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 2007 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
29 #pragma ident "%Z%%M% %I% %E% SMI"
34 #include <sys/zfs_vfsops.h>
36 #include <sys/zfs_acl.h>
44 * Define special zfs pflags
46 #define ZFS_XATTR 0x1 /* is an extended attribute */
47 #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */
48 #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */
50 #define MASTER_NODE_OBJ 1
53 * special attributes for master node.
56 #define ZFS_FSID "FSID"
57 #define ZFS_UNLINKED_SET "DELETE_QUEUE"
58 #define ZFS_ROOT_OBJ "ROOT"
59 #define ZPL_VERSION_OBJ "VERSION"
60 #define ZFS_PROP_BLOCKPERPAGE "BLOCKPERPAGE"
61 #define ZFS_PROP_NOGROWBLOCKS "NOGROWBLOCKS"
63 #define ZFS_FLAG_BLOCKPERPAGE 0x1
64 #define ZFS_FLAG_NOGROWBLOCKS 0x2
67 * ZPL version - rev'd whenever an incompatible on-disk format change
68 * occurs. Independent of SPA/DMU/ZAP versioning.
71 #define ZPL_VERSION 1ULL
73 #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE)
75 /* Path component length */
77 * The generic fs code uses MAXNAMELEN to represent
78 * what the largest component length is. Unfortunately,
79 * this length includes the terminating NULL. ZFS needs
80 * to tell the users via pathconf() and statvfs() what the
81 * true maximum length of a component is, excluding the NULL.
83 #define ZFS_MAXNAMELEN (MAXNAMELEN - 1)
86 * The directory entry has the type (currently unused on Solaris) in the
87 * top 4 bits, and the object number in the low 48 bits. The "middle"
90 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
91 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
92 #define ZFS_DIRENT_MAKE(type, obj) (((uint64_t)type << 60) | obj)
96 * This is the persistent portion of the znode. It is stored
97 * in the "bonus buffer" of the file. Short symbolic links
98 * are also stored in the bonus buffer.
100 typedef struct znode_phys {
101 uint64_t zp_atime[2]; /* 0 - last file access time */
102 uint64_t zp_mtime[2]; /* 16 - last file modification time */
103 uint64_t zp_ctime[2]; /* 32 - last file change time */
104 uint64_t zp_crtime[2]; /* 48 - creation time */
105 uint64_t zp_gen; /* 64 - generation (txg of creation) */
106 uint64_t zp_mode; /* 72 - file mode bits */
107 uint64_t zp_size; /* 80 - size of file */
108 uint64_t zp_parent; /* 88 - directory parent (`..') */
109 uint64_t zp_links; /* 96 - number of links to file */
110 uint64_t zp_xattr; /* 104 - DMU object for xattrs */
111 uint64_t zp_rdev; /* 112 - dev_t for VBLK & VCHR files */
112 uint64_t zp_flags; /* 120 - persistent flags */
113 uint64_t zp_uid; /* 128 - file owner */
114 uint64_t zp_gid; /* 136 - owning group */
115 uint64_t zp_pad[4]; /* 144 - future */
116 zfs_znode_acl_t zp_acl; /* 176 - 263 ACL */
118 * Data may pad out any remaining bytes in the znode buffer, eg:
120 * |<---------------------- dnode_phys (512) ------------------------>|
121 * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->|
122 * |<---- znode (264) ---->|<---- data (56) ---->|
124 * At present, we only use this space to store symbolic links.
129 * Directory entry locks control access to directory entries.
130 * They are used to protect creates, deletes, and renames.
131 * Each directory znode has a mutex and a list of locked names.
134 typedef struct zfs_dirlock {
135 char *dl_name; /* directory entry being locked */
136 uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */
137 uint16_t dl_namesize; /* set if dl_name was allocated */
138 kcondvar_t dl_cv; /* wait for entry to be unlocked */
139 struct znode *dl_dzp; /* directory znode */
140 struct zfs_dirlock *dl_next; /* next in z_dirlocks list */
143 typedef struct znode {
144 struct zfsvfs *z_zfsvfs;
146 uint64_t z_id; /* object ID for this znode */
147 kmutex_t z_lock; /* znode modification lock */
148 krwlock_t z_map_lock; /* page map lock */
149 krwlock_t z_parent_lock; /* parent lock for directories */
150 krwlock_t z_name_lock; /* "master" lock for dirent locks */
151 zfs_dirlock_t *z_dirlocks; /* directory entry lock list */
152 kmutex_t z_range_lock; /* protects changes to z_range_avl */
153 avl_tree_t z_range_avl; /* avl tree of file range locks */
154 uint8_t z_unlinked; /* file has been unlinked */
155 uint8_t z_atime_dirty; /* atime needs to be synced */
156 uint8_t z_dbuf_held; /* Is z_dbuf already held? */
157 uint8_t z_zn_prefetch; /* Prefetch znodes? */
158 uint_t z_blksz; /* block size in bytes */
159 uint_t z_seq; /* modification sequence number */
160 uint64_t z_mapcnt; /* number of pages mapped to file */
161 uint64_t z_last_itx; /* last ZIL itx on this znode */
162 uint32_t z_sync_cnt; /* synchronous open count */
163 kmutex_t z_acl_lock; /* acl data lock */
164 list_node_t z_link_node; /* all znodes in fs link */
166 * These are dmu managed fields.
