1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242
|
\documentstyle[11pt,warn]{article}
% Introduction to the Minix file system:
% (c) 1993 Scott D. Heavner (sdh@po.cwru.edu)
%
% This is a brief introduction to the Minix file system compiled by
% Scott D. Heavner (sdh@po.cwru.edu). Please direct any comments or
% corrections back to the author. You are free to distribute this
% document to anyone in any form as long as this notice remains intact.
%
% One day it might even make it to the LDP.
%
\parindent 0pt
\parskip 5pt
\topmargin 0pt
\textheight 9in
\textwidth 6.5in
\oddsidemargin 0pt
\evensidemargin 0pt
\pagestyle{myheadings}
\markboth{Introduction to the Minix File System}{Introduction to the
Minix File System}
\begin{document}
\section*{Introduction to the Minix File System}
This is a brief introduction to the Minix file system compiled by
Scott D. Heavner \linebreak[4]
({\bf sdh@po.cwru.edu}). Please direct any comments or
corrections back to the author. You are free to distribute this
document to anyone in any form as long as this notice remains intact.
\section{Definitions}
(All are probably relevant to this document only.)
{\bf inode:} Stores all the information about a file except its
name.
{\bf block:} A unit of size which is determined by the medium or
programmer. For example, most devices use 1024 byte (1k) blocks,
including hard disks and floppy disks (stored in {\tt BLOCK\_SIZE}).
{\bf zone:} probably used interchangeably with block in this document.
A zone is the part of the disk where the files data exists.
{\bf super block:} the first block on a disk which contains
information about the type and size of the file system.
\section{Physical Layout}
\small
\begin{tabular}{|r||c|p{3.5in}|}
\hline
\multicolumn{3}{|c|}{Minix FS Layout}
\\ \hline\hline
\multicolumn{1}{|c||}{}
& \multicolumn{1}{c|}{Size (Blocks)}
& \multicolumn{1}{c|}{Description}
\\ \hline
Blank Block & 1 & Reserved for partition boot code. \\ \hline
Super Block & 1 & Info about the file system. \\ \hline
Inode Map & \#Inodes/BLOCK\_SIZE &
Keeps track of used/unused inodes. \\ \hline
Zone Map & \#Data Zones/BLOCK\_SIZE &
Keeps track of used/unused zones. \\ \hline
Inode Table & ($(32 or 16) \times {\rm \#Inodes}$)/{BLOCK\_SIZE} &
Store info about files/devices. \\ \hline
Data Zones & Big & File/Directory contents. \\ \hline
\end{tabular}
\normalsize
\section{Inodes}
From $<$linux/minix.h$>$ (Isn't source code wonderful?)
\begin{verbatim}
struct minix_inode {
unsigned short i_mode;
unsigned short i_uid;
unsigned long i_size;
unsigned long i_time;
unsigned char i_gid;
unsigned char i_nlinks;
unsigned short i_zone[9];
};
\end{verbatim}
\begin{tabular}{|r||c|c|c|c|c|c|c|c||c|c|c|c|c|c|c|c|}
\hline
\multicolumn{17}{|c|}{Minix Inode Entry}
\\ \hline\hline
& 00 & 01 & 02 & 03 & 04 & 05 & 06 & 07 & 08 &
09 & 0A & 0B & 0C & 0D & 0E & 0F \\ \hline
0x00 & \multicolumn{2}{c|}{MODE} & \multicolumn{2}{c|}{UID} &
\multicolumn{4}{c||}{SIZE} & \multicolumn{4}{c|}{TIME} &
GID & LINKS & \multicolumn{2}{c|}{ZONE 0} \\ \hline
0x10 & \multicolumn{2}{c|}{ZONE 1} & \multicolumn{2}{c|}{ZONE 2} &
\multicolumn{2}{c|}{ZONE 3} & \multicolumn{2}{c||}{ZONE 4} &
\multicolumn{2}{c|}{ZONE 5} & \multicolumn{2}{c|}{ZONE 6} &
\multicolumn{2}{c|}{ZONE 7} & \multicolumn{2}{c|}{ZONE 8} \\ \hline
\end{tabular}
\vskip1eM
The inode contains all the important information about a file, except
its name. It contains the file permissions, file type, user, group,
size, modification time, number of links, and the location and order
of all the blocks in the file. All values are stored in low byte --
high byte order. The maximum number of links is 250 ({\tt
MINIX\_LINK\_MAX}). Notice that the group number is limited to one byte
whereas $<$gnu/types.h$>$ defines it as short.
Now, for the zones. The first seven zone pointers (0-6) are point to
file data. They are two byte numbers which point to a BLOCK on the
disk which contains the file's data. The eighth is a pointer to an
indirect block; this block continues the tradition of the first seven
zone pointers, and contains 512 ({\tt BLOCK\_SIZE/2}) zone pointers.
