md.h 27.2 KB
Newer Older
1
/* SPDX-License-Identifier: GPL-2.0-or-later */
L
Linus Torvalds 已提交
2
/*
3
   md.h : kernel internal structure of the Linux MD driver
L
Linus Torvalds 已提交
4
          Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
5

L
Linus Torvalds 已提交
6 7
*/

C
Christoph Hellwig 已提交
8 9 10 11
#ifndef _MD_MD_H
#define _MD_MD_H

#include <linux/blkdev.h>
12
#include <linux/backing-dev.h>
13
#include <linux/badblocks.h>
C
Christoph Hellwig 已提交
14 15 16 17 18 19 20
#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>
21
#include "md-cluster.h"
22

L
Linus Torvalds 已提交
23 24
#define MaxSector (~(sector_t)0)

25 26 27 28 29 30 31 32 33 34
/*
 * These flags should really be called "NO_RETRY" rather than
 * "FAILFAST" because they don't make any promise about time lapse,
 * only about the number of retries, which will be zero.
 * REQ_FAILFAST_DRIVER is not included because
 * Commit: 4a27446f3e39 ("[SCSI] modify scsi to handle new fail fast flags.")
 * seems to suggest that the errors it avoids retrying should usually
 * be retried.
 */
#define	MD_FAILFAST	(REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT)
35 36 37 38 39 40 41 42 43 44

/*
 * The struct embedded in rdev is used to serialize IO.
 */
struct serial_in_rdev {
	struct rb_root_cached serial_rb;
	spinlock_t serial_lock;
	wait_queue_head_t serial_io_wait;
};

L
Linus Torvalds 已提交
45 46 47
/*
 * MD's 'extended' device
 */
48
struct md_rdev {
L
Linus Torvalds 已提交
49 50
	struct list_head same_set;	/* RAID devices within the same set */

51
	sector_t sectors;		/* Device size (in 512bytes sectors) */
52
	struct mddev *mddev;		/* RAID array if running */
N
NeilBrown 已提交
53
	int last_events;		/* IO event timestamp */
L
Linus Torvalds 已提交
54

55 56 57 58 59 60
	/*
	 * 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
61 62
	struct block_device *bdev;	/* block device handle */

63
	struct page	*sb_page, *bb_page;
L
Linus Torvalds 已提交
64
	int		sb_loaded;
65
	__u64		sb_events;
L
Linus Torvalds 已提交
66
	sector_t	data_offset;	/* start of data in array */
67
	sector_t	new_data_offset;/* only relevant while reshaping */
68
	sector_t	sb_start;	/* offset of the super block (in 512byte sectors) */
69
	int		sb_size;	/* bytes in the superblock */
L
Linus Torvalds 已提交
70 71
	int		preferred_minor;	/* autorun support */

72 73
	struct kobject	kobj;

L
Linus Torvalds 已提交
74 75 76 77 78 79 80 81 82 83 84
	/* 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
	 */

85
	unsigned long	flags;	/* bit set of 'enum flag_bits' bits. */
86
	wait_queue_head_t blocked_wait;
87

L
Linus Torvalds 已提交
88 89
	int desc_nr;			/* descriptor index in the superblock */
	int raid_disk;			/* role of device in array */
90 91 92
	int new_raid_disk;		/* role that the device will have in
					 * the array after a level-change completes.
					 */
93 94 95 96
	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
					 */
97 98
	union {
		sector_t recovery_offset;/* If this device has been partially
99 100 101
					 * recovered, this is where we were
					 * up to.
					 */
102 103 104 105 106
		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
107 108 109 110 111

	atomic_t	nr_pending;	/* number of pending requests.
					 * only maintained for arrays that
					 * support hot removal
					 */
112 113 114
	atomic_t	read_errors;	/* number of consecutive read errors that
					 * we have tried to ignore.
					 */
115
	time64_t	last_read_error;	/* monotonic time since our
116 117
						 * last read error
						 */
118 119 120 121
	atomic_t	corrected_errors; /* number of corrected read errors,
					   * for reporting to userspace and storing
					   * in superblock.
					   */
122

