1fa9e4066Sahrens /* 2fa9e4066Sahrens * CDDL HEADER START 3fa9e4066Sahrens * 4fa9e4066Sahrens * The contents of this file are subject to the terms of the 5ea8dc4b6Seschrock * Common Development and Distribution License (the "License"). 6ea8dc4b6Seschrock * You may not use this file except in compliance with the License. 7fa9e4066Sahrens * 8fa9e4066Sahrens * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9fa9e4066Sahrens * or http://www.opensolaris.org/os/licensing. 10fa9e4066Sahrens * See the License for the specific language governing permissions 11fa9e4066Sahrens * and limitations under the License. 12fa9e4066Sahrens * 13fa9e4066Sahrens * When distributing Covered Code, include this CDDL HEADER in each 14fa9e4066Sahrens * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15fa9e4066Sahrens * If applicable, add the following below this CDDL HEADER, with the 16fa9e4066Sahrens * fields enclosed by brackets "[]" replaced with your own identifying 17fa9e4066Sahrens * information: Portions Copyright [yyyy] [name of copyright owner] 18fa9e4066Sahrens * 19fa9e4066Sahrens * CDDL HEADER END 20fa9e4066Sahrens */ 21fa9e4066Sahrens /* 2255434c77Sek * Copyright 2007 Sun Microsystems, Inc. All rights reserved. 23fa9e4066Sahrens * Use is subject to license terms. 24fa9e4066Sahrens */ 25fa9e4066Sahrens 26fa9e4066Sahrens #ifndef _SYS_FS_ZFS_ZNODE_H 27fa9e4066Sahrens #define _SYS_FS_ZFS_ZNODE_H 28fa9e4066Sahrens 29fa9e4066Sahrens #pragma ident "%Z%%M% %I% %E% SMI" 30fa9e4066Sahrens 31fa9e4066Sahrens #ifdef _KERNEL 32fa9e4066Sahrens #include <sys/isa_defs.h> 33fa9e4066Sahrens #include <sys/types32.h> 34fa9e4066Sahrens #include <sys/list.h> 35fa9e4066Sahrens #include <sys/dmu.h> 36fa9e4066Sahrens #include <sys/zfs_vfsops.h> 37fa9e4066Sahrens #endif 38fa9e4066Sahrens #include <sys/zfs_acl.h> 39fa9e4066Sahrens #include <sys/zil.h> 40fa9e4066Sahrens 41fa9e4066Sahrens #ifdef __cplusplus 42fa9e4066Sahrens extern "C" { 43fa9e4066Sahrens #endif 44fa9e4066Sahrens 45fa9e4066Sahrens /* 46fa9e4066Sahrens * Define special zfs pflags 47fa9e4066Sahrens */ 48fa9e4066Sahrens #define ZFS_XATTR 0x1 /* is an extended attribute */ 49fa9e4066Sahrens #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */ 50de122929Smarks #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */ 51fa9e4066Sahrens 52fa9e4066Sahrens #define MASTER_NODE_OBJ 1 53fa9e4066Sahrens 54fa9e4066Sahrens /* 55fa9e4066Sahrens * special attributes for master node. 56fa9e4066Sahrens */ 57fa9e4066Sahrens 58fa9e4066Sahrens #define ZFS_FSID "FSID" 59893a6d32Sahrens #define ZFS_UNLINKED_SET "DELETE_QUEUE" 60fa9e4066Sahrens #define ZFS_ROOT_OBJ "ROOT" 61eaca9bbdSeschrock #define ZPL_VERSION_OBJ "VERSION" 62fa9e4066Sahrens #define ZFS_PROP_BLOCKPERPAGE "BLOCKPERPAGE" 63fa9e4066Sahrens #define ZFS_PROP_NOGROWBLOCKS "NOGROWBLOCKS" 64fa9e4066Sahrens 65fa9e4066Sahrens #define ZFS_FLAG_BLOCKPERPAGE 0x1 66fa9e4066Sahrens #define ZFS_FLAG_NOGROWBLOCKS 0x2 67fa9e4066Sahrens 68fa9e4066Sahrens /* 69eaca9bbdSeschrock * ZPL version - rev'd whenever an incompatible on-disk format change 70fa9e4066Sahrens * occurs. Independent of SPA/DMU/ZAP versioning. 