1 /* $NetBSD: tmpfs.h,v 1.33 2008/04/28 20:24:02 martin Exp $ */ 2 3 /* 4 * Copyright (c) 2005, 2006, 2007 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 33 #ifndef _FS_TMPFS_TMPFS_H_ 34 #define _FS_TMPFS_TMPFS_H_ 35 36 /* --------------------------------------------------------------------- 37 * KERNEL-SPECIFIC DEFINITIONS 38 * --------------------------------------------------------------------- */ 39 #include <sys/dirent.h> 40 #include <sys/mount.h> 41 #include <sys/queue.h> 42 #include <sys/vnode.h> 43 44 #if defined(_KERNEL) 45 #include <fs/tmpfs/tmpfs_pool.h> 46 #endif /* defined(_KERNEL) */ 47 48 /* --------------------------------------------------------------------- */ 49 50 /* 51 * Internal representation of a tmpfs directory entry. 52 */ 53 struct tmpfs_dirent { 54 TAILQ_ENTRY(tmpfs_dirent) td_entries; 55 56 /* Length of the name stored in this directory entry. This avoids 57 * the need to recalculate it every time the name is used. */ 58 uint16_t td_namelen; 59 60 /* The name of the entry, allocated from a string pool. This 61 * string is not required to be zero-terminated; therefore, the 62 * td_namelen field must always be used when accessing its value. */ 63 char * td_name; 64 65 /* Pointer to the node this entry refers to. */ 66 struct tmpfs_node * td_node; 67 }; 68 69 /* A directory in tmpfs holds a sorted list of directory entries, which in 70 * turn point to other files (which can be directories themselves). 71 * 72 * In tmpfs, this list is managed by a tail queue, whose head is defined by 73 * the struct tmpfs_dir type. 74 * 75 * It is imporant to notice that directories do not have entries for . and 76 * .. as other file systems do. These can be generated when requested 77 * based on information available by other means, such as the pointer to 78 * the node itself in the former case or the pointer to the parent directory 79 * in the latter case. This is done to simplify tmpfs's code and, more 80 * importantly, to remove redundancy. */ 81 TAILQ_HEAD(tmpfs_dir, tmpfs_dirent); 82 83 /* Each entry in a directory has a cookie that identifies it. Cookies 84 * supersede offsets within directories because, given how tmpfs stores 85 * directories in memory, there is no such thing as an offset. (Emulating 86 * a real offset could be very difficult.) 87 * 88 * The '.', '..' and the end of directory markers have fixed cookies which 89 * cannot collide with the cookies generated by other entries. The cookies 90 * fot the other entries are generated based on the memory address on which 91 * stores their information is stored. 92 * 93 * Ideally, using the entry's memory pointer as the cookie would be enough 94 * to represent it and it wouldn't cause collisions in any system. 95 * Unfortunately, this results in "offsets" with very large values which 96 * later raise problems in the Linux compatibility layer (and maybe in other 97 * places) as described in PR kern/32034. Hence we need to workaround this 98 * with a rather ugly hack. 99 * 100 * Linux 32-bit binaries, unless built with _FILE_OFFSET_BITS=64, have off_t 101 * set to 'long', which is a 32-bit *signed* long integer. Regardless of 102 * the macro value, GLIBC (2.3 at least) always uses the getdents64 103 * system call (when calling readdir) which internally returns off64_t 104 * offsets. In order to make 32-bit binaries work, *GLIBC* converts the 105 * 64-bit values returned by the kernel to 32-bit ones and aborts with 106 * EOVERFLOW if the conversion results in values that won't fit in 32-bit 107 * integers (which it assumes is because the directory is extremely large). 108 * This wouldn't cause problems if we were dealing with unsigned integers, 109 * but as we have signed integers, this check fails due to sign expansion. 