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 2007 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 26 #ifndef _SYS_FS_ZFS_ZNODE_H 27 #define _SYS_FS_ZFS_ZNODE_H 28 29 #pragma ident "%Z%%M% %I% %E% SMI" 30 31 #ifdef _KERNEL 32 #include <sys/isa_defs.h> 33 #include <sys/types32.h> 34 #include <sys/list.h> 35 #include <sys/dmu.h> 36 #include <sys/zfs_vfsops.h> 37 #endif 38 #include <sys/zfs_acl.h> 39 #include <sys/zil.h> 40 41 #ifdef __cplusplus 42 extern "C" { 43 #endif 44 45 /* 46 * Define special zfs pflags 47 */ 48 #define ZFS_XATTR 0x1 /* is an extended attribute */ 49 #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */ 50 #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */ 51 52 #define MASTER_NODE_OBJ 1 53 54 /* 55 * special attributes for master node. 56 */ 57 58 #define ZFS_FSID "FSID" 59 #define ZFS_UNLINKED_SET "DELETE_QUEUE" 60 #define ZFS_ROOT_OBJ "ROOT" 61 #define ZPL_VERSION_OBJ "VERSION" 62 #define ZFS_PROP_BLOCKPERPAGE "BLOCKPERPAGE" 63 #define ZFS_PROP_NOGROWBLOCKS "NOGROWBLOCKS" 64 65 #define ZFS_FLAG_BLOCKPERPAGE 0x1 66 #define ZFS_FLAG_NOGROWBLOCKS 0x2 67 68 /* 69 * ZPL version - rev'd whenever an incompatible on-disk format change 70 * occurs. Independent of SPA/DMU/ZAP versioning. 71 */ 72 73 #define ZPL_VERSION 1ULL 74 75 #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE) 76 77 /* Path component length */ 78 /* 79 * The generic fs code uses MAXNAMELEN to represent 80 * what the largest component length is. Unfortunately, 81 * this length includes the terminating NULL. ZFS needs 82 * to tell the users via pathconf() and statvfs() what the 83 * true maximum length of a component is, excluding the NULL. 84 */ 85 #define ZFS_MAXNAMELEN (MAXNAMELEN - 1) 86 87 /* 88 * The directory entry has the type (currently unused on Solaris) in the 89 * top 4 bits, and the object number in the low 48 bits. The "middle" 90 * 12 bits are unused. 91 */ 92 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4) 93 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48) 94 #define ZFS_DIRENT_MAKE(type, obj) (((uint64_t)type << 60) | obj) 95 96 97 /* 98 * This is the persistent portion of the znode. It is stored 99 * in the "bonus buffer" of the file. Short symbolic links 100 * are also stored in the bonus buffer. 101 */ 102 typedef struct znode_phys { 103 uint64_t zp_atime[2]; /* 0 - last file access time */ 104 uint64_t zp_mtime[2]; /* 16 - last file modification time */ 105 uint64_t zp_ctime[2]; /* 32 - last file change time */ 106 uint64_t zp_crtime[2]; /* 48 - creation time */ 107 uint64_t zp_gen; /* 64 - generation (txg of creation) */ 108 uint64_t zp_mode; /* 72 - file mode bits */ 109 uint64_t zp_size; /* 80 - size of file */ 110 uint64_t zp_parent; /* 88 - directory parent (`..') */ 111 uint64_t zp_links; /* 96 - number of links to file */ 112 uint64_t zp_xattr; /* 104 - DMU object for xattrs */ 113 uint64_t zp_rdev; /* 112 - dev_t for VBLK & VCHR files */ 114 uint64_t zp_flags; /* 120 - persistent flags */ 115 uint64_t zp_uid; /* 128 - file owner */ 116 uint64_t zp_gid; /* 136 - owning group */ 117 uint64_t zp_pad[4]; /* 144 - future */ 118 zfs_znode_acl_t zp_acl; /* 176 - 263 ACL */ 119 /* 120 * Data may pad out any remaining bytes in the znode buffer, eg: 121 * 122 * |<---------------------- dnode_phys (512) ------------------------>| 123 * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->| 124 * |<---- znode (264) ---->|<---- data (56) ---->| 125 * 126 * At present, we only use this space to store symbolic links. 127 */ 128 } znode_phys_t; 129 130 /* 131 * Directory entry locks control access to directory entries. 132 * They are used to protect creates, deletes, and renames. 133 * Each directory znode has a mutex and a list of locked names. 134 */ 135 #ifdef _KERNEL 136 typedef struct zfs_dirlock { 137 char *dl_name; /* directory entry being locked */ 138 uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */ 139 uint16_t dl_namesize; /* set if dl_name was allocated */ 140 kcondvar_t dl_cv; /* wait for entry to be unlocked */ 141 struct znode *dl_dzp; /* directory znode */ 142 struct zfs_dirlock *dl_next; /* next in z_dirlocks list */ 143 } zfs_dirlock_t; 144 145 typedef struct znode { 146 struct zfsvfs *z_zfsvfs; 147 vnode_t *z_vnode; 148 uint64_t z_id; /* object ID for this znode */ 149 kmutex_t z_lock; /* znode modification lock */ 150 krwlock_t z_map_lock; /* page map lock */ 151 krwlock_t z_parent_lock; /* parent lock for directories */ 152 krwlock_t z_name_lock; /* "master" lock for dirent locks */ 153 zfs_dirlock_t *z_dirlocks; /* directory entry lock list */ 154 kmutex_t z_range_lock; /* protects changes to z_range_avl */ 155 avl_tree_t z_range_avl; /* avl tree of file range locks */ 156 uint8_t z_unlinked; /* file has been unlinked */ 157 uint8_t z_atime_dirty; /* atime needs to be synced */ 158 uint8_t z_dbuf_held; /* Is z_dbuf already held? */ 159 uint8_t z_zn_prefetch; /* Prefetch znodes? */ 160 uint_t z_blksz; /* block size in bytes */ 161 uint_t z_seq; /* modification sequence number */ 162 uint64_t z_mapcnt; /* number of pages mapped to file */ 163 uint64_t z_last_itx; /* last ZIL itx on this znode */ 164 uint32_t z_sync_cnt; /* synchronous open count */ 165 kmutex_t z_acl_lock; /* acl data lock */ 166 list_node_t z_link_node; /* all znodes in fs link */ 167 /* 168 * These are dmu managed fields. 