1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (c) 1989, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the University nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 * 31 * @(#)vnode.h 8.7 (Berkeley) 2/4/94 32 * $FreeBSD$ 33 */ 34 35 #ifndef _SYS_VNODE_H_ 36 #define _SYS_VNODE_H_ 37 38 #include <sys/bufobj.h> 39 #include <sys/queue.h> 40 #include <sys/lock.h> 41 #include <sys/lockmgr.h> 42 #include <sys/mutex.h> 43 #include <sys/rangelock.h> 44 #include <sys/selinfo.h> 45 #include <sys/uio.h> 46 #include <sys/acl.h> 47 #include <sys/ktr.h> 48 49 /* 50 * The vnode is the focus of all file activity in UNIX. There is a 51 * unique vnode allocated for each active file, each current directory, 52 * each mounted-on file, text file, and the root. 53 */ 54 55 /* 56 * Vnode types. VNON means no type. 57 */ 58 enum vtype { VNON, VREG, VDIR, VBLK, VCHR, VLNK, VSOCK, VFIFO, VBAD, 59 VMARKER }; 60 61 /* 62 * Each underlying filesystem allocates its own private area and hangs 63 * it from v_data. If non-null, this area is freed in getnewvnode(). 64 */ 65 66 struct namecache; 67 68 struct vpollinfo { 69 struct mtx vpi_lock; /* lock to protect below */ 70 struct selinfo vpi_selinfo; /* identity of poller(s) */ 71 short vpi_events; /* what they are looking for */ 72 short vpi_revents; /* what has happened */ 73 }; 74 75 /* 76 * Reading or writing any of these items requires holding the appropriate lock. 77 * 78 * Lock reference: 79 * c - namecache mutex 80 * i - interlock 81 * l - mp mnt_listmtx or freelist mutex 82 * I - updated with atomics, 0->1 and 1->0 transitions with interlock held 83 * m - mount point interlock 84 * p - pollinfo lock 85 * u - Only a reference to the vnode is needed to read. 86 * v - vnode lock 87 * 88 * Vnodes may be found on many lists. The general way to deal with operating 89 * on a vnode that is on a list is: 90 * 1) Lock the list and find the vnode. 91 * 2) Lock interlock so that the vnode does not go away. 92 * 3) Unlock the list to avoid lock order reversals. 93 * 4) vget with LK_INTERLOCK and check for ENOENT, or 94 * 5) Check for DOOMED if the vnode lock is not required. 95 * 6) Perform your operation, then vput(). 96 */ 97 98 #if defined(_KERNEL) || defined(_KVM_VNODE) 99 100 struct vnode { 101 /* 102 * Fields which define the identity of the vnode. These fields are 103 * owned by the filesystem (XXX: and vgone() ?) 104 */ 105 const char *v_tag; /* u type of underlying data */ 106 struct vop_vector *v_op; /* u vnode operations vector */ 107 void *v_data; /* u private data for fs */ 108 109 /* 110 * Filesystem instance stuff 111 */ 112 struct mount *v_mount; /* u ptr to vfs we are in */ 113 TAILQ_ENTRY(vnode) v_nmntvnodes; /* m vnodes for mount point */ 114 115 /* 116 * Type specific fields, only one applies to any given vnode. 117 */ 118 union { 119 struct mount *v_mountedhere; /* v ptr to mountpoint (VDIR) */ 120 struct unpcb *v_unpcb; /* v unix domain net (VSOCK) */ 121 struct cdev *v_rdev; /* v device (VCHR, VBLK) */ 122 struct fifoinfo *v_fifoinfo; /* v fifo (VFIFO) */ 123 }; 124 125 /* 126 * vfs_hash: (mount + inode) -> vnode hash. The hash value 127 * itself is grouped with other int fields, to avoid padding. 128 */ 129 LIST_ENTRY(vnode) v_hashlist; 130 131 /* 132 * VFS_namecache stuff 133 */ 134 LIST_HEAD(, namecache) v_cache_src; /* c Cache entries from us */ 135 TAILQ_HEAD(, namecache) v_cache_dst; /* c Cache entries to us */ 136 struct namecache *v_cache_dd; /* c Cache entry for .. vnode */ 137 138 /* 139 * Locking 140 */ 141 struct lock v_lock; /* u (if fs don't have one) */ 142 struct mtx v_interlock; /* lock for "i" things */ 143 struct lock *v_vnlock; /* u pointer to vnode lock */ 144 145 /* 146 * The machinery of being a vnode 147 */ 148 TAILQ_ENTRY(vnode) v_actfreelist; /* l vnode active/free lists */ 149 struct bufobj v_bufobj; /* * Buffer cache object */ 150 151 /* 152 * Hooks for various subsystems and features. 