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.
23 * Copyright (c) 2012, 2015 by Delphix. All rights reserved.
26 #ifndef _SYS_FS_ZFS_ZNODE_H
27 #define _SYS_FS_ZFS_ZNODE_H
30 #include <sys/isa_defs.h>
31 #include <sys/types32.h>
36 #include <sys/zfs_vfsops.h>
37 #include <sys/rrwlock.h>
38 #include <sys/zfs_sa.h>
39 #include <sys/zfs_stat.h>
40 #include <sys/zfs_rlock.h>
42 #include <sys/zfs_acl.h>
50 * Additional file level attributes, that are stored
51 * in the upper half of zp_flags
53 #define ZFS_READONLY 0x0000000100000000ull
54 #define ZFS_HIDDEN 0x0000000200000000ull
55 #define ZFS_SYSTEM 0x0000000400000000ull
56 #define ZFS_ARCHIVE 0x0000000800000000ull
57 #define ZFS_IMMUTABLE 0x0000001000000000ull
58 #define ZFS_NOUNLINK 0x0000002000000000ull
59 #define ZFS_APPENDONLY 0x0000004000000000ull
60 #define ZFS_NODUMP 0x0000008000000000ull
61 #define ZFS_OPAQUE 0x0000010000000000ull
62 #define ZFS_AV_QUARANTINED 0x0000020000000000ull
63 #define ZFS_AV_MODIFIED 0x0000040000000000ull
64 #define ZFS_REPARSE 0x0000080000000000ull
65 #define ZFS_OFFLINE 0x0000100000000000ull
66 #define ZFS_SPARSE 0x0000200000000000ull
68 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
74 VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
75 &pflags, sizeof (pflags), tx)); \
79 * Define special zfs pflags
81 #define ZFS_XATTR 0x1 /* is an extended attribute */
82 #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */
83 #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */
84 #define ZFS_ACL_OBJ_ACE 0x8 /* ACL has CMPLX Object ACE */
85 #define ZFS_ACL_PROTECTED 0x10 /* ACL protected */
86 #define ZFS_ACL_DEFAULTED 0x20 /* ACL should be defaulted */
87 #define ZFS_ACL_AUTO_INHERIT 0x40 /* ACL should be inherited */
88 #define ZFS_BONUS_SCANSTAMP 0x80 /* Scanstamp in bonus area */
89 #define ZFS_NO_EXECS_DENIED 0x100 /* exec was given to everyone */
91 #define SA_ZPL_ATIME(z) z->z_attr_table[ZPL_ATIME]
92 #define SA_ZPL_MTIME(z) z->z_attr_table[ZPL_MTIME]
93 #define SA_ZPL_CTIME(z) z->z_attr_table[ZPL_CTIME]
94 #define SA_ZPL_CRTIME(z) z->z_attr_table[ZPL_CRTIME]
95 #define SA_ZPL_GEN(z) z->z_attr_table[ZPL_GEN]
96 #define SA_ZPL_DACL_ACES(z) z->z_attr_table[ZPL_DACL_ACES]
97 #define SA_ZPL_XATTR(z) z->z_attr_table[ZPL_XATTR]
98 #define SA_ZPL_SYMLINK(z) z->z_attr_table[ZPL_SYMLINK]
99 #define SA_ZPL_RDEV(z) z->z_attr_table[ZPL_RDEV]
100 #define SA_ZPL_SCANSTAMP(z) z->z_attr_table[ZPL_SCANSTAMP]
101 #define SA_ZPL_UID(z) z->z_attr_table[ZPL_UID]
102 #define SA_ZPL_GID(z) z->z_attr_table[ZPL_GID]
103 #define SA_ZPL_PARENT(z) z->z_attr_table[ZPL_PARENT]
104 #define SA_ZPL_LINKS(z) z->z_attr_table[ZPL_LINKS]
105 #define SA_ZPL_MODE(z) z->z_attr_table[ZPL_MODE]
106 #define SA_ZPL_DACL_COUNT(z) z->z_attr_table[ZPL_DACL_COUNT]
107 #define SA_ZPL_FLAGS(z) z->z_attr_table[ZPL_FLAGS]
108 #define SA_ZPL_SIZE(z) z->z_attr_table[ZPL_SIZE]
109 #define SA_ZPL_ZNODE_ACL(z) z->z_attr_table[ZPL_ZNODE_ACL]
110 #define SA_ZPL_DXATTR(z) z->z_attr_table[ZPL_DXATTR]
111 #define SA_ZPL_PAD(z) z->z_attr_table[ZPL_PAD]
116 #define IS_EPHEMERAL(x) (x > MAXUID)
119 * Should we use FUIDs?
