1 /* $NetBSD: tmpfs.h,v 1.26 2007/02/22 06:37:00 thorpej Exp $ */ 2 3 /*- 4 * Copyright (c) 2005, 2006 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Julio M. Merino Vidal, developed as part of Google's Summer of Code 9 * 2005 program. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 * 32 * $FreeBSD$ 33 */ 34 35 #ifndef _FS_TMPFS_TMPFS_H_ 36 #define _FS_TMPFS_TMPFS_H_ 37 38 /* --------------------------------------------------------------------- 39 * KERNEL-SPECIFIC DEFINITIONS 40 * --------------------------------------------------------------------- */ 41 #include <sys/dirent.h> 42 #include <sys/mount.h> 43 #include <sys/queue.h> 44 #include <sys/vnode.h> 45 #include <sys/file.h> 46 #include <sys/lock.h> 47 #include <sys/mutex.h> 48 49 /* --------------------------------------------------------------------- */ 50 #include <sys/malloc.h> 51 #include <sys/systm.h> 52 #include <sys/tree.h> 53 #include <sys/vmmeter.h> 54 #include <vm/swap_pager.h> 55 56 MALLOC_DECLARE(M_TMPFSMNT); 57 MALLOC_DECLARE(M_TMPFSNAME); 58 59 /* --------------------------------------------------------------------- */ 60 61 /* 62 * Internal representation of a tmpfs directory entry. 63 */ 64 65 LIST_HEAD(tmpfs_dir_duphead, tmpfs_dirent); 66 67 struct tmpfs_dirent { 68 /* 69 * Depending on td_cookie flag entry can be of 3 types: 70 * - regular -- no hash collisions, stored in RB-Tree 71 * - duphead -- synthetic linked list head for dup entries 72 * - dup -- stored in linked list instead of RB-Tree 73 */ 74 union { 75 /* regular and duphead entry types */ 76 RB_ENTRY(tmpfs_dirent) td_entries; 77 78 /* dup entry type */ 79 struct { 80 LIST_ENTRY(tmpfs_dirent) entries; 81 LIST_ENTRY(tmpfs_dirent) index_entries; 82 } td_dup; 83 } uh; 84 85 uint32_t td_cookie; 86 uint32_t td_hash; 87 u_int td_namelen; 88 89 /* Pointer to the node this entry refers to. In case this field 90 * is NULL, the node is a whiteout. */ 91 struct tmpfs_node * td_node; 92 93 union { 94 /* 95 * The name of the entry, allocated from a string pool. This 96 * string is not required to be zero-terminated. 97 */ 98 char * td_name; /* regular, dup */ 99 struct tmpfs_dir_duphead td_duphead; /* duphead */ 100 } ud; 101 }; 102 103 /* A directory in tmpfs holds a list of directory entries, which in 104 * turn point to other files (which can be directories themselves). 105 * 106 * In tmpfs, this list is managed by a RB-Tree, whose head is defined by 107 * the struct tmpfs_dir type. 108 * 109 * It is important to notice that directories do not have entries for . and 110 * .. as other file systems do. These can be generated when requested 111 * based on information available by other means, such as the pointer to 112 * the node itself in the former case or the pointer to the parent directory 113 * in the latter case. This is done to simplify tmpfs's code and, more 114 * importantly, to remove redundancy. */ 115 RB_HEAD(tmpfs_dir, tmpfs_dirent); 116 117 /* Each entry in a directory has a cookie that identifies it. Cookies 118 * supersede offsets within directories because, given how tmpfs stores 119 * directories in memory, there is no such thing as an offset. 120 * 121 * The '.', '..' and the end of directory markers have fixed cookies which 122 * cannot collide with the cookies generated by other entries. The cookies 123 * for the other entries are generated based on the file name hash value or 124 * unique number in case of name hash collision. 