1/* 2 * Mach Operating System 3 * Copyright (c) 1991,1990 Carnegie Mellon University 4 * All Rights Reserved. 5 * 6 * Permission to use, copy, modify and distribute this software and its 7 * documentation is hereby granted, provided that both the copyright 8 * notice and this permission notice appear in all copies of the 9 * software, derivative works or modified versions, and any portions 10 * thereof, and that both notices appear in supporting documentation. 11 * 12 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS" 13 * CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR 14 * ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE. 15 * 16 * Carnegie Mellon requests users of this software to return to 17 * 18 * Software Distribution Coordinator or Software.Distribution@CS.CMU.EDU 19 * School of Computer Science 20 * Carnegie Mellon University 21 * Pittsburgh PA 15213-3890 22 * 23 * any improvements or extensions that they make and grant Carnegie Mellon 24 * the rights to redistribute these changes. 25 */ 26/* 27 * Copyright (c) 1982, 1986 Regents of the University of California. 28 * All rights reserved. 29 * 30 * Redistribution and use in source and binary forms are permitted 31 * provided that the above copyright notice and this paragraph are 32 * duplicated in all such forms and that any documentation, 33 * advertising materials, and other materials related to such 34 * distribution and use acknowledge that the software was developed 35 * by the University of California, Berkeley. The name of the 36 * University may not be used to endorse or promote products derived 37 * from this software without specific prior written permission. 38 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR 39 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED 40 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. 41 * 42 * @(#)fs.h 7.7 (Berkeley) 5/9/89 43 */ 44 45/* 46 * Each disk drive contains some number of file systems. 47 * A file system consists of a number of cylinder groups. 48 * Each cylinder group has inodes and data. 49 * 50 * A file system is described by its super-block, which in turn 51 * describes the cylinder groups. The super-block is critical 52 * data and is replicated in each cylinder group to protect against 53 * catastrophic loss. This is done at `newfs' time and the critical 54 * super-block data does not change, so the copies need not be 55 * referenced further unless disaster strikes. 56 * 57 * For file system fs, the offsets of the various blocks of interest 58 * are given in the super block as: 59 * [fs->fs_sblkno] Super-block 60 * [fs->fs_cblkno] Cylinder group block 61 * [fs->fs_iblkno] Inode blocks 62 * [fs->fs_dblkno] Data blocks 63 * The beginning of cylinder group cg in fs, is given by 64 * the ``cgbase(fs, cg)'' macro. 65 * 66 * The first boot and super blocks are given in absolute disk addresses. 67 * The byte-offset forms are preferred, as they don't imply a sector size. 68 */ 69#define BBSIZE 8192 70#define SBSIZE 8192 71#define BBOFF ((mach_off_t)(0)) 72#define SBOFF ((mach_off_t)(BBOFF + BBSIZE)) 73#define BBLOCK ((mach_daddr_t)(0)) 74#define SBLOCK ((mach_daddr_t)(BBLOCK + BBSIZE / DEV_BSIZE)) 75 76/* 77 * Addresses stored in inodes are capable of addressing fragments 78 * of `blocks'. File system blocks of at most size MAXBSIZE can 79 * be optionally broken into 2, 4, or 8 pieces, each of which is 80 * addressible; these pieces may be DEV_BSIZE, or some multiple of 81 * a DEV_BSIZE unit. 82 * 83 * Large files consist of exclusively large data blocks. To avoid 84 * undue wasted disk space, the last data block of a small file may be 85 * allocated as only as many fragments of a large block as are 86 * necessary. The file system format retains only a single pointer 87 * to such a fragment, which is a piece of a single large block that 88 * has been divided. The size of such a fragment is determinable from 89 * information in the inode, using the ``blksize(fs, ip, lbn)'' macro. 