[linux.git] / drivers / md / md.h

/*
   md_k.h : kernel internal structure of the Linux MD driver
          Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
	  
   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2, or (at your option)
   any later version.
   
   You should have received a copy of the GNU General Public License
   (for example /usr/src/linux/COPYING); if not, write to the Free
   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
*/

#ifndef _MD_MD_H
#define _MD_MD_H

#include <linux/blkdev.h>
#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>

#define MaxSector (~(sector_t)0)

typedef struct mddev_s mddev_t;
typedef struct mdk_rdev_s mdk_rdev_t;

/* generic plugging support - like that provided with request_queue,
 * but does not require a request_queue
 */
struct plug_handle {
	void			(*unplug_fn)(struct plug_handle *);
	struct timer_list	unplug_timer;
	struct work_struct	unplug_work;
	unsigned long		unplug_flag;
};
#define	PLUGGED_FLAG 1
void plugger_init(struct plug_handle *plug,
		  void (*unplug_fn)(struct plug_handle *));
void plugger_set_plug(struct plug_handle *plug);
int plugger_remove_plug(struct plug_handle *plug);
static inline void plugger_flush(struct plug_handle *plug)
{
	del_timer_sync(&plug->unplug_timer);
	cancel_work_sync(&plug->unplug_work);
}

/*
 * MD's 'extended' device
 */
struct mdk_rdev_s
{
	struct list_head same_set;	/* RAID devices within the same set */

	sector_t sectors;		/* Device size (in 512bytes sectors) */
	mddev_t *mddev;			/* RAID array if running */
	int last_events;		/* IO event timestamp */

	/*
	 * 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;
	struct block_device *bdev;	/* block device handle */

	struct page	*sb_page;
	int		sb_loaded;
	__u64		sb_events;
	sector_t	data_offset;	/* start of data in array */
	sector_t 	sb_start;	/* offset of the super block (in 512byte sectors) */
	int		sb_size;	/* bytes in the superblock */
	int		preferred_minor;	/* autorun support */

	struct kobject	kobj;

	/* 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
	 */

	unsigned long	flags;
#define	Faulty		1		/* device is known to have a fault */
#define	In_sync		2		/* device is in_sync with rest of array */
#define	WriteMostly	4		/* Avoid reading if at all possible */
#define	AutoDetected	7		/* added by auto-detect */
#define Blocked		8		/* An error occurred on an externally
					 * managed array, don't allow writes
					 * until it is cleared */
	wait_queue_head_t blocked_wait;

	int desc_nr;			/* descriptor index in the superblock */
	int raid_disk;			/* role of device in array */
	int new_raid_disk;		/* role that the device will have in
					 * the array after a level-change completes.
					 */
	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
					 */
	sector_t	recovery_offset;/* If this device has been partially
					 * recovered, this is where we were
					 * up to.
					 */

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

	struct sysfs_dirent *sysfs_state; /* handle for 'state'
					   * sysfs entry */
};

struct mddev_s
{
	void				*private;
	struct mdk_personality		*pers;
	dev_t				unit;
	int				md_minor;
	struct list_head 		disks;
	unsigned long			flags;
#define MD_CHANGE_DEVS	0	/* Some device status has changed */
#define MD_CHANGE_CLEAN 1	/* transition to or from 'clean' */
#define MD_CHANGE_PENDING 2	/* switch from 'clean' to 'active' in progress */

	int				suspended;
	atomic_t			active_io;
	int				ro;
	int				sysfs_active; /* set when sysfs deletes
						       * are happening, so run/
						       * takeover/stop are not safe
						       */
	int				ready; /* See when safe to pass 
						* IO requests down */
	struct gendisk			*gendisk;

	struct kobject			kobj;
	int				hold_active;
#define	UNTIL_IOCTL	1
#define	UNTIL_STOP	2

	/* Superblock information */
	int				major_version,
					minor_version,
					patch_version;
	int				persistent;
	int 				external;	/* metadata is
							 * managed externally */
	char				metadata_type[17]; /* externally set*/
	int				chunk_sectors;
	time_t				ctime, utime;
	int				level, layout;
	char				clevel[16];
	int				raid_disks;
	int				max_disks;
	sector_t			dev_sectors; 	/* used size of
							 * component devices */
	sector_t			array_sectors; /* exported array size */
	int				external_size; /* size managed
							* externally */
	__u64				events;
	/* 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;

	char				uuid[16];

	/* 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;
	int				delta_disks, new_level, new_layout;
	int				new_chunk_sectors;

	struct mdk_thread_s		*thread;	/* management thread */
	struct mdk_thread_s		*sync_thread;	/* doing resync or reconstruct */
	sector_t			curr_resync;	/* last block scheduled */
	/* 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;
	unsigned long			resync_mark;	/* a recent timestamp */
	sector_t			resync_mark_cnt;/* blocks written at resync_mark */
	sector_t			curr_mark_cnt; /* blocks scheduled now */

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

	sector_t			resync_mismatches; /* count of sectors where
							    * parity/replica mismatch found
							    */

