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 by Delphix. All rights reserved.
  24  */
  25
  26 #ifndef _SYS_FS_ZFS_ZNODE_H
  27 #define _SYS_FS_ZFS_ZNODE_H
  28
  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/sa.h>
  36 #include <sys/zfs_vfsops.h>
  37 #include <sys/rrwlock.h>
  38 #include <sys/zfs_sa.h>
  39 #include <sys/zfs_stat.h>
  40 #endif
  41 #include <sys/zfs_acl.h>
  42 #include <sys/zil.h>
  43
  44 #ifdef  __cplusplus
  45 extern "C" {
  46 #endif
  47
  48 /*
  49  * Additional file level attributes, that are stored
  50  * in the upper half of zp_flags
  51  */
  52 #define ZFS_READONLY            0x0000000100000000ull
  53 #define ZFS_HIDDEN              0x0000000200000000ull
  54 #define ZFS_SYSTEM              0x0000000400000000ull
  55 #define ZFS_ARCHIVE             0x0000000800000000ull
  56 #define ZFS_IMMUTABLE           0x0000001000000000ull
  57 #define ZFS_NOUNLINK            0x0000002000000000ull
  58 #define ZFS_APPENDONLY          0x0000004000000000ull
  59 #define ZFS_NODUMP              0x0000008000000000ull
  60 #define ZFS_OPAQUE              0x0000010000000000ull
  61 #define ZFS_AV_QUARANTINED      0x0000020000000000ull
  62 #define ZFS_AV_MODIFIED         0x0000040000000000ull
  63 #define ZFS_REPARSE             0x0000080000000000ull
  64 #define ZFS_OFFLINE             0x0000100000000000ull
  65 #define ZFS_SPARSE              0x0000200000000000ull
  66
  67 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
  68 { \
  69         if (value) \
  70                 pflags |= attr; \
  71         else \
  72                 pflags &= ~attr; \
  73         VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
  74             &pflags, sizeof (pflags), tx)); \
  75 }
  76
  77 /*
  78  * Define special zfs pflags
  79  */
  80 #define ZFS_XATTR               0x1             /* is an extended attribute */
  81 #define ZFS_INHERIT_ACE         0x2             /* ace has inheritable ACEs */
  82 #define ZFS_ACL_TRIVIAL         0x4             /* files ACL is trivial */
  83 #define ZFS_ACL_OBJ_ACE         0x8             /* ACL has CMPLX Object ACE */
  84 #define ZFS_ACL_PROTECTED       0x10            /* ACL protected */
  85 #define ZFS_ACL_DEFAULTED       0x20            /* ACL should be defaulted */
  86 #define ZFS_ACL_AUTO_INHERIT    0x40            /* ACL should be inherited */
  87 #define ZFS_BONUS_SCANSTAMP     0x80            /* Scanstamp in bonus area */
  88 #define ZFS_NO_EXECS_DENIED     0x100           /* exec was given to everyone */
  89
  90 #define SA_ZPL_ATIME(z)         z->z_attr_table[ZPL_ATIME]
  91 #define SA_ZPL_MTIME(z)         z->z_attr_table[ZPL_MTIME]
  92 #define SA_ZPL_CTIME(z)         z->z_attr_table[ZPL_CTIME]
  93 #define SA_ZPL_CRTIME(z)        z->z_attr_table[ZPL_CRTIME]
