]>
Commit | Line | Data |
---|---|---|
34dc7c2f BB |
1 | /* |
2 | * CDDL HEADER START | |
3 | * | |
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. | |
7 | * | |
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. | |
12 | * | |
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] | |
18 | * | |
19 | * CDDL HEADER END | |
20 | */ | |
21 | /* | |
d164b209 | 22 | * Copyright 2009 Sun Microsystems, Inc. All rights reserved. |
34dc7c2f BB |
23 | * Use is subject to license terms. |
24 | */ | |
25 | ||
26 | #ifndef _SYS_FS_ZFS_ZNODE_H | |
27 | #define _SYS_FS_ZFS_ZNODE_H | |
28 | ||
34dc7c2f BB |
29 | #ifdef _KERNEL |
30 | #include <sys/isa_defs.h> | |
31 | #include <sys/types32.h> | |
32 | #include <sys/attr.h> | |
33 | #include <sys/list.h> | |
34 | #include <sys/dmu.h> | |
35 | #include <sys/zfs_vfsops.h> | |
36 | #include <sys/rrwlock.h> | |
37 | #endif | |
38 | #include <sys/zfs_acl.h> | |
39 | #include <sys/zil.h> | |
40 | ||
41 | #ifdef __cplusplus | |
42 | extern "C" { | |
43 | #endif | |
44 | ||
45 | /* | |
46 | * Additional file level attributes, that are stored | |
47 | * in the upper half of zp_flags | |
48 | */ | |
49 | #define ZFS_READONLY 0x0000000100000000 | |
50 | #define ZFS_HIDDEN 0x0000000200000000 | |
51 | #define ZFS_SYSTEM 0x0000000400000000 | |
52 | #define ZFS_ARCHIVE 0x0000000800000000 | |
53 | #define ZFS_IMMUTABLE 0x0000001000000000 | |
54 | #define ZFS_NOUNLINK 0x0000002000000000 | |
55 | #define ZFS_APPENDONLY 0x0000004000000000 | |
56 | #define ZFS_NODUMP 0x0000008000000000 | |
57 | #define ZFS_OPAQUE 0x0000010000000000 | |
58 | #define ZFS_AV_QUARANTINED 0x0000020000000000 | |
59 | #define ZFS_AV_MODIFIED 0x0000040000000000 | |
60 | ||
61 | #define ZFS_ATTR_SET(zp, attr, value) \ | |
62 | { \ | |
63 | if (value) \ | |
64 | zp->z_phys->zp_flags |= attr; \ | |
65 | else \ | |
66 | zp->z_phys->zp_flags &= ~attr; \ | |
67 | } | |
68 | ||
69 | /* | |
70 | * Define special zfs pflags | |
71 | */ | |
72 | #define ZFS_XATTR 0x1 /* is an extended attribute */ | |
73 | #define ZFS_INHERIT_ACE 0x2 /* ace has inheritable ACEs */ | |
74 | #define ZFS_ACL_TRIVIAL 0x4 /* files ACL is trivial */ | |
75 | #define ZFS_ACL_OBJ_ACE 0x8 /* ACL has CMPLX Object ACE */ | |
76 | #define ZFS_ACL_PROTECTED 0x10 /* ACL protected */ | |
77 | #define ZFS_ACL_DEFAULTED 0x20 /* ACL should be defaulted */ | |
78 | #define ZFS_ACL_AUTO_INHERIT 0x40 /* ACL should be inherited */ | |
79 | #define ZFS_BONUS_SCANSTAMP 0x80 /* Scanstamp in bonus area */ | |
80 | ||
81 | /* | |
82 | * Is ID ephemeral? | |
83 | */ | |
84 | #define IS_EPHEMERAL(x) (x > MAXUID) | |
85 | ||
86 | /* | |
87 | * Should we use FUIDs? | |
88 | */ | |
89 | #define USE_FUIDS(version, os) (version >= ZPL_VERSION_FUID &&\ | |
90 | spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID) | |
91 | ||
92 | #define MASTER_NODE_OBJ 1 | |
93 | ||
94 | /* | |
95 | * Special attributes for master node. | |
96 | */ | |
97 | #define ZFS_FSID "FSID" | |
98 | #define ZFS_UNLINKED_SET "DELETE_QUEUE" | |
99 | #define ZFS_ROOT_OBJ "ROOT" | |
100 | #define ZPL_VERSION_STR "VERSION" | |
101 | #define ZFS_FUID_TABLES "FUID" | |
102 | ||
103 | #define ZFS_MAX_BLOCKSIZE (SPA_MAXBLOCKSIZE) | |
104 | ||
105 | /* Path component length */ | |
106 | /* | |
107 | * The generic fs code uses MAXNAMELEN to represent | |
108 | * what the largest component length is. Unfortunately, | |
109 | * this length includes the terminating NULL. ZFS needs | |
110 | * to tell the users via pathconf() and statvfs() what the | |
111 | * true maximum length of a component is, excluding the NULL. | |
112 | */ | |
113 | #define ZFS_MAXNAMELEN (MAXNAMELEN - 1) | |
114 | ||
