1. Silberschatz, Galvin and Gagne  2002 12.1 Operating System Concepts Chapter 12: File System Implementation File System Structure File System Implementation Directory Implementation Allocation Methods Free-Space Management Efficiency and Performance Recovery Log-Structured File Systems

2. Silberschatz, Galvin and Gagne  2002 12.2 Operating System Concepts File-System Structure File structure  Logical storage unit  Collection of related information File system resides on secondary storage (disks) File system organized into layerslayers File control block storage structure with of information about a file

3. Silberschatz, Galvin and Gagne  2002 12.3 Operating System Concepts Layered File System

4. Silberschatz, Galvin and Gagne  2002 12.4 Operating System Concepts A Typical File Control Block

5. Silberschatz, Galvin and Gagne  2002 12.5 Operating System Concepts File System Access example

6. Silberschatz, Galvin and Gagne  2002 12.6 Operating System Concepts Virtual File Systems Virtual File Systems (VFS) provide an object- oriented way of implementing file systems VFS allows the same system call interface (the API) to be used for different types of file systems The API is to the VFS interface, rather than any specific type of file system

7. Silberschatz, Galvin and Gagne  2002 12.7 Operating System Concepts Schematic View of Virtual File System

8. Silberschatz, Galvin and Gagne  2002 12.8 Operating System Concepts Directory Implementation Linear list of file names with pointer to the data blocks  simple to program  time-consuming to execute Hash Table  linear list with hash data structure  decreases directory search time  issues:  collisions: several file names may hash to same location  fixed size

9. Silberschatz, Galvin and Gagne  2002 12.9 Operating System Concepts Allocation Methods how disk blocks are allocated for files:  Contiguous allocation  Linked allocation  Indexed allocation

10. Silberschatz, Galvin and Gagne  2002 12.10 Operating System Concepts Contiguous Allocation Each file occupies set of contiguous blocks on disk simple  only starting location (block #) and length (number of blocks) are required random access external fragmentation files cannot grow

11. Silberschatz, Galvin and Gagne  2002 12.11 Operating System Concepts Contiguous Allocation of Disk Space

12. Silberschatz, Galvin and Gagne  2002 12.12 Operating System Concepts Extent-Based Systems modified contiguous allocation disk blocks allocated in contiguous extents  An extent is a contiguous block of disks  Extents are allocated for file allocation  A file consists of one or more extents

13. Silberschatz, Galvin and Gagne  2002 12.13 Operating System Concepts Linked Allocation Each file is a linked list of disk blocks: blocks may be scattered anywhere on the disk each block contains index of next block Simple – need only starting address Free-space management system – no waste of space No random access Logical block size reduced

14. Silberschatz, Galvin and Gagne  2002 12.14 Operating System Concepts Linked Allocation

15. Silberschatz, Galvin and Gagne  2002 12.15 Operating System Concepts File-Allocation Table  linked-allocation scheme used in MSDOS  cached in memory

16. Silberschatz, Galvin and Gagne  2002 12.16 Operating System Concepts Indexed Allocation Brings all pointers together into the index block

17. Silberschatz, Galvin and Gagne  2002 12.17 Operating System Concepts Indexed Allocation Needs extra space for index block  small files: large overhead  large files: multiple index blocks Supports random access No external fragmentation

18. Silberschatz, Galvin and Gagne  2002 12.18 Operating System Concepts Multi-level Indexed Allocation  outer-index index table file

19. Silberschatz, Galvin and Gagne  2002 12.19 Operating System Concepts Combined Scheme: UNIX (4KB blocks)

20. Silberschatz, Galvin and Gagne  2002 12.20 Operating System Concepts Free-Space Management Bit vector Linked list Grouping Counting

21. Silberschatz, Galvin and Gagne  2002 12.21 Operating System Concepts Bit Vector … 012n-1 bit[i] =  0  block[i] free 1  block[i] occupied Maintain bit vector reflecting block usage

22. Silberschatz, Galvin and Gagne  2002 12.22 Operating System Concepts Linked Free Space List on Disk

23. Silberschatz, Galvin and Gagne  2002 12.23 Operating System Concepts Free Space Issues Bit map requires extra space: block size = 2 12 bytes disk size = 2 30 bytes (1 gigabyte) n = 2 30 /2 12 = 2 18 bits (or 32K bytes)  Easy to get contiguous files Linked list (free list)  Cannot get contiguous space easily  No waste of space Grouping  index block of free blocks Counting  store address of free block plus count free contiguous blocks

24. Silberschatz, Galvin and Gagne  2002 12.24 Operating System Concepts Efficiency and Performance Efficiency depends on:  disk allocation and directory algorithms  types of data kept in file’s directory entry Performance  disk cache – separate section of main memory for frequently used blocks  free-behind and read-ahead – techniques to optimize sequential access  improve PC performance by dedicating section of memory as virtual disk, or RAM disk

25. Silberschatz, Galvin and Gagne  2002 12.25 Operating System Concepts Various Disk-Caching Locations

26. Silberschatz, Galvin and Gagne  2002 12.26 Operating System Concepts Double Buffering I. memory-mapped I/O uses page cache II. routine I/O via file system uses the disk buffer cache

27. Silberschatz, Galvin and Gagne  2002 12.27 Operating System Concepts Double Buffering better: unified buffer  uses page cache for both memory-mapped and routine file system I/O  fast: hardware supported

28. Silberschatz, Galvin and Gagne  2002 12.28 Operating System Concepts Recovery Consistency checking  compares data in directory structure with data blocks on disk: tries to fix inconsistencies back up  copy disk data to another storage device (floppy disk, magnetic tape)  Recover lost file or disk by restoring data from backup

29. Silberschatz, Galvin and Gagne  2002 12.29 Operating System Concepts Log-Structured File Systems also: journaling file system record each update to the file system as a transaction all transactions are written to a log transaction is considered committed once it is written to the log transactions in log are asynchronously written to the file system when file system is modified: transaction is removed from log after system crash: all remaining transactions in log must still be performed