1UFS(5) 386BSD Programmer's Manual UFS(5) 2 3NNAAMMEE 4 uuffss, uuffss__iinnooddee - format of UFS file system volume 5 6SSYYNNOOPPSSIISS 7 $$((NNOONNSSTTDDIINNCC)) ++== $$((IINNCCFFSS)) 8 ##iinncclluuddee <<ssyyss//ttyyppeess..hh>> 9 ##iinncclluuddee ""uuffss..hh"" 10 ##iinncclluuddee ""uuffss__iinnooddee..hh"" 11 12DDEESSCCRRIIPPTTIIOONN 13 The files <_u_f_s._h> and <_u_f_s__i_n_o_d_e._h> declare several structures, defined 14 variables and macros which are used to create and manage the underlying 15 format of file system objects on random access devices (disks). 16 17 The block size and number of blocks which comprise a file system are 18 parameters of the file system. Sectors beginning at BBLOCK and 19 continuing for BBSIZE are used for a disklabel and for some hardware 20 primary and secondary bootstrapping programs. 21 22 The actual file system begins at sector SBLOCK with the _s_u_p_e_r-_b_l_o_c_k that 23 is of size SBSIZE. The following structure described the super-block and 24 is from the file <_u_f_s._h>: 25 26 #define FS_MAGIC 0x011954 27 struct fs { 28 struct fs *fs_link; /* linked list of file systems */ 29 struct fs *fs_rlink; /* used for incore super blocks */ 30 daddr_t fs_sblkno; /* addr of super-block in filesys */ 31 daddr_t fs_cblkno; /* offset of cyl-block in filesys */ 32 daddr_t fs_iblkno; /* offset of inode-blocks in filesys */ 33 daddr_t fs_dblkno; /* offset of first data after cg */ 34 long fs_cgoffset; /* cylinder group offset in cylinder */ 35 long fs_cgmask; /* used to calc mod fs_ntrak */ 36 time_t fs_time; /* last time written */ 37 long fs_size; /* number of blocks in fs */ 38 long fs_dsize; /* number of data blocks in fs */ 39 long fs_ncg; /* number of cylinder groups */ 40 long fs_bsize; /* size of basic blocks in fs */ 41 long fs_fsize; /* size of frag blocks in fs */ 42 long fs_frag; /* number of frags in a block in fs */ 43 /* these are configuration parameters */ 44 long fs_minfree; /* minimum percentage of free blocks */ 45 long fs_rotdelay; /* num of ms for optimal next block */ 46 long fs_rps; /* disk revolutions per second */ 47 /* these fields can be computed from the others */ 48 long fs_bmask; /* ``blkoff'' calc of blk offsets */ 49 long fs_fmask; /* ``fragoff'' calc of frag offsets */ 50 long fs_bshift; /* ``lblkno'' calc of logical blkno */ 51 long fs_fshift; /* ``numfrags'' calc number of frags */ 52 /* these are configuration parameters */ 53 long fs_maxcontig; /* max number of contiguous blks */ 54 long fs_maxbpg; /* max number of blks per cyl group */ 55 /* these fields can be computed from the others */ 56 long fs_fragshift; /* block to frag shift */ 57 long fs_fsbtodb; /* fsbtodb and dbtofsb shift constant */ 58 long fs_sbsize; /* actual size of super block */ 59 long fs_csmask; /* csum block offset */ 60 long fs_csshift; /* csum block number */ 61 long fs_nindir; /* value of NINDIR */ 62 long fs_inopb; /* value of INOPB */ 63 long fs_nspf; /* value of NSPF */ 64 /* yet another configuration parameter */ 65 long fs_optim; /* optimization preference, see below */ 66 /* these fields are derived from the hardware */ 67 long fs_npsect; /* # sectors/track including spares */ 68 long fs_interleave; /* hardware sector interleave */ 69 long fs_trackskew; /* sector 0 skew, per track */ 70 long fs_headswitch; /* head switch time, usec */ 71 long fs_trkseek; /* track-to-track seek, usec */ 72 /* sizes determined by number of cylinder groups and their sizes */ 73 daddr_t fs_csaddr; /* blk addr of cyl grp summary area */ 74 long fs_cssize; /* size of cyl grp summary area */ 75 long fs_cgsize; /* cylinder group size */ 76 /* these fields are derived from the hardware */ 77 long fs_ntrak; /* tracks per cylinder */ 78 long fs_nsect; /* sectors per track */ 79 long fs_spc; /* sectors per cylinder */ 80 /* this comes from the disk driver partitioning */ 81 long fs_ncyl; /* cylinders in file system */ 82 /* these fields can be computed