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, 2018 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/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 #include <sys/zfs_rlock.h>
  41 #endif
  42 #include <sys/zfs_acl.h>
  43 #include <sys/zil.h>
  44 #include <sys/zfs_project.h>
  45
  46 #ifdef  __cplusplus
  47 extern "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            0x0000000100000000ull
  55 #define ZFS_HIDDEN              0x0000000200000000ull
  56 #define ZFS_SYSTEM              0x0000000400000000ull
  57 #define ZFS_ARCHIVE             0x0000000800000000ull
  58 #define ZFS_IMMUTABLE           0x0000001000000000ull
  59 #define ZFS_NOUNLINK            0x0000002000000000ull
  60 #define ZFS_APPENDONLY          0x0000004000000000ull
  61 #define ZFS_NODUMP              0x0000008000000000ull
  62 #define ZFS_OPAQUE              0x0000010000000000ull
  63 #define ZFS_AV_QUARANTINED      0x0000020000000000ull
  64 #define ZFS_AV_MODIFIED         0x0000040000000000ull
  65 #define ZFS_REPARSE             0x0000080000000000ull
  66 #define ZFS_OFFLINE             0x0000100000000000ull
  67 #define ZFS_SPARSE              0x0000200000000000ull
  68
  69 /*
  70  * PROJINHERIT attribute is used to indicate that the child object under the
  71  * directory which has the PROJINHERIT attribute needs to inherit its parent
  72  * project ID that is used by project quota.
  73  */
  74 #define ZFS_PROJINHERIT         0x0000400000000000ull
  75
  76 /*
  77  * PROJID attr is used internally to indicate that the object has project ID.
  78  */
  79 #define ZFS_PROJID              0x0000800000000000ull
  80
  81 #define ZFS_ATTR_SET(zp, attr, value, pflags, tx) \
  82 { \
  83         if (value) \
  84                 pflags |= attr; \
  85         else \
  86                 pflags &= ~attr; \
  87         VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_FLAGS(ZTOZSB(zp)), \
  88             &pflags, sizeof (pflags), tx)); \
  89 }
  90
  91 /*
  92  * Define special zfs pflags
  93  */
  94 #define ZFS_XATTR               0x1             /* is an extended attribute */
  95 #define ZFS_INHERIT_ACE         0x2             /* ace has inheritable ACEs */
  96 #define ZFS_ACL_TRIVIAL         0x4             /* files ACL is trivial */
  97 #define ZFS_ACL_OBJ_ACE         0x8             /* ACL has CMPLX Object ACE */
  98 #define ZFS_ACL_PROTECTED       0x10            /* ACL protected */
  99 #define ZFS_ACL_DEFAULTED       0x20            /* ACL should be defaulted */
 100 #define ZFS_ACL_AUTO_INHERIT    0x40            /* ACL should be inherited */
 101 #define ZFS_BONUS_SCANSTAMP     0x80            /* Scanstamp in bonus area */
 102 #define ZFS_NO_EXECS_DENIED     0x100           /* exec was given to everyone */
 103
 104 #define SA_ZPL_ATIME(z)         z->z_attr_table[ZPL_ATIME]
 105 #define SA_ZPL_MTIME(z)         z->z_attr_table[ZPL_MTIME]
 106 #define SA_ZPL_CTIME(z)         z->z_attr_table[ZPL_CTIME]
