md.h 22.9 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
   md.h : kernel internal structure of the Linux MD driver
L
Linus Torvalds 已提交
3
          Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
4

L
Linus Torvalds 已提交
5 6 7 8
   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.
9

L
Linus Torvalds 已提交
10 11
   You should have received a copy of the GNU General Public License
   (for example /usr/src/linux/COPYING); if not, write to the Free
12
   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
L
Linus Torvalds 已提交
13 14
*/

C
Christoph Hellwig 已提交
15 16 17 18
#ifndef _MD_MD_H
#define _MD_MD_H

#include <linux/blkdev.h>
19
#include <linux/backing-dev.h>
20
#include <linux/badblocks.h>
C
Christoph Hellwig 已提交
21 22 23 24 25 26 27
#include <linux/kobject.h>
#include <linux/list.h>
#include <linux/mm.h>
#include <linux/mutex.h>
#include <linux/timer.h>
#include <linux/wait.h>
#include <linux/workqueue.h>
28
#include "md-cluster.h"
29

L
Linus Torvalds 已提交
30 31 32 33 34
#define MaxSector (~(sector_t)0)

/*
 * MD's 'extended' device
 */
35
struct md_rdev {
L
Linus Torvalds 已提交
36 37
	struct list_head same_set;	/* RAID devices within the same set */

38
	sector_t sectors;		/* Device size (in 512bytes sectors) */
39
	struct mddev *mddev;		/* RAID array if running */
N
NeilBrown 已提交
40
	int last_events;		/* IO event timestamp */
L
Linus Torvalds 已提交
41

42 43 44 45 46 47
	/*
	 * If meta_bdev is non-NULL, it means that a separate device is
	 * being used to store the metadata (superblock/bitmap) which
	 * would otherwise be contained on the same device as the data (bdev).
	 */
	struct block_device *meta_bdev;
L
Linus Torvalds 已提交
48 49
	struct block_device *bdev;	/* block device handle */

50
	struct page	*sb_page, *bb_page;
L
Linus Torvalds 已提交
51
	int		sb_loaded;
52
	__u64		sb_events;
L
Linus Torvalds 已提交
53
	sector_t	data_offset;	/* start of data in array */
54
	sector_t	new_data_offset;/* only relevant while reshaping */
55
	sector_t	sb_start;	/* offset of the super block (in 512byte sectors) */
56
	int		sb_size;	/* bytes in the superblock */
L
Linus Torvalds 已提交
57 58
	int		preferred_minor;	/* autorun support */

59 60
	struct kobject	kobj;

L
Linus Torvalds 已提交
61 62 63 64 65 66 67 68 69 70 71
	/* A device can be in one of three states based on two flags:
	 * Not working:   faulty==1 in_sync==0
	 * Fully working: faulty==0 in_sync==1
	 * Working, but not
	 * in sync with array
	 *                faulty==0 in_sync==0
	 *
	 * It can never have faulty==1, in_sync==1
	 * This reduces the burden of testing multiple flags in many cases
	 */

72
	unsigned long	flags;	/* bit set of 'enum flag_bits' bits. */
73
	wait_queue_head_t blocked_wait;
74

L
Linus Torvalds 已提交
75 76
	int desc_nr;			/* descriptor index in the superblock */
	int raid_disk;			/* role of device in array */
77 78 79
	int new_raid_disk;		/* role that the device will have in
					 * the array after a level-change completes.
					 */
80 81 82 83
	int saved_raid_disk;		/* role that device used to have in the
					 * array and could again if we did a partial
					 * resync from the bitmap
					 */
84 85
	union {
		sector_t recovery_offset;/* If this device has been partially
86 87 88
					 * recovered, this is where we were
					 * up to.
					 */
89 90 91 92 93
		sector_t journal_tail;	/* If this device is a journal device,
					 * this is the journal tail (journal
					 * recovery start point)
					 */
	};
L
Linus Torvalds 已提交
94 95 96 97 98

	atomic_t	nr_pending;	/* number of pending requests.
					 * only maintained for arrays that
					 * support hot removal
					 */
99 100 101
	atomic_t	read_errors;	/* number of consecutive read errors that
					 * we have tried to ignore.
					 */
102
	time64_t	last_read_error;	/* monotonic time since our
103 104
						 * last read error
						 */
105 106 107 108
	atomic_t	corrected_errors; /* number of corrected read errors,
					   * for reporting to userspace and storing
					   * in superblock.
					   */
109
	struct work_struct del_work;	/* used for delayed sysfs removal */
110

