lfs.h revision 1.169
1 /* $NetBSD: lfs.h,v 1.169 2015/08/02 18:10:08 dholland Exp $ */ 2 3 /* from NetBSD: dinode.h,v 1.22 2013/01/22 09:39:18 dholland Exp */ 4 /* from NetBSD: dir.h,v 1.21 2009/07/22 04:49:19 dholland Exp */ 5 6 /*- 7 * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc. 8 * All rights reserved. 9 * 10 * This code is derived from software contributed to The NetBSD Foundation 11 * by Konrad E. Schroder <perseant (at) hhhh.org>. 12 * 13 * Redistribution and use in source and binary forms, with or without 14 * modification, are permitted provided that the following conditions 15 * are met: 16 * 1. Redistributions of source code must retain the above copyright 17 * notice, this list of conditions and the following disclaimer. 18 * 2. Redistributions in binary form must reproduce the above copyright 19 * notice, this list of conditions and the following disclaimer in the 20 * documentation and/or other materials provided with the distribution. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 23 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 24 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 25 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 26 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 27 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 28 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 29 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 30 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 31 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 32 * POSSIBILITY OF SUCH DAMAGE. 33 */ 34 /*- 35 * Copyright (c) 1991, 1993 36 * The Regents of the University of California. All rights reserved. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. Neither the name of the University nor the names of its contributors 47 * may be used to endorse or promote products derived from this software 48 * without specific prior written permission. 49 * 50 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 51 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 52 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 53 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 54 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 55 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 56 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 57 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 58 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 59 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 60 * SUCH DAMAGE. 61 * 62 * @(#)lfs.h 8.9 (Berkeley) 5/8/95 63 */ 64 /* 65 * Copyright (c) 2002 Networks Associates Technology, Inc. 66 * All rights reserved. 67 * 68 * This software was developed for the FreeBSD Project by Marshall 69 * Kirk McKusick and Network Associates Laboratories, the Security 70 * Research Division of Network Associates, Inc. under DARPA/SPAWAR 71 * contract N66001-01-C-8035 ("CBOSS"), as part of the DARPA CHATS 72 * research program 73 * 74 * Copyright (c) 1982, 1989, 1993 75 * The Regents of the University of California. All rights reserved. 76 * (c) UNIX System Laboratories, Inc. 77 * All or some portions of this file are derived from material licensed 78 * to the University of California by American Telephone and Telegraph 79 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 80 * the permission of UNIX System Laboratories, Inc. 81 * 82 * Redistribution and use in source and binary forms, with or without 83 * modification, are permitted provided that the following conditions 84 * are met: 85 * 1. Redistributions of source code must retain the above copyright 86 * notice, this list of conditions and the following disclaimer. 87 * 2. Redistributions in binary form must reproduce the above copyright 88 * notice, this list of conditions and the following disclaimer in the 89 * documentation and/or other materials provided with the distribution. 90 * 3. Neither the name of the University nor the names of its contributors 91 * may be used to endorse or promote products derived from this software 92 * without specific prior written permission. 93 * 94 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 95 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 96 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 97 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 98 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 99 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 100 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 101 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 102 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 103 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 104 * SUCH DAMAGE. 