1 /*- 2 * SPDX-License-Identifier: BSD-3-Clause 3 * 4 * Copyright (c) 1982, 1986, 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 * @(#)fs.h 8.13 (Berkeley) 3/21/95 32 * $FreeBSD$ 33 */ 34 35 #ifndef UFS2_H 36 #define UFS2_H 37 38 #include "system_dependencies.h" 39 40 typedef uint32 cluster_t; 41 typedef uint64_t ufs2_daddr_t; 42 43 #define SBLOCK_FLOPPY 0 44 #define SBLOCK_UFS1 8192 45 #define SBLOCK_UFS2 65536 46 #define SBLOCK_PIGGY 262144 47 #define SBLOCKSIZE 8192 48 #define SBLOCKSEARCH \ 49 { SBLOCK_UFS2, SBLOCK_UFS1, SBLOCK_FLOPPY, SBLOCK_PIGGY, -1 } 50 51 /* 52 * Max number of fragments per block. This value is NOT tweakable. 53 */ 54 #define MAXFRAG 8 55 56 /* 57 * Addresses stored in inodes are capable of addressing fragments 58 * of `blocks'. File system blocks of at most size MAXBSIZE can 59 * be optionally broken into 2, 4, or 8 pieces, each of which is 60 * addressable; these pieces may be DEV_BSIZE, or some multiple of 61 * a DEV_BSIZE unit. 62 * 63 * Large files consist of exclusively large data blocks. To avoid 64 * undue wasted disk space, the last data block of a small file may be 65 * allocated as only as many fragments of a large block as are 66 * necessary. The filesystem format retains only a single pointer 67 * to such a fragment, which is a piece of a single large block that 68 * has been divided. The size of such a fragment is determinable from 69 * information in the inode, using the ``blksize(fs, ip, lbn)'' macro. 70 * 71 * The filesystem records space availability at the fragment level; 72 * to determine block availability, aligned fragments are examined. 73 */ 74 75 /* 76 * The path name on which the filesystem is mounted is maintained 77 * in fs_fsmnt. MAXMNTLEN defines the amount of space allocated in 78 * the super block for this name. 79 */ 80 #define MAXMNTLEN 468 81 82 /* 83 * The volume name for this filesystem is maintained in fs_volname. 84 * MAXVOLLEN defines the length of the buffer allocated. 85 */ 86 #define MAXVOLLEN 32 87 88 /* 89 * There is a 128-byte region in the superblock reserved for in-core 90 * pointers to summary information. Originally this included an array 91 * of pointers to blocks of struct csum; now there are just a few 92 * pointers and the remaining space is padded with fs_ocsp[]. 93 * 94 * NOCSPTRS determines the size of this padding. One pointer (fs_csp) 95 * is taken away to point to a contiguous array of struct csum for 96 * all cylinder groups; a second (fs_maxcluster) points to an array 97 * of cluster sizes that is computed as cylinder groups are inspected, 98 * and the third points to an array that tracks the creation of new 99 * directories. A fourth pointer, fs_active, is used when creating 100 * snapshots; it points to a bitmap of cylinder groups for which the 101 * free-block bitmap has changed since the snapshot operation began. 102 */ 103 #define NOCSPTRS ((128 / sizeof(void *)) - 4) 104 105 /* 106 * A summary of contiguous blocks of various sizes is maintained 107 * in each cylinder group. Normally this is set by the initial 108 * value of fs_maxcontig. To conserve space, a maximum summary size 109 * is set by FS_MAXCONTIG. 110 */ 111 #define FS_MAXCONTIG 16 112 113 /* 114 * MINFREE gives the minimum acceptable percentage of filesystem 115 * blocks which may be free. If the freelist drops below this level 116 * only the superuser may continue to allocate blocks. This may 117 * be set to 0 if no reserve of free blocks is deemed necessary, 118 * however throughput drops by fifty percent if the filesystem 119 * is run at between 95% and 100% full; thus the minimum default 120 * value of fs_minfree is 5%. However, to get good clustering 121 * performance, 10% is a better choice. hence we use 10% as our 122 * default value. With 10% free space, fragmentation is not a 123 * problem, so we choose to optimize for time. 124 */ 125 #define MINFREE 8 126 #define DEFAULTOPT FS_OPTTIME 127 128 /* 129 * Grigoriy Orlov <gluk@ptci.ru> has done some extensive work to fine 130 * tune the layout preferences for directories within a filesystem. 