111
	struct kernfs_node *sysfs_state; /* handle for 'state'
112
					   * sysfs entry */
113

114
	struct badblocks badblocks;
L
Linus Torvalds 已提交
115
};
116 117 118
enum flag_bits {
	Faulty,			/* device is known to have a fault */
	In_sync,		/* device is in_sync with rest of array */
119 120 121
	Bitmap_sync,		/* ..actually, not quite In_sync.  Need a
				 * bitmap-based recovery to get fully in sync
				 */
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
	WriteMostly,		/* Avoid reading if at all possible */
	AutoDetected,		/* added by auto-detect */
	Blocked,		/* An error occurred but has not yet
				 * been acknowledged by the metadata
				 * handler, so don't allow writes
				 * until it is cleared */
	WriteErrorSeen,		/* A write error has been seen on this
				 * device
				 */
	FaultRecorded,		/* Intermediate state for clearing
				 * Blocked.  The Fault is/will-be
				 * recorded in the metadata, but that
				 * metadata hasn't been stored safely
				 * on disk yet.
				 */
	BlockedBadBlocks,	/* A writer is blocked because they
				 * found an unacknowledged bad-block.
				 * This can safely be cleared at any
				 * time, and the writer will re-check.
				 * It may be set at any time, and at
				 * worst the writer will timeout and
				 * re-check.  So setting it as
				 * accurately as possible is good, but
				 * not absolutely critical.
				 */
	WantReplacement,	/* This device is a candidate to be
				 * hot-replaced, either because it has
				 * reported some faults, or because
				 * of explicit request.
				 */
	Replacement,		/* This device is a replacement for
				 * a want_replacement device with same
				 * raid_disk number.
				 */
156 157 158 159
	Candidate,		/* For clustered environments only:
				 * This device is seen locally but not
				 * by the whole cluster
				 */
160 161 162 163 164
	Journal,		/* This device is used as journal for
				 * raid-5/6.
				 * Usually, this device should be faster
				 * than other devices in the array
				 */
165
	ClusterRemove,
166 167 168 169 170
	RemoveSynchronized,	/* synchronize_rcu() was called after
				 * this device was known to be faulty,
				 * so it is safe to remove without
				 * another synchronize_rcu() call.
				 */
171 172 173
	ExternalBbl,            /* External metadata provides bad
				 * block management for a disk
				 */
174
};
L
Linus Torvalds 已提交
175

176
static inline int is_badblock(struct md_rdev *rdev, sector_t s, int sectors,
177 178 179
			      sector_t *first_bad, int *bad_sectors)
{
	if (unlikely(rdev->badblocks.count)) {
180
		int rv = badblocks_check(&rdev->badblocks, rdev->data_offset + s,
181 182 183 184 185 186 187 188
					sectors,
					first_bad, bad_sectors);
		if (rv)
			*first_bad -= rdev->data_offset;
		return rv;
	}
	return 0;
}
189
extern int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
190 191 192
			      int is_new);
extern int rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
				int is_new);
G
Goldwyn Rodrigues 已提交
193 194
struct md_cluster_info;

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
enum mddev_flags {
	MD_CHANGE_DEVS,		/* Some device status has changed */
	MD_CHANGE_CLEAN,	/* transition to or from 'clean' */
	MD_CHANGE_PENDING,	/* switch from 'clean' to 'active' in progress */
	MD_ARRAY_FIRST_USE,	/* First use of array, needs initialization */
	MD_CLOSING,		/* If set, we are closing the array, do not open
				 * it then */
	MD_JOURNAL_CLEAN,	/* A raid with journal is already clean */
	MD_HAS_JOURNAL,		/* The raid array has journal feature set */
	MD_RELOAD_SB,		/* Reload the superblock because another node
				 * updated it.
				 */
	MD_CLUSTER_RESYNC_LOCKED, /* cluster raid only, which means node
				   * already took resync lock, need to
				   * release the lock */
};
#define MD_UPDATE_SB_FLAGS (BIT(MD_CHANGE_DEVS) | \
			    BIT(MD_CHANGE_CLEAN) | \
			    BIT(MD_CHANGE_PENDING))	/* If these are set, md_update_sb needed */
214
struct mddev {
L
Linus Torvalds 已提交
215
	void				*private;
216
	struct md_personality		*pers;
L
Linus Torvalds 已提交
217 218
	dev_t				unit;
	int				md_minor;
219
	struct list_head		disks;
220 221
	unsigned long			flags;

