include/sys/zfs_znode.h

   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 2016 Nexenta Systems, Inc. All rights reserved.
  25  */
  26
  27 #ifndef _SYS_FS_ZFS_ZNODE_H
  28 #define _SYS_FS_ZFS_ZNODE_H
  29
  30 #ifdef _KERNEL
  31 #include <sys/isa_defs.h>
  32 #include <sys/types32.h>
  33 #include <sys/attr.h>
  34 #include <sys/list.h>
  35 #include <sys/dmu.h>
  36 #include <sys/sa.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>
  42 #endif
  43 #include <sys/zfs_acl.h>
  44 #include <sys/zil.h>
  45 #include <sys/zfs_project.h>
  46
  47 #ifdef  __cplusplus
  48 extern "C" {
  49 #endif
  50
  51 /*
  52  * Additional file level attributes, that are stored
  53  * in the upper half of zp_flags
  54  */
  55 #define ZFS_READONLY            0x0000000100000000ull
  56 #define ZFS_HIDDEN              0x0000000200000000ull
  57 #define ZFS_SYSTEM              0x0000000400000000ull
  58 #define ZFS_ARCHIVE             0x0000000800000000ull
  59 #define ZFS_IMMUTABLE           0x0000001000000000ull
  60 #define ZFS_NOUNLINK            0x0000002000000000ull
  61 #define ZFS_APPENDONLY          0x0000004000000000ull
  62 #define ZFS_NODUMP              0x0000008000000000ull
  63 #define ZFS_OPAQUE              0x0000010000000000ull
  64 #define ZFS_AV_QUARANTINED      0x0000020000000000ull
  65 #define ZFS_AV_MODIFIED         0x0000040000000000ull
  66 #define ZFS_REPARSE             0x0000080000000000ull
  67 #define ZFS_OFFLINE             0x0000100000000000ull
  68 #define ZFS_SPARSE              0x0000200000000000ull
  69
  70 /*
  71  * PROJINHERIT attribute is used to indicate that the child object under the
  72  * directory which has the PROJINHERIT attribute needs to inherit its parent
  73  * project ID that is used by project quota.
  74  */
  75 #define ZFS_PROJINHERIT         0x0000400000000000ull
  76
  77 /*
  78  * PROJID attr is used internally to indicate that the object has project ID.
  79  */
  80 #define ZFS_PROJID              0x0000800000000000ull
  81
  82 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
  83 { \
  84         if (value) \
  85                 pflags |= attr; \
  86         else \
  87                 pflags &= ~attr; \
  88         VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
  89             &pflags, sizeof (pflags), tx)); \
  90 }
  91
  92 /*
  93  * Define special zfs pflags
  94  */
  95 #define ZFS_XATTR               0x1             /* is an extended attribute */
  96 #define ZFS_INHERIT_ACE         0x2             /* ace has inheritable ACEs */
  97 #define ZFS_ACL_TRIVIAL         0x4             /* files ACL is trivial */
  98 #define ZFS_ACL_OBJ_ACE         0x8             /* ACL has CMPLX Object ACE */
  99 #define ZFS_ACL_PROTECTED       0x10            /* ACL protected */
 100 #define ZFS_ACL_DEFAULTED       0x20            /* ACL should be defaulted */
 101 #define ZFS_ACL_AUTO_INHERIT    0x40            /* ACL should be inherited */
 102 #define ZFS_BONUS_SCANSTAMP     0x80            /* Scanstamp in bonus area */
 103 #define ZFS_NO_EXECS_DENIED     0x100           /* exec was given to everyone */
 104
 105 #define SA_ZPL_ATIME(z)         z->z_attr_table[ZPL_ATIME]