131 * His algorithm can be tuned by adjusting the following parameters 132 * which tell the system the average file size and the average number 133 * of files per directory. These defaults are well selected for typical 134 * filesystems, but may need to be tuned for odd cases like filesystems 135 * being used for squid caches or news spools. 136 */ 137 #define AVFILESIZ 16384 /* expected average file size */ 138 #define AFPDIR 64 /* expected number of files per directory */ 139 140 /* 141 * The maximum number of snapshot nodes that can be associated 142 * with each filesystem. This limit affects only the number of 143 * snapshot files that can be recorded within the superblock so 144 * that they can be found when the filesystem is mounted. However, 145 * maintaining too many will slow the filesystem performance, so 146 * having this limit is a good idea. 147 */ 148 #define FSMAXSNAP 20 149 150 /* 151 * Used to identify special blocks in snapshots: 152 * 153 * BLK_NOCOPY - A block that was unallocated at the time the snapshot 154 * was taken, hence does not need to be copied when written. 155 * BLK_SNAP - A block held by another snapshot that is not needed by this 156 * snapshot. When the other snapshot is freed, the BLK_SNAP entries 157 * are converted to BLK_NOCOPY. These are needed to allow fsck to 158 * identify blocks that are in use by other snapshots (which are 159 * expunged from this snapshot). 160 */ 161 #define BLK_NOCOPY ((ufs2_daddr_t)(1)) 162 #define BLK_SNAP ((ufs2_daddr_t)(2)) 163 164 /* 165 * Sysctl values for the fast filesystem. 166 */ 167 #define FFS_ADJ_REFCNT 1 /* adjust inode reference count */ 168 #define FFS_ADJ_BLKCNT 2 /* adjust inode used block count */ 169 #define FFS_BLK_FREE 3 /* free range of blocks in map */ 170 #define FFS_DIR_FREE 4 /* free specified dir inodes in map */ 171 #define FFS_FILE_FREE 5 /* free specified file inodes in map */ 172 #define FFS_SET_FLAGS 6 /* set filesystem flags */ 173 #define FFS_ADJ_NDIR 7 /* adjust number of directories */ 174 #define FFS_ADJ_NBFREE 8 /* adjust number of free blocks */ 175 #define FFS_ADJ_NIFREE 9 /* adjust number of free inodes */ 176 #define FFS_ADJ_NFFREE 10 /* adjust number of free frags */ 177 #define FFS_ADJ_NUMCLUSTERS 11 /* adjust number of free clusters */ 178 #define FFS_SET_CWD 12 /* set current directory */ 179 #define FFS_SET_DOTDOT 13 /* set inode number for ".." */ 180 #define FFS_UNLINK 14 /* remove a name in the filesystem */ 181 /* Was FFS_SET_INODE 15 */ 182 /* Was FFS_SET_BUFOUTPUT 16 */ 183 #define FFS_SET_SIZE 17 /* set inode size */ 184 #define FFS_MAXID 17 /* number of valid ffs ids */ 185 186 /* 187 * Command structure passed in to the filesystem to adjust filesystem values. 188 */ 189 #define FFS_CMD_VERSION 0x19790518 /* version ID */ 190 struct fsck_cmd { 191 int32_t version; /* version of command structure */ 192 int32_t handle; /* reference to filesystem to be changed */ 193 int64_t value; /* inode or block number to be affected */ 194 int64_t size; /* amount or range to be adjusted */ 195 int64_t spare; /* reserved for future use */ 196 }; 197 198 /* 199 * A recovery structure placed at the end of the boot block area by newfs 200 * that can be used by fsck to search for alternate superblocks. 201 */ 202 struct fsrecovery { 203 int32_t fsr_magic; /* magic number */ 204 int32_t fsr_fsbtodb; /* fsbtodb and dbtofsb shift constant */ 205 int32_t fsr_sblkno; /* offset of super-block in filesys */ 206 int32_t fsr_fpg; /* blocks per group * fs_frag */ 207 u_int32_t fsr_ncg; /* number of cylinder groups */ 208 }; 209 210 /* 211 * Per cylinder group information; summarized in blocks allocated 212 * from first cylinder group data blocks. These blocks have to be 213 * read in from fs_csaddr (size fs_cssize) in addition to the 214 * super block. 215 */ 216 struct csum { 217 int32_t cs_ndir; /* number of directories */ 218 int32_t cs_nbfree; /* number of free blocks */ 219 int32_t cs_nifree; /* number of free inodes */ 220 int32_t cs_nffree; /* number of free frags */ 221 }; 222 struct csum_total { 223 int64_t cs_ndir; /* number of directories */ 224 int64_t cs_nbfree; /* number of free blocks */ 225 int64_t cs_nifree; /* number of free inodes */ 226 int64_t cs_nffree; /* number of free frags */ 227 int64_t cs_numclusters; /* number of free clusters */ 228 int64_t cs_spare[3]; /* future expansion */ 229 }; 230 231 /* 232 * Super block for an FFS filesystem. 