168 znode_phys_t *z_phys; /* pointer to persistent znode */
169 dmu_buf_t *z_dbuf; /* buffer containing the z_phys */
174 * Range locking rules
175 * --------------------
176 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
177 * file range needs to be locked as RL_WRITER. Only then can the pages be
178 * freed etc and zp_size reset. zp_size must be set within range lock.
179 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
180 * being written or freed needs to be locked as RL_WRITER.
181 * Multiple writes at the end of the file must coordinate zp_size updates
182 * to ensure data isn't lost. A compare and swap loop is currently used
183 * to ensure the file size is at least the offset last written.
184 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
185 * read needs to be locked as RL_READER. A check against zp_size can then
186 * be made for reading beyond end of file.
190 * Convert between znode pointers and vnode pointers
192 #define ZTOV(ZP) ((ZP)->z_vnode)
193 #define VTOZ(VP) ((znode_t *)(VP)->v_data)
196 * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation.
197 * ZFS_EXIT() must be called before exitting the vop.
199 #define ZFS_ENTER(zfsvfs) \
201 atomic_add_32(&(zfsvfs)->z_op_cnt, 1); \
202 if ((zfsvfs)->z_unmounted1) { \
207 #define ZFS_EXIT(zfsvfs) atomic_add_32(&(zfsvfs)->z_op_cnt, -1)
210 * Macros for dealing with dmu_buf_hold
212 #define ZFS_OBJ_HASH(obj_num) (obj_num & (ZFS_OBJ_MTX_SZ - 1))
213 #define ZFS_OBJ_MUTEX(zp) \
214 (&zp->z_zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(zp->z_id)])
215 #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \
216 mutex_enter(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]);
218 #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \
219 mutex_exit(&zfsvfs->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
222 * Macros to encode/decode ZFS stored time values from/to struct timespec
224 #define ZFS_TIME_ENCODE(tp, stmp) \
226 stmp[0] = (uint64_t)(tp)->tv_sec; \
227 stmp[1] = (uint64_t)(tp)->tv_nsec; \
230 #define ZFS_TIME_DECODE(tp, stmp) \
232 (tp)->tv_sec = (time_t)stmp[0]; \
233 (tp)->tv_nsec = (long)stmp[1]; \
239 #define ACCESSED (AT_ATIME)
240 #define STATE_CHANGED (AT_CTIME)
241 #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME)
243 #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \
244 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \
245 zfs_time_stamper(zp, ACCESSED, NULL)
247 extern int zfs_init_fs(zfsvfs_t *, znode_t **, cred_t *);
248 extern void zfs_set_dataprop(objset_t *);
249 extern void zfs_create_fs(objset_t *os, cred_t *cr, dmu_tx_t *tx);
250 extern void zfs_time_stamper(znode_t *, uint_t, dmu_tx_t *);
251 extern void zfs_time_stamper_locked(znode_t *, uint_t, dmu_tx_t *);
252 extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
253 extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
254 extern void zfs_znode_init(void);
255 extern void zfs_znode_fini(void);
256 extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
257 extern void zfs_zinactive(znode_t *);
258 extern void zfs_znode_delete(znode_t *, dmu_tx_t *);
259 extern void zfs_znode_free(znode_t *);
260 extern void zfs_remove_op_tables();
261 extern int zfs_create_op_tables();
262 extern dev_t zfs_cmpldev(uint64_t);
264 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, int txtype,
265 znode_t *dzp, znode_t *zp, char *name);
266 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, int txtype,
267 znode_t *dzp, char *name);
268 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, int txtype,
269 znode_t *dzp, znode_t *zp, char *name);
270 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, int txtype,
271 znode_t *dzp, znode_t *zp, char *name, char *link);
272 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, int txtype,
273 znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
274 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
275 znode_t *zp, offset_t off, ssize_t len, int ioflag);
276 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
277 znode_t *zp, uint64_t off, uint64_t len);
278 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
279 znode_t *zp, vattr_t *vap, uint_t mask_applied);
281 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, int txtype,
282 znode_t *zp, int aclcnt, ace_t *z_ace);
285 extern zil_get_data_t zfs_get_data;
286 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
287 extern int zfsfstype;
291 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
297 #endif /* _SYS_FS_ZFS_ZNODE_H */