115 | /* | |
116 | * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in | |
117 | * the directory entries. | |
118 | */ | |
119 | #define IFTODT(mode) (((mode) & S_IFMT) >> 12) | |
120 | ||
121 | /* | |
122 | * The directory entry has the type (currently unused on Solaris) in the | |
123 | * top 4 bits, and the object number in the low 48 bits. The "middle" | |
124 | * 12 bits are unused. | |
125 | */ | |
126 | #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4) | |
127 | #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48) | |
128 | ||
129 | /* | |
130 | * This is the persistent portion of the znode. It is stored | |
131 | * in the "bonus buffer" of the file. Short symbolic links | |
132 | * are also stored in the bonus buffer. | |
133 | */ | |
134 | typedef struct znode_phys { | |
135 | uint64_t zp_atime[2]; /* 0 - last file access time */ | |
136 | uint64_t zp_mtime[2]; /* 16 - last file modification time */ | |
137 | uint64_t zp_ctime[2]; /* 32 - last file change time */ | |
138 | uint64_t zp_crtime[2]; /* 48 - creation time */ | |
139 | uint64_t zp_gen; /* 64 - generation (txg of creation) */ | |
140 | uint64_t zp_mode; /* 72 - file mode bits */ | |
141 | uint64_t zp_size; /* 80 - size of file */ | |
142 | uint64_t zp_parent; /* 88 - directory parent (`..') */ | |
143 | uint64_t zp_links; /* 96 - number of links to file */ | |
144 | uint64_t zp_xattr; /* 104 - DMU object for xattrs */ | |
145 | uint64_t zp_rdev; /* 112 - dev_t for VBLK & VCHR files */ | |
146 | uint64_t zp_flags; /* 120 - persistent flags */ | |
147 | uint64_t zp_uid; /* 128 - file owner */ | |
148 | uint64_t zp_gid; /* 136 - owning group */ | |
149 | uint64_t zp_zap; /* 144 - extra attributes */ | |
150 | uint64_t zp_pad[3]; /* 152 - future */ | |
151 | zfs_acl_phys_t zp_acl; /* 176 - 263 ACL */ | |
152 | /* | |
153 | * Data may pad out any remaining bytes in the znode buffer, eg: | |
154 | * | |
155 | * |<---------------------- dnode_phys (512) ------------------------>| | |
156 | * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->| | |
157 | * |<---- znode (264) ---->|<---- data (56) ---->| | |
158 | * | |
159 | * At present, we use this space for the following: | |
160 | * - symbolic links | |
161 | * - 32-byte anti-virus scanstamp (regular files only) | |
162 | */ | |
163 | } znode_phys_t; | |
164 | ||
165 | /* | |
166 | * Directory entry locks control access to directory entries. | |
167 | * They are used to protect creates, deletes, and renames. | |
168 | * Each directory znode has a mutex and a list of locked names. | |
169 | */ | |
170 | #ifdef _KERNEL | |
171 | typedef struct zfs_dirlock { | |
172 | char *dl_name; /* directory entry being locked */ | |
173 | uint32_t dl_sharecnt; /* 0 if exclusive, > 0 if shared */ | |
174 | uint16_t dl_namesize; /* set if dl_name was allocated */ | |
175 | kcondvar_t dl_cv; /* wait for entry to be unlocked */ | |
176 | struct znode *dl_dzp; /* directory znode */ | |
177 | struct zfs_dirlock *dl_next; /* next in z_dirlocks list */ | |
178 | } zfs_dirlock_t; | |
179 | ||
180 | typedef struct znode { | |
181 | struct zfsvfs *z_zfsvfs; | |
182 | vnode_t *z_vnode; | |
183 | uint64_t z_id; /* object ID for this znode */ | |
184 | kmutex_t z_lock; /* znode modification lock */ | |
34dc7c2f BB |
185 | krwlock_t z_parent_lock; /* parent lock for directories */ |
186 | krwlock_t z_name_lock; /* "master" lock for dirent locks */ | |
187 | zfs_dirlock_t *z_dirlocks; /* directory entry lock list */ | |
188 | kmutex_t z_range_lock; /* protects changes to z_range_avl */ | |
189 | avl_tree_t z_range_avl; /* avl tree of file range locks */ | |
190 | uint8_t z_unlinked; /* file has been unlinked */ | |