71fa9e4066Sahrens */ 72fa9e4066Sahrens 73eaca9bbdSeschrock #define ZPL_VERSION 1ULL 74fa9e4066Sahrens 75fa9e4066Sahrens #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE) 76fa9e4066Sahrens 77fa9e4066Sahrens /* Path component length */ 78fa9e4066Sahrens /* 79fa9e4066Sahrens * The generic fs code uses MAXNAMELEN to represent 80fa9e4066Sahrens * what the largest component length is. Unfortunately, 81fa9e4066Sahrens * this length includes the terminating NULL. ZFS needs 82fa9e4066Sahrens * to tell the users via pathconf() and statvfs() what the 83fa9e4066Sahrens * true maximum length of a component is, excluding the NULL. 84fa9e4066Sahrens */ 85fa9e4066Sahrens #define ZFS_MAXNAMELEN (MAXNAMELEN - 1) 86fa9e4066Sahrens 87*b1b8ab34Slling /* 88*b1b8ab34Slling * The directory entry has the type (currently unused on Solaris) in the 89*b1b8ab34Slling * top 4 bits, and the object number in the low 48 bits. The "middle" 90*b1b8ab34Slling * 12 bits are unused. 91*b1b8ab34Slling */ 92*b1b8ab34Slling #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4) 93*b1b8ab34Slling #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48) 94*b1b8ab34Slling #define ZFS_DIRENT_MAKE(type, obj) (((uint64_t)type << 60) | obj) 95*b1b8ab34Slling 96*b1b8ab34Slling 97fa9e4066Sahrens /* 98fa9e4066Sahrens * This is the persistent portion of the znode. It is stored 99fa9e4066Sahrens * in the "bonus buffer" of the file. Short symbolic links 100fa9e4066Sahrens * are also stored in the bonus buffer. 101fa9e4066Sahrens */ 102fa9e4066Sahrens typedef struct znode_phys { 103fa9e4066Sahrens uint64_t zp_atime[2]; /* 0 - last file access time */ 104fa9e4066Sahrens uint64_t zp_mtime[2]; /* 16 - last file modification time */ 105fa9e4066Sahrens uint64_t zp_ctime[2]; /* 32 - last file change time */ 106fa9e4066Sahrens uint64_t zp_crtime[2]; /* 48 - creation time */ 107fa9e4066Sahrens uint64_t zp_gen; /* 64 - generation (txg of creation) */ 108fa9e4066Sahrens uint64_t zp_mode; /* 72 - file mode bits */ 109fa9e4066Sahrens uint64_t zp_size; /* 80 - size of file */ 110fa9e4066Sahrens uint64_t zp_parent; /* 88 - directory parent (`..') */ 111fa9e4066Sahrens uint64_t zp_links; /* 96 - number of links to file */ 112fa9e4066Sahrens uint64_t zp_xattr; /* 104 - DMU object for xattrs */ 113fa9e4066Sahrens uint64_t zp_rdev; /* 112 - dev_t for VBLK & VCHR files */ 114fa9e4066Sahrens uint64_t zp_flags; /* 120 - persistent flags */ 115fa9e4066Sahrens uint64_t zp_uid; /* 128 - file owner */ 116fa9e4066Sahrens uint64_t zp_gid; /* 136 - owning group */ 117fa9e4066Sahrens uint64_t zp_pad[4]; /* 144 - future */ 118fa9e4066Sahrens zfs_znode_acl_t zp_acl; /* 176 - 263 ACL */ 119fa9e4066Sahrens /* 120fa9e4066Sahrens * Data may pad out any remaining bytes in the znode buffer, eg: 121fa9e4066Sahrens * 122fa9e4066Sahrens * |<---------------------- dnode_phys (512) ------------------------>| 123fa9e4066Sahrens * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->| 124fa9e4066Sahrens * |<---- znode (264) ---->|<---- data (56) ---->| 125fa9e4066Sahrens * 126fa9e4066Sahrens * At present, we only use this space to store symbolic links. 