233 */ 234 struct ufs2_super_block { 235 int32_t fs_firstfield; /* historic filesystem linked list, */ 236 int32_t fs_unused_1; /* used for incore super blocks */ 237 int32_t fs_sblkno; /* offset of super-block in filesys */ 238 int32_t fs_cblkno; /* offset of cyl-block in filesys */ 239 int32_t fs_iblkno; /* offset of inode-blocks in filesys */ 240 int32_t fs_dblkno; /* offset of first data after cg */ 241 int32_t fs_old_cgoffset; /* cylinder group offset in cylinder */ 242 int32_t fs_old_cgmask; /* used to calc mod fs_ntrak */ 243 int32_t fs_old_time; /* last time written */ 244 int32_t fs_old_size; /* number of blocks in fs */ 245 int32_t fs_old_dsize; /* number of data blocks in fs */ 246 u_int32_t fs_ncg; /* number of cylinder groups */ 247 int32_t fs_bsize; /* size of basic blocks in fs */ 248 int32_t fs_fsize; /* size of frag blocks in fs */ 249 int32_t fs_frag; /* number of frags in a block in fs */ 250 /* these are configuration parameters */ 251 int32_t fs_minfree; /* minimum percentage of free blocks */ 252 int32_t fs_old_rotdelay; /* num of ms for optimal next block */ 253 int32_t fs_old_rps; /* disk revolutions per second */ 254 /* these fields can be computed from the others */ 255 int32_t fs_bmask; /* ``blkoff'' calc of blk offsets */ 256 int32_t fs_fmask; /* ``fragoff'' calc of frag offsets */ 257 int32_t fs_bshift; /* ``lblkno'' calc of logical blkno */ 258 int32_t fs_fshift; /* ``numfrags'' calc number of frags */ 259 /* these are configuration parameters */ 260 int32_t fs_maxcontig; /* max number of contiguous blks */ 261 int32_t fs_maxbpg; /* max number of blks per cyl group */ 262 /* these fields can be computed from the others */ 263 int32_t fs_fragshift; /* block to frag shift */ 264 int32_t fs_fsbtodb; /* fsbtodb and dbtofsb shift constant */ 265 int32_t fs_sbsize; /* actual size of super block */ 266 int32_t fs_spare1[2]; /* old fs_csmask */ 267 /* old fs_csshift */ 268 int32_t fs_nindir; /* value of NINDIR */ 269 u_int32_t fs_inopb; /* value of INOPB */ 270 int32_t fs_old_nspf; /* value of NSPF */ 271 /* yet another configuration parameter */ 272 int32_t fs_optim; /* optimization preference, see below */ 273 int32_t fs_old_npsect; /* # sectors/track including spares */ 274 int32_t fs_old_interleave; /* hardware sector interleave */ 275 int32_t fs_old_trackskew; /* sector 0 skew, per track */ 276 int32_t fs_id[2]; /* unique filesystem id */ 277 /* sizes determined by number of cylinder groups and their sizes */ 278 int32_t fs_old_csaddr; /* blk addr of cyl grp summary area */ 279 int32_t fs_cssize; /* size of cyl grp summary area */ 280 int32_t fs_cgsize; /* cylinder group size */ 281 int32_t fs_spare2; /* old fs_ntrak */ 282 int32_t fs_old_nsect; /* sectors per track */ 283 int32_t fs_old_spc; /* sectors per cylinder */ 284 int32_t fs_old_ncyl; /* cylinders in filesystem */ 285 int32_t fs_old_cpg; /* cylinders per group */ 286 u_int32_t fs_ipg; /* inodes per group */ 287 int32_t fs_fpg; /* blocks per group * fs_frag */ 288 /* this data must be re-computed after crashes */ 289 struct csum fs_old_cstotal; /* cylinder summary information */ 290 /* these fields are cleared at mount time */ 291 int8_t fs_fmod; /* super block modified flag */ 292 int8_t fs_clean; /* filesystem is clean flag */ 293 int8_t fs_ronly; /* mounted read-only flag */ 294 int8_t fs_old_flags; /* old FS_ flags */ 295 u_char fs_fsmnt[MAXMNTLEN]; /* name mounted on */ 296 char fs_volname[MAXVOLLEN]; /* volume name */ 297 u_int64_t fs_swuid; /* system-wide uid */ 298 int32_t fs_pad; /* due to alignment of fs_swuid */ 299 /* these fields retain the current block allocation info */ 300 int32_t fs_cgrotor; /* last cg searched */ 301 void *fs_ocsp[NOCSPTRS]; /* padding; was list of fs_cs buffers */ 302 u_int8_t *fs_contigdirs; /* (u) # of contig. allocated dirs */ 303 struct csum *fs_csp; /* (u) cg summary info buffer */ 304 int32_t *fs_maxcluster; /* (u) max