153 */ 154 struct vpollinfo *v_pollinfo; /* i Poll events, p for *v_pi */ 155 struct label *v_label; /* MAC label for vnode */ 156 struct lockf *v_lockf; /* Byte-level advisory lock list */ 157 struct rangelock v_rl; /* Byte-range lock */ 158 159 /* 160 * clustering stuff 161 */ 162 daddr_t v_cstart; /* v start block of cluster */ 163 daddr_t v_lasta; /* v last allocation */ 164 daddr_t v_lastw; /* v last write */ 165 int v_clen; /* v length of cur. cluster */ 166 167 u_int v_holdcnt; /* I prevents recycling. */ 168 u_int v_usecount; /* I ref count of users */ 169 u_int v_iflag; /* i vnode flags (see below) */ 170 u_int v_vflag; /* v vnode flags */ 171 u_int v_mflag; /* l mnt-specific vnode flags */ 172 int v_writecount; /* v ref count of writers */ 173 u_int v_hash; 174 enum vtype v_type; /* u vnode type */ 175 }; 176 177 #endif /* defined(_KERNEL) || defined(_KVM_VNODE) */ 178 179 #define bo2vnode(bo) __containerof((bo), struct vnode, v_bufobj) 180 181 /* XXX: These are temporary to avoid a source sweep at this time */ 182 #define v_object v_bufobj.bo_object 183 184 /* 185 * Userland version of struct vnode, for sysctl. 186 */ 187 struct xvnode { 188 size_t xv_size; /* sizeof(struct xvnode) */ 189 void *xv_vnode; /* address of real vnode */ 190 u_long xv_flag; /* vnode vflags */ 191 int xv_usecount; /* reference count of users */ 192 int xv_writecount; /* reference count of writers */ 193 int xv_holdcnt; /* page & buffer references */ 194 u_long xv_id; /* capability identifier */ 195 void *xv_mount; /* address of parent mount */ 196 long xv_numoutput; /* num of writes in progress */ 197 enum vtype xv_type; /* vnode type */ 198 union { 199 void *xvu_socket; /* unpcb, if VSOCK */ 200 void *xvu_fifo; /* fifo, if VFIFO */ 201 dev_t xvu_rdev; /* maj/min, if VBLK/VCHR */ 202 struct { 203 dev_t xvu_dev; /* device, if VDIR/VREG/VLNK */ 204 ino_t xvu_ino; /* id, if VDIR/VREG/VLNK */ 205 } xv_uns; 206 } xv_un; 207 }; 208 #define xv_socket xv_un.xvu_socket 209 #define xv_fifo xv_un.xvu_fifo 210 #define xv_rdev xv_un.xvu_rdev 211 #define xv_dev xv_un.xv_uns.xvu_dev 212 #define xv_ino xv_un.xv_uns.xvu_ino 213 214 /* We don't need to lock the knlist */ 215 #define VN_KNLIST_EMPTY(vp) ((vp)->v_pollinfo == NULL || \ 216 KNLIST_EMPTY(&(vp)->v_pollinfo->vpi_selinfo.si_note)) 217 218 #define VN_KNOTE(vp, b, a) \ 219 do { \ 220 if (!VN_KNLIST_EMPTY(vp)) \ 221 KNOTE(&vp->v_pollinfo->vpi_selinfo.si_note, (b), \ 222 (a) | KNF_NOKQLOCK); \ 223 } while (0) 224 #define VN_KNOTE_LOCKED(vp, b) VN_KNOTE(vp, b, KNF_LISTLOCKED) 225 #define VN_KNOTE_UNLOCKED(vp, b) VN_KNOTE(vp, b, 0) 226 227 /* 228 * Vnode flags. 229 * VI flags are protected by interlock and live in v_iflag 230 * VV flags are protected by the vnode lock and live in v_vflag 231 * 232 * VI_DOOMED is doubly protected by the interlock and vnode lock. Both 233 * are required for writing but the status may be checked with either. 234 */ 235 #define VI_MOUNT 0x0020 /* Mount in progress */ 236 #define VI_DOOMED 0x0080 /* This vnode is being recycled */ 237 #define VI_FREE 0x0100 /* This vnode is on the freelist */ 238 #define VI_ACTIVE 0x0200 /* This vnode is on the active list */ 239 #define VI_DOINGINACT 0x0800 /* VOP_INACTIVE is in progress */ 240 #define VI_OWEINACT 0x1000 /* Need to call inactive */ 241 242 #define VV_ROOT 0x0001 /* root of its filesystem */ 243 #define VV_ISTTY 0x0002 /* vnode represents a tty */ 244 #define VV_NOSYNC 0x0004 /* unlinked, stop syncing */ 245 #define VV_ETERNALDEV 0x0008 /* device that is never destroyed */ 246 #define VV_CACHEDLABEL 0x0010 /* Vnode has valid cached MAC label */ 247 #define VV_TEXT 0x0020 /* vnode is a pure text prototype */ 248 #define VV_COPYONWRITE 0x0040 /* vnode is doing copy-on-write */ 249 #define VV_SYSTEM 0x0080 /* vnode being used by kernel */ 250 #define VV_PROCDEP 0x0100 /* vnode is process dependent */ 251 #define VV_NOKNOTE 0x0200 /* don't activate knotes on this vnode */ 252 #define VV_DELETED 0x0400 /* should be removed */ 253 #define VV_MD 0x0800 /* vnode backs the md device */ 254 #define VV_FORCEINSMQ 0x1000 /* force the insmntque to succeed */ 255 #define VV_READLINK 0x2000 /* fdescfs linux vnode */ 256 257 #define VMP_TMPMNTFREELIST 0x0001 /* Vnode is on mnt's tmp free list */ 258 259 /* 260 * Vnode attributes. A field value of VNOVAL represents a field whose value 261 * is unavailable (getattr) or which is not to be changed (setattr). 