222 223
	int				suspended;
	atomic_t			active_io;
L
Linus Torvalds 已提交
224
	int				ro;
225 226 227 228
	int				sysfs_active; /* set when sysfs deletes
						       * are happening, so run/
						       * takeover/stop are not safe
						       */
L
Linus Torvalds 已提交
229 230
	struct gendisk			*gendisk;

231
	struct kobject			kobj;
232 233
	int				hold_active;
#define	UNTIL_IOCTL	1
234
#define	UNTIL_STOP	2
235

L
Linus Torvalds 已提交
236 237 238 239 240
	/* Superblock information */
	int				major_version,
					minor_version,
					patch_version;
	int				persistent;
241
	int				external;	/* metadata is
242 243
							 * managed externally */
	char				metadata_type[17]; /* externally set*/
244
	int				chunk_sectors;
245
	time64_t			ctime, utime;
L
Linus Torvalds 已提交
246
	int				level, layout;
247
	char				clevel[16];
L
Linus Torvalds 已提交
248 249
	int				raid_disks;
	int				max_disks;
250
	sector_t			dev_sectors;	/* used size of
A
Andre Noll 已提交
251
							 * component devices */
252
	sector_t			array_sectors; /* exported array size */
D
Dan Williams 已提交
253 254
	int				external_size; /* size managed
							* externally */
L
Linus Torvalds 已提交
255
	__u64				events;
256 257 258 259 260 261
	/* If the last 'event' was simply a clean->dirty transition, and
	 * we didn't write it to the spares, then it is safe and simple
	 * to just decrement the event count on a dirty->clean transition.
	 * So we record that possibility here.
	 */
	int				can_decrease_events;
L
Linus Torvalds 已提交
262 263 264

	char				uuid[16];

265 266 267 268 269 270
	/* If the array is being reshaped, we need to record the
	 * new shape and an indication of where we are up to.
	 * This is written to the superblock.
	 * If reshape_position is MaxSector, then no reshape is happening (yet).
	 */
	sector_t			reshape_position;
271 272
	int				delta_disks, new_level, new_layout;
	int				new_chunk_sectors;
273
	int				reshape_backwards;
274

275 276
	struct md_thread		*thread;	/* management thread */
	struct md_thread		*sync_thread;	/* doing resync or reconstruct */
277 278 279 280 281 282 283 284

	/* 'last_sync_action' is initialized to "none".  It is set when a
	 * sync operation (i.e "data-check", "requested-resync", "resync",
	 * "recovery", or "reshape") is started.  It holds this value even
	 * when the sync thread is "frozen" (interrupted) or "idle" (stopped
	 * or finished).  It is overwritten when a new sync operation is begun.
	 */
	char				*last_sync_action;
285
	sector_t			curr_resync;	/* last block scheduled */
286 287 288 289 290 291 292
	/* As resync requests can complete out of order, we cannot easily track
	 * how much resync has been completed.  So we occasionally pause until
	 * everything completes, then set curr_resync_completed to curr_resync.
	 * As such it may be well behind the real resync mark, but it is a value
	 * we are certain of.
	 */
	sector_t			curr_resync_completed;
L
Linus Torvalds 已提交
293 294
	unsigned long			resync_mark;	/* a recent timestamp */
	sector_t			resync_mark_cnt;/* blocks written at resync_mark */
295
	sector_t			curr_mark_cnt; /* blocks scheduled now */
L
Linus Torvalds 已提交
296 297

	sector_t			resync_max_sectors; /* may be set by personality */
298

299
	atomic64_t			resync_mismatches; /* count of sectors where
300 301
							    * parity/replica mismatch found
							    */
302 303 304 305

