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.
24 * Copyright 2016 Nexenta Systems, Inc. All rights reserved.
27 #ifndef _SYS_FS_ZFS_ZNODE_H
28 #define _SYS_FS_ZFS_ZNODE_H
31 #include <sys/isa_defs.h>
32 #include <sys/types32.h>
37 #include <sys/zfs_vfsops.h>
38 #include <sys/rrwlock.h>
39 #include <sys/zfs_sa.h>
40 #include <sys/zfs_stat.h>
41 #include <sys/zfs_rlock.h>
43 #include <sys/zfs_acl.h>
51 * Additional file level attributes, that are stored
52 * in the upper half of zp_flags
54 #define ZFS_READONLY 0x0000000100000000ull
55 #define ZFS_HIDDEN 0x0000000200000000ull
56 #define ZFS_SYSTEM 0x0000000400000000ull
57 #define ZFS_ARCHIVE 0x0000000800000000ull
58 #define ZFS_IMMUTABLE 0x0000001000000000ull
59 #define ZFS_NOUNLINK 0x0000002000000000ull
60 #define ZFS_APPENDONLY 0x0000004000000000ull
61 #define ZFS_NODUMP 0x0000008000000000ull
62 #define ZFS_OPAQUE 0x0000010000000000ull
63 #define ZFS_AV_QUARANTINED 0x0000020000000000ull
64 #define ZFS_AV_MODIFIED 0x0000040000000000ull
65 #define ZFS_REPARSE 0x0000080000000000ull
66 #define ZFS_OFFLINE 0x0000100000000000ull
67 #define ZFS_SPARSE 0x0000200000000000ull
69 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
75 VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
76 &pflags, sizeof (pflags), tx)); \
80 * Define special zfs pflags
82 #define ZFS_XATTR 0x1 /* is an extended attribute */
83 #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */
84 #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */
85 #define ZFS_ACL_OBJ_ACE 0x8 /* ACL has CMPLX Object ACE */
86 #define ZFS_ACL_PROTECTED 0x10 /* ACL protected */
87 #define ZFS_ACL_DEFAULTED 0x20 /* ACL should be defaulted */
88 #define ZFS_ACL_AUTO_INHERIT 0x40 /* ACL should be inherited */
89 #define ZFS_BONUS_SCANSTAMP 0x80 /* Scanstamp in bonus area */
90 #define ZFS_NO_EXECS_DENIED 0x100 /* exec was given to everyone */
92 #define SA_ZPL_ATIME(z) z->z_attr_table[ZPL_ATIME]
93 #define SA_ZPL_MTIME(z) z->z_attr_table[ZPL_MTIME]
94 #define SA_ZPL_CTIME(z) z->z_attr_table[ZPL_CTIME]
95 #define SA_ZPL_CRTIME(z) z->z_attr_table[ZPL_CRTIME]
96 #define SA_ZPL_GEN(z) z->z_attr_table[ZPL_GEN]
97 #define SA_ZPL_DACL_ACES(z) z->z_attr_table[ZPL_DACL_ACES]
98 #define SA_ZPL_XATTR(z) z->z_attr_table[ZPL_XATTR]
99 #define SA_ZPL_SYMLINK(z) z->z_attr_table[ZPL_SYMLINK]
100 #define SA_ZPL_RDEV(z) z->z_attr_table[ZPL_RDEV]
101 #define SA_ZPL_SCANSTAMP(z) z->z_attr_table[ZPL_SCANSTAMP]
102 #define SA_ZPL_UID(z) z->z_attr_table[ZPL_UID]
103 #define SA_ZPL_GID(z) z->z_attr_table[ZPL_GID]
104 #define SA_ZPL_PARENT(z) z->z_attr_table[ZPL_PARENT]
105 #define SA_ZPL_LINKS(z) z->z_attr_table[ZPL_LINKS]
106 #define SA_ZPL_MODE(z) z->z_attr_table[ZPL_MODE]
107 #define SA_ZPL_DACL_COUNT(z) z->z_attr_table[ZPL_DACL_COUNT]
108 #define SA_ZPL_FLAGS(z) z->z_attr_table[ZPL_FLAGS]
109 #define SA_ZPL_SIZE(z) z->z_attr_table[ZPL_SIZE]
110 #define SA_ZPL_ZNODE_ACL(z) z->z_attr_table[ZPL_ZNODE_ACL]
111 #define SA_ZPL_DXATTR(z) z->z_attr_table[ZPL_DXATTR]
112 #define SA_ZPL_PAD(z) z->z_attr_table[ZPL_PAD]
117 #define IS_EPHEMERAL(x) (x > MAXUID)
120 * Should we use FUIDs?