262 */ 263 struct vattr { 264 enum vtype va_type; /* vnode type (for create) */ 265 u_short va_mode; /* files access mode and type */ 266 u_short va_padding0; 267 uid_t va_uid; /* owner user id */ 268 gid_t va_gid; /* owner group id */ 269 nlink_t va_nlink; /* number of references to file */ 270 dev_t va_fsid; /* filesystem id */ 271 ino_t va_fileid; /* file id */ 272 u_quad_t va_size; /* file size in bytes */ 273 long va_blocksize; /* blocksize preferred for i/o */ 274 struct timespec va_atime; /* time of last access */ 275 struct timespec va_mtime; /* time of last modification */ 276 struct timespec va_ctime; /* time file changed */ 277 struct timespec va_birthtime; /* time file created */ 278 u_long va_gen; /* generation number of file */ 279 u_long va_flags; /* flags defined for file */ 280 dev_t va_rdev; /* device the special file represents */ 281 u_quad_t va_bytes; /* bytes of disk space held by file */ 282 u_quad_t va_filerev; /* file modification number */ 283 u_int va_vaflags; /* operations flags, see below */ 284 long va_spare; /* remain quad aligned */ 285 }; 286 287 /* 288 * Flags for va_vaflags. 289 */ 290 #define VA_UTIMES_NULL 0x01 /* utimes argument was NULL */ 291 #define VA_EXCLUSIVE 0x02 /* exclusive create request */ 292 #define VA_SYNC 0x04 /* O_SYNC truncation */ 293 294 /* 295 * Flags for ioflag. (high 16 bits used to ask for read-ahead and 296 * help with write clustering) 297 * NB: IO_NDELAY and IO_DIRECT are linked to fcntl.h 298 */ 299 #define IO_UNIT 0x0001 /* do I/O as atomic unit */ 300 #define IO_APPEND 0x0002 /* append write to end */ 301 #define IO_NDELAY 0x0004 /* FNDELAY flag set in file table */ 302 #define IO_NODELOCKED 0x0008 /* underlying node already locked */ 303 #define IO_ASYNC 0x0010 /* bawrite rather then bdwrite */ 304 #define IO_VMIO 0x0020 /* data already in VMIO space */ 305 #define IO_INVAL 0x0040 /* invalidate after I/O */ 306 #define IO_SYNC 0x0080 /* do I/O synchronously */ 307 #define IO_DIRECT 0x0100 /* attempt to bypass buffer cache */ 308 #define IO_NOREUSE 0x0200 /* VMIO data won't be reused */ 309 #define IO_EXT 0x0400 /* operate on external attributes */ 310 #define IO_NORMAL 0x0800 /* operate on regular data */ 311 #define IO_NOMACCHECK 0x1000 /* MAC checks unnecessary */ 312 #define IO_BUFLOCKED 0x2000 /* ffs flag; indir buf is locked */ 313 #define IO_RANGELOCKED 0x4000 /* range locked */ 314 315 #define IO_SEQMAX 0x7F /* seq heuristic max value */ 316 #define IO_SEQSHIFT 16 /* seq heuristic in upper 16 bits */ 317 318 /* 319 * Flags for accmode_t. 320 */ 321 #define VEXEC 000000000100 /* execute/search permission */ 322 #define VWRITE 000000000200 /* write permission */ 323 #define VREAD 000000000400 /* read permission */ 324 #define VADMIN 000000010000 /* being the file owner */ 325 #define VAPPEND 000000040000 /* permission to write/append */ 326 /* 327 * VEXPLICIT_DENY makes VOP_ACCESSX(9) return EPERM or EACCES only 328 * if permission was denied explicitly, by a "deny" rule in NFSv4 ACL, 329 * and 0 otherwise. This never happens with ordinary unix access rights 330 * or POSIX.1e ACLs. Obviously, VEXPLICIT_DENY must be OR-ed with 331 * some other V* constant. 332 */ 333 #define VEXPLICIT_DENY 000000100000 334 #define VREAD_NAMED_ATTRS 000000200000 /* not used */ 335 #define VWRITE_NAMED_ATTRS 000000400000 /* not used */ 336 #define VDELETE_CHILD 000001000000 337 #define VREAD_ATTRIBUTES 000002000000 /* permission to stat(2) */ 338 #define VWRITE_ATTRIBUTES 000004000000 /* change {m,c,a}time */ 339 #define VDELETE 000010000000 340 #define VREAD_ACL 000020000000 /* read ACL and file mode */ 341 #define VWRITE_ACL 000040000000 /* change ACL and/or file mode */ 342 #define VWRITE_OWNER 000100000000 /* change file owner */ 343 #define VSYNCHRONIZE 000200000000 /* not used */ 344 #define VCREAT 000400000000 /* creating new file */ 345 #define VVERIFY 001000000000 /* verification required */ 346 347 /* 348 * Permissions that were traditionally granted only to the file owner. 349 */ 350 #define VADMIN_PERMS (VADMIN | VWRITE_ATTRIBUTES | VWRITE_ACL | \ 351 VWRITE_OWNER) 352 353 /* 354 * Permissions that were traditionally granted to everyone. 355 */ 356 #define VSTAT_PERMS (VREAD_ATTRIBUTES | VREAD_ACL) 357 358 /* 359 * Permissions that allow to change the state of the file in any way. 360 */ 361 #define VMODIFY_PERMS (VWRITE | VAPPEND | VADMIN_PERMS | VDELETE_CHILD | \ 362 VDELETE) 363 364 /* 365 * Token indicating no attribute value yet assigned. 