	/* allow user-space to request suspension of IO to regions of the array */
	sector_t			suspend_lo;
	sector_t			suspend_hi;
306 307 308 309
	/* if zero, use the system-wide default */
	int				sync_speed_min;
	int				sync_speed_max;

310 311 312
	/* resync even though the same disks are shared among md-devices */
	int				parallel_resync;

313
	int				ok_start_degraded;
314

L
Linus Torvalds 已提交
315
	unsigned long			recovery;
316 317 318 319 320 321
	/* If a RAID personality determines that recovery (of a particular
	 * device) will fail due to a read error on the source device, it
	 * takes a copy of this number and does not attempt recovery again
	 * until this number changes.
	 */
	int				recovery_disabled;
L
Linus Torvalds 已提交
322 323

	int				in_sync;	/* know to not need resync */
N
NeilBrown 已提交
324 325 326 327 328 329 330 331 332 333
	/* 'open_mutex' avoids races between 'md_open' and 'do_md_stop', so
	 * that we are never stopping an array while it is open.
	 * 'reconfig_mutex' protects all other reconfiguration.
	 * These locks are separate due to conflicting interactions
	 * with bdev->bd_mutex.
	 * Lock ordering is:
	 *  reconfig_mutex -> bd_mutex : e.g. do_md_run -> revalidate_disk
	 *  bd_mutex -> open_mutex:  e.g. __blkdev_get -> md_open
	 */
	struct mutex			open_mutex;
334
	struct mutex			reconfig_mutex;
335 336
	atomic_t			active;		/* general refcount */
	atomic_t			openers;	/* number of active opens */
L
Linus Torvalds 已提交
337

338 339
	int				changed;	/* True if we might need to
							 * reread partition info */
L
Linus Torvalds 已提交
340 341 342 343 344 345 346
	int				degraded;	/* whether md should consider
							 * adding a spare
							 */

	atomic_t			recovery_active; /* blocks scheduled, but not written */
	wait_queue_head_t		recovery_wait;
	sector_t			recovery_cp;
347 348
	sector_t			resync_min;	/* user requested sync
							 * starts here */
349 350
	sector_t			resync_max;	/* resync should pause
							 * when it gets here */
351

352
	struct kernfs_node		*sysfs_state;	/* handle for 'array_state'
353 354
							 * file in sysfs.
							 */
355
	struct kernfs_node		*sysfs_action;  /* handle for 'sync_action' */
356

357 358
	struct work_struct del_work;	/* used for delayed sysfs removal */

359 360 361 362 363
	/* "lock" protects:
	 *   flush_bio transition from NULL to !NULL
	 *   rdev superblocks, events
	 *   clearing MD_CHANGE_*
	 *   in_sync - and related safemode and MD_CHANGE changes
364
	 *   pers (also protected by reconfig_mutex and pending IO).
365
	 *   clearing ->bitmap
366
	 *   clearing ->bitmap_info.file
367 368
	 *   changing ->resync_{min,max}
	 *   setting MD_RECOVERY_RUNNING (which interacts with resync_{min,max})
369 370
	 */
	spinlock_t			lock;
371
	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
372
	atomic_t			pending_writes;	/* number of active superblock writes */
373

L
Linus Torvalds 已提交
374 375
	unsigned int			safemode;	/* if set, update "clean" superblock
							 * when no writes pending.
376
							 */
L
Linus Torvalds 已提交
377 378
	unsigned int			safemode_delay;
	struct timer_list		safemode_timer;
379
	atomic_t			writes_pending;
380
	struct request_queue		*queue;	/* for plugging ... */
L
Linus Torvalds 已提交
381

382
	struct bitmap			*bitmap; /* the bitmap for the device */
383 384
	struct {
		struct file		*file; /* the bitmap file */
385
		loff_t			offset; /* offset from superblock of
386 387
						 * start of bitmap. May be
						 * negative, but not '0'
388
						 * For external metadata, offset
389
						 * from start of device.
390
						 */
391
		unsigned long		space; /* space available at this offset */
392
		loff_t			default_offset; /* this is the offset to use when
393 394 395
							 * hot-adding a bitmap.  It should
							 * eventually be settable by sysfs.
							 */
396 397
		unsigned long		default_space; /* space available at
							* default offset */
398
		struct mutex		mutex;
399
		unsigned long		chunksize;
400
		unsigned long		daemon_sleep; /* how many jiffies between updates? */
401
		unsigned long		max_write_behind; /* write-behind mode */
402
		int			external;
G
Goldwyn Rodrigues 已提交
403
		int			nodes; /* Maximum number of nodes in the cluster */
G
Goldwyn Rodrigues 已提交
404
		char                    cluster_name[64]; /* Name of the cluster */
405
	} bitmap_info;
406

