
L
Linus Torvalds 已提交
4855
	mddev->in_sync = 1;
4856 4857
	smp_wmb();
	mddev->ready = 1;
4858
	list_for_each_entry(rdev, &mddev->disks, same_set)
4859 4860
		if (rdev->raid_disk >= 0)
			if (sysfs_link_rdev(mddev, rdev))
4861
				/* failure here is OK */;
L
Linus Torvalds 已提交
4862 4863 4864
	
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	
4865 4866
	if (mddev->flags)
		md_update_sb(mddev, 0);
L
Linus Torvalds 已提交
4867

4868
	md_new_event(mddev);
4869 4870
	sysfs_notify_dirent_safe(mddev->sysfs_state);
	sysfs_notify_dirent_safe(mddev->sysfs_action);
4871
	sysfs_notify(&mddev->kobj, NULL, "degraded");
L
Linus Torvalds 已提交
4872 4873
	return 0;
}
4874
EXPORT_SYMBOL_GPL(md_run);
L
Linus Torvalds 已提交
4875

4876 4877 4878 4879 4880 4881 4882
static int do_md_run(mddev_t *mddev)
{
	int err;

	err = md_run(mddev);
	if (err)
		goto out;
4883 4884 4885 4886 4887
	err = bitmap_load(mddev);
	if (err) {
		bitmap_destroy(mddev);
		goto out;
	}
4888 4889 4890 4891

	md_wakeup_thread(mddev->thread);
	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */

4892 4893
	set_capacity(mddev->gendisk, mddev->array_sectors);
	revalidate_disk(mddev->gendisk);
4894
	mddev->changed = 1;
4895 4896 4897 4898 4899
	kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE);
out:
	return err;
}

L
Linus Torvalds 已提交
4900 4901 4902 4903
static int restart_array(mddev_t *mddev)
{
	struct gendisk *disk = mddev->gendisk;

4904
	/* Complain if it has no devices */
L
Linus Torvalds 已提交
4905
	if (list_empty(&mddev->disks))
4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919
		return -ENXIO;
	if (!mddev->pers)
		return -EINVAL;
	if (!mddev->ro)
		return -EBUSY;
	mddev->safemode = 0;
	mddev->ro = 0;
	set_disk_ro(disk, 0);
	printk(KERN_INFO "md: %s switched to read-write mode.\n",
		mdname(mddev));
	/* Kick recovery or resync if necessary */
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
	md_wakeup_thread(mddev->sync_thread);
4920
	sysfs_notify_dirent_safe(mddev->sysfs_state);
4921
	return 0;
L
Linus Torvalds 已提交
4922 4923
}

4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940
/* similar to deny_write_access, but accounts for our holding a reference
 * to the file ourselves */
static int deny_bitmap_write_access(struct file * file)
{
	struct inode *inode = file->f_mapping->host;

	spin_lock(&inode->i_lock);
	if (atomic_read(&inode->i_writecount) > 1) {
		spin_unlock(&inode->i_lock);
		return -ETXTBSY;
	}
	atomic_set(&inode->i_writecount, -1);
	spin_unlock(&inode->i_lock);

	return 0;
}

4941
void restore_bitmap_write_access(struct file *file)
4942 4943 4944 4945 4946 4947 4948 4949
{
	struct inode *inode = file->f_mapping->host;

	spin_lock(&inode->i_lock);
	atomic_set(&inode->i_writecount, 1);
	spin_unlock(&inode->i_lock);
}

4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971
static void md_clean(mddev_t *mddev)
{
	mddev->array_sectors = 0;
	mddev->external_size = 0;
	mddev->dev_sectors = 0;
	mddev->raid_disks = 0;
	mddev->recovery_cp = 0;
	mddev->resync_min = 0;
	mddev->resync_max = MaxSector;
	mddev->reshape_position = MaxSector;
	mddev->external = 0;
	mddev->persistent = 0;
	mddev->level = LEVEL_NONE;
	mddev->clevel[0] = 0;
	mddev->flags = 0;
	mddev->ro = 0;
	mddev->metadata_type[0] = 0;
	mddev->chunk_sectors = 0;
	mddev->ctime = mddev->utime = 0;
	mddev->layout = 0;
	mddev->max_disks = 0;
	mddev->events = 0;
4972
	mddev->can_decrease_events = 0;
4973 4974 4975 4976 4977 4978 4979 4980 4981 4982
	mddev->delta_disks = 0;
	mddev->new_level = LEVEL_NONE;
	mddev->new_layout = 0;
	mddev->new_chunk_sectors = 0;
	mddev->curr_resync = 0;
	mddev->resync_mismatches = 0;
	mddev->suspend_lo = mddev->suspend_hi = 0;
	mddev->sync_speed_min = mddev->sync_speed_max = 0;
	mddev->recovery = 0;
	mddev->in_sync = 0;
4983
	mddev->changed = 0;
4984 4985 4986 4987 4988 4989 4990 4991 4992
	mddev->degraded = 0;
	mddev->safemode = 0;
	mddev->bitmap_info.offset = 0;
	mddev->bitmap_info.default_offset = 0;
	mddev->bitmap_info.chunksize = 0;
	mddev->bitmap_info.daemon_sleep = 0;
	mddev->bitmap_info.max_write_behind = 0;
}

4993
static void __md_stop_writes(mddev_t *mddev)
4994 4995 4996 4997
{
	if (mddev->sync_thread) {
		set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
4998
		reap_sync_thread(mddev);
4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011
	}

	del_timer_sync(&mddev->safemode_timer);

	bitmap_flush(mddev);
	md_super_wait(mddev);

	if (!mddev->in_sync || mddev->flags) {
		/* mark array as shutdown cleanly */
		mddev->in_sync = 1;
		md_update_sb(mddev, 1);
	}
}
5012 5013 5014 5015 5016 5017 5018

void md_stop_writes(mddev_t *mddev)
{
	mddev_lock(mddev);
	__md_stop_writes(mddev);
	mddev_unlock(mddev);
}
5019
EXPORT_SYMBOL_GPL(md_stop_writes);
5020

5021
void md_stop(mddev_t *mddev)
5022
{
5023
	mddev->ready = 0;
5024 5025 5026 5027 5028
	mddev->pers->stop(mddev);
	if (mddev->pers->sync_request && mddev->to_remove == NULL)
		mddev->to_remove = &md_redundancy_group;
	module_put(mddev->pers->owner);
	mddev->pers = NULL;
5029
	clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
5030
}
5031
EXPORT_SYMBOL_GPL(md_stop);
5032

5033 5034 5035 5036 5037 5038 5039 5040 5041 5042
static int md_set_readonly(mddev_t *mddev, int is_open)
{
	int err = 0;
	mutex_lock(&mddev->open_mutex);
	if (atomic_read(&mddev->openers) > is_open) {
		printk("md: %s still in use.\n",mdname(mddev));
		err = -EBUSY;
		goto out;
	}
	if (mddev->pers) {
5043
		__md_stop_writes(mddev);
5044 5045 5046 5047 5048 5049 5050

		err  = -ENXIO;
		if (mddev->ro==1)
			goto out;
		mddev->ro = 1;
		set_disk_ro(mddev->gendisk, 1);
		clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
5051
		sysfs_notify_dirent_safe(mddev->sysfs_state);
5052 5053 5054 5055 5056 5057 5058
		err = 0;	
	}
out:
	mutex_unlock(&mddev->open_mutex);
	return err;
}

5059 5060 5061 5062
/* mode:
 *   0 - completely stop and dis-assemble array
 *   2 - stop but do not disassemble array
 */
5063
static int do_md_stop(mddev_t * mddev, int mode, int is_open)
L
Linus Torvalds 已提交
5064 5065
{
	struct gendisk *disk = mddev->gendisk;
5066
	mdk_rdev_t *rdev;
L
Linus Torvalds 已提交
5067

5068
	mutex_lock(&mddev->open_mutex);
5069 5070
	if (atomic_read(&mddev->openers) > is_open ||
	    mddev->sysfs_active) {
5071
		printk("md: %s still in use.\n",mdname(mddev));
N
NeilBrown 已提交
5072 5073 5074
		mutex_unlock(&mddev->open_mutex);
		return -EBUSY;
	}
L
Linus Torvalds 已提交
5075

N
NeilBrown 已提交
5076
	if (mddev->pers) {
5077 5078
		if (mddev->ro)
			set_disk_ro(disk, 0);
5079

5080
		__md_stop_writes(mddev);
5081 5082 5083
		md_stop(mddev);
		mddev->queue->merge_bvec_fn = NULL;
		mddev->queue->backing_dev_info.congested_fn = NULL;
5084

5085
		/* tell userspace to handle 'inactive' */
5086
		sysfs_notify_dirent_safe(mddev->sysfs_state);
5087

5088
		list_for_each_entry(rdev, &mddev->disks, same_set)
5089 5090
			if (rdev->raid_disk >= 0)
				sysfs_unlink_rdev(mddev, rdev);
5091

5092
		set_capacity(disk, 0);
N
NeilBrown 已提交
5093
		mutex_unlock(&mddev->open_mutex);
5094
		mddev->changed = 1;
5095
		revalidate_disk(disk);
5096

5097 5098
		if (mddev->ro)
			mddev->ro = 0;
N
NeilBrown 已提交
5099 5100
	} else
		mutex_unlock(&mddev->open_mutex);
L
Linus Torvalds 已提交
5101 5102 5103
	/*
	 * Free resources if final stop
	 */
5104
	if (mode == 0) {
L
Linus Torvalds 已提交
5105 5106
		printk(KERN_INFO "md: %s stopped.\n", mdname(mddev));

5107
		bitmap_destroy(mddev);
5108 5109 5110 5111
		if (mddev->bitmap_info.file) {
			restore_bitmap_write_access(mddev->bitmap_info.file);
			fput(mddev->bitmap_info.file);
			mddev->bitmap_info.file = NULL;
5112
		}
5113
		mddev->bitmap_info.offset = 0;
5114

L
Linus Torvalds 已提交
5115 5116
		export_array(mddev);

5117
		md_clean(mddev);
5118
		kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE);
5119 5120
		if (mddev->hold_active == UNTIL_STOP)
			mddev->hold_active = 0;
5121
	}
5122
	blk_integrity_unregister(disk);
5123
	md_new_event(mddev);
5124
	sysfs_notify_dirent_safe(mddev->sysfs_state);
N
NeilBrown 已提交
5125
	return 0;
L
Linus Torvalds 已提交
5126 5127
}

5128
#ifndef MODULE
L
Linus Torvalds 已提交
5129 5130 5131 5132 5133
static void autorun_array(mddev_t *mddev)
{
	mdk_rdev_t *rdev;
	int err;

5134
	if (list_empty(&mddev->disks))
L
Linus Torvalds 已提交
5135 5136 5137 5138
		return;

	printk(KERN_INFO "md: running: ");

5139
	list_for_each_entry(rdev, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
5140 5141 5142 5143 5144
		char b[BDEVNAME_SIZE];
		printk("<%s>", bdevname(rdev->bdev,b));
	}
	printk("\n");

5145
	err = do_md_run(mddev);
L
Linus Torvalds 已提交
5146 5147
	if (err) {
		printk(KERN_WARNING "md: do_md_run() returned %d\n", err);
5148
		do_md_stop(mddev, 0, 0);
L
Linus Torvalds 已提交
5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165
	}
}

/*
 * lets try to run arrays based on all disks that have arrived
 * until now. (those are in pending_raid_disks)
 *
 * the method: pick the first pending disk, collect all disks with
 * the same UUID, remove all from the pending list and put them into
 * the 'same_array' list. Then order this list based on superblock
 * update time (freshest comes first), kick out 'old' disks and
 * compare superblocks. If everything's fine then run it.
 *
 * If "unit" is allocated, then bump its reference count
 */
static void autorun_devices(int part)
{
5166
	mdk_rdev_t *rdev0, *rdev, *tmp;
L
Linus Torvalds 已提交
5167 5168 5169 5170 5171
	mddev_t *mddev;
	char b[BDEVNAME_SIZE];

	printk(KERN_INFO "md: autorun ...\n");
	while (!list_empty(&pending_raid_disks)) {
5172
		int unit;
L
Linus Torvalds 已提交
5173
		dev_t dev;
5174
		LIST_HEAD(candidates);
L
Linus Torvalds 已提交
5175 5176 5177 5178 5179 5180
		rdev0 = list_entry(pending_raid_disks.next,
					 mdk_rdev_t, same_set);

		printk(KERN_INFO "md: considering %s ...\n",
			bdevname(rdev0->bdev,b));
		INIT_LIST_HEAD(&candidates);
5181
		rdev_for_each_list(rdev, tmp, &pending_raid_disks)
L
Linus Torvalds 已提交
5182 5183 5184 5185 5186 5187 5188 5189 5190 5191
			if (super_90_load(rdev, rdev0, 0) >= 0) {
				printk(KERN_INFO "md:  adding %s ...\n",
					bdevname(rdev->bdev,b));
				list_move(&rdev->same_set, &candidates);
			}
		/*
		 * now we have a set of devices, with all of them having
		 * mostly sane superblocks. It's time to allocate the
		 * mddev.
		 */
5192 5193 5194 5195 5196 5197 5198 5199 5200
		if (part) {
			dev = MKDEV(mdp_major,
				    rdev0->preferred_minor << MdpMinorShift);
			unit = MINOR(dev) >> MdpMinorShift;
		} else {
			dev = MKDEV(MD_MAJOR, rdev0->preferred_minor);
			unit = MINOR(dev);
		}
		if (rdev0->preferred_minor != unit) {
L
Linus Torvalds 已提交
5201 5202 5203 5204 5205 5206 5207
			printk(KERN_INFO "md: unit number in %s is bad: %d\n",
			       bdevname(rdev0->bdev, b), rdev0->preferred_minor);
			break;
		}

		md_probe(dev, NULL, NULL);
		mddev = mddev_find(dev);
5208 5209 5210 5211
		if (!mddev || !mddev->gendisk) {
			if (mddev)
				mddev_put(mddev);
			printk(KERN_ERR
L
Linus Torvalds 已提交
5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225
				"md: cannot allocate memory for md drive.\n");
			break;
		}
		if (mddev_lock(mddev)) 
			printk(KERN_WARNING "md: %s locked, cannot run\n",
			       mdname(mddev));
		else if (mddev->raid_disks || mddev->major_version
			 || !list_empty(&mddev->disks)) {
			printk(KERN_WARNING 
				"md: %s already running, cannot run %s\n",
				mdname(mddev), bdevname(rdev0->bdev,b));
			mddev_unlock(mddev);
		} else {
			printk(KERN_INFO "md: created %s\n", mdname(mddev));
5226
			mddev->persistent = 1;
5227
			rdev_for_each_list(rdev, tmp, &candidates) {
L
Linus Torvalds 已提交
5228 5229 5230 5231 5232 5233 5234 5235 5236 5237
				list_del_init(&rdev->same_set);
				if (bind_rdev_to_array(rdev, mddev))
					export_rdev(rdev);
			}
			autorun_array(mddev);
			mddev_unlock(mddev);
		}
		/* on success, candidates will be empty, on error
		 * it won't...
		 */
5238
		rdev_for_each_list(rdev, tmp, &candidates) {
5239
			list_del_init(&rdev->same_set);
L
Linus Torvalds 已提交
5240
			export_rdev(rdev);
5241
		}
L
Linus Torvalds 已提交
5242 5243 5244 5245
		mddev_put(mddev);
	}
	printk(KERN_INFO "md: ... autorun DONE.\n");
}
5246
#endif /* !MODULE */
L
Linus Torvalds 已提交
5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264

static int get_version(void __user * arg)
{
	mdu_version_t ver;

	ver.major = MD_MAJOR_VERSION;
	ver.minor = MD_MINOR_VERSION;
	ver.patchlevel = MD_PATCHLEVEL_VERSION;

	if (copy_to_user(arg, &ver, sizeof(ver)))
		return -EFAULT;

	return 0;
}

static int get_array_info(mddev_t * mddev, void __user * arg)
{
	mdu_array_info_t info;
5265
	int nr,working,insync,failed,spare;
L
Linus Torvalds 已提交
5266 5267
	mdk_rdev_t *rdev;