191 | uint8_t z_atime_dirty; /* atime needs to be synced */ | |
192 | uint8_t z_zn_prefetch; /* Prefetch znodes? */ | |
193 | uint_t z_blksz; /* block size in bytes */ | |
194 | uint_t z_seq; /* modification sequence number */ | |
195 | uint64_t z_mapcnt; /* number of pages mapped to file */ | |
196 | uint64_t z_last_itx; /* last ZIL itx on this znode */ | |
197 | uint64_t z_gen; /* generation (same as zp_gen) */ | |
198 | uint32_t z_sync_cnt; /* synchronous open count */ | |
199 | kmutex_t z_acl_lock; /* acl data lock */ | |
200 | list_node_t z_link_node; /* all znodes in fs link */ | |
201 | /* | |
202 | * These are dmu managed fields. | |
203 | */ | |
204 | znode_phys_t *z_phys; /* pointer to persistent znode */ | |
205 | dmu_buf_t *z_dbuf; /* buffer containing the z_phys */ | |
206 | } znode_t; | |
207 | ||
208 | ||
209 | /* | |
210 | * Range locking rules | |
211 | * -------------------- | |
212 | * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole | |
213 | * file range needs to be locked as RL_WRITER. Only then can the pages be | |
214 | * freed etc and zp_size reset. zp_size must be set within range lock. | |
215 | * 2. For writes and punching holes (zfs_write & zfs_space) just the range | |
216 | * being written or freed needs to be locked as RL_WRITER. | |
217 | * Multiple writes at the end of the file must coordinate zp_size updates | |
218 | * to ensure data isn't lost. A compare and swap loop is currently used | |
219 | * to ensure the file size is at least the offset last written. | |
220 | * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being | |
221 | * read needs to be locked as RL_READER. A check against zp_size can then | |
222 | * be made for reading beyond end of file. | |
223 | */ | |
224 | ||
225 | /* | |
226 | * Convert between znode pointers and vnode pointers | |
227 | */ | |
228 | #define ZTOV(ZP) ((ZP)->z_vnode) | |
229 | #define VTOZ(VP) ((znode_t *)(VP)->v_data) | |
230 | ||
231 | /* | |
232 | * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation. | |
233 | * ZFS_EXIT() must be called before exitting the vop. | |
234 | * ZFS_VERIFY_ZP() verifies the znode is valid. | |
235 | */ | |
236 | #define ZFS_ENTER(zfsvfs) \ | |
237 | { \ | |
238 | rrw_enter(&(zfsvfs)->z_teardown_lock, RW_READER, FTAG); \ | |
239 | if ((zfsvfs)->z_unmounted) { \ | |
240 | ZFS_EXIT(zfsvfs); \ | |
241 | return (EIO); \ | |
242 | } \ | |
243 | } | |
244 | ||
245 | #define ZFS_EXIT(zfsvfs) rrw_exit(&(zfsvfs)->z_teardown_lock, FTAG) | |
246 | ||
247 | #define ZFS_VERIFY_ZP(zp) \ | |
248 | if ((zp)->z_dbuf == NULL) { \ | |
249 | ZFS_EXIT((zp)->z_zfsvfs); \ | |
250 | return (EIO); \ | |
251 | } \ | |
252 | ||
253 | /* | |
254 | * Macros for dealing with dmu_buf_hold | |
255 | */ | |
256 | #define ZFS_OBJ_HASH(obj_num) ((obj_num) & (ZFS_OBJ_MTX_SZ - 1)) | |
b128c09f BB |
257 | #define ZFS_OBJ_MUTEX(zfsvfs, obj_num) \ |
258 | (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)]) | |
34dc7c2f | 259 | #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \ |
b128c09f BB |
260 | mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) |
261 | #define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \ | |
262 | mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) | |
34dc7c2f | 263 | #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \ |
b128c09f | 264 | mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num))) |
34dc7c2f BB |
265 | |
266 | /* | |
267 | * Macros to encode/decode ZFS stored time values from/to struct timespec | |
268 | */ | |
269 | #define ZFS_TIME_ENCODE(tp, stmp) \ | |
270 | { \ | |
b128c09f | 271 | (stmp)[0] = (uint64_t)(tp)->tv_sec; \ |
34dc7c2f BB |