366 */ 367 #define VNOVAL (-1) 368 369 /* 370 * LK_TIMELOCK timeout for vnode locks (used mainly by the pageout daemon) 371 */ 372 #define VLKTIMEOUT (hz / 20 + 1) 373 374 #ifdef _KERNEL 375 376 #ifdef MALLOC_DECLARE 377 MALLOC_DECLARE(M_VNODE); 378 #endif 379 380 extern u_int ncsizefactor; 381 382 /* 383 * Convert between vnode types and inode formats (since POSIX.1 384 * defines mode word of stat structure in terms of inode formats). 385 */ 386 extern enum vtype iftovt_tab[]; 387 extern int vttoif_tab[]; 388 #define IFTOVT(mode) (iftovt_tab[((mode) & S_IFMT) >> 12]) 389 #define VTTOIF(indx) (vttoif_tab[(int)(indx)]) 390 #define MAKEIMODE(indx, mode) (int)(VTTOIF(indx) | (mode)) 391 392 /* 393 * Flags to various vnode functions. 394 */ 395 #define SKIPSYSTEM 0x0001 /* vflush: skip vnodes marked VSYSTEM */ 396 #define FORCECLOSE 0x0002 /* vflush: force file closure */ 397 #define WRITECLOSE 0x0004 /* vflush: only close writable files */ 398 #define EARLYFLUSH 0x0008 /* vflush: early call for ffs_flushfiles */ 399 #define V_SAVE 0x0001 /* vinvalbuf: sync file first */ 400 #define V_ALT 0x0002 /* vinvalbuf: invalidate only alternate bufs */ 401 #define V_NORMAL 0x0004 /* vinvalbuf: invalidate only regular bufs */ 402 #define V_CLEANONLY 0x0008 /* vinvalbuf: invalidate only clean bufs */ 403 #define V_VMIO 0x0010 /* vinvalbuf: called during pageout */ 404 #define V_ALLOWCLEAN 0x0020 /* vinvalbuf: allow clean buffers after flush */ 405 #define REVOKEALL 0x0001 /* vop_revoke: revoke all aliases */ 406 #define V_WAIT 0x0001 /* vn_start_write: sleep for suspend */ 407 #define V_NOWAIT 0x0002 /* vn_start_write: don't sleep for suspend */ 408 #define V_XSLEEP 0x0004 /* vn_start_write: just return after sleep */ 409 #define V_MNTREF 0x0010 /* vn_start_write: mp is already ref-ed */ 410 411 #define VR_START_WRITE 0x0001 /* vfs_write_resume: start write atomically */ 412 #define VR_NO_SUSPCLR 0x0002 /* vfs_write_resume: do not clear suspension */ 413 414 #define VS_SKIP_UNMOUNT 0x0001 /* vfs_write_suspend: fail if the 415 filesystem is being unmounted */ 416 417 #define VREF(vp) vref(vp) 418 419 #ifdef DIAGNOSTIC 420 #define VATTR_NULL(vap) vattr_null(vap) 421 #else 422 #define VATTR_NULL(vap) (*(vap) = va_null) /* initialize a vattr */ 423 #endif /* DIAGNOSTIC */ 424 425 #define NULLVP ((struct vnode *)NULL) 426 427 /* 428 * Global vnode data. 429 */ 430 extern struct vnode *rootvnode; /* root (i.e. "/") vnode */ 431 extern struct mount *rootdevmp; /* "/dev" mount */ 432 extern int desiredvnodes; /* number of vnodes desired */ 433 extern struct uma_zone *namei_zone; 434 extern struct vattr va_null; /* predefined null vattr structure */ 435 436 #define VI_LOCK(vp) mtx_lock(&(vp)->v_interlock) 437 #define VI_LOCK_FLAGS(vp, flags) mtx_lock_flags(&(vp)->v_interlock, (flags)) 438 #define VI_TRYLOCK(vp) mtx_trylock(&(vp)->v_interlock) 439 #define VI_UNLOCK(vp) mtx_unlock(&(vp)->v_interlock) 440 #define VI_MTX(vp) (&(vp)->v_interlock) 441 442 #define VN_LOCK_AREC(vp) lockallowrecurse((vp)->v_vnlock) 443 #define VN_LOCK_ASHARE(vp) lockallowshare((vp)->v_vnlock) 444 #define VN_LOCK_DSHARE(vp) lockdisableshare((vp)->v_vnlock) 445 446 #endif /* _KERNEL */ 447 448 /* 449 * Mods for extensibility. 450 */ 451 452 /* 453 * Flags for vdesc_flags: 454 */ 455 #define VDESC_MAX_VPS 16 456 /* Low order 16 flag bits are reserved for willrele flags for vp arguments. */ 457 #define VDESC_VP0_WILLRELE 0x0001 458 #define VDESC_VP1_WILLRELE 0x0002 459 #define VDESC_VP2_WILLRELE 0x0004 460 #define VDESC_VP3_WILLRELE 0x0008 461 #define VDESC_NOMAP_VPP 0x0100 462 #define VDESC_VPP_WILLRELE 0x0200 463 464 /* 465 * A generic structure. 466 * This can be used by bypass routines to identify generic arguments. 467 */ 468 struct vop_generic_args { 469 struct vnodeop_desc *a_desc; 470 /* other random data follows, presumably */ 471 }; 472 473 typedef int vop_bypass_t(struct vop_generic_args *); 474 475 /* 476 * VDESC_NO_OFFSET is used to identify the end of the offset list 477 * and in places where no such field exists. 478 */ 479 #define VDESC_NO_OFFSET -1 480 481 /* 482 * This structure describes the vnode operation taking place. 