110 * 111 * For example, consider that the kernel returns the 0xc1234567 cookie to 112 * userspace in a off64_t integer. Later on, GLIBC casts this value to 113 * off_t (remember, signed) with code similar to: 114 * system call returns the offset in kernel_value; 115 * off_t casted_value = kernel_value; 116 * if (sizeof(off_t) != sizeof(off64_t) && 117 * kernel_value != casted_value) 118 * error! 119 * In this case, casted_value still has 0xc1234567, but when it is compared 120 * for equality against kernel_value, it is promoted to a 64-bit integer and 121 * becomes 0xffffffffc1234567, which is different than 0x00000000c1234567. 122 * Then, GLIBC assumes this is because the directory is very large. 123 * 124 * Given that all the above happens in user-space, we have no control over 125 * it; therefore we must workaround the issue here. We do this by 126 * truncating the pointer value to a 32-bit integer and hope that there 127 * won't be collisions. In fact, this will not cause any problems in 128 * 32-bit platforms but some might arise in 64-bit machines (I'm not sure 129 * if they can happen at all in practice). 130 * 131 * XXX A nicer solution shall be attempted. */ 132 #if defined(_KERNEL) 133 #define TMPFS_DIRCOOKIE_DOT 0 134 #define TMPFS_DIRCOOKIE_DOTDOT 1 135 #define TMPFS_DIRCOOKIE_EOF 2 136 static __inline 137 off_t 138 tmpfs_dircookie(struct tmpfs_dirent *de) 139 { 140 off_t cookie; 141 142 cookie = ((off_t)(uintptr_t)de >> 1) & 0x7FFFFFFF; 143 KASSERT(cookie != TMPFS_DIRCOOKIE_DOT); 144 KASSERT(cookie != TMPFS_DIRCOOKIE_DOTDOT); 145 KASSERT(cookie != TMPFS_DIRCOOKIE_EOF); 146 147 return cookie; 148 } 149 #endif /* defined(_KERNEL) */ 150 151 /* --------------------------------------------------------------------- */ 152 153 /* 154 * Internal representation of a tmpfs file system node. 155 * 156 * This structure is splitted in two parts: one holds attributes common 157 * to all file types and the other holds data that is only applicable to 158 * a particular type. The code must be careful to only access those 159 * attributes that are actually allowed by the node's type. 160 */ 161 struct tmpfs_node { 162 /* Doubly-linked list entry which links all existing nodes for a 163 * single file system. This is provided to ease the removal of 164 * all nodes during the unmount operation. */ 165 LIST_ENTRY(tmpfs_node) tn_entries; 166 167 /* The node's type. Any of 'VBLK', 'VCHR', 'VDIR', 'VFIFO', 168 * 'VLNK', 'VREG' and 'VSOCK' is allowed. The usage of vnode 169 * types instead of a custom enumeration is to make things simpler 170 * and faster, as we do not need to convert between two types. */ 171 enum vtype tn_type; 172 173 /* Node identifier. */ 174 ino_t tn_id; 175 176 /* Node's internal status. This is used by several file system 177 * operations to do modifications to the node in a delayed 178 * fashion. */ 179 int tn_status; 180 #define TMPFS_NODE_ACCESSED (1 << 1) 181 #define TMPFS_NODE_MODIFIED (1 << 2) 182 #define TMPFS_NODE_CHANGED (1 << 3) 183 184 /* The node size. It does not necessarily match the real amount 185 * of memory consumed by it. */ 186 off_t tn_size; 187 188 /* Generic node attributes. */ 189 uid_t tn_uid; 190 gid_t tn_gid; 191 mode_t tn_mode; 192 int tn_flags; 193 nlink_t tn_links; 194 struct timespec tn_atime; 195 struct timespec tn_mtime; 196 struct timespec tn_ctime; 197 struct timespec tn_birthtime; 198 unsigned long tn_gen; 199 200 /* Head of byte-level lock list (used by tmpfs_advlock). */ 201 struct lockf * tn_lockf; 202 203 /* As there is a single vnode for each active file within the 204 * system, care has to be taken to avoid allocating more than one 205 * vnode per file. In order to do this, a bidirectional association 206 * is kept between vnodes and nodes. 