123
	struct serial_in_rdev *serial;  /* used for raid1 io serialization */
124

125
	struct work_struct del_work;	/* used for delayed sysfs removal */
126

127
	struct kernfs_node *sysfs_state; /* handle for 'state'
128
					   * sysfs entry */
129

130
	struct badblocks badblocks;
131 132 133 134 135 136 137

	struct {
		short offset;	/* Offset from superblock to start of PPL.
				 * Not used by external metadata. */
		unsigned int size;	/* Size in sectors of the PPL space */
		sector_t sector;	/* First sector of the PPL space */
	} ppl;
L
Linus Torvalds 已提交
138
};
139 140 141
enum flag_bits {
	Faulty,			/* device is known to have a fault */
	In_sync,		/* device is in_sync with rest of array */
142
	Bitmap_sync,		/* ..actually, not quite In_sync.  Need a
143 144 145
				 * bitmap-based recovery to get fully in sync.
				 * The bit is only meaningful before device
				 * has been passed to pers->hot_add_disk.
146
				 */
147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
	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.
				 */
181 182 183 184
	Candidate,		/* For clustered environments only:
				 * This device is seen locally but not
				 * by the whole cluster
				 */
185 186 187 188 189
	Journal,		/* This device is used as journal for
				 * raid-5/6.
				 * Usually, this device should be faster
				 * than other devices in the array
				 */
190
	ClusterRemove,
191 192 193 194 195
	RemoveSynchronized,	/* synchronize_rcu() was called after
				 * this device was known to be faulty,
				 * so it is safe to remove without
				 * another synchronize_rcu() call.
				 */
196 197 198
	ExternalBbl,            /* External metadata provides bad
				 * block management for a disk
				 */
199 200 201 202 203 204
	FailFast,		/* Minimal retries should be attempted on
				 * this device, so use REQ_FAILFAST_DEV.
				 * Also don't try to repair failed reads.
				 * It is expects that no bad block log
				 * is present.
				 */
205 206 207 208
	LastDev,		/* Seems to be the last working dev as
				 * it didn't fail, so don't use FailFast
				 * any more for metadata
				 */
G
Guoqing Jiang 已提交
209 210 211
	CollisionCheck,		/*
				 * check if there is collision between raid1
				 * serial bios.
212
				 */
213
};
L
Linus Torvalds 已提交
214

215
static inline int is_badblock(struct md_rdev *rdev, sector_t s, int sectors,
216 217 218
			      sector_t *first_bad, int *bad_sectors)
{
	if (unlikely(rdev->badblocks.count)) {
219
		int rv = badblocks_check(&rdev->badblocks, rdev->data_offset + s,
220 221 222 223 224 225 226 227
					sectors,
					first_bad, bad_sectors);
		if (rv)
			*first_bad -= rdev->data_offset;
		return rv;
	}
	return 0;
}
228
extern int rdev_set_badblocks(struct md_rdev *rdev, sector_t s, int sectors,
229 230 231
			      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
232 233
struct md_cluster_info;

234
/* change UNSUPPORTED_MDDEV_FLAGS for each array type if new flag is added */
235 236 237 238 239 240 241 242 243
enum mddev_flags {
	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_CLUSTER_RESYNC_LOCKED, /* cluster raid only, which means node
				   * already took resync lock, need to
				   * release the lock */
244 245 246 247
	MD_FAILFAST_SUPPORTED,	/* Using MD_FAILFAST on metadata writes is
				 * supported as calls to md_error() will
				 * never cause the array to become failed.
				 */
248
	MD_HAS_PPL,		/* The raid array has PPL feature set */
249
	MD_HAS_MULTIPLE_PPLS,	/* The raid array has multiple PPLs feature set */
250 251 252 253 254 255
	MD_ALLOW_SB_UPDATE,	/* md_check_recovery is allowed to update
				 * the metadata without taking reconfig_mutex.
				 */
	MD_UPDATING_SB,		/* md_check_recovery is updating the metadata
				 * without explicitly holding reconfig_mutex.
				 */
256 257 258
	MD_NOT_READY,		/* do_md_run() is active, so 'array_state'
				 * must not report that array is ready yet
				 */
259 260 261
	MD_BROKEN,              /* This is used in RAID-0/LINEAR only, to stop
				 * I/O in case an array member is gone/failed.
				 */
262
};
263 264 265 266 267 268 269 270

