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.
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.
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]
22 * Copyright 2008 Sun Microsystems, Inc. All rights reserved.
23 * Use is subject to license terms.
26 #ifndef _SYS_FS_ZFS_ZNODE_H
27 #define _SYS_FS_ZFS_ZNODE_H
32 #include <sys/zfs_vfsops.h>
33 #include <sys/rrwlock.h>
35 #include <sys/zfs_acl.h>
43 * Additional file level attributes, that are stored
44 * in the upper half of zp_flags
46 #define ZFS_READONLY 0x0000000100000000
47 #define ZFS_HIDDEN 0x0000000200000000
48 #define ZFS_SYSTEM 0x0000000400000000
49 #define ZFS_ARCHIVE 0x0000000800000000
50 #define ZFS_IMMUTABLE 0x0000001000000000
51 #define ZFS_NOUNLINK 0x0000002000000000
52 #define ZFS_APPENDONLY 0x0000004000000000
53 #define ZFS_NODUMP 0x0000008000000000
54 #define ZFS_OPAQUE 0x0000010000000000
55 #define ZFS_AV_QUARANTINED 0x0000020000000000
56 #define ZFS_AV_MODIFIED 0x0000040000000000
58 #define ZFS_ATTR_SET(zp, attr, value) \
61 zp->z_phys->zp_flags |= attr; \
63 zp->z_phys->zp_flags &= ~attr; \
67 * Define special zfs pflags
69 #define ZFS_XATTR 0x1 /* is an extended attribute */
70 #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */
71 #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */
72 #define ZFS_ACL_OBJ_ACE 0x8 /* ACL has CMPLX Object ACE */
73 #define ZFS_ACL_PROTECTED 0x10 /* ACL protected */
74 #define ZFS_ACL_DEFAULTED 0x20 /* ACL should be defaulted */
75 #define ZFS_ACL_AUTO_INHERIT 0x40 /* ACL should be inherited */
76 #define ZFS_BONUS_SCANSTAMP 0x80 /* Scanstamp in bonus area */
82 #define IS_EPHEMERAL(x) (x > MAXUID)
84 #define IS_EPHEMERAL(x) (0)
88 * Should we use FUIDs?
90 #define USE_FUIDS(version, os) (version >= ZPL_VERSION_FUID &&\
91 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
93 #define MASTER_NODE_OBJ 1
96 * Special attributes for master node.
98 #define ZFS_FSID "FSID"
99 #define ZFS_UNLINKED_SET "DELETE_QUEUE"
100 #define ZFS_ROOT_OBJ "ROOT"
101 #define ZPL_VERSION_STR "VERSION"
102 #define ZFS_FUID_TABLES "FUID"
104 #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE)
106 /* Path component length */
108 * The generic fs code uses MAXNAMELEN to represent
109 * what the largest component length is. Unfortunately,
110 * this length includes the terminating NULL. ZFS needs
111 * to tell the users via pathconf() and statvfs() what the
112 * true maximum length of a component is, excluding the NULL.
114 #define ZFS_MAXNAMELEN (MAXNAMELEN - 1)
117 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
118 * the directory entries.
121 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
125 * The directory entry has the type (currently unused on Solaris) in the
126 * top 4 bits, and the object number in the low 48 bits. The "middle"
127 * 12 bits are unused.
129 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
130 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
133 * This is the persistent portion of the znode. It is stored
134 * in the "bonus buffer" of the file. Short symbolic links
135 * are also stored in the bonus buffer.
