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 #ifndef _SYS_ZAP_LEAF_H 27 #define _SYS_ZAP_LEAF_H 28 29 #ifdef __cplusplus 30 extern "C" { 31 #endif 32 33 struct zap; 34 35 #define ZAP_LEAF_MAGIC 0x2AB1EAF 36 37 /* chunk size = 24 bytes */ 38 #define ZAP_LEAF_CHUNKSIZE 24 39 40 /* 41 * The amount of space available for chunks is: 42 * block size (1<<l->l_bs) - hash entry size (2) * number of hash 43 * entries - header space (2*chunksize) 44 */ 45 #define ZAP_LEAF_NUMCHUNKS(l) \ 46 (((1<<(l)->l_bs) - 2*ZAP_LEAF_HASH_NUMENTRIES(l)) / \ 47 ZAP_LEAF_CHUNKSIZE - 2) 48 49 /* 50 * The amount of space within the chunk available for the array is: 51 * chunk size - space for type (1) - space for next pointer (2) 52 */ 53 #define ZAP_LEAF_ARRAY_BYTES (ZAP_LEAF_CHUNKSIZE - 3) 54 55 #define ZAP_LEAF_ARRAY_NCHUNKS(bytes) \ 56 (((bytes)+ZAP_LEAF_ARRAY_BYTES-1)/ZAP_LEAF_ARRAY_BYTES) 57 58 /* 59 * Low water mark: when there are only this many chunks free, start 60 * growing the ptrtbl. Ideally, this should be larger than a 61 * "reasonably-sized" entry. 20 chunks is more than enough for the 62 * largest directory entry (MAXNAMELEN (256) byte name, 8-byte value), 63 * while still being only around 3% for 16k blocks. 64 */ 65 #define ZAP_LEAF_LOW_WATER (20) 66 67 /* 68 * The leaf hash table has block size / 2^5 (32) number of entries, 69 * which should be more than enough for the maximum number of entries, 70 * which is less than block size / CHUNKSIZE (24) / minimum number of 71 * chunks per entry (3). 72 */ 73 #define ZAP_LEAF_HASH_SHIFT(l) ((l)->l_bs - 5) 74 #define ZAP_LEAF_HASH_NUMENTRIES(l) (1 << ZAP_LEAF_HASH_SHIFT(l)) 75 76 /* 77 * The chunks start immediately after the hash table. The end of the 78 * hash table is at l_hash + HASH_NUMENTRIES, which we simply cast to a 79 * chunk_t. 80 */ 81 #define ZAP_LEAF_CHUNK(l, idx) \ 82 ((zap_leaf_chunk_t *) \ 83 ((l)->l_phys->l_hash + ZAP_LEAF_HASH_NUMENTRIES(l)))[idx] 84 #define ZAP_LEAF_ENTRY(l, idx) (&ZAP_LEAF_CHUNK(l, idx).l_entry) 85 86 typedef enum zap_chunk_type { 87 ZAP_CHUNK_FREE = 253, 88 ZAP_CHUNK_ENTRY = 252, 89 ZAP_CHUNK_ARRAY = 251, 90 ZAP_CHUNK_TYPE_MAX = 250 91 } zap_chunk_type_t; 92 93 #define ZLF_ENTRIES_CDSORTED (1<<0) 94 95 /* 96 * TAKE NOTE: 97 * If zap_leaf_phys_t is modified, zap_leaf_byteswap() must be modified. 98 */ 99 typedef struct zap_leaf_phys { 100 struct zap_leaf_header { 101 uint64_t lh_block_type; /* ZBT_LEAF */ 102 uint64_t lh_pad1; 103 uint64_t lh_prefix; /* hash prefix of this leaf */ 104 uint32_t lh_magic; /* ZAP_LEAF_MAGIC */ 105 uint16_t lh_nfree; /* number free chunks */ 106 uint16_t lh_nentries; /* number of entries */ 107 uint16_t lh_prefix_len; /* num bits used to id this */ 108 109 /* above is accessable to zap, below is zap_leaf private */ 110 111 uint16_t lh_freelist; /* chunk head of free list */ 112 uint8_t lh_flags; /* ZLF_* flags */ 113 uint8_t lh_pad2[11]; 114 } l_hdr; /* 2 24-byte chunks */ 115 116 /* 117 * The header is followed by a hash table with 118 * ZAP_LEAF_HASH_NUMENTRIES(zap) entries. The hash table is 119 * followed by an array of ZAP_LEAF_NUMCHUNKS(zap) 120 * zap_leaf_chunk structures. These structures are accessed 121 * with the ZAP_LEAF_CHUNK() macro. 