207 * 208 * Whenever a vnode is allocated, its v_data field is updated to 209 * point to the node it references. At the same time, the node's 210 * tn_vnode field is modified to point to the new vnode representing 211 * it. Further attempts to allocate a vnode for this same node will 212 * result in returning a new reference to the value stored in 213 * tn_vnode. 214 * 215 * May be NULL when the node is unused (that is, no vnode has been 216 * allocated for it or it has been reclaimed). */ 217 kmutex_t tn_vlock; 218 struct vnode * tn_vnode; 219 220 union { 221 /* Valid when tn_type == VBLK || tn_type == VCHR. */ 222 struct { 223 dev_t tn_rdev; 224 } tn_dev; 225 226 /* Valid when tn_type == VDIR. */ 227 struct { 228 /* Pointer to the parent directory. The root 229 * directory has a pointer to itself in this field; 230 * this property identifies the root node. */ 231 struct tmpfs_node * tn_parent; 232 233 /* Head of a tail-queue that links the contents of 234 * the directory together. See above for a 235 * description of its contents. */ 236 struct tmpfs_dir tn_dir; 237 238 /* Number and pointer of the first directory entry 239 * returned by the readdir operation if it were 240 * called again to continue reading data from the 241 * same directory as before. This is used to speed 242 * up reads of long directories, assuming that no 243 * more than one read is in progress at a given time. 244 * Otherwise, these values are discarded and a linear 245 * scan is performed from the beginning up to the 246 * point where readdir starts returning values. */ 247 off_t tn_readdir_lastn; 248 struct tmpfs_dirent * tn_readdir_lastp; 249 } tn_dir; 250 251 /* Valid when tn_type == VLNK. */ 252 struct tn_lnk { 253 /* The link's target, allocated from a string pool. */ 254 char * tn_link; 255 } tn_lnk; 256 257 /* Valid when tn_type == VREG. */ 258 struct tn_reg { 259 /* The contents of regular files stored in a tmpfs 260 * file system are represented by a single anonymous 261 * memory object (aobj, for short). The aobj provides 262 * direct access to any position within the file, 263 * because its contents are always mapped in a 264 * contiguous region of virtual memory. It is a task 265 * of the memory management subsystem (see uvm(9)) to 266 * issue the required page ins or page outs whenever 267 * a position within the file is accessed. */ 268 struct uvm_object * tn_aobj; 269 size_t tn_aobj_pages; 270 } tn_reg; 271 } tn_spec; 272 }; 273 274 #if defined(_KERNEL) 275 276 LIST_HEAD(tmpfs_node_list, tmpfs_node); 277 278 /* --------------------------------------------------------------------- */ 279 280 /* 281 * Internal representation of a tmpfs mount point. 282 */ 283 struct tmpfs_mount { 284 /* Maximum number of memory pages available for use by the file 285 * system, set during mount time. This variable must never be 286 * used directly as it may be bigger than the current amount of 287 * free memory; in the extreme case, it will hold the SIZE_MAX 288 * value. Instead, use the TMPFS_PAGES_MAX macro. */ 289 unsigned int tm_pages_max; 290 291 /* Number of pages in use by the file system. Cannot be bigger 292 * than the value returned by TMPFS_PAGES_MAX in any case. */ 293 unsigned int tm_pages_used; 294 295 /* Pointer to the node representing the root directory of this 296 * file system. */ 297 struct tmpfs_node * tm_root; 298 299 /* Maximum number of possible nodes for this file system; set 300 * during mount time. We need a hard limit on the maximum number 301 * of nodes to avoid allocating too much of them; their objects 302 * cannot be released until the file system is unmounted. 