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

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

	int				ok_start_degraded;
	/* recovery/resync flags 
	 * 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
	 * RECOVER:  doing recovery, or need to try it.
	 * INTR:     resync needs to be aborted for some reason
	 * DONE:     thread is done and is waiting to be reaped
	 * REQUEST:  user-space has requested a sync (used with SYNC)
	 * CHECK:    user-space request for check-only, no repair
	 * RESHAPE:  A reshape is happening
	 *
	 * If neither SYNC or RESHAPE are set, then it is a recovery.
	 */
#define	MD_RECOVERY_RUNNING	0
#define	MD_RECOVERY_SYNC	1
#define	MD_RECOVERY_RECOVER	2
#define	MD_RECOVERY_INTR	3
#define	MD_RECOVERY_DONE	4
#define	MD_RECOVERY_NEEDED	5
#define	MD_RECOVERY_REQUESTED	6
#define	MD_RECOVERY_CHECK	7
#define MD_RECOVERY_RESHAPE	8
#define	MD_RECOVERY_FROZEN	9

	unsigned long			recovery;
	int				recovery_disabled; /* if we detect that recovery
							    * will always fail, set this
							    * so we don't loop trying */

	int				in_sync;	/* know to not need resync */
	/* '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;
	struct mutex			reconfig_mutex;
	atomic_t			active;		/* general refcount */
	atomic_t			openers;	/* number of active opens */

	int				changed;	/* True if we might need to
							 * reread partition info */
	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;
	sector_t			resync_min;	/* user requested sync
							 * starts here */
	sector_t			resync_max;	/* resync should pause
							 * when it gets here */

	struct sysfs_dirent		*sysfs_state;	/* handle for 'array_state'
							 * file in sysfs.
							 */
	struct sysfs_dirent		*sysfs_action;  /* handle for 'sync_action' */

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

	spinlock_t			write_lock;
	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
	atomic_t			pending_writes;	/* number of active superblock writes */

	unsigned int			safemode;	/* if set, update "clean" superblock
							 * when no writes pending.
							 */ 
	unsigned int			safemode_delay;
	struct timer_list		safemode_timer;
	atomic_t			writes_pending; 
	struct request_queue		*queue;	/* for plugging ... */

	struct bitmap                   *bitmap; /* the bitmap for the device */
	struct {
		struct file		*file; /* the bitmap file */
		loff_t			offset; /* offset from superblock of
						 * start of bitmap. May be
						 * negative, but not '0'
						 * For external metadata, offset
						 * from start of device. 
						 */
		loff_t			default_offset; /* this is the offset to use when
							 * hot-adding a bitmap.  It should
							 * eventually be settable by sysfs.
							 */
		/* When md is serving under dm, it might use a
		 * dirty_log to store the bits.
		 */
		struct dm_dirty_log *log;

		struct mutex		mutex;
		unsigned long		chunksize;
		unsigned long		daemon_sleep; /* how many jiffies between updates? */
		unsigned long		max_write_behind; /* write-behind mode */
		int			external;
	} bitmap_info;

	atomic_t 			max_corr_read_errors; /* max read retries */
	struct list_head		all_mddevs;

	struct attribute_group		*to_remove;
	struct plug_handle		*plug; /* if used by personality */

	struct bio_set			*bio_set;

	/* Generic flush handling.
	 * The last to finish preflush schedules a worker to submit
	 * the rest of the request (without the REQ_FLUSH flag).
	 */
	struct bio *flush_bio;
	atomic_t flush_pending;
	struct work_struct flush_work;
	struct work_struct event_work;	/* used by dm to report failure event */
};


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

static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
        atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
}

struct mdk_personality
{
	char *name;
	int level;
	struct list_head list;
	struct module *owner;
	int (*make_request)(mddev_t *mddev, struct bio *bio);
	int (*run)(mddev_t *mddev);
	int (*stop)(mddev_t *mddev);
	void (*status)(struct seq_file *seq, mddev_t *mddev);
	/* error_handler must set ->faulty and clear ->in_sync
	 * if appropriate, and should abort recovery if needed 
	 */
	void (*error_handler)(mddev_t *mddev, mdk_rdev_t *rdev);
	int (*hot_add_disk) (mddev_t *mddev, mdk_rdev_t *rdev);
	int (*hot_remove_disk) (mddev_t *mddev, int number);
	int (*spare_active) (mddev_t *mddev);
	sector_t (*sync_request)(mddev_t *mddev, sector_t sector_nr, int *skipped, int go_faster);
	int (*resize) (mddev_t *mddev, sector_t sectors);
	sector_t (*size) (mddev_t *mddev, sector_t sectors, int raid_disks);
	int (*check_reshape) (mddev_t *mddev);
	int (*start_reshape) (mddev_t *mddev);
	void (*finish_reshape) (mddev_t *mddev);
	/* quiesce moves between quiescence states
	 * 0 - fully active
	 * 1 - no new requests allowed
	 * others - reserved
	 */
	void (*quiesce) (mddev_t *mddev, int state);
	/* 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.
	 */
	void *(*takeover) (mddev_t *mddev);
};