  94 #define SA_ZPL_GEN(z)           z->z_attr_table[ZPL_GEN]
  95 #define SA_ZPL_DACL_ACES(z)     z->z_attr_table[ZPL_DACL_ACES]
  96 #define SA_ZPL_XATTR(z)         z->z_attr_table[ZPL_XATTR]
  97 #define SA_ZPL_SYMLINK(z)       z->z_attr_table[ZPL_SYMLINK]
  98 #define SA_ZPL_RDEV(z)          z->z_attr_table[ZPL_RDEV]
  99 #define SA_ZPL_SCANSTAMP(z)     z->z_attr_table[ZPL_SCANSTAMP]
 100 #define SA_ZPL_UID(z)           z->z_attr_table[ZPL_UID]
 101 #define SA_ZPL_GID(z)           z->z_attr_table[ZPL_GID]
 102 #define SA_ZPL_PARENT(z)        z->z_attr_table[ZPL_PARENT]
 103 #define SA_ZPL_LINKS(z)         z->z_attr_table[ZPL_LINKS]
 104 #define SA_ZPL_MODE(z)          z->z_attr_table[ZPL_MODE]
 105 #define SA_ZPL_DACL_COUNT(z)    z->z_attr_table[ZPL_DACL_COUNT]
 106 #define SA_ZPL_FLAGS(z)         z->z_attr_table[ZPL_FLAGS]
 107 #define SA_ZPL_SIZE(z)          z->z_attr_table[ZPL_SIZE]
 108 #define SA_ZPL_ZNODE_ACL(z)     z->z_attr_table[ZPL_ZNODE_ACL]
 109 #define SA_ZPL_DXATTR(z)        z->z_attr_table[ZPL_DXATTR]
 110 #define SA_ZPL_PAD(z)           z->z_attr_table[ZPL_PAD]
 111
 112 /*
 113  * Is ID ephemeral?
 114  */
 115 #define IS_EPHEMERAL(x)         (x > MAXUID)
 116
 117 /*
 118  * Should we use FUIDs?
 119  */
 120 #define USE_FUIDS(version, os)  (version >= ZPL_VERSION_FUID && \
 121     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
 122 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
 123     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
 124
 125 #define MASTER_NODE_OBJ 1
 126
 127 /*
 128  * Special attributes for master node.
 129  * "userquota@" and "groupquota@" are also valid (from
 130  * zfs_userquota_prop_prefixes[]).
 131  */
 132 #define ZFS_FSID                "FSID"
 133 #define ZFS_UNLINKED_SET        "DELETE_QUEUE"
 134 #define ZFS_ROOT_OBJ            "ROOT"
 135 #define ZPL_VERSION_STR         "VERSION"
 136 #define ZFS_FUID_TABLES         "FUID"
 137 #define ZFS_SHARES_DIR          "SHARES"
 138 #define ZFS_SA_ATTRS            "SA_ATTRS"
 139
 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.  On Linux systems this value is already
 154  * defined correctly as part of the /usr/include/dirent.h header file.
 155  */
 156 #ifndef IFTODT
 157 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
 158 #endif
 159
 160 /*
 161  * The directory entry has the type (currently unused on Solaris) in the
 162  * top 4 bits, and the object number in the low 48 bits.  The "middle"
 163  * 12 bits are unused.
 164  */
 165 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
 166 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
 167
 168 /*
 169  * Directory entry locks control access to directory entries.
 170  * They are used to protect creates, deletes, and renames.
 171  * Each directory znode has a mutex and a list of locked names.