121 #define USE_FUIDS(version, os) (version >= ZPL_VERSION_FUID && \
122 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
123 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
124 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
126 #define MASTER_NODE_OBJ 1
129 * Special attributes for master node.
130 * "userquota@" and "groupquota@" are also valid (from
131 * zfs_userquota_prop_prefixes[]).
133 #define ZFS_FSID "FSID"
134 #define ZFS_UNLINKED_SET "DELETE_QUEUE"
135 #define ZFS_ROOT_OBJ "ROOT"
136 #define ZPL_VERSION_STR "VERSION"
137 #define ZFS_FUID_TABLES "FUID"
138 #define ZFS_SHARES_DIR "SHARES"
139 #define ZFS_SA_ATTRS "SA_ATTRS"
142 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
143 * the directory entries. On Linux systems this value is already
144 * defined correctly as part of the /usr/include/dirent.h header file.
147 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
151 * The directory entry has the type (currently unused on Solaris) in the
152 * top 4 bits, and the object number in the low 48 bits. The "middle"
153 * 12 bits are unused.
155 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
156 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
159 * Directory entry locks control access to directory entries.
160 * They are used to protect creates, deletes, and renames.
161 * Each directory znode has a mutex and a list of locked names.
164 typedef struct zfs_dirlock
{
165 char *dl_name
; /* directory entry being locked */
166 uint32_t dl_sharecnt
; /* 0 if exclusive, > 0 if shared */
167 uint8_t dl_namelock
; /* 1 if z_name_lock is NOT held */
168 uint16_t dl_namesize
; /* set if dl_name was allocated */
169 kcondvar_t dl_cv
; /* wait for entry to be unlocked */
170 struct znode
*dl_dzp
; /* directory znode */
171 struct zfs_dirlock
*dl_next
; /* next in z_dirlocks list */
174 typedef struct znode
{
175 uint64_t z_id
; /* object ID for this znode */
176 kmutex_t z_lock
; /* znode modification lock */
177 krwlock_t z_parent_lock
; /* parent lock for directories */
178 krwlock_t z_name_lock
; /* "master" lock for dirent locks */
179 zfs_dirlock_t
*z_dirlocks
; /* directory entry lock list */
180 zfs_rlock_t z_range_lock
; /* file range lock */
181 uint8_t z_unlinked
; /* file has been unlinked */
182 uint8_t z_atime_dirty
; /* atime needs to be synced */
183 uint8_t z_zn_prefetch
; /* Prefetch znodes? */
184 uint8_t z_moved
; /* Has this znode been moved? */
185 uint_t z_blksz
; /* block size in bytes */
186 uint_t z_seq
; /* modification sequence number */
187 uint64_t z_mapcnt
; /* number of pages mapped to file */
188 uint64_t z_dnodesize
; /* dnode size */
189 uint64_t z_size
; /* file size (cached) */
190 uint64_t z_pflags
; /* pflags (cached) */
191 uint32_t z_sync_cnt
; /* synchronous open count */
192 mode_t z_mode
; /* mode (cached) */
193 kmutex_t z_acl_lock
; /* acl data lock */
194 zfs_acl_t
*z_acl_cached
; /* cached acl */
195 krwlock_t z_xattr_lock
; /* xattr data lock */
196 nvlist_t
*z_xattr_cached
; /* cached xattrs */
197 list_node_t z_link_node
; /* all znodes in fs link */
198 sa_handle_t
*z_sa_hdl
; /* handle to sa data */
199 boolean_t z_is_sa
; /* are we native sa? */
200 boolean_t z_is_mapped
; /* are we mmap'ed */
201 boolean_t z_is_ctldir
; /* are we .zfs entry */
202 boolean_t z_is_stale
; /* are we stale due to rollback? */
203 struct inode z_inode
; /* generic vfs inode */
206 typedef struct znode_hold
{
207 uint64_t zh_obj
; /* object id */
208 kmutex_t zh_lock
; /* lock serializing object access */
209 avl_node_t zh_node
; /* avl tree linkage */
210 refcount_t zh_refcount
; /* active consumer reference count */
214 * Range locking rules
215 * --------------------
216 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
217 * file range needs to be locked as RL_WRITER. Only then can the pages be
218 * freed etc and zp_size reset. zp_size must be set within range lock.