 106 #define SA_ZPL_MTIME(z)         z->z_attr_table[ZPL_MTIME]
 107 #define SA_ZPL_CTIME(z)         z->z_attr_table[ZPL_CTIME]
 108 #define SA_ZPL_CRTIME(z)        z->z_attr_table[ZPL_CRTIME]
 109 #define SA_ZPL_GEN(z)           z->z_attr_table[ZPL_GEN]
 110 #define SA_ZPL_DACL_ACES(z)     z->z_attr_table[ZPL_DACL_ACES]
 111 #define SA_ZPL_XATTR(z)         z->z_attr_table[ZPL_XATTR]
 112 #define SA_ZPL_SYMLINK(z)       z->z_attr_table[ZPL_SYMLINK]
 113 #define SA_ZPL_RDEV(z)          z->z_attr_table[ZPL_RDEV]
 114 #define SA_ZPL_SCANSTAMP(z)     z->z_attr_table[ZPL_SCANSTAMP]
 115 #define SA_ZPL_UID(z)           z->z_attr_table[ZPL_UID]
 116 #define SA_ZPL_GID(z)           z->z_attr_table[ZPL_GID]
 117 #define SA_ZPL_PARENT(z)        z->z_attr_table[ZPL_PARENT]
 118 #define SA_ZPL_LINKS(z)         z->z_attr_table[ZPL_LINKS]
 119 #define SA_ZPL_MODE(z)          z->z_attr_table[ZPL_MODE]
 120 #define SA_ZPL_DACL_COUNT(z)    z->z_attr_table[ZPL_DACL_COUNT]
 121 #define SA_ZPL_FLAGS(z)         z->z_attr_table[ZPL_FLAGS]
 122 #define SA_ZPL_SIZE(z)          z->z_attr_table[ZPL_SIZE]
 123 #define SA_ZPL_ZNODE_ACL(z)     z->z_attr_table[ZPL_ZNODE_ACL]
 124 #define SA_ZPL_DXATTR(z)        z->z_attr_table[ZPL_DXATTR]
 125 #define SA_ZPL_PAD(z)           z->z_attr_table[ZPL_PAD]
 126 #define SA_ZPL_PROJID(z)        z->z_attr_table[ZPL_PROJID]
 127
 128 /*
 129  * Is ID ephemeral?
 130  */
 131 #define IS_EPHEMERAL(x)         (x > MAXUID)
 132
 133 /*
 134  * Should we use FUIDs?
 135  */
 136 #define USE_FUIDS(version, os)  (version >= ZPL_VERSION_FUID && \
 137     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
 138 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
 139     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
 140
 141 #define MASTER_NODE_OBJ 1
 142
 143 /*
 144  * Special attributes for master node.
 145  * "userquota@", "groupquota@" and "projectquota@" are also valid (from
 146  * zfs_userquota_prop_prefixes[]).
 147  */
 148 #define ZFS_FSID                "FSID"
 149 #define ZFS_UNLINKED_SET        "DELETE_QUEUE"
 150 #define ZFS_ROOT_OBJ            "ROOT"
 151 #define ZPL_VERSION_STR         "VERSION"
 152 #define ZFS_FUID_TABLES         "FUID"
 153 #define ZFS_SHARES_DIR          "SHARES"
 154 #define ZFS_SA_ATTRS            "SA_ATTRS"
 155
 156 /*
 157  * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
 158  * the directory entries.  On Linux systems this value is already
 159  * defined correctly as part of the /usr/include/dirent.h header file.
 160  */
 161 #ifndef IFTODT
 162 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
 163 #endif
 164
 165 /*
 166  * The directory entry has the type (currently unused on Solaris) in the
 167  * top 4 bits, and the object number in the low 48 bits.  The "middle"
 168  * 12 bits are unused.
 169  */
 170 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
 171 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
 172
 173 /*
 174  * Directory entry locks control access to directory entries.
 175  * They are used to protect creates, deletes, and renames.
 176  * Each directory znode has a mutex and a list of locked names.
