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 (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
25 #ifndef _SYS_FS_ZFS_ZNODE_H
26 #define _SYS_FS_ZFS_ZNODE_H
29 #include <sys/isa_defs.h>
30 #include <sys/types32.h>
35 #include <sys/zfs_vfsops.h>
36 #include <sys/rrwlock.h>
37 #include <sys/zfs_sa.h>
38 #include <sys/zfs_stat.h>
40 #include <sys/zfs_acl.h>
48 * Additional file level attributes, that are stored
49 * in the upper half of zp_flags
51 #define ZFS_READONLY 0x0000000100000000ull
52 #define ZFS_HIDDEN 0x0000000200000000ull
53 #define ZFS_SYSTEM 0x0000000400000000ull
54 #define ZFS_ARCHIVE 0x0000000800000000ull
55 #define ZFS_IMMUTABLE 0x0000001000000000ull
56 #define ZFS_NOUNLINK 0x0000002000000000ull
57 #define ZFS_APPENDONLY 0x0000004000000000ull
58 #define ZFS_NODUMP 0x0000008000000000ull
59 #define ZFS_OPAQUE 0x0000010000000000ull
60 #define ZFS_AV_QUARANTINED 0x0000020000000000ull
61 #define ZFS_AV_MODIFIED 0x0000040000000000ull
62 #define ZFS_REPARSE 0x0000080000000000ull
63 #define ZFS_OFFLINE 0x0000100000000000ull
64 #define ZFS_SPARSE 0x0000200000000000ull
66 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
72 VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(zp->z_zfsvfs), \
73 &pflags, sizeof (pflags), tx)); \
77 * Define special zfs pflags
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 */
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]
113 #define IS_EPHEMERAL(x) (x > MAXUID)
116 * Should we use FUIDs?
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)
123 #define MASTER_NODE_OBJ 1
126 * Special attributes for master node.
127 * "userquota@" and "groupquota@" are also valid (from
128 * zfs_userquota_prop_prefixes[]).
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"
138 #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE)
140 /* Path component length */
142 * The generic fs code uses MAXNAMELEN to represent
143 * what the largest component length is. Unfortunately,
144 * this length includes the terminating NULL. ZFS needs
145 * to tell the users via pathconf() and statvfs() what the
146 * true maximum length of a component is, excluding the NULL.
148 #define ZFS_MAXNAMELEN (MAXNAMELEN - 1)
151 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
152 * the directory entries. On Linux systems this value is already
153 * defined correctly as part of the /usr/include/dirent.h header file.
156 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
160 * The directory entry has the type (currently unused on Solaris) in the
161 * top 4 bits, and the object number in the low 48 bits. The "middle"
162 * 12 bits are unused.
164 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
165 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
168 * Directory entry locks control access to directory entries.
169 * They are used to protect creates, deletes, and renames.
170 * Each directory znode has a mutex and a list of locked names.
173 typedef struct zfs_dirlock
{
174 char *dl_name
; /* directory entry being locked */
175 uint32_t dl_sharecnt
; /* 0 if exclusive, > 0 if shared */
176 uint8_t dl_namelock
; /* 1 if z_name_lock is NOT held */
177 uint16_t dl_namesize
; /* set if dl_name was allocated */
178 kcondvar_t dl_cv
; /* wait for entry to be unlocked */
179 struct znode
*dl_dzp
; /* directory znode */
180 struct zfs_dirlock
*dl_next
; /* next in z_dirlocks list */
183 typedef struct znode
{
184 struct zfsvfs
*z_zfsvfs
;
186 uint64_t z_id
; /* object ID for this znode */
187 kmutex_t z_lock
; /* znode modification lock */
188 krwlock_t z_parent_lock
; /* parent lock for directories */
189 krwlock_t z_name_lock
; /* "master" lock for dirent locks */
190 zfs_dirlock_t
