md.h 22.7 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.
					 */
102 103 104
	struct timespec last_read_error;	/* monotonic time since our
						 * 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
};
L
Linus Torvalds 已提交
167

168
static inline int is_badblock(struct md_rdev *rdev, sector_t s, int sectors,
169 170 171
			      sector_t *first_bad, int *bad_sectors)
{
	if (unlikely(rdev->badblocks.count)) {
172
		int rv = badblocks_check(&rdev->badblocks, rdev->data_offset + s,
173 174 175 176 177 178 179 180
					sectors,
					first_bad, bad_sectors);
		if (rv)
			*first_bad -= rdev->data_offset;
		return rv;
	}
	return 0;
}
181
extern int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
182 183 184
			      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
185 186
struct md_cluster_info;

187
struct mddev {
L
Linus Torvalds 已提交
188
	void				*private;
189
	struct md_personality		*pers;
L
Linus Torvalds 已提交
190 191
	dev_t				unit;
	int				md_minor;
192
	struct list_head		disks;
193 194 195
	unsigned long			flags;
#define MD_CHANGE_DEVS	0	/* Some device status has changed */
#define MD_CHANGE_CLEAN 1	/* transition to or from 'clean' */
196
#define MD_CHANGE_PENDING 2	/* switch from 'clean' to 'active' in progress */
197
#define MD_UPDATE_SB_FLAGS (1 | 2 | 4)	/* If these are set, md_update_sb needed */
198
#define MD_ARRAY_FIRST_USE 3    /* First use of array, needs initialization */
199 200 201
#define MD_STILL_CLOSED	4	/* If set, then array has not been opened since
				 * md_ioctl checked on it.
				 */
202
#define MD_JOURNAL_CLEAN 5	/* A raid with journal is already clean */
203
#define MD_HAS_JOURNAL	6	/* The raid array has journal feature set */
204 205 206
#define MD_RELOAD_SB	7	/* Reload the superblock because another node
				 * updated it.
				 */
207 208 209
#define MD_CLUSTER_RESYNC_LOCKED 8 /* cluster raid only, which means node
				    * already took resync lock, need to
				    * release the lock */
210

211 212
	int				suspended;
	atomic_t			active_io;
L
Linus Torvalds 已提交
213
	int				ro;
214 215 216 217
	int				sysfs_active; /* set when sysfs deletes
						       * are happening, so run/
						       * takeover/stop are not safe
						       */
L
Linus Torvalds 已提交
218 219
	struct gendisk			*gendisk;

220
	struct kobject			kobj;
221 222
	int				hold_active;
#define	UNTIL_IOCTL	1
223
#define	UNTIL_STOP	2
224

L
Linus Torvalds 已提交
225 226 227 228 229
	/* Superblock information */
	int				major_version,
					minor_version,
					patch_version;
	int				persistent;
230
	int				external;	/* metadata is
231 232
							 * managed externally */
	char				metadata_type[17]; /* externally set*/
233
	int				chunk_sectors;
234
	time64_t			ctime, utime;
L
Linus Torvalds 已提交
235
	int				level, layout;
236
	char				clevel[16];
L
Linus Torvalds 已提交
237 238
	int				raid_disks;
	int				max_disks;
239
	sector_t			dev_sectors;	/* used size of
A
Andre Noll 已提交
240
							 * component devices */
241
	sector_t			array_sectors; /* exported array size */
D
Dan Williams 已提交
242 243
	int				external_size; /* size managed
							* externally */
L
Linus Torvalds 已提交
244
	__u64				events;
245 246 247 248 249 250
	/* 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
251 252 253

	char				uuid[16];

254 255 256 257 258 259
	/* 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;
260 261
	int				delta_disks, new_level, new_layout;
	int				new_chunk_sectors;
262
	int				reshape_backwards;
263

264 265
	struct md_thread		*thread;	/* management thread */
	struct md_thread		*sync_thread;	/* doing resync or reconstruct */
266 267 268 269 270 271 272 273

	/* '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;
274
	sector_t			curr_resync;	/* last block scheduled */
275 276 277 278 279 280 281
	/* 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
282 283
	unsigned long			resync_mark;	/* a recent timestamp */
	sector_t			resync_mark_cnt;/* blocks written at resync_mark */
284
	sector_t			curr_mark_cnt; /* blocks scheduled now */
L
Linus Torvalds 已提交
285 286

	sector_t			resync_max_sectors; /* may be set by personality */
287

288
	atomic64_t			resync_mismatches; /* count of sectors where
289 290
							    * parity/replica mismatch found
							    */
291 292 293 294

