[mirror_zfs-debian.git] / include / sys / vdev_impl.h

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2011, 2015 by Delphix. All rights reserved.
 */

#ifndef _SYS_VDEV_IMPL_H
#define	_SYS_VDEV_IMPL_H

#include <sys/avl.h>
#include <sys/dmu.h>
#include <sys/metaslab.h>
#include <sys/nvpair.h>
#include <sys/space_map.h>
#include <sys/vdev.h>
#include <sys/dkio.h>
#include <sys/uberblock_impl.h>
#include <sys/zfs_ratelimit.h>

#ifdef	__cplusplus
extern "C" {
#endif

/*
 * Virtual device descriptors.
 *
 * All storage pool operations go through the virtual device framework,
 * which provides data replication and I/O scheduling.
 */

/*
 * Forward declarations that lots of things need.
 */
typedef struct vdev_queue vdev_queue_t;
typedef struct vdev_cache vdev_cache_t;
typedef struct vdev_cache_entry vdev_cache_entry_t;
struct abd;

extern int zfs_vdev_queue_depth_pct;
extern uint32_t zfs_vdev_async_write_max_active;

/*
 * Virtual device operations
 */
typedef int	vdev_open_func_t(vdev_t *vd, uint64_t *size, uint64_t *max_size,
    uint64_t *ashift);
typedef void	vdev_close_func_t(vdev_t *vd);
typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize);
typedef void	vdev_io_start_func_t(zio_t *zio);
typedef void	vdev_io_done_func_t(zio_t *zio);
typedef void	vdev_state_change_func_t(vdev_t *vd, int, int);
typedef boolean_t vdev_need_resilver_func_t(vdev_t *vd, uint64_t, size_t);
typedef void	vdev_hold_func_t(vdev_t *vd);
typedef void	vdev_rele_func_t(vdev_t *vd);

typedef const struct vdev_ops {
	vdev_open_func_t		*vdev_op_open;
	vdev_close_func_t		*vdev_op_close;
	vdev_asize_func_t		*vdev_op_asize;
	vdev_io_start_func_t		*vdev_op_io_start;
	vdev_io_done_func_t		*vdev_op_io_done;
	vdev_state_change_func_t	*vdev_op_state_change;
	vdev_need_resilver_func_t	*vdev_op_need_resilver;
	vdev_hold_func_t		*vdev_op_hold;
	vdev_rele_func_t		*vdev_op_rele;
	char				vdev_op_type[16];
	boolean_t			vdev_op_leaf;
} vdev_ops_t;

/*
 * Virtual device properties
 */
struct vdev_cache_entry {
	struct abd	*ve_abd;
	uint64_t	ve_offset;
	clock_t		ve_lastused;
	avl_node_t	ve_offset_node;
	avl_node_t	ve_lastused_node;
	uint32_t	ve_hits;
	uint16_t	ve_missed_update;
	zio_t		*ve_fill_io;
};

struct vdev_cache {
	avl_tree_t	vc_offset_tree;
	avl_tree_t	vc_lastused_tree;
	kmutex_t	vc_lock;
};

typedef struct vdev_queue_class {
	uint32_t	vqc_active;

	/*
	 * Sorted by offset or timestamp, depending on if the queue is
	 * LBA-ordered vs FIFO.
	 */
	avl_tree_t	vqc_queued_tree;
} vdev_queue_class_t;

struct vdev_queue {
	vdev_t		*vq_vdev;
	vdev_queue_class_t vq_class[ZIO_PRIORITY_NUM_QUEUEABLE];
	avl_tree_t	vq_active_tree;
	avl_tree_t	vq_read_offset_tree;
	avl_tree_t	vq_write_offset_tree;
	uint64_t	vq_last_offset;
	hrtime_t	vq_io_complete_ts; /* time last i/o completed */
	hrtime_t	vq_io_delta_ts;
	zio_t		vq_io_search; /* used as local for stack reduction */
	kmutex_t	vq_lock;
};