 107 #define SA_ZPL_CRTIME(z)        z->z_attr_table[ZPL_CRTIME]
 108 #define SA_ZPL_GEN(z)           z->z_attr_table[ZPL_GEN]
 109 #define SA_ZPL_DACL_ACES(z)     z->z_attr_table[ZPL_DACL_ACES]
 110 #define SA_ZPL_XATTR(z)         z->z_attr_table[ZPL_XATTR]
 111 #define SA_ZPL_SYMLINK(z)       z->z_attr_table[ZPL_SYMLINK]
 112 #define SA_ZPL_RDEV(z)          z->z_attr_table[ZPL_RDEV]
 113 #define SA_ZPL_SCANSTAMP(z)     z->z_attr_table[ZPL_SCANSTAMP]
 114 #define SA_ZPL_UID(z)           z->z_attr_table[ZPL_UID]
 115 #define SA_ZPL_GID(z)           z->z_attr_table[ZPL_GID]
 116 #define SA_ZPL_PARENT(z)        z->z_attr_table[ZPL_PARENT]
 117 #define SA_ZPL_LINKS(z)         z->z_attr_table[ZPL_LINKS]
 118 #define SA_ZPL_MODE(z)          z->z_attr_table[ZPL_MODE]
 119 #define SA_ZPL_DACL_COUNT(z)    z->z_attr_table[ZPL_DACL_COUNT]
 120 #define SA_ZPL_FLAGS(z)         z->z_attr_table[ZPL_FLAGS]
 121 #define SA_ZPL_SIZE(z)          z->z_attr_table[ZPL_SIZE]
 122 #define SA_ZPL_ZNODE_ACL(z)     z->z_attr_table[ZPL_ZNODE_ACL]
 123 #define SA_ZPL_DXATTR(z)        z->z_attr_table[ZPL_DXATTR]
 124 #define SA_ZPL_PAD(z)           z->z_attr_table[ZPL_PAD]
 125 #define SA_ZPL_PROJID(z)        z->z_attr_table[ZPL_PROJID]
 126
 127 /*
 128  * Is ID ephemeral?
 129  */
 130 #define IS_EPHEMERAL(x)         (x > MAXUID)
 131
 132 /*
 133  * Should we use FUIDs?
 134  */
 135 #define USE_FUIDS(version, os)  (version >= ZPL_VERSION_FUID && \
 136     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID)
 137 #define USE_SA(version, os) (version >= ZPL_VERSION_SA && \
 138     spa_version(dmu_objset_spa(os)) >= SPA_VERSION_SA)
 139
 140 #define MASTER_NODE_OBJ 1
 141
 142 /*
 143  * Special attributes for master node.
 144  * "userquota@", "groupquota@" and "projectquota@" are also valid (from
 145  * zfs_userquota_prop_prefixes[]).
 146  */
 147 #define ZFS_FSID                "FSID"
 148 #define ZFS_UNLINKED_SET        "DELETE_QUEUE"
 149 #define ZFS_ROOT_OBJ            "ROOT"
 150 #define ZPL_VERSION_STR         "VERSION"
 151 #define ZFS_FUID_TABLES         "FUID"
 152 #define ZFS_SHARES_DIR          "SHARES"
 153 #define ZFS_SA_ATTRS            "SA_ATTRS"
 154
 155 /*
 156  * Convert mode bits (zp_mode) to BSD-style DT_* values for storing in
 157  * the directory entries.  On Linux systems this value is already
 158  * defined correctly as part of the /usr/include/dirent.h header file.
 159  */
 160 #ifndef IFTODT
 161 #define IFTODT(mode) (((mode) & S_IFMT) >> 12)
 162 #endif
 163
 164 /*
 165  * The directory entry has the type (currently unused on Solaris) in the
 166  * top 4 bits, and the object number in the low 48 bits.  The "middle"
 167  * 12 bits are unused.
 168  */
 169 #define ZFS_DIRENT_TYPE(de) BF64_GET(de, 60, 4)
 170 #define ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
 171
 172 /*
 173  * Directory entry locks control access to directory entries.
 174  * They are used to protect creates, deletes, and renames.
 175  * Each directory znode has a mutex and a list of locked names.