struct md_sysfs_entry {
	struct attribute attr;
	ssize_t (*show)(mddev_t *, char *);
	ssize_t (*store)(mddev_t *, const char *, size_t);
};
extern struct attribute_group md_bitmap_group;

static inline struct sysfs_dirent *sysfs_get_dirent_safe(struct sysfs_dirent *sd, char *name)
{
	if (sd)
		return sysfs_get_dirent(sd, NULL, name);
	return sd;
}
static inline void sysfs_notify_dirent_safe(struct sysfs_dirent *sd)
{
	if (sd)
		sysfs_notify_dirent(sd);
}

static inline char * mdname (mddev_t * mddev)
{
	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
}

/*
 * iterates through some rdev ringlist. It's safe to remove the
 * current 'rdev'. Dont touch 'tmp' though.
 */
#define rdev_for_each_list(rdev, tmp, head)				\
	list_for_each_entry_safe(rdev, tmp, head, same_set)

/*
 * iterates through the 'same array disks' ringlist
 */
#define rdev_for_each(rdev, tmp, mddev)				\
	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)

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

typedef struct mdk_thread_s {
	void			(*run) (mddev_t *mddev);
	mddev_t			*mddev;
	wait_queue_head_t	wqueue;
	unsigned long           flags;
	struct task_struct	*tsk;
	unsigned long		timeout;
} mdk_thread_t;

#define THREAD_WAKEUP  0

#define __wait_event_lock_irq(wq, condition, lock, cmd) 		\
do {									\
	wait_queue_t __wait;						\
	init_waitqueue_entry(&__wait, current);				\
									\
	add_wait_queue(&wq, &__wait);					\
	for (;;) {							\
		set_current_state(TASK_UNINTERRUPTIBLE);		\
		if (condition)						\
			break;						\
		spin_unlock_irq(&lock);					\
		cmd;							\
		schedule();						\
		spin_lock_irq(&lock);					\
	}								\
	current->state = TASK_RUNNING;					\
	remove_wait_queue(&wq, &__wait);				\
} while (0)

#define wait_event_lock_irq(wq, condition, lock, cmd) 			\
do {									\
	if (condition)	 						\
		break;							\
	__wait_event_lock_irq(wq, condition, lock, cmd);		\
} while (0)

static inline void safe_put_page(struct page *p)
{
	if (p) put_page(p);
}

extern int register_md_personality(struct mdk_personality *p);
extern int unregister_md_personality(struct mdk_personality *p);
extern mdk_thread_t * md_register_thread(void (*run) (mddev_t *mddev),
				mddev_t *mddev, const char *name);
extern void md_unregister_thread(mdk_thread_t *thread);
extern void md_wakeup_thread(mdk_thread_t *thread);
extern void md_check_recovery(mddev_t *mddev);
extern void md_write_start(mddev_t *mddev, struct bio *bi);
extern void md_write_end(mddev_t *mddev);
extern void md_done_sync(mddev_t *mddev, int blocks, int ok);
extern void md_error(mddev_t *mddev, mdk_rdev_t *rdev);

extern int mddev_congested(mddev_t *mddev, int bits);
extern void md_flush_request(mddev_t *mddev, struct bio *bio);
extern void md_super_write(mddev_t *mddev, mdk_rdev_t *rdev,
			   sector_t sector, int size, struct page *page);
extern void md_super_wait(mddev_t *mddev);
extern int sync_page_io(mdk_rdev_t *rdev, sector_t sector, int size, 
			struct page *page, int rw, bool metadata_op);
extern void md_do_sync(mddev_t *mddev);
extern void md_new_event(mddev_t *mddev);
extern int md_allow_write(mddev_t *mddev);
extern void md_wait_for_blocked_rdev(mdk_rdev_t *rdev, mddev_t *mddev);
extern void md_set_array_sectors(mddev_t *mddev, sector_t array_sectors);
extern int md_check_no_bitmap(mddev_t *mddev);
extern int md_integrity_register(mddev_t *mddev);
extern void md_integrity_add_rdev(mdk_rdev_t *rdev, mddev_t *mddev);
extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
extern void restore_bitmap_write_access(struct file *file);
extern void md_unplug(mddev_t *mddev);

extern void mddev_init(mddev_t *mddev);
extern int md_run(mddev_t *mddev);
extern void md_stop(mddev_t *mddev);
extern void md_stop_writes(mddev_t *mddev);
extern void md_rdev_init(mdk_rdev_t *rdev);