 172  */
 173 #ifdef _KERNEL
 174 typedef struct zfs_dirlock {
 175         char            *dl_name;       /* directory entry being locked */
 176         uint32_t        dl_sharecnt;    /* 0 if exclusive, > 0 if shared */
 177         uint8_t         dl_namelock;    /* 1 if z_name_lock is NOT held */
 178         uint16_t        dl_namesize;    /* set if dl_name was allocated */
 179         kcondvar_t      dl_cv;          /* wait for entry to be unlocked */
 180         struct znode    *dl_dzp;        /* directory znode */
 181         struct zfs_dirlock *dl_next;    /* next in z_dirlocks list */
 182 } zfs_dirlock_t;
 183
 184 typedef struct znode {
 185         uint64_t        z_id;           /* object ID for this znode */
 186         kmutex_t        z_lock;         /* znode modification lock */
 187         krwlock_t       z_parent_lock;  /* parent lock for directories */
 188         krwlock_t       z_name_lock;    /* "master" lock for dirent locks */
 189         zfs_dirlock_t   *z_dirlocks;    /* directory entry lock list */
 190         kmutex_t        z_range_lock;   /* protects changes to z_range_avl */
 191         avl_tree_t      z_range_avl;    /* avl tree of file range locks */
 192         uint8_t         z_unlinked;     /* file has been unlinked */
 193         uint8_t         z_atime_dirty;  /* atime needs to be synced */
 194         uint8_t         z_zn_prefetch;  /* Prefetch znodes? */
 195         uint8_t         z_moved;        /* Has this znode been moved? */
 196         uint_t          z_blksz;        /* block size in bytes */
 197         uint_t          z_seq;          /* modification sequence number */
 198         uint64_t        z_mapcnt;       /* number of pages mapped to file */
 199         uint64_t        z_size;         /* file size (cached) */
 200         uint64_t        z_links;        /* file links (cached) */
 201         uint64_t        z_pflags;       /* pflags (cached) */
 202         uint64_t        z_uid;          /* uid fuid (cached) */
 203         uint64_t        z_gid;          /* gid fuid (cached) */
 204         uint32_t        z_sync_cnt;     /* synchronous open count */
 205         mode_t          z_mode;         /* mode (cached) */
 206         kmutex_t        z_acl_lock;     /* acl data lock */
 207         zfs_acl_t       *z_acl_cached;  /* cached acl */
 208         krwlock_t       z_xattr_lock;   /* xattr data lock */
 209         nvlist_t        *z_xattr_cached; /* cached xattrs */
 210         struct znode    *z_xattr_parent; /* xattr parent znode */
 211         list_node_t     z_link_node;    /* all znodes in fs link */
 212         sa_handle_t     *z_sa_hdl;      /* handle to sa data */
 213         boolean_t       z_is_sa;        /* are we native sa? */
 214         boolean_t       z_is_zvol;      /* are we used by the zvol */
 215         boolean_t       z_is_mapped;    /* are we mmap'ed */
 216         boolean_t       z_is_ctldir;    /* are we .zfs entry */
 217         boolean_t       z_is_stale;     /* are we stale due to rollback? */
 218         struct inode    z_inode;        /* generic vfs inode */
 219 } znode_t;
 220
 221 typedef struct znode_hold {
 222         uint64_t        zh_obj;         /* object id */
 223         kmutex_t        zh_lock;        /* lock serializing object access */
 224         avl_node_t      zh_node;        /* avl tree linkage */
 225         refcount_t      zh_refcount;    /* active consumer reference count */
 226 } znode_hold_t;
 227
 228 /*
 229  * Range locking rules
 230  * --------------------
 231  * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
 232  *    file range needs to be locked as RL_WRITER. Only then can the pages be
 233  *    freed etc and zp_size reset. zp_size must be set within range lock.
 234  * 2. For writes and punching holes (zfs_write & zfs_space) just the range
 235  *    being written or freed needs to be locked as RL_WRITER.
 236  *    Multiple writes at the end of the file must coordinate zp_size updates
 237  *    to ensure data isn't lost. A compare and swap loop is currently used
 238  *    to ensure the file size is at least the offset last written.
 239  * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
 240  *    read needs to be locked as RL_READER. A check against zp_size can then
 241  *    be made for reading beyond end of file.