 176  */
 177 #ifdef _KERNEL
 178 typedef struct zfs_dirlock {
 179         char            *dl_name;       /* directory entry being locked */
 180         uint32_t        dl_sharecnt;    /* 0 if exclusive, > 0 if shared */
 181         uint8_t         dl_namelock;    /* 1 if z_name_lock is NOT held */
 182         uint16_t        dl_namesize;    /* set if dl_name was allocated */
 183         kcondvar_t      dl_cv;          /* wait for entry to be unlocked */
 184         struct znode    *dl_dzp;        /* directory znode */
 185         struct zfs_dirlock *dl_next;    /* next in z_dirlocks list */
 186 } zfs_dirlock_t;
 187
 188 typedef struct znode {
 189         uint64_t        z_id;           /* object ID for this znode */
 190         kmutex_t        z_lock;         /* znode modification lock */
 191         krwlock_t       z_parent_lock;  /* parent lock for directories */
 192         krwlock_t       z_name_lock;    /* "master" lock for dirent locks */
 193         zfs_dirlock_t   *z_dirlocks;    /* directory entry lock list */
 194         rangelock_t     z_rangelock;    /* file range locks */
 195         uint8_t         z_unlinked;     /* file has been unlinked */
 196         uint8_t         z_atime_dirty;  /* atime needs to be synced */
 197         uint8_t         z_zn_prefetch;  /* Prefetch znodes? */
 198         uint8_t         z_moved;        /* Has this znode been moved? */
 199         uint_t          z_blksz;        /* block size in bytes */
 200         uint_t          z_seq;          /* modification sequence number */
 201         uint64_t        z_mapcnt;       /* number of pages mapped to file */
 202         uint64_t        z_dnodesize;    /* dnode size */
 203         uint64_t        z_size;         /* file size (cached) */
 204         uint64_t        z_pflags;       /* pflags (cached) */
 205         uint32_t        z_sync_cnt;     /* synchronous open count */
 206         mode_t          z_mode;         /* mode (cached) */
 207         kmutex_t        z_acl_lock;     /* acl data lock */
 208         zfs_acl_t       *z_acl_cached;  /* cached acl */
 209         krwlock_t       z_xattr_lock;   /* xattr data lock */
 210         nvlist_t        *z_xattr_cached; /* cached xattrs */
 211         uint64_t        z_xattr_parent; /* parent obj for this xattr */
 212         uint64_t        z_projid;       /* project ID */
 213         list_node_t     z_link_node;    /* all znodes in fs link */
 214         sa_handle_t     *z_sa_hdl;      /* handle to sa data */
 215         boolean_t       z_is_sa;        /* are we native sa? */
 216         boolean_t       z_is_mapped;    /* are we mmap'ed */
 217         boolean_t       z_is_ctldir;    /* are we .zfs entry */
 218         boolean_t       z_is_stale;     /* are we stale due to rollback? */
 219         struct inode    z_inode;        /* generic vfs inode */
 220 } znode_t;
 221
 222 typedef struct znode_hold {
 223         uint64_t        zh_obj;         /* object id */
 224         kmutex_t        zh_lock;        /* lock serializing object access */
 225         avl_node_t      zh_node;        /* avl tree linkage */
 226         zfs_refcount_t  zh_refcount;    /* active consumer reference count */
 227 } znode_hold_t;
 228
 229 static inline uint64_t
 230 zfs_inherit_projid(znode_t *dzp)
 231 {
 232         return ((dzp->z_pflags & ZFS_PROJINHERIT) ? dzp->z_projid :
 233             ZFS_DEFAULT_PROJID);
 234 }
 235
 236 /*
 237  * Range locking rules
 238  * --------------------
 239  * 1. When truncating a file (zfs_create, zfs_setattr, zfs_space) the whole
 240  *    file range needs to be locked as RL_WRITER. Only then can the pages be
 241  *    freed etc and zp_size reset. zp_size must be set within range lock.
 242  * 2. For writes and punching holes (zfs_write & zfs_space) just the range
 243  *    being written or freed needs to be locked as RL_WRITER.
 244  *    Multiple writes at the end of the file must coordinate zp_size updates
 245  *    to ensure data isn't lost. A compare and swap loop is currently used
 246  *    to ensure the file size is at least the offset last written.
 247  * 3. For reads (zfs_read, zfs_get_data & zfs_putapage) just the range being
 248  *    read needs to be locked as RL_READER. A check against zp_size can then
 249  *    be made for reading beyond end of file.
