1.\" Copyright (c) 1983, 1991, 1993 2.\" The Regents of the University of California. All rights reserved. 3.\" 4.\" Redistribution and use in source and binary forms, with or without 5.\" modification, are permitted provided that the following conditions 6.\" are met: 7.\" 1. Redistributions of source code must retain the above copyright 8.\" notice, this list of conditions and the following disclaimer. 9.\" 2. Redistributions in binary form must reproduce the above copyright 10.\" notice, this list of conditions and the following disclaimer in the 11.\" documentation and/or other materials provided with the distribution. 12.\" 3. All advertising materials mentioning features or use of this software 13.\" must display the following acknowledgement: 14.\" This product includes software developed by the University of 15.\" California, Berkeley and its contributors. 16.\" 4. Neither the name of the University nor the names of its contributors 17.\" may be used to endorse or promote products derived from this software 18.\" without specific prior written permission. 19.\" 20.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 21.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 23.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 24.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 25.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 26.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 27.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 28.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 29.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 30.\" SUCH DAMAGE. 31.\" 32.\" @(#)fs.5 8.2 (Berkeley) 4/19/94 33.\" $FreeBSD$ 34.\" 35.Dd April 19, 1994 36.Dt FS 5 37.Os BSD 4.2 38.Sh NAME 39.Nm fs , 40.Nm inode 41.Nd format of file system volume 42.Sh SYNOPSIS 43.Fd #include <sys/param.h> 44.Fd #include <ufs/ffs/fs.h> 45.Pp 46.Fd #include <sys/types.h> 47.Fd #include <sys/lock.h> 48.Fd #include <ufs/ufs/quota.h> 49.Fd #include <ufs/ufs/inode.h> 50.Sh DESCRIPTION 51The files 52.Aq Pa fs.h 53and 54.Aq Pa inode.h 55declare several structures, defined variables and macros 56which are used to create and manage the underlying format of 57file system objects on random access devices (disks). 58.Pp 59The block size and number of blocks which 60comprise a file system are parameters of the file system. 61Sectors beginning at 62.Dv BBLOCK 63and continuing for 64.Dv BBSIZE 65are used 66for a disklabel and for some hardware primary 67and secondary bootstrapping programs. 68.Pp 69The actual file system begins at sector 70.Dv SBLOCK 71with the 72.Em super-block 73that is of size 74.Dv SBSIZE . 75The following structure describes the super-block and is 76from the file 77.Aq Pa ufs/ffs/fs.h : 78.Bd -literal 79/* 80 * Super block for an FFS file system. 81 */ 82struct fs { 83 int32_t fs_firstfield; /* historic file system linked list, */ 84 int32_t fs_unused_1; /* used for incore super blocks */ 85 ufs_daddr_t fs_sblkno; /* addr of super-block in filesys */ 86 ufs_daddr_t fs_cblkno; /* offset of cyl-block in filesys */ 87 ufs_daddr_t fs_iblkno; /* offset of inode-blocks in filesys */ 88 ufs_daddr_t fs_dblkno; /* offset of first data after cg */ 89 int32_t fs_cgoffset; /* cylinder group offset in cylinder */ 90 int32_t fs_cgmask; /* used to calc mod fs_ntrak */ 91 time_t fs_time; /* last time written */ 92 int32_t fs_size; /* number of blocks in fs */ 93 int32_t fs_dsize; /* number of data blocks in fs */ 94 int32_t fs_ncg; /* number of cylinder groups */ 95 int32_t fs_bsize; /* size of basic blocks in fs */ 96 int32_t fs_fsize; /* size of frag blocks in fs */ 97 int32_t fs_frag; /* number of frags in a block in fs */ 98/* these are configuration parameters */ 99 int32_t fs_minfree; /* minimum percentage of free blocks */ 100 int32_t fs_rotdelay; /* num of ms for optimal next block */ 101 int32_t fs_rps; /* disk revolutions per second */ 102/* these fields can be computed from the others */ 103 int32_t fs_bmask; /* ``blkoff'' calc of blk offsets */ 104 int32_t fs_fmask; /* ``fragoff'' calc of frag offsets */ 105 int32_t fs_bshift; /* ``lblkno'' calc of logical blkno */ 106 int32_t fs_fshift; /* ``numfrags'' calc number of frags */ 107/* these are configuration parameters */ 108 int32_t fs_maxcontig; /* max number of contiguous blks */ 109 int32_t fs_maxbpg; /* max number of blks per cyl group */ 110/* these fields can be computed from the others */ 111 int32_t fs_fragshift; /* block to frag shift */ 112 int32_t fs_fsbtodb; /* fsbtodb and dbtofsb shift constant */ 113 int32_t fs_sbsize; /* actual size of super block */ 114 int32_t fs_csmask; /* csum block offset */ 115 int32_t fs_csshift; /* csum block number */ 116 int32_t fs_nindir; /* value of NINDIR */ 117 int32_t fs_inopb; /* value of INOPB */ 118 int32_t fs_nspf; /* value of NSPF */ 119/* yet another configuration parameter */ 120 int32_t fs_optim; /* optimization preference, see below */ 121/* these fields