5268
	nr=working=insync=failed=spare=0;
5269
	list_for_each_entry(rdev, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
5270
		nr++;
5271
		if (test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
5272 5273 5274
			failed++;
		else {
			working++;
5275
			if (test_bit(In_sync, &rdev->flags))
5276
				insync++;	
L
Linus Torvalds 已提交
5277 5278 5279 5280 5281 5282 5283 5284 5285 5286
			else
				spare++;
		}
	}

	info.major_version = mddev->major_version;
	info.minor_version = mddev->minor_version;
	info.patch_version = MD_PATCHLEVEL_VERSION;
	info.ctime         = mddev->ctime;
	info.level         = mddev->level;
5287 5288
	info.size          = mddev->dev_sectors / 2;
	if (info.size != mddev->dev_sectors / 2) /* overflow */
5289
		info.size = -1;
L
Linus Torvalds 已提交
5290 5291 5292 5293 5294 5295 5296 5297 5298
	info.nr_disks      = nr;
	info.raid_disks    = mddev->raid_disks;
	info.md_minor      = mddev->md_minor;
	info.not_persistent= !mddev->persistent;

	info.utime         = mddev->utime;
	info.state         = 0;
	if (mddev->in_sync)
		info.state = (1<<MD_SB_CLEAN);
5299
	if (mddev->bitmap && mddev->bitmap_info.offset)
5300
		info.state = (1<<MD_SB_BITMAP_PRESENT);
5301
	info.active_disks  = insync;
L
Linus Torvalds 已提交
5302 5303 5304 5305 5306
	info.working_disks = working;
	info.failed_disks  = failed;
	info.spare_disks   = spare;

	info.layout        = mddev->layout;
5307
	info.chunk_size    = mddev->chunk_sectors << 9;
L
Linus Torvalds 已提交
5308 5309 5310 5311 5312 5313 5314

	if (copy_to_user(arg, &info, sizeof(info)))
		return -EFAULT;

	return 0;
}

5315
static int get_bitmap_file(mddev_t * mddev, void __user * arg)
5316 5317 5318 5319 5320
{
	mdu_bitmap_file_t *file = NULL; /* too big for stack allocation */
	char *ptr, *buf = NULL;
	int err = -ENOMEM;

5321 5322 5323 5324
	if (md_allow_write(mddev))
		file = kmalloc(sizeof(*file), GFP_NOIO);
	else
		file = kmalloc(sizeof(*file), GFP_KERNEL);
5325

5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338
	if (!file)
		goto out;

	/* bitmap disabled, zero the first byte and copy out */
	if (!mddev->bitmap || !mddev->bitmap->file) {
		file->pathname[0] = '\0';
		goto copy_out;
	}

	buf = kmalloc(sizeof(file->pathname), GFP_KERNEL);
	if (!buf)
		goto out;

5339 5340
	ptr = d_path(&mddev->bitmap->file->f_path, buf, sizeof(file->pathname));
	if (IS_ERR(ptr))
5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354
		goto out;

	strcpy(file->pathname, ptr);

copy_out:
	err = 0;
	if (copy_to_user(arg, file, sizeof(*file)))
		err = -EFAULT;
out:
	kfree(buf);
	kfree(file);
	return err;
}

L
Linus Torvalds 已提交
5355 5356 5357 5358 5359 5360 5361 5362
static int get_disk_info(mddev_t * mddev, void __user * arg)
{
	mdu_disk_info_t info;
	mdk_rdev_t *rdev;

	if (copy_from_user(&info, arg, sizeof(info)))
		return -EFAULT;

5363
	rdev = find_rdev_nr(mddev, info.number);
L
Linus Torvalds 已提交
5364 5365 5366 5367 5368
	if (rdev) {
		info.major = MAJOR(rdev->bdev->bd_dev);
		info.minor = MINOR(rdev->bdev->bd_dev);
		info.raid_disk = rdev->raid_disk;
		info.state = 0;
5369
		if (test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
5370
			info.state |= (1<<MD_DISK_FAULTY);
5371
		else if (test_bit(In_sync, &rdev->flags)) {
L
Linus Torvalds 已提交
5372 5373 5374
			info.state |= (1<<MD_DISK_ACTIVE);
			info.state |= (1<<MD_DISK_SYNC);
		}
5375 5376
		if (test_bit(WriteMostly, &rdev->flags))
			info.state |= (1<<MD_DISK_WRITEMOSTLY);
L
Linus Torvalds 已提交
5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 5390 5391 5392 5393 5394 5395 5396 5397 5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409 5410
	} else {
		info.major = info.minor = 0;
		info.raid_disk = -1;
		info.state = (1<<MD_DISK_REMOVED);
	}

	if (copy_to_user(arg, &info, sizeof(info)))
		return -EFAULT;

	return 0;
}

static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info)
{
	char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
	mdk_rdev_t *rdev;
	dev_t dev = MKDEV(info->major,info->minor);

	if (info->major != MAJOR(dev) || info->minor != MINOR(dev))
		return -EOVERFLOW;

	if (!mddev->raid_disks) {
		int err;
		/* expecting a device which has a superblock */
		rdev = md_import_device(dev, mddev->major_version, mddev->minor_version);
		if (IS_ERR(rdev)) {
			printk(KERN_WARNING 
				"md: md_import_device returned %ld\n",
				PTR_ERR(rdev));
			return PTR_ERR(rdev);
		}
		if (!list_empty(&mddev->disks)) {
			mdk_rdev_t *rdev0 = list_entry(mddev->disks.next,
							mdk_rdev_t, same_set);
5411
			err = super_types[mddev->major_version]
L
Linus Torvalds 已提交
5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 5425 5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440
				.load_super(rdev, rdev0, mddev->minor_version);
			if (err < 0) {
				printk(KERN_WARNING 
					"md: %s has different UUID to %s\n",
					bdevname(rdev->bdev,b), 
					bdevname(rdev0->bdev,b2));
				export_rdev(rdev);
				return -EINVAL;
			}
		}
		err = bind_rdev_to_array(rdev, mddev);
		if (err)
			export_rdev(rdev);
		return err;
	}

	/*
	 * add_new_disk can be used once the array is assembled
	 * to add "hot spares".  They must already have a superblock
	 * written
	 */
	if (mddev->pers) {
		int err;
		if (!mddev->pers->hot_add_disk) {
			printk(KERN_WARNING 
				"%s: personality does not support diskops!\n",
			       mdname(mddev));
			return -EINVAL;
		}
5441 5442 5443 5444 5445
		if (mddev->persistent)
			rdev = md_import_device(dev, mddev->major_version,
						mddev->minor_version);
		else
			rdev = md_import_device(dev, -1, -1);
L
Linus Torvalds 已提交
5446 5447 5448 5449 5450 5451
		if (IS_ERR(rdev)) {
			printk(KERN_WARNING 
				"md: md_import_device returned %ld\n",
				PTR_ERR(rdev));
			return PTR_ERR(rdev);
		}
5452
		/* set saved_raid_disk if appropriate */
5453 5454
		if (!mddev->persistent) {
			if (info->state & (1<<MD_DISK_SYNC)  &&
5455
			    info->raid_disk < mddev->raid_disks) {
5456
				rdev->raid_disk = info->raid_disk;
5457 5458
				set_bit(In_sync, &rdev->flags);
			} else
5459 5460 5461 5462
				rdev->raid_disk = -1;
		} else
			super_types[mddev->major_version].
				validate_super(mddev, rdev);
5463 5464 5465 5466 5467 5468 5469 5470 5471 5472
		if ((info->state & (1<<MD_DISK_SYNC)) &&
		    (!test_bit(In_sync, &rdev->flags) ||
		     rdev->raid_disk != info->raid_disk)) {
			/* This was a hot-add request, but events doesn't
			 * match, so reject it.
			 */
			export_rdev(rdev);
			return -EINVAL;
		}

5473 5474 5475 5476
		if (test_bit(In_sync, &rdev->flags))
			rdev->saved_raid_disk = rdev->raid_disk;
		else
			rdev->saved_raid_disk = -1;
5477

5478
		clear_bit(In_sync, &rdev->flags); /* just to be sure */
5479 5480
		if (info->state & (1<<MD_DISK_WRITEMOSTLY))
			set_bit(WriteMostly, &rdev->flags);
5481 5482
		else
			clear_bit(WriteMostly, &rdev->flags);
5483

L
Linus Torvalds 已提交
5484 5485
		rdev->raid_disk = -1;
		err = bind_rdev_to_array(rdev, mddev);
5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496
		if (!err && !mddev->pers->hot_remove_disk) {
			/* If there is hot_add_disk but no hot_remove_disk
			 * then added disks for geometry changes,
			 * and should be added immediately.
			 */
			super_types[mddev->major_version].
				validate_super(mddev, rdev);
			err = mddev->pers->hot_add_disk(mddev, rdev);
			if (err)
				unbind_rdev_from_array(rdev);
		}
L
Linus Torvalds 已提交
5497 5498
		if (err)
			export_rdev(rdev);
5499
		else
5500
			sysfs_notify_dirent_safe(rdev->sysfs_state);
5501

5502
		md_update_sb(mddev, 1);
5503 5504
		if (mddev->degraded)
			set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
5505
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
5506 5507
		if (!err)
			md_new_event(mddev);
5508
		md_wakeup_thread(mddev->thread);
L
Linus Torvalds 已提交
5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522
		return err;
	}

	/* otherwise, add_new_disk is only allowed
	 * for major_version==0 superblocks
	 */
	if (mddev->major_version != 0) {
		printk(KERN_WARNING "%s: ADD_NEW_DISK not supported\n",
		       mdname(mddev));
		return -EINVAL;
	}

	if (!(info->state & (1<<MD_DISK_FAULTY))) {
		int err;
5523
		rdev = md_import_device(dev, -1, 0);
L
Linus Torvalds 已提交
5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536
		if (IS_ERR(rdev)) {
			printk(KERN_WARNING 
				"md: error, md_import_device() returned %ld\n",
				PTR_ERR(rdev));
			return PTR_ERR(rdev);
		}
		rdev->desc_nr = info->number;
		if (info->raid_disk < mddev->raid_disks)
			rdev->raid_disk = info->raid_disk;
		else
			rdev->raid_disk = -1;

		if (rdev->raid_disk < mddev->raid_disks)
5537 5538
			if (info->state & (1<<MD_DISK_SYNC))
				set_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
5539

5540 5541 5542
		if (info->state & (1<<MD_DISK_WRITEMOSTLY))
			set_bit(WriteMostly, &rdev->flags);

L
Linus Torvalds 已提交
5543 5544
		if (!mddev->persistent) {
			printk(KERN_INFO "md: nonpersistent superblock ...\n");
5545 5546
			rdev->sb_start = i_size_read(rdev->bdev->bd_inode) / 512;
		} else
5547
			rdev->sb_start = calc_dev_sboffset(rdev);
5548
		rdev->sectors = rdev->sb_start;
L
Linus Torvalds 已提交
5549

5550 5551 5552 5553 5554
		err = bind_rdev_to_array(rdev, mddev);
		if (err) {
			export_rdev(rdev);
			return err;
		}
L
Linus Torvalds 已提交
5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572
	}

	return 0;
}

static int hot_remove_disk(mddev_t * mddev, dev_t dev)
{
	char b[BDEVNAME_SIZE];
	mdk_rdev_t *rdev;

	rdev = find_rdev(mddev, dev);
	if (!rdev)
		return -ENXIO;

	if (rdev->raid_disk >= 0)
		goto busy;

	kick_rdev_from_array(rdev);
5573
	md_update_sb(mddev, 1);
5574
	md_new_event(mddev);
L
Linus Torvalds 已提交
5575 5576 5577

	return 0;
busy:
5578
	printk(KERN_WARNING "md: cannot remove active disk %s from %s ...\n",
L
Linus Torvalds 已提交
5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604
		bdevname(rdev->bdev,b), mdname(mddev));
	return -EBUSY;
}

static int hot_add_disk(mddev_t * mddev, dev_t dev)
{
	char b[BDEVNAME_SIZE];
	int err;
	mdk_rdev_t *rdev;

	if (!mddev->pers)
		return -ENODEV;

	if (mddev->major_version != 0) {
		printk(KERN_WARNING "%s: HOT_ADD may only be used with"
			" version-0 superblocks.\n",
			mdname(mddev));
		return -EINVAL;
	}
	if (!mddev->pers->hot_add_disk) {
		printk(KERN_WARNING 
			"%s: personality does not support diskops!\n",
			mdname(mddev));
		return -EINVAL;
	}

5605
	rdev = md_import_device(dev, -1, 0);
L
Linus Torvalds 已提交
5606 5607 5608 5609 5610 5611 5612 5613
	if (IS_ERR(rdev)) {
		printk(KERN_WARNING 
			"md: error, md_import_device() returned %ld\n",
			PTR_ERR(rdev));
		return -EINVAL;
	}

	if (mddev->persistent)
5614
		rdev->sb_start = calc_dev_sboffset(rdev);
L
Linus Torvalds 已提交
5615
	else
5616
		rdev->sb_start = i_size_read(rdev->bdev->bd_inode) / 512;
L
Linus Torvalds 已提交
5617

5618
	rdev->sectors = rdev->sb_start;
L
Linus Torvalds 已提交
5619

5620
	if (test_bit(Faulty, &rdev->flags)) {
L
Linus Torvalds 已提交
5621 5622 5623 5624 5625 5626
		printk(KERN_WARNING 
			"md: can not hot-add faulty %s disk to %s!\n",
			bdevname(rdev->bdev,b), mdname(mddev));
		err = -EINVAL;
		goto abort_export;
	}
5627
	clear_bit(In_sync, &rdev->flags);
L
Linus Torvalds 已提交
5628
	rdev->desc_nr = -1;
5629
	rdev->saved_raid_disk = -1;
5630 5631 5632
	err = bind_rdev_to_array(rdev, mddev);
	if (err)
		goto abort_export;
L
Linus Torvalds 已提交
5633 5634 5635 5636 5637 5638 5639 5640

	/*
	 * The rest should better be atomic, we can have disk failures
	 * noticed in interrupt contexts ...
	 */

	rdev->raid_disk = -1;

5641
	md_update_sb(mddev, 1);
L
Linus Torvalds 已提交
5642 5643 5644 5645 5646 5647 5648

	/*
	 * Kick recovery, maybe this spare has to be added to the
	 * array immediately.
	 */
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
5649
	md_new_event(mddev);
L
Linus Torvalds 已提交
5650 5651 5652 5653 5654 5655 5656
	return 0;

abort_export:
	export_rdev(rdev);
	return err;
}

5657 5658 5659 5660
static int set_bitmap_file(mddev_t *mddev, int fd)
{
	int err;

5661 5662 5663 5664 5665 5666 5667
	if (mddev->pers) {
		if (!mddev->pers->quiesce)
			return -EBUSY;
		if (mddev->recovery || mddev->sync_thread)
			return -EBUSY;
		/* we should be able to change the bitmap.. */
	}
5668 5669


5670 5671 5672
	if (fd >= 0) {
		if (mddev->bitmap)
			return -EEXIST; /* cannot add when bitmap is present */
5673
		mddev->bitmap_info.file = fget(fd);
5674

5675
		if (mddev->bitmap_info.file == NULL) {
5676 5677 5678 5679 5680
			printk(KERN_ERR "%s: error: failed to get bitmap file\n",
			       mdname(mddev));
			return -EBADF;
		}

5681
		err = deny_bitmap_write_access(mddev->bitmap_info.file);
5682 5683 5684
		if (err) {
			printk(KERN_ERR "%s: error: bitmap file is already in use\n",
			       mdname(mddev));
5685 5686
			fput(mddev->bitmap_info.file);
			mddev->bitmap_info.file = NULL;
5687 5688
			return err;
		}
5689
		mddev->bitmap_info.offset = 0; /* file overrides offset */
5690 5691 5692 5693 5694
	} else if (mddev->bitmap == NULL)
		return -ENOENT; /* cannot remove what isn't there */
	err = 0;
	if (mddev->pers) {
		mddev->pers->quiesce(mddev, 1);
5695
		if (fd >= 0) {
5696
			err = bitmap_create(mddev);
5697 5698 5699
			if (!err)
				err = bitmap_load(mddev);
		}
5700
		if (fd < 0 || err) {
5701
			bitmap_destroy(mddev);
5702 5703
			fd = -1; /* make sure to put the file */
		}
5704
		mddev->pers->quiesce(mddev, 0);
5705 5706
	}
	if (fd < 0) {
5707 5708 5709
		if (mddev->bitmap_info.file) {
			restore_bitmap_write_access(mddev->bitmap_info.file);
			fput(mddev->bitmap_info.file);
5710
		}
5711
		mddev->bitmap_info.file = NULL;
5712 5713
	}

