md.h 23.6 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>
C
Christoph Hellwig 已提交
20 21 22 23 24 25 26
#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>
27
#include "md-cluster.h"
28

L
Linus Torvalds 已提交
29 30
#define MaxSector (~(sector_t)0)

31 32 33 34 35 36 37
/* Bad block numbers are stored sorted in a single page.
 * 64bits is used for each block or extent.
 * 54 bits are sector number, 9 bits are extent size,
 * 1 bit is an 'acknowledged' flag.
 */
#define MD_MAX_BADBLOCKS	(PAGE_SIZE/8)

L
Linus Torvalds 已提交
38 39 40
/*
 * MD's 'extended' device
 */
41
struct md_rdev {
L
Linus Torvalds 已提交
42 43
	struct list_head same_set;	/* RAID devices within the same set */

44
	sector_t sectors;		/* Device size (in 512bytes sectors) */
45
	struct mddev *mddev;		/* RAID array if running */
N
NeilBrown 已提交
46
	int last_events;		/* IO event timestamp */
L
Linus Torvalds 已提交
47

48 49 50 51 52 53
	/*
	 * 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
54 55
	struct block_device *bdev;	/* block device handle */

56
	struct page	*sb_page, *bb_page;
L
Linus Torvalds 已提交
57
	int		sb_loaded;
58
	__u64		sb_events;
L
Linus Torvalds 已提交
59
	sector_t	data_offset;	/* start of data in array */
60
	sector_t	new_data_offset;/* only relevant while reshaping */
61
	sector_t	sb_start;	/* offset of the super block (in 512byte sectors) */
62
	int		sb_size;	/* bytes in the superblock */
L
Linus Torvalds 已提交
63 64
	int		preferred_minor;	/* autorun support */

65 66
	struct kobject	kobj;

L
Linus Torvalds 已提交
67 68 69 70 71 72 73 74 75 76 77
	/* 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
	 */

78
	unsigned long	flags;	/* bit set of 'enum flag_bits' bits. */
79
	wait_queue_head_t blocked_wait;
80

L
Linus Torvalds 已提交
81 82
	int desc_nr;			/* descriptor index in the superblock */
	int raid_disk;			/* role of device in array */
83 84 85
	int new_raid_disk;		/* role that the device will have in
					 * the array after a level-change completes.
					 */
86 87 88 89
	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
					 */
90 91
	union {
		sector_t recovery_offset;/* If this device has been partially
92 93 94
					 * recovered, this is where we were
					 * up to.
					 */
95 96 97 98 99
		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
100 101 102 103 104

	atomic_t	nr_pending;	/* number of pending requests.
					 * only maintained for arrays that
					 * support hot removal
					 */
105 106 107
	atomic_t	read_errors;	/* number of consecutive read errors that
					 * we have tried to ignore.
					 */
108 109 110
	struct timespec last_read_error;	/* monotonic time since our
						 * last read error
						 */
111 112 113 114
	atomic_t	corrected_errors; /* number of corrected read errors,
					   * for reporting to userspace and storing
					   * in superblock.
					   */
115
	struct work_struct del_work;	/* used for delayed sysfs removal */
116

117
	struct kernfs_node *sysfs_state; /* handle for 'state'
118
					   * sysfs entry */
119 120 121

	struct badblocks {
		int	count;		/* count of bad blocks */
122 123 124 125
		int	unacked_exist;	/* there probably are unacknowledged
					 * bad blocks.  This is only cleared
					 * when a read discovers none
					 */
126 127 128 129 130 131
		int	shift;		/* shift from sectors to block size
					 * a -ve shift means badblocks are
					 * disabled.*/
		u64	*page;		/* badblock list */
		int	changed;
		seqlock_t lock;
132 133 134

		sector_t sector;
		sector_t size;		/* in sectors */
135
	} badblocks;
L
Linus Torvalds 已提交
136
};
137 138 139
enum flag_bits {
	Faulty,			/* device is known to have a fault */
	In_sync,		/* device is in_sync with rest of array */
140 141 142
	Bitmap_sync,		/* ..actually, not quite In_sync.  Need a
				 * bitmap-based recovery to get fully in sync
				 */
143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
	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.
				 */
177 178 179 180
	Candidate,		/* For clustered environments only:
				 * This device is seen locally but not
				 * by the whole cluster
				 */
181 182 183 184 185
	Journal,		/* This device is used as journal for
				 * raid-5/6.
				 * Usually, this device should be faster
				 * than other devices in the array
				 */
186
};
L
Linus Torvalds 已提交
187