extern void mddev_suspend(mddev_t *mddev);
extern void mddev_resume(mddev_t *mddev);
extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
				   mddev_t *mddev);
extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
				   mddev_t *mddev);
#endif /* _MD_MD_H */
Commit	Line	Data
1da177e4 LT	1	/*
	2	md_k.h : kernel internal structure of the Linux MD driver
	3	Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
	4
	5	This program is free software; you can redistribute it and/or modify
	6	it under the terms of the GNU General Public License as published by
	7	the Free Software Foundation; either version 2, or (at your option)
	8	any later version.
	9
	10	You should have received a copy of the GNU General Public License
	11	(for example /usr/src/linux/COPYING); if not, write to the Free
	12	Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
	13	*/
	14
63fe0817 CH	15	#ifndef _MD_MD_H
	16	#define _MD_MD_H
	17
	18	#include <linux/blkdev.h>
	19	#include <linux/kobject.h>
	20	#include <linux/list.h>
	21	#include <linux/mm.h>
	22	#include <linux/mutex.h>
	23	#include <linux/timer.h>
	24	#include <linux/wait.h>
	25	#include <linux/workqueue.h>
9361401e	26
1da177e4	27	#define MaxSector (~(sector_t)0)
1da177e4	28
1da177e4 LT	29	typedef struct mddev_s mddev_t;
	30	typedef struct mdk_rdev_s mdk_rdev_t;
	31
2ac87401 N	32	/* generic plugging support - like that provided with request_queue,
	33	* but does not require a request_queue
	34	*/
	35	struct plug_handle {
	36	void (unplug_fn)(struct plug_handle );
	37	struct timer_list unplug_timer;
	38	struct work_struct unplug_work;
	39	unsigned long unplug_flag;
	40	};
	41	#define PLUGGED_FLAG 1
	42	void plugger_init(struct plug_handle *plug,
	43	void (unplug_fn)(struct plug_handle ));
	44	void plugger_set_plug(struct plug_handle *plug);
	45	int plugger_remove_plug(struct plug_handle *plug);
	46	static inline void plugger_flush(struct plug_handle *plug)
	47	{
	48	del_timer_sync(&plug->unplug_timer);
	49	cancel_work_sync(&plug->unplug_work);
	50	}
	51
1da177e4 LT	52	/*
	53	* MD's 'extended' device
	54	*/
	55	struct mdk_rdev_s
	56	{
	57	struct list_head same_set; /* RAID devices within the same set */
	58
dd8ac336	59	sector_t sectors; /* Device size (in 512bytes sectors) */
1da177e4	60	mddev_t mddev; / RAID array if running */
eea1bf38	61	int last_events; /* IO event timestamp */
1da177e4	62
a6ff7e08 JB	63	/*
	64	* If meta_bdev is non-NULL, it means that a separate device is
	65	* being used to store the metadata (superblock/bitmap) which
	66	* would otherwise be contained on the same device as the data (bdev).
	67	*/
	68	struct block_device *meta_bdev;
1da177e4 LT	69	struct block_device bdev; / block device handle */
	70
	71	struct page *sb_page;
	72	int sb_loaded;
42543769	73	__u64 sb_events;
1da177e4	74	sector_t data_offset; /* start of data in array */
0f420358	75	sector_t sb_start; /* offset of the super block (in 512byte sectors) */
0002b271	76	int sb_size; /* bytes in the superblock */
1da177e4 LT	77	int preferred_minor; /* autorun support */
1da177e4 LT	78
86e6ffdd N	79	struct kobject kobj;
86e6ffdd N	80
1da177e4 LT	81	/* A device can be in one of three states based on two flags:
	82	* Not working: faulty==1 in_sync==0
	83	* Fully working: faulty==0 in_sync==1
	84	* Working, but not
	85	* in sync with array
	86	* faulty==0 in_sync==0
	87	*
	88	* It can never have faulty==1, in_sync==1
	89	* This reduces the burden of testing multiple flags in many cases
	90	*/
1da177e4	91
b2d444d7 N	92	unsigned long flags;
	93	#define Faulty 1 /* device is known to have a fault */
	94	#define In_sync 2 /* device is in_sync with rest of array */
8ddf9efe	95	#define WriteMostly 4 /* Avoid reading if at all possible */
d0fae18f	96	#define AutoDetected 7 /* added by auto-detect */
25985edc	97	#define Blocked 8 /* An error occurred on an externally
6bfe0b49 DW	98	* managed array, don't allow writes
	99	* until it is cleared */
	100	wait_queue_head_t blocked_wait;
8ddf9efe	101
1da177e4 LT	102	int desc_nr; /* descriptor index in the superblock */
1da177e4 LT	103	int raid_disk; /* role of device in array */
e93f68a1 N	104	int new_raid_disk; /* role that the device will have in
	105	* the array after a level-change completes.
	106	*/