 177  */
 178 #ifdef _KERNEL
 179 typedef struct zfs_dirlock {
 180         char            *dl_name;       /* directory entry being locked */
 181         uint32_t        dl_sharecnt;    /* 0 if exclusive, > 0 if shared */
 182         uint8_t         dl_namelock;    /* 1 if z_name_lock is NOT held */
 183         uint16_t        dl_namesize;    /* set if dl_name was allocated */
 184         kcondvar_t      dl_cv;          /* wait for entry to be unlocked */
 185         struct znode    *dl_dzp;        /* directory znode */
 186         struct zfs_dirlock *dl_next;    /* next in z_dirlocks list */
 187 } zfs_dirlock_t;
 188
 189 typedef struct znode {
 190         uint64_t        z_id;           /* object ID for this znode */
 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         zfs_rlock_t     z_range_lock;   /* file range lock */
 196         uint8_t         z_unlinked;     /* file has been unlinked */
 197         uint8_t         z_atime_dirty;  /* atime needs to be synced */
 198         uint8_t         z_zn_prefetch;  /* Prefetch znodes? */
 199         uint8_t         z_moved;        /* Has this znode been moved? */
 200         uint_t          z_blksz;        /* block size in bytes */
 201         uint_t          z_seq;          /* modification sequence number */
 202         uint64_t        z_mapcnt;       /* number of pages mapped to file */
 203         uint64_t        z_dnodesize;    /* dnode size */
 204         uint64_t        z_size;         /* file size (cached) */
 205         uint64_t        z_pflags;       /* pflags (cached) */
 206         uint32_t        z_sync_cnt;     /* synchronous open count */
 207         mode_t          z_mode;         /* mode (cached) */
 208         kmutex_t        z_acl_lock;     /* acl data lock */
 209         zfs_acl_t       *z_acl_cached;  /* cached acl */
 210         krwlock_t       z_xattr_lock;   /* xattr data lock */
 211         nvlist_t        *z_xattr_cached; /* cached xattrs */
 212         uint64_t        z_xattr_parent; /* parent obj for this xattr */
 213         uint64_t        z_projid;       /* project ID */
 214         list_node_t     z_link_node;    /* all znodes in fs link */
 215         sa_handle_t     *z_sa_hdl;      /* handle to sa data */
 216         boolean_t       z_is_sa;        /* are we native sa? */
 217         boolean_t       z_is_mapped;    /* are we mmap'ed */
 218         boolean_t       z_is_ctldir;    /* are we .zfs entry */
 219         boolean_t       z_is_stale;     /* are we stale due to rollback? */
 220         struct inode    z_inode;        /* generic vfs inode */
 221 } znode_t;
 222
 223 typedef struct znode_hold {
 224         uint64_t        zh_obj;         /* object id */
 225         kmutex_t        zh_lock;        /* lock serializing object access */
 226         avl_node_t      zh_node;        /* avl tree linkage */
 227         refcount_t      zh_refcount;    /* active consumer reference count */
 228 } znode_hold_t;
 229
 230 static inline uint64_t
 231 zfs_inherit_projid(znode_t *dzp)
 232 {
 233         return ((dzp->z_pflags & ZFS_PROJINHERIT) ? dzp->z_projid :
 234             ZFS_DEFAULT_PROJID);
 235 }
 236
 237 /*
 238  * Range locking rules
 239  * --------------------
 240  * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
 241  *    file range needs to be locked as RL_WRITER. Only then can the pages be
 242  *    freed etc and zp_size reset. zp_size must be set within range lock.
 243  * 2. For writes and punching holes (zfs_write & zfs_space) just the range
 244  *    being written or freed needs to be locked as RL_WRITER.
 245  *    Multiple writes at the end of the file must coordinate zp_size updates
 246  *    to ensure data isn't lost. A compare and swap loop is currently used
 247  *    to ensure the file size is at least the offset last written.
 248  * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
 249  *    read needs to be locked as RL_READER. A check against zp_size can then
 250  *    be made for reading beyond end of file.
 251  */
 252
 253 /*
 254  * Convert between znode pointers and inode pointers
 255  */
 256 #define ZTOI(znode)     (&((znode)->z_inode))
 257 #define ITOZ(inode)     (container_of((inode), znode_t, z_inode))
 258 #define ZTOZSB(znode)   ((zfsvfs_t *)(ZTOI(znode)->i_sb->s_fs_info))
 259 #define ITOZSB(inode)   ((zfsvfs_t *)((inode)->i_sb->s_fs_info))
 260
 261 #define S_ISDEV(mode)   (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
 262
 263 /* Called on entry to each ZFS vnode and vfs operation  */
 264 #define ZFS_ENTER(zfsvfs) \
 265         { \
 266                 rrm_enter_read(&(zfsvfs)->z_teardown_lock, FTAG); \
 267                 if ((zfsvfs)->z_unmounted) { \
 268                         ZFS_EXIT(zfsvfs); \
 269                         return (EIO); \
 270                 } \
 271         }
 272
 273 /* Must be called before exiting the vop */
 274 #define ZFS_EXIT(zfsvfs) \