483 */ 484 struct vnodeop_desc { 485 char *vdesc_name; /* a readable name for debugging */ 486 int vdesc_flags; /* VDESC_* flags */ 487 int vdesc_vop_offset; 488 vop_bypass_t *vdesc_call; /* Function to call */ 489 490 /* 491 * These ops are used by bypass routines to map and locate arguments. 492 * Creds and procs are not needed in bypass routines, but sometimes 493 * they are useful to (for example) transport layers. 494 * Nameidata is useful because it has a cred in it. 495 */ 496 int *vdesc_vp_offsets; /* list ended by VDESC_NO_OFFSET */ 497 int vdesc_vpp_offset; /* return vpp location */ 498 int vdesc_cred_offset; /* cred location, if any */ 499 int vdesc_thread_offset; /* thread location, if any */ 500 int vdesc_componentname_offset; /* if any */ 501 }; 502 503 #ifdef _KERNEL 504 /* 505 * A list of all the operation descs. 506 */ 507 extern struct vnodeop_desc *vnodeop_descs[]; 508 509 #define VOPARG_OFFSETOF(s_type, field) __offsetof(s_type, field) 510 #define VOPARG_OFFSETTO(s_type, s_offset, struct_p) \ 511 ((s_type)(((char*)(struct_p)) + (s_offset))) 512 513 514 #ifdef DEBUG_VFS_LOCKS 515 /* 516 * Support code to aid in debugging VFS locking problems. Not totally 517 * reliable since if the thread sleeps between changing the lock 518 * state and checking it with the assert, some other thread could 519 * change the state. They are good enough for debugging a single 520 * filesystem using a single-threaded test. Note that the unreliability is 521 * limited to false negatives; efforts were made to ensure that false 522 * positives cannot occur. 523 */ 524 void assert_vi_locked(struct vnode *vp, const char *str); 525 void assert_vi_unlocked(struct vnode *vp, const char *str); 526 void assert_vop_elocked(struct vnode *vp, const char *str); 527 void assert_vop_locked(struct vnode *vp, const char *str); 528 void assert_vop_unlocked(struct vnode *vp, const char *str); 529 530 #define ASSERT_VI_LOCKED(vp, str) assert_vi_locked((vp), (str)) 531 #define ASSERT_VI_UNLOCKED(vp, str) assert_vi_unlocked((vp), (str)) 532 #define ASSERT_VOP_ELOCKED(vp, str) assert_vop_elocked((vp), (str)) 533 #define ASSERT_VOP_LOCKED(vp, str) assert_vop_locked((vp), (str)) 534 #define ASSERT_VOP_UNLOCKED(vp, str) assert_vop_unlocked((vp), (str)) 535 536 #else /* !DEBUG_VFS_LOCKS */ 537 538 #define ASSERT_VI_LOCKED(vp, str) ((void)0) 539 #define ASSERT_VI_UNLOCKED(vp, str) ((void)0) 540 #define ASSERT_VOP_ELOCKED(vp, str) ((void)0) 541 #define ASSERT_VOP_LOCKED(vp, str) ((void)0) 542 #define ASSERT_VOP_UNLOCKED(vp, str) ((void)0) 543 #endif /* DEBUG_VFS_LOCKS */ 544 545 546 /* 547 * This call works for vnodes in the kernel. 548 */ 549 #define VCALL(c) ((c)->a_desc->vdesc_call(c)) 550 551 #define DOINGASYNC(vp) \ 552 (((vp)->v_mount->mnt_kern_flag & MNTK_ASYNC) != 0 && \ 553 ((curthread->td_pflags & TDP_SYNCIO) == 0)) 554 555 /* 556 * VMIO support inline 557 */ 558 559 extern int vmiodirenable; 560 561 static __inline int 562 vn_canvmio(struct vnode *vp) 563 { 564 if (vp && (vp->v_type == VREG || (vmiodirenable && vp->v_type == VDIR))) 565 return(TRUE); 566 return(FALSE); 567 } 568 569 /* 570 * Finally, include the default set of vnode operations. 571 */ 572 typedef void vop_getpages_iodone_t(void *, vm_page_t *, int, int); 573 #include "vnode_if.h" 574 575 /* vn_open_flags */ 576 #define VN_OPEN_NOAUDIT 0x00000001 577 #define VN_OPEN_NOCAPCHECK 0x00000002 578 #define VN_OPEN_NAMECACHE 0x00000004 579 580 /* 581 * Public vnode manipulation functions. 582 */ 583 struct componentname; 584 struct file; 585 struct mount; 586 struct nameidata; 587 struct ostat; 588 struct freebsd11_stat; 589 struct thread; 590 struct proc; 591 struct stat; 592 struct nstat; 593 struct ucred; 594 struct uio; 595 struct vattr; 596 struct vfsops; 597 struct vnode; 598 599 typedef int (*vn_get_ino_t)(struct mount *, void *, int, struct vnode **); 600 601 int bnoreuselist(struct bufv *bufv, struct bufobj *bo, daddr_t startn, 602 daddr_t endn); 603 /* cache_* may belong in namei.h. */ 604 void cache_changesize(int newhashsize); 605 #define cache_enter(dvp, vp, cnp) \ 606 cache_enter_time(dvp, vp, cnp, NULL, NULL) 607 void cache_enter_time(struct vnode *dvp, struct vnode *vp, 608 struct componentname *cnp, struct timespec *tsp, 609 struct timespec *dtsp); 610 int cache_lookup(struct vnode *dvp, struct vnode **vpp, 611 struct