41158c7e N	107	int saved_raid_disk; /* role that device used to have in the
	108	* array and could again if we did a partial
	109	* resync from the bitmap
	110	*/
5fd6c1dc N	111	sector_t recovery_offset;/* If this device has been partially
	112	* recovered, this is where we were
	113	* up to.
	114	*/
1da177e4 LT	115
	116	atomic_t nr_pending; /* number of pending requests.
	117	* only maintained for arrays that
	118	* support hot removal
	119	*/
ba22dcbf N	120	atomic_t read_errors; /* number of consecutive read errors that
	121	* we have tried to ignore.
	122	*/
1e50915f RB	123	struct timespec last_read_error; /* monotonic time since our
	124	* last read error
	125	*/
4dbcdc75 N	126	atomic_t corrected_errors; /* number of corrected read errors,
	127	* for reporting to userspace and storing
	128	* in superblock.
	129	*/
5792a285	130	struct work_struct del_work; /* used for delayed sysfs removal */
3c0ee63a N	131
	132	struct sysfs_dirent sysfs_state; / handle for 'state'
	133	* sysfs entry */
1da177e4 LT	134	};
1da177e4 LT	135
1da177e4 LT	136	struct mddev_s
	137	{
	138	void *private;
2604b703	139	struct mdk_personality *pers;
1da177e4 LT	140	dev_t unit;
	141	int md_minor;
	142	struct list_head disks;
850b2b42 N	143	unsigned long flags;
	144	#define MD_CHANGE_DEVS 0 /* Some device status has changed */
	145	#define MD_CHANGE_CLEAN 1 /* transition to or from 'clean' */
070dc6dd	146	#define MD_CHANGE_PENDING 2 /* switch from 'clean' to 'active' in progress */
850b2b42	147
409c57f3 N	148	int suspended;
409c57f3 N	149	atomic_t active_io;
1da177e4	150	int ro;
bb4f1e9d N	151	int sysfs_active; /* set when sysfs deletes
	152	* are happening, so run/
	153	* takeover/stop are not safe
	154	*/
0ca69886 N	155	int ready; /* See when safe to pass
0ca69886 N	156	* IO requests down */
1da177e4 LT	157	struct gendisk *gendisk;
1da177e4 LT	158
eae1701f	159	struct kobject kobj;
d3374825 N	160	int hold_active;
d3374825 N	161	#define UNTIL_IOCTL 1
efeb53c0	162	#define UNTIL_STOP 2
eae1701f	163
1da177e4 LT	164	/* Superblock information */
	165	int major_version,
	166	minor_version,
	167	patch_version;
	168	int persistent;
e691063a N	169	int external; /* metadata is
	170	* managed externally */
	171	char metadata_type[17]; /* externally set*/
9d8f0363	172	int chunk_sectors;
1da177e4 LT	173	time_t ctime, utime;
1da177e4 LT	174	int level, layout;
d9d166c2	175	char clevel[16];
1da177e4 LT	176	int raid_disks;
1da177e4 LT	177	int max_disks;
58c0fed4 AN	178	sector_t dev_sectors; /* used size of
58c0fed4 AN	179	* component devices */
f233ea5c	180	sector_t array_sectors; /* exported array size */
b522adcd DW	181	int external_size; /* size managed
b522adcd DW	182	* externally */
1da177e4	183	__u64 events;
a8707c08 N	184	/* If the last 'event' was simply a clean->dirty transition, and
	185	* we didn't write it to the spares, then it is safe and simple
	186	* to just decrement the event count on a dirty->clean transition.
	187	* So we record that possibility here.
	188	*/
	189	int can_decrease_events;
1da177e4 LT	190
	191	char uuid[16];
	192
f6705578 N	193	/* If the array is being reshaped, we need to record the
	194	* new shape and an indication of where we are up to.
	195	* This is written to the superblock.
	196	* If reshape_position is MaxSector, then no reshape is happening (yet).
	197	*/
	198	sector_t reshape_position;
664e7c41 AN	199	int delta_disks, new_level, new_layout;
664e7c41 AN	200	int new_chunk_sectors;
f6705578	201
1da177e4 LT	202	struct mdk_thread_s thread; / management thread */
1da177e4 LT	203	struct mdk_thread_s sync_thread; / doing resync or reconstruct */
ff4e8d9a	204	sector_t curr_resync; /* last block scheduled */
97e4f42d N	205	/* As resync requests can complete out of order, we cannot easily track
	206	* how much resync has been completed. So we occasionally pause until
	207	* everything completes, then set curr_resync_completed to curr_resync.
	208	* As such it may be well behind the real resync mark, but it is a value
	209	* we are certain of.
	210	*/
	211	sector_t curr_resync_completed;
1da177e4 LT	212	unsigned long resync_mark; /* a recent timestamp */
1da177e4 LT	213	sector_t resync_mark_cnt;/* blocks written at resync_mark */
ff4e8d9a	214	sector_t curr_mark_cnt; /* blocks scheduled now */
1da177e4 LT	215