137 typedef struct znode_phys {
138 uint64_t zp_atime[2]; /* 0 - last file access time */
139 uint64_t zp_mtime[2]; /* 16 - last file modification time */
140 uint64_t zp_ctime[2]; /* 32 - last file change time */
141 uint64_t zp_crtime[2]; /* 48 - creation time */
142 uint64_t zp_gen; /* 64 - generation (txg of creation) */
143 uint64_t zp_mode; /* 72 - file mode bits */
144 uint64_t zp_size; /* 80 - size of file */
145 uint64_t zp_parent; /* 88 - directory parent (`..') */
146 uint64_t zp_links; /* 96 - number of links to file */
147 uint64_t zp_xattr; /* 104 - DMU object for xattrs */
148 uint64_t zp_rdev; /* 112 - dev_t for VBLK & VCHR files */
149 uint64_t zp_flags; /* 120 - persistent flags */
150 uint64_t zp_uid; /* 128 - file owner */
151 uint64_t zp_gid; /* 136 - owning group */
152 uint64_t zp_zap; /* 144 - extra attributes */
153 uint64_t zp_pad[3]; /* 152 - future */
154 zfs_acl_phys_t zp_acl; /* 176 - 263 ACL */
156 * Data may pad out any remaining bytes in the znode buffer, eg:
158 * |<---------------------- dnode_phys (512) ------------------------>|
159 * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->|
160 * |<---- znode (264) ---->|<---- data (56) ---->|
162 * At present, we use this space for the following:
164 * - 32-byte anti-virus scanstamp (regular files only)
169 * Directory entry locks control access to directory entries.
170 * They are used to protect creates, deletes, and renames.
171 * Each directory znode has a mutex and a list of locked names.
174 typedef struct zfs_dirlock {
175 char *dl_name; /* directory entry being locked */
176 uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */
177 uint8_t dl_namelock; /* 1 if z_name_lock is NOT held */
178 uint16_t dl_namesize; /* set if dl_name was allocated */
179 kcondvar_t dl_cv; /* wait for entry to be unlocked */
180 struct znode *dl_dzp; /* directory znode */
181 struct zfs_dirlock *dl_next; /* next in z_dirlocks list */
184 typedef struct znode {
185 struct zfsvfs *z_zfsvfs;
187 uint64_t z_id; /* object ID for this znode */
188 kmutex_t z_lock; /* znode modification lock */
189 krwlock_t z_map_lock; /* page map lock */
190 krwlock_t z_parent_lock; /* parent lock for directories */
191 krwlock_t z_name_lock; /* "master" lock for dirent locks */
192 zfs_dirlock_t *z_dirlocks; /* directory entry lock list */
193 kmutex_t z_range_lock; /* protects changes to z_range_avl */
194 avl_tree_t z_range_avl; /* avl tree of file range locks */
195 uint8_t z_unlinked; /* file has been unlinked */
196 uint8_t z_atime_dirty; /* atime needs to be synced */
197 uint8_t z_zn_prefetch; /* Prefetch znodes? */
198 uint_t z_blksz; /* block size in bytes */
199 uint_t z_seq; /* modification sequence number */
200 uint64_t z_mapcnt; /* number of pages mapped to file */
201 uint64_t z_last_itx; /* last ZIL itx on this znode */
202 uint64_t z_gen; /* generation (same as zp_gen) */
203 uint32_t z_sync_cnt; /* synchronous open count */
204 kmutex_t z_acl_lock; /* acl data lock */
205 list_node_t z_link_node; /* all znodes in fs link */
207 * These are dmu managed fields.
209 znode_phys_t *z_phys; /* pointer to persistent znode */
210 dmu_buf_t *z_dbuf; /* buffer containing the z_phys */
211 /* FreeBSD-specific field. */
217 * Range locking rules
218 * --------------------
219 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
220 * file range needs to be locked as RL_WRITER. Only then can the pages be
221 * freed etc and zp_size reset. zp_size must be set within range lock.
222 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
223 * being written or freed needs to be locked as RL_WRITER.
224 * Multiple writes at the end of the file must coordinate zp_size updates
225 * to ensure data isn't lost. A compare and swap loop is currently used
226 * to ensure the file size is at least the offset last written.
227 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
228 * read needs to be locked as RL_READER. A check against zp_size can then
229 * be made for reading beyond end of file.