125 * 126 * To preserve compatibility cookies are limited to 31 bits. 127 */ 128 129 #define TMPFS_DIRCOOKIE_DOT 0 130 #define TMPFS_DIRCOOKIE_DOTDOT 1 131 #define TMPFS_DIRCOOKIE_EOF 2 132 #define TMPFS_DIRCOOKIE_MASK ((off_t)0x3fffffffU) 133 #define TMPFS_DIRCOOKIE_MIN ((off_t)0x00000004U) 134 #define TMPFS_DIRCOOKIE_DUP ((off_t)0x40000000U) 135 #define TMPFS_DIRCOOKIE_DUPHEAD ((off_t)0x80000000U) 136 #define TMPFS_DIRCOOKIE_DUP_MIN TMPFS_DIRCOOKIE_DUP 137 #define TMPFS_DIRCOOKIE_DUP_MAX \ 138 (TMPFS_DIRCOOKIE_DUP | TMPFS_DIRCOOKIE_MASK) 139 140 /* --------------------------------------------------------------------- */ 141 142 /* 143 * Internal representation of a tmpfs file system node. 144 * 145 * This structure is splitted in two parts: one holds attributes common 146 * to all file types and the other holds data that is only applicable to 147 * a particular type. The code must be careful to only access those 148 * attributes that are actually allowed by the node's type. 149 * 150 * 151 * Below is the key of locks used to protected the fields in the following 152 * structures. 153 * 154 */ 155 struct tmpfs_node { 156 /* Doubly-linked list entry which links all existing nodes for a 157 * single file system. This is provided to ease the removal of 158 * all nodes during the unmount operation. */ 159 LIST_ENTRY(tmpfs_node) tn_entries; 160 161 /* The node's type. Any of 'VBLK', 'VCHR', 'VDIR', 'VFIFO', 162 * 'VLNK', 'VREG' and 'VSOCK' is allowed. The usage of vnode 163 * types instead of a custom enumeration is to make things simpler 164 * and faster, as we do not need to convert between two types. */ 165 enum vtype tn_type; 166 167 /* Node identifier. */ 168 ino_t tn_id; 169 170 /* Node's internal status. This is used by several file system 171 * operations to do modifications to the node in a delayed 172 * fashion. */ 173 int tn_status; 174 #define TMPFS_NODE_ACCESSED (1 << 1) 175 #define TMPFS_NODE_MODIFIED (1 << 2) 176 #define TMPFS_NODE_CHANGED (1 << 3) 177 178 /* The node size. It does not necessarily match the real amount 179 * of memory consumed by it. */ 180 off_t tn_size; 181 182 /* Generic node attributes. */ 183 uid_t tn_uid; 184 gid_t tn_gid; 185 mode_t tn_mode; 186 u_long tn_flags; 187 nlink_t tn_links; 188 struct timespec tn_atime; 189 struct timespec tn_mtime; 190 struct timespec tn_ctime; 191 struct timespec tn_birthtime; 192 unsigned long tn_gen; 193 194 /* As there is a single vnode for each active file within the 195 * system, care has to be taken to avoid allocating more than one 196 * vnode per file. In order to do this, a bidirectional association 197 * is kept between vnodes and nodes. 198 * 199 * Whenever a vnode is allocated, its v_data field is updated to 200 * point to the node it references. At the same time, the node's 201 * tn_vnode field is modified to point to the new vnode representing 202 * it. Further attempts to allocate a vnode for this same node will 203 * result in returning a new reference to the value stored in 204 * tn_vnode. 205 * 206 * May be NULL when the node is unused (that is, no vnode has been 207 * allocated for it or it has been reclaimed). */ 208 struct vnode * tn_vnode; 209 210 /* interlock to protect tn_vpstate */ 211 struct mtx tn_interlock; 212 213 /* Identify if current node has vnode assiocate with 214 * or allocating vnode. 