90 * 91 * The file system records space availability at the fragment level; 92 * to determine block availability, aligned fragments are examined. 93 * 94 * The root inode is the root of the file system. 95 * Inode 0 can't be used for normal purposes and 96 * historically bad blocks were linked to inode 1, 97 * thus the root inode is 2. (inode 1 is no longer used for 98 * this purpose, however numerous dump tapes make this 99 * assumption, so we are stuck with it) 100 */ 101#define ROOTINO ((mach_ino_t)2) /* i number of all roots */ 102 103/* 104 * MINBSIZE is the smallest allowable block size. 105 * In order to insure that it is possible to create files of size 106 * 2^32 with only two levels of indirection, MINBSIZE is set to 4096. 107 * MINBSIZE must be big enough to hold a cylinder group block, 108 * thus changes to (struct cg) must keep its size within MINBSIZE. 109 * Note that super blocks are always of size SBSIZE, 110 * and that both SBSIZE and MAXBSIZE must be >= MINBSIZE. 111 */ 112#define MINBSIZE 4096 113 114/* 115 * The path name on which the file system is mounted is maintained 116 * in fs_fsmnt. MAXMNTLEN defines the amount of space allocated in 117 * the super block for this name. 118 * The limit on the amount of summary information per file system 119 * is defined by MAXCSBUFS. It is currently parameterized for a 120 * maximum of two million cylinders. 121 */ 122#define MAXMNTLEN 512 123#define MAXCSBUFS 32 124 125/* 126 * Per cylinder group information; summarized in blocks allocated 127 * from first cylinder group data blocks. These blocks have to be 128 * read in from fs_csaddr (size fs_cssize) in addition to the 129 * super block. 130 * 131 * N.B. sizeof(struct csum) must be a power of two in order for 132 * the ``fs_cs'' macro to work (see below). 133 */ 134struct csum 135 { 136 int cs_ndir; /* number of directories */ 137 int cs_nbfree; /* number of free blocks */ 138 int cs_nifree; /* number of free inodes */ 139 int cs_nffree; /* number of free frags */ 140 }; 141 142/* 143 * Super block for a file system. 144 */ 145#define FS_MAGIC 0x011954 146struct fs 147 { 148 int xxx1; /* struct fs *fs_link; */ 149 int xxx2; /* struct fs *fs_rlink; */ 150 mach_daddr_t fs_sblkno; /* addr of super-block in filesys */ 151 mach_daddr_t fs_cblkno; /* offset of cyl-block in filesys */ 152 mach_daddr_t fs_iblkno; /* offset of inode-blocks in filesys */ 153 mach_daddr_t fs_dblkno; /* offset of first data after cg */ 154 int fs_cgoffset; /* cylinder group offset in cylinder */ 155 int fs_cgmask; /* used to calc mod fs_ntrak */ 156 mach_time_t fs_time; /* last time written */ 157 int fs_size; /* number of blocks in fs */ 158 int fs_dsize; /* number of data blocks in fs */ 159 int fs_ncg; /* number of cylinder groups */ 160 int fs_bsize; /* size of basic blocks in fs */ 161 int fs_fsize; /* size of frag blocks in fs */ 162 int fs_frag; /* number of frags in a block in fs */ 163/* these are configuration parameters */ 164 int fs_minfree; /* minimum percentage of free blocks */ 165 int fs_rotdelay; /* num of ms for optimal next block */ 166 int fs_rps; /* disk revolutions per second */ 167/* these fields can be computed from the others */ 168 int fs_bmask; /* ``blkoff'' calc of blk offsets */ 169 int fs_fmask; /* ``fragoff'' calc of frag offsets */ 170 int fs_bshift; /* ``lblkno'' calc of logical blkno */ 171 int fs_fshift; /* ``numfrags'' calc number of frags */ 172/* these are configuration parameters */ 173 int fs_maxcontig; /* max number of contiguous blks */ 174 int fs_maxbpg; /* max number of blks per cyl group */ 175/* these fields can be computed from the others */ 176 int fs_fragshift; /* block to frag shift */ 177 int fs_fsbtodb; /* fsbtodb and dbtofsb shift constant */ 178 