	/* allow user-space to request suspension of IO to regions of the array */
	sector_t			suspend_lo;
	sector_t			suspend_hi;
295 296 297 298
	/* if zero, use the system-wide default */
	int				sync_speed_min;
	int				sync_speed_max;

299 300 301
	/* resync even though the same disks are shared among md-devices */
	int				parallel_resync;

302
	int				ok_start_degraded;
303
	/* recovery/resync flags
L
Linus Torvalds 已提交
304 305 306
	 * NEEDED:   we might need to start a resync/recover
	 * RUNNING:  a thread is running, or about to be started
	 * SYNC:     actually doing a resync, not a recovery
307
	 * RECOVER:  doing recovery, or need to try it.
308
	 * INTR:     resync needs to be aborted for some reason
L
Linus Torvalds 已提交
309
	 * DONE:     thread is done and is waiting to be reaped
310
	 * REQUEST:  user-space has requested a sync (used with SYNC)
311
	 * CHECK:    user-space request for check-only, no repair
312
	 * RESHAPE:  A reshape is happening
313
	 * ERROR:    sync-action interrupted because io-error
314 315
	 *
	 * If neither SYNC or RESHAPE are set, then it is a recovery.
L
Linus Torvalds 已提交
316 317 318
	 */
#define	MD_RECOVERY_RUNNING	0
#define	MD_RECOVERY_SYNC	1
319
#define	MD_RECOVERY_RECOVER	2
L
Linus Torvalds 已提交
320 321 322
#define	MD_RECOVERY_INTR	3
#define	MD_RECOVERY_DONE	4
#define	MD_RECOVERY_NEEDED	5
323 324
#define	MD_RECOVERY_REQUESTED	6
#define	MD_RECOVERY_CHECK	7
325
#define MD_RECOVERY_RESHAPE	8
326
#define	MD_RECOVERY_FROZEN	9
327
#define	MD_RECOVERY_ERROR	10
328

L
Linus Torvalds 已提交
329
	unsigned long			recovery;
330 331 332 333 334 335
	/* 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
336 337

	int				in_sync;	/* know to not need resync */
N
NeilBrown 已提交
338 339 340 341 342 343 344 345 346 347
	/* '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;
348
	struct mutex			reconfig_mutex;
349 350
	atomic_t			active;		/* general refcount */
	atomic_t			openers;	/* number of active opens */
L
Linus Torvalds 已提交
351

352 353
	int				changed;	/* True if we might need to
							 * reread partition info */
L
Linus Torvalds 已提交
354 355 356 357 358 359 360
	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;
361 362
	sector_t			resync_min;	/* user requested sync
							 * starts here */
363 364
	sector_t			resync_max;	/* resync should pause
							 * when it gets here */
365

366
	struct kernfs_node		*sysfs_state;	/* handle for 'array_state'
367 368
							 * file in sysfs.
							 */
369
	struct kernfs_node		*sysfs_action;  /* handle for 'sync_action' */
370

371 372
	struct work_struct del_work;	/* used for delayed sysfs removal */

373 374 375 376 377
	/* "lock" protects:
	 *   flush_bio transition from NULL to !NULL
	 *   rdev superblocks, events
	 *   clearing MD_CHANGE_*
	 *   in_sync - and related safemode and MD_CHANGE changes
378
	 *   pers (also protected by reconfig_mutex and pending IO).
379
	 *   clearing ->bitmap
380
	 *   clearing ->bitmap_info.file
381 382
	 *   changing ->resync_{min,max}
	 *   setting MD_RECOVERY_RUNNING (which interacts with resync_{min,max})
383 384
	 */
	spinlock_t			lock;
385
	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
386
	atomic_t			pending_writes;	/* number of active superblock writes */
387

L
Linus Torvalds 已提交
388 389
	unsigned int			safemode;	/* if set, update "clean" superblock
							 * when no writes pending.
390
							 */
L
Linus Torvalds 已提交
391 392
	unsigned int			safemode_delay;
	struct timer_list		safemode_timer;
393
	atomic_t			writes_pending;
394
	struct request_queue		*queue;	/* for plugging ... */
L
Linus Torvalds 已提交
395