 275         { \
 276                 rrm_exit(&(zfsvfs)->z_teardown_lock, FTAG); \
 277         }
 278
 279 /* Verifies the znode is valid */
 280 #define ZFS_VERIFY_ZP(zp) \
 281         if ((zp)->z_sa_hdl == NULL) { \
 282                 ZFS_EXIT(ZTOZSB(zp)); \
 283                 return (EIO); \
 284         }
 285
 286 /*
 287  * Macros for dealing with dmu_buf_hold
 288  */
 289 #define ZFS_OBJ_MTX_SZ          64
 290 #define ZFS_OBJ_MTX_MAX         (1024 * 1024)
 291 #define ZFS_OBJ_HASH(zfsvfs, obj)       ((obj) & ((zfsvfs->z_hold_size) - 1))
 292
 293 extern unsigned int zfs_object_mutex_size;
 294
 295 /* Encode ZFS stored time values from a struct timespec */
 296 #define ZFS_TIME_ENCODE(tp, stmp)               \
 297 {                                               \
 298         (stmp)[0] = (uint64_t)(tp)->tv_sec;     \
 299         (stmp)[1] = (uint64_t)(tp)->tv_nsec;    \
 300 }
 301
 302 /* Decode ZFS stored time values to a struct timespec */
 303 #define ZFS_TIME_DECODE(tp, stmp)               \
 304 {                                               \
 305         (tp)->tv_sec = (time_t)(stmp)[0];               \
 306         (tp)->tv_nsec = (long)(stmp)[1];                \
 307 }
 308
 309 /*
 310  * Timestamp defines
 311  */
 312 #define ACCESSED                (ATTR_ATIME)
 313 #define STATE_CHANGED           (ATTR_CTIME)
 314 #define CONTENT_MODIFIED        (ATTR_MTIME | ATTR_CTIME)
 315
 316 extern int      zfs_init_fs(zfsvfs_t *, znode_t **);
 317 extern void     zfs_set_dataprop(objset_t *);
 318 extern void     zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
 319     dmu_tx_t *tx);
 320 extern void     zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
 321     uint64_t [2]);
 322 extern void     zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
 323 extern int      zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
 324 extern void     zfs_znode_init(void);
 325 extern void     zfs_znode_fini(void);
 326 extern int      zfs_znode_hold_compare(const void *, const void *);
 327 extern int      zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
 328 extern int      zfs_rezget(znode_t *);
 329 extern void     zfs_zinactive(znode_t *);
 330 extern void     zfs_znode_delete(znode_t *, dmu_tx_t *);
 331 extern void     zfs_remove_op_tables(void);
 332 extern int      zfs_create_op_tables(void);
 333 extern int      zfs_sync(struct super_block *, int, cred_t *);
 334 extern dev_t    zfs_cmpldev(uint64_t);
 335 extern int      zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
 336 extern int      zfs_get_stats(objset_t *os, nvlist_t *nv);
 337 extern boolean_t zfs_get_vfs_flag_unmounted(objset_t *os);
 338 extern void     zfs_znode_dmu_fini(znode_t *);
 339 extern int      zfs_inode_alloc(struct super_block *, struct inode **ip);
 340 extern void     zfs_inode_destroy(struct inode *);
 341 extern void     zfs_inode_update(znode_t *);
 342 extern void     zfs_mark_inode_dirty(struct inode *);
 343
 344 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 345     znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
 346     vattr_t *vap);
 347 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
 348     vattr_t *vap);
 349 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 350     znode_t *dzp, char *name, uint64_t foid);
 351 #define ZFS_NO_OBJECT   0       /* no object id */
 352 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 353     znode_t *dzp, znode_t *zp, char *name);
 354 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 355     znode_t *dzp, znode_t *zp, char *name, char *link);
 356 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 357     znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
 358 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 359     znode_t *zp, offset_t off, ssize_t len, int ioflag,
 360     zil_callback_t callback, void *callback_data);
 361 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 362     znode_t *zp, uint64_t off, uint64_t len);
 363 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 364     znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
 365 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
 366     vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
 367 extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx);
 368 extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
 369 extern int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
 370
 371 #if defined(HAVE_UIO_RW)
 372 extern caddr_t zfs_map_page(page_t *, enum seg_rw);
 373 extern void zfs_unmap_page(page_t *, caddr_t);
 374 #endif /* HAVE_UIO_RW */
 375
 376 extern zil_get_data_t zfs_get_data;
 377 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
 378 extern int zfsfstype;
 379
 380 #endif /* _KERNEL */
 381
 382 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
 383
 384 #ifdef  __cplusplus
 385 }
 386 #endif
 387
 388 #endif  /* _SYS_FS_ZFS_ZNODE_H */