int fs_sbsize; /* actual size of super block */ 179 int fs_csmask; /* csum block offset */ 180 int fs_csshift; /* csum block number */ 181 int fs_nindir; /* value of NINDIR */ 182 int fs_inopb; /* value of INOPB */ 183 int fs_nspf; /* value of NSPF */ 184/* yet another configuration parameter */ 185 int fs_optim; /* optimization preference, see below */ 186/* these fields are derived from the hardware */ 187 int fs_npsect; /* # sectors/track including spares */ 188 int fs_interleave; /* hardware sector interleave */ 189 int fs_trackskew; /* sector 0 skew, per track */ 190 int fs_headswitch; /* head switch time, usec */ 191 int fs_trkseek; /* track-to-track seek, usec */ 192/* sizes determined by number of cylinder groups and their sizes */ 193 mach_daddr_t fs_csaddr; /* blk addr of cyl grp summary area */ 194 int fs_cssize; /* size of cyl grp summary area */ 195 int fs_cgsize; /* cylinder group size */ 196/* these fields are derived from the hardware */ 197 int fs_ntrak; /* tracks per cylinder */ 198 int fs_nsect; /* sectors per track */ 199 int fs_spc; /* sectors per cylinder */ 200/* this comes from the disk driver partitioning */ 201 int fs_ncyl; /* cylinders in file system */ 202/* these fields can be computed from the others */ 203 int fs_cpg; /* cylinders per group */ 204 int fs_ipg; /* inodes per group */ 205 int fs_fpg; /* blocks per group * fs_frag */ 206/* this data must be re-computed after crashes */ 207 struct csum fs_cstotal; /* cylinder summary information */ 208/* these fields are cleared at mount time */ 209 char fs_fmod; /* super block modified flag */ 210 char fs_clean; /* file system is clean flag */ 211 char fs_ronly; /* mounted read-only flag */ 212 char fs_flags; /* currently unused flag */ 213 char fs_fsmnt[MAXMNTLEN]; /* name mounted on */ 214/* these fields retain the current block allocation info */ 215 int fs_cgrotor; /* last cg searched */ 216#if 1 217 int was_fs_csp[MAXCSBUFS]; 218#else 219 struct csum *fs_csp[MAXCSBUFS]; /* list of fs_cs info buffers */ 220#endif 221 int fs_cpc; /* cyl per cycle in postbl */ 222 short fs_opostbl[16][8]; /* old rotation block list head */ 223 long fs_sparecon[50]; /* reserved for future constants */ 224 long fs_contigsumsize; /* size of cluster summary array */ 225 long fs_maxsymlinklen; /* max length of an internal symlink */ 226 long fs_inodefmt; /* format of on-disk inodes */ 227 quad fs_maxfilesize; /* maximum representable file size */ 228 quad fs_qbmask; /* ~fs_bmask - for use with quad size */ 229 quad fs_qfmask; /* ~fs_fmask - for use with quad size */ 230 long fs_state; /* validate fs_clean field */ 231 int fs_postblformat; /* format of positional layout tables */ 232 int fs_nrpos; /* number of rotaional positions */ 233 int fs_postbloff; /* (short) rotation block list head */ 234 int fs_rotbloff; /* (u_char) blocks for each rotation */ 235 int fs_magic; /* magic number */ 236 u_char fs_space[1]; /* list of blocks for each rotation */ 237/* actually longer */ 238 }; 239/* 240 * Preference for optimization. 241 */ 242#define FS_OPTTIME 0 /* minimize allocation time */ 243#define FS_OPTSPACE 1 /* minimize disk fragmentation */ 244 245/* 246 * Rotational layout table format types 247 */ 248#define FS_42POSTBLFMT -1 /* 4.2BSD rotational table format */ 249#define FS_DYNAMICPOSTBLFMT 1 /* dynamic rotational table format */ 250/* 251 * Macros for access to superblock array structures 252 */ 253#define fs_postbl(fs, cylno) \ 254 (((fs)->fs_postblformat == FS_42POSTBLFMT) \ 255 ? ((fs)->fs_opostbl[cylno]) \ 256 : ((short *)((char *)(fs) + (fs)->fs_postbloff) + (cylno) * (fs)->fs_nrpos)) 257#define fs_rotbl(fs) \ 258 (((fs)->fs_postblformat == FS_42POSTBLFMT) \ 259 ? ((fs)->fs_space) \ 260 : ((u_char *)((char *)(fs) + (fs)->fs_rotbloff))) 261 262/* 263 * Convert cylinder group to base address of its global summary info. 264 * 265 * N.B. This macro assumes that sizeof(struct csum) is a power of two. 266 */ 267#define fs_cs(fs, indx) \ 268 fs_csp[(indx) >> (fs)->fs_csshift][(indx) & ~(fs)->fs_csmask] 269 270/* 271 * Cylinder group block for a file system. 