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  */
  24
  25 #ifndef _SYS_FS_ZFS_ZNODE_H
  26 #define _SYS_FS_ZFS_ZNODE_H
  27
  28 #ifdef _KERNEL
  29 #include <sys/isa_defs.h>
  30 #include <sys/types32.h>
  31 #include <sys/attr.h>
  32 #include <sys/list.h>
  33 #include <sys/dmu.h>
  34 #include <sys/sa.h>
  35 #include <sys/zfs_vfsops.h>
  36 #include <sys/rrwlock.h>
  37 #include <sys/zfs_sa.h>
  38 #include <sys/zfs_stat.h>
  39 #endif
  40 #include <sys/zfs_acl.h>
  41 #include <sys/zil.h>
  42
  43 #ifdef  __cplusplus
  44 extern "C" {
  45 #endif
  46
  47 /*
  48  * Additional file level attributes, that are stored
  49  * in the upper half of zp_flags
  50  */
  51 #define ZFS_READONLY            0x0000000100000000ull
  52 #define ZFS_HIDDEN              0x0000000200000000ull
  53 #define ZFS_SYSTEM              0x0000000400000000ull
  54 #define ZFS_ARCHIVE             0x0000000800000000ull
  55 #define ZFS_IMMUTABLE           0x0000001000000000ull
  56 #define ZFS_NOUNLINK            0x0000002000000000ull
  57 #define ZFS_APPENDONLY          0x0000004000000000ull
  58 #define ZFS_NODUMP              0x0000008000000000ull
  59 #define ZFS_OPAQUE              0x0000010000000000ull
  60 #define ZFS_AV_QUARANTINED      0x0000020000000000ull
  61 #define ZFS_AV_MODIFIED         0x0000040000000000ull
  62 #define ZFS_REPARSE             0x0000080000000000ull
  63 #define ZFS_OFFLINE             0x0000100000000000ull
  64 #define ZFS_SPARSE              0x0000200000000000ull
  65
  66 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
  67 { \
  68         if (value) \
  69                 pflags |= attr; \
  70         else \
  71                 pflags &= ~attr; \
  72         VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
  73             &pflags, sizeof (pflags), tx)); \
  74 }
  75
  76 /*
  77  * Define special zfs pflags
  78  */
  79 #define ZFS_XATTR               0x1             /* is an extended attribute */
  80 #define ZFS_INHERIT_ACE         0x2             /* ace has inheritable ACEs */
  81 #define ZFS_ACL_TRIVIAL         0x4             /* files ACL is trivial */
  82 #define ZFS_ACL_OBJ_ACE         0x8             /* ACL has CMPLX Object ACE */
  83 #define ZFS_ACL_PROTECTED       0x10            /* ACL protected */
  84 #define ZFS_ACL_DEFAULTED       0x20            /* ACL should be defaulted */
  85 #define ZFS_ACL_AUTO_INHERIT    0x40            /* ACL should be inherited */
  86 #define ZFS_BONUS_SCANSTAMP     0x80            /* Scanstamp in bonus area */
  87 #define ZFS_NO_EXECS_DENIED     0x100           /* exec was given to everyone */
  88
  89 #define SA_ZPL_ATIME(z)         z->z_attr_table[ZPL_ATIME]
  90 #define SA_ZPL_MTIME(z)         z->z_attr_table[ZPL_MTIME]
  91 #define SA_ZPL_CTIME(z)         z->z_attr_table[ZPL_CTIME]
  92 #define SA_ZPL_CRTIME(z)        z->z_attr_table[ZPL_CRTIME]
  93 #define SA_ZPL_GEN(z)           z->z_attr_table[ZPL_GEN]
  94 #define SA_ZPL_DACL_ACES(z)     z->z_attr_table[ZPL_DACL_ACES]
  95 #define SA_ZPL_XATTR(z)         z->z_attr_table[ZPL_XATTR]
  96 #define SA_ZPL_SYMLINK(z)       z->z_attr_table[ZPL_SYMLINK]
  97 #define SA_ZPL_RDEV(z)          z->z_attr_table[ZPL_RDEV]
  98 #define SA_ZPL_SCANSTAMP(z)     z->z_attr_table[ZPL_SCANSTAMP]
  99 #define SA_ZPL_UID(z)           z->z_attr_table[ZPL_UID]
 100 #define SA_ZPL_GID(z)           z->z_attr_table[ZPL_GID]
 101 #define SA_ZPL_PARENT(z)        z->z_attr_table[ZPL_PARENT]
 102 #define SA_ZPL_LINKS(z)         z->z_attr_table[ZPL_LINKS]
 103 #define SA_ZPL_MODE(z)          z->z_attr_table[ZPL_MODE]
 104 #define SA_ZPL_DACL_COUNT(z)    z->z_attr_table[ZPL_DACL_COUNT]
 105 #define SA_ZPL_FLAGS(z)         z->z_attr_table[ZPL_FLAGS]
 106 #define SA_ZPL_SIZE(z)          z->z_attr_table[ZPL_SIZE]
 107 #define SA_ZPL_ZNODE_ACL(z)     z->z_attr_table[ZPL_ZNODE_ACL]
 108 #define SA_ZPL_PAD(z)           z->z_attr_table[ZPL_PAD]
 109
 110 /*
 111  * Is ID ephemeral?