are derived from the hardware */ 122 int32_t fs_npsect; /* # sectors/track including spares */ 123 int32_t fs_interleave; /* hardware sector interleave */ 124 int32_t fs_trackskew; /* sector 0 skew, per track */ 125/* fs_id takes the space of the unused fs_headswitch and fs_trkseek fields */ 126 int32_t fs_id[2]; /* unique filesystem id*/ 127/* sizes determined by number of cylinder groups and their sizes */ 128 ufs_daddr_t fs_csaddr; /* blk addr of cyl grp summary area */ 129 int32_t fs_cssize; /* size of cyl grp summary area */ 130 int32_t fs_cgsize; /* cylinder group size */ 131/* these fields are derived from the hardware */ 132 int32_t fs_ntrak; /* tracks per cylinder */ 133 int32_t fs_nsect; /* sectors per track */ 134 int32_t fs_spc; /* sectors per cylinder */ 135/* this comes from the disk driver partitioning */ 136 int32_t fs_ncyl; /* cylinders in file system */ 137/* these fields can be computed from the others */ 138 int32_t fs_cpg; /* cylinders per group */ 139 int32_t fs_ipg; /* inodes per group */ 140 int32_t fs_fpg; /* blocks per group * fs_frag */ 141/* this data must be re-computed after crashes */ 142 struct csum fs_cstotal;/* cylinder summary information */ 143/* these fields are cleared at mount time */ 144 int8_t fs_fmod; /* super block modified flag */ 145 int8_t fs_clean; /* file system is clean flag */ 146 int8_t fs_ronly; /* mounted read-only flag */ 147 int8_t fs_flags; /* currently unused flag */ 148 u_char fs_fsmnt[MAXMNTLEN]; /* name mounted on */ 149/* these fields retain the current block allocation info */ 150 int32_t fs_cgrotor; /* last cg searched */ 151 struct csum *fs_csp[MAXCSBUFS];/* list of fs_cs info buffers */ 152 int32_t *fs_maxcluster;/* max cluster in each cyl group */ 153 int32_t fs_cpc; /* cyl per cycle in postbl */ 154 int16_t fs_opostbl[16][8]; /* old rotation block list head */ 155 int32_t fs_sparecon[50]; /* reserved for future constants */ 156 int32_t fs_contigsumsize; /* size of cluster summary array */ 157 int32_t fs_maxsymlinklen;/* max length of an internal symlink */ 158 int32_t fs_inodefmt; /* format of on-disk inodes */ 159 u_int64_t fs_maxfilesize;/* maximum representable file size */ 160 int64_t fs_qbmask; /* ~fs_bmask for use with 64-bit size */ 161 int64_t fs_qfmask; /* ~fs_fmask for use with 64-bit size */ 162 int32_t fs_state; /* validate fs_clean field */ 163 int32_t fs_postblformat;/* format of positional layout tables */ 164 int32_t fs_nrpos; /* number of rotational positions */ 165 int32_t fs_postbloff; /* (u_int16) rotation block list head */ 166 int32_t fs_rotbloff; /* (u_int8) blocks for each rotation */ 167 int32_t fs_magic; /* magic number */ 168 u_int8_t fs_space[1]; /* list of blocks for each rotation */ 169/* actually longer */ 170}; 171 172/* 173 * Filesystem identification 174 */ 175#define FS_MAGIC 0x011954 /* the fast filesystem magic number */ 176#define FS_OKAY 0x7c269d38 /* superblock checksum */ 177#define FS_42INODEFMT -1 /* 4.2BSD inode format */ 178#define FS_44INODEFMT 2 /* 4.4BSD inode format */ 179/* 180 * Preference for optimization. 181 */ 182#define FS_OPTTIME 0 /* minimize allocation time */ 183#define FS_OPTSPACE 1 /* minimize disk fragmentation */ 184 185/* 186 * Rotational layout table format types 187 */ 188#define FS_42POSTBLFMT -1 /* 4.2BSD rotational table format */ 189#define FS_DYNAMICPOSTBLFMT 1 /* dynamic rotational table format */ 190.Ed 191.Pp 192Each disk drive contains some number of file systems. 193A file system consists of a number of cylinder groups. 194Each cylinder group has inodes and data. 195.Pp 196A file system is described by its super-block, which in turn 197describes the cylinder groups. The super-block is critical 198data and is replicated in each cylinder group to protect against 199catastrophic loss. This is done at file system creation 200time and the critical 201super-block data does not change, so the copies need not be 202referenced further unless disaster strikes. 203.Pp 204Addresses stored in inodes are capable of addressing fragments 205of `blocks'. File system blocks of at most size 206.Dv MAXBSIZE 207can 208be optionally broken into 2, 4, or 8 pieces, each of which is 209addressable; these pieces may be 210.Dv DEV_BSIZE , 211or some multiple of 212a 213.Dv DEV_BSIZE 214unit. 215.Pp 216Large files consist of exclusively large data blocks. To avoid 217undue wasted disk space, the last data block of a small file is 218allocated as only as many fragments of a large block as are 219necessary. The file system format retains only a single pointer 220to such a fragment, which is a piece of a single large block that 221has been divided. The size of such a fragment is determinable from 222information in the inode, using the 223.Fn blksize fs ip lbn 224macro. 225.Pp 226The file system records space availability at the fragment level; 227to determine block availability, aligned fragments are examined. 228.Pp 229The root inode is the root of the file system. 