sys/vdev_impl.h

Go to the documentation of this file.

/*
 * 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) 2012 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>

#ifdef  __cplusplus
extern "C" {
#endif

/*
 * 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;

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 int     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 void    vdev_hold_func_t(vdev_t *vd);
typedef void    vdev_rele_func_t(vdev_t *vd);
/* \} */

typedef 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_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;

struct vdev_cache_entry {
        char            *ve_data;
        uint64_t        ve_offset;
        uint64_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;
};

struct vdev_queue {
        avl_tree_t      vq_deadline_tree;
        avl_tree_t      vq_read_tree;
        avl_tree_t      vq_write_tree;
        avl_tree_t      vq_pending_tree;
        kmutex_t        vq_lock;
};

struct vdev {
        uint64_t        vdev_id;        
        uint64_t        vdev_guid;      
        uint64_t        vdev_guid_sum;  
        uint64_t        vdev_orig_guid; 
        uint64_t        vdev_asize;     
        uint64_t        vdev_min_asize; 
        uint64_t        vdev_max_asize; 
        uint64_t        vdev_ashift;    
        uint64_t        vdev_state;     
        uint64_t        vdev_prevstate; 
        vdev_ops_t      *vdev_ops;      
        spa_t           *vdev_spa;      
        void            *vdev_tsd;      
        vnode_t         *vdev_name_vp;  
        vnode_t         *vdev_devid_vp; 
        vdev_t          *vdev_top;      
        vdev_t          *vdev_parent;   
        vdev_t          **vdev_child;   
        uint64_t        vdev_children;  
        space_map_t     vdev_dtl[DTL_TYPES]; 
        vdev_stat_t     vdev_stat;      
        boolean_t       vdev_expanding; 
        boolean_t       vdev_reopening; 
        int             vdev_open_error; 
        kthread_t       *vdev_open_thread; 
        uint64_t        vdev_crtxg;     
        /* \} */

        uint64_t        vdev_ms_array;  
        uint64_t        vdev_ms_shift;  
        uint64_t        vdev_ms_count;  
        metaslab_group_t *vdev_mg;      
        metaslab_t      **vdev_ms;      
        txg_list_t      vdev_ms_list;   
        txg_list_t      vdev_dtl_list;  
        txg_node_t      vdev_txg_node;  
        boolean_t       vdev_remove_wanted; 
        boolean_t       vdev_probe_wanted; 
        uint64_t        vdev_removing;  
        list_node_t     vdev_config_dirty_node; 
        list_node_t     vdev_state_dirty_node; 
        uint64_t        vdev_deflate_ratio; 
        uint64_t        vdev_islog;     
        uint64_t        vdev_ishole;    
        /* \} */

        uint64_t        vdev_psize;     
        space_map_obj_t vdev_dtl_smo;   
        txg_node_t      vdev_dtl_node;  
        uint64_t        vdev_wholedisk; 
        uint64_t        vdev_offline;   
        uint64_t        vdev_faulted;   
        uint64_t        vdev_degraded;  
        uint64_t        vdev_removed;   
        uint64_t        vdev_resilvering; 
        uint64_t        vdev_nparity;   
        char            *vdev_path;     
        char            *vdev_devid;    
        char            *vdev_physpath; 
        char            *vdev_fru;      
        uint64_t        vdev_not_present; 
        uint64_t        vdev_unspare;   
        hrtime_t        vdev_last_try;  
        boolean_t       vdev_nowritecache; 
        boolean_t       vdev_notrim;    
        boolean_t       vdev_checkremove; 
        boolean_t       vdev_forcefault; 
        boolean_t       vdev_splitting; 
        boolean_t       vdev_delayed_close; 
        uint8_t         vdev_tmpoffline; 
        uint8_t         vdev_detached;  
        uint8_t         vdev_cant_read; 
        uint8_t         vdev_cant_write; 
        uint64_t        vdev_isspare;   
        uint64_t        vdev_isl2cache; 
        vdev_queue_t    vdev_queue;     
        vdev_cache_t    vdev_cache;     
        spa_aux_vdev_t  *vdev_aux;      
        zio_t           *vdev_probe_zio; 
        vdev_aux_t      vdev_label_aux; 
        struct trim_map *vdev_trimmap;
        /* \} */

        /*
         * 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;  
        kmutex_t        vdev_stat_lock; 
        kmutex_t        vdev_probe_lock; 
};

#define VDEV_RAIDZ_MAXPARITY    3

#define VDEV_PAD_SIZE           (8 << 10)

#define VDEV_SKIP_SIZE          VDEV_PAD_SIZE * 2
#define VDEV_PHYS_SIZE          (112 << 10)
#define VDEV_UBERBLOCK_RING     (128 << 10)

#define VDEV_UBERBLOCK_SHIFT(vd)        \
        MAX((vd)->vdev_top->vdev_ashift, 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];                 
        char            vl_pad2[VDEV_PAD_SIZE];                 
        vdev_phys_t     vl_vdev_phys;                           
        char            vl_uberblock[VDEV_UBERBLOCK_RING];      
} vdev_label_t;

#define VDD_METASLAB    0x01
#define VDD_DTL         0x02

#define VDEV_BOOT_OFFSET        (2 * sizeof (vdev_label_t))

#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 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);

/*
 * 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;
#ifdef _KERNEL
extern vdev_ops_t vdev_geom_ops;
#else
extern vdev_ops_t vdev_disk_ops;
#endif
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);

/*
 * 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 */