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 (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. 23 * Copyright (c) 2012, 2015 by Delphix. All rights reserved. 24 * Copyright (c) 2014 Integros [integros.com] 25 * Copyright 2016 Nexenta Systems, Inc. All rights reserved. 26 */ 27 28#ifndef _SYS_FS_ZFS_ZNODE_H 29#define _SYS_FS_ZFS_ZNODE_H 30 31#ifdef _KERNEL 32#include <sys/list.h> 33#include <sys/dmu.h> 34#include <sys/sa.h> 35#include <sys/zfs_vfsops.h> 36#include <sys/rrwlock.h> 37#include <sys/zfs_sa.h> 38#include <sys/zfs_stat.h> 39#endif 40#include <sys/zfs_acl.h> 41#include <sys/zil.h> 42 43#ifdef __cplusplus 44extern "C" { 45#endif 46 47/* 48 * Additional file level attributes, that are stored 49 * in the upper half of zp_flags 50 */ 51#define ZFS_READONLY 0x0000000100000000 52#define ZFS_HIDDEN 0x0000000200000000 53#define ZFS_SYSTEM 0x0000000400000000 54#define ZFS_ARCHIVE 0x0000000800000000 55#define ZFS_IMMUTABLE 0x0000001000000000 56#define ZFS_NOUNLINK 0x0000002000000000 57#define ZFS_APPENDONLY 0x0000004000000000 58#define ZFS_NODUMP 0x0000008000000000 59#define ZFS_OPAQUE 0x0000010000000000 60#define ZFS_AV_QUARANTINED 0x0000020000000000 61#define ZFS_AV_MODIFIED 0x0000040000000000 62#define ZFS_REPARSE 0x0000080000000000 63#define ZFS_OFFLINE 0x0000100000000000 64#define ZFS_SPARSE 0x0000200000000000 65 66#define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \ 67{ \ 68 if (value) \ 69 pflags |= attr; \ 70 else \ 71 pflags &= ~attr; \ 72 VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(zp->z_zfsvfs), \ 73 &pflags, sizeof (pflags), tx)); \ 74} 75 76/* 77 * Define special zfs pflags 78 */ 79#define ZFS_XATTR 0x1 /* is an extended attribute */ 80#define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */ 81#define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */ 82#define ZFS_ACL_OBJ_ACE 0x8 /* ACL has CMPLX Object ACE */ 83#define ZFS_ACL_PROTECTED 0x10 /* ACL protected */ 84#define ZFS_ACL_DEFAULTED 0x20 /* ACL should be defaulted */ 85#define ZFS_ACL_AUTO_INHERIT 0x40 /* ACL should be inherited */ 86#define ZFS_BONUS_SCANSTAMP 0x80 /* Scanstamp in bonus area */ 87#define ZFS_NO_EXECS_DENIED 0x100 /* exec was given to everyone */ 88 89#define SA_ZPL_ATIME(z) z->z_attr_table[ZPL_ATIME] 90#define SA_ZPL_MTIME(z) z->z_attr_table[ZPL_MTIME] 91#define SA_ZPL_CTIME(z) z->z_attr_table[ZPL_CTIME] 92#define SA_ZPL_CRTIME(z) z->z_attr_table[ZPL_CRTIME] 93#define SA_ZPL_GEN(z) z->z_attr_table[ZPL_GEN] 94#define SA_ZPL_DACL_ACES(z) z->z_attr_table[ZPL_DACL_ACES] 95#define SA_ZPL_XATTR(z) z->z_attr_table[ZPL_XATTR] 96#define SA_ZPL_SYMLINK(z) z->z_attr_table[ZPL_SYMLINK] 97#define SA_ZPL_RDEV(z) z->z_attr_table[ZPL_RDEV] 98#define SA_ZPL_SCANSTAMP(z) z->z_attr_table[ZPL_SCANSTAMP] 99#define SA_ZPL_UID(z) z->z_attr_table[ZPL_UID] 100#define SA_ZPL_GID(z) z->z_attr_table[ZPL_GID] 101#define SA_ZPL_PARENT(z) z->z_attr_table[ZPL_PARENT] 102#define SA_ZPL_LINKS(z) z->z_attr_table[ZPL_LINKS] 103#define SA_ZPL_MODE(z) z->z_attr_table[ZPL_MODE] 104#define SA_ZPL_DACL_COUNT(z) z->z_attr_table[ZPL_DACL_COUNT] 105#define SA_ZPL_FLAGS(z) z->z_attr_table[ZPL_FLAGS] 106#define SA_ZPL_SIZE(z) z->z_attr_table[ZPL_SIZE] 107#define SA_ZPL_ZNODE_ACL(z) z->z_attr_table[ZPL_ZNODE_ACL] 108#define SA_ZPL_PAD(z) z->z_attr_table[ZPL_PAD] 109 110/* 111 * Is ID ephemeral? 112 */ 113#define IS_EPHEMERAL(x) (x > MAXUID) 114 115/* 116 * Should we use FUIDs? 117 */ 118#define USE_FUIDS(version, os) (version >= ZPL_VERSION_FUID && \ 119 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID) 120#define USE_SA(version, os) (version >= ZPL_VERSION_SA && \ 121 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA) 122 123#define MASTER_NODE_OBJ 1 124 125/* 126 * Special attributes for master node. 127 * "userquota@" and "groupquota@" are also valid (from 128 * zfs_userquota_prop_prefixes[]). 129 */ 130#define ZFS_FSID "FSID" 131#define ZFS_UNLINKED_SET "DELETE_QUEUE" 132#define ZFS_ROOT_OBJ "ROOT" 133#define ZPL_VERSION_STR "VERSION" 134#define ZFS_FUID_TABLES "FUID" 135#define ZFS_SHARES_DIR "SHARES" 136#define ZFS_SA_ATTRS "SA_ATTRS" 137 138/* 139 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in 140 * the directory entries. 141 */ 142#ifndef IFTODT 143#define IFTODT(mode) (((mode) & S_IFMT) >> 12) 144#endif 145 146/* 147 * The directory entry has the type (currently unused on Solaris) in the 148 * top 4 bits, and the object number in the low 48 bits. The "middle" 149 * 12 bits are unused. 