230Inode 0 can't be used for normal purposes and 231historically bad blocks were linked to inode 1, 232thus the root inode is 2 (inode 1 is no longer used for 233this purpose, however numerous dump tapes make this 234assumption, so we are stuck with it). 235.Pp 236The 237.Fa fs_minfree 238element gives the minimum acceptable percentage of file system 239blocks that may be free. 240If the freelist drops below this level 241only the super-user may continue to allocate blocks. 242The 243.Fa fs_minfree 244element 245may be set to 0 if no reserve of free blocks is deemed necessary, 246however severe performance degradations will be observed if the 247file system is run at greater than 90% full; thus the default 248value of 249.Fa fs_minfree 250is 10%. 251.Pp 252Empirically the best trade-off between block fragmentation and 253overall disk utilization at a loading of 90% comes with a 254fragmentation of 8, thus the default fragment size is an eighth 255of the block size. 256.Pp 257The element 258.Fa fs_optim 259specifies whether the file system should try to minimize the time spent 260allocating blocks, or if it should attempt to minimize the space 261fragmentation on the disk. 262If the value of fs_minfree (see above) is less than 10%, 263then the file system defaults to optimizing for space to avoid 264running out of full sized blocks. 265If the value of minfree is greater than or equal to 10%, 266fragmentation is unlikely to be problematical, and 267the file system defaults to optimizing for time. 268.Pp 269.Em Cylinder group related limits : 270Each cylinder keeps track of the availability of blocks at different 271rotational positions, so that sequential blocks can be laid out 272with minimum rotational latency. 273With the default of 8 distinguished 274rotational positions, the resolution of the 275summary information is 2ms for a typical 3600 rpm drive. 276.Pp 277The element 278.Fa fs_rotdelay 279gives the minimum number of milliseconds to initiate 280another disk transfer on the same cylinder. 281It is used in determining the rotationally optimal 282layout for disk blocks within a file; 283the default value for 284.Fa fs_rotdelay 285is 2ms. 286.Pp 287Each file system has a statically allocated number of inodes. 288An inode is allocated for each 289.Dv NBPI 290bytes of disk space. 291The inode allocation strategy is extremely conservative. 292.Pp 293.Dv MINBSIZE 294is the smallest allowable block size. 295With a 296.Dv MINBSIZE 297of 4096 298it is possible to create files of size 2992^32 with only two levels of indirection. 300.Dv MINBSIZE 301must be big enough to hold a cylinder group block, 302thus changes to 303.Pq Fa struct cg 304must keep its size within 305.Dv MINBSIZE . 306Note that super-blocks are never more than size 307.Dv SBSIZE . 308.Pp 309The path name on which the file system is mounted is maintained in 310.Fa fs_fsmnt . 311.Dv MAXMNTLEN 312defines the amount of space allocated in 313the super-block for this name. 314The limit on the amount of summary information per file system 315is defined by 316.Dv MAXCSBUFS. 317For a 4096 byte block size, it is currently parameterized for a 318maximum of two million cylinders. 319.Pp 320Per cylinder group information is summarized in blocks allocated 321from the first cylinder group's data blocks. 322These blocks are read in from 323.Fa fs_csaddr 324(size 325.Fa fs_cssize ) 326in addition to the super-block. 327.Pp 328.Sy N.B.: 329.Fn sizeof "struct csum" 330must be a power of two in order for 331the 332.Fn fs_cs 333macro to work. 334.Pp 335The 336.Em "Super-block for a file system" : 337The size of the rotational layout tables 338is limited by the fact that the super-block is of size 339.Dv SBSIZE . 340The size of these tables is 341.Em inversely 342proportional to the block 343size of the file system. 344The size of the tables is 345increased when sector sizes are not powers of two, 346as this increases the number of cylinders 347included before the rotational pattern repeats 348.Pq Fa fs_cpc . 349The size of the rotational layout 350tables is derived from the number of bytes remaining in 351.Pq Fa struct fs . 352.Pp 353The number of blocks of data per cylinder group 354is limited because cylinder groups are at most one block. 355The inode and free block tables 356must fit into a single block after deducting space for 357the cylinder group structure 358.Pq Fa struct cg . 359.Pp 360The 361.Em Inode : 362The inode is the focus of all file activity in the 363.Tn UNIX 364file system. 365There is a unique inode allocated 366for each active file, 367each current directory, each mounted-on file, 368text file, and the root. 369An inode is `named' by its device/i-number pair. 370For further information, see the include file 371.Aq Pa ufs/ufs/inode.h . 372.Sh HISTORY 373A super-block structure named filsys appeared in 374.At v6 . 375The file system described in this manual appeared 376in 377.Bx 4.2 . 378