5714 5715 5716
	return err;
}

L
Linus Torvalds 已提交
5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735
/*
 * set_array_info is used two different ways
 * The original usage is when creating a new array.
 * In this usage, raid_disks is > 0 and it together with
 *  level, size, not_persistent,layout,chunksize determine the
 *  shape of the array.
 *  This will always create an array with a type-0.90.0 superblock.
 * The newer usage is when assembling an array.
 *  In this case raid_disks will be 0, and the major_version field is
 *  use to determine which style super-blocks are to be found on the devices.
 *  The minor and patch _version numbers are also kept incase the
 *  super_block handler wishes to interpret them.
 */
static int set_array_info(mddev_t * mddev, mdu_array_info_t *info)
{

	if (info->raid_disks == 0) {
		/* just setting version number for superblock loading */
		if (info->major_version < 0 ||
5736
		    info->major_version >= ARRAY_SIZE(super_types) ||
L
Linus Torvalds 已提交
5737 5738 5739 5740 5741 5742 5743 5744 5745 5746
		    super_types[info->major_version].name == NULL) {
			/* maybe try to auto-load a module? */
			printk(KERN_INFO 
				"md: superblock version %d not known\n",
				info->major_version);
			return -EINVAL;
		}
		mddev->major_version = info->major_version;
		mddev->minor_version = info->minor_version;
		mddev->patch_version = info->patch_version;
5747
		mddev->persistent = !info->not_persistent;
5748 5749 5750 5751
		/* ensure mddev_put doesn't delete this now that there
		 * is some minimal configuration.
		 */
		mddev->ctime         = get_seconds();
L
Linus Torvalds 已提交
5752 5753 5754 5755 5756 5757 5758 5759
		return 0;
	}
	mddev->major_version = MD_MAJOR_VERSION;
	mddev->minor_version = MD_MINOR_VERSION;
	mddev->patch_version = MD_PATCHLEVEL_VERSION;
	mddev->ctime         = get_seconds();

	mddev->level         = info->level;
5760
	mddev->clevel[0]     = 0;
5761
	mddev->dev_sectors   = 2 * (sector_t)info->size;
L
Linus Torvalds 已提交
5762 5763 5764 5765 5766 5767 5768 5769 5770
	mddev->raid_disks    = info->raid_disks;
	/* don't set md_minor, it is determined by which /dev/md* was
	 * openned
	 */
	if (info->state & (1<<MD_SB_CLEAN))
		mddev->recovery_cp = MaxSector;
	else
		mddev->recovery_cp = 0;
	mddev->persistent    = ! info->not_persistent;
5771
	mddev->external	     = 0;
L
Linus Torvalds 已提交
5772 5773

	mddev->layout        = info->layout;
5774
	mddev->chunk_sectors = info->chunk_size >> 9;
L
Linus Torvalds 已提交
5775 5776 5777

	mddev->max_disks     = MD_SB_DISKS;

5778 5779
	if (mddev->persistent)
		mddev->flags         = 0;
5780
	set_bit(MD_CHANGE_DEVS, &mddev->flags);
L
Linus Torvalds 已提交
5781

5782 5783
	mddev->bitmap_info.default_offset = MD_SB_BYTES >> 9;
	mddev->bitmap_info.offset = 0;
5784

5785 5786
	mddev->reshape_position = MaxSector;

L
Linus Torvalds 已提交
5787 5788 5789 5790 5791
	/*
	 * Generate a 128 bit UUID
	 */
	get_random_bytes(mddev->uuid, 16);

5792
	mddev->new_level = mddev->level;
5793
	mddev->new_chunk_sectors = mddev->chunk_sectors;
5794 5795 5796
	mddev->new_layout = mddev->layout;
	mddev->delta_disks = 0;

L
Linus Torvalds 已提交
5797 5798 5799
	return 0;
}

5800 5801
void md_set_array_sectors(mddev_t *mddev, sector_t array_sectors)
{
5802 5803 5804 5805 5806
	WARN(!mddev_is_locked(mddev), "%s: unlocked mddev!\n", __func__);

	if (mddev->external_size)
		return;

5807 5808 5809 5810
	mddev->array_sectors = array_sectors;
}
EXPORT_SYMBOL(md_set_array_sectors);

5811
static int update_size(mddev_t *mddev, sector_t num_sectors)
5812
{
5813
	mdk_rdev_t *rdev;
5814
	int rv;
5815
	int fit = (num_sectors == 0);
5816 5817 5818

	if (mddev->pers->resize == NULL)
		return -EINVAL;
5819 5820 5821 5822 5823
	/* The "num_sectors" is the number of sectors of each device that
	 * is used.  This can only make sense for arrays with redundancy.
	 * linear and raid0 always use whatever space is available. We can only
	 * consider changing this number if no resync or reconstruction is
	 * happening, and if the new size is acceptable. It must fit before the
5824
	 * sb_start or, if that is <data_offset, it must fit before the size
5825 5826
	 * of each device.  If num_sectors is zero, we find the largest size
	 * that fits.
5827 5828 5829
	 */
	if (mddev->sync_thread)
		return -EBUSY;
5830 5831 5832 5833 5834
	if (mddev->bitmap)
		/* Sorry, cannot grow a bitmap yet, just remove it,
		 * grow, and re-add.
		 */
		return -EBUSY;
5835
	list_for_each_entry(rdev, &mddev->disks, same_set) {
5836
		sector_t avail = rdev->sectors;
5837

5838 5839 5840
		if (fit && (num_sectors == 0 || num_sectors > avail))
			num_sectors = avail;
		if (avail < num_sectors)
5841 5842
			return -ENOSPC;
	}
5843
	rv = mddev->pers->resize(mddev, num_sectors);
5844 5845
	if (!rv)
		revalidate_disk(mddev->gendisk);
5846 5847 5848
	return rv;
}

5849 5850 5851 5852
static int update_raid_disks(mddev_t *mddev, int raid_disks)
{
	int rv;
	/* change the number of raid disks */
5853
	if (mddev->pers->check_reshape == NULL)
5854 5855
		return -EINVAL;
	if (raid_disks <= 0 ||
5856
	    (mddev->max_disks && raid_disks >= mddev->max_disks))
5857
		return -EINVAL;
5858
	if (mddev->sync_thread || mddev->reshape_position != MaxSector)
5859
		return -EBUSY;
5860 5861 5862
	mddev->delta_disks = raid_disks - mddev->raid_disks;

	rv = mddev->pers->check_reshape(mddev);
5863 5864
	if (rv < 0)
		mddev->delta_disks = 0;
5865 5866 5867 5868
	return rv;
}


L
Linus Torvalds 已提交
5869 5870 5871 5872 5873 5874 5875 5876 5877 5878 5879 5880
/*
 * update_array_info is used to change the configuration of an
 * on-line array.
 * The version, ctime,level,size,raid_disks,not_persistent, layout,chunk_size
 * fields in the info are checked against the array.
 * Any differences that cannot be handled will cause an error.
 * Normally, only one change can be managed at a time.
 */
static int update_array_info(mddev_t *mddev, mdu_array_info_t *info)
{
	int rv = 0;
	int cnt = 0;
5881 5882 5883
	int state = 0;

	/* calculate expected state,ignoring low bits */
5884
	if (mddev->bitmap && mddev->bitmap_info.offset)
5885
		state |= (1 << MD_SB_BITMAP_PRESENT);
L
Linus Torvalds 已提交
5886 5887 5888 5889 5890 5891 5892 5893

	if (mddev->major_version != info->major_version ||
	    mddev->minor_version != info->minor_version ||
/*	    mddev->patch_version != info->patch_version || */
	    mddev->ctime         != info->ctime         ||
	    mddev->level         != info->level         ||
/*	    mddev->layout        != info->layout        || */
	    !mddev->persistent	 != info->not_persistent||
5894
	    mddev->chunk_sectors != info->chunk_size >> 9 ||
5895 5896 5897
	    /* ignore bottom 8 bits of state, and allow SB_BITMAP_PRESENT to change */
	    ((state^info->state) & 0xfffffe00)
		)
L
Linus Torvalds 已提交
5898 5899
		return -EINVAL;
	/* Check there is only one change */
5900 5901 5902 5903 5904 5905 5906 5907 5908 5909 5910 5911
	if (info->size >= 0 && mddev->dev_sectors / 2 != info->size)
		cnt++;
	if (mddev->raid_disks != info->raid_disks)
		cnt++;
	if (mddev->layout != info->layout)
		cnt++;
	if ((state ^ info->state) & (1<<MD_SB_BITMAP_PRESENT))
		cnt++;
	if (cnt == 0)
		return 0;
	if (cnt > 1)
		return -EINVAL;
L
Linus Torvalds 已提交
5912 5913 5914 5915 5916 5917

	if (mddev->layout != info->layout) {
		/* Change layout
		 * we don't need to do anything at the md level, the
		 * personality will take care of it all.
		 */
5918
		if (mddev->pers->check_reshape == NULL)
L
Linus Torvalds 已提交
5919
			return -EINVAL;
5920 5921
		else {
			mddev->new_layout = info->layout;
5922
			rv = mddev->pers->check_reshape(mddev);
5923 5924 5925 5926
			if (rv)
				mddev->new_layout = mddev->layout;
			return rv;
		}
L
Linus Torvalds 已提交
5927
	}
5928
	if (info->size >= 0 && mddev->dev_sectors / 2 != info->size)
5929
		rv = update_size(mddev, (sector_t)info->size * 2);
5930

5931 5932 5933
	if (mddev->raid_disks    != info->raid_disks)
		rv = update_raid_disks(mddev, info->raid_disks);

5934 5935 5936 5937 5938 5939 5940 5941 5942
	if ((state ^ info->state) & (1<<MD_SB_BITMAP_PRESENT)) {
		if (mddev->pers->quiesce == NULL)
			return -EINVAL;
		if (mddev->recovery || mddev->sync_thread)
			return -EBUSY;
		if (info->state & (1<<MD_SB_BITMAP_PRESENT)) {
			/* add the bitmap */
			if (mddev->bitmap)
				return -EEXIST;
5943
			if (mddev->bitmap_info.default_offset == 0)
5944
				return -EINVAL;
5945 5946
			mddev->bitmap_info.offset =
				mddev->bitmap_info.default_offset;
5947 5948
			mddev->pers->quiesce(mddev, 1);
			rv = bitmap_create(mddev);
5949 5950
			if (!rv)
				rv = bitmap_load(mddev);
5951 5952 5953 5954 5955 5956 5957 5958 5959 5960 5961 5962
			if (rv)
				bitmap_destroy(mddev);
			mddev->pers->quiesce(mddev, 0);
		} else {
			/* remove the bitmap */
			if (!mddev->bitmap)
				return -ENOENT;
			if (mddev->bitmap->file)
				return -EINVAL;
			mddev->pers->quiesce(mddev, 1);
			bitmap_destroy(mddev);
			mddev->pers->quiesce(mddev, 0);
5963
			mddev->bitmap_info.offset = 0;
5964 5965
		}
	}
5966
	md_update_sb(mddev, 1);
L
Linus Torvalds 已提交
5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984
	return rv;
}

static int set_disk_faulty(mddev_t *mddev, dev_t dev)
{
	mdk_rdev_t *rdev;

	if (mddev->pers == NULL)
		return -ENODEV;

	rdev = find_rdev(mddev, dev);
	if (!rdev)
		return -ENODEV;

	md_error(mddev, rdev);
	return 0;
}

5985 5986 5987 5988 5989 5990
/*
 * We have a problem here : there is no easy way to give a CHS
 * virtual geometry. We currently pretend that we have a 2 heads
 * 4 sectors (with a BIG number of cylinders...). This drives
 * dosfs just mad... ;-)
 */
5991 5992 5993 5994 5995 5996
static int md_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
	mddev_t *mddev = bdev->bd_disk->private_data;

	geo->heads = 2;
	geo->sectors = 4;
5997
	geo->cylinders = mddev->array_sectors / 8;
5998 5999 6000
	return 0;
}

A
Al Viro 已提交
6001
static int md_ioctl(struct block_device *bdev, fmode_t mode,
L
Linus Torvalds 已提交
6002 6003 6004 6005 6006
			unsigned int cmd, unsigned long arg)
{
	int err = 0;
	void __user *argp = (void __user *)arg;
	mddev_t *mddev = NULL;
6007
	int ro;
L
Linus Torvalds 已提交
6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039

	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;

	/*
	 * Commands dealing with the RAID driver but not any
	 * particular array:
	 */
	switch (cmd)
	{
		case RAID_VERSION:
			err = get_version(argp);
			goto done;

		case PRINT_RAID_DEBUG:
			err = 0;
			md_print_devices();
			goto done;

#ifndef MODULE
		case RAID_AUTORUN:
			err = 0;
			autostart_arrays(arg);
			goto done;
#endif
		default:;
	}

	/*
	 * Commands creating/starting a new array:
	 */

A
Al Viro 已提交
6040
	mddev = bdev->bd_disk->private_data;
L
Linus Torvalds 已提交
6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103

	if (!mddev) {
		BUG();
		goto abort;
	}

	err = mddev_lock(mddev);
	if (err) {
		printk(KERN_INFO 
			"md: ioctl lock interrupted, reason %d, cmd %d\n",
			err, cmd);
		goto abort;
	}

	switch (cmd)
	{
		case SET_ARRAY_INFO:
			{
				mdu_array_info_t info;
				if (!arg)
					memset(&info, 0, sizeof(info));
				else if (copy_from_user(&info, argp, sizeof(info))) {
					err = -EFAULT;
					goto abort_unlock;
				}
				if (mddev->pers) {
					err = update_array_info(mddev, &info);
					if (err) {
						printk(KERN_WARNING "md: couldn't update"
						       " array info. %d\n", err);
						goto abort_unlock;
					}
					goto done_unlock;
				}
				if (!list_empty(&mddev->disks)) {
					printk(KERN_WARNING
					       "md: array %s already has disks!\n",
					       mdname(mddev));
					err = -EBUSY;
					goto abort_unlock;
				}
				if (mddev->raid_disks) {
					printk(KERN_WARNING
					       "md: array %s already initialised!\n",
					       mdname(mddev));
					err = -EBUSY;
					goto abort_unlock;
				}
				err = set_array_info(mddev, &info);
				if (err) {
					printk(KERN_WARNING "md: couldn't set"
					       " array info. %d\n", err);
					goto abort_unlock;
				}
			}
			goto done_unlock;

		default:;
	}

	/*
	 * Commands querying/configuring an existing array:
	 */
6104
	/* if we are not initialised yet, only ADD_NEW_DISK, STOP_ARRAY,
6105
	 * RUN_ARRAY, and GET_ and SET_BITMAP_FILE are allowed */
6106 6107 6108 6109
	if ((!mddev->raid_disks && !mddev->external)
	    && cmd != ADD_NEW_DISK && cmd != STOP_ARRAY
	    && cmd != RUN_ARRAY && cmd != SET_BITMAP_FILE
	    && cmd != GET_BITMAP_FILE) {
L
Linus Torvalds 已提交
6110 6111 6112 6113 6114 6115 6116 6117 6118 6119 6120 6121 6122
		err = -ENODEV;
		goto abort_unlock;
	}

	/*
	 * Commands even a read-only array can execute:
	 */
	switch (cmd)
	{
		case GET_ARRAY_INFO:
			err = get_array_info(mddev, argp);
			goto done_unlock;

6123
		case GET_BITMAP_FILE:
6124
			err = get_bitmap_file(mddev, argp);
6125 6126
			goto done_unlock;

L
Linus Torvalds 已提交
6127 6128 6129 6130 6131 6132 6133 6134 6135
		case GET_DISK_INFO:
			err = get_disk_info(mddev, argp);
			goto done_unlock;

		case RESTART_ARRAY_RW:
			err = restart_array(mddev);
			goto done_unlock;

		case STOP_ARRAY:
6136
			err = do_md_stop(mddev, 0, 1);
L
Linus Torvalds 已提交
6137 6138 6139
			goto done_unlock;

		case STOP_ARRAY_RO:
6140
			err = md_set_readonly(mddev, 1);
L
Linus Torvalds 已提交
6141 6142
			goto done_unlock;

6143 6144 6145 6146 6147 6148 6149 6150 6151 6152 6153 6154 6155 6156 6157 6158 6159 6160 6161 6162 6163 6164 6165 6166 6167 6168 6169 6170
		case BLKROSET:
			if (get_user(ro, (int __user *)(arg))) {
				err = -EFAULT;
				goto done_unlock;
			}
			err = -EINVAL;

			/* if the bdev is going readonly the value of mddev->ro
			 * does not matter, no writes are coming
			 */
			if (ro)
				goto done_unlock;

			/* are we are already prepared for writes? */
			if (mddev->ro != 1)
				goto done_unlock;

			/* transitioning to readauto need only happen for
			 * arrays that call md_write_start
			 */
			if (mddev->pers) {
				err = restart_array(mddev);
				if (err == 0) {
					mddev->ro = 2;
					set_disk_ro(mddev->gendisk, 0);
				}
			}
			goto done_unlock;
L
Linus Torvalds 已提交
6171 6172 6173 6174
	}

