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 /*
  22  * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
  23  * Use is subject to license terms.
  24  */
  25 
  26 /*
  27  * Copyright (c) 2012, 2017 by Delphix. All rights reserved.
  28  */
  29 
  30 #ifndef _SYS_SPACE_MAP_H
  31 #define _SYS_SPACE_MAP_H
  32 
  33 #include <sys/avl.h>
  34 #include <sys/range_tree.h>
  35 #include <sys/dmu.h>
  36 
  37 #ifdef  __cplusplus
  38 extern "C" {
  39 #endif
  40 
  41 /*
  42  * The size of the space map object has increased to include a histogram.
  43  * The SPACE_MAP_SIZE_V0 designates the original size and is used to
  44  * maintain backward compatibility.
  45  */
  46 #define SPACE_MAP_SIZE_V0       (3 * sizeof (uint64_t))
  47 #define SPACE_MAP_HISTOGRAM_SIZE        32
  48 
  49 /*
  50  * The space_map_phys is the on-disk representation of the space map.
  51  * Consumers of space maps should never reference any of the members of this
  52  * structure directly. These members may only be updated in syncing context.
  53  *
  54  * Note the smp_object is no longer used but remains in the structure
  55  * for backward compatibility.
  56  */
  57 typedef struct space_map_phys {
  58         uint64_t        smp_object;     /* on-disk space map object */
  59         uint64_t        smp_objsize;    /* size of the object */
  60         int64_t         smp_alloc;      /* space allocated from the map */
  61         uint64_t        smp_pad[5];     /* reserved */
  62 
  63         /*
  64          * The smp_histogram maintains a histogram of free regions. Each
  65          * bucket, smp_histogram[i], contains the number of free regions
  66          * whose size is:
  67          * 2^(i+sm_shift) <= size of free region in bytes < 2^(i+sm_shift+1)
  68          */
  69         uint64_t        smp_histogram[SPACE_MAP_HISTOGRAM_SIZE];
  70 } space_map_phys_t;
  71 
  72 /*
  73  * The space map object defines a region of space, its size, how much is
  74  * allocated, and the on-disk object that stores this information.
  75  * Consumers of space maps may only access the members of this structure.
  76  *
  77  * Note: the space_map may not be accessed concurrently; consumers
  78  * must provide external locking if required.
  79  */
  80 typedef struct space_map {
  81         uint64_t        sm_start;       /* start of map */
  82         uint64_t        sm_size;        /* size of map */
  83         uint8_t         sm_shift;       /* unit shift */
  84         uint64_t        sm_length;      /* synced length */
  85         int64_t         sm_alloc;       /* synced space allocated */
  86         objset_t        *sm_os;         /* objset for this map */
  87         uint64_t        sm_object;      /* object id for this map */
  88         uint32_t        sm_blksz;       /* block size for space map */
  89         dmu_buf_t       *sm_dbuf;       /* space_map_phys_t dbuf */
  90         space_map_phys_t *sm_phys;      /* on-disk space map */
  91 } space_map_t;
  92 
  93 /*
  94  * debug entry
  95  *
  96  *     2     2        10                     50
  97  *  +-----+-----+------------+----------------------------------+
  98  *  | 1 0 | act |  syncpass  |        txg (lower bits)          |
  99  *  +-----+-----+------------+----------------------------------+
 100  *   63 62 61 60 59        50 49                                0
 101  *
 102  *
 103  * one-word entry
 104  *
 105  *    1               47                   1           15
 106  *  +-----------------------------------------------------------+
 107  *  | 0 |   offset (sm_shift units)    | type |       run       |
 108  *  +-----------------------------------------------------------+
 109  *   63  62                          16   15   14               0
 110  *
 111  *
 112  * two-word entry
 113  *
 114  *     2     2               36                      24
 115  *  +-----+-----+---------------------------+-------------------+
 116  *  | 1 1 | pad |            run            |       vdev        |
 117  *  +-----+-----+---------------------------+-------------------+
 118  *   63 62 61 60 59                       24 23                 0
 119  *
 120  *     1                            63
 121  *  +------+----------------------------------------------------+
 122  *  | type |                      offset                        |
 123  *  +------+----------------------------------------------------+
 124  *     63   62                                                  0
 125  *
 126  * Note that a two-word entry will not strandle a block boundary.
 127  * If necessary, the last word of a block will be padded with a
 128  * debug entry (with act = syncpass = txg = 0).