enum mddev_sb_flags {
	MD_SB_CHANGE_DEVS,		/* Some device status has changed */
	MD_SB_CHANGE_CLEAN,	/* transition to or from 'clean' */
	MD_SB_CHANGE_PENDING,	/* switch from 'clean' to 'active' in progress */
	MD_SB_NEED_REWRITE,	/* metadata write needs to be repeated */
};

G
Guoqing Jiang 已提交
271 272 273
#define NR_SERIAL_INFOS		8
/* record current range of serialize IOs */
struct serial_info {
274 275 276 277
	struct rb_node node;
	sector_t start;		/* start sector of rb node */
	sector_t last;		/* end sector of rb node */
	sector_t _subtree_last; /* highest sector in subtree of rb node */
278 279
};

280
struct mddev {
L
Linus Torvalds 已提交
281
	void				*private;
282
	struct md_personality		*pers;
L
Linus Torvalds 已提交
283 284
	dev_t				unit;
	int				md_minor;
285
	struct list_head		disks;
286
	unsigned long			flags;
287
	unsigned long			sb_flags;
288

289 290
	int				suspended;
	atomic_t			active_io;
L
Linus Torvalds 已提交
291
	int				ro;
292 293 294 295
	int				sysfs_active; /* set when sysfs deletes
						       * are happening, so run/
						       * takeover/stop are not safe
						       */
L
Linus Torvalds 已提交
296 297
	struct gendisk			*gendisk;

298
	struct kobject			kobj;
299 300
	int				hold_active;
#define	UNTIL_IOCTL	1
301
#define	UNTIL_STOP	2
302

L
Linus Torvalds 已提交
303 304 305 306 307
	/* Superblock information */
	int				major_version,
					minor_version,
					patch_version;
	int				persistent;
308
	int				external;	/* metadata is
309 310
							 * managed externally */
	char				metadata_type[17]; /* externally set*/
311
	int				chunk_sectors;
312
	time64_t			ctime, utime;
L
Linus Torvalds 已提交
313
	int				level, layout;
314
	char				clevel[16];
L
Linus Torvalds 已提交
315 316
	int				raid_disks;
	int				max_disks;
317
	sector_t			dev_sectors;	/* used size of
A
Andre Noll 已提交
318
							 * component devices */
319
	sector_t			array_sectors; /* exported array size */
D
Dan Williams 已提交
320 321
	int				external_size; /* size managed
							* externally */
L
Linus Torvalds 已提交
322
	__u64				events;
323 324 325 326 327 328
	/* 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
329 330 331

	char				uuid[16];

332 333 334 335 336 337
	/* 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;
338 339
	int				delta_disks, new_level, new_layout;
	int				new_chunk_sectors;
340
	int				reshape_backwards;
341

342 343
	struct md_thread		*thread;	/* management thread */
	struct md_thread		*sync_thread;	/* doing resync or reconstruct */
344 345 346 347 348 349 350 351

	/* '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;
352
	sector_t			curr_resync;	/* last block scheduled */
353 354 355 356 357 358 359
	/* 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
360 361
	unsigned long			resync_mark;	/* a recent timestamp */
	sector_t			resync_mark_cnt;/* blocks written at resync_mark */
362
	sector_t			curr_mark_cnt; /* blocks scheduled now */
L
Linus Torvalds 已提交
363 364

	sector_t			resync_max_sectors; /* may be set by personality */
365

366
	atomic64_t			resync_mismatches; /* count of sectors where
367 368
							    * parity/replica mismatch found
							    */
369 370 371 372