169 */ 170 znode_phys_t *z_phys; /* pointer to persistent znode */ 171 dmu_buf_t *z_dbuf; /* buffer containing the z_phys */ 172 } znode_t; 173 174 175 /* 176 * Range locking rules 177 * -------------------- 178 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole 179 * file range needs to be locked as RL_WRITER. Only then can the pages be 180 * freed etc and zp_size reset. zp_size must be set within range lock. 181 * 2. For writes and punching holes (zfs_write & zfs_space) just the range 182 * being written or freed needs to be locked as RL_WRITER. 183 * Multiple writes at the end of the file must coordinate zp_size updates 184 * to ensure data isn't lost. A compare and swap loop is currently used 185 * to ensure the file size is at least the offset last written. 186 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being 187 * read needs to be locked as RL_READER. A check against zp_size can then 188 * be made for reading beyond end of file. 189 */ 190 191 /* 192 * Convert between znode pointers and vnode pointers 193 */ 194 #define ZTOV(ZP) ((ZP)->z_vnode) 195 #define VTOZ(VP) ((znode_t *)(VP)->v_data) 196 197 /* 198 * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation. 199 * ZFS_EXIT() must be called before exitting the vop. 200 */ 201 #define ZFS_ENTER(zfsvfs) \ 202 { \ 203 atomic_add_32(&(zfsvfs)->z_op_cnt, 1); \ 204 if ((zfsvfs)->z_unmounted1) { \ 205 ZFS_EXIT(zfsvfs); \ 206 return (EIO); \ 207 } \ 208 } 209 #define ZFS_EXIT(zfsvfs) atomic_add_32(&(zfsvfs)->z_op_cnt, -1) 210 211 /* 212 * Macros for dealing with dmu_buf_hold 213 */ 214 #define ZFS_OBJ_HASH(obj_num) (obj_num & (ZFS_OBJ_MTX_SZ - 1)) 215 #define ZFS_OBJ_MUTEX(zp) \ 216 (&zp->z_zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(zp->z_id)]) 217 #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \ 218 mutex_enter(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]); 219 220 #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \ 221 mutex_exit(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]) 222 223 /* 224 * Macros to encode/decode ZFS stored time values from/to struct timespec 225 */ 226 #define ZFS_TIME_ENCODE(tp, stmp) \ 227 { \ 228 stmp[0] = (uint64_t)(tp)->tv_sec; \ 229 stmp[1] = (uint64_t)(tp)->tv_nsec; \ 230 } 231 232 #define ZFS_TIME_DECODE(tp, stmp) \ 233 { \ 234 (tp)->tv_sec = (time_t)stmp[0]; \ 235 (tp)->tv_nsec = (long)stmp[1]; \ 236 } 237 238 /* 239 * Timestamp defines 240 */ 241 #define ACCESSED (AT_ATIME) 242 #define STATE_CHANGED (AT_CTIME) 243 #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME) 244 245 #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \ 246 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \ 247 zfs_time_stamper(zp, ACCESSED, NULL) 248 249 extern int zfs_init_fs(zfsvfs_t *, znode_t **, cred_t *); 250 extern void zfs_set_dataprop(objset_t *); 251 extern void zfs_create_fs(objset_t *os, cred_t *cr, dmu_tx_t *tx); 252 extern void zfs_time_stamper(znode_t *, uint_t, dmu_tx_t *); 253 extern void zfs_time_stamper_locked(znode_t *, uint_t, dmu_tx_t *); 254 extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *); 255 extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t); 256 extern void zfs_znode_init(void); 257 extern void zfs_znode_fini(void); 258 extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **); 259 extern void zfs_zinactive(znode_t *); 260 extern void zfs_znode_delete(znode_t *, dmu_tx_t *); 261 extern void zfs_znode_free(znode_t *); 262 extern void zfs_remove_op_tables(); 263 extern int zfs_create_op_tables(); 264 extern int zfs_sync(vfs_t *vfsp, short flag, cred_t *cr); 265 extern dev_t zfs_cmpldev(uint64_t); 266 267 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, int txtype, 268 znode_t *dzp, znode_t *zp, char *name); 269 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, int txtype, 270 znode_t *dzp, char *name); 271 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, int txtype, 272 znode_t *dzp, znode_t *zp, char *name); 273 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, int txtype, 274 znode_t *dzp, znode_t *zp, char *name, char *link); 275 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, int txtype, 276 znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp); 277 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype, 278 znode_t *zp, offset_t off, ssize_t len, int ioflag); 279 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype, 280 znode_t *zp, uint64_t off, uint64_t len); 281 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype, 282 znode_t *zp, vattr_t *vap, uint_t mask_applied); 283 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, int txtype, 284 znode_t *zp, int aclcnt, ace_t *z_ace); 285 286 extern zil_get_data_t zfs_get_data; 287 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE]; 288 extern int zfsfstype; 289 290 #endif /* _KERNEL */ 291 292 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len); 293 294 #ifdef __cplusplus 295 } 296 #endif 297 298 #endif /* _SYS_FS_ZFS_ZNODE_H */ 299