1da177e4 LT	216	sector_t resync_max_sectors; /* may be set by personality */
9d88883e N	217
	218	sector_t resync_mismatches; /* count of sectors where
	219	* parity/replica mismatch found
	220	*/
e464eafd N	221
	222	/* allow user-space to request suspension of IO to regions of the array */
	223	sector_t suspend_lo;
	224	sector_t suspend_hi;
88202a0c N	225	/* if zero, use the system-wide default */
	226	int sync_speed_min;
	227	int sync_speed_max;
	228
90b08710 BS	229	/* resync even though the same disks are shared among md-devices */
	230	int parallel_resync;
	231
6ff8d8ec	232	int ok_start_degraded;
1da177e4 LT	233	/* recovery/resync flags
	234	* NEEDED: we might need to start a resync/recover
	235	* RUNNING: a thread is running, or about to be started
	236	* SYNC: actually doing a resync, not a recovery
72a23c21	237	* RECOVER: doing recovery, or need to try it.
dfc70645	238	* INTR: resync needs to be aborted for some reason
1da177e4	239	* DONE: thread is done and is waiting to be reaped
24dd469d	240	* REQUEST: user-space has requested a sync (used with SYNC)
411c9403	241	* CHECK: user-space request for check-only, no repair
ccfcc3c1 N	242	* RESHAPE: A reshape is happening
	243	*
	244	* If neither SYNC or RESHAPE are set, then it is a recovery.
1da177e4 LT	245	*/
	246	#define MD_RECOVERY_RUNNING 0
	247	#define MD_RECOVERY_SYNC 1
72a23c21	248	#define MD_RECOVERY_RECOVER 2
1da177e4 LT	249	#define MD_RECOVERY_INTR 3
	250	#define MD_RECOVERY_DONE 4
	251	#define MD_RECOVERY_NEEDED 5
24dd469d N	252	#define MD_RECOVERY_REQUESTED 6
24dd469d N	253	#define MD_RECOVERY_CHECK 7
ccfcc3c1	254	#define MD_RECOVERY_RESHAPE 8
5fd6c1dc N	255	#define MD_RECOVERY_FROZEN 9
5fd6c1dc N	256
1da177e4	257	unsigned long recovery;
4044ba58 N	258	int recovery_disabled; /* if we detect that recovery
	259	* will always fail, set this
	260	* so we don't loop trying */
1da177e4 LT	261
1da177e4 LT	262	int in_sync; /* know to not need resync */
c8c00a69 N	263	/* 'open_mutex' avoids races between 'md_open' and 'do_md_stop', so
	264	* that we are never stopping an array while it is open.
	265	* 'reconfig_mutex' protects all other reconfiguration.
	266	* These locks are separate due to conflicting interactions
	267	* with bdev->bd_mutex.
	268	* Lock ordering is:
	269	* reconfig_mutex -> bd_mutex : e.g. do_md_run -> revalidate_disk
	270	* bd_mutex -> open_mutex: e.g. __blkdev_get -> md_open
	271	*/
	272	struct mutex open_mutex;
df5b89b3	273	struct mutex reconfig_mutex;
f2ea68cf N	274	atomic_t active; /* general refcount */
f2ea68cf N	275	atomic_t openers; /* number of active opens */
1da177e4	276
f0b4f7e2 N	277	int changed; /* True if we might need to
f0b4f7e2 N	278	* reread partition info */
1da177e4 LT	279	int degraded; /* whether md should consider
	280	* adding a spare
	281	*/
	282
	283	atomic_t recovery_active; /* blocks scheduled, but not written */
	284	wait_queue_head_t recovery_wait;
	285	sector_t recovery_cp;
5e96ee65 NB	286	sector_t resync_min; /* user requested sync
5e96ee65 NB	287	* starts here */
c6207277 N	288	sector_t resync_max; /* resync should pause
c6207277 N	289	* when it gets here */
06d91a5f	290
b62b7590 N	291	struct sysfs_dirent sysfs_state; / handle for 'array_state'
	292	* file in sysfs.
	293	*/
0c3573f1	294	struct sysfs_dirent sysfs_action; / handle for 'sync_action' */
b62b7590	295
d3374825 N	296	struct work_struct del_work; /* used for delayed sysfs removal */
d3374825 N	297
06d91a5f	298	spinlock_t write_lock;
3d310eb7	299	wait_queue_head_t sb_wait; /* for waiting on superblock updates */
7bfa19f2	300	atomic_t pending_writes; /* number of active superblock writes */
06d91a5f	301
1da177e4 LT	302	unsigned int safemode; /* if set, update "clean" superblock
	303	* when no writes pending.
	304	*/
	305	unsigned int safemode_delay;
	306	struct timer_list safemode_timer;
	307	atomic_t writes_pending;
165125e1	308	struct request_queue queue; / for plugging ... */
1da177e4	309
32a7627c	310	struct bitmap bitmap; / the bitmap for the device */
c3d9714e N	311	struct {
c3d9714e N	312	struct file file; / the bitmap file */
f6af949c	313	loff_t offset; /* offset from superblock of
c3d9714e N	314	* start of bitmap. May be
c3d9714e N	315	* negative, but not '0'
f6af949c N	316	* For external metadata, offset
