POSIX file system basics

POSIX is the Portable Operating System Interface standard, IEEE Std 1003.1-1988 and related.  These standards, based on the Unix operating system, define a set of programming and command interfaces.  Programs and scripts following these standards are supposed to be easily portable between operating system platforms providing these interfaces.

The POSIX standards imply a model for file system organization: POSIX file systems are organized as directories (folders) containing files (documents).  The POSIX programming API defined a number of operations to work with files, directories, and entire file systems:

• mount, umount the file system
• open, close file descriptor
• write, read to/from file descriptor
• mkdir, rmdir, creat, unlink, link, symlink
• fcntl (byte range locks, etc.)
• stat, utimes, chmod, chown, chgrp
• Path names are case sensitive, components are separated with forward slash (/).

Essentially, this is the API that evolved for Unix and adopted by Linux.

POSIX file systems are treated as a sequence of bytes.  However, internally the data content of a file is stored as logical sequence of file system blocks:

• Each block is a fixed number of bytes.  The last block might not be full.
• Within a block, all the bytes are sequential.  However, within a file, the blocks might not reside sequentially on the disk.
• Underlying storage systems are usually organized as blocks, and ideally file system blocks are aligned with the storage system’s blocks.

File systems also contain metadata, i.e., “data about the data”:

• There is an inode for each file or directory, providing:
  • Locations of the data blocks for the file
  • Attributes about the file, like “time last accessed” or “owner of the file”
• The inode table records where each inode is located, indexed by number.
• Directories are special files listing the names and inode numbers of files under the folder.