	/* allow user-space to request suspension of IO to regions of the array */
	sector_t			suspend_lo;
	sector_t			suspend_hi;
373 374 375 376
	/* if zero, use the system-wide default */
	int				sync_speed_min;
	int				sync_speed_max;

377 378 379
	/* resync even though the same disks are shared among md-devices */
	int				parallel_resync;

380
	int				ok_start_degraded;
381

L
Linus Torvalds 已提交
382
	unsigned long			recovery;
383 384 385 386 387 388
	/* 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
389 390

	int				in_sync;	/* know to not need resync */
N
NeilBrown 已提交
391 392 393 394 395 396 397 398 399 400
	/* '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;
401
	struct mutex			reconfig_mutex;
402 403
	atomic_t			active;		/* general refcount */
	atomic_t			openers;	/* number of active opens */
L
Linus Torvalds 已提交
404

405 406
	int				changed;	/* True if we might need to
							 * reread partition info */
L
Linus Torvalds 已提交
407 408 409 410 411 412 413
	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;
414 415
	sector_t			resync_min;	/* user requested sync
							 * starts here */
416 417
	sector_t			resync_max;	/* resync should pause
							 * when it gets here */
418

419
	struct kernfs_node		*sysfs_state;	/* handle for 'array_state'
420 421
							 * file in sysfs.
							 */
422
	struct kernfs_node		*sysfs_action;  /* handle for 'sync_action' */
423

424 425
	struct work_struct del_work;	/* used for delayed sysfs removal */

426 427 428 429 430
	/* "lock" protects:
	 *   flush_bio transition from NULL to !NULL
	 *   rdev superblocks, events
	 *   clearing MD_CHANGE_*
	 *   in_sync - and related safemode and MD_CHANGE changes
431
	 *   pers (also protected by reconfig_mutex and pending IO).
432
	 *   clearing ->bitmap
433
	 *   clearing ->bitmap_info.file
434 435
	 *   changing ->resync_{min,max}
	 *   setting MD_RECOVERY_RUNNING (which interacts with resync_{min,max})
436 437
	 */
	spinlock_t			lock;
438
	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
439
	atomic_t			pending_writes;	/* number of active superblock writes */
440

L
Linus Torvalds 已提交
441 442
	unsigned int			safemode;	/* if set, update "clean" superblock
							 * when no writes pending.
443
							 */
L
Linus Torvalds 已提交
444 445
	unsigned int			safemode_delay;
	struct timer_list		safemode_timer;
446 447
	struct percpu_ref		writes_pending;
	int				sync_checkers;	/* # of threads checking writes_pending */
448
	struct request_queue		*queue;	/* for plugging ... */
L
Linus Torvalds 已提交
449

450
	struct bitmap			*bitmap; /* the bitmap for the device */
451 452
	struct {
		struct file		*file; /* the bitmap file */
453
		loff_t			offset; /* offset from superblock of
454 455
						 * start of bitmap. May be
						 * negative, but not '0'
456
						 * For external metadata, offset
457
						 * from start of device.
458
						 */
459
		unsigned long		space; /* space available at this offset */
460
		loff_t			default_offset; /* this is the offset to use when
461 462 463
							 * hot-adding a bitmap.  It should
							 * eventually be settable by sysfs.
							 */
464 465
		unsigned long		default_space; /* space available at
							* default offset */
466
		struct mutex		mutex;
467
		unsigned long		chunksize;
468
		unsigned long		daemon_sleep; /* how many jiffies between updates? */
469
		unsigned long		max_write_behind; /* write-behind mode */
470
		int			external;
G
Goldwyn Rodrigues 已提交
471
		int			nodes; /* Maximum number of nodes in the cluster */
G
Goldwyn Rodrigues 已提交
472
		char                    cluster_name[64]; /* Name of the cluster */
473
	} bitmap_info;
474

475
	atomic_t			max_corr_read_errors; /* max read retries */
L
Linus Torvalds 已提交
476
	struct list_head		all_mddevs;
477