188 189 190 191 192 193 194 195 196 197 198
#define BB_LEN_MASK	(0x00000000000001FFULL)
#define BB_OFFSET_MASK	(0x7FFFFFFFFFFFFE00ULL)
#define BB_ACK_MASK	(0x8000000000000000ULL)
#define BB_MAX_LEN	512
#define BB_OFFSET(x)	(((x) & BB_OFFSET_MASK) >> 9)
#define BB_LEN(x)	(((x) & BB_LEN_MASK) + 1)
#define BB_ACK(x)	(!!((x) & BB_ACK_MASK))
#define BB_MAKE(a, l, ack) (((a)<<9) | ((l)-1) | ((u64)(!!(ack)) << 63))

extern int md_is_badblock(struct badblocks *bb, sector_t s, int sectors,
			  sector_t *first_bad, int *bad_sectors);
199
static inline int is_badblock(struct md_rdev *rdev, sector_t s, int sectors,
200 201 202 203 204 205 206 207 208 209 210 211
			      sector_t *first_bad, int *bad_sectors)
{
	if (unlikely(rdev->badblocks.count)) {
		int rv = md_is_badblock(&rdev->badblocks, rdev->data_offset + s,
					sectors,
					first_bad, bad_sectors);
		if (rv)
			*first_bad -= rdev->data_offset;
		return rv;
	}
	return 0;
}
212
extern int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
213 214 215
			      int is_new);
extern int rdev_clear_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
				int is_new);
216 217
extern void md_ack_all_badblocks(struct badblocks *bb);

G
Goldwyn Rodrigues 已提交
218 219
struct md_cluster_info;

220
struct mddev {
L
Linus Torvalds 已提交
221
	void				*private;
222
	struct md_personality		*pers;
L
Linus Torvalds 已提交
223 224
	dev_t				unit;
	int				md_minor;
225
	struct list_head		disks;
226 227 228
	unsigned long			flags;
#define MD_CHANGE_DEVS	0	/* Some device status has changed */
#define MD_CHANGE_CLEAN 1	/* transition to or from 'clean' */
229
#define MD_CHANGE_PENDING 2	/* switch from 'clean' to 'active' in progress */
230
#define MD_UPDATE_SB_FLAGS (1 | 2 | 4)	/* If these are set, md_update_sb needed */
231
#define MD_ARRAY_FIRST_USE 3    /* First use of array, needs initialization */
232 233 234
#define MD_STILL_CLOSED	4	/* If set, then array has not been opened since
				 * md_ioctl checked on it.
				 */
235
#define MD_JOURNAL_CLEAN 5	/* A raid with journal is already clean */
236
#define MD_HAS_JOURNAL	6	/* The raid array has journal feature set */
237

238 239
	int				suspended;
	atomic_t			active_io;
L
Linus Torvalds 已提交
240
	int				ro;
241 242 243 244
	int				sysfs_active; /* set when sysfs deletes
						       * are happening, so run/
						       * takeover/stop are not safe
						       */
245
	int				ready; /* See when safe to pass
246
						* IO requests down */
L
Linus Torvalds 已提交
247 248
	struct gendisk			*gendisk;

249
	struct kobject			kobj;
250 251
	int				hold_active;
#define	UNTIL_IOCTL	1
252
#define	UNTIL_STOP	2
253

L
Linus Torvalds 已提交
254 255 256 257 258
	/* Superblock information */
	int				major_version,
					minor_version,
					patch_version;
	int				persistent;
259
	int				external;	/* metadata is
260 261
							 * managed externally */
	char				metadata_type[17]; /* externally set*/
262
	int				chunk_sectors;
L
Linus Torvalds 已提交
263 264
	time_t				ctime, utime;
	int				level, layout;
265
	char				clevel[16];
L
Linus Torvalds 已提交
266 267
	int				raid_disks;
	int				max_disks;
268
	sector_t			dev_sectors;	/* used size of
A
Andre Noll 已提交
269
							 * component devices */
270
	sector_t			array_sectors; /* exported array size */
D
Dan Williams 已提交
271 272
	int				external_size; /* size managed
							* externally */
L
Linus Torvalds 已提交
273
	__u64				events;
274 275 276 277 278 279
	/* 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
280 281 282

	char				uuid[16];

283 284 285 286 287 288
	/* 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;
289 290
	int				delta_disks, new_level, new_layout;
	int				new_chunk_sectors;
291
	int				reshape_backwards;
292

