File Management

COMP124 Lectures 2021-03-16

Disk Hardware

Disks are based around the concept of blocks:

• Disks are split into block of a certain size.
• Files fill up one or more blocks.
• Only one file can fill each block, so some wasted space if a file is smaller than the block.

Operating system determines the block size and imposes programmatic access to physical devices.

File System

• Logical vs. physical organisation of file.

• Most file systems are tree-structured:

    graph TD
    / --- etc
    / --- bin
    / --- usr
    / --- home
    usr --- ub[bin]
    home --- staff
    home --- students
    students --- folder1
    students --- folder2
    folder1 --- file1
    folder1 --- file2
    folder1 --- file3
  

  The logical organisation is not related to the physical structure.

Directories

Non-leaf nodes are directories:

• Contain list of filenames plus further into about each file.

Unix Inodes

Directory entry consists of filename and inode number:

• Inode number references (points to) an inode file descriptor.

An inode contains info such as:

• File owner
• Permissions
• Modify and Access Times
• Size
• Type
  • Regular
  • Special (such as /dev and links)
• Location on Disk

File Types

Windows uses file extensions to indicate application.

On Unix you can execute any file:

• Exec will look for magic number at head of a valid executable binary file.

• If the number is not present, exec looks for #! followed by a name of a program to execute:

    #!/bin/bash
  

• Otherwise it assumes it is a shell script and creates an instance of a user’s shell to process it.

File-store Allocation

File-store is divided into fixed-size blocks.

Blocks not currently allocated to files are maintained in a free list.

Several strategies for allocating blocks:

• Contiguous
• Linked
• File allocation table (FAT)
• Indexed (NTFS)

The Free List

This represents the free block on a block device.

It can be a simple bit vector. Bit vectors are simple and efficient if kept in memory. There can be issues though:

• Needs writing to the disk every so often.
• May be huge:
  • 80GB disk with 512 byte blocks needs 20MB for a free list.

An alternative is to chain all free blocks into a linked list.

These methods use the disk to store information about the allocation of the disk.