The ninth zone pointer in the inode points to a double indirect
block. The double indirect block contains 512 pointers to more
indirect blocks, each of which points to 512 data zones. Each
indirect block adds 1k to the file. It's no big deal, but it's nice
that small files (under 8k) don't have this overhead. Technically you
can make 262M files ($7 + 512 + 512 \times 512$ k, see also
s\_max\_size in the superblock information), but
since the Minix fs uses unsigned shorts for block pointers,
it is limited to 64M partitions.
To determine the meaning of the mode entry, consult
$<$linux/types.h$>$ which is included via $<$sys/types.h$>$. Below is
a list of octal numbers which can be extracted from the mode entry. This
information can also be found in the stat(2) and chmod(2) man pages.
\begin{verbatim}
#define S_IFSOCK 0140000 /* Socket */
#define S_IFLNK 0120000 /* Symbolic Linux */
#define S_IFREG 0100000 /* Regular file */
#define S_IFBLK 0060000 /* Block device */
#define S_IFDIR 0040000 /* Directory */
#define S_IFCHR 0020000 /* Character device */
#define S_IFIFO 0010000 /* FIFO/Named pipe */
#define S_ISUID 0004000 /* Set user id upon execution */
#define S_ISGID 0002000 /* Set group id upon execution */
#define S_ISVTX 0001000 /* Sticky bit */
#define S_IRUSR 00400 /* User readable */
#define S_IWUSR 00200 /* User writable */
#define S_IXUSR 00100 /* User executable */
#define S_IRGRP 00040 /* Group readable */
#define S_IWGRP 00020 /* Group writable */
#define S_IXGRP 00010 /* Group executable */
#define S_IROTH 00004 /* World/Other readable */
#define S_IWOTH 00002 /* World/Other writable */
#define S_IXOTH 00001 /* World/Other executable */
\end{verbatim}
\section{Directories}
From $<$linux/minix.h$>$
\begin{verbatim}
struct minix_dir_entry {
unsigned short inode;
char name[0];
};
\end{verbatim}
\begin{center}
\begin{tabular}{|r||c|c|c|}
\hline
\multicolumn{4}{|c|}{Minix Directory Entry}
\\ \hline\hline
& 00 & 01 & 02 -- 1F \\ \hline
0x00 & \multicolumn{2}{c|}{INODE NUMBER} & char[30/14] FILENAME \\ \hline
\end{tabular}
\end{center}
The directory entry associates a filename with an inode. The first two
bytes of a directory entry indicate the inode, the remainder is the
filename. On the Minix FS, the root directory is inode 1
({\tt MINIX\_ROOT\_INO}). You can fit 32 directory entries on a 1k block
which can exist anywhere in the data zone. Hard links are made when two
entries point to the same node. Soft links are stored in the data zone,
indexed by an inode entry. A hard links uses only a directory entry,
whereas a soft link uses a directory entry, an inode, and at least one
data zone.
\section{Super Block}
Again, from $<$linux/minix.h$>$:
\begin{verbatim}
struct minix_super_block {
unsigned short s_ninodes; /* Number of inodes */
unsigned short s_nzones; /* Number of data zones */
unsigned short s_imap_blocks; /* Space used by inode map (blocks) */
unsigned short s_zmap_blocks; /* Space used by zone map (blocks) */
unsigned short s_firstdatazone; /* First zone with ``file'' data */
unsigned short s_log_zone_size; /* Size of a data zone =
(1024 << s_log_zone_size) */
unsigned long s_max_size; /* Maximum file size (bytes) */
unsigned short s_magic; /* Minix 14/30 ID number */
unsigned short s_state; /* Mount state, was it
cleanly unmounted */
};
\end{verbatim}
\scriptsize
\begin{tabular}{|r||c|c|c|c|c|c|c|c||c|c|c|c|c|c|c|c|}
\hline
\multicolumn{17}{|c|}{Minix Inode Entry}
\\ \hline\hline
& 00 & 01 & 02 & 03 & 04 & 05 & 06 & 07 & 08 &
09 & 0A & 0B & 0C & 0D & 0E & 0F \\ \hline
0x00 & \multicolumn{2}{c|}{No. inodes} & \multicolumn{2}{c|}{No. zones} &
\multicolumn{2}{c|}{Imap blocks} & \multicolumn{2}{c||}{ZMap blocks} &
\multicolumn{2}{c|}{First data zone} &
\multicolumn{2}{c|}{Log zone size} &
\multicolumn{4}{c|}{Max. file size} \\ \hline
0x10 & \multicolumn{2}{c|}{Minix magic} & \multicolumn{2}{c|}{Mount state} &
\multicolumn{4}{c||}{} & \multicolumn{8}{c|}{} \\ \hline
\end{tabular}
\normalsize
\vskip1eM
The superblock for the minix fs is offset 1024 bytes from the start of
the disk, this is to leave room for things like LILO and other boot code.
The superblock basically details the size of the file system. It contains
the number of inodes, numner of data zones, space used by the inode and zone
map, the blocksize of the filesystem and a two character ID number indicating
that it is indeed a Minix file system.
\end{document}
% Local Variables:
% TeX-parse-self: nil
% TeX-auto-save: nil
% End:
|