478
	struct attribute_group		*to_remove;
479

480 481
	struct bio_set			bio_set;
	struct bio_set			sync_set; /* for sync operations like
482 483
						   * metadata and bitmap writes
						   */
484

485 486 487 488 489 490
	/* Generic flush handling.
	 * The last to finish preflush schedules a worker to submit
	 * the rest of the request (without the REQ_PREFLUSH flag).
	 */
	struct bio *flush_bio;
	atomic_t flush_pending;
N
NeilBrown 已提交
491 492 493
	ktime_t start_flush, last_flush; /* last_flush is when the last completed
					  * flush was started.
					  */
494
	struct work_struct flush_work;
495
	struct work_struct event_work;	/* used by dm to report failure event */
G
Guoqing Jiang 已提交
496
	mempool_t *serial_info_pool;
497
	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
G
Goldwyn Rodrigues 已提交
498
	struct md_cluster_info		*cluster_info;
499
	unsigned int			good_device_nr;	/* good device num within cluster raid */
500
	unsigned int			noio_flag; /* for memalloc scope API */
501 502

	bool	has_superblocks:1;
503
	bool	fail_last_dev:1;
504
	bool	serialize_policy:1;
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521
};

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 */
522
	MD_RECOVERY_WAIT,	/* waiting for pers->start() to finish */
523
	MD_RESYNCING_REMOTE,	/* remote node is running resync thread */
L
Linus Torvalds 已提交
524 525
};

526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
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_trylock(struct mddev *mddev)
{
	return mutex_trylock(&mddev->reconfig_mutex);
}
extern void mddev_unlock(struct mddev *mddev);

L
Linus Torvalds 已提交
545 546
static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
547
	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
L
Linus Torvalds 已提交
548 549
}

550 551 552 553 554
static inline void md_sync_acct_bio(struct bio *bio, unsigned long nr_sectors)
{
	atomic_add(nr_sectors, &bio->bi_disk->sync_io);
}

555
struct md_personality
L
Linus Torvalds 已提交
556 557
{
	char *name;
558 559
	int level;
	struct list_head list;
L
Linus Torvalds 已提交
560
	struct module *owner;
561
	bool __must_check (*make_request)(struct mddev *mddev, struct bio *bio);
562 563 564 565
	/*
	 * start up works that do NOT require md_thread. tasks that
	 * requires md_thread should go into start()
	 */
566
	int (*run)(struct mddev *mddev);
567 568
	/* start up works that require md threads */
	int (*start)(struct mddev *mddev);
N
NeilBrown 已提交
569
	void (*free)(struct mddev *mddev, void *priv);
570
	void (*status)(struct seq_file *seq, struct mddev *mddev);
L
Linus Torvalds 已提交
571
	/* error_handler must set ->faulty and clear ->in_sync
572
	 * if appropriate, and should abort recovery if needed
L
Linus Torvalds 已提交
573
	 */
574 575
	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
576
	int (*hot_remove_disk) (struct mddev *mddev, struct md_rdev *rdev);
577
	int (*spare_active) (struct mddev *mddev);
578
	sector_t (*sync_request)(struct mddev *mddev, sector_t sector_nr, int *skipped);
579 580 581 582 583
	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);
584
	void (*update_reshape_pos) (struct mddev *mddev);
585 586 587
	/* quiesce suspends or resumes internal processing.
	 * 1 - stop new actions and wait for action io to complete
	 * 0 - return to normal behaviour
588
	 */
589
	void (*quiesce) (struct mddev *mddev, int quiesce);
590 591 592 593 594 595 596 597 598
	/* 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.
	 */
599
	void *(*takeover) (struct mddev *mddev);
600 601 602
	/* congested implements bdi.congested_fn().
	 * Will not be called while array is 'suspended' */
	int (*congested)(struct mddev *mddev, int bits);
603 604
	/* Changes the consistency policy of an active array. */
	int (*change_consistency_policy)(struct mddev *mddev, const char *buf);
L
Linus Torvalds 已提交
605 606
};