	/*
	 * The remaining ioctls are changing the state of the
6175 6176 6177 6178
	 * superblock, so we do not allow them on read-only arrays.
	 * However non-MD ioctls (e.g. get-size) will still come through
	 * here and hit the 'default' below, so only disallow
	 * 'md' ioctls, and switch to rw mode if started auto-readonly.
L
Linus Torvalds 已提交
6179
	 */
6180
	if (_IOC_TYPE(cmd) == MD_MAJOR && mddev->ro && mddev->pers) {
6181 6182
		if (mddev->ro == 2) {
			mddev->ro = 0;
6183
			sysfs_notify_dirent_safe(mddev->sysfs_state);
6184 6185
			set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
			md_wakeup_thread(mddev->thread);
6186 6187 6188 6189
		} else {
			err = -EROFS;
			goto abort_unlock;
		}
L
Linus Torvalds 已提交
6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216
	}

	switch (cmd)
	{
		case ADD_NEW_DISK:
		{
			mdu_disk_info_t info;
			if (copy_from_user(&info, argp, sizeof(info)))
				err = -EFAULT;
			else
				err = add_new_disk(mddev, &info);
			goto done_unlock;
		}

		case HOT_REMOVE_DISK:
			err = hot_remove_disk(mddev, new_decode_dev(arg));
			goto done_unlock;

		case HOT_ADD_DISK:
			err = hot_add_disk(mddev, new_decode_dev(arg));
			goto done_unlock;

		case SET_DISK_FAULTY:
			err = set_disk_faulty(mddev, new_decode_dev(arg));
			goto done_unlock;

		case RUN_ARRAY:
6217
			err = do_md_run(mddev);
L
Linus Torvalds 已提交
6218 6219
			goto done_unlock;

6220 6221 6222 6223
		case SET_BITMAP_FILE:
			err = set_bitmap_file(mddev, (int)arg);
			goto done_unlock;

L
Linus Torvalds 已提交
6224 6225 6226 6227 6228 6229 6230
		default:
			err = -EINVAL;
			goto abort_unlock;
	}

done_unlock:
abort_unlock:
6231 6232 6233
	if (mddev->hold_active == UNTIL_IOCTL &&
	    err != -EINVAL)
		mddev->hold_active = 0;
L
Linus Torvalds 已提交
6234 6235 6236 6237 6238 6239 6240 6241 6242
	mddev_unlock(mddev);

	return err;
done:
	if (err)
		MD_BUG();
abort:
	return err;
}
6243 6244 6245 6246 6247 6248 6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260 6261
#ifdef CONFIG_COMPAT
static int md_compat_ioctl(struct block_device *bdev, fmode_t mode,
		    unsigned int cmd, unsigned long arg)
{
	switch (cmd) {
	case HOT_REMOVE_DISK:
	case HOT_ADD_DISK:
	case SET_DISK_FAULTY:
	case SET_BITMAP_FILE:
		/* These take in integer arg, do not convert */
		break;
	default:
		arg = (unsigned long)compat_ptr(arg);
		break;
	}

	return md_ioctl(bdev, mode, cmd, arg);
}
#endif /* CONFIG_COMPAT */
L
Linus Torvalds 已提交
6262

A
Al Viro 已提交
6263
static int md_open(struct block_device *bdev, fmode_t mode)
L
Linus Torvalds 已提交
6264 6265 6266 6267 6268
{
	/*
	 * Succeed if we can lock the mddev, which confirms that
	 * it isn't being stopped right now.
	 */
6269
	mddev_t *mddev = mddev_find(bdev->bd_dev);
L
Linus Torvalds 已提交
6270 6271
	int err;

6272 6273 6274 6275 6276 6277
	if (mddev->gendisk != bdev->bd_disk) {
		/* we are racing with mddev_put which is discarding this
		 * bd_disk.
		 */
		mddev_put(mddev);
		/* Wait until bdev->bd_disk is definitely gone */
6278
		flush_workqueue(md_misc_wq);
6279 6280 6281 6282 6283
		/* Then retry the open from the top */
		return -ERESTARTSYS;
	}
	BUG_ON(mddev != bdev->bd_disk->private_data);

6284
	if ((err = mutex_lock_interruptible(&mddev->open_mutex)))
L
Linus Torvalds 已提交
6285 6286 6287
		goto out;

	err = 0;
6288
	atomic_inc(&mddev->openers);
6289
	mutex_unlock(&mddev->open_mutex);
L
Linus Torvalds 已提交
6290

6291
	check_disk_change(bdev);
L
Linus Torvalds 已提交
6292 6293 6294 6295
 out:
	return err;
}

A
Al Viro 已提交
6296
static int md_release(struct gendisk *disk, fmode_t mode)
L
Linus Torvalds 已提交
6297
{
A
Al Viro 已提交
6298
 	mddev_t *mddev = disk->private_data;
L
Linus Torvalds 已提交
6299

6300
	BUG_ON(!mddev);
6301
	atomic_dec(&mddev->openers);
L
Linus Torvalds 已提交
6302 6303 6304 6305
	mddev_put(mddev);

	return 0;
}
6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320

static int md_media_changed(struct gendisk *disk)
{
	mddev_t *mddev = disk->private_data;

	return mddev->changed;
}

static int md_revalidate(struct gendisk *disk)
{
	mddev_t *mddev = disk->private_data;

	mddev->changed = 0;
	return 0;
}
6321
static const struct block_device_operations md_fops =
L
Linus Torvalds 已提交
6322 6323
{
	.owner		= THIS_MODULE,
A
Al Viro 已提交
6324 6325
	.open		= md_open,
	.release	= md_release,
N
NeilBrown 已提交
6326
	.ioctl		= md_ioctl,
6327 6328 6329
#ifdef CONFIG_COMPAT
	.compat_ioctl	= md_compat_ioctl,
#endif
6330
	.getgeo		= md_getgeo,
6331 6332
	.media_changed  = md_media_changed,
	.revalidate_disk= md_revalidate,
L
Linus Torvalds 已提交
6333 6334
};

6335
static int md_thread(void * arg)
L
Linus Torvalds 已提交
6336 6337 6338 6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350
{
	mdk_thread_t *thread = arg;

	/*
	 * md_thread is a 'system-thread', it's priority should be very
	 * high. We avoid resource deadlocks individually in each
	 * raid personality. (RAID5 does preallocation) We also use RR and
	 * the very same RT priority as kswapd, thus we will never get
	 * into a priority inversion deadlock.
	 *
	 * we definitely have to have equal or higher priority than
	 * bdflush, otherwise bdflush will deadlock if there are too
	 * many dirty RAID5 blocks.
	 */

6351
	allow_signal(SIGKILL);
6352
	while (!kthread_should_stop()) {
L
Linus Torvalds 已提交
6353

6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366
		/* We need to wait INTERRUPTIBLE so that
		 * we don't add to the load-average.
		 * That means we need to be sure no signals are
		 * pending
		 */
		if (signal_pending(current))
			flush_signals(current);

		wait_event_interruptible_timeout
			(thread->wqueue,
			 test_bit(THREAD_WAKEUP, &thread->flags)
			 || kthread_should_stop(),
			 thread->timeout);
L
Linus Torvalds 已提交
6367

6368 6369
		clear_bit(THREAD_WAKEUP, &thread->flags);
		if (!kthread_should_stop())
6370
			thread->run(thread->mddev);
L
Linus Torvalds 已提交
6371
	}
6372

L
Linus Torvalds 已提交
6373 6374 6375 6376 6377 6378 6379 6380 6381 6382 6383 6384 6385 6386 6387 6388 6389
	return 0;
}

void md_wakeup_thread(mdk_thread_t *thread)
{
	if (thread) {
		dprintk("md: waking up MD thread %s.\n", thread->tsk->comm);
		set_bit(THREAD_WAKEUP, &thread->flags);
		wake_up(&thread->wqueue);
	}
}

mdk_thread_t *md_register_thread(void (*run) (mddev_t *), mddev_t *mddev,
				 const char *name)
{
	mdk_thread_t *thread;

6390
	thread = kzalloc(sizeof(mdk_thread_t), GFP_KERNEL);
L
Linus Torvalds 已提交
6391 6392 6393 6394 6395 6396 6397
	if (!thread)
		return NULL;

	init_waitqueue_head(&thread->wqueue);

	thread->run = run;
	thread->mddev = mddev;
6398
	thread->timeout = MAX_SCHEDULE_TIMEOUT;
6399 6400 6401 6402
	thread->tsk = kthread_run(md_thread, thread,
				  "%s_%s",
				  mdname(thread->mddev),
				  name ?: mddev->pers->name);
6403
	if (IS_ERR(thread->tsk)) {
L
Linus Torvalds 已提交
6404 6405 6406 6407 6408 6409 6410 6411
		kfree(thread);
		return NULL;
	}
	return thread;
}

void md_unregister_thread(mdk_thread_t *thread)
{
6412 6413
	if (!thread)
		return;
6414
	dprintk("interrupting MD-thread pid %d\n", task_pid_nr(thread->tsk));
6415 6416

	kthread_stop(thread->tsk);
L
Linus Torvalds 已提交
6417 6418 6419 6420 6421 6422 6423 6424 6425 6426
	kfree(thread);
}

void md_error(mddev_t *mddev, mdk_rdev_t *rdev)
{
	if (!mddev) {
		MD_BUG();
		return;
	}

6427
	if (!rdev || test_bit(Faulty, &rdev->flags))
L
Linus Torvalds 已提交
6428
		return;
6429

6430
	if (!mddev->pers || !mddev->pers->error_handler)
L
Linus Torvalds 已提交
6431 6432
		return;
	mddev->pers->error_handler(mddev,rdev);
6433 6434
	if (mddev->degraded)
		set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
6435
	sysfs_notify_dirent_safe(rdev->sysfs_state);
L
Linus Torvalds 已提交
6436 6437 6438
	set_bit(MD_RECOVERY_INTR, &mddev->recovery);
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
6439
	if (mddev->event_work.func)
6440
		queue_work(md_misc_wq, &mddev->event_work);
6441
	md_new_event_inintr(mddev);
L
Linus Torvalds 已提交
6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452
}

/* seq_file implementation /proc/mdstat */

static void status_unused(struct seq_file *seq)
{
	int i = 0;
	mdk_rdev_t *rdev;

	seq_printf(seq, "unused devices: ");

6453
	list_for_each_entry(rdev, &pending_raid_disks, same_set) {
L
Linus Torvalds 已提交
6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 6465 6466 6467
		char b[BDEVNAME_SIZE];
		i++;
		seq_printf(seq, "%s ",
			      bdevname(rdev->bdev,b));
	}
	if (!i)
		seq_printf(seq, "<none>");

	seq_printf(seq, "\n");
}


static void status_resync(struct seq_file *seq, mddev_t * mddev)
{
6468 6469 6470
	sector_t max_sectors, resync, res;
	unsigned long dt, db;
	sector_t rt;
6471 6472
	int scale;
	unsigned int per_milli;
L
Linus Torvalds 已提交
6473

6474
	resync = mddev->curr_resync - atomic_read(&mddev->recovery_active);
L
Linus Torvalds 已提交
6475 6476

	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
6477
		max_sectors = mddev->resync_max_sectors;
L
Linus Torvalds 已提交
6478
	else
6479
		max_sectors = mddev->dev_sectors;
L
Linus Torvalds 已提交
6480 6481 6482 6483

	/*
	 * Should not happen.
	 */
6484
	if (!max_sectors) {
L
Linus Torvalds 已提交
6485 6486 6487
		MD_BUG();
		return;
	}
6488
	/* Pick 'scale' such that (resync>>scale)*1000 will fit
6489
	 * in a sector_t, and (max_sectors>>scale) will fit in a
6490 6491 6492 6493 6494
	 * u32, as those are the requirements for sector_div.
	 * Thus 'scale' must be at least 10
	 */
	scale = 10;
	if (sizeof(sector_t) > sizeof(unsigned long)) {
6495
		while ( max_sectors/2 > (1ULL<<(scale+32)))
6496 6497 6498
			scale++;
	}
	res = (resync>>scale)*1000;
6499
	sector_div(res, (u32)((max_sectors>>scale)+1));
6500 6501

	per_milli = res;
L
Linus Torvalds 已提交
6502
	{
6503
		int i, x = per_milli/50, y = 20-x;
L
Linus Torvalds 已提交
6504 6505 6506 6507 6508 6509 6510 6511
		seq_printf(seq, "[");
		for (i = 0; i < x; i++)
			seq_printf(seq, "=");
		seq_printf(seq, ">");
		for (i = 0; i < y; i++)
			seq_printf(seq, ".");
		seq_printf(seq, "] ");
	}
6512
	seq_printf(seq, " %s =%3u.%u%% (%llu/%llu)",
6513 6514
		   (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)?
		    "reshape" :
6515 6516 6517 6518 6519
		    (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)?
		     "check" :
		     (test_bit(MD_RECOVERY_SYNC, &mddev->recovery) ?
		      "resync" : "recovery"))),
		   per_milli/10, per_milli % 10,
6520 6521
		   (unsigned long long) resync/2,
		   (unsigned long long) max_sectors/2);
L
Linus Torvalds 已提交
6522 6523 6524 6525 6526

	/*
	 * dt: time from mark until now
	 * db: blocks written from mark until now
	 * rt: remaining time
6527 6528 6529 6530
	 *
	 * rt is a sector_t, so could be 32bit or 64bit.
	 * So we divide before multiply in case it is 32bit and close
	 * to the limit.
L
Lucas De Marchi 已提交
6531
	 * We scale the divisor (db) by 32 to avoid losing precision
6532 6533 6534 6535
	 * near the end of resync when the number of remaining sectors
	 * is close to 'db'.
	 * We then divide rt by 32 after multiplying by db to compensate.
	 * The '+1' avoids division by zero if db is very small.
L
Linus Torvalds 已提交
6536 6537 6538
	 */
	dt = ((jiffies - mddev->resync_mark) / HZ);
	if (!dt) dt++;
6539 6540
	db = (mddev->curr_mark_cnt - atomic_read(&mddev->recovery_active))
		- mddev->resync_mark_cnt;
L
Linus Torvalds 已提交
6541

6542 6543 6544 6545 6546 6547 6548
	rt = max_sectors - resync;    /* number of remaining sectors */
	sector_div(rt, db/32+1);
	rt *= dt;
	rt >>= 5;

	seq_printf(seq, " finish=%lu.%lumin", (unsigned long)rt / 60,
		   ((unsigned long)rt % 60)/6);
L
Linus Torvalds 已提交
6549

6550
	seq_printf(seq, " speed=%ldK/sec", db/2/dt);
L
Linus Torvalds 已提交
6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566 6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580 6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591 6592 6593 6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608 6609 6610 6611 6612 6613 6614 6615 6616 6617
}

static void *md_seq_start(struct seq_file *seq, loff_t *pos)
{
	struct list_head *tmp;
	loff_t l = *pos;
	mddev_t *mddev;

	if (l >= 0x10000)
		return NULL;
	if (!l--)
		/* header */
		return (void*)1;

	spin_lock(&all_mddevs_lock);
	list_for_each(tmp,&all_mddevs)
		if (!l--) {
			mddev = list_entry(tmp, mddev_t, all_mddevs);
			mddev_get(mddev);
			spin_unlock(&all_mddevs_lock);
			return mddev;
		}
	spin_unlock(&all_mddevs_lock);
	if (!l--)
		return (void*)2;/* tail */
	return NULL;
}

static void *md_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
	struct list_head *tmp;
	mddev_t *next_mddev, *mddev = v;
	
	++*pos;
	if (v == (void*)2)
		return NULL;

	spin_lock(&all_mddevs_lock);
	if (v == (void*)1)
		tmp = all_mddevs.next;
	else
		tmp = mddev->all_mddevs.next;
	if (tmp != &all_mddevs)
		next_mddev = mddev_get(list_entry(tmp,mddev_t,all_mddevs));
	else {
		next_mddev = (void*)2;
		*pos = 0x10000;
	}		
	spin_unlock(&all_mddevs_lock);

	if (v != (void*)1)
		mddev_put(mddev);
	return next_mddev;

}

static void md_seq_stop(struct seq_file *seq, void *v)
{
	mddev_t *mddev = v;

	if (mddev && v != (void*)1 && v != (void*)2)
		mddev_put(mddev);
}

static int md_seq_show(struct seq_file *seq, void *v)
{
	mddev_t *mddev = v;
6618
	sector_t sectors;
L
Linus Torvalds 已提交
6619
	mdk_rdev_t *rdev;
6620
	struct bitmap *bitmap;
L
Linus Torvalds 已提交
6621 6622

	if (v == (void*)1) {
6623
		struct mdk_personality *pers;
L
Linus Torvalds 已提交
6624 6625
		seq_printf(seq, "Personalities : ");
		spin_lock(&pers_lock);
6626 6627
		list_for_each_entry(pers, &pers_list, list)
			seq_printf(seq, "[%s] ", pers->name);
L
Linus Torvalds 已提交
6628 6629 6630

		spin_unlock(&pers_lock);
		seq_printf(seq, "\n");
6631
		seq->poll_event = atomic_read(&md_event_count);
L
Linus Torvalds 已提交
6632 6633 6634 6635 6636 6637 6638
		return 0;
	}
	if (v == (void*)2) {
		status_unused(seq);
		return 0;
	}