303 * Otherwise, we could easily run out of memory by creating lots 304 * of empty files and then simply removing them. */ 305 unsigned int tm_nodes_max; 306 307 /* Number of nodes currently allocated. This number only grows. 308 * When it reaches tm_nodes_max, no more new nodes can be allocated. 309 * Of course, the old, unused ones can be reused. */ 310 unsigned int tm_nodes_cnt; 311 312 /* Node list. */ 313 kmutex_t tm_lock; 314 struct tmpfs_node_list tm_nodes; 315 316 /* Pools used to store file system meta data. These are not shared 317 * across several instances of tmpfs for the reasons described in 318 * tmpfs_pool.c. */ 319 struct tmpfs_pool tm_dirent_pool; 320 struct tmpfs_pool tm_node_pool; 321 struct tmpfs_str_pool tm_str_pool; 322 }; 323 324 /* --------------------------------------------------------------------- */ 325 326 /* 327 * This structure maps a file identifier to a tmpfs node. Used by the 328 * NFS code. 329 */ 330 struct tmpfs_fid { 331 uint16_t tf_len; 332 uint16_t tf_pad; 333 uint32_t tf_gen; 334 ino_t tf_id; 335 }; 336 337 /* --------------------------------------------------------------------- */ 338 339 /* 340 * Prototypes for tmpfs_subr.c. 341 */ 342 343 int tmpfs_alloc_node(struct tmpfs_mount *, enum vtype, 344 uid_t uid, gid_t gid, mode_t mode, struct tmpfs_node *, 345 char *, dev_t, struct tmpfs_node **); 346 void tmpfs_free_node(struct tmpfs_mount *, struct tmpfs_node *); 347 int tmpfs_alloc_dirent(struct tmpfs_mount *, struct tmpfs_node *, 348 const char *, uint16_t, struct tmpfs_dirent **); 349 void tmpfs_free_dirent(struct tmpfs_mount *, struct tmpfs_dirent *, 350 bool); 351 int tmpfs_alloc_vp(struct mount *, struct tmpfs_node *, struct vnode **); 352 void tmpfs_free_vp(struct vnode *); 353 int tmpfs_alloc_file(struct vnode *, struct vnode **, struct vattr *, 354 struct componentname *, char *); 355 void tmpfs_dir_attach(struct vnode *, struct tmpfs_dirent *); 356 void tmpfs_dir_detach(struct vnode *, struct tmpfs_dirent *); 357 struct tmpfs_dirent * tmpfs_dir_lookup(struct tmpfs_node *node, 358 struct componentname *cnp); 359 int tmpfs_dir_getdotdent(struct tmpfs_node *, struct uio *); 360 int tmpfs_dir_getdotdotdent(struct tmpfs_node *, struct uio *); 361 struct tmpfs_dirent * tmpfs_dir_lookupbycookie(struct tmpfs_node *, off_t); 362 int tmpfs_dir_getdents(struct tmpfs_node *, struct uio *, off_t *); 363 int tmpfs_reg_resize(struct vnode *, off_t); 364 size_t tmpfs_mem_info(bool); 365 int tmpfs_chflags(struct vnode *, int, kauth_cred_t, struct lwp *); 366 int tmpfs_chmod(struct vnode *, mode_t, kauth_cred_t, struct lwp *); 367 int tmpfs_chown(struct vnode *, uid_t, gid_t, kauth_cred_t, struct lwp *); 368 int tmpfs_chsize(struct vnode *, u_quad_t, kauth_cred_t, struct lwp *); 369 int tmpfs_chtimes(struct vnode *, struct timespec *, struct timespec *, 370 int, kauth_cred_t, struct lwp *); 371 void tmpfs_itimes(struct vnode *, const struct timespec *, 372 const struct timespec *); 373 374 void tmpfs_update(struct vnode *, const struct timespec *, 375 const struct timespec *, int); 376 int tmpfs_truncate(struct vnode *, off_t); 377 378 /* --------------------------------------------------------------------- */ 379 380 /* 381 * Convenience macros to simplify some logical expressions. 382 */ 383 #define IMPLIES(a, b) (!(a) || (b)) 384 #define IFF(a, b) (IMPLIES(a, b) && IMPLIES(b, a)) 385 386 /* --------------------------------------------------------------------- */ 387 388 /* 389 * Checks that the directory entry pointed by 'de' matches the name 'name' 390 * with a length of 'len'. 391 */ 392 #define TMPFS_DIRENT_MATCHES(de, name, len) \ 393 (de->td_namelen == (uint16_t)len && \ 394 memcmp((de)->td_name, (name), (de)->td_namelen) == 0) 395 396 /* --------------------------------------------------------------------- */ 397 398 /* 399 * Ensures that the node pointed by 'node' is a directory and that its 400 * contents are consistent with respect to directories. 