105 * 106 * @(#)dinode.h 8.9 (Berkeley) 3/29/95 107 */ 108 /* 109 * Copyright (c) 1982, 1986, 1989, 1993 110 * The Regents of the University of California. All rights reserved. 111 * (c) UNIX System Laboratories, Inc. 112 * All or some portions of this file are derived from material licensed 113 * to the University of California by American Telephone and Telegraph 114 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 115 * the permission of UNIX System Laboratories, Inc. 116 * 117 * Redistribution and use in source and binary forms, with or without 118 * modification, are permitted provided that the following conditions 119 * are met: 120 * 1. Redistributions of source code must retain the above copyright 121 * notice, this list of conditions and the following disclaimer. 122 * 2. Redistributions in binary form must reproduce the above copyright 123 * notice, this list of conditions and the following disclaimer in the 124 * documentation and/or other materials provided with the distribution. 125 * 3. Neither the name of the University nor the names of its contributors 126 * may be used to endorse or promote products derived from this software 127 * without specific prior written permission. 128 * 129 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 130 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 131 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 132 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 133 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 134 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 135 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 136 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 137 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 138 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 139 * SUCH DAMAGE. 140 * 141 * @(#)dir.h 8.5 (Berkeley) 4/27/95 142 */ 143 144 /* 145 * NOTE: COORDINATE ON-DISK FORMAT CHANGES WITH THE FREEBSD PROJECT. 146 */ 147 148 #ifndef _UFS_LFS_LFS_H_ 149 #define _UFS_LFS_LFS_H_ 150 151 #include <sys/rwlock.h> 152 #include <sys/mutex.h> 153 #include <sys/queue.h> 154 #include <sys/condvar.h> 155 #include <sys/mount.h> 156 #include <sys/pool.h> 157 158 /* 159 * Compile-time options for LFS. 160 */ 161 #define LFS_IFIND_RETRIES 16 162 #define LFS_LOGLENGTH 1024 /* size of debugging log */ 163 #define LFS_MAX_ACTIVE 10 /* Dirty segments before ckp forced */ 164 165 /* 166 * Fixed filesystem layout parameters 167 */ 168 #define LFS_LABELPAD 8192 /* LFS label size */ 169 #define LFS_SBPAD 8192 /* LFS superblock size */ 170 171 #define LFS_UNUSED_INUM 0 /* 0: out of band inode number */ 172 #define LFS_IFILE_INUM 1 /* 1: IFILE inode number */ 173 /* 2: Root inode number */ 174 #define LFS_LOSTFOUNDINO 3 /* 3: lost+found inode number */ 175 #define LFS_FIRST_INUM 4 /* 4: first free inode number */ 176 177 /* 178 * The root inode is the root of the file system. Inode 0 can't be used for 179 * normal purposes and historically bad blocks were linked to inode 1, thus 180 * the root inode is 2. (Inode 1 is no longer used for this purpose, however 181 * numerous dump tapes make this assumption, so we are stuck with it). 182 */ 183 #define ULFS_ROOTINO ((ino_t)2) 184 185 /* 186 * The Whiteout inode# is a dummy non-zero inode number which will 187 * never be allocated to a real file. It is used as a place holder 188 * in the directory entry which has been tagged as a LFS_DT_WHT entry. 189 * See the comments about ULFS_ROOTINO above. 190 */ 191 #define ULFS_WINO ((ino_t)1) 192 193 194 #define LFS_V1_SUMMARY_SIZE 512 /* V1 fixed summary size */ 195 #define LFS_DFL_SUMMARY_SIZE 512 /* Default summary size */ 196 197 #define LFS_MAX_DADDR 0x7fffffff /* Highest addressable fsb */ 198 199 #define LFS_MAXNAMLEN 255 /* maximum name length in a dir */ 200 201 #define ULFS_NXADDR 2 202 #define ULFS_NDADDR 12 /* Direct addresses in inode. */ 203 #define ULFS_NIADDR 3 /* Indirect addresses in inode. */ 204 205 /* 206 * Adjustable filesystem parameters 207 */ 208 #ifndef LFS_ATIME_IFILE 209 # define LFS_ATIME_IFILE 0 /* Store atime info in ifile (optional in LFSv1) */ 210 #endif 211 #define LFS_MARKV_MAXBLKCNT 65536 /* Max block count for lfs_markv() */ 212 213 /* 214 * Directories 215 */ 216 217 /* 218 * A directory consists of some number of blocks of LFS_DIRBLKSIZ 219 * bytes, where LFS_DIRBLKSIZ is chosen such that it can be transferred 220 * to disk in a single atomic operation (e.g. 512 bytes on most machines). 