272 */ 273#define CG_MAGIC 0x090255 274struct cg 275 { 276 int xxx1; /* struct cg *cg_link; */ 277 int cg_magic; /* magic number */ 278 mach_time_t cg_time; /* time last written */ 279 int cg_cgx; /* we are the cgx'th cylinder group */ 280 short cg_ncyl; /* number of cyl's this cg */ 281 short cg_niblk; /* number of inode blocks this cg */ 282 int cg_ndblk; /* number of data blocks this cg */ 283 struct csum cg_cs; /* cylinder summary information */ 284 int cg_rotor; /* position of last used block */ 285 int cg_frotor; /* position of last used frag */ 286 int cg_irotor; /* position of last used inode */ 287 int cg_frsum[MAXFRAG]; /* counts of available frags */ 288 int cg_btotoff; /* (long) block totals per cylinder */ 289 int cg_boff; /* (short) free block positions */ 290 int cg_iusedoff; /* (char) used inode map */ 291 int cg_freeoff; /* (u_char) free block map */ 292 int cg_nextfreeoff; /* (u_char) next available space */ 293 int cg_sparecon[16]; /* reserved for future use */ 294 u_char cg_space[1]; /* space for cylinder group maps */ 295/* actually longer */ 296 }; 297/* 298 * Macros for access to cylinder group array structures 299 */ 300#define cg_blktot(cgp) \ 301 (((cgp)->cg_magic != CG_MAGIC) \ 302 ? (((struct ocg *)(cgp))->cg_btot) \ 303 : ((int *)((char *)(cgp) + (cgp)->cg_btotoff))) 304#define cg_blks(fs, cgp, cylno) \ 305 (((cgp)->cg_magic != CG_MAGIC) \ 306 ? (((struct ocg *)(cgp))->cg_b[cylno]) \ 307 : ((short *)((char *)(cgp) + (cgp)->cg_boff) + (cylno) * (fs)->fs_nrpos)) 308#define cg_inosused(cgp) \ 309 (((cgp)->cg_magic != CG_MAGIC) \ 310 ? (((struct ocg *)(cgp))->cg_iused) \ 311 : ((char *)((char *)(cgp) + (cgp)->cg_iusedoff))) 312#define cg_blksfree(cgp) \ 313 (((cgp)->cg_magic != CG_MAGIC) \ 314 ? (((struct ocg *)(cgp))->cg_free) \ 315 : ((u_char *)((char *)(cgp) + (cgp)->cg_freeoff))) 316#define cg_chkmagic(cgp) \ 317 ((cgp)->cg_magic == CG_MAGIC || ((struct ocg *)(cgp))->cg_magic == CG_MAGIC) 318 319/* 320 * The following structure is defined 321 * for compatibility with old file systems. 322 */ 323struct ocg 324 { 325 int xxx1; /* struct ocg *cg_link; */ 326 int xxx2; /* struct ocg *cg_rlink; */ 327 mach_time_t cg_time; /* time last written */ 328 int cg_cgx; /* we are the cgx'th cylinder group */ 329 short cg_ncyl; /* number of cyl's this cg */ 330 short cg_niblk; /* number of inode blocks this cg */ 331 int cg_ndblk; /* number of data blocks this cg */ 332 struct csum cg_cs; /* cylinder summary information */ 333 int cg_rotor; /* position of last used block */ 334 int cg_frotor; /* position of last used frag */ 335 int cg_irotor; /* position of last used inode */ 336 int cg_frsum[8]; /* counts of available frags */ 337 int cg_btot[32]; /* block totals per cylinder */ 338 short cg_b[32][8]; /* positions of free blocks */ 339 char cg_iused[256]; /* used inode map */ 340 int cg_magic; /* magic number */ 341 u_char cg_free[1]; /* free block map */ 342/* actually longer */ 343 }; 344 345/* 346 * Turn file system block numbers into disk block addresses. 347 * This maps file system blocks to device size blocks. 348 */ 349#define fsbtodb(fs, b) ((b) << (fs)->fs_fsbtodb) 350#define dbtofsb(fs, b) ((b) >> (fs)->fs_fsbtodb) 351 352/* 353 * Cylinder group macros to locate things in cylinder groups. 354 * They calc file system addresses of cylinder group data structures. 355 */ 356#define cgbase(fs, c) ((mach_daddr_t)((fs)->fs_fpg * (c))) 357#define cgstart(fs, c) \ 358 (cgbase(fs, c) + (fs)->fs_cgoffset * ((c) & ~((fs)->fs_cgmask))) 359#define cgsblock(fs, c) (cgstart(fs, c) + (fs)->fs_sblkno) /* super blk */ 360#define cgtod(fs, c) (cgstart(fs, c) + (fs)->fs_cblkno) /* cg block */ 361#define cgimin(fs, c) (cgstart(fs, c) + (fs)->fs_iblkno) /* inode blk */ 362#define cgdmin(fs, c) (cgstart(fs, c) + (fs)->fs_dblkno) /* 1st data */ 363 364/* 365 * Macros for handling inode numbers: 366 * inode number to file system block offset. 