407
	atomic_t			max_corr_read_errors; /* max read retries */
L
Linus Torvalds 已提交
408
	struct list_head		all_mddevs;
409

410
	struct attribute_group		*to_remove;
411

412 413
	struct bio_set			*bio_set;

T
Tejun Heo 已提交
414 415
	/* Generic flush handling.
	 * The last to finish preflush schedules a worker to submit
416
	 * the rest of the request (without the REQ_PREFLUSH flag).
417
	 */
T
Tejun Heo 已提交
418
	struct bio *flush_bio;
419
	atomic_t flush_pending;
T
Tejun Heo 已提交
420
	struct work_struct flush_work;
421
	struct work_struct event_work;	/* used by dm to report failure event */
422
	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
G
Goldwyn Rodrigues 已提交
423
	struct md_cluster_info		*cluster_info;
424
	unsigned int			good_device_nr;	/* good device num within cluster raid */
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
};

enum recovery_flags {
	/*
	 * If neither SYNC or RESHAPE are set, then it is a recovery.
	 */
	MD_RECOVERY_RUNNING,	/* a thread is running, or about to be started */
	MD_RECOVERY_SYNC,	/* actually doing a resync, not a recovery */
	MD_RECOVERY_RECOVER,	/* doing recovery, or need to try it. */
	MD_RECOVERY_INTR,	/* resync needs to be aborted for some reason */
	MD_RECOVERY_DONE,	/* thread is done and is waiting to be reaped */
	MD_RECOVERY_NEEDED,	/* we might need to start a resync/recover */
	MD_RECOVERY_REQUESTED,	/* user-space has requested a sync (used with SYNC) */
	MD_RECOVERY_CHECK,	/* user-space request for check-only, no repair */
	MD_RECOVERY_RESHAPE,	/* A reshape is happening */
	MD_RECOVERY_FROZEN,	/* User request to abort, and not restart, any action */
	MD_RECOVERY_ERROR,	/* sync-action interrupted because io-error */
L
Linus Torvalds 已提交
442 443
};

444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
static inline int __must_check mddev_lock(struct mddev *mddev)
{
	return mutex_lock_interruptible(&mddev->reconfig_mutex);
}

/* Sometimes we need to take the lock in a situation where
 * failure due to interrupts is not acceptable.
 */
static inline void mddev_lock_nointr(struct mddev *mddev)
{
	mutex_lock(&mddev->reconfig_mutex);
}

static inline int mddev_is_locked(struct mddev *mddev)
{
	return mutex_is_locked(&mddev->reconfig_mutex);
}

static inline int mddev_trylock(struct mddev *mddev)
{
	return mutex_trylock(&mddev->reconfig_mutex);
}
extern void mddev_unlock(struct mddev *mddev);

L
Linus Torvalds 已提交
468 469
static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
470
	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
L
Linus Torvalds 已提交
471 472
}