componentname *cnp, struct timespec *tsp, int *ticksp); 612 void cache_purge(struct vnode *vp); 613 void cache_purge_negative(struct vnode *vp); 614 void cache_purgevfs(struct mount *mp, bool force); 615 int change_dir(struct vnode *vp, struct thread *td); 616 void cvtstat(struct stat *st, struct ostat *ost); 617 void freebsd11_cvtnstat(struct stat *sb, struct nstat *nsb); 618 int freebsd11_cvtstat(struct stat *st, struct freebsd11_stat *ost); 619 int getnewvnode(const char *tag, struct mount *mp, struct vop_vector *vops, 620 struct vnode **vpp); 621 void getnewvnode_reserve(u_int count); 622 void getnewvnode_drop_reserve(void); 623 int insmntque1(struct vnode *vp, struct mount *mp, 624 void (*dtr)(struct vnode *, void *), void *dtr_arg); 625 int insmntque(struct vnode *vp, struct mount *mp); 626 u_quad_t init_va_filerev(void); 627 int speedup_syncer(void); 628 int vn_vptocnp(struct vnode **vp, struct ucred *cred, char *buf, 629 u_int *buflen); 630 int vn_fullpath(struct thread *td, struct vnode *vn, 631 char **retbuf, char **freebuf); 632 int vn_fullpath_global(struct thread *td, struct vnode *vn, 633 char **retbuf, char **freebuf); 634 struct vnode * 635 vn_dir_dd_ino(struct vnode *vp); 636 int vn_commname(struct vnode *vn, char *buf, u_int buflen); 637 int vn_path_to_global_path(struct thread *td, struct vnode *vp, 638 char *path, u_int pathlen); 639 int vaccess(enum vtype type, mode_t file_mode, uid_t file_uid, 640 gid_t file_gid, accmode_t accmode, struct ucred *cred, 641 int *privused); 642 int vaccess_acl_nfs4(enum vtype type, uid_t file_uid, gid_t file_gid, 643 struct acl *aclp, accmode_t accmode, struct ucred *cred, 644 int *privused); 645 int vaccess_acl_posix1e(enum vtype type, uid_t file_uid, 646 gid_t file_gid, struct acl *acl, accmode_t accmode, 647 struct ucred *cred, int *privused); 648 void vattr_null(struct vattr *vap); 649 int vcount(struct vnode *vp); 650 #define vdrop(vp) _vdrop((vp), 0) 651 #define vdropl(vp) _vdrop((vp), 1) 652 void _vdrop(struct vnode *, bool); 653 int vflush(struct mount *mp, int rootrefs, int flags, struct thread *td); 654 int vget(struct vnode *vp, int lockflag, struct thread *td); 655 void vgone(struct vnode *vp); 656 #define vhold(vp) _vhold((vp), 0) 657 #define vholdl(vp) _vhold((vp), 1) 658 void _vhold(struct vnode *, bool); 659 void vinactive(struct vnode *, struct thread *); 660 int vinvalbuf(struct vnode *vp, int save, int slpflag, int slptimeo); 661 int vtruncbuf(struct vnode *vp, struct ucred *cred, off_t length, 662 int blksize); 663 void vunref(struct vnode *); 664 void vn_printf(struct vnode *vp, const char *fmt, ...) __printflike(2,3); 665 int vrecycle(struct vnode *vp); 666 int vrecyclel(struct vnode *vp); 667 int vn_bmap_seekhole(struct vnode *vp, u_long cmd, off_t *off, 668 struct ucred *cred); 669 int vn_close(struct vnode *vp, 670 int flags, struct ucred *file_cred, struct thread *td); 671 void vn_finished_write(struct mount *mp); 672 void vn_finished_secondary_write(struct mount *mp); 673 int vn_fsync_buf(struct vnode *vp, int waitfor); 674 int vn_isdisk(struct vnode *vp, int *errp); 675 int _vn_lock(struct vnode *vp, int flags, char *file, int line); 676 #define vn_lock(vp, flags) _vn_lock(vp, flags, __FILE__, __LINE__) 677 int vn_open(struct nameidata *ndp, int *flagp, int cmode, struct file *fp); 678 int vn_open_cred(struct nameidata *ndp, int *flagp, int cmode, 679 u_int vn_open_flags, struct ucred *cred, struct file *fp); 680 int vn_open_vnode(struct vnode *vp, int fmode, struct ucred *cred, 681 struct thread *td, struct file *fp); 682 void vn_pages_remove(struct vnode *vp, vm_pindex_t start, vm_pindex_t end); 683 int vn_pollrecord(struct vnode *vp, struct thread *p, int events); 684 int vn_rdwr(enum uio_rw rw, struct vnode *vp, void *base, 685 int len, off_t offset, enum uio_seg segflg, int ioflg, 686 struct ucred *active_cred, struct ucred *file_cred, ssize_t *aresid, 687 struct thread *td); 688 int vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, void *base, 689 size_t len, off_t offset, enum uio_seg segflg, int ioflg, 690 struct ucred *active_cred, struct ucred *file_cred, size_t *aresid, 691 struct thread *td); 692 int vn_rlimit_fsize(const struct vnode *vn, const struct uio *uio, 693 struct thread *td); 694 int