cluster in each cyl group */ 305 u_int *fs_active; /* (u) used by snapshots to track fs */ 306 int32_t fs_old_cpc; /* cyl per cycle in postbl */ 307 int32_t fs_maxbsize; /* maximum blocking factor permitted */ 308 int64_t fs_unrefs; /* number of unreferenced inodes */ 309 int64_t fs_providersize; /* size of underlying GEOM provider */ 310 int64_t fs_metaspace; /* size of area reserved for metadata */ 311 int64_t fs_sparecon64[13]; /* old rotation block list head */ 312 int64_t fs_sblockactualloc; /* byte offset of this superblock */ 313 int64_t fs_sblockloc; /* byte offset of standard superblock */ 314 struct csum_total fs_cstotal; /* (u) cylinder summary information */ 315 int64_t fs_time; /* last time written */ 316 int64_t fs_size; /* number of blocks in fs */ 317 int64_t fs_dsize; /* number of data blocks in fs */ 318 int64_t fs_csaddr; /* blk addr of cyl grp summary area */ 319 int64_t fs_pendingblocks; /* (u) blocks being freed */ 320 u_int32_t fs_pendinginodes; /* (u) inodes being freed */ 321 uint32_t fs_snapinum[FSMAXSNAP];/* list of snapshot inode numbers */ 322 u_int32_t fs_avgfilesize; /* expected average file size */ 323 u_int32_t fs_avgfpdir; /* expected # of files per directory */ 324 int32_t fs_save_cgsize; /* save real cg size to use fs_bsize */ 325 int64_t fs_mtime; /* Last mount or fsck time. */ 326 int32_t fs_sujfree; /* SUJ free list */ 327 int32_t fs_sparecon32[21]; /* reserved for future constants */ 328 u_int32_t fs_ckhash; /* if CK_SUPERBLOCK, its check-hash */ 329 u_int32_t fs_metackhash; /* metadata check-hash, see CK_ below */ 330 int32_t fs_flags; /* see FS_ flags below */ 331 int32_t fs_contigsumsize; /* size of cluster summary array */ 332 int32_t fs_maxsymlinklen; /* max length of an internal symlink */ 333 int32_t fs_old_inodefmt; /* format of on-disk inodes */ 334 u_int64_t fs_maxfilesize; /* maximum representable file size */ 335 int64_t fs_qbmask; /* ~fs_bmask for use with 64-bit size */ 336 int64_t fs_qfmask; /* ~fs_fmask for use with 64-bit size */ 337 int32_t fs_state; /* validate fs_clean field */ 338 int32_t fs_old_postblformat; /* format of positional layout tables */ 339 int32_t fs_old_nrpos; /* number of rotational positions */ 340 int32_t fs_spare5[2]; /* old fs_postbloff */ 341 /* old fs_rotbloff */ 342 int32_t fs_magic; /* magic number */ 343 bool IsValid(); 344 }; 345 346 /* Sanity checking. */ 347 #ifdef CTASSERT 348 CTASSERT(sizeof(struct fs) == 1376); 349 #endif 350 351 352 struct file_cookie { 353 bigtime_t last_notification; 354 off_t last_size; 355 int open_mode; 356 }; 357 358 359 /* 360 * Filesystem identification 361 */ 362 #define FS_UFS1_MAGIC 0x011954 /* UFS1 fast filesystem magic number */ 363 #define FS_UFS2_MAGIC 0x19540119 /* UFS2 fast filesystem magic number */ 364 #define FS_BAD_MAGIC 0x19960408 /* UFS incomplete newfs magic number */ 365 #define FS_42INODEFMT -1 /* 4.2BSD inode format */ 366 #define FS_44INODEFMT 2 /* 4.4BSD inode format */ 367 368 /* 369 * Preference for optimization. 370 */ 371 #define FS_OPTTIME 0 /* minimize allocation time */ 372 #define FS_OPTSPACE 1 /* minimize disk fragmentation */ 373 374 /* 375 * Filesystem flags. 376 * 377 * The FS_UNCLEAN flag is set by the kernel when the filesystem was 378 * mounted with fs_clean set to zero. The FS_DOSOFTDEP flag indicates 379 * that the filesystem should be managed by the soft updates code. 380 * Note that the FS_NEEDSFSCK flag is set and cleared by the fsck 381 * utility. It is set when background fsck finds an unexpected 382 * inconsistency which requires a traditional foreground fsck to be 383 * run. Such inconsistencies should only be found after an uncorrectable 384 * disk error. The FS_NEEDSFSCK can also be set when a mounted filesystem 385 * discovers an internal inconsistency such as freeing a freed inode. 