221 * 222 * Each LFS_DIRBLKSIZ byte block contains some number of directory entry 223 * structures, which are of variable length. Each directory entry has 224 * a struct lfs_direct at the front of it, containing its inode number, 225 * the length of the entry, and the length of the name contained in 226 * the entry. These are followed by the name padded to a 4 byte boundary. 227 * All names are guaranteed null terminated. 228 * The maximum length of a name in a directory is LFS_MAXNAMLEN. 229 * 230 * The macro DIRSIZ(fmt, dp) gives the amount of space required to represent 231 * a directory entry. Free space in a directory is represented by 232 * entries which have dp->d_reclen > DIRSIZ(fmt, dp). All LFS_DIRBLKSIZ bytes 233 * in a directory block are claimed by the directory entries. This 234 * usually results in the last entry in a directory having a large 235 * dp->d_reclen. When entries are deleted from a directory, the 236 * space is returned to the previous entry in the same directory 237 * block by increasing its dp->d_reclen. If the first entry of 238 * a directory block is free, then its dp->d_ino is set to 0. 239 * Entries other than the first in a directory do not normally have 240 * dp->d_ino set to 0. 241 */ 242 243 /* 244 * Directory block size. 245 */ 246 #undef LFS_DIRBLKSIZ 247 #define LFS_DIRBLKSIZ DEV_BSIZE 248 249 /* 250 * Convert between stat structure types and directory types. 251 */ 252 #define LFS_IFTODT(mode) (((mode) & 0170000) >> 12) 253 #define LFS_DTTOIF(dirtype) ((dirtype) << 12) 254 255 /* 256 * The LFS_DIRSIZ macro gives the minimum record length which will hold 257 * the directory entry. This requires the amount of space in struct lfs_direct 258 * without the d_name field, plus enough space for the name with a terminating 259 * null byte (dp->d_namlen+1), rounded up to a 4 byte boundary. 260 */ 261 #define LFS_DIRECTSIZ(namlen) \ 262 ((sizeof(struct lfs_direct) - (LFS_MAXNAMLEN+1)) + (((namlen)+1 + 3) &~ 3)) 263 264 #if (BYTE_ORDER == LITTLE_ENDIAN) 265 #define LFS_DIRSIZ(oldfmt, dp, needswap) \ 266 (((oldfmt) && !(needswap)) ? \ 267 LFS_DIRECTSIZ((dp)->d_type) : LFS_DIRECTSIZ((dp)->d_namlen)) 268 #else 269 #define LFS_DIRSIZ(oldfmt, dp, needswap) \ 270 (((oldfmt) && (needswap)) ? \ 271 LFS_DIRECTSIZ((dp)->d_type) : LFS_DIRECTSIZ((dp)->d_namlen)) 272 #endif 273 274 /* Constants for the first argument of LFS_DIRSIZ */ 275 #define LFS_OLDDIRFMT 1 276 #define LFS_NEWDIRFMT 0 277 278 /* 279 * Theoretically, directories can be more than 2Gb in length; however, in 280 * practice this seems unlikely. So, we define the type doff_t as a 32-bit 281 * quantity to keep down the cost of doing lookup on a 32-bit machine. 282 */ 283 #define doff_t int32_t 284 #define lfs_doff_t int32_t 285 #define LFS_MAXDIRSIZE (0x7fffffff) 286 287 /* 288 * File types for d_type 289 */ 290 #define LFS_DT_UNKNOWN 0 291 #define LFS_DT_FIFO 1 292 #define LFS_DT_CHR 2 293 #define LFS_DT_DIR 4 294 #define LFS_DT_BLK 6 295 #define LFS_DT_REG 8 296 #define LFS_DT_LNK 10 297 #define LFS_DT_SOCK 12 298 #define LFS_DT_WHT 14 299 300 /* 301 * (See notes above) 302 */ 303 #define d_ino d_fileno 304 struct lfs_direct { 305 u_int32_t d_fileno; /* inode number of entry */ 306 u_int16_t d_reclen; /* length of this record */ 307 u_int8_t d_type; /* file type, see below */ 308 u_int8_t d_namlen; /* length of string in d_name */ 309 char d_name[LFS_MAXNAMLEN + 1];/* name with length <= LFS_MAXNAMLEN */ 310 }; 311 312 /* 313 * Template for manipulating directories. Should use struct lfs_direct's, 314 * but the name field is LFS_MAXNAMLEN - 1, and this just won't do. 315 */ 316 struct lfs_dirtemplate { 317 u_int32_t dot_ino; 318 int16_t dot_reclen; 319 u_int8_t dot_type; 320 u_int8_t dot_namlen; 321 char dot_name[4]; /* must be multiple of 4 */ 322 u_int32_t dotdot_ino; 323 int16_t dotdot_reclen; 324 u_int8_t dotdot_type; 325 u_int8_t dotdot_namlen; 326 char dotdot_name[4]; /* ditto */ 327 }; 328 329 /* 330 * This is the old format of directories, sans type element. 331 */ 332 struct lfs_odirtemplate { 333 u_int32_t dot_ino; 334 int16_t dot_reclen; 335 u_int16_t dot_namlen; 336 char dot_name[4]; /* must be multiple of 4 */ 337 u_int32_t dotdot_ino; 338 int16_t dotdot_reclen; 339 u_int16_t dotdot_namlen; 340 char dotdot_name[4]; /* ditto */ 341 }; 342 343 /* 344 * Inodes 345 */ 346 347 /* 348 * A dinode contains all the meta-data associated with a LFS file. 349 * This structure defines the on-disk format of a dinode. Since 350 * this structure describes an on-disk structure, all its fields 351 * are defined by types with precise widths. 