/*
 * Virtual device descriptor
 */
struct vdev {
	/*
	 * Common to all vdev types.
	 */
	uint64_t	vdev_id;	/* child number in vdev parent	*/
	uint64_t	vdev_guid;	/* unique ID for this vdev	*/
	uint64_t	vdev_guid_sum;	/* self guid + all child guids	*/
	uint64_t	vdev_orig_guid;	/* orig. guid prior to remove	*/
	uint64_t	vdev_asize;	/* allocatable device capacity	*/
	uint64_t	vdev_min_asize;	/* min acceptable asize		*/
	uint64_t	vdev_max_asize;	/* max acceptable asize		*/
	uint64_t	vdev_ashift;	/* block alignment shift	*/
	uint64_t	vdev_state;	/* see VDEV_STATE_* #defines	*/
	uint64_t	vdev_prevstate;	/* used when reopening a vdev	*/
	vdev_ops_t	*vdev_ops;	/* vdev operations		*/
	spa_t		*vdev_spa;	/* spa for this vdev		*/
	void		*vdev_tsd;	/* type-specific data		*/
	vnode_t		*vdev_name_vp;	/* vnode for pathname		*/
	vnode_t		*vdev_devid_vp;	/* vnode for devid		*/
	vdev_t		*vdev_top;	/* top-level vdev		*/
	vdev_t		*vdev_parent;	/* parent vdev			*/
	vdev_t		**vdev_child;	/* array of children		*/
	uint64_t	vdev_children;	/* number of children		*/
	vdev_stat_t	vdev_stat;	/* virtual device statistics	*/
	vdev_stat_ex_t	vdev_stat_ex;	/* extended statistics		*/
	boolean_t	vdev_expanding;	/* expand the vdev?		*/
	boolean_t	vdev_reopening;	/* reopen in progress?		*/
	boolean_t	vdev_nonrot;	/* true if solid state		*/
	int		vdev_open_error; /* error on last open		*/
	kthread_t	*vdev_open_thread; /* thread opening children	*/
	uint64_t	vdev_crtxg;	/* txg when top-level was added */

	/*
	 * Top-level vdev state.
	 */
	uint64_t	vdev_ms_array;	/* metaslab array object	*/
	uint64_t	vdev_ms_shift;	/* metaslab size shift		*/
	uint64_t	vdev_ms_count;	/* number of metaslabs		*/
	metaslab_group_t *vdev_mg;	/* metaslab group		*/
	metaslab_t	**vdev_ms;	/* metaslab array		*/
	uint64_t	vdev_pending_fastwrite; /* allocated fastwrites */
	txg_list_t	vdev_ms_list;	/* per-txg dirty metaslab lists	*/
	txg_list_t	vdev_dtl_list;	/* per-txg dirty DTL lists	*/
	txg_node_t	vdev_txg_node;	/* per-txg dirty vdev linkage	*/
	boolean_t	vdev_remove_wanted; /* async remove wanted?	*/
	boolean_t	vdev_probe_wanted; /* async probe wanted?	*/
	list_node_t	vdev_config_dirty_node; /* config dirty list	*/
	list_node_t	vdev_state_dirty_node; /* state dirty list	*/
	uint64_t	vdev_deflate_ratio; /* deflation ratio (x512)	*/
	uint64_t	vdev_islog;	/* is an intent log device	*/
	uint64_t	vdev_removing;	/* device is being removed?	*/
	boolean_t	vdev_ishole;	/* is a hole in the namespace	*/
	kmutex_t	vdev_queue_lock; /* protects vdev_queue_depth	*/
	uint64_t	vdev_top_zap;

	/*
	 * The queue depth parameters determine how many async writes are
	 * still pending (i.e. allocated by net yet issued to disk) per
	 * top-level (vdev_async_write_queue_depth) and the maximum allowed
	 * (vdev_max_async_write_queue_depth). These values only apply to
	 * top-level vdevs.
	 */
	uint64_t	vdev_async_write_queue_depth;
	uint64_t	vdev_max_async_write_queue_depth;

	/*
	 * Leaf vdev state.
	 */
	range_tree_t	*vdev_dtl[DTL_TYPES]; /* dirty time logs	*/
	space_map_t	*vdev_dtl_sm;	/* dirty time log space map	*/
	txg_node_t	vdev_dtl_node;	/* per-txg dirty DTL linkage	*/
	uint64_t	vdev_dtl_object; /* DTL object			*/
	uint64_t	vdev_psize;	/* physical device capacity	*/
	uint64_t	vdev_wholedisk;	/* true if this is a whole disk */
	uint64_t	vdev_offline;	/* persistent offline state	*/
	uint64_t	vdev_faulted;	/* persistent faulted state	*/
	uint64_t	vdev_degraded;	/* persistent degraded state	*/
	uint64_t	vdev_removed;	/* persistent removed state	*/
	uint64_t	vdev_resilver_txg; /* persistent resilvering state */
	uint64_t	vdev_nparity;	/* number of parity devices for raidz */
	char		*vdev_path;	/* vdev path (if any)		*/
	char		*vdev_devid;	/* vdev devid (if any)		*/
	char		*vdev_physpath;	/* vdev device path (if any)	*/
	char		*vdev_enc_sysfs_path;	/* enclosure sysfs path */
	char		*vdev_fru;	/* physical FRU location	*/
	uint64_t	vdev_not_present; /* not present during import	*/
	uint64_t	vdev_unspare;	/* unspare when resilvering done */
	boolean_t	vdev_nowritecache; /* true if flushwritecache failed */
	boolean_t	vdev_checkremove; /* temporary online test	*/
	boolean_t	vdev_forcefault; /* force online fault		*/
	boolean_t	vdev_splitting;	/* split or repair in progress  */
	boolean_t	vdev_delayed_close; /* delayed device close?	*/
	boolean_t	vdev_tmpoffline; /* device taken offline temporarily? */
	boolean_t	vdev_detached;	/* device detached?		*/
	boolean_t	vdev_cant_read;	/* vdev is failing all reads	*/
	boolean_t	vdev_cant_write; /* vdev is failing all writes	*/
	boolean_t	vdev_isspare;	/* was a hot spare		*/
	boolean_t	vdev_isl2cache;	/* was a l2cache device		*/
	boolean_t	vdev_copy_uberblocks;  /* post expand copy uberblocks */
	vdev_queue_t	vdev_queue;	/* I/O deadline schedule queue	*/
	vdev_cache_t	vdev_cache;	/* physical block cache		*/
	spa_aux_vdev_t	*vdev_aux;	/* for l2cache and spares vdevs	*/
	zio_t		*vdev_probe_zio; /* root of current probe	*/
	vdev_aux_t	vdev_label_aux;	/* on-disk aux state		*/
	uint64_t	vdev_leaf_zap;
	hrtime_t	vdev_mmp_pending; /* 0 if write finished	*/