386 * A foreground fsck will clear the FS_NEEDSFSCK flag when it has 387 * successfully cleaned up the filesystem. The kernel uses this 388 * flag to enforce that inconsistent filesystems be mounted read-only. 389 * 390 * The FS_METACKHASH flag when set indicates that the kernel maintains 391 * one or more check hashes. The actual set of supported check hashes 392 * is stored in the fs_metackhash field. Kernels that do not support 393 * check hashes clear the FS_METACKHASH flag to indicate that the 394 * check hashes need to be rebuilt (by fsck) before they can be used. 395 * 396 * When a filesystem is mounted, any flags not included in FS_SUPPORTED 397 * are cleared. This lets newer features know that the filesystem has 398 * been run on an older version of the filesystem and thus that data 399 * structures associated with those features are out-of-date and need 400 * to be rebuilt. 401 * 402 * FS_ACLS indicates that POSIX.1e ACLs are administratively enabled 403 * for the file system, so they should be loaded from extended attributes, 404 * observed for access control purposes, and be administered by object 405 * owners. FS_NFS4ACLS indicates that NFSv4 ACLs are administratively 406 * enabled. This flag is mutually exclusive with FS_ACLS. FS_MULTILABEL 407 * indicates that the TrustedBSD MAC Framework should attempt to back MAC 408 * labels into extended attributes on the file system rather than maintain 409 * a single mount label for all objects. 410 */ 411 #define FS_UNCLEAN 0x00000001 /* filesystem not clean at mount */ 412 #define FS_DOSOFTDEP 0x00000002 /* filesystem using soft dependencies */ 413 /* filesystem needs sync fsck before mount */ 414 #define FS_NEEDSFSCK 0x00000004 415 #define FS_SUJ 0x00000008 /* Filesystem using softupdate journal */ 416 /* file system has POSIX.1e ACLs enabled */ 417 #define FS_ACLS 0x00000010 418 #define FS_MULTILABEL 0x00000020 /* file system is MAC multi-label */ 419 #define FS_GJOURNAL 0x00000040 /* gjournaled file system */ 420 /* flags have been moved to new location */ 421 #define FS_FLAGS_UPDATED 0x0000080 422 #define FS_NFS4ACLS 0x00000100 /* file system has NFSv4 ACLs enabled */ 423 /* kernel supports metadata check hashes */ 424 #define FS_METACKHASH 0x00000200 425 #define FS_TRIM 0x00000400 /* issue BIO_DELETE for deleted blocks */ 426 /* supported flags, others cleared at mount */ 427 #define FS_SUPPORTED 0x00FFFFFF 428 /* 429 * Things that we may someday support, but currently do not. 430 * These flags are all cleared so we know if we ran on a kernel 431 * that does not support them. 432 */ 433 #define FS_INDEXDIRS 0x01000000 /* kernel supports indexed directories */ 434 /* kernel supports variable block sizes */ 435 #define FS_VARBLKSIZE 0x02000000 436 #define FS_COOLOPT1 0x04000000 /* kernel supports cool option 1 */ 437 #define FS_COOLOPT2 0x08000000 /* kernel supports cool option 2 */ 438 #define FS_COOLOPT3 0x10000000 /* kernel supports cool option 3 */ 439 #define FS_COOLOPT4 0x20000000 /* kernel supports cool option 4 */ 440 #define FS_COOLOPT5 0x40000000 /* kernel supports cool option 5 */ 441 #define FS_COOLOPT6 0x80000000 /* kernel supports cool option 6 */ 442 443 /* 444 * The fs_metackhash field indicates the types of metadata check-hash 445 * that are maintained for a filesystem. Not all filesystems check-hash 446 * all metadata. 447 */ 448 #define CK_SUPERBLOCK 0x0001 /* the superblock */ 449 #define CK_CYLGRP 0x0002 /* the cylinder groups */ 450 #define CK_INODE 0x0004 /* inodes */ 451 #define CK_INDIR 0x0008 /* indirect blocks */ 452 #define CK_DIR 0x0010 /* directory contents */ 453 /* 454 * The BX_FSPRIV buffer b_xflags are used to track types of data in buffers. 455 */ 456 #define BX_SUPERBLOCK 0x00010000 /* superblock */ 457 #define BX_CYLGRP 0x00020000 /* cylinder groups */ 458 #define BX_INODE 0x00040000 /* inodes */ 459 #define BX_INDIR 0x00080000 /* indirect blocks */ 460 #define BX_DIR 0x00100000 /* directory contents */ 461 462 /* 463 #define PRINT_UFS_BUF_XFLAGS "\20\25dir\24indir\23inode\22cylgrp\21superblock" 464 */ 465 466 /* 467 * Macros to access bits in the fs_active array. 468 */ 469 #define ACTIVECGNUM(fs, cg) ((fs)->fs_active[(cg) / (NBBY * sizeof(int))]) 470 #define ACTIVECGOFF(cg) (1 << ((cg) % (NBBY * sizeof(int)))) 471 #define ACTIVESET(fs, cg) do { \ 472 if ((fs)->fs_active) \ 473 ACTIVECGNUM((fs), (cg)) |= ACTIVECGOFF((cg)); \ 474 } while (0) 475 #define ACTIVECLEAR(fs, cg) do { \ 476 if ((fs)->fs_active) \ 477 ACTIVECGNUM((fs), (cg)) &= ~ACTIVECGOFF((cg)); \ 478 } while (0) 479 480 /* 481 * The size of a cylinder group is calculated by CGSIZE. The maximum size 482 * is limited by the fact that cylinder groups are at most one block. 