 112  */
 113 #define IS_EPHEMERAL(x)         (x > MAXUID)
 114
 115 /*
 116  * Should we use FUIDs?
 117  */
 118 #define USE_FUIDS(version, os)  (version >= ZPL_VERSION_FUID && \
 119     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
 120 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
 121     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
 122
 123 #define MASTER_NODE_OBJ 1
 124
 125 /*
 126  * Special attributes for master node.
 127  * "userquota@" and "groupquota@" are also valid (from
 128  * zfs_userquota_prop_prefixes[]).
 129  */
 130 #define ZFS_FSID                "FSID"
 131 #define ZFS_UNLINKED_SET        "DELETE_QUEUE"
 132 #define ZFS_ROOT_OBJ            "ROOT"
 133 #define ZPL_VERSION_STR         "VERSION"
 134 #define ZFS_FUID_TABLES         "FUID"
 135 #define ZFS_SHARES_DIR          "SHARES"
 136 #define ZFS_SA_ATTRS            "SA_ATTRS"
 137
 138 #define ZFS_MAX_BLOCKSIZE       (SPA_MAXBLOCKSIZE)
 139
 140 /* Path component length */
 141 /*
 142  * The generic fs code uses MAXNAMELEN to represent
 143  * what the largest component length is.  Unfortunately,
 144  * this length includes the terminating NULL.  ZFS needs
 145  * to tell the users via pathconf() and statvfs() what the
 146  * true maximum length of a component is, excluding the NULL.
 147  */
 148 #define ZFS_MAXNAMELEN  (MAXNAMELEN - 1)
 149
 150 /*
 151  * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
 152  * the directory entries.  On Linux systems this value is already
 153  * defined correctly as part of the /usr/include/dirent.h header file.
 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
 173 typedef 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
 183 typedef struct znode {
 184         uint64_t        z_id;           /* object ID for this znode */
 185         kmutex_t        z_lock;         /* znode modification lock */
 186         krwlock_t       z_parent_lock;  /* parent lock for directories */
 187         krwlock_t       z_name_lock;    /* "master" lock for dirent locks */
 188         zfs_dirlock_t   *z_dirlocks;    /* directory entry lock list */
 189         kmutex_t        z_range_lock;   /* protects changes to z_range_avl */
 190         avl_tree_t      z_range_avl;    /* avl tree of file range locks */
 191         uint8_t         z_unlinked;     /* file has been unlinked */
 192         uint8_t         z_atime_dirty;  /* atime needs to be synced */
 193         uint8_t         z_zn_prefetch;  /* Prefetch znodes? */
 194         uint8_t         z_moved;        /* Has this znode been moved? */
 195         uint_t          z_blksz;        /* block size in bytes */
 196         uint_t          z_seq;          /* modification sequence number */
 197         uint64_t        z_mapcnt;       /* number of pages mapped to file */
 198         uint64_t        z_gen;          /* generation (cached) */
 199         uint64_t        z_size;         /* file size (cached) */
 200         uint64_t        z_atime[2];     /* atime (cached) */
 201         uint64_t        z_links;        /* file links (cached) */
 202         uint64_t        z_pflags;       /* pflags (cached) */
 203         uint64_t        z_uid;          /* uid fuid (cached) */
 204         uint64_t        z_gid;          /* gid fuid (cached) */
 205         mode_t          z_mode;         /* mode (cached) */
 206         uint32_t        z_sync_cnt;     /* synchronous open count */
 207         kmutex_t        z_acl_lock;     /* acl data lock */
 208         zfs_acl_t       *z_acl_cached;  /* cached acl */
 209         list_node_t     z_link_node;    /* all znodes in fs link */
 210         sa_handle_t     *z_sa_hdl;      /* handle to sa data */
 211         boolean_t       z_is_sa;        /* are we native sa? */
 212         boolean_t       z_is_zvol;      /* are we used by the zvol */
 213         boolean_t       z_is_mapped;    /* are we mmap'ed */
 214         struct inode    z_inode;        /* generic vfs inode */
 215 } znode_t;
 216
 217
 218 /*
 219  * Range locking rules
 220  * --------------------
 221  * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
 222  *    file range needs to be locked as RL_WRITER. Only then can the pages be
 223  *    freed etc and zp_size reset. zp_size must be set within range lock.