	/*
	 * For DTrace to work in userland (libzpool) context, these fields must
	 * remain at the end of the structure.  DTrace will use the kernel's
	 * CTF definition for 'struct vdev', and since the size of a kmutex_t is
	 * larger in userland, the offsets for the rest of the fields would be
	 * incorrect.
	 */
	kmutex_t	vdev_dtl_lock;	/* vdev_dtl_{map,resilver}	*/
	kmutex_t	vdev_stat_lock;	/* vdev_stat			*/
	kmutex_t	vdev_probe_lock; /* protects vdev_probe_zio	*/

	/*
	 * We rate limit ZIO delay and ZIO checksum events, since they
	 * can flood ZED with tons of events when a drive is acting up.
	 */
#define	DELAYS_PER_SECOND 5
#define	CHECKSUMS_PER_SECOND 5
	zfs_ratelimit_t vdev_delay_rl;
	zfs_ratelimit_t vdev_checksum_rl;
};

#define	VDEV_RAIDZ_MAXPARITY	3

#define	VDEV_PAD_SIZE		(8 << 10)
/* 2 padding areas (vl_pad1 and vl_pad2) to skip */
#define	VDEV_SKIP_SIZE		VDEV_PAD_SIZE * 2
#define	VDEV_PHYS_SIZE		(112 << 10)
#define	VDEV_UBERBLOCK_RING	(128 << 10)

/*
 * MMP blocks occupy the last MMP_BLOCKS_PER_LABEL slots in the uberblock
 * ring when MMP is enabled.
 */
#define	MMP_BLOCKS_PER_LABEL	1

/* The largest uberblock we support is 8k. */
#define	MAX_UBERBLOCK_SHIFT (13)
#define	VDEV_UBERBLOCK_SHIFT(vd)	\
	MIN(MAX((vd)->vdev_top->vdev_ashift, UBERBLOCK_SHIFT), \
	    MAX_UBERBLOCK_SHIFT)
#define	VDEV_UBERBLOCK_COUNT(vd)	\
	(VDEV_UBERBLOCK_RING >> VDEV_UBERBLOCK_SHIFT(vd))
#define	VDEV_UBERBLOCK_OFFSET(vd, n)	\
	offsetof(vdev_label_t, vl_uberblock[(n) << VDEV_UBERBLOCK_SHIFT(vd)])
#define	VDEV_UBERBLOCK_SIZE(vd)		(1ULL << VDEV_UBERBLOCK_SHIFT(vd))

typedef struct vdev_phys {
	char		vp_nvlist[VDEV_PHYS_SIZE - sizeof (zio_eck_t)];
	zio_eck_t	vp_zbt;
} vdev_phys_t;

typedef struct vdev_label {
	char		vl_pad1[VDEV_PAD_SIZE];			/*  8K */
	char		vl_pad2[VDEV_PAD_SIZE];			/*  8K */
	vdev_phys_t	vl_vdev_phys;				/* 112K	*/
	char		vl_uberblock[VDEV_UBERBLOCK_RING];	/* 128K	*/
} vdev_label_t;							/* 256K total */

/*
 * vdev_dirty() flags
 */
#define	VDD_METASLAB	0x01
#define	VDD_DTL		0x02

/* Offset of embedded boot loader region on each label */
#define	VDEV_BOOT_OFFSET	(2 * sizeof (vdev_label_t))
/*
 * Size of embedded boot loader region on each label.
 * The total size of the first two labels plus the boot area is 4MB.
 */
#define	VDEV_BOOT_SIZE		(7ULL << 19)			/* 3.5M */

/*
 * Size of label regions at the start and end of each leaf device.
 */
#define	VDEV_LABEL_START_SIZE	(2 * sizeof (vdev_label_t) + VDEV_BOOT_SIZE)
#define	VDEV_LABEL_END_SIZE	(2 * sizeof (vdev_label_t))
#define	VDEV_LABELS		4
#define	VDEV_BEST_LABEL		VDEV_LABELS