483 * Its size is derived from the size of the maps maintained in the 484 * cylinder group and the (struct cg) size. 485 */ 486 #define CGSIZE(fs) \ 487 /* base cg */ (sizeof(struct cg) + sizeof(int32_t) + \ 488 /* old btotoff */ (fs)->fs_old_cpg * sizeof(int32_t) + \ 489 /* old boff */ (fs)->fs_old_cpg * sizeof(u_int16_t) + \ 490 /* inode map */ howmany((fs)->fs_ipg, NBBY) + \ 491 /* block map */ howmany((fs)->fs_fpg, NBBY) +\ 492 /* if present */ ((fs)->fs_contigsumsize <= 0 ? 0 : \ 493 /* cluster sum */ (fs)->fs_contigsumsize * sizeof(int32_t) + \ 494 /* cluster map */ howmany(fragstoblks(fs, (fs)->fs_fpg), NBBY))) 495 496 /* 497 * The minimal number of cylinder groups that should be created. 498 */ 499 #define MINCYLGRPS 4 500 501 /* 502 * Convert cylinder group to base address of its global summary info. 503 */ 504 #define fs_cs(fs, indx) fs_csp[indx] 505 506 /* 507 * Cylinder group block for a filesystem. 508 */ 509 #define CG_MAGIC 0x090255 510 struct cg { 511 int32_t cg_firstfield; /* historic cyl groups linked list */ 512 int32_t cg_magic; /* magic number */ 513 int32_t cg_old_time; /* time last written */ 514 u_int32_t cg_cgx; /* we are the cgx'th cylinder group */ 515 int16_t cg_old_ncyl; /* number of cyl's this cg */ 516 int16_t cg_old_niblk; /* number of inode blocks this cg */ 517 u_int32_t cg_ndblk; /* number of data blocks this cg */ 518 struct csum cg_cs; /* cylinder summary information */ 519 u_int32_t cg_rotor; /* position of last used block */ 520 u_int32_t cg_frotor; /* position of last used frag */ 521 u_int32_t cg_irotor; /* position of last used inode */ 522 u_int32_t cg_frsum[MAXFRAG]; /* counts of available frags */ 523 int32_t cg_old_btotoff; /* (int32) block totals per cylinder */ 524 int32_t cg_old_boff; /* (u_int16) free block positions */ 525 u_int32_t cg_iusedoff; /* (u_int8) used inode map */ 526 u_int32_t cg_freeoff; /* (u_int8) free block map */ 527 u_int32_t cg_nextfreeoff; /* (u_int8) next available space */ 528 u_int32_t cg_clustersumoff; /* (u_int32) counts of avail clusters */ 529 u_int32_t cg_clusteroff; /* (u_int8) free cluster map */ 530 u_int32_t cg_nclusterblks; /* number of clusters this cg */ 531 u_int32_t cg_niblk; /* number of inode blocks this cg */ 532 u_int32_t cg_initediblk; /* last initialized inode */ 533 u_int32_t cg_unrefs; /* number of unreferenced inodes */ 534 int32_t cg_sparecon32[1]; /* reserved for future use */ 535 u_int32_t cg_ckhash; /* check-hash of this cg */ 536 int64_t cg_time; /* time last written */ 537 int64_t cg_sparecon64[3]; /* reserved for future use */ 538 u_int8_t cg_space[1]; /* space for cylinder group maps */ 539 /* actually longer */ 540 }; 541 542 /* 543 * Macros for access to cylinder group array structures 544 */ 545 #define cg_chkmagic(cgp) ((cgp)->cg_magic == CG_MAGIC) 546 #define cg_inosused(cgp) \ 547 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_iusedoff)) 548 #define cg_blksfree(cgp) \ 549 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_freeoff)) 550 #define cg_clustersfree(cgp) \ 551 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_clusteroff)) 552 #define cg_clustersum(cgp) \ 553 ((int32_t *)((uintptr_t)(cgp) + (cgp)->cg_clustersumoff)) 554 555 /* 556 * Turn filesystem block numbers into disk block addresses. 557 * This maps filesystem blocks to device size blocks. 558 */ 559 #define fsbtodb(fs, b) ((daddr_t)(b) << (fs)->fs_fsbtodb) 560 #define dbtofsb(fs, b) ((b) >> (fs)->fs_fsbtodb) 561 562 /* 563 * Cylinder group macros to locate things in cylinder groups. 564 * They calc filesystem addresses of cylinder group data structures. 