127fa9e4066Sahrens */ 128fa9e4066Sahrens } znode_phys_t; 129fa9e4066Sahrens 130fa9e4066Sahrens /* 131fa9e4066Sahrens * Directory entry locks control access to directory entries. 132fa9e4066Sahrens * They are used to protect creates, deletes, and renames. 133fa9e4066Sahrens * Each directory znode has a mutex and a list of locked names. 134fa9e4066Sahrens */ 135fa9e4066Sahrens #ifdef _KERNEL 136fa9e4066Sahrens typedef struct zfs_dirlock { 137fa9e4066Sahrens char *dl_name; /* directory entry being locked */ 138fa9e4066Sahrens uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */ 139fa9e4066Sahrens uint16_t dl_namesize; /* set if dl_name was allocated */ 140fa9e4066Sahrens kcondvar_t dl_cv; /* wait for entry to be unlocked */ 141fa9e4066Sahrens struct znode *dl_dzp; /* directory znode */ 142fa9e4066Sahrens struct zfs_dirlock *dl_next; /* next in z_dirlocks list */ 143fa9e4066Sahrens } zfs_dirlock_t; 144fa9e4066Sahrens 145fa9e4066Sahrens typedef struct znode { 146fa9e4066Sahrens struct zfsvfs *z_zfsvfs; 147fa9e4066Sahrens vnode_t *z_vnode; 148fa9e4066Sahrens uint64_t z_id; /* object ID for this znode */ 149fa9e4066Sahrens kmutex_t z_lock; /* znode modification lock */ 150fa9e4066Sahrens krwlock_t z_map_lock; /* page map lock */ 151104e2ed7Sperrin krwlock_t z_parent_lock; /* parent lock for directories */ 152af2c4821Smaybee krwlock_t z_name_lock; /* "master" lock for dirent locks */ 153fa9e4066Sahrens zfs_dirlock_t *z_dirlocks; /* directory entry lock list */ 154104e2ed7Sperrin kmutex_t z_range_lock; /* protects changes to z_range_avl */ 155104e2ed7Sperrin avl_tree_t z_range_avl; /* avl tree of file range locks */ 156893a6d32Sahrens uint8_t z_unlinked; /* file has been unlinked */ 157fa9e4066Sahrens uint8_t z_atime_dirty; /* atime needs to be synced */ 158fa9e4066Sahrens uint8_t z_dbuf_held; /* Is z_dbuf already held? */ 1597f6e3e7dSperrin uint8_t z_zn_prefetch; /* Prefetch znodes? */ 160fa9e4066Sahrens uint_t z_blksz; /* block size in bytes */ 161fa9e4066Sahrens uint_t z_seq; /* modification sequence number */ 162ea8dc4b6Seschrock uint64_t z_mapcnt; /* number of pages mapped to file */ 163fa9e4066Sahrens uint64_t z_last_itx; /* last ZIL itx on this znode */ 16467bd71c6Sperrin uint32_t z_sync_cnt; /* synchronous open count */ 165fa9e4066Sahrens kmutex_t z_acl_lock; /* acl data lock */ 166fa9e4066Sahrens list_node_t z_link_node; /* all znodes in fs link */ 167fa9e4066Sahrens /* 168fa9e4066Sahrens * These are dmu managed fields. 169fa9e4066Sahrens */ 170fa9e4066Sahrens znode_phys_t *z_phys; /* pointer to persistent znode */ 171fa9e4066Sahrens dmu_buf_t *z_dbuf; /* buffer containing the z_phys */ 172fa9e4066Sahrens } znode_t; 173fa9e4066Sahrens 174104e2ed7Sperrin 175fa9e4066Sahrens /* 176104e2ed7Sperrin * Range locking rules 177104e2ed7Sperrin * -------------------- 178104e2ed7Sperrin * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole 179104e2ed7Sperrin * file range needs to be locked as RL_WRITER. Only then can the pages be 180104e2ed7Sperrin * freed etc and zp_size reset. zp_size must be set within range lock. 181104e2ed7Sperrin * 2. For writes and punching holes (zfs_write & zfs_space) just the range 182104e2ed7Sperrin * being written or freed needs to be locked as RL_WRITER. 