215 */ 216 int tn_vpstate; 217 218 /* misc data field for different tn_type node */ 219 union { 220 /* Valid when tn_type == VBLK || tn_type == VCHR. */ 221 dev_t tn_rdev; 222 223 /* Valid when tn_type == VDIR. */ 224 struct tn_dir { 225 /* Pointer to the parent directory. The root 226 * directory has a pointer to itself in this field; 227 * this property identifies the root node. */ 228 struct tmpfs_node * tn_parent; 229 230 /* Head of a tree that links the contents of 231 * the directory together. */ 232 struct tmpfs_dir tn_dirhead; 233 234 /* Head of a list the contains fake directory entries 235 * heads, i.e. entries with TMPFS_DIRCOOKIE_DUPHEAD 236 * flag. */ 237 struct tmpfs_dir_duphead tn_dupindex; 238 239 /* Number and pointer of the first directory entry 240 * returned by the readdir operation if it were 241 * called again to continue reading data from the 242 * same directory as before. This is used to speed 243 * up reads of long directories, assuming that no 244 * more than one read is in progress at a given time. 245 * Otherwise, these values are discarded. */ 246 off_t tn_readdir_lastn; 247 struct tmpfs_dirent * tn_readdir_lastp; 248 } tn_dir; 249 250 /* Valid when tn_type == VLNK. */ 251 /* The link's target, allocated from a string pool. */ 252 char * tn_link; 253 254 /* Valid when tn_type == VREG. */ 255 struct tn_reg { 256 /* The contents of regular files stored in a tmpfs 257 * file system are represented by a single anonymous 258 * memory object (aobj, for short). The aobj provides 259 * direct access to any position within the file, 260 * because its contents are always mapped in a 261 * contiguous region of virtual memory. It is a task 262 * of the memory management subsystem (see uvm(9)) to 263 * issue the required page ins or page outs whenever 264 * a position within the file is accessed. */ 265 vm_object_t tn_aobj; 266 267 }tn_reg; 268 269 /* Valid when tn_type = VFIFO */ 270 struct tn_fifo { 271 fo_rdwr_t *tn_fo_read; 272 fo_rdwr_t *tn_fo_write; 273 }tn_fifo; 274 }tn_spec; 275 }; 276 LIST_HEAD(tmpfs_node_list, tmpfs_node); 277 278 #define tn_rdev tn_spec.tn_rdev 279 #define tn_dir tn_spec.tn_dir 280 #define tn_link tn_spec.tn_link 281 #define tn_reg tn_spec.tn_reg 282 #define tn_fifo tn_spec.tn_fifo 283 284 #define TMPFS_NODE_LOCK(node) mtx_lock(&(node)->tn_interlock) 285 #define TMPFS_NODE_UNLOCK(node) mtx_unlock(&(node)->tn_interlock) 286 #define TMPFS_NODE_MTX(node) (&(node)->tn_interlock) 287 288 #ifdef INVARIANTS 289 #define TMPFS_ASSERT_LOCKED(node) do { \ 290 MPASS(node != NULL); \ 291 MPASS(node->tn_vnode != NULL); \ 292 if (!VOP_ISLOCKED(node->tn_vnode) && \ 293 !mtx_owned(TMPFS_NODE_MTX(node))) \ 294 panic("tmpfs: node is not locked: %p", node); \ 295 } while (0) 296 #define TMPFS_ASSERT_ELOCKED(node) do { \ 297 MPASS((node) != NULL); \ 298 MPASS((node)->tn_vnode != NULL); \ 299 mtx_assert(TMPFS_NODE_MTX(node), MA_OWNED); \ 300 ASSERT_VOP_LOCKED((node)->tn_vnode, "tmpfs"); \ 301 } while (0) 302 #else 303 #define TMPFS_ASSERT_LOCKED(node) (void)0 304 #define TMPFS_ASSERT_ELOCKED(node) (void)0 305 #endif 306 307 #define TMPFS_VNODE_ALLOCATING 1 308 #define TMPFS_VNODE_WANT 2 309 #define TMPFS_VNODE_DOOMED 4 310 #define TMPFS_VNODE_WRECLAIM 8 311 /* --------------------------------------------------------------------- */ 312 313 /* 314 * Internal representation of a tmpfs mount point. 