565 */ 566 #define cgbase(fs, c) (((ufs2_daddr_t)(fs)->fs_fpg) * (c)) 567 #define cgdata(fs, c) (cgdmin(fs, c) + (fs)->fs_metaspace) /* data zone */ 568 #define cgmeta(fs, c) (cgdmin(fs, c)) /* meta data */ 569 #define cgdmin(fs, c) (cgstart(fs, c) + (fs)->fs_dblkno) /* 1st data */ 570 #define cgimin(fs, c) (cgstart(fs, c) + (fs)->fs_iblkno) /* inode blk */ 571 #define cgsblock(fs, c) (cgstart(fs, c) + (fs)->fs_sblkno) /* super blk */ 572 #define cgtod(fs, c) (cgstart(fs, c) + (fs)->fs_cblkno) /* cg block */ 573 #define cgstart(fs, c) \ 574 ((fs)->fs_magic == FS_UFS2_MAGIC ? cgbase(fs, c) : \ 575 (cgbase(fs, c) + (fs)->fs_old_cgoffset * ((c) & ~((fs)->fs_old_cgmask)))) 576 577 /* 578 * Macros for handling inode numbers: 579 * inode number to filesystem block offset. 580 * inode number to cylinder group number. 581 * inode number to filesystem block address. 582 */ 583 #define ino_to_cg(fs, x) (((ino_t)(x)) / (fs)->fs_ipg) 584 #define ino_to_fsba(fs, x) \ 585 ((ufs2_daddr_t)(cgimin(fs, ino_to_cg(fs, (ino_t)(x))) + \ 586 (blkstofrags((fs), ((((ino_t)(x)) % (fs)->fs_ipg) / INOPB(fs)))))) 587 #define ino_to_fsbo(fs, x) (((ino_t)(x)) % INOPB(fs)) 588 589 /* 590 * Give cylinder group number for a filesystem block. 591 * Give cylinder group block number for a filesystem block. 592 */ 593 #define dtog(fs, d) ((d) / (fs)->fs_fpg) 594 #define dtogd(fs, d) ((d) % (fs)->fs_fpg) 595 596 /* 597 * Extract the bits for a block from a map. 598 * Compute the cylinder and rotational position of a cyl block addr. 599 */ 600 #define blkmap(fs, map, loc) \ 601 (((map)[(loc) / NBBY] >> ((loc) % NBBY)) & (0xff >> (NBBY - (fs)->fs_frag))) 602 603 /* 604 * The following macros optimize certain frequently calculated 605 * quantities by using shifts and masks in place of divisions 606 * modulos and multiplications. 607 */ 608 #define blkoff(fs, loc) /* calculates (loc % fs->fs_bsize) */ \ 609 ((loc) & (fs)->fs_qbmask) 610 #define fragoff(fs, loc) /* calculates (loc % fs->fs_fsize) */ \ 611 ((loc) & (fs)->fs_qfmask) 612 #define lfragtosize(fs, frag) /* calculates ((off_t)frag * fs->fs_fsize) */ \ 613 (((off_t)(frag)) << (fs)->fs_fshift) 614 #define lblktosize(fs, blk) /* calculates ((off_t)blk * fs->fs_bsize) */ \ 615 (((off_t)(blk)) << (fs)->fs_bshift) 616 /* Use this only when `blk' is known to be small, e.g., < UFS_NDADDR. */ 617 #define smalllblktosize(fs, blk) /* calculates (blk * fs->fs_bsize) */ \ 618 ((blk) << (fs)->fs_bshift) 619 #define lblkno(fs, loc) /* calculates (loc / fs->fs_bsize) */ \ 620 ((loc) >> (fs)->fs_bshift) 621 #define numfrags(fs, loc) /* calculates (loc / fs->fs_fsize) */ \ 622 ((loc) >> (fs)->fs_fshift) 623 #define blkroundup(fs, size) /* calculates roundup(size, fs->fs_bsize) */ \ 624 (((size) + (fs)->fs_qbmask) & (fs)->fs_bmask) 625 #define fragroundup(fs, size) /* calculates roundup(size, fs->fs_fsize) */ \ 626 (((size) + (fs)->fs_qfmask) & (fs)->fs_fmask) 627 #define fragstoblks(fs, frags) /* calculates (frags / fs->fs_frag) */ \ 628 ((frags) >> (fs)->fs_fragshift) 629 #define blkstofrags(fs, blks) /* calculates (blks * fs->fs_frag) */ \ 630 ((blks) << (fs)->fs_fragshift) 631 #define fragnum(fs, fsb) /* calculates (fsb % fs->fs_frag) */ \ 632 ((fsb) & ((fs)->fs_frag - 1)) 633 #define blknum(fs, fsb) /* calculates rounddown(fsb, fs->fs_frag) */ \ 634 ((fsb) &~ ((fs)->fs_frag - 1)) 635 636 /* 637 * Determine the number of available frags given a 638 * percentage to hold in reserve. 639 */ 640 #define freespace(fs, percentreserved) \ 641 (blkstofrags((fs), (fs)->fs_cstotal.cs_nbfree) + \ 642 (fs)->fs_cstotal.cs_nffree - \ 643 (((off_t)((fs)->fs_dsize)) * (percentreserved) / 100)) 644 645 /* 646 * Determining the size of a file block in the filesystem. 647 */ 648 #define blksize(fs, ip, lbn) \ 649 (((lbn) >= UFS_NDADDR || (ip)->i_size >= \ 650 (uint64_t)smalllblktosize(fs, (lbn) + 1)) \ 651 ? (fs)->fs_bsize \ 652 : (fragroundup(fs, blkoff(fs, (ip)->i_size)))) 653 #define sblksize(fs, size, lbn) \ 654 (((lbn) >= UFS_NDADDR || (size) >= ((lbn) + 1) << (fs)->fs_bshift) \ 655 ? (fs)->fs_bsize \ 656 : (fragroundup(fs, blkoff(fs, (size))))) 657 658 /* 659 * Number of indirects in a filesystem block. 660 */ 661 #define NINDIR(fs) ((fs)->fs_nindir) 662 663 /* 664 * Indirect lbns are aligned on UFS_NDADDR addresses where single indirects 665 * are the negated address of the lowest lbn reachable, double indirects 666 * are this lbn - 1 and triple indirects are this lbn - 2. This yields 667 * an unusual bit order to determine level. 