I
Ingo Molnar 已提交
6639
	if (mddev_lock(mddev) < 0)
L
Linus Torvalds 已提交
6640
		return -EINTR;
I
Ingo Molnar 已提交
6641

L
Linus Torvalds 已提交
6642 6643 6644 6645
	if (mddev->pers || mddev->raid_disks || !list_empty(&mddev->disks)) {
		seq_printf(seq, "%s : %sactive", mdname(mddev),
						mddev->pers ? "" : "in");
		if (mddev->pers) {
6646
			if (mddev->ro==1)
L
Linus Torvalds 已提交
6647
				seq_printf(seq, " (read-only)");
6648
			if (mddev->ro==2)
6649
				seq_printf(seq, " (auto-read-only)");
L
Linus Torvalds 已提交
6650 6651 6652
			seq_printf(seq, " %s", mddev->pers->name);
		}

6653
		sectors = 0;
6654
		list_for_each_entry(rdev, &mddev->disks, same_set) {
L
Linus Torvalds 已提交
6655 6656 6657
			char b[BDEVNAME_SIZE];
			seq_printf(seq, " %s[%d]",
				bdevname(rdev->bdev,b), rdev->desc_nr);
6658 6659
			if (test_bit(WriteMostly, &rdev->flags))
				seq_printf(seq, "(W)");
6660
			if (test_bit(Faulty, &rdev->flags)) {
L
Linus Torvalds 已提交
6661 6662
				seq_printf(seq, "(F)");
				continue;
6663 6664
			} else if (rdev->raid_disk < 0)
				seq_printf(seq, "(S)"); /* spare */
6665
			sectors += rdev->sectors;
L
Linus Torvalds 已提交
6666 6667 6668 6669 6670
		}

		if (!list_empty(&mddev->disks)) {
			if (mddev->pers)
				seq_printf(seq, "\n      %llu blocks",
6671 6672
					   (unsigned long long)
					   mddev->array_sectors / 2);
L
Linus Torvalds 已提交
6673 6674
			else
				seq_printf(seq, "\n      %llu blocks",
6675
					   (unsigned long long)sectors / 2);
L
Linus Torvalds 已提交
6676
		}
6677 6678 6679 6680 6681 6682 6683
		if (mddev->persistent) {
			if (mddev->major_version != 0 ||
			    mddev->minor_version != 90) {
				seq_printf(seq," super %d.%d",
					   mddev->major_version,
					   mddev->minor_version);
			}
6684 6685 6686 6687
		} else if (mddev->external)
			seq_printf(seq, " super external:%s",
				   mddev->metadata_type);
		else
6688
			seq_printf(seq, " super non-persistent");
L
Linus Torvalds 已提交
6689 6690

		if (mddev->pers) {
6691
			mddev->pers->status(seq, mddev);
L
Linus Torvalds 已提交
6692
	 		seq_printf(seq, "\n      ");
6693 6694
			if (mddev->pers->sync_request) {
				if (mddev->curr_resync > 2) {
6695
					status_resync(seq, mddev);
6696 6697 6698 6699 6700 6701
					seq_printf(seq, "\n      ");
				} else if (mddev->curr_resync == 1 || mddev->curr_resync == 2)
					seq_printf(seq, "\tresync=DELAYED\n      ");
				else if (mddev->recovery_cp < MaxSector)
					seq_printf(seq, "\tresync=PENDING\n      ");
			}
6702 6703 6704 6705 6706 6707 6708
		} else
			seq_printf(seq, "\n       ");

		if ((bitmap = mddev->bitmap)) {
			unsigned long chunk_kb;
			unsigned long flags;
			spin_lock_irqsave(&bitmap->lock, flags);
6709
			chunk_kb = mddev->bitmap_info.chunksize >> 10;
6710 6711 6712 6713 6714 6715
			seq_printf(seq, "bitmap: %lu/%lu pages [%luKB], "
				"%lu%s chunk",
				bitmap->pages - bitmap->missing_pages,
				bitmap->pages,
				(bitmap->pages - bitmap->missing_pages)
					<< (PAGE_SHIFT - 10),
6716
				chunk_kb ? chunk_kb : mddev->bitmap_info.chunksize,
6717
				chunk_kb ? "KB" : "B");
6718 6719
			if (bitmap->file) {
				seq_printf(seq, ", file: ");
6720
				seq_path(seq, &bitmap->file->f_path, " \t\n");
6721
			}
6722

6723 6724
			seq_printf(seq, "\n");
			spin_unlock_irqrestore(&bitmap->lock, flags);
L
Linus Torvalds 已提交
6725 6726 6727 6728 6729 6730 6731 6732 6733
		}

		seq_printf(seq, "\n");
	}
	mddev_unlock(mddev);
	
	return 0;
}

J
Jan Engelhardt 已提交
6734
static const struct seq_operations md_seq_ops = {
L
Linus Torvalds 已提交
6735 6736 6737 6738 6739 6740 6741 6742
	.start  = md_seq_start,
	.next   = md_seq_next,
	.stop   = md_seq_stop,
	.show   = md_seq_show,
};

static int md_seq_open(struct inode *inode, struct file *file)
{
6743
	struct seq_file *seq;
L
Linus Torvalds 已提交
6744 6745 6746
	int error;

	error = seq_open(file, &md_seq_ops);
6747
	if (error)
6748 6749 6750 6751
		return error;

	seq = file->private_data;
	seq->poll_event = atomic_read(&md_event_count);
L
Linus Torvalds 已提交
6752 6753 6754
	return error;
}

6755 6756
static unsigned int mdstat_poll(struct file *filp, poll_table *wait)
{
6757
	struct seq_file *seq = filp->private_data;
6758 6759 6760 6761 6762 6763 6764
	int mask;

	poll_wait(filp, &md_event_waiters, wait);

	/* always allow read */
	mask = POLLIN | POLLRDNORM;

6765
	if (seq->poll_event != atomic_read(&md_event_count))
6766 6767 6768 6769
		mask |= POLLERR | POLLPRI;
	return mask;
}

6770
static const struct file_operations md_seq_fops = {
6771
	.owner		= THIS_MODULE,
L
Linus Torvalds 已提交
6772 6773 6774
	.open           = md_seq_open,
	.read           = seq_read,
	.llseek         = seq_lseek,
6775
	.release	= seq_release_private,
6776
	.poll		= mdstat_poll,
L
Linus Torvalds 已提交
6777 6778
};

6779
int register_md_personality(struct mdk_personality *p)
L
Linus Torvalds 已提交
6780 6781
{
	spin_lock(&pers_lock);
6782 6783
	list_add_tail(&p->list, &pers_list);
	printk(KERN_INFO "md: %s personality registered for level %d\n", p->name, p->level);
L
Linus Torvalds 已提交
6784 6785 6786 6787
	spin_unlock(&pers_lock);
	return 0;
}

6788
int unregister_md_personality(struct mdk_personality *p)
L
Linus Torvalds 已提交
6789
{
6790
	printk(KERN_INFO "md: %s personality unregistered\n", p->name);
L
Linus Torvalds 已提交
6791
	spin_lock(&pers_lock);
6792
	list_del_init(&p->list);
L
Linus Torvalds 已提交
6793 6794 6795 6796
	spin_unlock(&pers_lock);
	return 0;
}

6797
static int is_mddev_idle(mddev_t *mddev, int init)
L
Linus Torvalds 已提交
6798 6799 6800
{
	mdk_rdev_t * rdev;
	int idle;
6801
	int curr_events;
L
Linus Torvalds 已提交
6802 6803

	idle = 1;
6804 6805
	rcu_read_lock();
	rdev_for_each_rcu(rdev, mddev) {
L
Linus Torvalds 已提交
6806
		struct gendisk *disk = rdev->bdev->bd_contains->bd_disk;
6807 6808 6809
		curr_events = (int)part_stat_read(&disk->part0, sectors[0]) +
			      (int)part_stat_read(&disk->part0, sectors[1]) -
			      atomic_read(&disk->sync_io);
6810 6811 6812 6813 6814 6815 6816 6817 6818 6819 6820 6821 6822 6823 6824 6825 6826 6827 6828 6829
		/* sync IO will cause sync_io to increase before the disk_stats
		 * as sync_io is counted when a request starts, and
		 * disk_stats is counted when it completes.
		 * So resync activity will cause curr_events to be smaller than
		 * when there was no such activity.
		 * non-sync IO will cause disk_stat to increase without
		 * increasing sync_io so curr_events will (eventually)
		 * be larger than it was before.  Once it becomes
		 * substantially larger, the test below will cause
		 * the array to appear non-idle, and resync will slow
		 * down.
		 * If there is a lot of outstanding resync activity when
		 * we set last_event to curr_events, then all that activity
		 * completing might cause the array to appear non-idle
		 * and resync will be slowed down even though there might
		 * not have been non-resync activity.  This will only
		 * happen once though.  'last_events' will soon reflect
		 * the state where there is little or no outstanding
		 * resync requests, and further resync activity will
		 * always make curr_events less than last_events.
6830
		 *
L
Linus Torvalds 已提交
6831
		 */
6832
		if (init || curr_events - rdev->last_events > 64) {
L
Linus Torvalds 已提交
6833 6834 6835 6836
			rdev->last_events = curr_events;
			idle = 0;
		}
	}
6837
	rcu_read_unlock();
L
Linus Torvalds 已提交
6838 6839 6840 6841 6842 6843 6844 6845 6846
	return idle;
}

void md_done_sync(mddev_t *mddev, int blocks, int ok)
{
	/* another "blocks" (512byte) blocks have been synced */
	atomic_sub(blocks, &mddev->recovery_active);
	wake_up(&mddev->recovery_wait);
	if (!ok) {
6847
		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
L
Linus Torvalds 已提交
6848 6849 6850 6851 6852 6853
		md_wakeup_thread(mddev->thread);
		// stop recovery, signal do_sync ....
	}
}


6854 6855
/* md_write_start(mddev, bi)
 * If we need to update some array metadata (e.g. 'active' flag
6856 6857
 * in superblock) before writing, schedule a superblock update
 * and wait for it to complete.
6858
 */
6859
void md_write_start(mddev_t *mddev, struct bio *bi)
L
Linus Torvalds 已提交
6860
{
6861
	int did_change = 0;
6862
	if (bio_data_dir(bi) != WRITE)
6863
		return;
6864

6865 6866 6867 6868 6869 6870
	BUG_ON(mddev->ro == 1);
	if (mddev->ro == 2) {
		/* need to switch to read/write */
		mddev->ro = 0;
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
		md_wakeup_thread(mddev->thread);
6871
		md_wakeup_thread(mddev->sync_thread);
6872
		did_change = 1;
6873
	}
6874
	atomic_inc(&mddev->writes_pending);
6875 6876
	if (mddev->safemode == 1)
		mddev->safemode = 0;
6877
	if (mddev->in_sync) {
6878
		spin_lock_irq(&mddev->write_lock);
6879 6880
		if (mddev->in_sync) {
			mddev->in_sync = 0;
6881
			set_bit(MD_CHANGE_CLEAN, &mddev->flags);
6882
			set_bit(MD_CHANGE_PENDING, &mddev->flags);
6883
			md_wakeup_thread(mddev->thread);
6884
			did_change = 1;
6885
		}
6886
		spin_unlock_irq(&mddev->write_lock);
6887
	}
6888
	if (did_change)
6889
		sysfs_notify_dirent_safe(mddev->sysfs_state);
6890 6891
	wait_event(mddev->sb_wait,
		   !test_bit(MD_CHANGE_PENDING, &mddev->flags));
L
Linus Torvalds 已提交
6892 6893 6894 6895 6896 6897 6898
}

void md_write_end(mddev_t *mddev)
{
	if (atomic_dec_and_test(&mddev->writes_pending)) {
		if (mddev->safemode == 2)
			md_wakeup_thread(mddev->thread);
6899
		else if (mddev->safemode_delay)
L
Linus Torvalds 已提交
6900 6901 6902 6903
			mod_timer(&mddev->safemode_timer, jiffies + mddev->safemode_delay);
	}
}

6904 6905 6906 6907 6908
/* md_allow_write(mddev)
 * Calling this ensures that the array is marked 'active' so that writes
 * may proceed without blocking.  It is important to call this before
 * attempting a GFP_KERNEL allocation while holding the mddev lock.
 * Must be called with mddev_lock held.
6909 6910 6911
 *
 * In the ->external case MD_CHANGE_CLEAN can not be cleared until mddev->lock
 * is dropped, so return -EAGAIN after notifying userspace.
6912
 */
6913
int md_allow_write(mddev_t *mddev)
6914 6915
{
	if (!mddev->pers)
6916
		return 0;
6917
	if (mddev->ro)
6918
		return 0;
6919
	if (!mddev->pers->sync_request)
6920
		return 0;
6921 6922 6923 6924 6925

	spin_lock_irq(&mddev->write_lock);
	if (mddev->in_sync) {
		mddev->in_sync = 0;
		set_bit(MD_CHANGE_CLEAN, &mddev->flags);
6926
		set_bit(MD_CHANGE_PENDING, &mddev->flags);
6927 6928 6929 6930 6931
		if (mddev->safemode_delay &&
		    mddev->safemode == 0)
			mddev->safemode = 1;
		spin_unlock_irq(&mddev->write_lock);
		md_update_sb(mddev, 0);
6932
		sysfs_notify_dirent_safe(mddev->sysfs_state);
6933 6934
	} else
		spin_unlock_irq(&mddev->write_lock);
6935

6936
	if (test_bit(MD_CHANGE_PENDING, &mddev->flags))
6937 6938 6939
		return -EAGAIN;
	else
		return 0;
6940 6941 6942
}
EXPORT_SYMBOL_GPL(md_allow_write);

L
Linus Torvalds 已提交
6943 6944
#define SYNC_MARKS	10
#define	SYNC_MARK_STEP	(3*HZ)
6945
void md_do_sync(mddev_t *mddev)
L
Linus Torvalds 已提交
6946 6947 6948 6949
{
	mddev_t *mddev2;
	unsigned int currspeed = 0,
		 window;
6950
	sector_t max_sectors,j, io_sectors;
L
Linus Torvalds 已提交
6951 6952 6953 6954 6955
	unsigned long mark[SYNC_MARKS];
	sector_t mark_cnt[SYNC_MARKS];
	int last_mark,m;
	struct list_head *tmp;
	sector_t last_check;
6956
	int skipped = 0;
6957
	mdk_rdev_t *rdev;
6958
	char *desc;
L
Linus Torvalds 已提交
6959 6960 6961 6962

	/* just incase thread restarts... */
	if (test_bit(MD_RECOVERY_DONE, &mddev->recovery))
		return;
6963 6964
	if (mddev->ro) /* never try to sync a read-only array */
		return;
L
Linus Torvalds 已提交
6965

6966 6967 6968 6969 6970 6971 6972 6973 6974 6975 6976 6977
	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
		if (test_bit(MD_RECOVERY_CHECK, &mddev->recovery))
			desc = "data-check";
		else if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
			desc = "requested-resync";
		else
			desc = "resync";
	} else if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
		desc = "reshape";
	else
		desc = "recovery";

L
Linus Torvalds 已提交
6978 6979 6980 6981 6982 6983 6984 6985 6986 6987 6988 6989 6990 6991 6992 6993 6994 6995 6996 6997
	/* we overload curr_resync somewhat here.
	 * 0 == not engaged in resync at all
	 * 2 == checking that there is no conflict with another sync
	 * 1 == like 2, but have yielded to allow conflicting resync to
	 *		commense
	 * other == active in resync - this many blocks
	 *
	 * Before starting a resync we must have set curr_resync to
	 * 2, and then checked that every "conflicting" array has curr_resync
	 * less than ours.  When we find one that is the same or higher
	 * we wait on resync_wait.  To avoid deadlock, we reduce curr_resync
	 * to 1 if we choose to yield (based arbitrarily on address of mddev structure).
	 * This will mean we have to start checking from the beginning again.
	 *
	 */

	do {
		mddev->curr_resync = 2;

	try_again:
6998
		if (kthread_should_stop())
6999
			set_bit(MD_RECOVERY_INTR, &mddev->recovery);
7000 7001

		if (test_bit(MD_RECOVERY_INTR, &mddev->recovery))
L
Linus Torvalds 已提交
7002
			goto skip;
7003
		for_each_mddev(mddev2, tmp) {
L
Linus Torvalds 已提交
7004 7005
			if (mddev2 == mddev)
				continue;
7006 7007 7008
			if (!mddev->parallel_resync
			&&  mddev2->curr_resync
			&&  match_mddev_units(mddev, mddev2)) {
L
Linus Torvalds 已提交
7009 7010 7011 7012 7013 7014 7015 7016 7017 7018 7019
				DEFINE_WAIT(wq);
				if (mddev < mddev2 && mddev->curr_resync == 2) {
					/* arbitrarily yield */
					mddev->curr_resync = 1;
					wake_up(&resync_wait);
				}
				if (mddev > mddev2 && mddev->curr_resync == 1)
					/* no need to wait here, we can wait the next
					 * time 'round when curr_resync == 2
					 */
					continue;
7020 7021 7022 7023 7024
				/* We need to wait 'interruptible' so as not to
				 * contribute to the load average, and not to
				 * be caught by 'softlockup'
				 */
				prepare_to_wait(&resync_wait, &wq, TASK_INTERRUPTIBLE);
7025
				if (!kthread_should_stop() &&
7026
				    mddev2->curr_resync >= mddev->curr_resync) {
7027 7028
					printk(KERN_INFO "md: delaying %s of %s"
					       " until %s has finished (they"
L
Linus Torvalds 已提交
7029
					       " share one or more physical units)\n",
7030
					       desc, mdname(mddev), mdname(mddev2));
L
Linus Torvalds 已提交
7031
					mddev_put(mddev2);
7032 7033
					if (signal_pending(current))
						flush_signals(current);
L
Linus Torvalds 已提交
7034 7035 7036 7037 7038 7039 7040 7041 7042
					schedule();
					finish_wait(&resync_wait, &wq);
					goto try_again;
				}
				finish_wait(&resync_wait, &wq);
			}
		}
	} while (mddev->curr_resync < 2);