f6af949c N	317	* from start of device.
c3d9714e	318	*/
f6af949c	319	loff_t default_offset; /* this is the offset to use when
c3d9714e N	320	* hot-adding a bitmap. It should
	321	* eventually be settable by sysfs.
	322	*/
e384e585 N	323	/* When md is serving under dm, it might use a
	324	* dirty_log to store the bits.
	325	*/
	326	struct dm_dirty_log *log;
	327
c3d9714e	328	struct mutex mutex;
42a04b50	329	unsigned long chunksize;
ac2f40be	330	unsigned long daemon_sleep; /* how many jiffies between updates? */
42a04b50	331	unsigned long max_write_behind; /* write-behind mode */
ece5cff0	332	int external;
c3d9714e	333	} bitmap_info;
32a7627c	334
1e50915f	335	atomic_t max_corr_read_errors; /* max read retries */
1da177e4	336	struct list_head all_mddevs;
a2826aa9	337
a64c876f	338	struct attribute_group *to_remove;
252ac522 N	339	struct plug_handle plug; / if used by personality */
252ac522 N	340
a167f663 N	341	struct bio_set *bio_set;
a167f663 N	342
e9c7469b TH	343	/* Generic flush handling.
	344	* The last to finish preflush schedules a worker to submit
	345	* the rest of the request (without the REQ_FLUSH flag).
a2826aa9	346	*/
e9c7469b	347	struct bio *flush_bio;
a2826aa9	348	atomic_t flush_pending;
e9c7469b	349	struct work_struct flush_work;
768a418d	350	struct work_struct event_work; /* used by dm to report failure event */
1da177e4 LT	351	};
	352
	353
	354	static inline void rdev_dec_pending(mdk_rdev_t rdev, mddev_t mddev)
	355	{
b2d444d7	356	int faulty = test_bit(Faulty, &rdev->flags);
1da177e4 LT	357	if (atomic_dec_and_test(&rdev->nr_pending) && faulty)
	358	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
	359	}
	360
	361	static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
	362	{
	363	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
	364	}
	365
2604b703	366	struct mdk_personality
1da177e4 LT	367	{
1da177e4 LT	368	char *name;
2604b703 N	369	int level;
2604b703 N	370	struct list_head list;
1da177e4	371	struct module *owner;
21a52c6d	372	int (make_request)(mddev_t mddev, struct bio *bio);
1da177e4 LT	373	int (run)(mddev_t mddev);
	374	int (stop)(mddev_t mddev);
	375	void (status)(struct seq_file seq, mddev_t *mddev);
	376	/* error_handler must set ->faulty and clear ->in_sync
	377	* if appropriate, and should abort recovery if needed
	378	*/
	379	void (error_handler)(mddev_t mddev, mdk_rdev_t *rdev);
	380	int (hot_add_disk) (mddev_t mddev, mdk_rdev_t *rdev);
	381	int (hot_remove_disk) (mddev_t mddev, int number);
	382	int (spare_active) (mddev_t mddev);
57afd89f	383	sector_t (sync_request)(mddev_t mddev, sector_t sector_nr, int *skipped, int go_faster);
1da177e4	384	int (resize) (mddev_t mddev, sector_t sectors);
80c3a6ce	385	sector_t (size) (mddev_t mddev, sector_t sectors, int raid_disks);
63c70c4f N	386	int (check_reshape) (mddev_t mddev);
63c70c4f N	387	int (start_reshape) (mddev_t mddev);
cea9c228	388	void (finish_reshape) (mddev_t mddev);
36fa3063 N	389	/* quiesce moves between quiescence states
	390	* 0 - fully active
	391	* 1 - no new requests allowed
	392	* others - reserved
	393	*/
	394	void (quiesce) (mddev_t mddev, int state);
245f46c2 N	395	/* takeover is used to transition an array from one
	396	* personality to another. The new personality must be able
	397	* to handle the data in the current layout.
	398	* e.g. 2drive raid1 -> 2drive raid5
	399	* ndrive raid5 -> degraded n+1drive raid6 with special layout
	400	* If the takeover succeeds, a new 'private' structure is returned.
	401	* This needs to be installed and then ->run used to activate the
	402	* array.
	403	*/
	404	void (takeover) (mddev_t *mddev);
1da177e4 LT	405	};
	406
	407
007583c9 N	408	struct md_sysfs_entry {
	409	struct attribute attr;
	410	ssize_t (show)(mddev_t , char *);
	411	ssize_t (store)(mddev_t , const char *, size_t);
	412	};
43a70507	413	extern struct attribute_group md_bitmap_group;
007583c9	414
00bcb4ac N	415	static inline struct sysfs_dirent sysfs_get_dirent_safe(struct sysfs_dirent sd, char *name)
	416	{
	417	if (sd)
	418	return sysfs_get_dirent(sd, NULL, name);
	419	return sd;
	420	}
	421	static inline void sysfs_notify_dirent_safe(struct sysfs_dirent *sd)
	422	{
	423	if (sd)
	424	sysfs_notify_dirent(sd);
	425	}
	426
1da177e4 LT	427	static inline char * mdname (mddev_t * mddev)