352 */ 353 354 struct ulfs1_dinode { 355 u_int16_t di_mode; /* 0: IFMT, permissions; see below. */ 356 int16_t di_nlink; /* 2: File link count. */ 357 u_int32_t di_inumber; /* 4: Inode number. */ 358 u_int64_t di_size; /* 8: File byte count. */ 359 int32_t di_atime; /* 16: Last access time. */ 360 int32_t di_atimensec; /* 20: Last access time. */ 361 int32_t di_mtime; /* 24: Last modified time. */ 362 int32_t di_mtimensec; /* 28: Last modified time. */ 363 int32_t di_ctime; /* 32: Last inode change time. */ 364 int32_t di_ctimensec; /* 36: Last inode change time. */ 365 int32_t di_db[ULFS_NDADDR]; /* 40: Direct disk blocks. */ 366 int32_t di_ib[ULFS_NIADDR]; /* 88: Indirect disk blocks. */ 367 u_int32_t di_flags; /* 100: Status flags (chflags). */ 368 u_int32_t di_blocks; /* 104: Blocks actually held. */ 369 int32_t di_gen; /* 108: Generation number. */ 370 u_int32_t di_uid; /* 112: File owner. */ 371 u_int32_t di_gid; /* 116: File group. */ 372 u_int64_t di_modrev; /* 120: i_modrev for NFSv4 */ 373 }; 374 375 struct ulfs2_dinode { 376 u_int16_t di_mode; /* 0: IFMT, permissions; see below. */ 377 int16_t di_nlink; /* 2: File link count. */ 378 u_int32_t di_uid; /* 4: File owner. */ 379 u_int32_t di_gid; /* 8: File group. */ 380 u_int32_t di_blksize; /* 12: Inode blocksize. */ 381 u_int64_t di_size; /* 16: File byte count. */ 382 u_int64_t di_blocks; /* 24: Bytes actually held. */ 383 int64_t di_atime; /* 32: Last access time. */ 384 int64_t di_mtime; /* 40: Last modified time. */ 385 int64_t di_ctime; /* 48: Last inode change time. */ 386 int64_t di_birthtime; /* 56: Inode creation time. */ 387 int32_t di_mtimensec; /* 64: Last modified time. */ 388 int32_t di_atimensec; /* 68: Last access time. */ 389 int32_t di_ctimensec; /* 72: Last inode change time. */ 390 int32_t di_birthnsec; /* 76: Inode creation time. */ 391 int32_t di_gen; /* 80: Generation number. */ 392 u_int32_t di_kernflags; /* 84: Kernel flags. */ 393 u_int32_t di_flags; /* 88: Status flags (chflags). */ 394 int32_t di_extsize; /* 92: External attributes block. */ 395 int64_t di_extb[ULFS_NXADDR];/* 96: External attributes block. */ 396 int64_t di_db[ULFS_NDADDR]; /* 112: Direct disk blocks. */ 397 int64_t di_ib[ULFS_NIADDR]; /* 208: Indirect disk blocks. */ 398 u_int64_t di_modrev; /* 232: i_modrev for NFSv4 */ 399 u_int32_t di_inumber; /* 240: Inode number */ 400 u_int32_t di_spare[3]; /* 244: Reserved; currently unused */ 401 }; 402 403 /* 404 * The di_db fields may be overlaid with other information for 405 * file types that do not have associated disk storage. Block 406 * and character devices overlay the first data block with their 407 * dev_t value. Short symbolic links place their path in the 408 * di_db area. 409 */ 410 #define di_rdev di_db[0] 411 412 /* Size of the on-disk inode. */ 413 #define LFS_DINODE1_SIZE (sizeof(struct ulfs1_dinode)) /* 128 */ 414 #define LFS_DINODE2_SIZE (sizeof(struct ulfs2_dinode)) 415 416 /* File types, found in the upper bits of di_mode. */ 417 #define LFS_IFMT 0170000 /* Mask of file type. */ 418 #define LFS_IFIFO 0010000 /* Named pipe (fifo). */ 419 #define LFS_IFCHR 0020000 /* Character device. */ 420 #define LFS_IFDIR 0040000 /* Directory file. */ 421 #define LFS_IFBLK 0060000 /* Block device. */ 422 #define LFS_IFREG 0100000 /* Regular file. */ 423 #define LFS_IFLNK 0120000 /* Symbolic link. */ 424 #define LFS_IFSOCK 0140000 /* UNIX domain socket. */ 425 #define LFS_IFWHT 0160000 /* Whiteout. */ 426 427 /* 428 * "struct buf" associated definitions 429 */ 430 431 /* Unassigned disk addresses. */ 432 #define UNASSIGNED -1 433 #define UNWRITTEN -2 434 435 /* Unused logical block number */ 436 #define LFS_UNUSED_LBN -1 437 438 /* 439 * "struct inode" associated definitions 440 */ 441 442 /* For convenience */ 443 #define IN_ALLMOD (IN_MODIFIED|IN_ACCESS|IN_CHANGE|IN_UPDATE|IN_MODIFY|IN_ACCESSED|IN_CLEANING) 444 445 /* 446 * On-disk and in-memory checkpoint segment usage structure. 447 */ 448 typedef struct segusage SEGUSE; 449 struct segusage { 450 u_int32_t su_nbytes; /* 0: number of live bytes */ 451 u_int32_t su_olastmod; /* 4: SEGUSE last modified timestamp */ 452 u_int16_t su_nsums; /* 8: number of summaries in segment */ 453 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 454 455 #define SEGUSE_ACTIVE 0x01 /* segment currently being written */ 456 #define SEGUSE_DIRTY 0x02 /* segment has data in it */ 457 #define SEGUSE_SUPERBLOCK 0x04 /* segment contains a superblock */ 458 #define SEGUSE_ERROR 0x08 /* cleaner: do not clean segment */ 459 #define SEGUSE_EMPTY 0x10 /* segment is empty */ 460 #define SEGUSE_INVAL 0x20 /* segment is invalid */ 461 u_int32_t su_flags; /* 12: segment flags */ 462 u_int64_t su_lastmod; /* 16: last modified timestamp */ 463 }; 464 465 typedef struct segusage_v1 SEGUSE_V1; 466 struct segusage_v1 { 467 u_int32_t su_nbytes; /* 0: number of live bytes */ 468 u_int32_t su_lastmod; /* 4: SEGUSE last modified timestamp */ 469 u_int16_t su_nsums; /* 8: number of summaries in segment */ 470 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 471 u_int32_t su_flags; /* 12: segment flags */ 472 }; 473 474 /* 475 * On-disk file information. One per file with data blocks in the segment. 