vn_stat(struct vnode *vp, struct stat *sb, struct ucred *active_cred, 695 struct ucred *file_cred, struct thread *td); 696 int vn_start_write(struct vnode *vp, struct mount **mpp, int flags); 697 int vn_start_secondary_write(struct vnode *vp, struct mount **mpp, 698 int flags); 699 int vn_writechk(struct vnode *vp); 700 int vn_extattr_get(struct vnode *vp, int ioflg, int attrnamespace, 701 const char *attrname, int *buflen, char *buf, struct thread *td); 702 int vn_extattr_set(struct vnode *vp, int ioflg, int attrnamespace, 703 const char *attrname, int buflen, char *buf, struct thread *td); 704 int vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace, 705 const char *attrname, struct thread *td); 706 int vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags, 707 struct vnode **rvp); 708 int vn_vget_ino_gen(struct vnode *vp, vn_get_ino_t alloc, 709 void *alloc_arg, int lkflags, struct vnode **rvp); 710 int vn_utimes_perm(struct vnode *vp, struct vattr *vap, 711 struct ucred *cred, struct thread *td); 712 713 int vn_io_fault_uiomove(char *data, int xfersize, struct uio *uio); 714 int vn_io_fault_pgmove(vm_page_t ma[], vm_offset_t offset, int xfersize, 715 struct uio *uio); 716 717 #define vn_rangelock_unlock(vp, cookie) \ 718 rangelock_unlock(&(vp)->v_rl, (cookie), VI_MTX(vp)) 719 #define vn_rangelock_unlock_range(vp, cookie, start, end) \ 720 rangelock_unlock_range(&(vp)->v_rl, (cookie), (start), (end), \ 721 VI_MTX(vp)) 722 #define vn_rangelock_rlock(vp, start, end) \ 723 rangelock_rlock(&(vp)->v_rl, (start), (end), VI_MTX(vp)) 724 #define vn_rangelock_wlock(vp, start, end) \ 725 rangelock_wlock(&(vp)->v_rl, (start), (end), VI_MTX(vp)) 726 727 int vfs_cache_lookup(struct vop_lookup_args *ap); 728 void vfs_timestamp(struct timespec *); 729 void vfs_write_resume(struct mount *mp, int flags); 730 int vfs_write_suspend(struct mount *mp, int flags); 731 int vfs_write_suspend_umnt(struct mount *mp); 732 void vnlru_free(int, struct vfsops *); 733 int vop_stdbmap(struct vop_bmap_args *); 734 int vop_stdfdatasync_buf(struct vop_fdatasync_args *); 735 int vop_stdfsync(struct vop_fsync_args *); 736 int vop_stdgetwritemount(struct vop_getwritemount_args *); 737 int vop_stdgetpages(struct vop_getpages_args *); 738 int vop_stdinactive(struct vop_inactive_args *); 739 int vop_stdislocked(struct vop_islocked_args *); 740 int vop_stdkqfilter(struct vop_kqfilter_args *); 741 int vop_stdlock(struct vop_lock1_args *); 742 int vop_stdputpages(struct vop_putpages_args *); 743 int vop_stdunlock(struct vop_unlock_args *); 744 int vop_nopoll(struct vop_poll_args *); 745 int vop_stdaccess(struct vop_access_args *ap); 746 int vop_stdaccessx(struct vop_accessx_args *ap); 747 int vop_stdadvise(struct vop_advise_args *ap); 748 int vop_stdadvlock(struct vop_advlock_args *ap); 749 int vop_stdadvlockasync(struct vop_advlockasync_args *ap); 750 int vop_stdadvlockpurge(struct vop_advlockpurge_args *ap); 751 int vop_stdallocate(struct vop_allocate_args *ap); 752 int vop_stdpathconf(struct vop_pathconf_args *); 753 int vop_stdpoll(struct vop_poll_args *); 754 int vop_stdvptocnp(struct vop_vptocnp_args *ap); 755 int vop_stdvptofh(struct vop_vptofh_args *ap); 756 int vop_stdunp_bind(struct vop_unp_bind_args *ap); 757 int vop_stdunp_connect(struct vop_unp_connect_args *ap); 758 int vop_stdunp_detach(struct vop_unp_detach_args *ap); 759 int vop_eopnotsupp(struct vop_generic_args *ap); 760 int vop_ebadf(struct vop_generic_args *ap); 761 int vop_einval(struct vop_generic_args *ap); 762 int vop_enoent(struct vop_generic_args *ap); 763 int vop_enotty(struct vop_generic_args *ap); 764 int vop_null(struct vop_generic_args *ap); 765 int vop_panic(struct vop_generic_args *ap); 766 int dead_poll(struct vop_poll_args *ap); 767 int dead_read(struct vop_read_args *ap); 768 int dead_write(struct vop_write_args *ap); 769 770 /* These are called from within the actual VOPS. */ 771 void vop_close_post(void *a, int rc); 772 void vop_create_post(void *a, int rc); 773 void vop_deleteextattr_post(void *a, int rc); 774 void vop_link_post(void *a, int rc); 775 void vop_lookup_post(void *a, int rc); 776 void vop_lookup_pre(void *a); 777 void vop_mkdir_post(void *a, int rc); 778 void vop_mknod_post(void *a, int rc); 779 void vop_open_post(void *a, int rc); 