#define	VDEV_ALLOC_LOAD		0
#define	VDEV_ALLOC_ADD		1
#define	VDEV_ALLOC_SPARE	2
#define	VDEV_ALLOC_L2CACHE	3
#define	VDEV_ALLOC_ROOTPOOL	4
#define	VDEV_ALLOC_SPLIT	5
#define	VDEV_ALLOC_ATTACH	6

/*
 * Allocate or free a vdev
 */
extern vdev_t *vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid,
    vdev_ops_t *ops);
extern int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *config,
    vdev_t *parent, uint_t id, int alloctype);
extern void vdev_free(vdev_t *vd);

/*
 * Add or remove children and parents
 */
extern void vdev_add_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_remove_child(vdev_t *pvd, vdev_t *cvd);
extern void vdev_compact_children(vdev_t *pvd);
extern vdev_t *vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops);
extern void vdev_remove_parent(vdev_t *cvd);

/*
 * vdev sync load and sync
 */
extern void vdev_load_log_state(vdev_t *nvd, vdev_t *ovd);
extern boolean_t vdev_log_state_valid(vdev_t *vd);
extern void vdev_load(vdev_t *vd);
extern int vdev_dtl_load(vdev_t *vd);
extern void vdev_sync(vdev_t *vd, uint64_t txg);
extern void vdev_sync_done(vdev_t *vd, uint64_t txg);
extern void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg);
extern void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg);

/*
 * Available vdev types.
 */
extern vdev_ops_t vdev_root_ops;
extern vdev_ops_t vdev_mirror_ops;
extern vdev_ops_t vdev_replacing_ops;
extern vdev_ops_t vdev_raidz_ops;
extern vdev_ops_t vdev_disk_ops;
extern vdev_ops_t vdev_file_ops;
extern vdev_ops_t vdev_missing_ops;
extern vdev_ops_t vdev_hole_ops;
extern vdev_ops_t vdev_spare_ops;

/*
 * Common size functions
 */
extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize);
extern uint64_t vdev_get_min_asize(vdev_t *vd);
extern void vdev_set_min_asize(vdev_t *vd);

/*
 * Global variables
 */
/* zdb uses this tunable, so it must be declared here to make lint happy. */
extern int zfs_vdev_cache_size;