293 294
	struct md_thread		*thread;	/* management thread */
	struct md_thread		*sync_thread;	/* doing resync or reconstruct */
295 296 297 298 299 300 301 302

	/* '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;
303
	sector_t			curr_resync;	/* last block scheduled */
304 305 306 307 308 309 310
	/* 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
311 312
	unsigned long			resync_mark;	/* a recent timestamp */
	sector_t			resync_mark_cnt;/* blocks written at resync_mark */
313
	sector_t			curr_mark_cnt; /* blocks scheduled now */
L
Linus Torvalds 已提交
314 315

	sector_t			resync_max_sectors; /* may be set by personality */
316

317
	atomic64_t			resync_mismatches; /* count of sectors where
318 319
							    * parity/replica mismatch found
							    */
320 321 322 323

	/* allow user-space to request suspension of IO to regions of the array */
	sector_t			suspend_lo;
	sector_t			suspend_hi;
324 325 326 327
	/* if zero, use the system-wide default */
	int				sync_speed_min;
	int				sync_speed_max;

328 329 330
	/* resync even though the same disks are shared among md-devices */
	int				parallel_resync;

331
	int				ok_start_degraded;
332
	/* recovery/resync flags
L
Linus Torvalds 已提交
333 334 335
	 * 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
336
	 * RECOVER:  doing recovery, or need to try it.
337
	 * INTR:     resync needs to be aborted for some reason
L
Linus Torvalds 已提交
338
	 * DONE:     thread is done and is waiting to be reaped
339
	 * REQUEST:  user-space has requested a sync (used with SYNC)
340
	 * CHECK:    user-space request for check-only, no repair
341
	 * RESHAPE:  A reshape is happening
342
	 * ERROR:    sync-action interrupted because io-error
343 344
	 *
	 * If neither SYNC or RESHAPE are set, then it is a recovery.
L
Linus Torvalds 已提交
345 346 347
	 */
#define	MD_RECOVERY_RUNNING	0
#define	MD_RECOVERY_SYNC	1
348
#define	MD_RECOVERY_RECOVER	2
L
Linus Torvalds 已提交
349 350 351
#define	MD_RECOVERY_INTR	3
#define	MD_RECOVERY_DONE	4
#define	MD_RECOVERY_NEEDED	5
352 353
#define	MD_RECOVERY_REQUESTED	6
#define	MD_RECOVERY_CHECK	7
354
#define MD_RECOVERY_RESHAPE	8
355
#define	MD_RECOVERY_FROZEN	9
356
#define	MD_RECOVERY_ERROR	10
357

L
Linus Torvalds 已提交
358
	unsigned long			recovery;
359 360 361 362 363 364
	/* 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
365 366

	int				in_sync;	/* know to not need resync */
N
NeilBrown 已提交
367 368 369 370 371 372 373 374 375 376
	/* '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;
377
	struct mutex			reconfig_mutex;
378 379
	atomic_t			active;		/* general refcount */
	atomic_t			openers;	/* number of active opens */
L
Linus Torvalds 已提交
380

381 382
	int				changed;	/* True if we might need to
							 * reread partition info */
L
Linus Torvalds 已提交
383 384 385 386 387 388 389
	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;
390 391
	sector_t			resync_min;	/* user requested sync
							 * starts here */
392 393
	sector_t			resync_max;	/* resync should pause
							 * when it gets here */
394

395
	struct kernfs_node		*sysfs_state;	/* handle for 'array_state'
396 397
							 * file in sysfs.
							 */
398
	struct kernfs_node		*sysfs_action;  /* handle for 'sync_action' */
399

400 401
	struct work_struct del_work;	/* used for delayed sysfs removal */

402 403 404 405 406
	/* "lock" protects:
	 *   flush_bio transition from NULL to !NULL
	 *   rdev superblocks, events
	 *   clearing MD_CHANGE_*
	 *   in_sync - and related safemode and MD_CHANGE changes
407
	 *   pers (also protected by reconfig_mutex and pending IO).
408
	 *   clearing ->bitmap
409
	 *   clearing ->bitmap_info.file
410 411
	 *   changing ->resync_{min,max}
	 *   setting MD_RECOVERY_RUNNING (which interacts with resync_{min,max})
412 413
	 */
	spinlock_t			lock;
414
	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
415
	atomic_t			pending_writes;	/* number of active superblock writes */
416

L
Linus Torvalds 已提交
417 418
	unsigned int			safemode;	/* if set, update "clean" superblock
							 * when no writes pending.
419
							 */
L
Linus Torvalds 已提交
420 421
	unsigned int			safemode_delay;
	struct timer_list		safemode_timer;
422
	atomic_t			writes_pending;
423
	struct request_queue		*queue;	/* for plugging ... */
L
Linus Torvalds 已提交
424