668 */ 669 /*static inline int 670 lbn_level(ufs_lbn_t lbn) 671 { 672 if (lbn >= 0) 673 return 0; 674 switch (lbn & 0x3) { 675 case 0: 676 return (0); 677 case 1: 678 break; 679 case 2: 680 return (2); 681 case 3: 682 return (1); 683 default: 684 break; 685 } 686 return (-1); 687 } 688 689 static inline ufs_lbn_t 690 lbn_offset(struct fs *fs, int level) 691 { 692 ufs_lbn_t res; 693 694 for (res = 1; level > 0; level--) 695 res *= NINDIR(fs); 696 return (res); 697 }*/ 698 699 /* 700 * Number of inodes in a secondary storage block/fragment. 701 */ 702 #define INOPB(fs) ((fs)->fs_inopb) 703 #define INOPF(fs) ((fs)->fs_inopb >> (fs)->fs_fragshift) 704 705 /* 706 * Softdep journal record format. 707 */ 708 709 #define JOP_ADDREF 1 /* Add a reference to an inode. */ 710 #define JOP_REMREF 2 /* Remove a reference from an inode. */ 711 #define JOP_NEWBLK 3 /* Allocate a block. */ 712 #define JOP_FREEBLK 4 /* Free a block or a tree of blocks. */ 713 #define JOP_MVREF 5 /* Move a reference from one off to another. */ 714 #define JOP_TRUNC 6 /* Partial truncation record. */ 715 #define JOP_SYNC 7 /* fsync() complete record. */ 716 717 #define JREC_SIZE 32 /* Record and segment header size. */ 718 719 #define SUJ_MIN (4 * 1024 * 1024) /* Minimum journal size */ 720 #define SUJ_MAX (32 * 1024 * 1024) /* Maximum journal size */ 721 #define SUJ_FILE ".sujournal" /* Journal file name */ 722 723 /* 724 * Size of the segment record header. There is at most one for each disk 725 * block in the journal. The segment header is followed by an array of 726 * records. fsck depends on the first element in each record being 'op' 727 * and the second being 'ino'. Segments may span multiple disk blocks but 728 * the header is present on each. 729 */ 730 struct jsegrec { 731 uint64_t jsr_seq; /* Our sequence number */ 732 uint64_t jsr_oldest; /* Oldest valid sequence number */ 733 uint16_t jsr_cnt; /* Count of valid records */ 734 uint16_t jsr_blocks; /* Count of device bsize blocks. */ 735 uint32_t jsr_crc; /* 32bit crc of the valid space */ 736 int64_t jsr_time; /* timestamp for mount instance */ 737 }; 738 739 /* 740 * Reference record. Records a single link count modification. 741 */ 742 struct jrefrec { 743 uint32_t jr_op; 744 uint32_t jr_ino; 745 uint32_t jr_parent; 746 uint16_t jr_nlink; 747 uint16_t jr_mode; 748 int64_t jr_diroff; 749 uint64_t jr_unused; 750 }; 751 752 /* 753 * Move record. Records a reference moving within a directory block. The 754 * nlink is unchanged but we must search both locations. 755 */ 756 struct jmvrec { 757 uint32_t jm_op; 758 uint32_t jm_ino; 759 uint32_t jm_parent; 760 uint16_t jm_unused; 761 int64_t jm_oldoff; 762 int64_t jm_newoff; 763 }; 764 765 /* 766 * Block record. A set of frags or tree of blocks starting at an indirect are 767 * freed or a set of frags are allocated. 768 */ 769 struct jblkrec { 770 uint32_t jb_op; 771 uint32_t jb_ino; 772 int64_t jb_blkno; 773 int64_t jb_lbn; 774 uint16_t jb_frags; 775 uint16_t jb_oldfrags; 776 uint32_t jb_unused; 777 }; 778 779 /* 780 * Truncation record. Records a partial truncation so that it may be 781 * completed at check time. Also used for sync records. 782 */ 783 struct jtrncrec { 784 uint32_t jt_op; 785 uint32_t jt_ino; 786 int64_t jt_size; 787 uint32_t jt_extsize; 788 uint32_t jt_pad[3]; 789 }; 790 791 union jrec { 792 struct jsegrec rec_jsegrec; 793 struct jrefrec rec_jrefrec; 794 struct jmvrec rec_jmvrec; 795 struct jblkrec rec_jblkrec; 796 struct jtrncrec rec_jtrncrec; 797 }; 798 799 #ifdef CTASSERT 800 CTASSERT(sizeof(struct jsegrec) == JREC_SIZE); 801 CTASSERT(sizeof(struct jrefrec) == JREC_SIZE); 802 CTASSERT(sizeof(struct jmvrec) == JREC_SIZE); 803 CTASSERT(sizeof(struct jblkrec) == JREC_SIZE); 804 CTASSERT(sizeof(struct jtrncrec) == JREC_SIZE); 805 CTASSERT(sizeof(union jrec) == JREC_SIZE); 806 #endif 807 808 extern int inside[], around[]; 809 extern u_char *fragtbl[]; 810 811 /* 812 * IOCTLs used for filesystem write suspension. 813 */ 814 #define UFSSUSPEND _IOW('U', 1, fsid_t) 815 #define UFSRESUME _IO('U', 2) 816 817 #endif 818 819 820 821