476 */ 477 typedef struct finfo FINFO; 478 struct finfo { 479 u_int32_t fi_nblocks; /* number of blocks */ 480 u_int32_t fi_version; /* version number */ 481 u_int32_t fi_ino; /* inode number */ 482 u_int32_t fi_lastlength; /* length of last block in array */ 483 int32_t fi_blocks[1]; /* array of logical block numbers */ 484 }; 485 /* sizeof FINFO except fi_blocks */ 486 #define FINFOSIZE (sizeof(FINFO) - sizeof(int32_t)) 487 488 /* 489 * Index file inode entries. 490 */ 491 typedef struct ifile IFILE; 492 struct ifile { 493 u_int32_t if_version; /* inode version number */ 494 #define LFS_UNUSED_DADDR 0 /* out-of-band daddr */ 495 int32_t if_daddr; /* inode disk address */ 496 #define LFS_ORPHAN_NEXTFREE (~(u_int32_t)0) /* indicate orphaned file */ 497 u_int32_t if_nextfree; /* next-unallocated inode */ 498 u_int32_t if_atime_sec; /* Last access time, seconds */ 499 u_int32_t if_atime_nsec; /* and nanoseconds */ 500 }; 501 502 typedef struct ifile_v1 IFILE_V1; 503 struct ifile_v1 { 504 u_int32_t if_version; /* inode version number */ 505 int32_t if_daddr; /* inode disk address */ 506 u_int32_t if_nextfree; /* next-unallocated inode */ 507 #if LFS_ATIME_IFILE 508 struct timespec if_atime; /* Last access time */ 509 #endif 510 }; 511 512 /* 513 * Cleaner information structure. This resides in the ifile and is used 514 * to pass information from the kernel to the cleaner. 515 */ 516 typedef struct _cleanerinfo { 517 u_int32_t clean; /* number of clean segments */ 518 u_int32_t dirty; /* number of dirty segments */ 519 /* XXX64 bfree and avail must -> 64 */ 520 int32_t bfree; /* disk blocks free */ 521 int32_t avail; /* disk blocks available */ 522 u_int32_t free_head; /* head of the inode free list */ 523 u_int32_t free_tail; /* tail of the inode free list */ 524 #define LFS_CLEANER_MUST_CLEAN 0x01 525 u_int32_t flags; /* status word from the kernel */ 526 } CLEANERINFO; 527 528 /* 529 * On-disk segment summary information 530 */ 531 typedef struct segsum_v1 SEGSUM_V1; 532 struct segsum_v1 { 533 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 534 u_int32_t ss_datasum; /* 4: check sum of data */ 535 u_int32_t ss_magic; /* 8: segment summary magic number */ 536 #define SS_MAGIC 0x061561 537 int32_t ss_next; /* 12: next segment */ 538 u_int32_t ss_create; /* 16: creation time stamp */ 539 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 540 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 541 542 #define SS_DIROP 0x01 /* segment begins a dirop */ 543 #define SS_CONT 0x02 /* more partials to finish this write*/ 544 #define SS_CLEAN 0x04 /* written by the cleaner */ 545 #define SS_RFW 0x08 /* written by the roll-forward agent */ 546 #define SS_RECLAIM 0x10 /* written by the roll-forward agent */ 547 u_int16_t ss_flags; /* 24: used for directory operations */ 548 u_int16_t ss_pad; /* 26: extra space */ 549 /* FINFO's and inode daddr's... */ 550 }; 551 552 typedef struct segsum SEGSUM; 553 struct segsum { 554 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 555 u_int32_t ss_datasum; /* 4: check sum of data */ 556 u_int32_t ss_magic; /* 8: segment summary magic number */ 557 int32_t ss_next; /* 12: next segment (disk address) */ 558 u_int32_t ss_ident; /* 16: roll-forward fsid */ 559 #define ss_ocreate ss_ident /* ident is where create was in v1 */ 560 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 561 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 562 u_int16_t ss_flags; /* 24: used for directory operations */ 563 u_int8_t ss_pad[2]; /* 26: extra space */ 564 u_int32_t ss_reclino; /* 28: inode being reclaimed */ 565 u_int64_t ss_serial; /* 32: serial number */ 566 u_int64_t ss_create; /* 40: time stamp */ 567 /* FINFO's and inode daddr's... */ 568 }; 569 570 571 /* 572 * On-disk super block. 