425
	struct bitmap			*bitmap; /* the bitmap for the device */
426 427
	struct {
		struct file		*file; /* the bitmap file */
428
		loff_t			offset; /* offset from superblock of
429 430
						 * start of bitmap. May be
						 * negative, but not '0'
431
						 * For external metadata, offset
432
						 * from start of device.
433
						 */
434
		unsigned long		space; /* space available at this offset */
435
		loff_t			default_offset; /* this is the offset to use when
436 437 438
							 * hot-adding a bitmap.  It should
							 * eventually be settable by sysfs.
							 */
439 440
		unsigned long		default_space; /* space available at
							* default offset */
441
		struct mutex		mutex;
442
		unsigned long		chunksize;
443
		unsigned long		daemon_sleep; /* how many jiffies between updates? */
444
		unsigned long		max_write_behind; /* write-behind mode */
445
		int			external;
G
Goldwyn Rodrigues 已提交
446
		int			nodes; /* Maximum number of nodes in the cluster */
G
Goldwyn Rodrigues 已提交
447
		char                    cluster_name[64]; /* Name of the cluster */
448
	} bitmap_info;
449

450
	atomic_t			max_corr_read_errors; /* max read retries */
L
Linus Torvalds 已提交
451
	struct list_head		all_mddevs;
452

453
	struct attribute_group		*to_remove;
454

455 456
	struct bio_set			*bio_set;

T
Tejun Heo 已提交
457 458 459
	/* Generic flush handling.
	 * The last to finish preflush schedules a worker to submit
	 * the rest of the request (without the REQ_FLUSH flag).
460
	 */
T
Tejun Heo 已提交
461
	struct bio *flush_bio;
462
	atomic_t flush_pending;
T
Tejun Heo 已提交
463
	struct work_struct flush_work;
464
	struct work_struct event_work;	/* used by dm to report failure event */
465
	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
G
Goldwyn Rodrigues 已提交
466
	struct md_cluster_info		*cluster_info;
L
Linus Torvalds 已提交
467 468
};

469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
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
493 494
static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
495
	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
L
Linus Torvalds 已提交
496 497
}

498
struct md_personality
L
Linus Torvalds 已提交
499 500
{
	char *name;
501 502
	int level;
	struct list_head list;
L
Linus Torvalds 已提交
503
	struct module *owner;
504
	void (*make_request)(struct mddev *mddev, struct bio *bio);
505
	int (*run)(struct mddev *mddev);
N
NeilBrown 已提交
506
	void (*free)(struct mddev *mddev, void *priv);
507
	void (*status)(struct seq_file *seq, struct mddev *mddev);
L
Linus Torvalds 已提交
508
	/* error_handler must set ->faulty and clear ->in_sync
509
	 * if appropriate, and should abort recovery if needed
L
Linus Torvalds 已提交
510
	 */
511 512
	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
513
	int (*hot_remove_disk) (struct mddev *mddev, struct md_rdev *rdev);
514
	int (*spare_active) (struct mddev *mddev);
515
	sector_t (*sync_request)(struct mddev *mddev, sector_t sector_nr, int *skipped);
516 517 518 519 520
	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);
521 522 523 524 525
	/* quiesce moves between quiescence states
	 * 0 - fully active
	 * 1 - no new requests allowed
	 * others - reserved
	 */
526
	void (*quiesce) (struct mddev *mddev, int state);
527 528 529 530 531 532 533 534 535
	/* 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.
	 */
536
	void *(*takeover) (struct mddev *mddev);
537 538 539
	/* 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
540 541
};

542 543
struct md_sysfs_entry {
	struct attribute attr;
544 545
	ssize_t (*show)(struct mddev *, char *);
	ssize_t (*store)(struct mddev *, const char *, size_t);
546
};
547
extern struct attribute_group md_bitmap_group;
548

549
static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node *sd, char *name)
N
NeilBrown 已提交
550 551
{
	if (sd)
T
Tejun Heo 已提交
552
		return sysfs_get_dirent(sd, name);
N
NeilBrown 已提交
553 554
	return sd;
}
555
static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd)
N
NeilBrown 已提交
556 557 558 559 560
{
	if (sd)
		sysfs_notify_dirent(sd);
}

561
static inline char * mdname (struct mddev * mddev)
L
Linus Torvalds 已提交
562 563 564 565
{
	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
}