122 #define USE_FUIDS(version, os) (version >= ZPL_VERSION_FUID && \
123 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
124 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
125 spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
127 #define MASTER_NODE_OBJ 1
130 * Special attributes for master node.
131 * "userquota@" and "groupquota@" are also valid (from
132 * zfs_userquota_prop_prefixes[]).
134 #define ZFS_FSID "FSID"
135 #define ZFS_UNLINKED_SET "DELETE_QUEUE"
136 #define ZFS_ROOT_OBJ "ROOT"
137 #define ZPL_VERSION_STR "VERSION"
138 #define ZFS_FUID_TABLES "FUID"
139 #define ZFS_SHARES_DIR "SHARES"
140 #define ZFS_SA_ATTRS "SA_ATTRS"
143 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
144 * the directory entries. On Linux systems this value is already
145 * defined correctly as part of the /usr/include/dirent.h header file.
148 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
152 * The directory entry has the type (currently unused on Solaris) in the
153 * top 4 bits, and the object number in the low 48 bits. The "middle"
154 * 12 bits are unused.
156 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
157 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
160 * Directory entry locks control access to directory entries.
161 * They are used to protect creates, deletes, and renames.
162 * Each directory znode has a mutex and a list of locked names.
165 typedef struct zfs_dirlock
{
166 char *dl_name
; /* directory entry being locked */
167 uint32_t dl_sharecnt
; /* 0 if exclusive, > 0 if shared */
168 uint8_t dl_namelock
; /* 1 if z_name_lock is NOT held */
169 uint16_t dl_namesize
; /* set if dl_name was allocated */
170 kcondvar_t dl_cv
; /* wait for entry to be unlocked */
171 struct znode
*dl_dzp
; /* directory znode */
172 struct zfs_dirlock
*dl_next
; /* next in z_dirlocks list */
175 typedef struct znode
{
176 uint64_t z_id
; /* object ID for this znode */
177 kmutex_t z_lock
; /* znode modification lock */
178 krwlock_t z_parent_lock
; /* parent lock for directories */
179 krwlock_t z_name_lock
; /* "master" lock for dirent locks */
180 zfs_dirlock_t
*z_dirlocks
; /* directory entry lock list */
181 zfs_rlock_t z_range_lock
; /* file range lock */
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_dnodesize
; /* dnode size */
190 uint64_t z_size
; /* file size (cached) */
191 uint64_t z_pflags
; /* pflags (cached) */
192 uint32_t z_sync_cnt
; /* synchronous open count */
193 mode_t z_mode
; /* mode (cached) */
194 kmutex_t z_acl_lock
; /* acl data lock */
195 zfs_acl_t
*z_acl_cached
; /* cached acl */
196 krwlock_t z_xattr_lock
; /* xattr data lock */
197 nvlist_t
*z_xattr_cached
; /* cached xattrs */
198 uint64_t z_xattr_parent
; /* parent obj for this xattr */
199 list_node_t z_link_node
; /* all znodes in fs link */
200 sa_handle_t
*z_sa_hdl
; /* handle to sa data */
201 boolean_t z_is_sa
; /* are we native sa? */
202 boolean_t z_is_mapped
; /* are we mmap'ed */
203 boolean_t z_is_ctldir
; /* are we .zfs entry */
204 boolean_t z_is_stale
; /* are we stale due to rollback? */
205 struct inode z_inode
; /* generic vfs inode */
208 typedef struct znode_hold
{
209 uint64_t zh_obj
; /* object id */
210 kmutex_t zh_lock
; /* lock serializing object access */
211 avl_node_t zh_node
; /* avl tree linkage */
212 refcount_t zh_refcount
; /* active consumer reference count */
216 * Range locking rules
217 * --------------------
218 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
219 * file range needs to be locked as RL_WRITER. Only then can the pages be
220 * freed etc and zp_size reset. zp_size must be set within range lock.