396
	struct bitmap			*bitmap; /* the bitmap for the device */
397 398
	struct {
		struct file		*file; /* the bitmap file */
399
		loff_t			offset; /* offset from superblock of
400 401
						 * start of bitmap. May be
						 * negative, but not '0'
402
						 * For external metadata, offset
403
						 * from start of device.
404
						 */
405
		unsigned long		space; /* space available at this offset */
406
		loff_t			default_offset; /* this is the offset to use when
407 408 409
							 * hot-adding a bitmap.  It should
							 * eventually be settable by sysfs.
							 */
410 411
		unsigned long		default_space; /* space available at
							* default offset */
412
		struct mutex		mutex;
413
		unsigned long		chunksize;
414
		unsigned long		daemon_sleep; /* how many jiffies between updates? */
415
		unsigned long		max_write_behind; /* write-behind mode */
416
		int			external;
G
Goldwyn Rodrigues 已提交
417
		int			nodes; /* Maximum number of nodes in the cluster */
G
Goldwyn Rodrigues 已提交
418
		char                    cluster_name[64]; /* Name of the cluster */
419
	} bitmap_info;
420

421
	atomic_t			max_corr_read_errors; /* max read retries */
L
Linus Torvalds 已提交
422
	struct list_head		all_mddevs;
423

424
	struct attribute_group		*to_remove;
425

426 427
	struct bio_set			*bio_set;

T
Tejun Heo 已提交
428 429 430
	/* Generic flush handling.
	 * The last to finish preflush schedules a worker to submit
	 * the rest of the request (without the REQ_FLUSH flag).
431
	 */
T
Tejun Heo 已提交
432
	struct bio *flush_bio;
433
	atomic_t flush_pending;
T
Tejun Heo 已提交
434
	struct work_struct flush_work;
435
	struct work_struct event_work;	/* used by dm to report failure event */
436
	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
G
Goldwyn Rodrigues 已提交
437
	struct md_cluster_info		*cluster_info;
438
	unsigned int			good_device_nr;	/* good device num within cluster raid */
L
Linus Torvalds 已提交
439 440
};

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
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
465 466
static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
467
	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
L
Linus Torvalds 已提交
468 469
}

470
struct md_personality
L
Linus Torvalds 已提交
471 472
{
	char *name;
473 474
	int level;
	struct list_head list;
L
Linus Torvalds 已提交
475
	struct module *owner;
476
	void (*make_request)(struct mddev *mddev, struct bio *bio);
477
	int (*run)(struct mddev *mddev);
N
NeilBrown 已提交
478
	void (*free)(struct mddev *mddev, void *priv);
479
	void (*status)(struct seq_file *seq, struct mddev *mddev);
L
Linus Torvalds 已提交
480
	/* error_handler must set ->faulty and clear ->in_sync
481
	 * if appropriate, and should abort recovery if needed
L
Linus Torvalds 已提交
482
	 */
483 484
	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
485
	int (*hot_remove_disk) (struct mddev *mddev, struct md_rdev *rdev);
486
	int (*spare_active) (struct mddev *mddev);
487
	sector_t (*sync_request)(struct mddev *mddev, sector_t sector_nr, int *skipped);
488 489 490 491 492
	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);
493 494 495 496 497
	/* quiesce moves between quiescence states
	 * 0 - fully active
	 * 1 - no new requests allowed
	 * others - reserved
	 */
498
	void (*quiesce) (struct mddev *mddev, int state);
499 500 501 502 503 504 505 506 507
	/* 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.
	 */
508
	void *(*takeover) (struct mddev *mddev);
509 510 511
	/* 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
512 513
};

514 515
struct md_sysfs_entry {
	struct attribute attr;
516 517
	ssize_t (*show)(struct mddev *, char *);
	ssize_t (*store)(struct mddev *, const char *, size_t);
518
};
519
extern struct attribute_group md_bitmap_group;
520

521
static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node *sd, char *name)
N
NeilBrown 已提交
522 523
{
	if (sd)
T
Tejun Heo 已提交
524
		return sysfs_get_dirent(sd, name);
N
NeilBrown 已提交
525 526
	return sd;
}
527
static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd)
N
NeilBrown 已提交
528 529 530 531 532
{
	if (sd)
		sysfs_notify_dirent(sd);
}

533
static inline char * mdname (struct mddev * mddev)
L
Linus Torvalds 已提交
534 535 536 537
{
	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
}

538
static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)
539 540
{
	char nm[20];
541 542 543
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
544 545 546 547
		sprintf(nm, "rd%d", rdev->raid_disk);
		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
	} else
		return 0;
548 549
}