122 */ 123 124 uint16_t l_hash[1]; 125 } zap_leaf_phys_t; 126 127 typedef union zap_leaf_chunk { 128 struct zap_leaf_entry { 129 uint8_t le_type; /* always ZAP_CHUNK_ENTRY */ 130 uint8_t le_int_size; /* size of ints */ 131 uint16_t le_next; /* next entry in hash chain */ 132 uint16_t le_name_chunk; /* first chunk of the name */ 133 uint16_t le_name_length; /* ints in name (incl null) */ 134 uint16_t le_value_chunk; /* first chunk of the value */ 135 uint16_t le_value_length; /* value length in ints */ 136 uint32_t le_cd; /* collision differentiator */ 137 uint64_t le_hash; /* hash value of the name */ 138 } l_entry; 139 struct zap_leaf_array { 140 uint8_t la_type; /* always ZAP_CHUNK_ARRAY */ 141 uint8_t la_array[ZAP_LEAF_ARRAY_BYTES]; 142 uint16_t la_next; /* next blk or CHAIN_END */ 143 } l_array; 144 struct zap_leaf_free { 145 uint8_t lf_type; /* always ZAP_CHUNK_FREE */ 146 uint8_t lf_pad[ZAP_LEAF_ARRAY_BYTES]; 147 uint16_t lf_next; /* next in free list, or CHAIN_END */ 148 } l_free; 149 } zap_leaf_chunk_t; 150 151 typedef struct zap_leaf { 152 krwlock_t l_rwlock; 153 uint64_t l_blkid; /* 1<<ZAP_BLOCK_SHIFT byte block off */ 154 int l_bs; /* block size shift */ 155 dmu_buf_t *l_dbuf; 156 zap_leaf_phys_t *l_phys; 157 } zap_leaf_t; 158 159 160 typedef struct zap_entry_handle { 161 /* below is set by zap_leaf.c and is public to zap.c */ 162 uint64_t zeh_num_integers; 163 uint64_t zeh_hash; 164 uint32_t zeh_cd; 165 uint8_t zeh_integer_size; 166 167 /* below is private to zap_leaf.c */ 168 uint16_t zeh_fakechunk; 169 uint16_t *zeh_chunkp; 170 zap_leaf_t *zeh_leaf; 171 } zap_entry_handle_t; 172 173 /* 174 * Return a handle to the named entry, or ENOENT if not found. The hash 175 * value must equal zap_hash(name). 176 */ 177 extern int zap_leaf_lookup(zap_leaf_t *l, 178 zap_name_t *zn, zap_entry_handle_t *zeh); 179 180 /* 181 * Return a handle to the entry with this hash+cd, or the entry with the 182 * next closest hash+cd. 183 */ 184 extern int zap_leaf_lookup_closest(zap_leaf_t *l, 185 uint64_t hash, uint32_t cd, zap_entry_handle_t *zeh); 186 187 /* 188 * Read the first num_integers in the attribute. Integer size 189 * conversion will be done without sign extension. Return EINVAL if 190 * integer_size is too small. Return EOVERFLOW if there are more than 191 * num_integers in the attribute. 192 */ 193 extern int zap_entry_read(const zap_entry_handle_t *zeh, 194 uint8_t integer_size, uint64_t num_integers, void *buf); 195 196 extern int zap_entry_read_name(zap_t *zap, const zap_entry_handle_t *zeh, 197 uint16_t buflen, char *buf); 198 199 /* 200 * Replace the value of an existing entry. 201 * 202 * zap_entry_update may fail if it runs out of space (ENOSPC). 203 */ 204 extern int zap_entry_update(zap_entry_handle_t *zeh, 205 uint8_t integer_size, uint64_t num_integers, const void *buf); 206 207 /* 208 * Remove an entry. 209 */ 210 extern void zap_entry_remove(zap_entry_handle_t *zeh); 211 212 /* 213 * Create an entry. An equal entry must not exist, and this entry must 214 * belong in this leaf (according to its hash value). Fills in the 215 * entry handle on success. Returns 0 on success or ENOSPC on failure. 216 */ 217 extern int zap_entry_create(zap_leaf_t *l, zap_name_t *zn, uint32_t cd, 218 uint8_t integer_size, uint64_t num_integers, const void *buf, 219 zap_entry_handle_t *zeh); 220 221 /* 222 * Return true if there are additional entries with the same normalized 223 * form. 224 */ 225 extern boolean_t zap_entry_normalization_conflict(zap_entry_handle_t *zeh, 226 zap_name_t *zn, const char *name, zap_t *zap); 227 228 /* 229 * Other stuff. 230 */ 231 232 extern void zap_leaf_init(zap_leaf_t *l, boolean_t sort); 233 extern void zap_leaf_byteswap(zap_leaf_phys_t *buf, int len); 234 extern void zap_leaf_split(zap_leaf_t *l, zap_leaf_t *nl, boolean_t sort); 235 extern void zap_leaf_stats(zap_t *zap, zap_leaf_t *l, zap_stats_t *zs); 236 237 #ifdef __cplusplus 238 } 239 #endif 240 241 #endif /* _SYS_ZAP_LEAF_H */ 242