473
struct md_personality
L
Linus Torvalds 已提交
474 475
{
	char *name;
476 477
	int level;
	struct list_head list;
L
Linus Torvalds 已提交
478
	struct module *owner;
479
	void (*make_request)(struct mddev *mddev, struct bio *bio);
480
	int (*run)(struct mddev *mddev);
N
NeilBrown 已提交
481
	void (*free)(struct mddev *mddev, void *priv);
482
	void (*status)(struct seq_file *seq, struct mddev *mddev);
L
Linus Torvalds 已提交
483
	/* error_handler must set ->faulty and clear ->in_sync
484
	 * if appropriate, and should abort recovery if needed
L
Linus Torvalds 已提交
485
	 */
486 487
	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
488
	int (*hot_remove_disk) (struct mddev *mddev, struct md_rdev *rdev);
489
	int (*spare_active) (struct mddev *mddev);
490
	sector_t (*sync_request)(struct mddev *mddev, sector_t sector_nr, int *skipped);
491 492 493 494 495
	int (*resize) (struct mddev *mddev, sector_t sectors);
	sector_t (*size) (struct mddev *mddev, sector_t sectors, int raid_disks);
	int (*check_reshape) (struct mddev *mddev);
	int (*start_reshape) (struct mddev *mddev);
	void (*finish_reshape) (struct mddev *mddev);
496 497 498 499 500
	/* quiesce moves between quiescence states
	 * 0 - fully active
	 * 1 - no new requests allowed
	 * others - reserved
	 */
501
	void (*quiesce) (struct mddev *mddev, int state);
502 503 504 505 506 507 508 509 510
	/* takeover is used to transition an array from one
	 * personality to another.  The new personality must be able
	 * to handle the data in the current layout.
	 * e.g. 2drive raid1 -> 2drive raid5
	 *      ndrive raid5 -> degraded n+1drive raid6 with special layout
	 * If the takeover succeeds, a new 'private' structure is returned.
	 * This needs to be installed and then ->run used to activate the
	 * array.
	 */
511
	void *(*takeover) (struct mddev *mddev);
512 513 514
	/* congested implements bdi.congested_fn().
	 * Will not be called while array is 'suspended' */
	int (*congested)(struct mddev *mddev, int bits);
L
Linus Torvalds 已提交
515 516
};

517 518
struct md_sysfs_entry {
	struct attribute attr;
519 520
	ssize_t (*show)(struct mddev *, char *);
	ssize_t (*store)(struct mddev *, const char *, size_t);
521
};
522
extern struct attribute_group md_bitmap_group;
523

524
static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node *sd, char *name)
N
NeilBrown 已提交
525 526
{
	if (sd)
T
Tejun Heo 已提交
527
		return sysfs_get_dirent(sd, name);
N
NeilBrown 已提交
528 529
	return sd;
}
530
static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd)
N
NeilBrown 已提交
531 532 533 534 535
{
	if (sd)
		sysfs_notify_dirent(sd);
}

536
static inline char * mdname (struct mddev * mddev)
L
Linus Torvalds 已提交
537 538 539 540
{
	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
}

541
static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)
542 543
{
	char nm[20];
544 545 546
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
547 548 549 550
		sprintf(nm, "rd%d", rdev->raid_disk);
		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
	} else
		return 0;
551 552
}

553
static inline void sysfs_unlink_rdev(struct mddev *mddev, struct md_rdev *rdev)
554 555
{
	char nm[20];
556 557 558
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
559 560 561
		sprintf(nm, "rd%d", rdev->raid_disk);
		sysfs_remove_link(&mddev->kobj, nm);
	}
562 563
}

L
Linus Torvalds 已提交
564 565 566 567
/*
 * iterates through some rdev ringlist. It's safe to remove the
 * current 'rdev'. Dont touch 'tmp' though.
 */
568 569 570
#define rdev_for_each_list(rdev, tmp, head)				\
	list_for_each_entry_safe(rdev, tmp, head, same_set)

L
Linus Torvalds 已提交
571 572 573
/*
 * iterates through the 'same array disks' ringlist
 */
N
NeilBrown 已提交
574 575 576 577
#define rdev_for_each(rdev, mddev)				\
	list_for_each_entry(rdev, &((mddev)->disks), same_set)

#define rdev_for_each_safe(rdev, tmp, mddev)				\
578
	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
L
Linus Torvalds 已提交
579

580 581 582
#define rdev_for_each_rcu(rdev, mddev)				\
	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)

583
struct md_thread {
S
Shaohua Li 已提交
584
	void			(*run) (struct md_thread *thread);
585
	struct mddev		*mddev;
L
Linus Torvalds 已提交
586
	wait_queue_head_t	wqueue;
587
	unsigned long		flags;
L
Linus Torvalds 已提交
588
	struct task_struct	*tsk;
589
	unsigned long		timeout;
S
Shaohua Li 已提交
590
	void			*private;
591
};
L
Linus Torvalds 已提交
592 593 594