 242  */
 243
 244 /*
 245  * Convert between znode pointers and inode pointers
 246  */
 247 #define ZTOI(znode)     (&((znode)->z_inode))
 248 #define ITOZ(inode)     (container_of((inode), znode_t, z_inode))
 249 #define ZTOZSB(znode)   ((zfs_sb_t *)(ZTOI(znode)->i_sb->s_fs_info))
 250 #define ITOZSB(inode)   ((zfs_sb_t *)((inode)->i_sb->s_fs_info))
 251
 252 #define S_ISDEV(mode)   (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
 253
 254 /* Called on entry to each ZFS vnode and vfs operation  */
 255 #define ZFS_ENTER(zsb) \
 256         { \
 257                 rrm_enter_read(&(zsb)->z_teardown_lock, FTAG); \
 258                 if ((zsb)->z_unmounted) { \
 259                         ZFS_EXIT(zsb); \
 260                         return (EIO); \
 261                 } \
 262         }
 263
 264 /* Must be called before exiting the vop */
 265 #define ZFS_EXIT(zsb) \
 266         { \
 267                 rrm_exit(&(zsb)->z_teardown_lock, FTAG); \
 268         }
 269
 270 /* Verifies the znode is valid */
 271 #define ZFS_VERIFY_ZP(zp) \
 272         if ((zp)->z_sa_hdl == NULL) { \
 273                 ZFS_EXIT(ZTOZSB(zp)); \
 274                 return (EIO); \
 275         }
 276
 277 /*
 278  * Macros for dealing with dmu_buf_hold
 279  */
 280 #define ZFS_OBJ_MTX_SZ          64
 281 #define ZFS_OBJ_MTX_MAX         (1024 * 1024)
 282 #define ZFS_OBJ_HASH(zsb, obj)  ((obj) & ((zsb->z_hold_size) - 1))
 283
 284 extern unsigned int zfs_object_mutex_size;
 285
 286 /* Encode ZFS stored time values from a struct timespec */
 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 /* Decode ZFS stored time values to a struct timespec */
 294 #define ZFS_TIME_DECODE(tp, stmp)               \
 295 {                                               \
 296         (tp)->tv_sec = (time_t)(stmp)[0];               \
 297         (tp)->tv_nsec = (long)(stmp)[1];                \
 298 }
 299
 300 /*
 301  * Timestamp defines
 302  */
 303 #define ACCESSED                (ATTR_ATIME)
 304 #define STATE_CHANGED           (ATTR_CTIME)
 305 #define CONTENT_MODIFIED        (ATTR_MTIME | ATTR_CTIME)
 306
 307 extern int      zfs_init_fs(zfs_sb_t *, znode_t **);
 308 extern void     zfs_set_dataprop(objset_t *);
 309 extern void     zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
 310     dmu_tx_t *tx);
 311 extern void     zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
 312     uint64_t [2]);
 313 extern void     zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
 314 extern int      zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
 315 extern void     zfs_znode_init(void);
 316 extern void     zfs_znode_fini(void);
 317 extern int      zfs_znode_hold_compare(const void *, const void *);
 318 extern int      zfs_zget(zfs_sb_t *, uint64_t, znode_t **);
 319 extern int      zfs_rezget(znode_t *);
 320 extern void     zfs_zinactive(znode_t *);
 321 extern void     zfs_znode_delete(znode_t *, dmu_tx_t *);
 322 extern void     zfs_remove_op_tables(void);
 323 extern int      zfs_create_op_tables(void);
 324 extern int      zfs_sync(struct super_block *, int, cred_t *);
 325 extern dev_t    zfs_cmpldev(uint64_t);
 326 extern int      zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
 327 extern int      zfs_get_stats(objset_t *os, nvlist_t *nv);
 328 extern void     zfs_znode_dmu_fini(znode_t *);
 329 extern int      zfs_inode_alloc(struct super_block *, struct inode **ip);
 330 extern void     zfs_inode_destroy(struct inode *);
 331 extern void     zfs_inode_update(znode_t *);
 332 extern void     zfs_mark_inode_dirty(struct inode *);
 333
 334 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 335     znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
 336     vattr_t *vap);
 337 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
 338     vattr_t *vap);
 339 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 340     znode_t *dzp, char *name, uint64_t foid);
 341 #define ZFS_NO_OBJECT   0       /* no object id */
 342 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 343     znode_t *dzp, znode_t *zp, char *name);
 344 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 345     znode_t *dzp, znode_t *zp, char *name, char *link);
 346 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 347     znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
 348 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 349     znode_t *zp, offset_t off, ssize_t len, int ioflag,
 350     zil_callback_t callback, void *callback_data);
 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 */