7043
	j = 0;
7044
	if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
L
Linus Torvalds 已提交
7045
		/* resync follows the size requested by the personality,
7046
		 * which defaults to physical size, but can be virtual size
L
Linus Torvalds 已提交
7047 7048
		 */
		max_sectors = mddev->resync_max_sectors;
7049
		mddev->resync_mismatches = 0;
7050
		/* we don't use the checkpoint if there's a bitmap */
7051 7052 7053
		if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
			j = mddev->resync_min;
		else if (!mddev->bitmap)
7054
			j = mddev->recovery_cp;
7055

7056
	} else if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery))
7057
		max_sectors = mddev->dev_sectors;
7058
	else {
L
Linus Torvalds 已提交
7059
		/* recovery follows the physical size of devices */
7060
		max_sectors = mddev->dev_sectors;
7061
		j = MaxSector;
7062 7063
		rcu_read_lock();
		list_for_each_entry_rcu(rdev, &mddev->disks, same_set)
7064 7065 7066 7067 7068
			if (rdev->raid_disk >= 0 &&
			    !test_bit(Faulty, &rdev->flags) &&
			    !test_bit(In_sync, &rdev->flags) &&
			    rdev->recovery_offset < j)
				j = rdev->recovery_offset;
7069
		rcu_read_unlock();
7070
	}
L
Linus Torvalds 已提交
7071

7072 7073 7074
	printk(KERN_INFO "md: %s of RAID array %s\n", desc, mdname(mddev));
	printk(KERN_INFO "md: minimum _guaranteed_  speed:"
		" %d KB/sec/disk.\n", speed_min(mddev));
7075
	printk(KERN_INFO "md: using maximum available idle IO bandwidth "
7076 7077
	       "(but not more than %d KB/sec) for %s.\n",
	       speed_max(mddev), desc);
L
Linus Torvalds 已提交
7078

7079
	is_mddev_idle(mddev, 1); /* this initializes IO event counters */
7080

7081
	io_sectors = 0;
L
Linus Torvalds 已提交
7082 7083
	for (m = 0; m < SYNC_MARKS; m++) {
		mark[m] = jiffies;
7084
		mark_cnt[m] = io_sectors;
L
Linus Torvalds 已提交
7085 7086 7087 7088 7089 7090 7091 7092 7093
	}
	last_mark = 0;
	mddev->resync_mark = mark[last_mark];
	mddev->resync_mark_cnt = mark_cnt[last_mark];

	/*
	 * Tune reconstruction:
	 */
	window = 32*(PAGE_SIZE/512);
J
Jonathan Brassow 已提交
7094 7095
	printk(KERN_INFO "md: using %dk window, over a total of %lluk.\n",
		window/2, (unsigned long long)max_sectors/2);
L
Linus Torvalds 已提交
7096 7097 7098 7099 7100 7101

	atomic_set(&mddev->recovery_active, 0);
	last_check = 0;

	if (j>2) {
		printk(KERN_INFO 
7102 7103
		       "md: resuming %s of %s from checkpoint.\n",
		       desc, mdname(mddev));
L
Linus Torvalds 已提交
7104 7105
		mddev->curr_resync = j;
	}
7106
	mddev->curr_resync_completed = j;
L
Linus Torvalds 已提交
7107 7108

	while (j < max_sectors) {
7109
		sector_t sectors;
L
Linus Torvalds 已提交
7110

7111
		skipped = 0;
7112

7113 7114 7115 7116 7117 7118 7119
		if (!test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
		    ((mddev->curr_resync > mddev->curr_resync_completed &&
		      (mddev->curr_resync - mddev->curr_resync_completed)
		      > (max_sectors >> 4)) ||
		     (j - mddev->curr_resync_completed)*2
		     >= mddev->resync_max - mddev->curr_resync_completed
			    )) {
7120 7121 7122
			/* time to update curr_resync_completed */
			wait_event(mddev->recovery_wait,
				   atomic_read(&mddev->recovery_active) == 0);
7123
			mddev->curr_resync_completed = j;
7124
			set_bit(MD_CHANGE_CLEAN, &mddev->flags);
7125
			sysfs_notify(&mddev->kobj, NULL, "sync_completed");
7126
		}
7127

7128 7129 7130 7131 7132 7133 7134 7135 7136 7137
		while (j >= mddev->resync_max && !kthread_should_stop()) {
			/* As this condition is controlled by user-space,
			 * we can block indefinitely, so use '_interruptible'
			 * to avoid triggering warnings.
			 */
			flush_signals(current); /* just in case */
			wait_event_interruptible(mddev->recovery_wait,
						 mddev->resync_max > j
						 || kthread_should_stop());
		}
7138 7139 7140 7141

		if (kthread_should_stop())
			goto interrupted;

7142
		sectors = mddev->pers->sync_request(mddev, j, &skipped,
7143
						  currspeed < speed_min(mddev));
7144
		if (sectors == 0) {
7145
			set_bit(MD_RECOVERY_INTR, &mddev->recovery);
L
Linus Torvalds 已提交
7146 7147
			goto out;
		}
7148 7149 7150 7151 7152 7153

		if (!skipped) { /* actual IO requested */
			io_sectors += sectors;
			atomic_add(sectors, &mddev->recovery_active);
		}

7154 7155 7156
		if (test_bit(MD_RECOVERY_INTR, &mddev->recovery))
			break;

L
Linus Torvalds 已提交
7157 7158
		j += sectors;
		if (j>1) mddev->curr_resync = j;
7159
		mddev->curr_mark_cnt = io_sectors;
7160
		if (last_check == 0)
7161
			/* this is the earliest that rebuild will be
7162 7163 7164
			 * visible in /proc/mdstat
			 */
			md_new_event(mddev);
7165 7166

		if (last_check + window > io_sectors || j == max_sectors)
L
Linus Torvalds 已提交
7167 7168
			continue;

7169
		last_check = io_sectors;
L
Linus Torvalds 已提交
7170 7171 7172 7173 7174 7175 7176 7177
	repeat:
		if (time_after_eq(jiffies, mark[last_mark] + SYNC_MARK_STEP )) {
			/* step marks */
			int next = (last_mark+1) % SYNC_MARKS;

			mddev->resync_mark = mark[next];
			mddev->resync_mark_cnt = mark_cnt[next];
			mark[next] = jiffies;
7178
			mark_cnt[next] = io_sectors - atomic_read(&mddev->recovery_active);
L
Linus Torvalds 已提交
7179 7180 7181 7182
			last_mark = next;
		}


7183 7184 7185
		if (kthread_should_stop())
			goto interrupted;

L
Linus Torvalds 已提交
7186 7187 7188 7189 7190 7191 7192 7193 7194 7195 7196

		/*
		 * this loop exits only if either when we are slower than
		 * the 'hard' speed limit, or the system was IO-idle for
		 * a jiffy.
		 * the system might be non-idle CPU-wise, but we only care
		 * about not overloading the IO subsystem. (things like an
		 * e2fsck being done on the RAID array should execute fast)
		 */
		cond_resched();

7197 7198
		currspeed = ((unsigned long)(io_sectors-mddev->resync_mark_cnt))/2
			/((jiffies-mddev->resync_mark)/HZ +1) +1;
L
Linus Torvalds 已提交
7199

7200 7201
		if (currspeed > speed_min(mddev)) {
			if ((currspeed > speed_max(mddev)) ||
7202
					!is_mddev_idle(mddev, 0)) {
7203
				msleep(500);
L
Linus Torvalds 已提交
7204 7205 7206 7207
				goto repeat;
			}
		}
	}
7208
	printk(KERN_INFO "md: %s: %s done.\n",mdname(mddev), desc);
L
Linus Torvalds 已提交
7209 7210 7211 7212 7213 7214 7215
	/*
	 * this also signals 'finished resyncing' to md_stop
	 */
 out:
	wait_event(mddev->recovery_wait, !atomic_read(&mddev->recovery_active));

	/* tell personality that we are finished */
7216
	mddev->pers->sync_request(mddev, max_sectors, &skipped, 1);
L
Linus Torvalds 已提交
7217

7218
	if (!test_bit(MD_RECOVERY_CHECK, &mddev->recovery) &&
7219 7220 7221 7222 7223
	    mddev->curr_resync > 2) {
		if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
			if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) {
				if (mddev->curr_resync >= mddev->recovery_cp) {
					printk(KERN_INFO
7224 7225
					       "md: checkpointing %s of %s.\n",
					       desc, mdname(mddev));
7226 7227 7228 7229 7230 7231 7232
					mddev->recovery_cp = mddev->curr_resync;
				}
			} else
				mddev->recovery_cp = MaxSector;
		} else {
			if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery))
				mddev->curr_resync = MaxSector;
7233 7234
			rcu_read_lock();
			list_for_each_entry_rcu(rdev, &mddev->disks, same_set)
7235
				if (rdev->raid_disk >= 0 &&
7236
				    mddev->delta_disks >= 0 &&
7237 7238 7239 7240
				    !test_bit(Faulty, &rdev->flags) &&
				    !test_bit(In_sync, &rdev->flags) &&
				    rdev->recovery_offset < mddev->curr_resync)
					rdev->recovery_offset = mddev->curr_resync;
7241
			rcu_read_unlock();
7242
		}
L
Linus Torvalds 已提交
7243
	}
7244
	set_bit(MD_CHANGE_DEVS, &mddev->flags);
L
Linus Torvalds 已提交
7245 7246

 skip:
7247 7248 7249 7250 7251 7252 7253
	if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) {
		/* We completed so min/max setting can be forgotten if used. */
		if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
			mddev->resync_min = 0;
		mddev->resync_max = MaxSector;
	} else if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
		mddev->resync_min = mddev->curr_resync_completed;
L
Linus Torvalds 已提交
7254 7255 7256 7257
	mddev->curr_resync = 0;
	wake_up(&resync_wait);
	set_bit(MD_RECOVERY_DONE, &mddev->recovery);
	md_wakeup_thread(mddev->thread);
7258 7259 7260 7261 7262 7263 7264 7265 7266 7267 7268
	return;

 interrupted:
	/*
	 * got a signal, exit.
	 */
	printk(KERN_INFO
	       "md: md_do_sync() got signal ... exiting\n");
	set_bit(MD_RECOVERY_INTR, &mddev->recovery);
	goto out;

L
Linus Torvalds 已提交
7269
}
7270
EXPORT_SYMBOL_GPL(md_do_sync);
L
Linus Torvalds 已提交
7271

7272 7273 7274 7275 7276
static int remove_and_add_spares(mddev_t *mddev)
{
	mdk_rdev_t *rdev;
	int spares = 0;

7277 7278
	mddev->curr_resync_completed = 0;

7279
	list_for_each_entry(rdev, &mddev->disks, same_set)
7280
		if (rdev->raid_disk >= 0 &&
7281
		    !test_bit(Blocked, &rdev->flags) &&
7282 7283 7284 7285 7286
		    (test_bit(Faulty, &rdev->flags) ||
		     ! test_bit(In_sync, &rdev->flags)) &&
		    atomic_read(&rdev->nr_pending)==0) {
			if (mddev->pers->hot_remove_disk(
				    mddev, rdev->raid_disk)==0) {
7287
				sysfs_unlink_rdev(mddev, rdev);
7288 7289 7290 7291
				rdev->raid_disk = -1;
			}
		}

7292
	if (mddev->degraded) {
7293
		list_for_each_entry(rdev, &mddev->disks, same_set) {
7294
			if (rdev->raid_disk >= 0 &&
7295
			    !test_bit(In_sync, &rdev->flags) &&
7296
			    !test_bit(Faulty, &rdev->flags))
7297
				spares++;
7298 7299 7300
			if (rdev->raid_disk < 0
			    && !test_bit(Faulty, &rdev->flags)) {
				rdev->recovery_offset = 0;
7301 7302
				if (mddev->pers->
				    hot_add_disk(mddev, rdev) == 0) {
7303
					if (sysfs_link_rdev(mddev, rdev))
7304
						/* failure here is OK */;
7305 7306
					spares++;
					md_new_event(mddev);
7307
					set_bit(MD_CHANGE_DEVS, &mddev->flags);
7308 7309 7310
				} else
					break;
			}
7311
		}
7312 7313 7314
	}
	return spares;
}
7315 7316 7317 7318 7319 7320 7321 7322 7323 7324 7325 7326 7327 7328 7329 7330 7331 7332 7333 7334 7335 7336 7337 7338 7339 7340 7341 7342 7343 7344 7345 7346 7347 7348 7349 7350 7351

static void reap_sync_thread(mddev_t *mddev)
{
	mdk_rdev_t *rdev;

	/* resync has finished, collect result */
	md_unregister_thread(mddev->sync_thread);
	mddev->sync_thread = NULL;
	if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) &&
	    !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) {
		/* success...*/
		/* activate any spares */
		if (mddev->pers->spare_active(mddev))
			sysfs_notify(&mddev->kobj, NULL,
				     "degraded");
	}
	if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery) &&
	    mddev->pers->finish_reshape)
		mddev->pers->finish_reshape(mddev);
	md_update_sb(mddev, 1);

	/* if array is no-longer degraded, then any saved_raid_disk
	 * information must be scrapped
	 */
	if (!mddev->degraded)
		list_for_each_entry(rdev, &mddev->disks, same_set)
			rdev->saved_raid_disk = -1;

	clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
	clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
	clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
	clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
	clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
	/* flag recovery needed just to double check */
	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	sysfs_notify_dirent_safe(mddev->sysfs_action);
	md_new_event(mddev);
7352 7353
	if (mddev->event_work.func)
		queue_work(md_misc_wq, &mddev->event_work);
7354 7355
}

L
Linus Torvalds 已提交
7356 7357 7358 7359 7360 7361 7362 7363 7364 7365
/*
 * This routine is regularly called by all per-raid-array threads to
 * deal with generic issues like resync and super-block update.
 * Raid personalities that don't have a thread (linear/raid0) do not
 * need this as they never do any recovery or update the superblock.
 *
 * It does not do any resync itself, but rather "forks" off other threads
 * to do that as needed.
 * When it is determined that resync is needed, we set MD_RECOVERY_RUNNING in
 * "->recovery" and create a thread at ->sync_thread.
7366
 * When the thread finishes it sets MD_RECOVERY_DONE
L
Linus Torvalds 已提交
7367 7368 7369 7370 7371 7372 7373 7374 7375 7376 7377 7378 7379
 * and wakeups up this thread which will reap the thread and finish up.
 * This thread also removes any faulty devices (with nr_pending == 0).
 *
 * The overall approach is:
 *  1/ if the superblock needs updating, update it.
 *  2/ If a recovery thread is running, don't do anything else.
 *  3/ If recovery has finished, clean up, possibly marking spares active.
 *  4/ If there are any faulty devices, remove them.
 *  5/ If array is degraded, try to add spares devices
 *  6/ If array has spares or is not in-sync, start a resync thread.
 */
void md_check_recovery(mddev_t *mddev)
{
7380 7381 7382
	if (mddev->suspended)
		return;

7383
	if (mddev->bitmap)
7384
		bitmap_daemon_work(mddev);
L
Linus Torvalds 已提交
7385

7386
	if (signal_pending(current)) {
7387
		if (mddev->pers->sync_request && !mddev->external) {
7388 7389 7390 7391 7392 7393 7394
			printk(KERN_INFO "md: %s in immediate safe mode\n",
			       mdname(mddev));
			mddev->safemode = 2;
		}
		flush_signals(current);
	}