573 */ 574 struct dlfs { 575 #define LFS_MAGIC 0x070162 576 u_int32_t dlfs_magic; /* 0: magic number */ 577 #define LFS_VERSION 2 578 u_int32_t dlfs_version; /* 4: version number */ 579 580 u_int32_t dlfs_size; /* 8: number of blocks in fs (v1) */ 581 /* number of frags in fs (v2) */ 582 u_int32_t dlfs_ssize; /* 12: number of blocks per segment (v1) */ 583 /* number of bytes per segment (v2) */ 584 u_int32_t dlfs_dsize; /* 16: number of disk blocks in fs */ 585 u_int32_t dlfs_bsize; /* 20: file system block size */ 586 u_int32_t dlfs_fsize; /* 24: size of frag blocks in fs */ 587 u_int32_t dlfs_frag; /* 28: number of frags in a block in fs */ 588 589 /* Checkpoint region. */ 590 u_int32_t dlfs_freehd; /* 32: start of the free inode list */ 591 int32_t dlfs_bfree; /* 36: number of free frags */ 592 u_int32_t dlfs_nfiles; /* 40: number of allocated inodes */ 593 int32_t dlfs_avail; /* 44: blocks available for writing */ 594 int32_t dlfs_uinodes; /* 48: inodes in cache not yet on disk */ 595 int32_t dlfs_idaddr; /* 52: inode file disk address */ 596 u_int32_t dlfs_ifile; /* 56: inode file inode number */ 597 int32_t dlfs_lastseg; /* 60: address of last segment written */ 598 int32_t dlfs_nextseg; /* 64: address of next segment to write */ 599 int32_t dlfs_curseg; /* 68: current segment being written */ 600 int32_t dlfs_offset; /* 72: offset in curseg for next partial */ 601 int32_t dlfs_lastpseg; /* 76: address of last partial written */ 602 u_int32_t dlfs_inopf; /* 80: v1: time stamp; v2: inodes per frag */ 603 604 /* These are configuration parameters. */ 605 u_int32_t dlfs_minfree; /* 84: minimum percentage of free blocks */ 606 607 /* These fields can be computed from the others. */ 608 u_int64_t dlfs_maxfilesize; /* 88: maximum representable file size */ 609 u_int32_t dlfs_fsbpseg; /* 96: frags (fsb) per segment */ 610 u_int32_t dlfs_inopb; /* 100: inodes per block */ 611 u_int32_t dlfs_ifpb; /* 104: IFILE entries per block */ 612 u_int32_t dlfs_sepb; /* 108: SEGUSE entries per block */ 613 u_int32_t dlfs_nindir; /* 112: indirect pointers per block */ 614 u_int32_t dlfs_nseg; /* 116: number of segments */ 615 u_int32_t dlfs_nspf; /* 120: number of sectors per fragment */ 616 u_int32_t dlfs_cleansz; /* 124: cleaner info size in blocks */ 617 u_int32_t dlfs_segtabsz; /* 128: segment table size in blocks */ 618 u_int32_t dlfs_segmask; /* 132: calculate offset within a segment */ 619 u_int32_t dlfs_segshift; /* 136: fast mult/div for segments */ 620 u_int32_t dlfs_bshift; /* 140: calc block number from file offset */ 621 u_int32_t dlfs_ffshift; /* 144: fast mult/div for frag from file */ 622 u_int32_t dlfs_fbshift; /* 148: fast mult/div for frag from block */ 623 u_int64_t dlfs_bmask; /* 152: calc block offset from file offset */ 624 u_int64_t dlfs_ffmask; /* 160: calc frag offset from file offset */ 625 u_int64_t dlfs_fbmask; /* 168: calc frag offset from block offset */ 626 u_int32_t dlfs_blktodb; /* 176: blktodb and dbtoblk shift constant */ 627 u_int32_t dlfs_sushift; /* 180: fast mult/div for segusage table */ 628 629 int32_t dlfs_maxsymlinklen; /* 184: max length of an internal symlink */ 630 #define LFS_MIN_SBINTERVAL 5 /* minimum superblock segment spacing */ 631 #define LFS_MAXNUMSB 10 /* 188: superblock disk offsets */ 632 int32_t dlfs_sboffs[LFS_MAXNUMSB]; 633 634 u_int32_t dlfs_nclean; /* 228: Number of clean segments */ 635 u_char dlfs_fsmnt[MNAMELEN]; /* 232: name mounted on */ 636 #define LFS_PF_CLEAN 0x1 637 u_int16_t dlfs_pflags; /* 322: file system persistent flags */ 638 int32_t dlfs_dmeta; /* 324: total number of dirty summaries */ 639 u_int32_t dlfs_minfreeseg; /* 328: segments not counted in bfree */ 640 u_int32_t dlfs_sumsize; /* 332: size of summary blocks */ 641 u_int64_t dlfs_serial; /* 336: serial number */ 642 u_int32_t dlfs_ibsize; /* 344: size of inode blocks */ 643 int32_t dlfs_s0addr; /* 348: start of segment 0 */ 644 u_int64_t dlfs_tstamp; /* 352: time stamp */ 645 #define LFS_44INODEFMT 0 646 #define LFS_MAXINODEFMT 0 647 u_int32_t dlfs_inodefmt; /* 360: inode format version */ 648 u_int32_t dlfs_interleave; /* 364: segment interleave */ 649 u_int32_t dlfs_ident; /* 368: per-fs identifier */ 650 u_int32_t dlfs_fsbtodb; /* 372: fsbtodb and dbtodsb shift constant */ 651 u_int32_t dlfs_resvseg; /* 376: segments reserved for the cleaner */ 652 int8_t dlfs_pad[128]; /* 380: round to 512 bytes */ 653 /* Checksum -- last valid disk field. */ 654 u_int32_t dlfs_cksum; /* 508: checksum for superblock checking */ 655 }; 656 657 /* Type used for the inode bitmap */ 658 typedef u_int32_t lfs_bm_t; 659 660 /* 661 * Linked list of segments whose byte count needs updating following a 662 * file truncation. 663 */ 664 struct segdelta { 665 long segnum; 666 size_t num; 667 LIST_ENTRY(segdelta) list; 668 }; 669 670 /* 671 * In-memory super block. 