from the others */ 83 long fs_cpg; /* cylinders per group */ 84 long fs_ipg; /* inodes per group */ 85 long fs_fpg; /* blocks per group * fs_frag */ 86 /* this data must be re-computed after crashes */ 87 struct csum fs_cstotal; /* cylinder summary information */ 88 /* these fields are cleared at mount time */ 89 char fs_fmod; /* super block modified flag */ 90 char fs_clean; /* file system is clean flag */ 91 char fs_ronly; /* mounted read-only flag */ 92 char fs_flags; /* currently unused flag */ 93 char fs_fsmnt[MAXMNTLEN]; /* name mounted on */ 94 /* these fields retain the current block allocation info */ 95 long fs_cgrotor; /* last cg searched */ 96 struct csum *fs_csp[MAXCSBUFS]; /* list of fs_cs info buffers */ 97 long fs_cpc; /* cyl per cycle in postbl */ 98 short fs_opostbl[16][8]; /* old rotation block list head */ 99 long fs_sparecon[56]; /* reserved for future constants */ 100 quad fs_qbmask; /* ~fs_bmask - for use with quad size */ 101 quad fs_qfmask; /* ~fs_fmask - for use with quad size */ 102 long fs_postblformat; /* format of positional layout tables */ 103 long fs_nrpos; /* number of rotaional positions */ 104 long fs_postbloff; /* (short) rotation block list head */ 105 long fs_rotbloff; /* (u_char) blocks for each rotation */ 106 long fs_magic; /* magic number */ 107 u_char fs_space[1]; /* list of blocks for each rotation */ 108 /* actually longer */ 109 }; 110 111 Each disk drive contains some number of file systems. A file system 112 consists of a number of cylinder groups. Each cylinder group has inodes 113 and data. 114 115 A file system is described by its super-block, which in turn describes 116 the cylinder groups. The super-block is critical data and is replicated 117 in each cylinder group to protect against catastrophic loss. This is 118 done at file system creation time and the critical super-block data does 119 not change, so the copies need not be referenced further unless disaster 120 strikes. 121 122 Addresses stored in inodes are capable of addressing fragments of 123 `blocks'. File system blocks of at most size MAXBSIZE can be optionally 124 broken into 2, 4, or 8 pieces, each of which is addressable; these pieces 125 may be DEV_BSIZE, or some multiple of a DEV_BSIZE unit. 126 127 Large files consist of exclusively large data blocks. To avoid undue 128 wasted disk space, the last data block of a small file is allocated as 129 only as many fragments of a large block as are necessary. The file 130 system format retains only a single pointer to such a fragment, which is 131 a piece of a single large block that has been divided. The size of such 132 a fragment is determinable from information in the inode, using the 133 bbllkkssiizzee(_f_s, _i_p, _l_b_n) macro. 134 135 The file system records space availability at the fragment level; to 136 determine block availability, aligned fragments are examined. 137 138 The root inode is the root of the file system. Inode 0 can't be used for 139 normal purposes and historically bad blocks were linked to inode 1, thus 140 the root inode is 2 (inode 1 is no longer used for this purpose, however 141 numerous dump tapes make this assumption, so we are stuck with it). 142 143 The _f_s__m_i_n_f_r_e_e element gives the minimum acceptable percentage of file 144 system blocks that may be free. If the freelist drops below this level 145 only the super-user may continue to allocate blocks. The _f_s__m_i_n_f_r_e_e 146 element may be set to 0 if no reserve of free blocks is deemed necessary, 147 however severe performance degradations will be observed if the file 148 system is run at greater than 90% full; thus the default value of 149 _f_s__m_i_n_f_r_e_e is 10%. 