780 void vop_read_post(void *a, int rc); 781 void vop_readdir_post(void *a, int rc); 782 void vop_reclaim_post(void *a, int rc); 783 void vop_remove_post(void *a, int rc); 784 void vop_rename_post(void *a, int rc); 785 void vop_rename_pre(void *a); 786 void vop_rmdir_post(void *a, int rc); 787 void vop_setattr_post(void *a, int rc); 788 void vop_setextattr_post(void *a, int rc); 789 void vop_symlink_post(void *a, int rc); 790 int vop_sigdefer(struct vop_vector *vop, struct vop_generic_args *a); 791 792 #ifdef DEBUG_VFS_LOCKS 793 void vop_strategy_pre(void *a); 794 void vop_lock_pre(void *a); 795 void vop_lock_post(void *a, int rc); 796 void vop_unlock_post(void *a, int rc); 797 void vop_unlock_pre(void *a); 798 #else 799 #define vop_strategy_pre(x) do { } while (0) 800 #define vop_lock_pre(x) do { } while (0) 801 #define vop_lock_post(x, y) do { } while (0) 802 #define vop_unlock_post(x, y) do { } while (0) 803 #define vop_unlock_pre(x) do { } while (0) 804 #endif 805 806 void vop_rename_fail(struct vop_rename_args *ap); 807 808 #define VOP_WRITE_PRE(ap) \ 809 struct vattr va; \ 810 int error; \ 811 off_t osize, ooffset, noffset; \ 812 \ 813 osize = ooffset = noffset = 0; \ 814 if (!VN_KNLIST_EMPTY((ap)->a_vp)) { \ 815 error = VOP_GETATTR((ap)->a_vp, &va, (ap)->a_cred); \ 816 if (error) \ 817 return (error); \ 818 ooffset = (ap)->a_uio->uio_offset; \ 819 osize = (off_t)va.va_size; \ 820 } 821 822 #define VOP_WRITE_POST(ap, ret) \ 823 noffset = (ap)->a_uio->uio_offset; \ 824 if (noffset > ooffset && !VN_KNLIST_EMPTY((ap)->a_vp)) { \ 825 VFS_KNOTE_LOCKED((ap)->a_vp, NOTE_WRITE \ 826 | (noffset > osize ? NOTE_EXTEND : 0)); \ 827 } 828 829 #define VOP_LOCK(vp, flags) VOP_LOCK1(vp, flags, __FILE__, __LINE__) 830 831 832 void vput(struct vnode *vp); 833 void vrele(struct vnode *vp); 834 void vref(struct vnode *vp); 835 void vrefl(struct vnode *vp); 836 void vrefact(struct vnode *vp); 837 int vrefcnt(struct vnode *vp); 838 void v_addpollinfo(struct vnode *vp); 839 840 int vnode_create_vobject(struct vnode *vp, off_t size, struct thread *td); 841 void vnode_destroy_vobject(struct vnode *vp); 842 843 extern struct vop_vector fifo_specops; 844 extern struct vop_vector dead_vnodeops; 845 extern struct vop_vector default_vnodeops; 846 847 #define VOP_PANIC ((void*)(uintptr_t)vop_panic) 848 #define VOP_NULL ((void*)(uintptr_t)vop_null) 849 #define VOP_EBADF ((void*)(uintptr_t)vop_ebadf) 850 #define VOP_ENOTTY ((void*)(uintptr_t)vop_enotty) 851 #define VOP_EINVAL ((void*)(uintptr_t)vop_einval) 852 #define VOP_ENOENT ((void*)(uintptr_t)vop_enoent) 853 #define VOP_EOPNOTSUPP ((void*)(uintptr_t)vop_eopnotsupp) 854 855 /* fifo_vnops.c */ 856 int fifo_printinfo(struct vnode *); 857 858 /* vfs_hash.c */ 859 typedef int vfs_hash_cmp_t(struct vnode *vp, void *arg); 860 861 void vfs_hash_changesize(int newhashsize); 862 int vfs_hash_get(const struct mount *mp, u_int hash, int flags, 863 struct thread *td, struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg); 864 u_int vfs_hash_index(struct vnode *vp); 865 int vfs_hash_insert(struct vnode *vp, u_int hash, int flags, struct thread *td, 866 struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg); 867 void vfs_hash_ref(const struct mount *mp, u_int hash, struct thread *td, 868 struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg); 869 void vfs_hash_rehash(struct vnode *vp, u_int hash); 870 void vfs_hash_remove(struct vnode *vp); 871 872 int vfs_kqfilter(struct vop_kqfilter_args *); 873 void vfs_mark_atime(struct vnode *vp, struct ucred *cred); 874 struct dirent; 875 int vfs_read_dirent(struct vop_readdir_args *ap, struct dirent *dp, off_t off); 876 877 int vfs_unixify_accmode(accmode_t *accmode); 878 879 void vfs_unp_reclaim(struct vnode *vp); 880 881 int setfmode(struct thread *td, struct ucred *cred, struct vnode *vp, int mode); 882 int setfown(struct thread *td, struct ucred *cred, struct vnode *vp, uid_t uid, 883 gid_t gid); 884 int vn_chmod(struct file *fp, mode_t mode, struct ucred *active_cred, 885 struct thread *td); 886 int vn_chown(struct file *fp, uid_t uid, gid_t gid, struct ucred *active_cred, 887 struct thread *td); 888 889 void vn_fsid(struct vnode *vp, struct vattr *va); 890 891 #endif /* _KERNEL */ 892 893 #endif /* !_SYS_VNODE_H_ */ 894