*z_dirlocks
; /* directory entry lock list */
191 kmutex_t z_range_lock
; /* protects changes to z_range_avl */
192 avl_tree_t z_range_avl
; /* avl tree of file range locks */
193 uint8_t z_unlinked
; /* file has been unlinked */
194 uint8_t z_atime_dirty
; /* atime needs to be synced */
195 uint8_t z_zn_prefetch
; /* Prefetch znodes? */
196 uint8_t z_moved
; /* Has this znode been moved? */
197 uint_t z_blksz
; /* block size in bytes */
198 uint_t z_seq
; /* modification sequence number */
199 uint64_t z_mapcnt
; /* number of pages mapped to file */
200 uint64_t z_gen
; /* generation (cached) */
201 uint64_t z_size
; /* file size (cached) */
202 uint64_t z_atime
[2]; /* atime (cached) */
203 uint64_t z_links
; /* file links (cached) */
204 uint64_t z_pflags
; /* pflags (cached) */
205 uint64_t z_uid
; /* uid fuid (cached) */
206 uint64_t z_gid
; /* gid fuid (cached) */
207 mode_t z_mode
; /* mode (cached) */
208 uint32_t z_sync_cnt
; /* synchronous open count */
209 kmutex_t z_acl_lock
; /* acl data lock */
210 zfs_acl_t
*z_acl_cached
; /* cached acl */
211 list_node_t z_link_node
; /* all znodes in fs link */
212 sa_handle_t
*z_sa_hdl
; /* handle to sa data */
213 boolean_t z_is_sa
; /* are we native sa? */
214 void (*z_set_ops_inode
) (struct inode
*); /* set inode ops */
215 struct inode z_inode
; /* generic vfs inode */
220 * Range locking rules
221 * --------------------
222 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
223 * file range needs to be locked as RL_WRITER. Only then can the pages be
224 * freed etc and zp_size reset. zp_size must be set within range lock.
225 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
226 * being written or freed needs to be locked as RL_WRITER.
227 * Multiple writes at the end of the file must coordinate zp_size updates
228 * to ensure data isn't lost. A compare and swap loop is currently used
229 * to ensure the file size is at least the offset last written.
230 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
231 * read needs to be locked as RL_READER. A check against zp_size can then
232 * be made for reading beyond end of file.
236 * Convert between znode pointers and inode pointers
238 #define ZTOI(ZP) (&((ZP)->z_inode))
239 #define ITOZ(IP) (container_of((IP), znode_t, z_inode))
241 /* XXX - REMOVE ME ONCE THE OTHER BUILD ISSUES ARE RESOLVED */
242 #define ZTOV(ZP) ((ZP)->z_vnode)
243 #define VTOZ(VP) ((znode_t *)(VP)->v_data)
247 * ZFS_ENTER() is called on entry to each ZFS inode and vfs operation.
248 * ZFS_EXIT() must be called before exitting the vop.
249 * ZFS_VERIFY_ZP() verifies the znode is valid.
251 #define ZFS_ENTER(zfsvfs) \
253 rrw_enter(&(zfsvfs)->z_teardown_lock, RW_READER, FTAG); \
254 if ((zfsvfs)->z_unmounted) { \
260 #define ZFS_EXIT(zfsvfs) rrw_exit(&(zfsvfs)->z_teardown_lock, FTAG)
262 #define ZFS_VERIFY_ZP(zp) \
263 if ((zp)->z_sa_hdl == NULL) { \
264 ZFS_EXIT((zp)->z_zfsvfs); \
269 * Macros for dealing with dmu_buf_hold
271 #define ZFS_OBJ_HASH(obj_num) ((obj_num) & (ZFS_OBJ_MTX_SZ - 1))
272 #define ZFS_OBJ_MUTEX(zfsvfs, obj_num) \
273 (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
274 #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \
275 mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
276 #define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \
277 mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
278 #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \
279 mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
282 * Macros to encode/decode ZFS stored time values from/to struct timespec
284 #define ZFS_TIME_ENCODE(tp, stmp) \
286 (stmp)[0] = (uint64_t)(tp)->tv_sec; \