272 | (stmp)[1] = (uint64_t)(tp)->tv_nsec; \ |
273 | } | |
274 | ||
275 | #define ZFS_TIME_DECODE(tp, stmp) \ | |
276 | { \ | |
277 | (tp)->tv_sec = (time_t)(stmp)[0]; \ | |
278 | (tp)->tv_nsec = (long)(stmp)[1]; \ | |
279 | } | |
280 | ||
281 | /* | |
282 | * Timestamp defines | |
283 | */ | |
284 | #define ACCESSED (AT_ATIME) | |
285 | #define STATE_CHANGED (AT_CTIME) | |
286 | #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME) | |
287 | ||
288 | #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \ | |
289 | if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \ | |
290 | zfs_time_stamper(zp, ACCESSED, NULL) | |
291 | ||
b128c09f | 292 | extern int zfs_init_fs(zfsvfs_t *, znode_t **); |
34dc7c2f BB |
293 | extern void zfs_set_dataprop(objset_t *); |
294 | extern void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *, | |
295 | dmu_tx_t *tx); | |
296 | extern void zfs_time_stamper(znode_t *, uint_t, dmu_tx_t *); | |
297 | extern void zfs_time_stamper_locked(znode_t *, uint_t, dmu_tx_t *); | |
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_zget(zfsvfs_t *, uint64_t, znode_t **); | |
303 | extern int zfs_rezget(znode_t *); | |
304 | extern void zfs_zinactive(znode_t *); | |
305 | extern void zfs_znode_delete(znode_t *, dmu_tx_t *); | |
306 | extern void zfs_znode_free(znode_t *); | |
307 | extern void zfs_remove_op_tables(); | |
308 | extern int zfs_create_op_tables(); | |
309 | extern int zfs_sync(vfs_t *vfsp, short flag, cred_t *cr); | |
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_set_version(const char *name, uint64_t newvers); | |
313 | extern int zfs_get_stats(objset_t *os, nvlist_t *nv); | |
314 | extern void zfs_znode_dmu_fini(znode_t *); | |
315 | ||
316 | extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, | |
317 | znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *, | |
318 | vattr_t *vap); | |
319 | extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp, | |
320 | vattr_t *vap); | |
321 | extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, | |
322 | znode_t *dzp, char *name); | |
323 | extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, | |
324 | znode_t *dzp, znode_t *zp, char *name); | |
325 | extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, | |
326 | znode_t *dzp, znode_t *zp, char *name, char *link); | |
327 | extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype, | |
328 | znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp); | |
329 | extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype, | |
330 | znode_t *zp, offset_t off, ssize_t len, int ioflag); | |
331 | extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype, | |
332 | znode_t *zp, uint64_t off, uint64_t len); | |
333 | extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype, | |
334 | znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp); | |
335 | extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp, | |
336 | vsecattr_t *vsecp, zfs_fuid_info_t *fuidp); | |
337 | extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap); | |
338 | extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx); | |
339 | ||
b128c09f BB |
340 | extern caddr_t zfs_map_page(page_t *, enum seg_rw); |
341 | extern void zfs_unmap_page(page_t *, caddr_t); | |
342 | ||
34dc7c2f BB |
343 | extern zil_get_data_t zfs_get_data; |
344 | extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE]; | |
345 | extern int zfsfstype; | |
346 | ||
347 | #endif /* _KERNEL */ | |
348 | ||
349 | extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len); | |
350 | ||
351 | #ifdef __cplusplus | |
352 | } | |
353 | #endif | |
354 | ||
355 | #endif /* _SYS_FS_ZFS_ZNODE_H */ |