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/*
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 (c) 2014 Integros [integros.com]
25 */
26
27#ifndef _SYS_FS_ZFS_ZNODE_H
28#define _SYS_FS_ZFS_ZNODE_H
29
30#ifdef _KERNEL
31#include <sys/list.h>
32#include <sys/dmu.h>
33#include <sys/sa.h>
34#include <sys/zfs_vfsops.h>
35#include <sys/rrwlock.h>
36#include <sys/zfs_sa.h>
37#include <sys/zfs_stat.h>
38#endif
39#include <sys/zfs_acl.h>
40#include <sys/zil.h>
41
42#ifdef __NetBSD__
43#include <miscfs/genfs/genfs_node.h>
44#endif
45
46#ifdef __cplusplus
47extern "C" {
48#endif
49
50/*
51 * Additional file level attributes, that are stored
52 * in the upper half of zp_flags
53 */
54#define ZFS_READONLY 0x0000000100000000
55#define ZFS_HIDDEN 0x0000000200000000
56#define ZFS_SYSTEM 0x0000000400000000
57#define ZFS_ARCHIVE 0x0000000800000000
58#define ZFS_IMMUTABLE 0x0000001000000000
59#define ZFS_NOUNLINK 0x0000002000000000
60#define ZFS_APPENDONLY 0x0000004000000000
61#define ZFS_NODUMP 0x0000008000000000
62#define ZFS_OPAQUE 0x0000010000000000
63#define ZFS_AV_QUARANTINED 0x0000020000000000
64#define ZFS_AV_MODIFIED 0x0000040000000000
65#define ZFS_REPARSE 0x0000080000000000
66#define ZFS_OFFLINE 0x0000100000000000
67#define ZFS_SPARSE 0x0000200000000000
68
69#define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
70{ \
71 if (value) \
72 pflags |= attr; \
73 else \
74 pflags &= ~attr; \
75 VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(zp->z_zfsvfs), \
76 &pflags, sizeof (pflags), tx)); \
77}
78
79/*
80 * Define special zfs pflags
81 */
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 */
91
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_PAD(z) z->z_attr_table[ZPL_PAD]
112
113/*
114 * Is ID ephemeral?
115 */
116#define IS_EPHEMERAL(x) (x > MAXUID)
117
118/*
119 * Should we use FUIDs?
120 */
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)
125
126#define MASTER_NODE_OBJ 1
127
128/*
129 * Special attributes for master node.
130 * "userquota@" and "groupquota@" are also valid (from
131 * zfs_userquota_prop_prefixes[]).
132 */
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"
140
141/* Path component length */
142/*
143 * The generic fs code uses MAXNAMELEN to represent
144 * what the largest component length is. Unfortunately,
145 * this length includes the terminating NULL. ZFS needs
146 * to tell the users via pathconf() and statvfs() what the
147 * true maximum length of a component is, excluding the NULL.
148 */
149#define ZFS_MAXNAMELEN (MAXNAMELEN - 1)
150
151/*
152 * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
153 * the directory entries.
154 */
155#ifndef IFTODT
156#define IFTODT(mode) (((mode) & S_IFMT) >> 12)
157#endif
158
159/*
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.
163 */
164#define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
165#define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
166
167/*
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.