7395 7396
	if (mddev->ro && !test_bit(MD_RECOVERY_NEEDED, &mddev->recovery))
		return;
L
Linus Torvalds 已提交
7397
	if ( ! (
7398
		(mddev->flags & ~ (1<<MD_CHANGE_PENDING)) ||
L
Linus Torvalds 已提交
7399
		test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||
7400
		test_bit(MD_RECOVERY_DONE, &mddev->recovery) ||
7401
		(mddev->external == 0 && mddev->safemode == 1) ||
7402 7403
		(mddev->safemode == 2 && ! atomic_read(&mddev->writes_pending)
		 && !mddev->in_sync && mddev->recovery_cp == MaxSector)
L
Linus Torvalds 已提交
7404 7405
		))
		return;
7406

7407
	if (mddev_trylock(mddev)) {
7408
		int spares = 0;
7409

7410 7411 7412 7413
		if (mddev->ro) {
			/* Only thing we do on a ro array is remove
			 * failed devices.
			 */
7414 7415 7416 7417 7418 7419 7420 7421
			mdk_rdev_t *rdev;
			list_for_each_entry(rdev, &mddev->disks, same_set)
				if (rdev->raid_disk >= 0 &&
				    !test_bit(Blocked, &rdev->flags) &&
				    test_bit(Faulty, &rdev->flags) &&
				    atomic_read(&rdev->nr_pending)==0) {
					if (mddev->pers->hot_remove_disk(
						    mddev, rdev->raid_disk)==0) {
7422
						sysfs_unlink_rdev(mddev, rdev);
7423 7424 7425
						rdev->raid_disk = -1;
					}
				}
7426 7427 7428 7429
			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
			goto unlock;
		}

7430
		if (!mddev->external) {
7431
			int did_change = 0;
7432 7433 7434 7435 7436 7437
			spin_lock_irq(&mddev->write_lock);
			if (mddev->safemode &&
			    !atomic_read(&mddev->writes_pending) &&
			    !mddev->in_sync &&
			    mddev->recovery_cp == MaxSector) {
				mddev->in_sync = 1;
7438
				did_change = 1;
7439
				set_bit(MD_CHANGE_CLEAN, &mddev->flags);
7440 7441 7442 7443
			}
			if (mddev->safemode == 1)
				mddev->safemode = 0;
			spin_unlock_irq(&mddev->write_lock);
7444
			if (did_change)
7445
				sysfs_notify_dirent_safe(mddev->sysfs_state);
7446 7447
		}

7448 7449
		if (mddev->flags)
			md_update_sb(mddev, 0);
7450

L
Linus Torvalds 已提交
7451 7452 7453 7454 7455 7456 7457
		if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
		    !test_bit(MD_RECOVERY_DONE, &mddev->recovery)) {
			/* resync/recovery still happening */
			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
			goto unlock;
		}
		if (mddev->sync_thread) {
7458
			reap_sync_thread(mddev);
L
Linus Torvalds 已提交
7459 7460
			goto unlock;
		}
7461 7462 7463 7464 7465
		/* Set RUNNING before clearing NEEDED to avoid
		 * any transients in the value of "sync_action".
		 */
		set_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
		clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
7466 7467 7468 7469 7470
		/* Clear some bits that don't mean anything, but
		 * might be left set
		 */
		clear_bit(MD_RECOVERY_INTR, &mddev->recovery);
		clear_bit(MD_RECOVERY_DONE, &mddev->recovery);
L
Linus Torvalds 已提交
7471

7472 7473
		if (test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))
			goto unlock;
L
Linus Torvalds 已提交
7474 7475 7476 7477 7478 7479 7480
		/* no recovery is running.
		 * remove any failed drives, then
		 * add spares if possible.
		 * Spare are also removed and re-added, to allow
		 * the personality to fail the re-add.
		 */

7481
		if (mddev->reshape_position != MaxSector) {
7482 7483
			if (mddev->pers->check_reshape == NULL ||
			    mddev->pers->check_reshape(mddev) != 0)
7484 7485 7486
				/* Cannot proceed */
				goto unlock;
			set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
7487
			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
7488
		} else if ((spares = remove_and_add_spares(mddev))) {
7489 7490
			clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
			clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
7491
			clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
7492
			set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
7493 7494
		} else if (mddev->recovery_cp < MaxSector) {
			set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
7495
			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
7496 7497
		} else if (!test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
			/* nothing to be done ... */
L
Linus Torvalds 已提交
7498
			goto unlock;
7499

L
Linus Torvalds 已提交
7500
		if (mddev->pers->sync_request) {
7501 7502 7503 7504 7505 7506 7507
			if (spares && mddev->bitmap && ! mddev->bitmap->file) {
				/* We are adding a device or devices to an array
				 * which has the bitmap stored on all devices.
				 * So make sure all bitmap pages get written
				 */
				bitmap_write_all(mddev->bitmap);
			}
L
Linus Torvalds 已提交
7508 7509
			mddev->sync_thread = md_register_thread(md_do_sync,
								mddev,
7510
								"resync");
L
Linus Torvalds 已提交
7511 7512 7513 7514 7515
			if (!mddev->sync_thread) {
				printk(KERN_ERR "%s: could not start resync"
					" thread...\n", 
					mdname(mddev));
				/* leave the spares where they are, it shouldn't hurt */
7516 7517 7518 7519 7520
				clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
				clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
				clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
				clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
				clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
7521
			} else
L
Linus Torvalds 已提交
7522
				md_wakeup_thread(mddev->sync_thread);
7523
			sysfs_notify_dirent_safe(mddev->sysfs_action);
7524
			md_new_event(mddev);
L
Linus Torvalds 已提交
7525 7526
		}
	unlock:
7527 7528 7529 7530
		if (!mddev->sync_thread) {
			clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
			if (test_and_clear_bit(MD_RECOVERY_RECOVER,
					       &mddev->recovery))
7531
				if (mddev->sysfs_action)
7532
					sysfs_notify_dirent_safe(mddev->sysfs_action);
7533
		}
L
Linus Torvalds 已提交
7534 7535 7536 7537
		mddev_unlock(mddev);
	}
}

7538 7539
void md_wait_for_blocked_rdev(mdk_rdev_t *rdev, mddev_t *mddev)
{
7540
	sysfs_notify_dirent_safe(rdev->sysfs_state);
7541
	wait_event_timeout(rdev->blocked_wait,
7542 7543
			   !test_bit(Blocked, &rdev->flags) &&
			   !test_bit(BlockedBadBlocks, &rdev->flags),
7544 7545 7546 7547 7548
			   msecs_to_jiffies(5000));
	rdev_dec_pending(rdev, mddev);
}
EXPORT_SYMBOL(md_wait_for_blocked_rdev);

7549 7550 7551 7552 7553 7554 7555 7556 7557 7558 7559 7560 7561 7562 7563 7564 7565 7566 7567 7568 7569 7570 7571 7572 7573 7574 7575 7576 7577 7578 7579 7580 7581 7582 7583 7584 7585 7586 7587 7588 7589 7590 7591 7592 7593 7594 7595 7596 7597 7598 7599 7600 7601 7602 7603 7604 7605 7606 7607 7608 7609 7610 7611 7612 7613 7614 7615 7616 7617 7618 7619 7620 7621 7622 7623 7624 7625 7626 7627 7628 7629 7630 7631 7632 7633 7634 7635 7636 7637 7638 7639 7640 7641 7642 7643 7644 7645 7646 7647 7648 7649 7650 7651 7652 7653 7654 7655 7656 7657 7658 7659 7660 7661 7662 7663 7664 7665 7666 7667 7668 7669 7670 7671 7672 7673 7674 7675 7676 7677 7678 7679 7680 7681 7682 7683 7684 7685 7686 7687 7688 7689 7690 7691 7692 7693 7694 7695 7696 7697 7698 7699 7700 7701 7702 7703 7704 7705 7706 7707 7708 7709 7710 7711 7712 7713 7714 7715 7716 7717 7718 7719 7720 7721 7722 7723 7724 7725 7726 7727 7728 7729 7730 7731 7732 7733 7734 7735 7736 7737 7738 7739 7740 7741 7742 7743 7744 7745 7746 7747 7748 7749 7750 7751 7752 7753 7754 7755 7756 7757 7758 7759 7760 7761 7762 7763 7764 7765 7766 7767 7768 7769 7770 7771 7772 7773 7774 7775 7776 7777 7778 7779 7780 7781 7782 7783 7784 7785 7786 7787 7788

/* Bad block management.
 * We can record which blocks on each device are 'bad' and so just
 * fail those blocks, or that stripe, rather than the whole device.
 * Entries in the bad-block table are 64bits wide.  This comprises:
 * Length of bad-range, in sectors: 0-511 for lengths 1-512
 * Start of bad-range, sector offset, 54 bits (allows 8 exbibytes)
 *  A 'shift' can be set so that larger blocks are tracked and
 *  consequently larger devices can be covered.
 * 'Acknowledged' flag - 1 bit. - the most significant bit.
 *
 * Locking of the bad-block table uses a seqlock so md_is_badblock
 * might need to retry if it is very unlucky.
 * We will sometimes want to check for bad blocks in a bi_end_io function,
 * so we use the write_seqlock_irq variant.
 *
 * When looking for a bad block we specify a range and want to
 * know if any block in the range is bad.  So we binary-search
 * to the last range that starts at-or-before the given endpoint,
 * (or "before the sector after the target range")
 * then see if it ends after the given start.
 * We return
 *  0 if there are no known bad blocks in the range
 *  1 if there are known bad block which are all acknowledged
 * -1 if there are bad blocks which have not yet been acknowledged in metadata.
 * plus the start/length of the first bad section we overlap.
 */
int md_is_badblock(struct badblocks *bb, sector_t s, int sectors,
		   sector_t *first_bad, int *bad_sectors)
{
	int hi;
	int lo = 0;
	u64 *p = bb->page;
	int rv = 0;
	sector_t target = s + sectors;
	unsigned seq;

	if (bb->shift > 0) {
		/* round the start down, and the end up */
		s >>= bb->shift;
		target += (1<<bb->shift) - 1;
		target >>= bb->shift;
		sectors = target - s;
	}
	/* 'target' is now the first block after the bad range */

retry:
	seq = read_seqbegin(&bb->lock);

	hi = bb->count;

	/* Binary search between lo and hi for 'target'
	 * i.e. for the last range that starts before 'target'
	 */
	/* INVARIANT: ranges before 'lo' and at-or-after 'hi'
	 * are known not to be the last range before target.
	 * VARIANT: hi-lo is the number of possible
	 * ranges, and decreases until it reaches 1
	 */
	while (hi - lo > 1) {
		int mid = (lo + hi) / 2;
		sector_t a = BB_OFFSET(p[mid]);
		if (a < target)
			/* This could still be the one, earlier ranges
			 * could not. */
			lo = mid;
		else
			/* This and later ranges are definitely out. */
			hi = mid;
	}
	/* 'lo' might be the last that started before target, but 'hi' isn't */
	if (hi > lo) {
		/* need to check all range that end after 's' to see if
		 * any are unacknowledged.
		 */
		while (lo >= 0 &&
		       BB_OFFSET(p[lo]) + BB_LEN(p[lo]) > s) {
			if (BB_OFFSET(p[lo]) < target) {
				/* starts before the end, and finishes after
				 * the start, so they must overlap
				 */
				if (rv != -1 && BB_ACK(p[lo]))
					rv = 1;
				else
					rv = -1;
				*first_bad = BB_OFFSET(p[lo]);
				*bad_sectors = BB_LEN(p[lo]);
			}
			lo--;
		}
	}

	if (read_seqretry(&bb->lock, seq))
		goto retry;

	return rv;
}
EXPORT_SYMBOL_GPL(md_is_badblock);

/*
 * Add a range of bad blocks to the table.
 * This might extend the table, or might contract it
 * if two adjacent ranges can be merged.
 * We binary-search to find the 'insertion' point, then
 * decide how best to handle it.
 */
static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors,
			    int acknowledged)
{
	u64 *p;
	int lo, hi;
	int rv = 1;

	if (bb->shift < 0)
		/* badblocks are disabled */
		return 0;

	if (bb->shift) {
		/* round the start down, and the end up */
		sector_t next = s + sectors;
		s >>= bb->shift;
		next += (1<<bb->shift) - 1;
		next >>= bb->shift;
		sectors = next - s;
	}

	write_seqlock_irq(&bb->lock);

	p = bb->page;
	lo = 0;
	hi = bb->count;
	/* Find the last range that starts at-or-before 's' */
	while (hi - lo > 1) {
		int mid = (lo + hi) / 2;
		sector_t a = BB_OFFSET(p[mid]);
		if (a <= s)
			lo = mid;
		else
			hi = mid;
	}
	if (hi > lo && BB_OFFSET(p[lo]) > s)
		hi = lo;

	if (hi > lo) {
		/* we found a range that might merge with the start
		 * of our new range
		 */
		sector_t a = BB_OFFSET(p[lo]);
		sector_t e = a + BB_LEN(p[lo]);
		int ack = BB_ACK(p[lo]);
		if (e >= s) {
			/* Yes, we can merge with a previous range */
			if (s == a && s + sectors >= e)
				/* new range covers old */
				ack = acknowledged;
			else
				ack = ack && acknowledged;

			if (e < s + sectors)
				e = s + sectors;
			if (e - a <= BB_MAX_LEN) {
				p[lo] = BB_MAKE(a, e-a, ack);
				s = e;
			} else {
				/* does not all fit in one range,
				 * make p[lo] maximal
				 */
				if (BB_LEN(p[lo]) != BB_MAX_LEN)
					p[lo] = BB_MAKE(a, BB_MAX_LEN, ack);
				s = a + BB_MAX_LEN;
			}
			sectors = e - s;
		}
	}
	if (sectors && hi < bb->count) {
		/* 'hi' points to the first range that starts after 's'.
		 * Maybe we can merge with the start of that range */
		sector_t a = BB_OFFSET(p[hi]);
		sector_t e = a + BB_LEN(p[hi]);
		int ack = BB_ACK(p[hi]);
		if (a <= s + sectors) {
			/* merging is possible */
			if (e <= s + sectors) {
				/* full overlap */
				e = s + sectors;
				ack = acknowledged;
			} else
				ack = ack && acknowledged;

			a = s;
			if (e - a <= BB_MAX_LEN) {
				p[hi] = BB_MAKE(a, e-a, ack);
				s = e;
			} else {
				p[hi] = BB_MAKE(a, BB_MAX_LEN, ack);
				s = a + BB_MAX_LEN;
			}
			sectors = e - s;
			lo = hi;
			hi++;
		}
	}
	if (sectors == 0 && hi < bb->count) {
		/* we might be able to combine lo and hi */
		/* Note: 's' is at the end of 'lo' */
		sector_t a = BB_OFFSET(p[hi]);
		int lolen = BB_LEN(p[lo]);
		int hilen = BB_LEN(p[hi]);
		int newlen = lolen + hilen - (s - a);
		if (s >= a && newlen < BB_MAX_LEN) {
			/* yes, we can combine them */
			int ack = BB_ACK(p[lo]) && BB_ACK(p[hi]);
			p[lo] = BB_MAKE(BB_OFFSET(p[lo]), newlen, ack);
			memmove(p + hi, p + hi + 1,
				(bb->count - hi - 1) * 8);
			bb->count--;
		}
	}
	while (sectors) {
		/* didn't merge (it all).
		 * Need to add a range just before 'hi' */
		if (bb->count >= MD_MAX_BADBLOCKS) {
			/* No room for more */
			rv = 0;
			break;
		} else {
			int this_sectors = sectors;
			memmove(p + hi + 1, p + hi,
				(bb->count - hi) * 8);
			bb->count++;

			if (this_sectors > BB_MAX_LEN)
				this_sectors = BB_MAX_LEN;
			p[hi] = BB_MAKE(s, this_sectors, acknowledged);
			sectors -= this_sectors;
			s += this_sectors;
		}
	}

	bb->changed = 1;
7789 7790
	if (!acknowledged)
		bb->unacked_exist = 1;
7791 7792 7793 7794 7795 7796 7797 7798 7799 7800 7801 7802 7803 7804 7805 7806 7807 7808 7809 7810 7811 7812 7813 7814 7815 7816 7817 7818 7819 7820 7821 7822 7823 7824 7825 7826 7827 7828 7829 7830 7831 7832 7833 7834 7835 7836 7837 7838 7839 7840 7841 7842 7843 7844 7845 7846 7847 7848 7849 7850 7851 7852 7853 7854 7855 7856 7857 7858 7859 7860 7861 7862 7863 7864 7865 7866 7867 7868 7869 7870 7871 7872 7873 7874 7875 7876 7877 7878 7879 7880 7881 7882 7883 7884 7885 7886 7887 7888 7889 7890 7891 7892 7893 7894 7895 7896 7897 7898 7899 7900 7901 7902 7903 7904 7905 7906 7907 7908 7909 7910 7911 7912 7913 7914 7915 7916 7917 7918 7919 7920 7921 7922 7923 7924 7925 7926 7927 7928 7929 7930 7931 7932 7933 7934
	write_sequnlock_irq(&bb->lock);

	return rv;
}

int rdev_set_badblocks(mdk_rdev_t *rdev, sector_t s, int sectors,
		       int acknowledged)
{
	int rv = md_set_badblocks(&rdev->badblocks,
				  s + rdev->data_offset, sectors, acknowledged);
	if (rv) {
		/* Make sure they get written out promptly */
		set_bit(MD_CHANGE_CLEAN, &rdev->mddev->flags);
		md_wakeup_thread(rdev->mddev->thread);
	}
	return rv;
}
EXPORT_SYMBOL_GPL(rdev_set_badblocks);