607 608
struct md_sysfs_entry {
	struct attribute attr;
609 610
	ssize_t (*show)(struct mddev *, char *);
	ssize_t (*store)(struct mddev *, const char *, size_t);
611
};
612
extern struct attribute_group md_bitmap_group;
613

614
static inline struct kernfs_node *sysfs_get_dirent_safe(struct kernfs_node *sd, char *name)
N
NeilBrown 已提交
615 616
{
	if (sd)
T
Tejun Heo 已提交
617
		return sysfs_get_dirent(sd, name);
N
NeilBrown 已提交
618 619
	return sd;
}
620
static inline void sysfs_notify_dirent_safe(struct kernfs_node *sd)
N
NeilBrown 已提交
621 622 623 624 625
{
	if (sd)
		sysfs_notify_dirent(sd);
}

626
static inline char * mdname (struct mddev * mddev)
L
Linus Torvalds 已提交
627 628 629 630
{
	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
}

631
static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)
632 633
{
	char nm[20];
634 635 636
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
637 638 639 640
		sprintf(nm, "rd%d", rdev->raid_disk);
		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
	} else
		return 0;
641 642
}

643
static inline void sysfs_unlink_rdev(struct mddev *mddev, struct md_rdev *rdev)
644 645
{
	char nm[20];
646 647 648
	if (!test_bit(Replacement, &rdev->flags) &&
	    !test_bit(Journal, &rdev->flags) &&
	    mddev->kobj.sd) {
649 650 651
		sprintf(nm, "rd%d", rdev->raid_disk);
		sysfs_remove_link(&mddev->kobj, nm);
	}
652 653
}

L
Linus Torvalds 已提交
654 655 656 657
/*
 * iterates through some rdev ringlist. It's safe to remove the
 * current 'rdev'. Dont touch 'tmp' though.
 */
658 659 660
#define rdev_for_each_list(rdev, tmp, head)				\
	list_for_each_entry_safe(rdev, tmp, head, same_set)

L
Linus Torvalds 已提交
661 662 663
/*
 * iterates through the 'same array disks' ringlist
 */
N
NeilBrown 已提交
664 665 666 667
#define rdev_for_each(rdev, mddev)				\
	list_for_each_entry(rdev, &((mddev)->disks), same_set)

#define rdev_for_each_safe(rdev, tmp, mddev)				\
668
	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
L
Linus Torvalds 已提交
669

670 671 672
#define rdev_for_each_rcu(rdev, mddev)				\
	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)

673
struct md_thread {
S
Shaohua Li 已提交
674
	void			(*run) (struct md_thread *thread);
675
	struct mddev		*mddev;
L
Linus Torvalds 已提交
676
	wait_queue_head_t	wqueue;
677
	unsigned long		flags;
L
Linus Torvalds 已提交
678
	struct task_struct	*tsk;
679
	unsigned long		timeout;
S
Shaohua Li 已提交
680
	void			*private;
681
};
L
Linus Torvalds 已提交
682 683 684

#define THREAD_WAKEUP  0

685 686 687 688 689
static inline void safe_put_page(struct page *p)
{
	if (p) put_page(p);
}

690 691
extern int register_md_personality(struct md_personality *p);
extern int unregister_md_personality(struct md_personality *p);
692 693 694 695 696
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);
697
extern struct md_thread *md_register_thread(
S
Shaohua Li 已提交
698
	void (*run)(struct md_thread *thread),
699 700 701 702
	struct mddev *mddev,
	const char *name);
extern void md_unregister_thread(struct md_thread **threadp);
extern void md_wakeup_thread(struct md_thread *thread);
703
extern void md_check_recovery(struct mddev *mddev);
704
extern void md_reap_sync_thread(struct mddev *mddev);
705
extern int mddev_init_writes_pending(struct mddev *mddev);
706
extern bool md_write_start(struct mddev *mddev, struct bio *bi);
707
extern void md_write_inc(struct mddev *mddev, struct bio *bi);
708 709 710
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);
711
extern void md_finish_reshape(struct mddev *mddev);
712 713