 250  */
 251
 252 /*
 253  * Convert between znode pointers and inode pointers
 254  */
 255 #define ZTOI(znode)     (&((znode)->z_inode))
 256 #define ITOZ(inode)     (container_of((inode), znode_t, z_inode))
 257 #define ZTOZSB(znode)   ((zfsvfs_t *)(ZTOI(znode)->i_sb->s_fs_info))
 258 #define ITOZSB(inode)   ((zfsvfs_t *)((inode)->i_sb->s_fs_info))
 259
 260 #define S_ISDEV(mode)   (S_ISCHR(mode) || S_ISBLK(mode) || S_ISFIFO(mode))
 261
 262 /* Called on entry to each ZFS inode and vfs operation. */
 263 #define ZFS_ENTER_ERROR(zfsvfs, error)                          \
 264 do {                                                            \
 265         rrm_enter_read(&(zfsvfs)->z_teardown_lock, FTAG);       \
 266         if ((zfsvfs)->z_unmounted) {                            \
 267                 ZFS_EXIT(zfsvfs);                               \
 268                 return (error);                                 \
 269         }                                                       \
 270 } while (0)
 271 #define ZFS_ENTER(zfsvfs)       ZFS_ENTER_ERROR(zfsvfs, EIO)
 272 #define ZPL_ENTER(zfsvfs)       ZFS_ENTER_ERROR(zfsvfs, -EIO)
 273
 274 /* Must be called before exiting the operation. */
 275 #define ZFS_EXIT(zfsvfs)                                        \
 276 do {                                                            \
 277         rrm_exit(&(zfsvfs)->z_teardown_lock, FTAG);             \
 278 } while (0)
 279 #define ZPL_EXIT(zfsvfs)        ZFS_EXIT(zfsvfs)
 280
 281 /* Verifies the znode is valid. */
 282 #define ZFS_VERIFY_ZP_ERROR(zp, error)                          \
 283 do {                                                            \
 284         if ((zp)->z_sa_hdl == NULL) {                           \
 285                 ZFS_EXIT(ZTOZSB(zp));                           \
 286                 return (error);                                 \
 287         }                                                       \
 288 } while (0)
 289 #define ZFS_VERIFY_ZP(zp)       ZFS_VERIFY_ZP_ERROR(zp, EIO)
 290 #define ZPL_VERIFY_ZP(zp)       ZFS_VERIFY_ZP_ERROR(zp, -EIO)
 291
 292 /*
 293  * Macros for dealing with dmu_buf_hold
 294  */
 295 #define ZFS_OBJ_MTX_SZ          64
 296 #define ZFS_OBJ_MTX_MAX         (1024 * 1024)
 297 #define ZFS_OBJ_HASH(zfsvfs, obj)       ((obj) & ((zfsvfs->z_hold_size) - 1))
 298
 299 extern unsigned int zfs_object_mutex_size;
 300
 301 /*
 302  * Encode ZFS stored time values from a struct timespec / struct timespec64.
 303  */
 304 #define ZFS_TIME_ENCODE(tp, stmp)               \
 305 do {                                            \
 306         (stmp)[0] = (uint64_t)(tp)->tv_sec;     \
 307         (stmp)[1] = (uint64_t)(tp)->tv_nsec;    \
 308 } while (0)
 309
 310 #if defined(HAVE_INODE_TIMESPEC64_TIMES)
 311 /*
 312  * Decode ZFS stored time values to a struct timespec64
 313  * 4.18 and newer kernels.
 314  */
 315 #define ZFS_TIME_DECODE(tp, stmp)               \
 316 do {                                            \
 317         (tp)->tv_sec = (time64_t)(stmp)[0];     \
 318         (tp)->tv_nsec = (long)(stmp)[1];        \
 319 } while (0)
 320 #else
 321 /*
 322  * Decode ZFS stored time values to a struct timespec
 323  * 4.17 and older kernels.