/*
 * Remove a range of bad blocks from the table.
 * This may involve extending the table if we spilt a region,
 * but it must not fail.  So if the table becomes full, we just
 * drop the remove request.
 */
static int md_clear_badblocks(struct badblocks *bb, sector_t s, int sectors)
{
	u64 *p;
	int lo, hi;
	sector_t target = s + sectors;
	int rv = 0;

	if (bb->shift > 0) {
		/* When clearing we round the start up and the end down.
		 * This should not matter as the shift should align with
		 * the block size and no rounding should ever be needed.
		 * However it is better the think a block is bad when it
		 * isn't than to think a block is not bad when it is.
		 */
		s += (1<<bb->shift) - 1;
		s >>= bb->shift;
		target >>= bb->shift;
		sectors = target - s;
	}

	write_seqlock_irq(&bb->lock);

	p = bb->page;
	lo = 0;
	hi = bb->count;
	/* Find the last range that starts before 'target' */
	while (hi - lo > 1) {
		int mid = (lo + hi) / 2;
		sector_t a = BB_OFFSET(p[mid]);
		if (a < target)
			lo = mid;
		else
			hi = mid;
	}
	if (hi > lo) {
		/* p[lo] is the last range that could overlap the
		 * current range.  Earlier ranges could also overlap,
		 * but only this one can overlap the end of the range.
		 */
		if (BB_OFFSET(p[lo]) + BB_LEN(p[lo]) > target) {
			/* Partial overlap, leave the tail of this range */
			int ack = BB_ACK(p[lo]);
			sector_t a = BB_OFFSET(p[lo]);
			sector_t end = a + BB_LEN(p[lo]);

			if (a < s) {
				/* we need to split this range */
				if (bb->count >= MD_MAX_BADBLOCKS) {
					rv = 0;
					goto out;
				}
				memmove(p+lo+1, p+lo, (bb->count - lo) * 8);
				bb->count++;
				p[lo] = BB_MAKE(a, s-a, ack);
				lo++;
			}
			p[lo] = BB_MAKE(target, end - target, ack);
			/* there is no longer an overlap */
			hi = lo;
			lo--;
		}
		while (lo >= 0 &&
		       BB_OFFSET(p[lo]) + BB_LEN(p[lo]) > s) {
			/* This range does overlap */
			if (BB_OFFSET(p[lo]) < s) {
				/* Keep the early parts of this range. */
				int ack = BB_ACK(p[lo]);
				sector_t start = BB_OFFSET(p[lo]);
				p[lo] = BB_MAKE(start, s - start, ack);
				/* now low doesn't overlap, so.. */
				break;
			}
			lo--;
		}
		/* 'lo' is strictly before, 'hi' is strictly after,
		 * anything between needs to be discarded
		 */
		if (hi - lo > 1) {
			memmove(p+lo+1, p+hi, (bb->count - hi) * 8);
			bb->count -= (hi - lo - 1);
		}
	}

	bb->changed = 1;
out:
	write_sequnlock_irq(&bb->lock);
	return rv;
}

int rdev_clear_badblocks(mdk_rdev_t *rdev, sector_t s, int sectors)
{
	return md_clear_badblocks(&rdev->badblocks,
				  s + rdev->data_offset,
				  sectors);
}
EXPORT_SYMBOL_GPL(rdev_clear_badblocks);

/*
 * Acknowledge all bad blocks in a list.
 * This only succeeds if ->changed is clear.  It is used by
 * in-kernel metadata updates
 */
void md_ack_all_badblocks(struct badblocks *bb)
{
	if (bb->page == NULL || bb->changed)
		/* no point even trying */
		return;
	write_seqlock_irq(&bb->lock);

	if (bb->changed == 0) {
		u64 *p = bb->page;
		int i;
		for (i = 0; i < bb->count ; i++) {
			if (!BB_ACK(p[i])) {
				sector_t start = BB_OFFSET(p[i]);
				int len = BB_LEN(p[i]);
				p[i] = BB_MAKE(start, len, 1);
			}
		}
7935
		bb->unacked_exist = 0;
7936 7937 7938 7939 7940
	}
	write_sequnlock_irq(&bb->lock);
}
EXPORT_SYMBOL_GPL(md_ack_all_badblocks);

7941 7942 7943 7944 7945 7946 7947 7948 7949 7950 7951 7952 7953 7954 7955 7956 7957 7958 7959 7960 7961 7962 7963 7964 7965 7966 7967 7968 7969 7970 7971 7972 7973 7974 7975 7976 7977 7978 7979 7980 7981 7982
/* sysfs access to bad-blocks list.
 * We present two files.
 * 'bad-blocks' lists sector numbers and lengths of ranges that
 *    are recorded as bad.  The list is truncated to fit within
 *    the one-page limit of sysfs.
 *    Writing "sector length" to this file adds an acknowledged
 *    bad block list.
 * 'unacknowledged-bad-blocks' lists bad blocks that have not yet
 *    been acknowledged.  Writing to this file adds bad blocks
 *    without acknowledging them.  This is largely for testing.
 */

static ssize_t
badblocks_show(struct badblocks *bb, char *page, int unack)
{
	size_t len;
	int i;
	u64 *p = bb->page;
	unsigned seq;

	if (bb->shift < 0)
		return 0;

retry:
	seq = read_seqbegin(&bb->lock);

	len = 0;
	i = 0;

	while (len < PAGE_SIZE && i < bb->count) {
		sector_t s = BB_OFFSET(p[i]);
		unsigned int length = BB_LEN(p[i]);
		int ack = BB_ACK(p[i]);
		i++;

		if (unack && ack)
			continue;

		len += snprintf(page+len, PAGE_SIZE-len, "%llu %u\n",
				(unsigned long long)s << bb->shift,
				length << bb->shift);
	}
7983 7984
	if (unack && len == 0)
		bb->unacked_exist = 0;
7985 7986 7987 7988 7989 7990 7991 7992 7993 7994 7995 7996 7997 7998 7999 8000 8001 8002 8003 8004 8005 8006 8007 8008 8009 8010 8011 8012 8013 8014 8015 8016 8017 8018 8019 8020 8021 8022 8023 8024 8025 8026 8027 8028 8029 8030 8031 8032 8033 8034

	if (read_seqretry(&bb->lock, seq))
		goto retry;

	return len;
}

#define DO_DEBUG 1

static ssize_t
badblocks_store(struct badblocks *bb, const char *page, size_t len, int unack)
{
	unsigned long long sector;
	int length;
	char newline;
#ifdef DO_DEBUG
	/* Allow clearing via sysfs *only* for testing/debugging.
	 * Normally only a successful write may clear a badblock
	 */
	int clear = 0;
	if (page[0] == '-') {
		clear = 1;
		page++;
	}
#endif /* DO_DEBUG */

	switch (sscanf(page, "%llu %d%c", &sector, &length, &newline)) {
	case 3:
		if (newline != '\n')
			return -EINVAL;
	case 2:
		if (length <= 0)
			return -EINVAL;
		break;
	default:
		return -EINVAL;
	}

#ifdef DO_DEBUG
	if (clear) {
		md_clear_badblocks(bb, sector, length);
		return len;
	}
#endif /* DO_DEBUG */
	if (md_set_badblocks(bb, sector, length, !unack))
		return len;
	else
		return -ENOSPC;
}

8035 8036
static int md_notify_reboot(struct notifier_block *this,
			    unsigned long code, void *x)
L
Linus Torvalds 已提交
8037 8038 8039 8040 8041 8042 8043 8044
{
	struct list_head *tmp;
	mddev_t *mddev;

	if ((code == SYS_DOWN) || (code == SYS_HALT) || (code == SYS_POWER_OFF)) {

		printk(KERN_INFO "md: stopping all md devices.\n");

8045
		for_each_mddev(mddev, tmp)
8046
			if (mddev_trylock(mddev)) {
8047 8048 8049 8050
				/* Force a switch to readonly even array
				 * appears to still be in use.  Hence
				 * the '100'.
				 */
8051
				md_set_readonly(mddev, 100);
8052 8053
				mddev_unlock(mddev);
			}
L
Linus Torvalds 已提交
8054 8055 8056 8057 8058 8059 8060 8061 8062 8063 8064
		/*
		 * certain more exotic SCSI devices are known to be
		 * volatile wrt too early system reboots. While the
		 * right place to handle this issue is the given
		 * driver, we do want to have a safe RAID driver ...
		 */
		mdelay(1000*1);
	}
	return NOTIFY_DONE;
}

8065
static struct notifier_block md_notifier = {
L
Linus Torvalds 已提交
8066 8067 8068 8069 8070 8071 8072 8073 8074
	.notifier_call	= md_notify_reboot,
	.next		= NULL,
	.priority	= INT_MAX, /* before any real devices */
};

static void md_geninit(void)
{
	dprintk("md: sizeof(mdp_super_t) = %d\n", (int)sizeof(mdp_super_t));

8075
	proc_create("mdstat", S_IRUGO, NULL, &md_seq_fops);
L
Linus Torvalds 已提交
8076 8077
}

8078
static int __init md_init(void)
L
Linus Torvalds 已提交
8079
{
8080 8081
	int ret = -ENOMEM;

8082
	md_wq = alloc_workqueue("md", WQ_MEM_RECLAIM, 0);
8083 8084 8085 8086 8087 8088 8089 8090 8091 8092 8093 8094 8095 8096
	if (!md_wq)
		goto err_wq;

	md_misc_wq = alloc_workqueue("md_misc", 0, 0);
	if (!md_misc_wq)
		goto err_misc_wq;

	if ((ret = register_blkdev(MD_MAJOR, "md")) < 0)
		goto err_md;

	if ((ret = register_blkdev(0, "mdp")) < 0)
		goto err_mdp;
	mdp_major = ret;

8097
	blk_register_region(MKDEV(MD_MAJOR, 0), 1UL<<MINORBITS, THIS_MODULE,
8098 8099
			    md_probe, NULL, NULL);
	blk_register_region(MKDEV(mdp_major, 0), 1UL<<MINORBITS, THIS_MODULE,
L
Linus Torvalds 已提交
8100 8101 8102
			    md_probe, NULL, NULL);

	register_reboot_notifier(&md_notifier);
8103
	raid_table_header = register_sysctl_table(raid_root_table);
L
Linus Torvalds 已提交
8104 8105

	md_geninit();
8106
	return 0;
L
Linus Torvalds 已提交
8107

8108 8109 8110 8111 8112 8113 8114 8115 8116
err_mdp:
	unregister_blkdev(MD_MAJOR, "md");
err_md:
	destroy_workqueue(md_misc_wq);
err_misc_wq:
	destroy_workqueue(md_wq);
err_wq:
	return ret;
}
L
Linus Torvalds 已提交
8117 8118 8119 8120 8121 8122 8123

#ifndef MODULE

/*
 * Searches all registered partitions for autorun RAID arrays
 * at boot time.
 */
8124 8125 8126 8127 8128 8129

static LIST_HEAD(all_detected_devices);
struct detected_devices_node {
	struct list_head list;
	dev_t dev;
};
L
Linus Torvalds 已提交
8130 8131 8132

void md_autodetect_dev(dev_t dev)
{
8133 8134 8135 8136 8137 8138 8139 8140 8141 8142
	struct detected_devices_node *node_detected_dev;

	node_detected_dev = kzalloc(sizeof(*node_detected_dev), GFP_KERNEL);
	if (node_detected_dev) {
		node_detected_dev->dev = dev;
		list_add_tail(&node_detected_dev->list, &all_detected_devices);
	} else {
		printk(KERN_CRIT "md: md_autodetect_dev: kzalloc failed"
			", skipping dev(%d,%d)\n", MAJOR(dev), MINOR(dev));
	}
L
Linus Torvalds 已提交
8143 8144 8145 8146 8147 8148
}


static void autostart_arrays(int part)
{
	mdk_rdev_t *rdev;
8149 8150 8151
	struct detected_devices_node *node_detected_dev;
	dev_t dev;
	int i_scanned, i_passed;
L
Linus Torvalds 已提交
8152

8153 8154
	i_scanned = 0;
	i_passed = 0;
L
Linus Torvalds 已提交
8155

8156
	printk(KERN_INFO "md: Autodetecting RAID arrays.\n");
L
Linus Torvalds 已提交
8157

8158 8159 8160 8161 8162 8163 8164
	while (!list_empty(&all_detected_devices) && i_scanned < INT_MAX) {
		i_scanned++;
		node_detected_dev = list_entry(all_detected_devices.next,
					struct detected_devices_node, list);
		list_del(&node_detected_dev->list);
		dev = node_detected_dev->dev;
		kfree(node_detected_dev);
8165
		rdev = md_import_device(dev,0, 90);
L
Linus Torvalds 已提交
8166 8167 8168
		if (IS_ERR(rdev))
			continue;

8169
		if (test_bit(Faulty, &rdev->flags)) {
L
Linus Torvalds 已提交
8170 8171 8172
			MD_BUG();
			continue;
		}
8173
		set_bit(AutoDetected, &rdev->flags);
L
Linus Torvalds 已提交
8174
		list_add(&rdev->same_set, &pending_raid_disks);
8175
		i_passed++;
L
Linus Torvalds 已提交
8176
	}
8177 8178 8179

	printk(KERN_INFO "md: Scanned %d and added %d devices.\n",
						i_scanned, i_passed);
L
Linus Torvalds 已提交
8180 8181 8182 8183

	autorun_devices(part);
}

8184
#endif /* !MODULE */
L
Linus Torvalds 已提交
8185 8186 8187 8188 8189

static __exit void md_exit(void)
{
	mddev_t *mddev;
	struct list_head *tmp;
8190

8191
	blk_unregister_region(MKDEV(MD_MAJOR,0), 1U << MINORBITS);
8192
	blk_unregister_region(MKDEV(mdp_major,0), 1U << MINORBITS);
L
Linus Torvalds 已提交
8193

8194
	unregister_blkdev(MD_MAJOR,"md");
L
Linus Torvalds 已提交
8195 8196 8197 8198
	unregister_blkdev(mdp_major, "mdp");
	unregister_reboot_notifier(&md_notifier);
	unregister_sysctl_table(raid_table_header);
	remove_proc_entry("mdstat", NULL);
8199
	for_each_mddev(mddev, tmp) {
L
Linus Torvalds 已提交
8200
		export_array(mddev);
8201
		mddev->hold_active = 0;
L
Linus Torvalds 已提交
8202
	}
8203 8204
	destroy_workqueue(md_misc_wq);
	destroy_workqueue(md_wq);
L
Linus Torvalds 已提交
8205 8206
}

8207
subsys_initcall(md_init);
L
Linus Torvalds 已提交
8208 8209
module_exit(md_exit)

8210 8211 8212 8213 8214 8215 8216 8217 8218 8219
static int get_ro(char *buffer, struct kernel_param *kp)
{
	return sprintf(buffer, "%d", start_readonly);
}
static int set_ro(const char *val, struct kernel_param *kp)
{
	char *e;
	int num = simple_strtoul(val, &e, 10);
	if (*val && (*e == '\0' || *e == '\n')) {
		start_readonly = num;
8220
		return 0;
8221 8222 8223 8224
	}
	return -EINVAL;
}

8225 8226
module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR);
module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR);
8227

8228
module_param_call(new_array, add_named_array, NULL, NULL, S_IWUSR);
8229

L
Linus Torvalds 已提交
8230 8231 8232 8233 8234 8235 8236 8237 8238 8239 8240
EXPORT_SYMBOL(register_md_personality);
EXPORT_SYMBOL(unregister_md_personality);
EXPORT_SYMBOL(md_error);
EXPORT_SYMBOL(md_done_sync);
EXPORT_SYMBOL(md_write_start);
EXPORT_SYMBOL(md_write_end);
EXPORT_SYMBOL(md_register_thread);
EXPORT_SYMBOL(md_unregister_thread);
EXPORT_SYMBOL(md_wakeup_thread);
EXPORT_SYMBOL(md_check_recovery);
MODULE_LICENSE("GPL");
8241
MODULE_DESCRIPTION("MD RAID framework");
8242
MODULE_ALIAS("md");
8243
MODULE_ALIAS_BLOCKDEV_MAJOR(MD_MAJOR);
新手
引导
客服 返回
顶部