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            0x0000000100000000
  53 #define ZFS_HIDDEN              0x0000000200000000
  54 #define ZFS_SYSTEM              0x0000000400000000
  55 #define ZFS_ARCHIVE             0x0000000800000000
  56 #define ZFS_IMMUTABLE           0x0000001000000000
  57 #define ZFS_NOUNLINK            0x0000002000000000
  58 #define ZFS_APPENDONLY          0x0000004000000000
  59 #define ZFS_NODUMP              0x0000008000000000
  60 #define ZFS_OPAQUE              0x0000010000000000
  61 #define ZFS_AV_QUARANTINED      0x0000020000000000
  62 #define ZFS_AV_MODIFIED         0x0000040000000000
  63 #define ZFS_REPARSE             0x0000080000000000
  64 #define ZFS_OFFLINE             0x0000100000000000
  65 #define ZFS_SPARSE              0x0000200000000000
  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(zp->z_zfsvfs), \
  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_PAD(z)           z->z_attr_table[ZPL_PAD]
 110 
 111 /*
 112  * Is ID ephemeral?
 113  */
 114 #define IS_EPHEMERAL(x)         (x > MAXUID)
 115 
 116 /*
 117  * Should we use FUIDs?
 118  */
 119 #define USE_FUIDS(version, os)  (version >= ZPL_VERSION_FUID && \
 120     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
 121 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
 122     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
 123 
 124 #define MASTER_NODE_OBJ 1
 125 
 126 /*
 127  * Special attributes for master node.
 128  * "userquota@" and "groupquota@" are also valid (from
 129  * zfs_userquota_prop_prefixes[]).
 130  */
 131 #define ZFS_FSID                "FSID"
 132 #define ZFS_UNLINKED_SET        "DELETE_QUEUE"
 133 #define ZFS_ROOT_OBJ            "ROOT"
 134 #define ZPL_VERSION_STR         "VERSION"
 135 #define ZFS_FUID_TABLES         "FUID"
 136 #define ZFS_SHARES_DIR          "SHARES"
 137 #define ZFS_SA_ATTRS            "SA_ATTRS"
 138 
 139 #define ZFS_MAX_BLOCKSIZE       (SPA_MAXBLOCKSIZE)
 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 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
 156 
 157 /*
 158  * The directory entry has the type (currently unused on Solaris) in the
 159  * top 4 bits, and the object number in the low 48 bits.  The "middle"
 160  * 12 bits are unused.
 161  */
 162 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
 163 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
 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         uint8_t         dl_namelock;    /* 1 if z_name_lock is NOT held */
 175         uint16_t        dl_namesize;    /* set if dl_name was allocated */
 176         kcondvar_t      dl_cv;          /* wait for entry to be unlocked */
 177         struct znode    *dl_dzp;        /* directory znode */
 178         struct zfs_dirlock *dl_next;    /* next in z_dirlocks list */
 179 } zfs_dirlock_t;
 180 
 181 typedef struct znode {
 182         struct zfsvfs   *z_zfsvfs;
 183         vnode_t         *z_vnode;
 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 } znode_t;
 213 
 214 
 215 /*
 216  * Range locking rules
 217  * --------------------
 218  * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
 219  *    file range needs to be locked as RL_WRITER. Only then can the pages be
 220  *    freed etc and zp_size reset. zp_size must be set within range lock.
 221  * 2. For writes and punching holes (zfs_write & zfs_space) just the range
 222  *    being written or freed needs to be locked as RL_WRITER.
 223  *    Multiple writes at the end of the file must coordinate zp_size updates
 224  *    to ensure data isn't lost. A compare and swap loop is currently used
 225  *    to ensure the file size is at least the offset last written.
 226  * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
 227  *    read needs to be locked as RL_READER. A check against zp_size can then
 228  *    be made for reading beyond end of file.
 229  */
 230 
 231 /*
 232  * Convert between znode pointers and vnode pointers
 233  */
 234 #define ZTOV(ZP)        ((ZP)->z_vnode)
 235 #define VTOZ(VP)        ((znode_t *)(VP)->v_data)
 236 
 237 /*
 238  * ZFS_ENTER() is called on entry to each ZFS vnode and vfs operation.
 239  * ZFS_EXIT() must be called before exitting the vop.
 240  * ZFS_VERIFY_ZP() verifies the znode is valid.