 324  */
 325 #define ZFS_TIME_DECODE(tp, stmp)               \
 326 do {                                            \
 327         (tp)->tv_sec = (time_t)(stmp)[0];       \
 328         (tp)->tv_nsec = (long)(stmp)[1];        \
 329 } while (0)
 330 #endif /* HAVE_INODE_TIMESPEC64_TIMES */
 331
 332 /*
 333  * Timestamp defines
 334  */
 335 #define ACCESSED                (ATTR_ATIME)
 336 #define STATE_CHANGED           (ATTR_CTIME)
 337 #define CONTENT_MODIFIED        (ATTR_MTIME | ATTR_CTIME)
 338
 339 extern int      zfs_init_fs(zfsvfs_t *, znode_t **);
 340 extern void     zfs_set_dataprop(objset_t *);
 341 extern void     zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *,
 342     dmu_tx_t *tx);
 343 extern void     zfs_tstamp_update_setup(znode_t *, uint_t, uint64_t [2],
 344     uint64_t [2]);
 345 extern void     zfs_grow_blocksize(znode_t *, uint64_t, dmu_tx_t *);
 346 extern int      zfs_freesp(znode_t *, uint64_t, uint64_t, int, boolean_t);
 347 extern void     zfs_znode_init(void);
 348 extern void     zfs_znode_fini(void);
 349 extern int      zfs_znode_hold_compare(const void *, const void *);
 350 extern int      zfs_zget(zfsvfs_t *, uint64_t, znode_t **);
 351 extern int      zfs_rezget(znode_t *);
 352 extern void     zfs_zinactive(znode_t *);
 353 extern void     zfs_znode_delete(znode_t *, dmu_tx_t *);
 354 extern void     zfs_remove_op_tables(void);
 355 extern int      zfs_create_op_tables(void);
 356 extern int      zfs_sync(struct super_block *, int, cred_t *);
 357 extern dev_t    zfs_cmpldev(uint64_t);
 358 extern int      zfs_get_zplprop(objset_t *os, zfs_prop_t prop, uint64_t *value);
 359 extern int      zfs_get_stats(objset_t *os, nvlist_t *nv);
 360 extern boolean_t zfs_get_vfs_flag_unmounted(objset_t *os);
 361 extern void     zfs_znode_dmu_fini(znode_t *);
 362 extern int      zfs_inode_alloc(struct super_block *, struct inode **ip);
 363 extern void     zfs_inode_destroy(struct inode *);
 364 extern void     zfs_inode_update(znode_t *);
 365 extern void     zfs_mark_inode_dirty(struct inode *);
 366
 367 extern void zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 368     znode_t *dzp, znode_t *zp, char *name, vsecattr_t *, zfs_fuid_info_t *,
 369     vattr_t *vap);
 370 extern int zfs_log_create_txtype(zil_create_t, vsecattr_t *vsecp,
 371     vattr_t *vap);
 372 extern void zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 373     znode_t *dzp, char *name, uint64_t foid);
 374 #define ZFS_NO_OBJECT   0       /* no object id */
 375 extern void zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 376     znode_t *dzp, znode_t *zp, char *name);
 377 extern void zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 378     znode_t *dzp, znode_t *zp, char *name, char *link);
 379 extern void zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
 380     znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp);
 381 extern void zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 382     znode_t *zp, offset_t off, ssize_t len, int ioflag,
 383     zil_callback_t callback, void *callback_data);
 384 extern void zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 385     znode_t *zp, uint64_t off, uint64_t len);
 386 extern void zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
 387     znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp);
 388 extern void zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
 389     vsecattr_t *vsecp, zfs_fuid_info_t *fuidp);
 390 extern void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx);
 391 extern void zfs_upgrade(zfsvfs_t *zfsvfs, dmu_tx_t *tx);
 392
 393 #if defined(HAVE_UIO_RW)
 394 extern caddr_t zfs_map_page(page_t *, enum seg_rw);
 395 extern void zfs_unmap_page(page_t *, caddr_t);
 396 #endif /* HAVE_UIO_RW */
 397
 398 extern zil_get_data_t zfs_get_data;
 399 extern zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE];
 400 extern int zfsfstype;
 401
 402 #endif /* _KERNEL */
 403
 404 extern int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len);
 405
 406 #ifdef  __cplusplus
 407 }
 408 #endif
 409
 410 #endif  /* _SYS_FS_ZFS_ZNODE_H */