672 */ 673 struct lfs { 674 struct dlfs lfs_dlfs; /* on-disk parameters */ 675 676 /* These fields are set at mount time and are meaningless on disk. */ 677 struct segment *lfs_sp; /* current segment being written */ 678 struct vnode *lfs_ivnode; /* vnode for the ifile */ 679 u_int32_t lfs_seglock; /* single-thread the segment writer */ 680 pid_t lfs_lockpid; /* pid of lock holder */ 681 lwpid_t lfs_locklwp; /* lwp of lock holder */ 682 u_int32_t lfs_iocount; /* number of ios pending */ 683 u_int32_t lfs_writer; /* don't allow any dirops to start */ 684 u_int32_t lfs_dirops; /* count of active directory ops */ 685 u_int32_t lfs_dirvcount; /* count of VDIROP nodes in this fs */ 686 u_int32_t lfs_doifile; /* Write ifile blocks on next write */ 687 u_int32_t lfs_nactive; /* Number of segments since last ckp */ 688 int8_t lfs_fmod; /* super block modified flag */ 689 int8_t lfs_ronly; /* mounted read-only flag */ 690 #define LFS_NOTYET 0x01 691 #define LFS_IFDIRTY 0x02 692 #define LFS_WARNED 0x04 693 #define LFS_UNDIROP 0x08 694 int8_t lfs_flags; /* currently unused flag */ 695 u_int16_t lfs_activesb; /* toggle between superblocks */ 696 daddr_t lfs_sbactive; /* disk address of current sb write */ 697 struct vnode *lfs_flushvp; /* vnode being flushed */ 698 int lfs_flushvp_fakevref; /* fake vref count for flushvp */ 699 struct vnode *lfs_unlockvp; /* being inactivated in lfs_segunlock */ 700 u_int32_t lfs_diropwait; /* # procs waiting on dirop flush */ 701 size_t lfs_devbsize; /* Device block size */ 702 size_t lfs_devbshift; /* Device block shift */ 703 krwlock_t lfs_fraglock; 704 krwlock_t lfs_iflock; /* Ifile lock */ 705 kcondvar_t lfs_stopcv; /* Wrap lock */ 706 struct lwp *lfs_stoplwp; 707 pid_t lfs_rfpid; /* Process ID of roll-forward agent */ 708 int lfs_nadirop; /* number of active dirop nodes */ 709 long lfs_ravail; /* blocks pre-reserved for writing */ 710 long lfs_favail; /* blocks pre-reserved for writing */ 711 struct lfs_res_blk *lfs_resblk; /* Reserved memory for pageout */ 712 TAILQ_HEAD(, inode) lfs_dchainhd; /* dirop vnodes */ 713 TAILQ_HEAD(, inode) lfs_pchainhd; /* paging vnodes */ 714 #define LFS_RESHASH_WIDTH 17 715 LIST_HEAD(, lfs_res_blk) lfs_reshash[LFS_RESHASH_WIDTH]; 716 int lfs_pdflush; /* pagedaemon wants us to flush */ 717 u_int32_t **lfs_suflags; /* Segment use flags */ 718 #ifdef _KERNEL 719 struct pool lfs_clpool; /* Pool for struct lfs_cluster */ 720 struct pool lfs_bpppool; /* Pool for bpp */ 721 struct pool lfs_segpool; /* Pool for struct segment */ 722 #endif /* _KERNEL */ 723 #define LFS_MAX_CLEANIND 64 724 int32_t lfs_cleanint[LFS_MAX_CLEANIND]; /* Active cleaning intervals */ 725 int lfs_cleanind; /* Index into intervals */ 726 int lfs_sleepers; /* # procs sleeping this fs */ 727 int lfs_pages; /* dirty pages blaming this fs */ 728 lfs_bm_t *lfs_ino_bitmap; /* Inuse inodes bitmap */ 729 int lfs_nowrap; /* Suspend log wrap */ 730 int lfs_wrappass; /* Allow first log wrap requester to pass */ 731 int lfs_wrapstatus; /* Wrap status */ 732 int lfs_reclino; /* Inode being reclaimed */ 733 daddr_t lfs_startseg; /* Segment we started writing at */ 734 LIST_HEAD(, segdelta) lfs_segdhd; /* List of pending trunc accounting events */ 735 736 #ifdef _KERNEL 737 /* ULFS-level information */ 738 u_int32_t um_flags; /* ULFS flags (below) */ 739 u_long um_nindir; /* indirect ptrs per block */ 740 u_long um_lognindir; /* log2 of um_nindir */ 741 u_long um_bptrtodb; /* indir ptr to disk block */ 742 u_long um_seqinc; /* inc between seq blocks */ 743 int um_maxsymlinklen; 744 int um_dirblksiz; 745 u_int64_t um_maxfilesize; 746 747 /* Stuff used by quota2 code, not currently operable */ 748 unsigned lfs_use_quota2 : 1; 749 uint32_t lfs_quota_magic; 750 uint8_t lfs_quota_flags; 751 uint64_t lfs_quotaino[2]; 752 753 /* Sleep address replacing &lfs_avail inside the on-disk superblock */ 754 /* XXX: should be replaced with a condvar */ 755 int lfs_availsleep; 756 /* This one replaces &lfs_nextseg... all ditto */ 757 int lfs_nextsegsleep; 758 #endif 759 }; 760 761 /* 762 * Structures used by lfs_bmapv and lfs_markv to communicate information 763 * about inodes and data blocks. 