315 */ 316 struct tmpfs_mount { 317 /* Maximum number of memory pages available for use by the file 318 * system, set during mount time. This variable must never be 319 * used directly as it may be bigger than the current amount of 320 * free memory; in the extreme case, it will hold the SIZE_MAX 321 * value. */ 322 size_t tm_pages_max; 323 324 /* Number of pages in use by the file system. */ 325 size_t tm_pages_used; 326 327 /* Pointer to the node representing the root directory of this 328 * file system. */ 329 struct tmpfs_node * tm_root; 330 331 /* Maximum number of possible nodes for this file system; set 332 * during mount time. We need a hard limit on the maximum number 333 * of nodes to avoid allocating too much of them; their objects 334 * cannot be released until the file system is unmounted. 335 * Otherwise, we could easily run out of memory by creating lots 336 * of empty files and then simply removing them. */ 337 ino_t tm_nodes_max; 338 339 /* unrhdr used to allocate inode numbers */ 340 struct unrhdr * tm_ino_unr; 341 342 /* Number of nodes currently that are in use. */ 343 ino_t tm_nodes_inuse; 344 345 /* maximum representable file size */ 346 u_int64_t tm_maxfilesize; 347 348 /* Nodes are organized in two different lists. The used list 349 * contains all nodes that are currently used by the file system; 350 * i.e., they refer to existing files. The available list contains 351 * all nodes that are currently available for use by new files. 352 * Nodes must be kept in this list (instead of deleting them) 353 * because we need to keep track of their generation number (tn_gen 354 * field). 355 * 356 * Note that nodes are lazily allocated: if the available list is 357 * empty and we have enough space to create more nodes, they will be 358 * created and inserted in the used list. Once these are released, 359 * they will go into the available list, remaining alive until the 360 * file system is unmounted. */ 361 struct tmpfs_node_list tm_nodes_used; 362 363 /* All node lock to protect the node list and tmp_pages_used */ 364 struct mtx allnode_lock; 365 366 /* Pools used to store file system meta data. These are not shared 367 * across several instances of tmpfs for the reasons described in 368 * tmpfs_pool.c. */ 369 uma_zone_t tm_dirent_pool; 370 uma_zone_t tm_node_pool; 371 372 /* Read-only status. */ 373 int tm_ronly; 374 }; 375 #define TMPFS_LOCK(tm) mtx_lock(&(tm)->allnode_lock) 376 #define TMPFS_UNLOCK(tm) mtx_unlock(&(tm)->allnode_lock) 377 378 /* --------------------------------------------------------------------- */ 379 380 /* 381 * This structure maps a file identifier to a tmpfs node. Used by the 382 * NFS code. 383 */ 384 struct tmpfs_fid { 385 uint16_t tf_len; 386 uint16_t tf_pad; 387 ino_t tf_id; 388 unsigned long tf_gen; 389 }; 390 391 /* --------------------------------------------------------------------- */ 392 393 #ifdef _KERNEL 394 /* 395 * Prototypes for tmpfs_subr.c. 396 */ 397 398 int tmpfs_alloc_node(struct tmpfs_mount *, enum vtype, 399 uid_t uid, gid_t gid, mode_t mode, struct tmpfs_node *, 400 char *, dev_t, struct tmpfs_node **); 401 void tmpfs_free_node(struct tmpfs_mount *, struct tmpfs_node *); 402 int tmpfs_alloc_dirent(struct tmpfs_mount *, struct tmpfs_node *, 403 const char *, u_int, struct tmpfs_dirent **); 404 void tmpfs_free_dirent(struct tmpfs_mount *, struct tmpfs_dirent *); 405 void tmpfs_dirent_init(struct tmpfs_dirent *, const char *, u_int); 406 void tmpfs_destroy_vobject(struct vnode *vp, vm_object_t obj); 407 int tmpfs_alloc_vp(struct mount *, struct tmpfs_node *, int, 408 struct vnode **); 409 void tmpfs_free_vp(struct vnode *); 410 int tmpfs_alloc_file(struct vnode *, struct vnode **, struct vattr *, 411 struct componentname *, char *); 412 void tmpfs_dir_attach(struct vnode *, struct tmpfs_dirent *); 413 void tmpfs_dir_detach(struct vnode *, struct