 241  */
 242 #define ZFS_ENTER(zfsvfs) \
 243         { \
 244                 rrw_enter_read(&(zfsvfs)->z_teardown_lock, FTAG); \
 245                 if ((zfsvfs)->z_unmounted) { \
 246                         ZFS_EXIT(zfsvfs); \
 247                         return (EIO); \
 248                 } \
 249         }
 250 
 251 #define ZFS_EXIT(zfsvfs) rrw_exit(&(zfsvfs)->z_teardown_lock, FTAG)
 252 
 253 #define ZFS_VERIFY_ZP(zp) \
 254         if ((zp)->z_sa_hdl == NULL) { \
 255                 ZFS_EXIT((zp)->z_zfsvfs); \
 256                 return (EIO); \
 257         } \
 258 
 259 /*
 260  * Macros for dealing with dmu_buf_hold
 261  */
 262 #define ZFS_OBJ_HASH(obj_num)   ((obj_num) & (ZFS_OBJ_MTX_SZ - 1))
 263 #define ZFS_OBJ_MUTEX(zfsvfs, obj_num)  \
 264         (&(zfsvfs)->z_hold_mtx[ZFS_OBJ_HASH(obj_num)])
 265 #define ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num) \
 266         mutex_enter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
 267 #define ZFS_OBJ_HOLD_TRYENTER(zfsvfs, obj_num) \
 268         mutex_tryenter(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
 269 #define ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num) \
 270         mutex_exit(ZFS_OBJ_MUTEX((zfsvfs), (obj_num)))
 271 
 272 /*
 273  * Macros to encode/decode ZFS stored time values from/to struct timespec
 274  */
 275 #define ZFS_TIME_ENCODE(tp, stmp)               \
 276 {                                               \
 277         (stmp)[0] = (uint64_t)(tp)->tv_sec;  \
 278         (stmp)[1] = (uint64_t)(tp)->tv_nsec; \
 279 }
 280 
 281 #define ZFS_TIME_DECODE(tp, stmp)               \
 282 {                                               \
 283         (tp)->tv_sec = (time_t)(stmp)[0];            \
 284         (tp)->tv_nsec = (long)(stmp)[1];             \
 285 }
 286 
 287 /*
 288  * Timestamp defines
 289  */
 290 #define ACCESSED                (AT_ATIME)
 291 #define STATE_CHANGED           (AT_CTIME)
 292 #define CONTENT_MODIFIED        (AT_MTIME | AT_CTIME)
 293 
 294 #define ZFS_ACCESSTIME_STAMP(zfsvfs, zp) \
 295         if ((zfsvfs)->z_atime && !((zfsvfs)->z_vfs->vfs_flag & VFS_RDONLY)) \
 296                 zfs_tstamp_update_setup(zp, ACCESSED, NULL, NULL, B_FALSE);
 297 
 298 extern int      zfs_init_fs(zfsvfs_t *, znode_t **);
 299 extern void     zfs_set_dataprop(objset_t *);
 300 extern void     zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
 301     dmu_tx_t *tx);
 302 extern void     zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
 303     uint64_t [2], boolean_t);
 304 extern void     zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
 305 extern int      zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
 306 extern void     zfs_znode_init(void);
 307 extern void     zfs_znode_fini(void);
 308 extern int      zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
 309 extern int      zfs_rezget(znode_t *);
 310 extern void     zfs_zinactive(znode_t *);
 311 extern void     zfs_znode_delete(znode_t *, dmu_tx_t *);
 312 extern void     zfs_znode_free(znode_t *);
 313 extern void     zfs_remove_op_tables();
 314 extern int      zfs_create_op_tables();
 315 extern int      zfs_sync(vfs_t *vfsp, short flag, cred_t *cr);
 316 extern dev_t    zfs_cmpldev(uint64_t);
 317 extern int      zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
 318 extern int      zfs_get_stats(objset_t *os, nvlist_t *nv);
 319 extern void     zfs_znode_dmu_fini(znode_t *);
 320 
 321 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 322     znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
 323     vattr_t *vap);
 324 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
 325     vattr_t *vap);
 326 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 327     znode_t *dzp, char *name, uint64_t foid);
 328 #define ZFS_NO_OBJECT   0       /* no object id */
 329 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 330     znode_t *dzp, znode_t *zp, char *name);
 331 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 332     znode_t *dzp, znode_t *zp, char *name, char *link);
 333 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 334     znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
 335 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 336     znode_t *zp, offset_t off, ssize_t len, int ioflag);
 337 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 338     znode_t *zp, uint64_t off, uint64_t len);
 339 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 340     znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
 341 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
 342     vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
 343 extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx);
 344 extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
 345 extern int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
 346 
 347 extern caddr_t zfs_map_page(page_t *, enum seg_rw);
 348 extern void zfs_unmap_page(page_t *, caddr_t);
 349 
 350 extern zil_get_data_t zfs_get_data;
 351 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
 352 extern int zfsfstype;
 353 
 354 #endif /* _KERNEL */
 355 
 356 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
 357 
 358 #ifdef  __cplusplus
 359 }
 360 #endif
 361 
 362 #endif  /* _SYS_FS_ZFS_ZNODE_H */