287 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
290 #define ZFS_TIME_DECODE(tp, stmp) \
292 (tp)->tv_sec = (time_t)(stmp)[0]; \
293 (tp)->tv_nsec = (long)(stmp)[1]; \
299 #define ACCESSED (AT_ATIME)
300 #define STATE_CHANGED (AT_CTIME)
301 #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME)
303 #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \
304 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \
305 zfs_tstamp_update_setup(zp, ACCESSED, NULL, NULL, B_FALSE);
307 extern int zfs_init_fs(zfsvfs_t
*, znode_t
**);
308 extern void zfs_set_dataprop(objset_t
*);
309 extern void zfs_create_fs(objset_t
*os
, cred_t
*cr
, nvlist_t
*,
311 extern void zfs_tstamp_update_setup(znode_t
*, uint_t
, uint64_t [2],
312 uint64_t [2], boolean_t
);
313 extern void zfs_grow_blocksize(znode_t
*, uint64_t, dmu_tx_t
*);
314 extern int zfs_freesp(znode_t
*, uint64_t, uint64_t, int, boolean_t
);
315 extern void zfs_znode_init(void);
316 extern void zfs_znode_fini(void);
317 extern int zfs_zget(zfsvfs_t
*, uint64_t, znode_t
**);
318 extern int zfs_rezget(znode_t
*);
319 extern void zfs_zinactive(znode_t
*);
320 extern void zfs_znode_delete(znode_t
*, dmu_tx_t
*);
321 extern void zfs_znode_free(znode_t
*);
322 extern void zfs_remove_op_tables(void);
323 extern int zfs_create_op_tables(void);
324 extern int zfs_sync(vfs_t
*vfsp
, short flag
, cred_t
*cr
);
325 extern dev_t
zfs_cmpldev(uint64_t);
326 extern int zfs_get_zplprop(objset_t
*os
, zfs_prop_t prop
, uint64_t *value
);
327 extern int zfs_get_stats(objset_t
*os
, nvlist_t
*nv
);
328 extern void zfs_znode_dmu_fini(znode_t
*);
329 extern void zfs_inode_update(znode_t
*);
331 extern void zfs_log_create(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
332 znode_t
*dzp
, znode_t
*zp
, char *name
, vsecattr_t
*, zfs_fuid_info_t
*,
334 extern int zfs_log_create_txtype(zil_create_t
, vsecattr_t
*vsecp
,
336 extern void zfs_log_remove(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
337 znode_t
*dzp
, char *name
, uint64_t foid
);
338 #define ZFS_NO_OBJECT 0 /* no object id */
339 extern void zfs_log_link(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
340 znode_t
*dzp
, znode_t
*zp
, char *name
);
341 extern void zfs_log_symlink(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
342 znode_t
*dzp
, znode_t
*zp
, char *name
, char *link
);
343 extern void zfs_log_rename(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
344 znode_t
*sdzp
, char *sname
, znode_t
*tdzp
, char *dname
, znode_t
*szp
);
345 extern void zfs_log_write(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
346 znode_t
*zp
, offset_t off
, ssize_t len
, int ioflag
);
347 extern void zfs_log_truncate(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
348 znode_t
*zp
, uint64_t off
, uint64_t len
);
349 extern void zfs_log_setattr(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
350 znode_t
*zp
, vattr_t
*vap
, uint_t mask_applied
, zfs_fuid_info_t
*fuidp
);
351 extern void zfs_log_acl(zilog_t
*zilog
, dmu_tx_t
*tx
, znode_t
*zp
,
352 vsecattr_t
*vsecp
, zfs_fuid_info_t
*fuidp
);
353 extern void zfs_xvattr_set(znode_t
*zp
, xvattr_t
*xvap
, dmu_tx_t
*tx
);
354 extern void zfs_upgrade(zfsvfs_t
*zfsvfs
, dmu_tx_t
*tx
);
355 extern int zfs_create_share_dir(zfsvfs_t
*zfsvfs
, dmu_tx_t
*tx
);
357 #if defined(HAVE_UIO_RW)
358 extern caddr_t
zfs_map_page(page_t
*, enum seg_rw
);
359 extern void zfs_unmap_page(page_t
*, caddr_t
);
360 #endif /* HAVE_UIO_RW */
362 extern zil_get_data_t zfs_get_data
;
363 extern zil_replay_func_t
*zfs_replay_vector
[TX_MAX_TYPE
];
364 extern int zfsfstype
;
368 extern int zfs_obj_to_path(objset_t
*osp
, uint64_t obj
, char *buf
, int len
);
374 #endif /* _SYS_FS_ZFS_ZNODE_H */