566
static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)
567 568
{
	char nm[20];
569 570 571
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
572 573 574 575
		sprintf(nm, "rd%d", rdev->raid_disk);
		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
	} else
		return 0;
576 577
}

578
static inline void sysfs_unlink_rdev(struct mddev *mddev, struct md_rdev *rdev)
579 580
{
	char nm[20];
581 582 583
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
584 585 586
		sprintf(nm, "rd%d", rdev->raid_disk);
		sysfs_remove_link(&mddev->kobj, nm);
	}
587 588
}

L
Linus Torvalds 已提交
589 590 591 592
/*
 * iterates through some rdev ringlist. It's safe to remove the
 * current 'rdev'. Dont touch 'tmp' though.
 */
593 594 595
#define rdev_for_each_list(rdev, tmp, head)				\
	list_for_each_entry_safe(rdev, tmp, head, same_set)

L
Linus Torvalds 已提交
596 597 598
/*
 * iterates through the 'same array disks' ringlist
 */
N
NeilBrown 已提交
599 600 601 602
#define rdev_for_each(rdev, mddev)				\
	list_for_each_entry(rdev, &((mddev)->disks), same_set)

#define rdev_for_each_safe(rdev, tmp, mddev)				\
603
	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
L
Linus Torvalds 已提交
604

605 606 607
#define rdev_for_each_rcu(rdev, mddev)				\
	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)

608
struct md_thread {
S
Shaohua Li 已提交
609
	void			(*run) (struct md_thread *thread);
610
	struct mddev		*mddev;
L
Linus Torvalds 已提交
611
	wait_queue_head_t	wqueue;
612
	unsigned long		flags;
L
Linus Torvalds 已提交
613
	struct task_struct	*tsk;
614
	unsigned long		timeout;
S
Shaohua Li 已提交
615
	void			*private;
616
};
L
Linus Torvalds 已提交
617 618 619

#define THREAD_WAKEUP  0

620 621 622 623 624
static inline void safe_put_page(struct page *p)
{
	if (p) put_page(p);
}

625 626
extern int register_md_personality(struct md_personality *p);
extern int unregister_md_personality(struct md_personality *p);
627 628 629 630 631
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);
632
extern struct md_thread *md_register_thread(
S
Shaohua Li 已提交
633
	void (*run)(struct md_thread *thread),
634 635 636 637
	struct mddev *mddev,
	const char *name);
extern void md_unregister_thread(struct md_thread **threadp);
extern void md_wakeup_thread(struct md_thread *thread);
638
extern void md_check_recovery(struct mddev *mddev);
639
extern void md_reap_sync_thread(struct mddev *mddev);
640 641 642 643
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);
644
extern void md_finish_reshape(struct mddev *mddev);
645 646 647 648

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,
649
			   sector_t sector, int size, struct page *page);
650
extern void md_super_wait(struct mddev *mddev);
651
extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
J
Jonathan Brassow 已提交
652
			struct page *page, int rw, bool metadata_op);
S
Shaohua Li 已提交
653
extern void md_do_sync(struct md_thread *thread);
654 655 656 657 658 659 660
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);
extern void md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev);
661
extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
C
Christoph Hellwig 已提交
662

663 664 665 666
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);
667
extern int md_rdev_init(struct md_rdev *rdev);
668
extern void md_rdev_clear(struct md_rdev *rdev);
C
Christoph Hellwig 已提交
669

670 671
extern void mddev_suspend(struct mddev *mddev);
extern void mddev_resume(struct mddev *mddev);
672
extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
673
				   struct mddev *mddev);
674
extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
675
				   struct mddev *mddev);
676

677
extern void md_unplug(struct blk_plug_cb *cb, bool from_schedule);
678
extern void md_reload_sb(struct mddev *mddev, int raid_disk);
679
extern void md_update_sb(struct mddev *mddev, int force);
680
extern void md_kick_rdev_from_array(struct md_rdev * rdev);
681
struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr);
682 683 684 685 686
static inline int mddev_check_plugged(struct mddev *mddev)
{
	return !!blk_check_plugged(md_unplug, mddev,
				   sizeof(struct blk_plug_cb));
}
687 688 689 690 691 692 693 694 695 696

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);
	}
}

697
extern struct md_cluster_operations *md_cluster_ops;
G
Goldwyn Rodrigues 已提交
698 699 700 701
static inline int mddev_is_clustered(struct mddev *mddev)
{
	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;
}
C
Christoph Hellwig 已提交
702
#endif /* _MD_MD_H */