tmpfs_dirent *); 414 void tmpfs_dir_destroy(struct tmpfs_mount *, struct tmpfs_node *); 415 struct tmpfs_dirent * tmpfs_dir_lookup(struct tmpfs_node *node, 416 struct tmpfs_node *f, 417 struct componentname *cnp); 418 int tmpfs_dir_getdents(struct tmpfs_node *, struct uio *, int, 419 u_long *, int *); 420 int tmpfs_dir_whiteout_add(struct vnode *, struct componentname *); 421 void tmpfs_dir_whiteout_remove(struct vnode *, struct componentname *); 422 int tmpfs_reg_resize(struct vnode *, off_t, boolean_t); 423 int tmpfs_chflags(struct vnode *, u_long, struct ucred *, struct thread *); 424 int tmpfs_chmod(struct vnode *, mode_t, struct ucred *, struct thread *); 425 int tmpfs_chown(struct vnode *, uid_t, gid_t, struct ucred *, 426 struct thread *); 427 int tmpfs_chsize(struct vnode *, u_quad_t, struct ucred *, struct thread *); 428 int tmpfs_chtimes(struct vnode *, struct timespec *, struct timespec *, 429 struct timespec *, int, struct ucred *, struct thread *); 430 void tmpfs_itimes(struct vnode *, const struct timespec *, 431 const struct timespec *); 432 433 void tmpfs_update(struct vnode *); 434 int tmpfs_truncate(struct vnode *, off_t); 435 436 /* --------------------------------------------------------------------- */ 437 438 /* 439 * Convenience macros to simplify some logical expressions. 440 */ 441 #define IMPLIES(a, b) (!(a) || (b)) 442 #define IFF(a, b) (IMPLIES(a, b) && IMPLIES(b, a)) 443 444 /* --------------------------------------------------------------------- */ 445 446 /* 447 * Checks that the directory entry pointed by 'de' matches the name 'name' 448 * with a length of 'len'. 449 */ 450 #define TMPFS_DIRENT_MATCHES(de, name, len) \ 451 (de->td_namelen == len && \ 452 bcmp((de)->ud.td_name, (name), (de)->td_namelen) == 0) 453 454 /* --------------------------------------------------------------------- */ 455 456 /* 457 * Ensures that the node pointed by 'node' is a directory and that its 458 * contents are consistent with respect to directories. 459 */ 460 #define TMPFS_VALIDATE_DIR(node) do { \ 461 MPASS((node)->tn_type == VDIR); \ 462 MPASS((node)->tn_size % sizeof(struct tmpfs_dirent) == 0); \ 463 } while (0) 464 465 /* --------------------------------------------------------------------- */ 466 467 /* 468 * Memory management stuff. 469 */ 470 471 /* 472 * Amount of memory pages to reserve for the system (e.g., to not use by 473 * tmpfs). 474 */ 475 #define TMPFS_PAGES_MINRESERVED (4 * 1024 * 1024 / PAGE_SIZE) 476 477 size_t tmpfs_mem_avail(void); 478 479 size_t tmpfs_pages_used(struct tmpfs_mount *tmp); 480 481 #endif 482 483 /* --------------------------------------------------------------------- */ 484 485 /* 486 * Macros/functions to convert from generic data structures to tmpfs 487 * specific ones. 488 */ 489 490 static inline 491 struct tmpfs_mount * 492 VFS_TO_TMPFS(struct mount *mp) 493 { 494 struct tmpfs_mount *tmp; 495 496 MPASS((mp) != NULL && (mp)->mnt_data != NULL); 497 tmp = (struct tmpfs_mount *)(mp)->mnt_data; 498 return tmp; 499 } 500 501 static inline 502 struct tmpfs_node * 503 VP_TO_TMPFS_NODE(struct vnode *vp) 504 { 505 struct tmpfs_node *node; 506 507 MPASS((vp) != NULL && (vp)->v_data != NULL); 508 node = (struct tmpfs_node *)vp->v_data; 509 return node; 510 } 511 512 static inline 513 struct tmpfs_node * 514 VP_TO_TMPFS_DIR(struct vnode *vp) 515 { 516 struct tmpfs_node *node; 517 518 node = VP_TO_TMPFS_NODE(vp); 519 TMPFS_VALIDATE_DIR(node); 520 return node; 521 } 522 523 #endif /* _FS_TMPFS_TMPFS_H_ */ 524