171 */
172#ifdef _KERNEL
173typedef 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 */
181} zfs_dirlock_t;
182
183typedef struct znode {
184#ifdef __NetBSD__
185 struct genfs_node z_gnode;
186#endif
187 struct zfsvfs *z_zfsvfs;
188 vnode_t *z_vnode;
189 uint64_t z_id; /* object ID for this znode */
190#ifdef illumos
191 kmutex_t z_lock; /* znode modification lock */
192 krwlock_t z_parent_lock; /* parent lock for directories */
193 krwlock_t z_name_lock; /* "master" lock for dirent locks */
194 zfs_dirlock_t *z_dirlocks; /* directory entry lock list */
195#endif
196 kmutex_t z_range_lock; /* protects changes to z_range_avl */
197 avl_tree_t z_range_avl; /* avl tree of file range locks */
198 uint8_t z_unlinked; /* file has been unlinked */
199 uint8_t z_atime_dirty; /* atime needs to be synced */
200 uint8_t z_zn_prefetch; /* Prefetch znodes? */
201 uint8_t z_moved; /* Has this znode been moved? */
202 uint_t z_blksz; /* block size in bytes */
203 uint_t z_seq; /* modification sequence number */
204 uint64_t z_mapcnt; /* number of pages mapped to file */
205 uint64_t z_gen; /* generation (cached) */
206 uint64_t z_size; /* file size (cached) */
207 uint64_t z_atime[2]; /* atime (cached) */
208 uint64_t z_links; /* file links (cached) */
209 uint64_t z_pflags; /* pflags (cached) */
210 uint64_t z_uid; /* uid fuid (cached) */
211 uint64_t z_gid; /* gid fuid (cached) */
212 mode_t z_mode; /* mode (cached) */
213 uint32_t z_sync_cnt; /* synchronous open count */
214 kmutex_t z_acl_lock; /* acl data lock */
215 zfs_acl_t *z_acl_cached; /* cached acl */
216 list_node_t z_link_node; /* all znodes in fs link */
217 sa_handle_t *z_sa_hdl; /* handle to sa data */
218 boolean_t z_is_sa; /* are we native sa? */
219#ifdef __NetBSD__
220 struct lockf *z_lockf; /* head of byte-level lock list */
221#endif
222} znode_t;
223
224
225/*
226 * Range locking rules
227 * --------------------
228 * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
229 * file range needs to be locked as RL_WRITER. Only then can the pages be
230 * freed etc and zp_size reset. zp_size must be set within range lock.
231 * 2. For writes and punching holes (zfs_write & zfs_space) just the range
232 * being written or freed needs to be locked as RL_WRITER.
233 * Multiple writes at the end of the file must coordinate zp_size updates
234 * to ensure data isn't lost. A compare and swap loop is currently used
235 * to ensure the file size is at least the offset last written.
236 * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
237 * read needs to be locked as RL_READER. A check against zp_size can then
238 * be made for reading beyond end of file.
239 */
240
241/*
242 * Convert between znode pointers and vnode pointers
243 */
244#ifdef DEBUG
245static __inline vnode_t *
246ZTOV(znode_t *zp)
247{
248 vnode_t *vp = zp->z_vnode;
249
250 ASSERT(vp == NULL || vp->v_data == NULL || vp->v_data == zp);
251 return (vp);
252}
253static __inline znode_t *
254VTOZ(vnode_t *vp)
255{
256 znode_t *zp = (znode_t *)vp->v_data;
257
258 ASSERT(zp == NULL || zp->z_vnode == NULL || zp->z_vnode == vp);
259 return (zp);
260}
261#else
262#define ZTOV(ZP) ((ZP)->z_vnode)
263#define VTOZ(VP) ((znode_t *)(VP)->v_data)
264#endif
265
266/* Called on entry to each ZFS vnode and vfs operation */
267#define ZFS_ENTER(zfsvfs) \
268 { \
269 rrm_enter_read(&(zfsvfs)->z_teardown_lock, FTAG); \
270 if ((zfsvfs)->z_unmounted) { \
271 ZFS_EXIT(zfsvfs); \
272 return (EIO); \
273 } \
274 }
275
276/* Must be called before exiting the vop */
277#define ZFS_EXIT(zfsvfs) rrm_exit(&(zfsvfs)->z_teardown_lock, FTAG)
278
279/* Verifies the znode is valid */
280#define ZFS_VERIFY_ZP(zp) \
281 if ((zp)->z_sa_hdl == NULL) { \
282 ZFS_EXIT((zp)->z_zfsvfs); \
283 return (EIO); \
284 } \
285
286/*
287 * Macros for dealing with dmu_buf_hold
288 */
289#define ZFS_OBJ_HASH(obj_num) ((obj_num) & (ZFS_OBJ_MTX_SZ - 1))
290#define ZFS_OBJ_MUTEX(zfsvfs, obj_num) \
291 (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
292#define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \
293 mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
294#define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \
295 mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
296#define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \
297 mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
298
299/* Encode ZFS stored time values from a struct timespec */
300#define ZFS_TIME_ENCODE(tp, stmp) \
301{ \
302 (stmp)[0] = (uint64_t)(tp)->tv_sec; \
303 (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
304}
305
306/* Decode ZFS stored time values to a struct timespec */
307#define ZFS_TIME_DECODE(tp, stmp) \
308{ \
309 (tp)->tv_sec = (time_t)(stmp)[0]; \
310 (tp)->tv_nsec = (long)(stmp)[1]; \
311}
312
313/*
314 * Timestamp defines
315 */
316#define ACCESSED (AT_ATIME)
317#define STATE_CHANGED (AT_CTIME)
318#define CONTENT_MODIFIED (AT_MTIME | AT_CTIME)
319
320#define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \
321 if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \
322 zfs_tstamp_update_setup(zp, ACCESSED, NULL, NULL, B_FALSE);
323
324extern int zfs_init_fs(zfsvfs_t *, znode_t **);
325extern void zfs_set_dataprop(objset_t *);
326extern void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
327 dmu_tx_t *tx);
328extern void zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
329 uint64_t [2], boolean_t);
330extern void zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
331extern int zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
332extern void zfs_znode_init(void);
333extern void zfs_znode_fini(void);
334#ifdef __NetBSD__
335extern int zfs_loadvnode(struct mount *, struct vnode *,
336 const void *, size_t, const void **);
337extern int zfs_newvnode(struct mount *, struct vnode *, struct vnode *,
338 struct vattr *, kauth_cred_t, void *, size_t *, const void **);
339extern int zfs_zget_cleaner(zfsvfs_t *, uint64_t, znode_t **);
340#endif
341extern int zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
342extern int zfs_rezget(znode_t *);
343extern void zfs_zinactive(znode_t *);
344extern void zfs_znode_delete(znode_t *, dmu_tx_t *);
345extern void zfs_znode_free(znode_t *);
346extern void zfs_remove_op_tables();
347extern int zfs_create_op_tables();
348extern dev_t zfs_cmpldev(uint64_t);
349extern int zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
350extern int zfs_get_stats(objset_t *os, nvlist_t *nv);
351extern void zfs_znode_dmu_fini(znode_t *);
352
353extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
354 znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
355 vattr_t *vap);
356extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
357 vattr_t *vap);
358extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
359 znode_t *dzp, char *name, uint64_t foid);
360#define ZFS_NO_OBJECT 0 /* no object id */
361extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
362 znode_t *dzp, znode_t *zp, char *name);
363extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
364 znode_t *dzp, znode_t *zp, char *name, char *link);
365extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
366 znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
367extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
368 znode_t *zp, offset_t off, ssize_t len, int ioflag);
369extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
370 znode_t *zp, uint64_t off, uint64_t len);
371extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
372 znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
373#ifndef ZFS_NO_ACL
374extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
375 vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
376#endif
377extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx);
378extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
379extern int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
380
381extern caddr_t zfs_map_page(page_t *, enum seg_rw);
382extern void zfs_unmap_page(page_t *, caddr_t);
383
384extern zil_get_data_t zfs_get_data;
385extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
386extern int zfsfstype;
387
388extern int zfs_znode_parent_and_name(znode_t *zp, znode_t **dzpp, char *buf);
389
390#endif /* _KERNEL */
391
392extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
393
394#ifdef __cplusplus
395}
396#endif
397
398#endif /* _SYS_FS_ZFS_ZNODE_H */
399