Lecture Outline

• To discuss functions performed by the Filing System

• File - Logical representation of data, program, report, etc.

• Filing System has 3 main functions:
  • Facilities for file manipulation and long-term storage of files
    
  • to provide secondary storage managment (covered in earlier lectures)
    
  • to provide support for system integrity
    

• Users should be able to manipulate files using a symbolic name (Virtual Machine)

• Files should be protected against accidental or malicious misuse

• Files should be protected against hardware errors and malfunctions

• System should be responsible for managing space allocation and deallocation

Manipulating Files

• What sort of things might a user want to do to files?

• What sort of things might a user want to do to the contents of files?

Hierarchical Organisation

• Useful organisation because system facilities can be placed in one directory and write protected, while all users on the system can each have a directory (Home Directory, Account, etc.)

• Users need to be able to create / remove directories, rename them, copy / move files between them, see a directory listing, etc.

Referencing Files

• Absolute Pathname, e.g., edit /usr/chris/file1

• Relative Pathname, e.g., if current working directory is /usr/chris, then can refer to file1 as in edit file1.

• Filenames need to be unique only in a single directory.

• Identical Directory Names can be used as long as their absolute pathnames differ.

Keeping track of files.

• File System uses File Directories

• Root directory has a Master File Directory (MFD) which is available to the File System at system start-up.

• MFD points to User File Directories and Files.

• Each UFD points to the UFDs of any subdirectories or contains pointers to physical start addresses of files.

• UFD also contains additional information about each entry - type (text, dir, executable, etc) and access information (RO, RW, EO, RX, RWX, etc)

• Organisation of Secondary Storage can be different from the logical organisation of files which it stores.

• SS is typically organised into fixed-sized blocks which are a multiple of 512 Bytes.

• Each block can contain one or more whole logical records. Records cannot extend over blocks.

• Sequential - records are retrieved in their physical order on SS. New records are appended to the file.

• Direct - Records are accessed randomly using their phyical address. Physical ordering of records is not important (to the user!)

• Indexed Sequential - records are ordered logically according to some logical record key. They can be retrieved sequentially, or directly through the key (via an index table).

Next Lecture ...