150 */ 151#define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4) 152#define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48) 153 154/* 155 * Directory entry locks control access to directory entries. 156 * They are used to protect creates, deletes, and renames. 157 * Each directory znode has a mutex and a list of locked names. 158 */ 159#ifdef _KERNEL 160typedef struct zfs_dirlock { 161 char *dl_name; /* directory entry being locked */ 162 uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */ 163 uint8_t dl_namelock; /* 1 if z_name_lock is NOT held */ 164 uint16_t dl_namesize; /* set if dl_name was allocated */ 165 kcondvar_t dl_cv; /* wait for entry to be unlocked */ 166 struct znode *dl_dzp; /* directory znode */ 167 struct zfs_dirlock *dl_next; /* next in z_dirlocks list */ 168} zfs_dirlock_t; 169 170typedef struct znode { 171 struct zfsvfs *z_zfsvfs; 172 vnode_t *z_vnode; 173 uint64_t z_id; /* object ID for this znode */ 174#ifdef illumos 175 kmutex_t z_lock; /* znode modification lock */ 176 krwlock_t z_parent_lock; /* parent lock for directories */ 177 krwlock_t z_name_lock; /* "master" lock for dirent locks */ 178 zfs_dirlock_t *z_dirlocks; /* directory entry lock list */ 179#endif 180 kmutex_t z_range_lock; /* protects changes to z_range_avl */ 181 avl_tree_t z_range_avl; /* avl tree of file range locks */ 182 uint8_t z_unlinked; /* file has been unlinked */ 183 uint8_t z_atime_dirty; /* atime needs to be synced */ 184 uint8_t z_zn_prefetch; /* Prefetch znodes? */ 185 uint8_t z_moved; /* Has this znode been moved? */ 186 uint_t z_blksz; /* block size in bytes */ 187 uint_t z_seq; /* modification sequence number */ 188 uint64_t z_mapcnt; /* number of pages mapped to file */ 189 uint64_t z_gen; /* generation (cached) */ 190 uint64_t z_size; /* file size (cached) */ 191 uint64_t z_atime[2]; /* atime (cached) */ 192 uint64_t z_links; /* file links (cached) */ 193 uint64_t z_pflags; /* pflags (cached) */ 194 uint64_t z_uid; /* uid fuid (cached) */ 195 uint64_t z_gid; /* gid fuid (cached) */ 196 mode_t z_mode; /* mode (cached) */ 197 uint32_t z_sync_cnt; /* synchronous open count */ 198 kmutex_t z_acl_lock; /* acl data lock */ 199 zfs_acl_t *z_acl_cached; /* cached acl */ 200 list_node_t z_link_node; /* all znodes in fs link */ 201 sa_handle_t *z_sa_hdl; /* handle to sa data */ 202 boolean_t z_is_sa; /* are we native sa? */ 203} znode_t; 204 205 206/* 207 * Range locking rules 208 * -------------------- 209 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole 210 * file range needs to be locked as RL_WRITER. Only then can the pages be 211 * freed etc and zp_size reset. zp_size must be set within range lock. 212 * 2. For writes and punching holes (zfs_write & zfs_space) just the range 213 * being written or freed needs to be locked as RL_WRITER. 214 * Multiple writes at the end of the file must coordinate zp_size updates 215 * to ensure data isn't lost. A compare and swap loop is currently used 216 * to ensure the file size is at least the offset last written. 217 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being 218 * read needs to be locked as RL_READER. A check against zp_size can then 219 * be made for reading beyond end of file. 220 */ 221 222/* 223 * Convert between znode pointers and vnode pointers 224 */ 225#ifdef DEBUG 226static __inline vnode_t * 227ZTOV(znode_t *zp) 228{ 229 vnode_t *vp = zp->z_vnode; 230 231 ASSERT(vp != NULL && vp->v_data == zp); 232 return (vp); 233} 234static __inline znode_t * 235VTOZ(vnode_t *vp) 236{ 237 znode_t *zp = (znode_t *)vp->v_data; 238 239 ASSERT(zp != NULL && zp->z_vnode == vp); 240 return (zp); 241} 242#else 243#define ZTOV(ZP) ((ZP)->z_vnode) 244#define VTOZ(VP) ((znode_t *)(VP)->v_data) 245#endif 246 247/* Called on entry to each ZFS vnode and vfs operation */ 248#define ZFS_ENTER(zfsvfs) \ 249 { \ 250 rrm_enter_read(&(zfsvfs)->z_teardown_lock, FTAG); \ 251 if ((zfsvfs)->z_unmounted) { \ 252 ZFS_EXIT(zfsvfs); \ 253 return (EIO); \ 254 } \ 255 } 256 257/* Must be called before exiting the vop */ 258#define ZFS_EXIT(zfsvfs) rrm_exit(&(zfsvfs)->z_teardown_lock, FTAG) 259 260/* Verifies the znode is valid */ 261#define ZFS_VERIFY_ZP(zp) \ 262 if ((zp)->z_sa_hdl == NULL) { \ 263 ZFS_EXIT((zp)->z_zfsvfs); \ 264 return (EIO); \ 265 } \ 266 267/* 268 * Macros for dealing with dmu_buf_hold 269 */ 270#define ZFS_OBJ_HASH(obj_num) ((obj_num) & (ZFS_OBJ_MTX_SZ - 1)) 271#define ZFS_OBJ_MUTEX(zfsvfs, obj_num) \ 272 (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]) 273#define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \ 274 mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) 275#define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \ 276 mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) 277#define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \ 278 mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) 279 280/* Encode ZFS stored time values from a struct timespec */ 281#define ZFS_TIME_ENCODE(tp, stmp) \ 282{ \ 283 (stmp)[0] = (uint64_t)(tp)->tv_sec; \ 284 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \ 285} 286 287/* Decode ZFS stored time values to a struct timespec */ 288#define ZFS_TIME_DECODE(tp, stmp) \ 289{ \ 290 (tp)->tv_sec = (time_t)(stmp)[0]; \ 291 (tp)->tv_nsec = (long)(stmp)[1]; \ 292} 293 294/* 295 * Timestamp defines 296 */ 297#define ACCESSED (AT_ATIME) 298#define STATE_CHANGED (AT_CTIME) 299#define CONTENT_MODIFIED (AT_MTIME | AT_CTIME) 300 301#define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \ 302 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \ 303 zfs_tstamp_update_setup(zp, ACCESSED, NULL, NULL, B_FALSE); 304 305extern int zfs_init_fs(zfsvfs_t *, znode_t **); 306extern void zfs_set_dataprop(objset_t *); 307extern void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *, 308 dmu_tx_t *tx); 309extern void zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2], 310 uint64_t [2], boolean_t); 311extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *); 312extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t); 313extern void zfs_znode_init(void); 314extern void zfs_znode_fini(void); 315extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **); 316extern int zfs_rezget(znode_t *); 317extern void zfs_zinactive(znode_t *); 318extern void zfs_znode_delete(znode_t *, dmu_tx_t *); 319extern void zfs_znode_free(znode_t *); 320extern void zfs_remove_op_tables(); 321extern int zfs_create_op_tables(); 322extern dev_t zfs_cmpldev(uint64_t); 323extern int zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value); 324extern int zfs_get_stats(objset_t *os, nvlist_t *nv); 325extern boolean_t zfs_get_vfs_flag_unmounted(objset_t *os); 326extern void zfs_znode_dmu_fini(znode_t *); 327 328extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, 329 znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *, 330 vattr_t *vap); 331extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp, 332 vattr_t *vap); 333extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, 334 znode_t *dzp, char *name, uint64_t foid); 335#define ZFS_NO_OBJECT 0 /* no object id */ 336extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, 337 znode_t *dzp, znode_t *zp, char *name); 338extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, 339 znode_t *dzp, znode_t *zp, char *name, char *link); 340extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, 341 znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp); 342extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype, 343 znode_t *zp, offset_t off, ssize_t len, int ioflag); 344extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype, 345 znode_t *zp, uint64_t off, uint64_t len); 346extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype, 347 znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp); 348#ifndef ZFS_NO_ACL 349extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp, 350 vsecattr_t *vsecp, zfs_fuid_info_t *fuidp); 351#endif 352extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx); 353extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx); 354extern int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx); 355 356extern zil_get_data_t zfs_get_data; 357extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE]; 358extern int zfsfstype; 359 360extern int zfs_znode_parent_and_name(znode_t *zp, znode_t **dzpp, char *buf); 361 362#endif /* _KERNEL */ 363 364extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len); 365 366#ifdef __cplusplus 367} 368#endif 369 370#endif /* _SYS_FS_ZFS_ZNODE_H */ 371