764 */ 765 typedef struct block_info { 766 u_int32_t bi_inode; /* inode # */ 767 int32_t bi_lbn; /* logical block w/in file */ 768 int32_t bi_daddr; /* disk address of block */ 769 u_int64_t bi_segcreate; /* origin segment create time */ 770 int bi_version; /* file version number */ 771 void *bi_bp; /* data buffer */ 772 int bi_size; /* size of the block (if fragment) */ 773 } BLOCK_INFO; 774 775 /* Compatibility for 1.5 binaries */ 776 typedef struct block_info_15 { 777 u_int32_t bi_inode; /* inode # */ 778 int32_t bi_lbn; /* logical block w/in file */ 779 int32_t bi_daddr; /* disk address of block */ 780 u_int32_t bi_segcreate; /* origin segment create time */ 781 int bi_version; /* file version number */ 782 void *bi_bp; /* data buffer */ 783 int bi_size; /* size of the block (if fragment) */ 784 } BLOCK_INFO_15; 785 786 /* In-memory description of a segment about to be written. */ 787 struct segment { 788 struct lfs *fs; /* file system pointer */ 789 struct buf **bpp; /* pointer to buffer array */ 790 struct buf **cbpp; /* pointer to next available bp */ 791 struct buf **start_bpp; /* pointer to first bp in this set */ 792 struct buf *ibp; /* buffer pointer to inode page */ 793 struct ulfs1_dinode *idp; /* pointer to ifile dinode */ 794 struct finfo *fip; /* current fileinfo pointer */ 795 struct vnode *vp; /* vnode being gathered */ 796 void *segsum; /* segment summary info */ 797 u_int32_t ninodes; /* number of inodes in this segment */ 798 int32_t seg_bytes_left; /* bytes left in segment */ 799 int32_t sum_bytes_left; /* bytes left in summary block */ 800 u_int32_t seg_number; /* number of this segment */ 801 int32_t *start_lbp; /* beginning lbn for this set */ 802 803 #define SEGM_CKP 0x0001 /* doing a checkpoint */ 804 #define SEGM_CLEAN 0x0002 /* cleaner call; don't sort */ 805 #define SEGM_SYNC 0x0004 /* wait for segment */ 806 #define SEGM_PROT 0x0008 /* don't inactivate at segunlock */ 807 #define SEGM_PAGEDAEMON 0x0010 /* pagedaemon called us */ 808 #define SEGM_WRITERD 0x0020 /* LFS writed called us */ 809 #define SEGM_FORCE_CKP 0x0040 /* Force checkpoint right away */ 810 #define SEGM_RECLAIM 0x0080 /* Writing to reclaim vnode */ 811 #define SEGM_SINGLE 0x0100 /* Opportunistic writevnodes */ 812 u_int16_t seg_flags; /* run-time flags for this segment */ 813 u_int32_t seg_iocount; /* number of ios pending */ 814 int ndupino; /* number of duplicate inodes */ 815 }; 816 817 /* Statistics Counters */ 818 struct lfs_stats { /* Must match sysctl list in lfs_vfsops.h ! */ 819 u_int segsused; 820 u_int psegwrites; 821 u_int psyncwrites; 822 u_int pcleanwrites; 823 u_int blocktot; 824 u_int cleanblocks; 825 u_int ncheckpoints; 826 u_int nwrites; 827 u_int nsync_writes; 828 u_int wait_exceeded; 829 u_int write_exceeded; 830 u_int flush_invoked; 831 u_int vflush_invoked; 832 u_int clean_inlocked; 833 u_int clean_vnlocked; 834 u_int segs_reclaimed; 835 }; 836 837 /* Fcntls to take the place of the lfs syscalls */ 838 struct lfs_fcntl_markv { 839 BLOCK_INFO *blkiov; /* blocks to relocate */ 840 int blkcnt; /* number of blocks */ 841 }; 842 843 #define LFCNSEGWAITALL _FCNR_FSPRIV('L', 14, struct timeval) 844 #define LFCNSEGWAIT _FCNR_FSPRIV('L', 15, struct timeval) 845 #define LFCNBMAPV _FCNRW_FSPRIV('L', 2, struct lfs_fcntl_markv) 846 #define LFCNMARKV _FCNRW_FSPRIV('L', 3, struct lfs_fcntl_markv) 847 #define LFCNRECLAIM _FCNO_FSPRIV('L', 4) 848 849 struct lfs_fhandle { 850 char space[28]; /* FHANDLE_SIZE_COMPAT (but used from userland too) */ 851 }; 852 #define LFCNREWIND _FCNR_FSPRIV('L', 6, int) 853 #define LFCNINVAL _FCNR_FSPRIV('L', 7, int) 854 #define LFCNRESIZE _FCNR_FSPRIV('L', 8, int) 855 #define LFCNWRAPSTOP _FCNR_FSPRIV('L', 9, int) 856 #define LFCNWRAPGO _FCNR_FSPRIV('L', 10, int) 857 #define LFCNIFILEFH _FCNW_FSPRIV('L', 11, struct lfs_fhandle) 858 #define LFCNWRAPPASS _FCNR_FSPRIV('L', 12, int) 859 # define LFS_WRAP_GOING 0x0 860 # define LFS_WRAP_WAITING 0x1 861 #define LFCNWRAPSTATUS _FCNW_FSPRIV('L', 13, int) 862 863 /* Debug segment lock */ 864 #ifdef notyet 865 # define ASSERT_SEGLOCK(fs) KASSERT(LFS_SEGLOCK_HELD(fs)) 866 # define ASSERT_NO_SEGLOCK(fs) KASSERT(!LFS_SEGLOCK_HELD(fs)) 867 # define ASSERT_DUNNO_SEGLOCK(fs) 868 # define ASSERT_MAYBE_SEGLOCK(fs) 869 #else /* !notyet */ 870 # define ASSERT_DUNNO_SEGLOCK(fs) \ 871 DLOG((DLOG_SEG, "lfs func %s seglock wrong (%d)\n", __func__, \ 872 LFS_SEGLOCK_HELD(fs))) 873 # define ASSERT_SEGLOCK(fs) do { \ 874 if (!LFS_SEGLOCK_HELD(fs)) { \ 875 DLOG((DLOG_SEG, "lfs func %s seglock wrong (0)\n", __func__)); \ 876 } \ 877 } while(0) 878 # define ASSERT_NO_SEGLOCK(fs) do { \ 879 if (LFS_SEGLOCK_HELD(fs)) { \ 880 DLOG((DLOG_SEG, "lfs func %s seglock wrong (1)\n", __func__)); \ 881 } \ 882 } while(0) 883 # define ASSERT_MAYBE_SEGLOCK(x) 884 #endif /* !notyet */ 885 886 /* 887 * Arguments to mount LFS filesystems 888 */ 889 struct ulfs_args { 890 char *fspec; /* block special device to mount */ 891 }; 892 893 __BEGIN_DECLS 894 void lfs_itimes(struct inode *, const struct timespec *, 895 const struct timespec *, const struct timespec *); 896 __END_DECLS 897 898 #endif /* !_UFS_LFS_LFS_H_ */ 899
Indexes created Fri Sep 26 20:09:58 GMT 2025