550
static inline void sysfs_unlink_rdev(struct mddev *mddev, struct md_rdev *rdev)
551 552
{
	char nm[20];
553 554 555
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
556 557 558
		sprintf(nm, "rd%d", rdev->raid_disk);
		sysfs_remove_link(&mddev->kobj, nm);
	}
559 560
}

L
Linus Torvalds 已提交
561 562 563 564
/*
 * iterates through some rdev ringlist. It's safe to remove the
 * current 'rdev'. Dont touch 'tmp' though.
 */
565 566 567
#define rdev_for_each_list(rdev, tmp, head)				\
	list_for_each_entry_safe(rdev, tmp, head, same_set)

L
Linus Torvalds 已提交
568 569 570
/*
 * iterates through the 'same array disks' ringlist
 */
N
NeilBrown 已提交
571 572 573 574
#define rdev_for_each(rdev, mddev)				\
	list_for_each_entry(rdev, &((mddev)->disks), same_set)

#define rdev_for_each_safe(rdev, tmp, mddev)				\
575
	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
L
Linus Torvalds 已提交
576

577 578 579
#define rdev_for_each_rcu(rdev, mddev)				\
	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)

580
struct md_thread {
S
Shaohua Li 已提交
581
	void			(*run) (struct md_thread *thread);
582
	struct mddev		*mddev;
L
Linus Torvalds 已提交
583
	wait_queue_head_t	wqueue;
584
	unsigned long		flags;
L
Linus Torvalds 已提交
585
	struct task_struct	*tsk;
586
	unsigned long		timeout;
S
Shaohua Li 已提交
587
	void			*private;
588
};
L
Linus Torvalds 已提交
589 590 591

#define THREAD_WAKEUP  0

592 593 594 595 596
static inline void safe_put_page(struct page *p)
{
	if (p) put_page(p);
}

597 598
extern int register_md_personality(struct md_personality *p);
extern int unregister_md_personality(struct md_personality *p);
599 600 601 602 603
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);
604
extern struct md_thread *md_register_thread(
S
Shaohua Li 已提交
605
	void (*run)(struct md_thread *thread),
606 607 608 609
	struct mddev *mddev,
	const char *name);
extern void md_unregister_thread(struct md_thread **threadp);
extern void md_wakeup_thread(struct md_thread *thread);
610
extern void md_check_recovery(struct mddev *mddev);
611
extern void md_reap_sync_thread(struct mddev *mddev);
612 613 614 615
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);
616
extern void md_finish_reshape(struct mddev *mddev);
617 618 619 620

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,
621
			   sector_t sector, int size, struct page *page);
622
extern void md_super_wait(struct mddev *mddev);
623
extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
J
Jonathan Brassow 已提交
624
			struct page *page, int rw, bool metadata_op);
S
Shaohua Li 已提交
625
extern void md_do_sync(struct md_thread *thread);
626 627 628 629 630 631
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);
632
extern int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev);
633
extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
C
Christoph Hellwig 已提交
634

635 636 637 638
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);
639
extern int md_rdev_init(struct md_rdev *rdev);
640
extern void md_rdev_clear(struct md_rdev *rdev);
C
Christoph Hellwig 已提交
641

642 643
extern void mddev_suspend(struct mddev *mddev);
extern void mddev_resume(struct mddev *mddev);
644
extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
645
				   struct mddev *mddev);
646
extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
647
				   struct mddev *mddev);
648

649
extern void md_unplug(struct blk_plug_cb *cb, bool from_schedule);
650
extern void md_reload_sb(struct mddev *mddev, int raid_disk);
651
extern void md_update_sb(struct mddev *mddev, int force);
652
extern void md_kick_rdev_from_array(struct md_rdev * rdev);
653
struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr);
654 655 656 657 658
static inline int mddev_check_plugged(struct mddev *mddev)
{
	return !!blk_check_plugged(md_unplug, mddev,
				   sizeof(struct blk_plug_cb));
}
659 660 661 662 663 664 665 666 667 668

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);
	}
}

669
extern struct md_cluster_operations *md_cluster_ops;
G
Goldwyn Rodrigues 已提交
670 671 672 673
static inline int mddev_is_clustered(struct mddev *mddev)
{
	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;
}
C
Christoph Hellwig 已提交
674
#endif /* _MD_MD_H */