#define THREAD_WAKEUP  0

595 596 597 598 599
static inline void safe_put_page(struct page *p)
{
	if (p) put_page(p);
}

600 601
extern int register_md_personality(struct md_personality *p);
extern int unregister_md_personality(struct md_personality *p);
602 603 604 605 606
extern int register_md_cluster_operations(struct md_cluster_operations *ops,
		struct module *module);
extern int unregister_md_cluster_operations(void);
extern int md_setup_cluster(struct mddev *mddev, int nodes);
extern void md_cluster_stop(struct mddev *mddev);
607
extern struct md_thread *md_register_thread(
S
Shaohua Li 已提交
608
	void (*run)(struct md_thread *thread),
609 610 611 612
	struct mddev *mddev,
	const char *name);
extern void md_unregister_thread(struct md_thread **threadp);
extern void md_wakeup_thread(struct md_thread *thread);
613
extern void md_check_recovery(struct mddev *mddev);
614
extern void md_reap_sync_thread(struct mddev *mddev);
615 616 617 618
extern void md_write_start(struct mddev *mddev, struct bio *bi);
extern void md_write_end(struct mddev *mddev);
extern void md_done_sync(struct mddev *mddev, int blocks, int ok);
extern void md_error(struct mddev *mddev, struct md_rdev *rdev);
619
extern void md_finish_reshape(struct mddev *mddev);
620 621 622 623

extern int mddev_congested(struct mddev *mddev, int bits);
extern void md_flush_request(struct mddev *mddev, struct bio *bio);
extern void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
624
			   sector_t sector, int size, struct page *page);
625
extern void md_super_wait(struct mddev *mddev);
626
extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
M
Mike Christie 已提交
627 628
			struct page *page, int op, int op_flags,
			bool metadata_op);
S
Shaohua Li 已提交
629
extern void md_do_sync(struct md_thread *thread);
630 631 632 633 634 635
extern void md_new_event(struct mddev *mddev);
extern int md_allow_write(struct mddev *mddev);
extern void md_wait_for_blocked_rdev(struct md_rdev *rdev, struct mddev *mddev);
extern void md_set_array_sectors(struct mddev *mddev, sector_t array_sectors);
extern int md_check_no_bitmap(struct mddev *mddev);
extern int md_integrity_register(struct mddev *mddev);
636
extern int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev);
637
extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
C
Christoph Hellwig 已提交
638

639 640 641 642
extern void mddev_init(struct mddev *mddev);
extern int md_run(struct mddev *mddev);
extern void md_stop(struct mddev *mddev);
extern void md_stop_writes(struct mddev *mddev);
643
extern int md_rdev_init(struct md_rdev *rdev);
644
extern void md_rdev_clear(struct md_rdev *rdev);
C
Christoph Hellwig 已提交
645

646 647
extern void mddev_suspend(struct mddev *mddev);
extern void mddev_resume(struct mddev *mddev);
648
extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
649
				   struct mddev *mddev);
650
extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
651
				   struct mddev *mddev);
652

653
extern void md_unplug(struct blk_plug_cb *cb, bool from_schedule);
654
extern void md_reload_sb(struct mddev *mddev, int raid_disk);
655
extern void md_update_sb(struct mddev *mddev, int force);
656
extern void md_kick_rdev_from_array(struct md_rdev * rdev);
657
struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr);
658 659 660 661 662
static inline int mddev_check_plugged(struct mddev *mddev)
{
	return !!blk_check_plugged(md_unplug, mddev,
				   sizeof(struct blk_plug_cb));
}
663 664 665 666 667 668 669 670 671 672

static inline void rdev_dec_pending(struct md_rdev *rdev, struct mddev *mddev)
{
	int faulty = test_bit(Faulty, &rdev->flags);
	if (atomic_dec_and_test(&rdev->nr_pending) && faulty) {
		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
		md_wakeup_thread(mddev->thread);
	}
}

673
extern struct md_cluster_operations *md_cluster_ops;
G
Goldwyn Rodrigues 已提交
674 675 676 677
static inline int mddev_is_clustered(struct mddev *mddev)
{
	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;
}
C
Christoph Hellwig 已提交
678
#endif /* _MD_MD_H */