#ifdef	__cplusplus
}
#endif

#endif	/* _SYS_VDEV_IMPL_H */
Commit	Line	Data
34dc7c2f BB	1	/*
	2	* CDDL HEADER START
	3	*
	4	* The contents of this file are subject to the terms of the
	5	* Common Development and Distribution License (the "License").
	6	* You may not use this file except in compliance with the License.
	7	*
	8	* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
	9	* or http://www.opensolaris.org/os/licensing.
	10	* See the License for the specific language governing permissions
	11	* and limitations under the License.
	12	*
	13	* When distributing Covered Code, include this CDDL HEADER in each
	14	* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
	15	* If applicable, add the following below this CDDL HEADER, with the
	16	* fields enclosed by brackets "[]" replaced with your own identifying
	17	* information: Portions Copyright [yyyy] [name of copyright owner]
	18	*
	19	* CDDL HEADER END
	20	*/
	21	/*
428870ff	22	* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
e10b0808	23	* Copyright (c) 2011, 2015 by Delphix. All rights reserved.
34dc7c2f BB	24	*/
	25
	26	#ifndef _SYS_VDEV_IMPL_H
	27	#define _SYS_VDEV_IMPL_H
	28
34dc7c2f BB	29	#include <sys/avl.h>
	30	#include <sys/dmu.h>
	31	#include <sys/metaslab.h>
	32	#include <sys/nvpair.h>
	33	#include <sys/space_map.h>
	34	#include <sys/vdev.h>
	35	#include <sys/dkio.h>
	36	#include <sys/uberblock_impl.h>
cae5b340	37	#include <sys/zfs_ratelimit.h>
34dc7c2f BB	38
	39	#ifdef __cplusplus
	40	extern "C" {
	41	#endif
	42
	43	/*
	44	* Virtual device descriptors.
	45	*
	46	* All storage pool operations go through the virtual device framework,
	47	* which provides data replication and I/O scheduling.
	48	*/
	49
	50	/*
	51	* Forward declarations that lots of things need.
	52	*/
	53	typedef struct vdev_queue vdev_queue_t;
	54	typedef struct vdev_cache vdev_cache_t;
	55	typedef struct vdev_cache_entry vdev_cache_entry_t;
cae5b340 AX	56	struct abd;
	57
	58	extern int zfs_vdev_queue_depth_pct;
	59	extern uint32_t zfs_vdev_async_write_max_active;
34dc7c2f BB	60
	61	/*
	62	* Virtual device operations
	63	*/
1bd201e7 CS	64	typedef int vdev_open_func_t(vdev_t vd, uint64_t size, uint64_t *max_size,
1bd201e7 CS	65	uint64_t *ashift);
34dc7c2f	66	typedef void vdev_close_func_t(vdev_t *vd);
34dc7c2f	67	typedef uint64_t vdev_asize_func_t(vdev_t *vd, uint64_t psize);
e10b0808	68	typedef void vdev_io_start_func_t(zio_t *zio);
b128c09f	69	typedef void vdev_io_done_func_t(zio_t *zio);
34dc7c2f	70	typedef void vdev_state_change_func_t(vdev_t *vd, int, int);
cae5b340	71	typedef boolean_t vdev_need_resilver_func_t(vdev_t *vd, uint64_t, size_t);
428870ff BB	72	typedef void vdev_hold_func_t(vdev_t *vd);
428870ff BB	73	typedef void vdev_rele_func_t(vdev_t *vd);
34dc7c2f	74
b01615d5	75	typedef const struct vdev_ops {
34dc7c2f BB	76	vdev_open_func_t *vdev_op_open;
34dc7c2f BB	77	vdev_close_func_t *vdev_op_close;
34dc7c2f BB	78	vdev_asize_func_t *vdev_op_asize;
	79	vdev_io_start_func_t *vdev_op_io_start;
	80	vdev_io_done_func_t *vdev_op_io_done;
	81	vdev_state_change_func_t *vdev_op_state_change;
cae5b340	82	vdev_need_resilver_func_t *vdev_op_need_resilver;
428870ff BB	83	vdev_hold_func_t *vdev_op_hold;
428870ff BB	84	vdev_rele_func_t *vdev_op_rele;
34dc7c2f BB	85	char vdev_op_type[16];
	86	boolean_t vdev_op_leaf;
	87	} vdev_ops_t;
	88
	89	/*
	90	* Virtual device properties
	91	*/
	92	struct vdev_cache_entry {
cae5b340	93	struct abd *ve_abd;
34dc7c2f	94	uint64_t ve_offset;
ea04106b	95	clock_t ve_lastused;
34dc7c2f BB	96	avl_node_t ve_offset_node;
	97	avl_node_t ve_lastused_node;
	98	uint32_t ve_hits;
	99	uint16_t ve_missed_update;
	100	zio_t *ve_fill_io;
	101	};
	102
	103	struct vdev_cache {
	104	avl_tree_t vc_offset_tree;
	105	avl_tree_t vc_lastused_tree;
	106	kmutex_t vc_lock;
	107	};
	108
a08ee875 LG	109	typedef struct vdev_queue_class {
	110	uint32_t vqc_active;
	111
	112	/*
	113	* Sorted by offset or timestamp, depending on if the queue is
	114	* LBA-ordered vs FIFO.
	115	*/
	116	avl_tree_t vqc_queued_tree;
	117	} vdev_queue_class_t;
	118
