The "ls"-like utility whatfor is designed to help understand the purpose of all those files.
The Filesystem Hierarchy Standard (144 kB) formalizes the confusion, but is helpful.
"owner" of a file is referred to as "user" in chmod u=...,o=..., where the "o=" means "other users", not "owner".
The word "root" has different meanings in these two phrases. Similarly changing root filesystems with "chroot" is not related to changing "root user" with the "su root" command.
The first process started in the Linux kernel is start_kernel() in linux/init/main.c. As Process 0, start_kernel() does some kernel initialization and then spawns (forks) kernel routine init(), Process 1, which does some additional initialization and then hands control over to the program /sbin/init. /sbin/init carries on as Process 1, organizing the system administration daemons (disk caching, printing, logging,...) and establishing a family of login sessions. init's choreography of the system is directed by the file /etc/inittab and the scripts it invokes. Thus init() is the kernel phase of Process 1 and /sbin/init is an executable program that continues as Process 1 and spawns all other processes.
The kernel image has boot code built-into its front end that can be used for an instant boot diskette of restricted usefulness - it still depends on a healthy root filesystem on the hard disk.
CPU->VGA->Power-On-Self-Test->SCSI->Boot Manager->Lilo boot loader->kernel->init->bash
This is really useful. You can have a complete useful command-line oriented Linux system operating in memory to play with or fix your disk-based Linux setup.
It is a lookup table made to look like a directory of files. Each name represents a particular channel (minor number) of a particular block or character device driver (major number).
A powerful, but dangerous copy command that can do raw copying of data in the absence of a filesystem. This command can copy from devices that are not mounted and can read and write to boot sectors, including the master boot record. Be very careful when using it.
A filesystem is a way of organizing computer data that hides the device-dependent physical storage details (head, track, sectors, integrity checks, ...) from the user, and presents what really matters -- human-readable names for chunks of data arranged in a hierarchy of directories. The Linux Virtual Filesystem (VFS) takes this one step farther, by hiding (as much as possible) differences between filesystems. Hiding implementation details is an important general principle of programming technique.
A loop filesystem is a filesystem that is written onto a file, not a device. The file can then be mounted and used like an ordinary filesystem, filling it with files and subdirectories. You can probably even mount an additional layer of loop filesystems within it... I guess that's why it's called a loop filesystem. It's kind of is circular - a file with a filesystem that contains files. The loop filesystem trick was useful for making our compressed root filesystem, but is also used for pre-packaging files in iso-9660 file format for storage on CD-ROM drives.
The loop filesystem has no connection with the loopback address that allows a network driver to talk to itself.
The "map installer" sets up the boot code and tells it the physical locations of the data it needs. The "boot loader" code makes the boot happen. Some entries in lilo.conf are directives for the "map installer" to use when installing and other directives are for the "boot loader" and loaded kernel image to be used during the actual boot. The actual arrangement of the systems components might be different at those two different times.
Understanding what Lilo does makes it easier to understand the purpose of the different entries in lilo.conf.
When any boot-related files have been changed in any way, you need to re-create the map files by running /sbin/lilo. When in doubt, run it!
Linux has man pages, kernel documentation in linux/Documentation, lots of books, and the ultimate answer book, its source code. You may feel overwhelmed at first, but it's worth the effort; gradually you'll begin to feel the power of the Source! References specific for this talk are here.
Content last revised July 14, 2006
Last updated: April 8, 2006
Contact Craig Van Degrift if you have problems or questions with this web site.