221 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
222 * being written or freed needs to be locked as RL_WRITER.
223 * Multiple writes at the end of the file must coordinate zp_size updates
224 * to ensure data isn't lost. A compare and swap loop is currently used
225 * to ensure the file size is at least the offset last written.
226 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
227 * read needs to be locked as RL_READER. A check against zp_size can then
228 * be made for reading beyond end of file.
232 * Convert between znode pointers and inode pointers
234 #define ZTOI(znode) (&((znode)->z_inode))
235 #define ITOZ(inode) (container_of((inode), znode_t, z_inode))
236 #define ZTOZSB(znode) ((zfsvfs_t *)(ZTOI(znode)->i_sb->s_fs_info))
237 #define ITOZSB(inode) ((zfsvfs_t *)((inode)->i_sb->s_fs_info))
239 #define S_ISDEV(mode) (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
241 /* Called on entry to each ZFS vnode and vfs operation */
242 #define ZFS_ENTER(zfsvfs) \
244 rrm_enter_read(&(zfsvfs)->z_teardown_lock, FTAG); \
245 if ((zfsvfs)->z_unmounted) { \
251 /* Must be called before exiting the vop */
252 #define ZFS_EXIT(zfsvfs) \
254 rrm_exit(&(zfsvfs)->z_teardown_lock, FTAG); \
257 /* Verifies the znode is valid */
258 #define ZFS_VERIFY_ZP(zp) \
259 if ((zp)->z_sa_hdl == NULL) { \
260 ZFS_EXIT(ZTOZSB(zp)); \
265 * Macros for dealing with dmu_buf_hold
267 #define ZFS_OBJ_MTX_SZ 64
268 #define ZFS_OBJ_MTX_MAX (1024 * 1024)
269 #define ZFS_OBJ_HASH(zfsvfs, obj) ((obj) & ((zfsvfs->z_hold_size) - 1))
271 extern unsigned int zfs_object_mutex_size
;
273 /* Encode ZFS stored time values from a struct timespec */
274 #define ZFS_TIME_ENCODE(tp, stmp) \
276 (stmp)[0] = (uint64_t)(tp)->tv_sec; \
277 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
280 /* Decode ZFS stored time values to a struct timespec */
281 #define ZFS_TIME_DECODE(tp, stmp) \
283 (tp)->tv_sec = (time_t)(stmp)[0]; \
284 (tp)->tv_nsec = (long)(stmp)[1]; \
290 #define ACCESSED (ATTR_ATIME)
291 #define STATE_CHANGED (ATTR_CTIME)
292 #define CONTENT_MODIFIED (ATTR_MTIME | ATTR_CTIME)
294 extern int zfs_init_fs(zfsvfs_t
*, znode_t
**);
295 extern void zfs_set_dataprop(objset_t
*);
296 extern void zfs_create_fs(objset_t
*os
, cred_t
*cr
, nvlist_t
*,
298 extern void zfs_tstamp_update_setup(znode_t
*, uint_t
, uint64_t [2],
300 extern void zfs_grow_blocksize(znode_t
*, uint64_t, dmu_tx_t
*);
301 extern int zfs_freesp(znode_t
*, uint64_t, uint64_t, int, boolean_t
);
302 extern void zfs_znode_init(void);
303 extern void zfs_znode_fini(void);
304 extern int zfs_znode_hold_compare(const void *, const void *);
305 extern int zfs_zget(zfsvfs_t
*, uint64_t, znode_t
**);
306 extern int zfs_rezget(znode_t
*);
307 extern void zfs_zinactive(znode_t
*);
308 extern void zfs_znode_delete(znode_t