233 * Convert between znode pointers and vnode pointers
236 static __inline vnode_t *
239 vnode_t *vp = zp->z_vnode;
241 ASSERT(vp == NULL || vp->v_data == NULL || vp->v_data == zp);
244 static __inline znode_t *
247 znode_t *zp = (znode_t *)vp->v_data;
249 ASSERT(zp == NULL || zp->z_vnode == NULL || zp->z_vnode == vp);
253 #define ZTOV(ZP) ((ZP)->z_vnode)
254 #define VTOZ(VP) ((znode_t *)(VP)->v_data)
258 * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation.
259 * ZFS_ENTER_NOERROR() is called when we can't return EIO.
260 * ZFS_EXIT() must be called before exitting the vop.
261 * ZFS_VERIFY_ZP() verifies the znode is valid.
263 #define ZFS_ENTER(zfsvfs) \
265 rrw_enter(&(zfsvfs)->z_teardown_lock, RW_READER, FTAG); \
266 if ((zfsvfs)->z_unmounted) { \
272 #define ZFS_ENTER_NOERROR(zfsvfs) \
273 rrw_enter(&(zfsvfs)->z_teardown_lock, RW_READER, FTAG)
275 #define ZFS_EXIT(zfsvfs) rrw_exit(&(zfsvfs)->z_teardown_lock, FTAG)
277 #define ZFS_VERIFY_ZP(zp) \
278 if ((zp)->z_dbuf == NULL) { \
279 ZFS_EXIT((zp)->z_zfsvfs); \
284 * Macros for dealing with dmu_buf_hold
286 #define ZFS_OBJ_HASH(obj_num) ((obj_num) & (ZFS_OBJ_MTX_SZ - 1))
287 #define ZFS_OBJ_MUTEX(zfsvfs, obj_num) \
288 (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
289 #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \
290 mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
291 #define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \
292 mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
293 #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \
294 mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
297 * Macros to encode/decode ZFS stored time values from/to struct timespec
299 #define ZFS_TIME_ENCODE(tp, stmp) \
301 (stmp)[0] = (uint64_t)(tp)->tv_sec; \
302 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
305 #define ZFS_TIME_DECODE(tp, stmp) \
307 (tp)->tv_sec = (time_t)(stmp)[0]; \
308 (tp)->tv_nsec = (long)(stmp)[1]; \
314 #define ACCESSED (AT_ATIME)
315 #define STATE_CHANGED (AT_CTIME)
316 #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME)
318 #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \
319 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \
320 zfs_time_stamper(zp, ACCESSED, NULL)
322 extern int zfs_init_fs(zfsvfs_t *, znode_t **);
323 extern void zfs_set_dataprop(objset_t *);
324 extern void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
326 extern void zfs_time_stamper(znode_t *, uint_t, dmu_tx_t *);
327 extern void zfs_time_stamper_locked(znode_t *, uint_t, dmu_tx_t *);
328 extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
329 extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
330 extern void zfs_znode_init(void);
331 extern void zfs_znode_fini(void);
332 extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
333 extern int zfs_rezget(znode_t *);
334 extern void zfs_zinactive(znode_t *);
335 extern void zfs_znode_delete(znode_t *, dmu_tx_t *);
336 extern void zfs_znode_free(znode_t *);
337 extern void zfs_remove_op_tables();
338 extern int zfs_create_op_tables();
339 extern dev_t zfs_cmpldev(uint64_t);
340 extern int zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
341 extern int zfs_set_version(const char *name, uint64_t newvers);
342 extern int zfs_get_stats(objset_t *os, nvlist_t *nv);
343 extern void zfs_znode_dmu_fini(znode_t *);
345 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
346 znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
348 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
350 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
351 znode_t *dzp, char *name);
352 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
353 znode_t *dzp, znode_t *zp, char *name);
354 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
355 znode_t *dzp, znode_t *zp, char *name, char *link);
356 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
357 znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
358 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
359 znode_t *zp, offset_t off, ssize_t len, int ioflag);
360 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
361 znode_t *zp, uint64_t off, uint64_t len);
362 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
363 znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
365 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
366 vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
368 extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap);
369 extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
371 extern zil_get_data_t zfs_get_data;
372 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
373 extern int zfsfstype;
377 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
383 #endif /* _SYS_FS_ZFS_ZNODE_H */