 224  * 2. For writes and punching holes (zfs_write & zfs_space) just the range
 225  *    being written or freed needs to be locked as RL_WRITER.
 226  *    Multiple writes at the end of the file must coordinate zp_size updates
 227  *    to ensure data isn't lost. A compare and swap loop is currently used
 228  *    to ensure the file size is at least the offset last written.
 229  * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
 230  *    read needs to be locked as RL_READER. A check against zp_size can then
 231  *    be made for reading beyond end of file.
 232  */
 233
 234 /*
 235  * Convert between znode pointers and inode pointers
 236  */
 237 #define ZTOI(znode)     (&((znode)->z_inode))
 238 #define ITOZ(inode)     (container_of((inode), znode_t, z_inode))
 239 #define VTOZSB(vfs)     ((zfs_sb_t *)((vfs)->mnt_sb->s_fs_info))
 240 #define ZTOZSB(znode)   ((zfs_sb_t *)(ZTOI(znode)->i_sb->s_fs_info))
 241 #define ITOZSB(inode)   ((zfs_sb_t *)((inode)->i_sb->s_fs_info))
 242
 243 #define S_ISDEV(mode)   (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
 244
 245 /*
 246  * ZFS_ENTER() is called on entry to each ZFS inode and vfs operation.
 247  * ZFS_EXIT() must be called before exitting the vop.
 248  * ZFS_VERIFY_ZP() verifies the znode is valid.
 249  */
 250 #define ZFS_ENTER(zsb) \
 251         { \
 252                 rrw_enter(&(zsb)->z_teardown_lock, RW_READER, FTAG); \
 253                 if ((zsb)->z_unmounted) { \
 254                         ZFS_EXIT(zsb); \
 255                         return (EIO); \
 256                 } \
 257         }
 258
 259 #define ZFS_EXIT(zsb) \
 260         { \
 261                 rrw_exit(&(zsb)->z_teardown_lock, FTAG); \
 262                 tsd_exit(); \
 263         }
 264
 265 #define ZFS_VERIFY_ZP(zp) \
 266         if ((zp)->z_sa_hdl == NULL) { \
 267                 ZFS_EXIT(ZTOZSB(zp)); \
 268                 return (EIO); \
 269         }
 270
 271 /*
 272  * Macros for dealing with dmu_buf_hold
 273  */
 274 #define ZFS_OBJ_HASH(obj_num)   ((obj_num) & (ZFS_OBJ_MTX_SZ - 1))
 275 #define ZFS_OBJ_MUTEX(zsb, obj_num)     \
 276         (&(zsb)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
 277 #define ZFS_OBJ_HOLD_ENTER(zsb, obj_num) \
 278         mutex_enter(ZFS_OBJ_MUTEX((zsb), (obj_num)))
 279 #define ZFS_OBJ_HOLD_TRYENTER(zsb, obj_num) \
 280         mutex_tryenter(ZFS_OBJ_MUTEX((zsb), (obj_num)))
 281 #define ZFS_OBJ_HOLD_EXIT(zsb, obj_num) \
 282         mutex_exit(ZFS_OBJ_MUTEX((zsb), (obj_num)))
 283
 284 /*
 285  * Macros to encode/decode ZFS stored time values from/to struct timespec
 286  */
 287 #define ZFS_TIME_ENCODE(tp, stmp)               \
 288 {                                               \
 289         (stmp)[0] = (uint64_t)(tp)->tv_sec;     \
 290         (stmp)[1] = (uint64_t)(tp)->tv_nsec;    \
 291 }
 292
 293 #define ZFS_TIME_DECODE(tp, stmp)               \
 294 {                                               \
 295         (tp)->tv_sec = (time_t)(stmp)[0];               \
 296         (tp)->tv_nsec = (long)(stmp)[1];                \
 297 }
 298
 299 /*
 300  * Timestamp defines
 301  */
 302 #define ACCESSED                (ATTR_ATIME)
 303 #define STATE_CHANGED           (ATTR_CTIME)
 304 #define CONTENT_MODIFIED        (ATTR_MTIME | ATTR_CTIME)
 305
 306 #define ZFS_ACCESSTIME_STAMP(zsb, zp) \