34dc7c2f	119	struct vdev_queue {
a08ee875 LG	120	vdev_t *vq_vdev;
	121	vdev_queue_class_t vq_class[ZIO_PRIORITY_NUM_QUEUEABLE];
	122	avl_tree_t vq_active_tree;
e10b0808 AX	123	avl_tree_t vq_read_offset_tree;
e10b0808 AX	124	avl_tree_t vq_write_offset_tree;
a08ee875 LG	125	uint64_t vq_last_offset;
a08ee875 LG	126	hrtime_t vq_io_complete_ts; /* time last i/o completed */
c06d4368	127	hrtime_t vq_io_delta_ts;
ea04106b	128	zio_t vq_io_search; /* used as local for stack reduction */
34dc7c2f BB	129	kmutex_t vq_lock;
	130	};
	131
	132	/*
	133	* Virtual device descriptor
	134	*/
	135	struct vdev {
	136	/*
	137	* Common to all vdev types.
	138	*/
	139	uint64_t vdev_id; /* child number in vdev parent */
	140	uint64_t vdev_guid; /* unique ID for this vdev */
	141	uint64_t vdev_guid_sum; /* self guid + all child guids */
428870ff	142	uint64_t vdev_orig_guid; /* orig. guid prior to remove */
34dc7c2f	143	uint64_t vdev_asize; /* allocatable device capacity */
9babb374	144	uint64_t vdev_min_asize; /* min acceptable asize */
1bd201e7	145	uint64_t vdev_max_asize; /* max acceptable asize */
34dc7c2f BB	146	uint64_t vdev_ashift; /* block alignment shift */
	147	uint64_t vdev_state; /* see VDEV_STATE_* #defines */
	148	uint64_t vdev_prevstate; /* used when reopening a vdev */
	149	vdev_ops_t vdev_ops; / vdev operations */
	150	spa_t vdev_spa; / spa for this vdev */
	151	void vdev_tsd; / type-specific data */
428870ff BB	152	vnode_t vdev_name_vp; / vnode for pathname */
428870ff BB	153	vnode_t vdev_devid_vp; / vnode for devid */
34dc7c2f BB	154	vdev_t vdev_top; / top-level vdev */
	155	vdev_t vdev_parent; / parent vdev */
	156	vdev_t *vdev_child; / array of children */
	157	uint64_t vdev_children; /* number of children */
34dc7c2f	158	vdev_stat_t vdev_stat; /* virtual device statistics */
cae5b340	159	vdev_stat_ex_t vdev_stat_ex; /* extended statistics */
9babb374	160	boolean_t vdev_expanding; /* expand the vdev? */
428870ff	161	boolean_t vdev_reopening; /* reopen in progress? */
e10b0808	162	boolean_t vdev_nonrot; /* true if solid state */
45d1cae3 BB	163	int vdev_open_error; /* error on last open */
45d1cae3 BB	164	kthread_t vdev_open_thread; / thread opening children */
428870ff	165	uint64_t vdev_crtxg; /* txg when top-level was added */
34dc7c2f BB	166
	167	/*
	168	* Top-level vdev state.
	169	*/
	170	uint64_t vdev_ms_array; /* metaslab array object */
	171	uint64_t vdev_ms_shift; /* metaslab size shift */
	172	uint64_t vdev_ms_count; /* number of metaslabs */
	173	metaslab_group_t vdev_mg; / metaslab group */
	174	metaslab_t *vdev_ms; / metaslab array */
920dd524	175	uint64_t vdev_pending_fastwrite; /* allocated fastwrites */
34dc7c2f BB	176	txg_list_t vdev_ms_list; /* per-txg dirty metaslab lists */
	177	txg_list_t vdev_dtl_list; /* per-txg dirty DTL lists */
	178	txg_node_t vdev_txg_node; /* per-txg dirty vdev linkage */
	179	boolean_t vdev_remove_wanted; /* async remove wanted? */
b128c09f BB	180	boolean_t vdev_probe_wanted; /* async probe wanted? */
	181	list_node_t vdev_config_dirty_node; /* config dirty list */
	182	list_node_t vdev_state_dirty_node; /* state dirty list */
34dc7c2f BB	183	uint64_t vdev_deflate_ratio; /* deflation ratio (x512) */
34dc7c2f BB	184	uint64_t vdev_islog; /* is an intent log device */
ea04106b	185	uint64_t vdev_removing; /* device is being removed? */
cae5b340 AX	186	boolean_t vdev_ishole; /* is a hole in the namespace */
	187	kmutex_t vdev_queue_lock; /* protects vdev_queue_depth */
	188	uint64_t vdev_top_zap;
	189
	190	/*
	191	* The queue depth parameters determine how many async writes are
	192	* still pending (i.e. allocated by net yet issued to disk) per
	193	* top-level (vdev_async_write_queue_depth) and the maximum allowed
	194	* (vdev_max_async_write_queue_depth). These values only apply to
	195	* top-level vdevs.
	196	*/
	197	uint64_t vdev_async_write_queue_depth;
	198	uint64_t vdev_max_async_write_queue_depth;
34dc7c2f BB	199
	200	/*
	201	* Leaf vdev state.
	202	*/
ea04106b AX	203	range_tree_t vdev_dtl[DTL_TYPES]; / dirty time logs */
ea04106b AX	204	space_map_t vdev_dtl_sm; / dirty time log space map */