*, dmu_tx_t
*);
309 extern void zfs_remove_op_tables(void);
310 extern int zfs_create_op_tables(void);
311 extern int zfs_sync(struct super_block
*, int, cred_t
*);
312 extern dev_t
zfs_cmpldev(uint64_t);
313 extern int zfs_get_zplprop(objset_t
*os
, zfs_prop_t prop
, uint64_t *value
);
314 extern int zfs_get_stats(objset_t
*os
, nvlist_t
*nv
);
315 extern boolean_t
zfs_get_vfs_flag_unmounted(objset_t
*os
);
316 extern void zfs_znode_dmu_fini(znode_t
*);
317 extern int zfs_inode_alloc(struct super_block
*, struct inode
**ip
);
318 extern void zfs_inode_destroy(struct inode
*);
319 extern void zfs_inode_update(znode_t
*);
320 extern void zfs_mark_inode_dirty(struct inode
*);
322 extern void zfs_log_create(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
323 znode_t
*dzp
, znode_t
*zp
, char *name
, vsecattr_t
*, zfs_fuid_info_t
*,
325 extern int zfs_log_create_txtype(zil_create_t
, vsecattr_t
*vsecp
,
327 extern void zfs_log_remove(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
328 znode_t
*dzp
, char *name
, uint64_t foid
);
329 #define ZFS_NO_OBJECT 0 /* no object id */
330 extern void zfs_log_link(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
331 znode_t
*dzp
, znode_t
*zp
, char *name
);
332 extern void zfs_log_symlink(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
333 znode_t
*dzp
, znode_t
*zp
, char *name
, char *link
);
334 extern void zfs_log_rename(zilog_t
*zilog
, dmu_tx_t
*tx
, uint64_t txtype
,
335 znode_t
*sdzp
, char *sname
, znode_t
*tdzp
, char *dname
, znode_t
*szp
);
336 extern void zfs_log_write(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
337 znode_t
*zp
, offset_t off
, ssize_t len
, int ioflag
,
338 zil_callback_t callback
, void *callback_data
);
339 extern void zfs_log_truncate(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
340 znode_t
*zp
, uint64_t off
, uint64_t len
);
341 extern void zfs_log_setattr(zilog_t
*zilog
, dmu_tx_t
*tx
, int txtype
,
342 znode_t
*zp
, vattr_t
*vap
, uint_t mask_applied
, zfs_fuid_info_t
*fuidp
);
343 extern void zfs_log_acl(zilog_t
*zilog
, dmu_tx_t
*tx
, znode_t
*zp
,
344 vsecattr_t
*vsecp
, zfs_fuid_info_t
*fuidp
);
345 extern void zfs_xvattr_set(znode_t
*zp
, xvattr_t
*xvap
, dmu_tx_t
*tx
);
346 extern void zfs_upgrade(zfsvfs_t
*zfsvfs
, dmu_tx_t
*tx
);
347 extern int zfs_create_share_dir(zfsvfs_t
*zfsvfs
, dmu_tx_t
*tx
);
349 #if defined(HAVE_UIO_RW)
350 extern caddr_t
zfs_map_page(page_t
*, enum seg_rw
);
351 extern void zfs_unmap_page(page_t
*, caddr_t
);
352 #endif /* HAVE_UIO_RW */
354 extern zil_get_data_t zfs_get_data
;
355 extern zil_replay_func_t zfs_replay_vector
[TX_MAX_TYPE
];
356 extern int zfsfstype
;
360 extern int zfs_obj_to_path(objset_t
*osp
, uint64_t obj
, char *buf
, int len
);
366 #endif /* _SYS_FS_ZFS_ZNODE_H */