 307         if ((zsb)->z_atime && !((zsb)->z_vfs->mnt_flags & MNT_READONLY)) \
 308                 zfs_tstamp_update_setup(zp, ACCESSED, NULL, NULL, B_FALSE);
 309
 310 extern int      zfs_init_fs(zfs_sb_t *, znode_t **);
 311 extern void     zfs_set_dataprop(objset_t *);
 312 extern void     zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
 313     dmu_tx_t *tx);
 314 extern void     zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
 315     uint64_t [2], boolean_t);
 316 extern void     zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
 317 extern int      zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
 318 extern void     zfs_znode_init(void);
 319 extern void     zfs_znode_fini(void);
 320 extern int      zfs_zget(zfs_sb_t *, uint64_t, znode_t **);
 321 extern int      zfs_rezget(znode_t *);
 322 extern void     zfs_zinactive(znode_t *);
 323 extern void     zfs_znode_delete(znode_t *, dmu_tx_t *);
 324 extern void     zfs_remove_op_tables(void);
 325 extern int      zfs_create_op_tables(void);
 326 extern int      zfs_sync(zfs_sb_t *, short, cred_t *);
 327 extern dev_t    zfs_cmpldev(uint64_t);
 328 extern int      zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
 329 extern int      zfs_get_stats(objset_t *os, nvlist_t *nv);
 330 extern void     zfs_znode_dmu_fini(znode_t *);
 331 extern int      zfs_inode_alloc(struct super_block *, struct inode **ip);
 332 extern void     zfs_inode_destroy(struct inode *);
 333 extern void     zfs_inode_update(znode_t *);
 334
 335 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 336     znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
 337     vattr_t *vap);
 338 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
 339     vattr_t *vap);
 340 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 341     znode_t *dzp, char *name, uint64_t foid);
 342 #define ZFS_NO_OBJECT   0       /* no object id */
 343 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 344     znode_t *dzp, znode_t *zp, char *name);
 345 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 346     znode_t *dzp, znode_t *zp, char *name, char *link);
 347 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 348     znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
 349 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 350     znode_t *zp, offset_t off, ssize_t len, int ioflag);
 351 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 352     znode_t *zp, uint64_t off, uint64_t len);
 353 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 354     znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
 355 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
 356     vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
 357 extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx);
 358 extern void zfs_upgrade(zfs_sb_t *zsb, dmu_tx_t *tx);
 359 extern int zfs_create_share_dir(zfs_sb_t *zsb, dmu_tx_t *tx);
 360
 361 #if defined(HAVE_UIO_RW)
 362 extern caddr_t zfs_map_page(page_t *, enum seg_rw);
 363 extern void zfs_unmap_page(page_t *, caddr_t);
 364 #endif /* HAVE_UIO_RW */
 365
 366 extern zil_get_data_t zfs_get_data;
 367 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
 368 extern int zfsfstype;
 369
 370 #endif /* _KERNEL */
 371
 372 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
 373
 374 #ifdef  __cplusplus
 375 }
 376 #endif
 377
 378 #endif  /* _SYS_FS_ZFS_ZNODE_H */