extern int mddev_congested(struct mddev *mddev, int bits);
714
extern bool __must_check md_flush_request(struct mddev *mddev, struct bio *bio);
715
extern void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
716
			   sector_t sector, int size, struct page *page);
717
extern int md_super_wait(struct mddev *mddev);
718
extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
M
Mike Christie 已提交
719 720
			struct page *page, int op, int op_flags,
			bool metadata_op);
S
Shaohua Li 已提交
721
extern void md_do_sync(struct md_thread *thread);
722
extern void md_new_event(struct mddev *mddev);
723
extern void md_allow_write(struct mddev *mddev);
724 725 726 727
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);
728
extern int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev);
729
extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
C
Christoph Hellwig 已提交
730

731 732
extern void mddev_init(struct mddev *mddev);
extern int md_run(struct mddev *mddev);
733
extern int md_start(struct mddev *mddev);
734 735
extern void md_stop(struct mddev *mddev);
extern void md_stop_writes(struct mddev *mddev);
736
extern int md_rdev_init(struct md_rdev *rdev);
737
extern void md_rdev_clear(struct md_rdev *rdev);
C
Christoph Hellwig 已提交
738

S
Shaohua Li 已提交
739
extern void md_handle_request(struct mddev *mddev, struct bio *bio);
740 741
extern void mddev_suspend(struct mddev *mddev);
extern void mddev_resume(struct mddev *mddev);
742
extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
743
				   struct mddev *mddev);
744

745
extern void md_reload_sb(struct mddev *mddev, int raid_disk);
746
extern void md_update_sb(struct mddev *mddev, int force);
747
extern void md_kick_rdev_from_array(struct md_rdev * rdev);
G
Guoqing Jiang 已提交
748
extern void mddev_create_serial_pool(struct mddev *mddev, struct md_rdev *rdev,
749
				     bool is_suspend);
750 751
extern void mddev_destroy_serial_pool(struct mddev *mddev, struct md_rdev *rdev,
				      bool is_suspend);
752
struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr);
753
struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev);
754

755 756 757 758 759 760 761 762 763 764 765 766 767
static inline bool is_mddev_broken(struct md_rdev *rdev, const char *md_type)
{
	int flags = rdev->bdev->bd_disk->flags;

	if (!(flags & GENHD_FL_UP)) {
		if (!test_and_set_bit(MD_BROKEN, &rdev->mddev->flags))
			pr_warn("md: %s: %s array has a missing/failed member\n",
				mdname(rdev->mddev), md_type);
		return true;
	}
	return false;
}

768 769 770 771 772 773 774 775 776
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);
	}
}

777
extern struct md_cluster_operations *md_cluster_ops;
G
Goldwyn Rodrigues 已提交
778 779 780 781
static inline int mddev_is_clustered(struct mddev *mddev)
{
	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;
}
782 783 784 785 786 787 788

/* clear unsupported mddev_flags */
static inline void mddev_clear_unsupported_flags(struct mddev *mddev,
	unsigned long unsupported_flags)
{
	mddev->flags &= ~unsupported_flags;
}
789 790 791 792

static inline void mddev_check_writesame(struct mddev *mddev, struct bio *bio)
{
	if (bio_op(bio) == REQ_OP_WRITE_SAME &&
793
	    !bio->bi_disk->queue->limits.max_write_same_sectors)
794 795
		mddev->queue->limits.max_write_same_sectors = 0;
}
796 797 798 799

static inline void mddev_check_write_zeroes(struct mddev *mddev, struct bio *bio)
{
	if (bio_op(bio) == REQ_OP_WRITE_ZEROES &&
800
	    !bio->bi_disk->queue->limits.max_write_zeroes_sectors)
801 802
		mddev->queue->limits.max_write_zeroes_sectors = 0;
}
803 804 805

void md_autostart_arrays(int part);

C
Christoph Hellwig 已提交
806
#endif /* _MD_MD_H */