219 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
220 * being written or freed needs to be locked as RL_WRITER.
221 * Multiple writes at the end of the file must coordinate zp_size updates
222 * to ensure data isn't lost. A compare and swap loop is currently used
223 * to ensure the file size is at least the offset last written.
224 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
225 * read needs to be locked as RL_READER. A check against zp_size can then
226 * be made for reading beyond end of file.
230 * Convert between znode pointers and inode pointers
232 #define ZTOI(znode) (&((znode)->z_inode))
233 #define ITOZ(inode) (container_of((inode), znode_t, z_inode))
234 #define ZTOZSB(znode) ((zfs_sb_t *)(ZTOI(znode)->i_sb->s_fs_info))
235 #define ITOZSB(inode) ((zfs_sb_t *)((inode)->i_sb->s_fs_info))
237 #define S_ISDEV(mode) (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
239 /* Called on entry to each ZFS vnode and vfs operation */
240 #define ZFS_ENTER(zsb) \
242 rrm_enter_read(&(zsb)->z_teardown_lock, FTAG); \
243 if ((zsb)->z_unmounted) { \
249 /* Must be called before exiting the vop */
250 #define ZFS_EXIT(zsb) \
252 rrm_exit(&(zsb)->z_teardown_lock, FTAG); \
255 /* Verifies the znode is valid */
256 #define ZFS_VERIFY_ZP(zp) \
257 if ((zp)->z_sa_hdl == NULL) { \
258 ZFS_EXIT(ZTOZSB(zp)); \
263 * Macros for dealing with dmu_buf_hold
265 #define ZFS_OBJ_MTX_SZ 64
266 #define ZFS_OBJ_MTX_MAX (1024 * 1024)
267 #define ZFS_OBJ_HASH(zsb, obj) ((obj) & ((zsb->z_hold_size) - 1))
269 extern unsigned int zfs_object_mutex_size
;
271 /* Encode ZFS stored time values from a struct timespec */
272 #define ZFS_TIME_ENCODE(tp, stmp) \
274 (stmp)[0] = (uint64_t)(tp)->tv_sec; \
275 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
278 /* Decode ZFS stored time values to a struct timespec */
279 #define ZFS_TIME_DECODE(tp, stmp) \
281 (tp)->tv_sec = (time_t)(stmp)[0]; \
282 (tp)->tv_nsec = (long)(stmp)[1]; \
288 #define ACCESSED (ATTR_ATIME)
289 #define STATE_CHANGED (ATTR_CTIME)
290 #define CONTENT_MODIFIED (ATTR_MTIME | ATTR_CTIME)
292 extern int zfs_init_fs(zfs_sb_t
*, znode_t
**);
293 extern void zfs_set_dataprop(objset_t
*);
294 extern void zfs_create_fs(objset_t
*os
, cred_t
*cr
, nvlist_t
*,
296 extern void zfs_tstamp_update_setup(znode_t
*, uint_t
, uint64_t [2],
298 extern void zfs_grow_blocksize(znode_t
*, uint64_t, dmu_tx_t
*);
299 extern int zfs_freesp(znode_t
*, uint64_t, uint64_t, int, boolean_t
);
300 extern void zfs_znode_init(void);
301 extern void zfs_znode_fini(void);
302 extern int zfs_znode_hold_compare(const void *, const void *);
303 extern int zfs_zget(zfs_sb_t
*, uint64_t, znode_t
**);
304 extern int zfs_rezget(znode_t
*);
305 extern void zfs_zinactive(znode_t
*);