367 * inode number to cylinder group number. 368 * inode number to file system block address. 369 */ 370#define itoo(fs, x) ((x) % INOPB(fs)) 371#define itog(fs, x) ((x) / (fs)->fs_ipg) 372#define itod(fs, x) \ 373 ((mach_daddr_t)(cgimin(fs, itog(fs, x)) + \ 374 (blkstofrags((fs), (((x) % (fs)->fs_ipg) / INOPB(fs)))))) 375 376/* 377 * Give cylinder group number for a file system block. 378 * Give cylinder group block number for a file system block. 379 */ 380#define dtog(fs, d) ((d) / (fs)->fs_fpg) 381#define dtogd(fs, d) ((d) % (fs)->fs_fpg) 382 383/* 384 * Extract the bits for a block from a map. 385 * Compute the cylinder and rotational position of a cyl block addr. 386 */ 387#define blkmap(fs, map, loc) \ 388 (((map)[(loc) / NBBY] >> ((loc) % NBBY)) & (0xff >> (NBBY - (fs)->fs_frag))) 389#define cbtocylno(fs, bno) \ 390 ((bno) * NSPF(fs) / (fs)->fs_spc) 391#define cbtorpos(fs, bno) \ 392 (((bno) * NSPF(fs) % (fs)->fs_spc / (fs)->fs_nsect * (fs)->fs_trackskew + \ 393 (bno) * NSPF(fs) % (fs)->fs_spc % (fs)->fs_nsect * (fs)->fs_interleave) % \ 394 (fs)->fs_nsect * (fs)->fs_nrpos / (fs)->fs_npsect) 395 396/* 397 * The following macros optimize certain frequently calculated 398 * quantities by using shifts and masks in place of divisions 399 * modulos and multiplications. 400 */ 401#define blkoff(fs, loc) /* calculates (loc % fs->fs_bsize) */ \ 402 ((loc) & ~(fs)->fs_bmask) 403#define fragoff(fs, loc) /* calculates (loc % fs->fs_fsize) */ \ 404 ((loc) & ~(fs)->fs_fmask) 405#define lblkno(fs, loc) /* calculates (loc / fs->fs_bsize) */ \ 406 ((loc) >> (fs)->fs_bshift) 407#define numfrags(fs, loc) /* calculates (loc / fs->fs_fsize) */ \ 408 ((loc) >> (fs)->fs_fshift) 409#define blkroundup(fs, size) /* calculates roundup(size, fs->fs_bsize) */ \ 410 (((size) + (fs)->fs_bsize - 1) & (fs)->fs_bmask) 411#define fragroundup(fs, size) /* calculates roundup(size, fs->fs_fsize) */ \ 412 (((size) + (fs)->fs_fsize - 1) & (fs)->fs_fmask) 413#define fragstoblks(fs, frags) /* calculates (frags / fs->fs_frag) */ \ 414 ((frags) >> (fs)->fs_fragshift) 415#define blkstofrags(fs, blks) /* calculates (blks * fs->fs_frag) */ \ 416 ((blks) << (fs)->fs_fragshift) 417#define fragnum(fs, fsb) /* calculates (fsb % fs->fs_frag) */ \ 418 ((fsb) & ((fs)->fs_frag - 1)) 419#define blknum(fs, fsb) /* calculates rounddown(fsb, fs->fs_frag) */ \ 420 ((fsb) &~ ((fs)->fs_frag - 1)) 421 422/* 423 * Determine the number of available frags given a 424 * percentage to hold in reserve 425 */ 426#define freespace(fs, percentreserved) \ 427 (blkstofrags((fs), (fs)->fs_cstotal.cs_nbfree) + \ 428 (fs)->fs_cstotal.cs_nffree - ((fs)->fs_dsize * (percentreserved) / 100)) 429 430/* 431 * Determining the size of a file block in the file system. 432 */ 433#define blksize(fs, ip, lbn) \ 434 (((lbn) >= NDADDR || (ip)->i_size >= ((lbn) + 1) << (fs)->fs_bshift) \ 435 ? (fs)->fs_bsize \ 436 : (fragroundup(fs, blkoff(fs, (ip)->i_size)))) 437#define dblksize(fs, dip, lbn) \ 438 (((lbn) >= NDADDR || (dip)->di_size >= ((lbn) + 1) << (fs)->fs_bshift) \ 439 ? (fs)->fs_bsize \ 440 : (fragroundup(fs, blkoff(fs, (dip)->di_size)))) 441 442/* 443 * Number of disk sectors per block; assumes DEV_BSIZE byte sector size. 444 */ 445#define NSPB(fs) ((fs)->fs_nspf << (fs)->fs_fragshift) 446#define NSPF(fs) ((fs)->fs_nspf) 447 448/* 449 * INOPB is the number of inodes in a secondary storage block. 450 */ 451#define INOPB(fs) ((fs)->fs_inopb) 452#define INOPF(fs) ((fs)->fs_inopb >> (fs)->fs_fragshift) 453 454/* 455 * NINDIR is the number of indirects in a file system block. 456 */ 457#define NINDIR(fs) ((fs)->fs_nindir) 458