	428	{
	429	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";
	430	}
	431
1da177e4 LT	432	/*
	433	* iterates through some rdev ringlist. It's safe to remove the
	434	* current 'rdev'. Dont touch 'tmp' though.
	435	*/
159ec1fc CR	436	#define rdev_for_each_list(rdev, tmp, head) \
	437	list_for_each_entry_safe(rdev, tmp, head, same_set)
	438
1da177e4 LT	439	/*
	440	* iterates through the 'same array disks' ringlist
	441	*/
d089c6af	442	#define rdev_for_each(rdev, tmp, mddev) \
159ec1fc	443	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
1da177e4	444
4b80991c N	445	#define rdev_for_each_rcu(rdev, mddev) \
	446	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)
	447
1da177e4 LT	448	typedef struct mdk_thread_s {
	449	void (run) (mddev_t mddev);
	450	mddev_t *mddev;
	451	wait_queue_head_t wqueue;
	452	unsigned long flags;
1da177e4	453	struct task_struct *tsk;
32a7627c	454	unsigned long timeout;
1da177e4 LT	455	} mdk_thread_t;
	456
	457	#define THREAD_WAKEUP 0
	458
	459	#define __wait_event_lock_irq(wq, condition, lock, cmd) \
	460	do { \
	461	wait_queue_t __wait; \
	462	init_waitqueue_entry(&__wait, current); \
	463	\
	464	add_wait_queue(&wq, &__wait); \
	465	for (;;) { \
	466	set_current_state(TASK_UNINTERRUPTIBLE); \
	467	if (condition) \
	468	break; \
	469	spin_unlock_irq(&lock); \
	470	cmd; \
	471	schedule(); \
	472	spin_lock_irq(&lock); \
	473	} \
	474	current->state = TASK_RUNNING; \
	475	remove_wait_queue(&wq, &__wait); \
	476	} while (0)
	477
	478	#define wait_event_lock_irq(wq, condition, lock, cmd) \
	479	do { \
	480	if (condition) \
	481	break; \
	482	__wait_event_lock_irq(wq, condition, lock, cmd); \
	483	} while (0)
	484
1345b1d8 N	485	static inline void safe_put_page(struct page *p)
	486	{
	487	if (p) put_page(p);
	488	}
	489
92022950 N	490	extern int register_md_personality(struct mdk_personality *p);
	491	extern int unregister_md_personality(struct mdk_personality *p);
	492	extern mdk_thread_t * md_register_thread(void (run) (mddev_t mddev),
	493	mddev_t mddev, const char name);
	494	extern void md_unregister_thread(mdk_thread_t *thread);
	495	extern void md_wakeup_thread(mdk_thread_t *thread);
	496	extern void md_check_recovery(mddev_t *mddev);
	497	extern void md_write_start(mddev_t mddev, struct bio bi);
	498	extern void md_write_end(mddev_t *mddev);
	499	extern void md_done_sync(mddev_t *mddev, int blocks, int ok);
	500	extern void md_error(mddev_t mddev, mdk_rdev_t rdev);
	501
3fa841d7	502	extern int mddev_congested(mddev_t *mddev, int bits);
e9c7469b	503	extern void md_flush_request(mddev_t mddev, struct bio bio);
92022950 N	504	extern void md_super_write(mddev_t mddev, mdk_rdev_t rdev,
	505	sector_t sector, int size, struct page *page);
	506	extern void md_super_wait(mddev_t *mddev);
ccebd4c4 JB	507	extern int sync_page_io(mdk_rdev_t *rdev, sector_t sector, int size,
ccebd4c4 JB	508	struct page *page, int rw, bool metadata_op);
92022950 N	509	extern void md_do_sync(mddev_t *mddev);
	510	extern void md_new_event(mddev_t *mddev);
	511	extern int md_allow_write(mddev_t *mddev);
	512	extern void md_wait_for_blocked_rdev(mdk_rdev_t rdev, mddev_t mddev);
1f403624	513	extern void md_set_array_sectors(mddev_t *mddev, sector_t array_sectors);
0894cc30	514	extern int md_check_no_bitmap(mddev_t *mddev);
ac5e7113	515	extern int md_integrity_register(mddev_t *mddev);
72e02075 N	516	extern void md_integrity_add_rdev(mdk_rdev_t rdev, mddev_t mddev);
72e02075 N	517	extern int strict_strtoul_scaled(const char cp, unsigned long res, int scale);
43a70507	518	extern void restore_bitmap_write_access(struct file *file);
b63d7c2e	519	extern void md_unplug(mddev_t *mddev);
63fe0817	520
390ee602 N	521	extern void mddev_init(mddev_t *mddev);
	522	extern int md_run(mddev_t *mddev);
	523	extern void md_stop(mddev_t *mddev);
	524	extern void md_stop_writes(mddev_t *mddev);
e8bb9a83	525	extern void md_rdev_init(mdk_rdev_t *rdev);
63fe0817	526
390ee602 N	527	extern void mddev_suspend(mddev_t *mddev);
390ee602 N	528	extern void mddev_resume(mddev_t *mddev);
a167f663 N	529	extern struct bio bio_clone_mddev(struct bio bio, gfp_t gfp_mask,
	530	mddev_t *mddev);
	531	extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
	532	mddev_t *mddev);
63fe0817	533	#endif /* _MD_MD_H */