306 extern void zfs_znode_delete(znode_t
*, dmu_tx_t
*);
307 extern void zfs_remove_op_tables(void);
308 extern int zfs_create_op_tables(void);
309 extern int zfs_sync(struct super_block
*, int, cred_t
*);
310 extern dev_t
zfs_cmpldev(uint64_t);
311 extern int zfs_get_zplprop(objset_t
*os
, zfs_prop_t prop
, uint64_t *value
);
312 extern int zfs_get_stats(objset_t
*os
, nvlist_t
*nv
);
313 extern void zfs_znode_dmu_fini(znode_t
*);
314 extern int zfs_inode_alloc(struct super_block
*, struct inode
**ip
);
315 extern void zfs_inode_destroy(struct inode
*);
316 extern void zfs_inode_update(znode_t
*);
317 extern void zfs_mark_inode_dirty(struct inode
*);
319 extern void zfs_log_create(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
320 znode_t
*dzp
, znode_t
*zp
, char *name
, vsecattr_t
*, zfs_fuid_info_t
*,
322 extern int zfs_log_create_txtype(zil_create_t
, vsecattr_t
*vsecp
,
324 extern void zfs_log_remove(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
325 znode_t
*dzp
, char *name
, uint64_t foid
);
326 #define ZFS_NO_OBJECT 0 /* no object id */
327 extern void zfs_log_link(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
328 znode_t
*dzp
, znode_t
*zp
, char *name
);
329 extern void zfs_log_symlink(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
330 znode_t
*dzp
, znode_t
*zp
, char *name
, char *link
);
331 extern void zfs_log_rename(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
332 znode_t
*sdzp
, char *sname
, znode_t
*tdzp
, char *dname
, znode_t
*szp
);
333 extern void zfs_log_write(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
334 znode_t
*zp
, offset_t off
, ssize_t len
, int ioflag
,
335 zil_callback_t callback
, void *callback_data
);
336 extern void zfs_log_truncate(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
337 znode_t
*zp
, uint64_t off
, uint64_t len
);
338 extern void zfs_log_setattr(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
339 znode_t
*zp
, vattr_t
*vap
, uint_t mask_applied
, zfs_fuid_info_t
*fuidp
);
340 extern void zfs_log_acl(zilog_t
*zilog
, dmu_tx_t
*tx
, znode_t
*zp
,
341 vsecattr_t
*vsecp
, zfs_fuid_info_t
*fuidp
);
342 extern void zfs_xvattr_set(znode_t
*zp
, xvattr_t
*xvap
, dmu_tx_t
*tx
);
343 extern void zfs_upgrade(zfs_sb_t
*zsb
, dmu_tx_t
*tx
);
344 extern int zfs_create_share_dir(zfs_sb_t
*zsb
, dmu_tx_t
*tx
);
346 #if defined(HAVE_UIO_RW)
347 extern caddr_t
zfs_map_page(page_t
*, enum seg_rw
);
348 extern void zfs_unmap_page(page_t
*, caddr_t
);
349 #endif /* HAVE_UIO_RW */
351 extern zil_get_data_t zfs_get_data
;
352 extern zil_replay_func_t zfs_replay_vector
[TX_MAX_TYPE
];
353 extern int zfsfstype
;
357 extern int zfs_obj_to_path(objset_t
*osp
, uint64_t obj
, char *buf
, int len
);
363 #endif /* _SYS_FS_ZFS_ZNODE_H */