150 151 Empirically the best trade-off between block fragmentation and overall 152 disk utilization at a loading of 90% comes with a fragmentation of 8, 153 thus the default fragment size is an eighth of the block size. 154 155 The element _f_s__o_p_t_i_m specifies whether the file system should try to 156 minimize the time spent allocating blocks, or if it should attempt to 157 minimize the space fragmentation on the disk. If the value of fs_minfree 158 (see above) is less than 10%, then the file system defaults to optimizing 159 for space to avoid running out of full sized blocks. If the value of 160 minfree is greater than or equal to 10%, fragmentation is unlikely to be 161 problematical, and the file system defaults to optimizing for time. 162 163 _C_y_l_i_n_d_e_r _g_r_o_u_p _r_e_l_a_t_e_d _l_i_m_i_t_s: Each cylinder keeps track of the 164 availability of blocks at different rotational positions, so that 165 sequential blocks can be laid out with minimum rotational latency. With 166 the default of 8 distinguished rotational positions, the resolution of 167 the summary information is 2ms for a typical 3600 rpm drive. 168 169 The element _f_s__r_o_t_d_e_l_a_y gives the minimum number of milliseconds to 170 initiate another disk transfer on the same cylinder. It is used in 171 determining the rotationally optimal layout for disk blocks within a 172 file; the default value for _f_s__r_o_t_d_e_l_a_y is 2ms. 173 174 Each file system has a statically allocated number of inodes. An inode 175 is allocated for each NBPI bytes of disk space. The inode allocation 176 strategy is extremely conservative. 177 178 MINBSIZE is the smallest allowable block size. With a MINBSIZE of 4096 179 it is possible to create files of size 2^32 with only two levels of 180 indirection. MINBSIZE must be big enough to hold a cylinder group block, 181 thus changes to (_s_t_r_u_c_t _c_g) must keep its size within MINBSIZE. Note that 182 super-blocks are never more than size SBSIZE. 183 184 The path name on which the file system is mounted is maintained in 185 _f_s__f_s_m_n_t. MAXMNTLEN defines the amount of space allocated in the super- 186 block for this name. The limit on the amount of summary information per 187 file system is defined by MAXCSBUFS. For a 4096 byte block size, it is 188 currently parameterized for a maximum of two million cylinders. 189 190 Per cylinder group information is summarized in blocks allocated from the 191 first cylinder group's data blocks. These blocks are read in from 192 _f_s__c_s_a_d_d_r (size _f_s__c_s_s_i_z_e) in addition to the super-block. 193 194 NN..BB..:: sizeof(_s_t_r_u_c_t _c_s_u_m) must be a power of two in order for the 195 ffss__ccss() macro to work. 196 197 198 The _S_u_p_e_r-_b_l_o_c_k _f_o_r _a _f_i_l_e _s_y_s_t_e_m: The size of the rotational layout 199 tables is limited by the fact that the super-block is of size SBSIZE. The 200 size of these tables is _i_n_v_e_r_s_e_l_y proportional to the block size of the 201 file system. The size of the tables is increased when sector sizes are 202 not powers of two, as this increases the number of cylinders included 203 before the rotational pattern repeats (_f_s__c_p_c). The size of the 204 rotational layout tables is derived from the number of bytes remaining in 205 (_s_t_r_u_c_t _f_s). 206 207 The number of blocks of data per cylinder group is limited because 208 cylinder groups are at most one block. The inode and free block tables 209 must fit into a single block after deducting space for the cylinder group 210 structure (_s_t_r_u_c_t _c_g). 211 212 The _I_n_o_d_e: The inode is the focus of all file activity in the UNIX file 213 system. There is a unique inode allocated for each active file, each 214 current directory, each mounted-on file, text file, and the root. An 215 inode is `named' by its device/i-number pair. For further information, 216 see the include file <_u_f_s__i_n_o_d_e._h>. 217 218HHIISSTTOORRYY 219 A super-block structure named filsys appeared in Version 6 AT&T UNIX. 220 The file system described in this manual appeared in 4.2BSD. 221 2224.2 Berkeley Distribution April 28, 1994 4 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265