 129  */
 130 
 131 typedef enum {
 132         SM_ALLOC,
 133         SM_FREE
 134 } maptype_t;
 135 
 136 typedef struct space_map_entry {
 137         maptype_t sme_type;
 138         uint32_t sme_vdev;      /* max is 2^24-1; SM_NO_VDEVID if not present */
 139         uint64_t sme_offset;    /* max is 2^63-1; units of sm_shift */
 140         uint64_t sme_run;       /* max is 2^36; units of sm_shift */
 141 } space_map_entry_t;
 142 
 143 #define SM_NO_VDEVID    (1 << SPA_VDEVBITS)
 144 
 145 /* one-word entry constants */
 146 #define SM_DEBUG_PREFIX 2
 147 #define SM_OFFSET_BITS  47
 148 #define SM_RUN_BITS     15
 149 
 150 /* two-word entry constants */
 151 #define SM2_PREFIX      3
 152 #define SM2_OFFSET_BITS 63
 153 #define SM2_RUN_BITS    36
 154 
 155 #define SM_PREFIX_DECODE(x)     BF64_DECODE(x, 62, 2)
 156 #define SM_PREFIX_ENCODE(x)     BF64_ENCODE(x, 62, 2)
 157 
 158 #define SM_DEBUG_ACTION_DECODE(x)       BF64_DECODE(x, 60, 2)
 159 #define SM_DEBUG_ACTION_ENCODE(x)       BF64_ENCODE(x, 60, 2)
 160 #define SM_DEBUG_SYNCPASS_DECODE(x)     BF64_DECODE(x, 50, 10)
 161 #define SM_DEBUG_SYNCPASS_ENCODE(x)     BF64_ENCODE(x, 50, 10)
 162 #define SM_DEBUG_TXG_DECODE(x)          BF64_DECODE(x, 0, 50)
 163 #define SM_DEBUG_TXG_ENCODE(x)          BF64_ENCODE(x, 0, 50)
 164 
 165 #define SM_OFFSET_DECODE(x)     BF64_DECODE(x, 16, SM_OFFSET_BITS)
 166 #define SM_OFFSET_ENCODE(x)     BF64_ENCODE(x, 16, SM_OFFSET_BITS)
 167 #define SM_TYPE_DECODE(x)       BF64_DECODE(x, 15, 1)
 168 #define SM_TYPE_ENCODE(x)       BF64_ENCODE(x, 15, 1)
 169 #define SM_RUN_DECODE(x)        (BF64_DECODE(x, 0, SM_RUN_BITS) + 1)
 170 #define SM_RUN_ENCODE(x)        BF64_ENCODE((x) - 1, 0, SM_RUN_BITS)
 171 #define SM_RUN_MAX              SM_RUN_DECODE(~0ULL)
 172 #define SM_OFFSET_MAX           SM_OFFSET_DECODE(~0ULL)
 173 
 174 #define SM2_RUN_DECODE(x)       (BF64_DECODE(x, SPA_VDEVBITS, SM2_RUN_BITS) + 1)
 175 #define SM2_RUN_ENCODE(x)       BF64_ENCODE((x) - 1, SPA_VDEVBITS, SM2_RUN_BITS)
 176 #define SM2_VDEV_DECODE(x)      BF64_DECODE(x, 0, SPA_VDEVBITS)
 177 #define SM2_VDEV_ENCODE(x)      BF64_ENCODE(x, 0, SPA_VDEVBITS)
 178 #define SM2_TYPE_DECODE(x)      BF64_DECODE(x, SM2_OFFSET_BITS, 1)
 179 #define SM2_TYPE_ENCODE(x)      BF64_ENCODE(x, SM2_OFFSET_BITS, 1)
 180 #define SM2_OFFSET_DECODE(x)    BF64_DECODE(x, 0, SM2_OFFSET_BITS)
 181 #define SM2_OFFSET_ENCODE(x)    BF64_ENCODE(x, 0, SM2_OFFSET_BITS)
 182 #define SM2_RUN_MAX             SM2_RUN_DECODE(~0ULL)
 183 #define SM2_OFFSET_MAX          SM2_OFFSET_DECODE(~0ULL)
 184 
 185 boolean_t sm_entry_is_debug(uint64_t e);
 186 boolean_t sm_entry_is_single_word(uint64_t e);
 187 boolean_t sm_entry_is_double_word(uint64_t e);
 188 
 189 typedef int (*sm_cb_t)(space_map_entry_t *sme, void *arg);
 190 
 191 int space_map_load(space_map_t *sm, range_tree_t *rt, maptype_t maptype);
 192 int space_map_iterate(space_map_t *sm, sm_cb_t callback, void *arg);
 193 int space_map_incremental_destroy(space_map_t *sm, sm_cb_t callback, void *arg,
 194     dmu_tx_t *tx);
 195 
 196 void space_map_histogram_clear(space_map_t *sm);
 197 void space_map_histogram_add(space_map_t *sm, range_tree_t *rt,
 198     dmu_tx_t *tx);
 199 
 200 void space_map_update(space_map_t *sm);
 201 
 202 uint64_t space_map_object(space_map_t *sm);
 203 uint64_t space_map_allocated(space_map_t *sm);
 204 uint64_t space_map_length(space_map_t *sm);
 205 
 206 void space_map_write(space_map_t *sm, range_tree_t *rt, maptype_t maptype,
 207     uint64_t vdev_id, dmu_tx_t *tx);
 208 uint64_t space_map_estimate_optimal_size(space_map_t *sm, range_tree_t *rt,
 209     uint64_t vdev_id);
 210 void space_map_truncate(space_map_t *sm, int blocksize, dmu_tx_t *tx);
 211 uint64_t space_map_alloc(objset_t *os, int blocksize, dmu_tx_t *tx);
 212 void space_map_free(space_map_t *sm, dmu_tx_t *tx);
 213 void space_map_free_obj(objset_t *os, uint64_t smobj, dmu_tx_t *tx);
 214 
 215 int space_map_open(space_map_t **smp, objset_t *os, uint64_t object,
 216     uint64_t start, uint64_t size, uint8_t shift);
 217 void space_map_close(space_map_t *sm);
 218 
 219 int64_t space_map_alloc_delta(space_map_t *sm);
 220 
 221 #ifdef  __cplusplus
 222 }
 223 #endif
 224 
 225 #endif  /* _SYS_SPACE_MAP_H */