401 */ 402 #define TMPFS_VALIDATE_DIR(node) \ 403 KASSERT((node)->tn_type == VDIR); \ 404 KASSERT((node)->tn_size % sizeof(struct tmpfs_dirent) == 0); \ 405 KASSERT((node)->tn_spec.tn_dir.tn_readdir_lastp == NULL || \ 406 tmpfs_dircookie((node)->tn_spec.tn_dir.tn_readdir_lastp) == \ 407 (node)->tn_spec.tn_dir.tn_readdir_lastn); 408 409 /* --------------------------------------------------------------------- */ 410 411 /* 412 * Memory management stuff. 413 */ 414 415 /* Amount of memory pages to reserve for the system (e.g., to not use by 416 * tmpfs). 417 * XXX: Should this be tunable through sysctl, for instance? */ 418 #define TMPFS_PAGES_RESERVED (4 * 1024 * 1024 / PAGE_SIZE) 419 420 /* Returns the maximum size allowed for a tmpfs file system. This macro 421 * must be used instead of directly retrieving the value from tm_pages_max. 422 * The reason is that the size of a tmpfs file system is dynamic: it lets 423 * the user store files as long as there is enough free memory (including 424 * physical memory and swap space). Therefore, the amount of memory to be 425 * used is either the limit imposed by the user during mount time or the 426 * amount of available memory, whichever is lower. To avoid consuming all 427 * the memory for a given mount point, the system will always reserve a 428 * minimum of TMPFS_PAGES_RESERVED pages, which is also taken into account 429 * by this macro (see above). */ 430 static __inline size_t 431 TMPFS_PAGES_MAX(struct tmpfs_mount *tmp) 432 { 433 size_t freepages; 434 435 freepages = tmpfs_mem_info(false); 436 if (freepages < TMPFS_PAGES_RESERVED) 437 freepages = 0; 438 else 439 freepages -= TMPFS_PAGES_RESERVED; 440 441 return MIN(tmp->tm_pages_max, freepages + tmp->tm_pages_used); 442 } 443 444 /* Returns the available space for the given file system. */ 445 #define TMPFS_PAGES_AVAIL(tmp) \ 446 ((ssize_t)(TMPFS_PAGES_MAX(tmp) - (tmp)->tm_pages_used)) 447 448 /* --------------------------------------------------------------------- */ 449 450 /* 451 * Macros/functions to convert from generic data structures to tmpfs 452 * specific ones. 453 */ 454 455 static __inline 456 struct tmpfs_mount * 457 VFS_TO_TMPFS(struct mount *mp) 458 { 459 struct tmpfs_mount *tmp; 460 461 #ifdef KASSERT 462 KASSERT((mp) != NULL && (mp)->mnt_data != NULL); 463 #endif 464 tmp = (struct tmpfs_mount *)(mp)->mnt_data; 465 return tmp; 466 } 467 468 #endif /* defined(_KERNEL) */ 469 470 static __inline 471 struct tmpfs_node * 472 VP_TO_TMPFS_NODE(struct vnode *vp) 473 { 474 struct tmpfs_node *node; 475 476 #ifdef KASSERT 477 KASSERT((vp) != NULL && (vp)->v_data != NULL); 478 #endif 479 node = (struct tmpfs_node *)vp->v_data; 480 return node; 481 } 482 483 #if defined(_KERNEL) 484 485 static __inline 486 struct tmpfs_node * 487 VP_TO_TMPFS_DIR(struct vnode *vp) 488 { 489 struct tmpfs_node *node; 490 491 node = VP_TO_TMPFS_NODE(vp); 492 #ifdef KASSERT 493 TMPFS_VALIDATE_DIR(node); 494 #endif 495 return node; 496 } 497 498 #endif /* defined(_KERNEL) */ 499 500 /* --------------------------------------------------------------------- 501 * USER AND KERNEL DEFINITIONS 502 * --------------------------------------------------------------------- */ 503 504 /* 505 * This structure is used to communicate mount parameters between userland 506 * and kernel space. 507 */ 508 #define TMPFS_ARGS_VERSION 1 509 struct tmpfs_args { 510 int ta_version; 511 512 /* Size counters. */ 513 ino_t ta_nodes_max; 514 off_t ta_size_max; 515 516 /* Root node attributes. */ 517 uid_t ta_root_uid; 518 gid_t ta_root_gid; 519 mode_t ta_root_mode; 520 }; 521 #endif /* _FS_TMPFS_TMPFS_H_ */ 522