183104e2ed7Sperrin * Multiple writes at the end of the file must coordinate zp_size updates 184104e2ed7Sperrin * to ensure data isn't lost. A compare and swap loop is currently used 185104e2ed7Sperrin * to ensure the file size is at least the offset last written. 186104e2ed7Sperrin * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being 187104e2ed7Sperrin * read needs to be locked as RL_READER. A check against zp_size can then 188104e2ed7Sperrin * be made for reading beyond end of file. 189fa9e4066Sahrens */ 190fa9e4066Sahrens 191fa9e4066Sahrens /* 192fa9e4066Sahrens * Convert between znode pointers and vnode pointers 193fa9e4066Sahrens */ 194fa9e4066Sahrens #define ZTOV(ZP) ((ZP)->z_vnode) 195fa9e4066Sahrens #define VTOZ(VP) ((znode_t *)(VP)->v_data) 196fa9e4066Sahrens 197fa9e4066Sahrens /* 198fa9e4066Sahrens * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation. 199fa9e4066Sahrens * ZFS_EXIT() must be called before exitting the vop. 200fa9e4066Sahrens */ 201fa9e4066Sahrens #define ZFS_ENTER(zfsvfs) \ 202fa9e4066Sahrens { \ 203fa9e4066Sahrens atomic_add_32(&(zfsvfs)->z_op_cnt, 1); \ 204fa9e4066Sahrens if ((zfsvfs)->z_unmounted1) { \ 205fa9e4066Sahrens ZFS_EXIT(zfsvfs); \ 206fa9e4066Sahrens return (EIO); \ 207fa9e4066Sahrens } \ 208fa9e4066Sahrens } 209fa9e4066Sahrens #define ZFS_EXIT(zfsvfs) atomic_add_32(&(zfsvfs)->z_op_cnt, -1) 210fa9e4066Sahrens 211fa9e4066Sahrens /* 212fa9e4066Sahrens * Macros for dealing with dmu_buf_hold 213fa9e4066Sahrens */ 214fa9e4066Sahrens #define ZFS_OBJ_HASH(obj_num) (obj_num & (ZFS_OBJ_MTX_SZ - 1)) 215fa9e4066Sahrens #define ZFS_OBJ_MUTEX(zp) \ 216fa9e4066Sahrens (&zp->z_zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(zp->z_id)]) 217fa9e4066Sahrens #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \ 218fa9e4066Sahrens mutex_enter(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]); 219fa9e4066Sahrens 220fa9e4066Sahrens #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \ 221fa9e4066Sahrens mutex_exit(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]) 222fa9e4066Sahrens 223fa9e4066Sahrens /* 224fa9e4066Sahrens * Macros to encode/decode ZFS stored time values from/to struct timespec 225fa9e4066Sahrens */ 226fa9e4066Sahrens #define ZFS_TIME_ENCODE(tp, stmp) \ 227fa9e4066Sahrens { \ 228fa9e4066Sahrens stmp[0] = (uint64_t)(tp)->tv_sec; \ 229fa9e4066Sahrens stmp[1] = (uint64_t)(tp)->tv_nsec; \ 230fa9e4066Sahrens } 231fa9e4066Sahrens 232fa9e4066Sahrens #define ZFS_TIME_DECODE(tp, stmp) \ 233fa9e4066Sahrens { \ 234fa9e4066Sahrens (tp)->tv_sec = (time_t)stmp[0]; \ 235fa9e4066Sahrens (tp)->tv_nsec = (long)stmp[1]; \ 236fa9e4066Sahrens } 237fa9e4066Sahrens 238fa9e4066Sahrens /* 239fa9e4066Sahrens * Timestamp defines 240fa9e4066Sahrens */ 241fa9e4066Sahrens #define ACCESSED (AT_ATIME) 242fa9e4066Sahrens #define STATE_CHANGED (AT_CTIME) 243fa9e4066Sahrens #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME) 244fa9e4066Sahrens 245fa9e4066Sahrens #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \ 246fa9e4066Sahrens if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \ 247fa9e4066Sahrens zfs_time_stamper(zp, ACCESSED, NULL) 248fa9e4066Sahrens 249fa9e4066Sahrens extern int zfs_init_fs(zfsvfs_t *, znode_t **, cred_t *); 250fa9e4066Sahrens extern void zfs_set_dataprop(objset_t *); 251fa9e4066Sahrens extern void zfs_create_fs(objset_t *os, cred_t *cr, dmu_tx_t *tx); 252fa9e4066Sahrens extern void zfs_time_stamper(znode_t *, uint_t, dmu_tx_t *); 253fa9e4066Sahrens extern void zfs_time_stamper_locked(znode_t *, uint_t, dmu_tx_t *); 254104e2ed7Sperrin extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *); 2555730cc9aSmaybee extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t); 256fa9e4066Sahrens extern void zfs_znode_init(void); 257fa9e4066Sahrens extern void zfs_znode_fini(void); 258fa9e4066Sahrens extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **); 259fa9e4066Sahrens extern void zfs_zinactive(znode_t *); 260fa9e4066Sahrens extern void zfs_znode_delete(znode_t *, dmu_tx_t *); 261fa9e4066Sahrens extern void zfs_znode_free(znode_t *); 262fa9e4066Sahrens extern void zfs_remove_op_tables(); 263fa9e4066Sahrens extern int zfs_create_op_tables(); 264fa9e4066Sahrens extern int zfs_sync(vfs_t *vfsp, short flag, cred_t *cr); 26572fc53bcSmarks extern dev_t zfs_cmpldev(uint64_t); 266fa9e4066Sahrens 267b19a79ecSperrin extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, int txtype, 268fa9e4066Sahrens znode_t *dzp, znode_t *zp, char *name); 269b19a79ecSperrin extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, int txtype, 270fa9e4066Sahrens znode_t *dzp, char *name); 271b19a79ecSperrin extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, int txtype, 272fa9e4066Sahrens znode_t *dzp, znode_t *zp, char *name); 273b19a79ecSperrin extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, int txtype, 274fa9e4066Sahrens znode_t *dzp, znode_t *zp, char *name, char *link); 275b19a79ecSperrin extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, int txtype, 276fa9e4066Sahrens znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp); 277b19a79ecSperrin extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype, 278feb08c6bSbillm znode_t *zp, offset_t off, ssize_t len, int ioflag); 279b19a79ecSperrin extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype, 280fa9e4066Sahrens znode_t *zp, uint64_t off, uint64_t len); 281b19a79ecSperrin extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype, 282fa9e4066Sahrens znode_t *zp, vattr_t *vap, uint_t mask_applied); 283b19a79ecSperrin extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, int txtype, 284fa9e4066Sahrens znode_t *zp, int aclcnt, ace_t *z_ace); 285fa9e4066Sahrens 286fa9e4066Sahrens extern zil_get_data_t zfs_get_data; 287fa9e4066Sahrens extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE]; 288fa9e4066Sahrens extern int zfsfstype; 289fa9e4066Sahrens 290fa9e4066Sahrens #endif /* _KERNEL */ 291fa9e4066Sahrens 29255434c77Sek extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len); 29355434c77Sek 294fa9e4066Sahrens #ifdef __cplusplus 295fa9e4066Sahrens } 296fa9e4066Sahrens #endif 297fa9e4066Sahrens 298fa9e4066Sahrens #endif /* _SYS_FS_ZFS_ZNODE_H */ 299