34dc7c2f	205	txg_node_t vdev_dtl_node; /* per-txg dirty DTL linkage */
ea04106b AX	206	uint64_t vdev_dtl_object; /* DTL object */
ea04106b AX	207	uint64_t vdev_psize; /* physical device capacity */
34dc7c2f BB	208	uint64_t vdev_wholedisk; /* true if this is a whole disk */
	209	uint64_t vdev_offline; /* persistent offline state */
	210	uint64_t vdev_faulted; /* persistent faulted state */
	211	uint64_t vdev_degraded; /* persistent degraded state */
	212	uint64_t vdev_removed; /* persistent removed state */
a08ee875	213	uint64_t vdev_resilver_txg; /* persistent resilvering state */
34dc7c2f BB	214	uint64_t vdev_nparity; /* number of parity devices for raidz */
	215	char vdev_path; / vdev path (if any) */
	216	char vdev_devid; / vdev devid (if any) */
	217	char vdev_physpath; / vdev device path (if any) */
cae5b340	218	char vdev_enc_sysfs_path; / enclosure sysfs path */
9babb374	219	char vdev_fru; / physical FRU location */
b128c09f BB	220	uint64_t vdev_not_present; /* not present during import */
b128c09f BB	221	uint64_t vdev_unspare; /* unspare when resilvering done */
b128c09f BB	222	boolean_t vdev_nowritecache; /* true if flushwritecache failed */
	223	boolean_t vdev_checkremove; /* temporary online test */
	224	boolean_t vdev_forcefault; /* force online fault */
428870ff BB	225	boolean_t vdev_splitting; /* split or repair in progress */
428870ff BB	226	boolean_t vdev_delayed_close; /* delayed device close? */
ea04106b AX	227	boolean_t vdev_tmpoffline; /* device taken offline temporarily? */
	228	boolean_t vdev_detached; /* device detached? */
	229	boolean_t vdev_cant_read; /* vdev is failing all reads */
	230	boolean_t vdev_cant_write; /* vdev is failing all writes */
	231	boolean_t vdev_isspare; /* was a hot spare */
	232	boolean_t vdev_isl2cache; /* was a l2cache device */
cae5b340	233	boolean_t vdev_copy_uberblocks; /* post expand copy uberblocks */
34dc7c2f BB	234	vdev_queue_t vdev_queue; /* I/O deadline schedule queue */
34dc7c2f BB	235	vdev_cache_t vdev_cache; /* physical block cache */
e10b0808	236	spa_aux_vdev_t vdev_aux; / for l2cache and spares vdevs */
b128c09f	237	zio_t vdev_probe_zio; / root of current probe */
428870ff	238	vdev_aux_t vdev_label_aux; /* on-disk aux state */
cae5b340 AX	239	uint64_t vdev_leaf_zap;
cae5b340 AX	240	hrtime_t vdev_mmp_pending; /* 0 if write finished */
34dc7c2f BB	241
	242	/*
	243	* For DTrace to work in userland (libzpool) context, these fields must
	244	* remain at the end of the structure. DTrace will use the kernel's
	245	* CTF definition for 'struct vdev', and since the size of a kmutex_t is
9ae529ec	246	* larger in userland, the offsets for the rest of the fields would be
34dc7c2f BB	247	* incorrect.
	248	*/
	249	kmutex_t vdev_dtl_lock; /* vdev_dtl_{map,resilver} */
	250	kmutex_t vdev_stat_lock; /* vdev_stat */
b128c09f	251	kmutex_t vdev_probe_lock; /* protects vdev_probe_zio */
cae5b340 AX	252
	253	/*
	254	* We rate limit ZIO delay and ZIO checksum events, since they
	255	* can flood ZED with tons of events when a drive is acting up.
	256	*/
	257	#define DELAYS_PER_SECOND 5
	258	#define CHECKSUMS_PER_SECOND 5
	259	zfs_ratelimit_t vdev_delay_rl;
	260	zfs_ratelimit_t vdev_checksum_rl;
34dc7c2f BB	261	};
34dc7c2f BB	262
428870ff BB	263	#define VDEV_RAIDZ_MAXPARITY 3
428870ff BB	264
9babb374 BB	265	#define VDEV_PAD_SIZE (8 << 10)
	266	/* 2 padding areas (vl_pad1 and vl_pad2) to skip */
	267	#define VDEV_SKIP_SIZE VDEV_PAD_SIZE * 2
34dc7c2f BB	268	#define VDEV_PHYS_SIZE (112 << 10)
	269	#define VDEV_UBERBLOCK_RING (128 << 10)
	270
cae5b340 AX	271	/*
	272	* MMP blocks occupy the last MMP_BLOCKS_PER_LABEL slots in the uberblock
	273	* ring when MMP is enabled.
	274	*/
	275	#define MMP_BLOCKS_PER_LABEL 1
	276
ea04106b AX	277	/* The largest uberblock we support is 8k. */
ea04106b AX	278	#define MAX_UBERBLOCK_SHIFT (13)
34dc7c2f	279	#define VDEV_UBERBLOCK_SHIFT(vd) \
ea04106b AX	280	MIN(MAX((vd)->vdev_top->vdev_ashift, UBERBLOCK_SHIFT), \
ea04106b AX	281	MAX_UBERBLOCK_SHIFT)
34dc7c2f BB	282	#define VDEV_UBERBLOCK_COUNT(vd) \
	283	(VDEV_UBERBLOCK_RING >> VDEV_UBERBLOCK_SHIFT(vd))
	284	#define VDEV_UBERBLOCK_OFFSET(vd, n) \
	285	offsetof(vdev_label_t, vl_uberblock[(n) << VDEV_UBERBLOCK_SHIFT(vd)])
	286	#define VDEV_UBERBLOCK_SIZE(vd) (1ULL << VDEV_UBERBLOCK_SHIFT(vd))
	287
34dc7c2f	288	typedef struct vdev_phys {
428870ff BB	289	char vp_nvlist[VDEV_PHYS_SIZE - sizeof (zio_eck_t)];
428870ff BB	290	zio_eck_t vp_zbt;
34dc7c2f BB	291	} vdev_phys_t;
	292
	293	typedef struct vdev_label {
9babb374 BB	294	char vl_pad1[VDEV_PAD_SIZE]; /* 8K */
9babb374 BB	295	char vl_pad2[VDEV_PAD_SIZE]; /* 8K */
34dc7c2f BB	296	vdev_phys_t vl_vdev_phys; /* 112K */
	297	char vl_uberblock[VDEV_UBERBLOCK_RING]; /* 128K */
	298	} vdev_label_t; /* 256K total */
	299
	300	/*
	301	* vdev_dirty() flags
	302	*/
	303	#define VDD_METASLAB 0x01
	304	#define VDD_DTL 0x02
	305
a08ee875 LG	306	/* Offset of embedded boot loader region on each label */
a08ee875 LG	307	#define VDEV_BOOT_OFFSET (2 * sizeof (vdev_label_t))
34dc7c2f	308	/*
a08ee875	309	* Size of embedded boot loader region on each label.
34dc7c2f BB	310	* The total size of the first two labels plus the boot area is 4MB.
34dc7c2f BB	311	*/
a08ee875	312	#define VDEV_BOOT_SIZE (7ULL << 19) /* 3.5M */
34dc7c2f BB	313
	314	/*
	315	* Size of label regions at the start and end of each leaf device.
	316	*/
	317	#define VDEV_LABEL_START_SIZE (2 * sizeof (vdev_label_t) + VDEV_BOOT_SIZE)
	318	#define VDEV_LABEL_END_SIZE (2 * sizeof (vdev_label_t))
	319	#define VDEV_LABELS 4
9ae529ec	320	#define VDEV_BEST_LABEL VDEV_LABELS
34dc7c2f BB	321
	322	#define VDEV_ALLOC_LOAD 0
	323	#define VDEV_ALLOC_ADD 1
	324	#define VDEV_ALLOC_SPARE 2
	325	#define VDEV_ALLOC_L2CACHE 3
9babb374	326	#define VDEV_ALLOC_ROOTPOOL 4
428870ff	327	#define VDEV_ALLOC_SPLIT 5
5ffb9d1d	328	#define VDEV_ALLOC_ATTACH 6
34dc7c2f BB	329
	330	/*
	331	* Allocate or free a vdev
	332	*/
428870ff BB	333	extern vdev_t vdev_alloc_common(spa_t spa, uint_t id, uint64_t guid,
428870ff BB	334	vdev_ops_t *ops);
34dc7c2f BB	335	extern int vdev_alloc(spa_t spa, vdev_t vdp, nvlist_t config,
	336	vdev_t *parent, uint_t id, int alloctype);
	337	extern void vdev_free(vdev_t *vd);
	338
	339	/*
	340	* Add or remove children and parents
	341	*/
	342	extern void vdev_add_child(vdev_t pvd, vdev_t cvd);
	343	extern void vdev_remove_child(vdev_t pvd, vdev_t cvd);
	344	extern void vdev_compact_children(vdev_t *pvd);
	345	extern vdev_t vdev_add_parent(vdev_t cvd, vdev_ops_t *ops);
	346	extern void vdev_remove_parent(vdev_t *cvd);
	347
	348	/*
	349	* vdev sync load and sync
	350	*/
428870ff	351	extern void vdev_load_log_state(vdev_t nvd, vdev_t ovd);
572e2857	352	extern boolean_t vdev_log_state_valid(vdev_t *vd);
34dc7c2f	353	extern void vdev_load(vdev_t *vd);
ea04106b	354	extern int vdev_dtl_load(vdev_t *vd);
34dc7c2f BB	355	extern void vdev_sync(vdev_t *vd, uint64_t txg);
	356	extern void vdev_sync_done(vdev_t *vd, uint64_t txg);
	357	extern void vdev_dirty(vdev_t vd, int flags, void arg, uint64_t txg);
ea04106b	358	extern void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg);
34dc7c2f BB	359
	360	/*
	361	* Available vdev types.
	362	*/
	363	extern vdev_ops_t vdev_root_ops;
	364	extern vdev_ops_t vdev_mirror_ops;
	365	extern vdev_ops_t vdev_replacing_ops;
	366	extern vdev_ops_t vdev_raidz_ops;
	367	extern vdev_ops_t vdev_disk_ops;
	368	extern vdev_ops_t vdev_file_ops;
	369	extern vdev_ops_t vdev_missing_ops;
428870ff	370	extern vdev_ops_t vdev_hole_ops;
34dc7c2f BB	371	extern vdev_ops_t vdev_spare_ops;
	372
	373	/*
	374	* Common size functions
	375	*/
	376	extern uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize);
9babb374 BB	377	extern uint64_t vdev_get_min_asize(vdev_t *vd);
9babb374 BB	378	extern void vdev_set_min_asize(vdev_t *vd);
34dc7c2f BB	379
34dc7c2f BB	380	/*
a08ee875	381	* Global variables
34dc7c2f	382	*/
a08ee875	383	/* zdb uses this tunable, so it must be declared here to make lint happy. */
34dc7c2f BB	384	extern int zfs_vdev_cache_size;
	385
	386	#ifdef __cplusplus
	387	}
	388	#endif
	389
	390	#endif /* _SYS_VDEV_IMPL_H */